- Management Summary
- Research Design & Time Line
- Environment & Native American Culture
- GIS Design
- Archaeological Database
- Archaeological & Environmental Variables
- Model Development & Evaluation
- Model Results & Interpretation
- Project Applications
- Model Enhancements
- Model Implementation
- Landscape Suitability Models
- Summary & Recommendations
- Archaeological Predictive Modeling: An Overview
- GIS Standards & Procedures
- Archaeology Field Survey Standards, Procedures & Rationale
- Archaeology Field Survey Results
- Geomorphology Survey Profiles, Sections, & Lists
- Building a Macrophysical Climate Model for the State of Minnesota
- Correspondence of Support for Mn/Model
- List of Figures
- List of Tables
Landscape Suitability Models for Geologically Buried Precontact Cultural Resources
By Curtis M. Hudak and Edwin R. Hajic, with contributions by Patricia A. Trocki and Rose A. Kluth
12 Table of Contents
12.1 General Introduction
12.2 General Methods
12.2.1 Summary of Methodology
12.2.2 Archival Research
12.2.3 Field Program
12.2.4 Core Descriptions
12.2.5 Geologic Cross-Sections and Long-Sections
12.2.7 Radiocarbon Chronology
12.2.8 Geomorphological Quality Assurance
12.3 Minnesota River Model
12.3.2 Overview of Past Work
12.3.3 Specific Methods and Data
12.3.4 Landform-Sediment Assemblages (LfSA)
12.3.5 Landscape Suitability Rankings
12.4 Mississippi River Model
12.4.2 Overview of Past Work
12.4.3 Specific Methods and Data
12.4.4 Landform-Sediment Assemblages (LfSA)
12.4.5 Landscape Suitability Rankings
12.5 Rainy River Model
12.5.2 Overview of Past Work
12.5.3 Specific Methods and Data
12.5.4 Landform-Sediment Assemblages (LfSA)
12.5.5 Landscape Suitability Rankings
12.6 Red Lake Bog Model
12.6.2 Overview of Past Work
12.6.3 Specific Methods and Data
12.6.4 Landform-Sediment Assemblages (LfSA)
12.6.5 Landscape Suitability Rankings
12.7 Red River of the North Model
12.7.2 Overview of Past Work
12.7.3 Specific Methods and Data
12.7.4 Landform-Sediment Assemblages (LfSA)
12.7.5 Landscape Suitability Rankings
12.8 Rock River Model
12.8.2 Overview of Past Work
12.8.3 Specific Methods and Data
12.8.4 Landform-Sediment Assemblages (LfSA)
12.8.5 Landscape Suitability Rankings
12.9 Root River Model
12.9.2 Overview of Past Work
12.9.3 Specific Methods and Data
12.9.4 Landform-Sediment Assemblages (LfSA)
12.9.5 Landscape Suitability Rankings
12.10 St. Croix River Model
12.10.2 Overview of Past Work
12.10.3 Specific Methods and Data
12.10.4 Landform-Sediment Assemblages (LfSA)
12.10.5 Landscape Suitability Rankings
12.11 Anoka Uplands
12.11.2 Overview of Past Work
12.11.3 Mapping Methods and Data
12.11.4 Landform-Sediment Assemblages (LfSA)
12.11.5 Landscape Suitability Rankings
12.12 Bemidji Uplands Model
12.12.2 Overview of Past Work
12.12.3 Mapping Methods and Data
12.12.4 Landform-Sediment Assemblages (LfSA)
12.12.5 Landscape Suitability Rankings
12.13 Glacial Lake Agassiz Basin Model
12.13.2 Overview of Past Work
12.13.3 Mapping Methods and Data
12.13.4 Landform-Sediment Assemblages (LfSA)
12.13.5 Landscape Suitability Rankings
12.14 Lake Benton Uplands Model
12.14.2 Overview of Past Work
12.14.3 Mapping Methods and Data
12.14.4 Landform-Sediment Assemblages (LfSA)
12.14.5 Landscape Suitability Rankings
12.15 Mountain Lake Uplands Model
12.15.2 Overview of Past Work
12.15.3 Mapping Methods and Data
12.15.4 Landform-Sediment Assemblages (LfSA)
12.15.5 Landscape Suitability Rankings
12.16 Nicollet Uplands
12.16.2 Overview of Past Work
12.16.3 Mapping Methods and Data
12.16.4 Landform-Sediment Assemblages (LfSA)
12.16.5 Landscape Suitability Rankings
12.17 Testing the Geomorphic Models
12.17.2 Archaeology Site Location Test
12.17.3 Archaeology Temporal Test
12.17.4 Ongoing Tests
12.18 Discussion on Radiocarbon Chronology and Linking the Study Areas
12.18.1 Radiocarbon Dates from the Valleys and Bog
12.18.2 Linking the Valley Project Areas
12.18.3 Comparing the Valley Project Areas
12.19 Research Questions for the Future
12.20 General Summary
Geomorphology is a geologic discipline that studies the form and evolution of both ancient and present-day landscapes. Fundamentally, cultural resources are part of the geologic record, as in the case of surface and buried prehistoric and historic resources. People who have inhabited the earth utilized specific areas based on essential needs such as food, water, shelter, tools, and safety. If people could not grow or harvest it on the landscape, they had to "mine" it from the earth. The geology and landscape have also influenced the spiritual aspects of cultures. Understanding the geology and geomorphology of an area is therefore critical to understanding cultural resource patterns and trends predicted by any model.
Eight areas were initially selected to analyze and interpret for the potentials of geologically buried suitable habitats or environments that could contain prehistoric cultural resources. These eight areas include seven river valleys and one ancient lake bed, now a bog, and were selected for mapping because of one or more of the following reasons:
- They will be impacted by significant road or bridge construction during the next five to ten years;
- They are their own major physiographic province or are a significant feature of a major physiographic province in Minnesota;
- They appeared to have a relatively dynamic geologic history represented by the deposits (late Pleistocene and Holocene) within their physiographic province(s); and/or
- They have had little systematic archaeological/geological work conducted to adequately characterize the environment.
The goal for the Mn/Model geomorphology project was to develop a working landform sediment assemblage model for each of the fourteen project areas. An additional goal was to establish probabilities for geologically buried "suitable habitats" represented by deposits beneath recognizable landforms. The probabilities were interpreted after an assessment of the stratum’s geologic age, depositional environment, and post-depositional alterations. The latter two factors contribute to archaeological site integrity.
An archaeological property must meet four criteria to be eligible to the National Register of Historic Places. Along with these four criteria the cultural resources must also be in situ or, in the case of an archaeological site, the artifact assemblage must be in its original environmental context (i.e., depositional environment) and not severely disturbed by post-depositional processes. Archaeologists must consider whether they are examining redeposited cultural material. Archaeological properties that have been removed from their original locations by, for example, colluvial or alluvial processes are rarely ever considered significant. An essential component of Mn/Model’s geomorphology, therefore, is to identify locations where natural processes have reduced the probability of an intact site.
The geomorphic models also help to determine the third (vertical) and fourth (temporal) dimensions of the 14 study areas. Mn/Model’s early work will guide more large-scale (i.e., higher resolution) investigations by archaeologists and geologists in each of these 14 areas. These larger scale cultural resource projects will therefore provide more detailed information about the Mn/Model landforms and underlying sediments (Landforms Sediment Assemblages), and also test these smaller-scaled assemblages. As more data are collected, the geomorphic models, resolution, and hence archaeological predictability will be improved.
The methods discussed below have proven successful across the United States for systematic geologic surveys. These methods are a combination of two types of studies commonly used across the United States for geologic mapping. One end of the mapping spectrum is to map from existing topographic maps, soil maps, and aerial photography, without any ground truth, whereas the other end of the spectrum is to construct detailed geologic cross-sections by coring at regular intervals across the entire project area. The former method is relatively inexpensive and the least reliable; the latter method is relatively reliable and the most costly. Mn/Model’s methods fall into the middle of the spectrum because the geomorphic mapping product for the first eight areas was in part extrapolated from a moderate amount of ground truth (i.e., field work). The mapping of the six upland areas was accomplished with little geologic ground truth information, although these areas have a relative abundance of archaeological sites in and around them.
The geomorphic team included many field and laboratory members who worked on various tasks of the Landscape Suitability models. The numbers of staff meant that both standard procedures and a Quality Assurance program were needed for difficult and/or tedious large tasks. The basic tasks were documented to help remind the staff of the project’s preferred methods. Field crew chiefs and mappers (often the same staff members) did not have extensive knowledge of, nor did they review the literature pertaining to, each of the specific project areas prior to the completion of the mapping. Some literature was, however, reviewed to adopt select terminology. This blind approach was intentionally adopted to generate new hypotheses about the project areas’ geologic ages and landscape evolution. Some archival research was conducted prior to each field project to make better decisions on placement of core locations and to plan resources for coring. This research included a review of geotechnical data and aerial photographs.
Cores were described as soon as possible after the field work was completed in a project area. The mappers were occasionally the drilling crew chiefs, who spent multiple days to weeks in the field examining the terrain and outcrops, depending on the project area size. Cores were described and graphically displayed according to sedimentary textures and soil profiles. The graphic logs were created on a CADD system and displayed on a cross-section according to the field workers elevation notes. The cross-sectional data and soil series data from the NRCS (when available) were used to interpret the subsurface sedimentology and stratigraphy. Organic samples from the cores yielded radiocarbon dates that were eventually incorporated into the mapping units and Landscape Suitability models.
The mappers relied mostly on high altitude color infrared aerial photography, NRCS soil series maps, and Mn/Model’s geological field work and cross-sections to construct basic mapping units. These mapping units link the surface characteristics (e.g., landforms, tonal contrasts, elevation, etc.) with the subsurface characteristics (e.g., NRCS soils and Mn/Model cross-sections) to form Landform Sediment Assemblages.
The Landform Sediment Assemblage maps were delineated on USGS 7.5' topographic maps and sent to the GIS team for digitizing with descriptive codes attached (see Section 12.2.6). The mappers reviewed their own maps for the initial Quality Control. The geomorphology models are incorporated into GIS layers for each of the 14 study areas. These inductive models have assigned high, medium, low, and nil landscape suitability rankings for geologically buried cultural resources. Each model weighs the likelihood of the strata containing suitable habitats capable of preserving cultural resources because of the geological age and both the depositional and post-depositional environments. The Landscape Suitability models are expert system models based upon reproducible data documented by the Mn/Model field sampling and some pre-Mn/Model geotechnical boring data (on file at the Minnesota Department of Transportation). These Landscape Suitability models are intended to be used in tandem with the statistical archaeological predictive models previously discussed. One does not supersede the other, but should provide a check and balance. Furthermore, the Landscape Suitability models are designed to address the potential for deeply buried sites, which are not addressed by the archaeological predictive models.
Archival research included discussions with scientists working in Minnesota, and also the soil mapping staff at selected county Natural Resource and Conservation Service (NRCS) offices of each of the study areas. These discussions provided information on outcrops that could yield datable materials indicating the absolute or relative ages of the landforms in question and their underlying sediments. Geotechnical boring logs collected by MnDOT were also reviewed. These MnDOT data offer glimpses of deeper strata than the Mn/Model field work and provided data to construct both valley cross-sections and long-sections. An aerial photographic search was conducted for the basins in question to determine if relative geologic ages of landforms or additional landforms may be recognizable. These data and interpretations were used to further define the areas of geological testing. Some of the Minnesota Department of Natural Resources (DNR) surficial mapping projects (for aggregate inventories, peat inventories, and ecosystem management) were consulted; however, these mapping projects relied mostly upon existing well and boring logs, which typically provide poor details (e.g., radiocarbon ages, soil descriptions, etc.) on the younger unconsolidated sediments that could contain cultural resources. Archival work also included a search for publications and gray literature on studies that may have pertinent data on the geologic history of the study areas including, but not limited to, radiocarbon dating. The archival work was not exhaustive given the volumes of difficult-to-obtain "gray" literature.
Field work included describing outcrops, collecting cores, and field-checking geomorphic maps (Section 12.2.6). Continuous cores were collected along upland to lowland transects (when feasible) within the lowland project areas. Borings allowed for the chance to recover radiocarbon-datable materials in areas that have no outcrops.
The field work included collecting core samples from beneath each type of landform when possible. Selected samples were chosen from the cores for radiocarbon analyses and blowsand analyses. The soils were described according to United States Department of Agriculture methods (Section 12.2.4).
Three-hundred sixty-nine cores were collected for Mn/Model from both public and private properties. The MnDOT Foundations Department drilling crews were used in areas where and when the collection of borings required a tracked vehicle (i.e., Red Lake Bog, and the Red and Mississippi rivers). Most of the coring/drilling was accomplished with a Giddings soil-probe. Core samples were collected in 1.55 m (5-foot) long core barrels, wrapped in plastic wrap and foil, and boxed in either 0.78 m (2.5-foot) or 0.62 m (2.0-foot) long core boxes. The core diameters varied between 6.6 and 7.9 cm (2.5-3.0 inches). Drillers logged the recovered depths versus actual depths of each boring in field notebooks.
Boring elevations and locations were supposed to be determined by a geographic positioning system (GPS) within a +/- 1 m margin of error for x-y coordinates. However, statewide coverage of an FM radio frequency was not yet available for GPS units, and satellite coverage was difficult if not impossible in the deeper valleys (especially with tree-covered valley walls and bluff tops). Most data were therefore tape-measured from a feature that was recognizable on the USGS 7.5 minute topographic maps. Relative elevation differences were calculated between the ground surfaces at each core location. Standard surveying methods were not practical for the vegetated and rugged terrain of the valleys, which comprised the majority of the project areas.
Each profile was described by a geologist with experience in soil descriptions using USDA terminology. These descriptions were reviewed for quality control by one of two senior geomorphologists on the team. Core segments from each core were unwrapped from their foil and plastic wrap sheaths and laid out lengthwise in proper down-hole order. Core segments were split longitudinally by inserting a trowel edge slightly into the core and twisting the trowel to "pop" open a core segment. This way, the largely natural breakage exposes undisturbed soil and sedimentary structures for description.
The core was then described in its moist state using primarily standard pedologic and sedimentologic techniques and terminology (Soil Survey Staff 1994; Hallberg et al. 1978). The core was initially divided into soil horizons and, beneath the solum, weathering zones. Weathering zones are essentially extensions of the soil profile well below the soil profile as traditionally recognized by the USDA for agricultural purposes. Soil horizon, soil color, texture, mottling, soil structure, ped coatings, sedimentary structure and bedding characteristics, moist consistency, effervescence (carbonates), roots and pores, pore coatings, and inclusions such as organic material or shell fragments were noted for each soil horizon on a form developed by the team and designed for the purpose of using standard USDA terminology. Soil samples were also visually examined for particle-sizes and frosted quartz grains to help determine if the landform has been created, buried, disturbed, or destroyed by eolian, fluvial, colluvial, soil-forming, freeze-thaw, or plant/animal processes.
Several modifications to the standard terminology have been added to accommodate multiple buried soils and emphasize their significance. Whereas the Soil Survey Staff (1994) recognizes the use of a "b" for a buried genetic horizon in mineral soils, the Mn/Model project extended its use to organic soils as well. An Arabic number prefix to master horizons is used by the Soil Survey Staff to mark a discontinuity in a soil caused by a significant change in particle size, mineralogy and/or age, unless the difference in age is indicated by the suffix "b" indicating a buried horizon. Even if the material in which a buried soil is developed is the same as that of an overlying soil, the Mn/Model project preferred to highlight the buried soil and package of material it developed in with a different Arabic number prefix. Otherwise, when the Arabic number prefix is used, it denotes a significant change in texture as proposed by the Soil Survey Staff (1994). Where multiple buried soils were present within a single profile, the "b" was given a subscript Arabic number coinciding with the number of the buried soil counting down the profile. The surface soil is not counted, and the first buried soil is noted as "b1," the second as "b2," and so forth. When only one buried soil was present, no subscript Arabic number was used.
A graphic sediment/soil log was simultaneously constructed as the horizons were being described. The graphic log illustrates vertical sedimentological trends, which helped to interpret lithofacies and depositional environments. The sedimentary textures are displayed on a histogram, with the vertical axis of the histogram representing core depth. Each bar width, therefore, represents the depth of each geologic/pedologic unit. The length of each bar (x-axis) represents a specific texture or grouping of textures. Appendix E shows all the geologic cross-sections with the graphic logs. Space-saving measures on these graphics necessitated the grouping of less common textures within certain project areas. The graphic bar lengths are therefore not standardized between project areas, but remain flexible to accommodate varying textures.
Soil horizons, soil texture, horizon color shading, sedimentary structures and bedding characteristics, location and relative abundance of inclusions, and the character of soil horizon boundaries and sediment contacts are represented on the log. Although sedimentological characteristics are noted, the USDA soil texture terminology was used throughout because it is more precise than sedimentological terminology and allows subdivision of the generic ‘muds’ of sedimentologists. Interpretations of stratigraphic units, depositional environments, correlations to other cores, etc., are noted on both the graphic sediment logs on each cross-section and the core description sheets.
The core was allowed to dry in some cases before being re-wrapped. Sometimes drying enhances certain pedogenic and sedimentologic features that are not evident in the moist state. These are duly noted with an indication they were observed in the dry state. All cores have been saved and are being used by the Minnesota Geological Survey in their statewide geomorphic mapping project.
Organic matter in the core that could potentially yield a radiocarbon age was sampled. Sampling techniques varied depending on the kind and amount of material available. In general, individual fragile charcoal fragments were picked out of the core with a small amount of surrounding matrix, sealed in a plastic zip-lock bag, and sent for plant identification. More sturdy charcoal and uncarbonized organic matter were sampled in appropriate length segments of core. The sample was placed in a plastic container with a sodium-hexametaphosphate solution to disperse the surrounding mineral matrix from the organic material. The sample was then wet-sieved through a series of nested standard sieves (generally 2 mm, 1 mm and 0.5 mm), rinsed with distilled water, and allowed to air dry on foil. Sample fractions were then bagged in zip-lock plastic bags and sent for further cleaning and identification. Following description and sampling, core segments were re-wrapped in cellophane plastic and aluminum foil for storage. Many organic soil samples were also dispersed and reviewed under microscope for any macrobotanical fragments that may be taxonomically identified for possible radiocarbon dating.
Cross-sections provide a quick glimpse of the subsurface for the cultural resource manager. Ninety-eight geologic cross-sections were constructed from the geologic logs on a CADD system (Appendix E). These cross-sections graphically represent both the textural data of the logs, and the interpreted stratigraphic units or depositional environments. Radiocarbon data are also presented on the logs when available. The textural key is found in the lower left corner of cross-section figures found in Appendix E. The textural keys are usually standardized for each project area (e.g., Minnesota River Valley); however, they can vary between project areas to better accentuate the local textural variations.
For ease of use, the cross-section and boring log shape files are set up for hotlinking within ArcView. Double-clicking on the cross-section line will pull up the digital version of the cross-section for quick reference. Double-clicking on the core location (point symbol) will pull up the digital core log. The cultural resource manager can then get some "ground truth" to apply to the planning stage of their respective project.
Seven long-sections were compiled for the river valleys to help trace surfaces of similar age. The long-section was constructed by measuring up the centerlines of each valley and not necessarily the river channel. This method was used to avoid the extreme meanders of many of Minnesota’s rivers, and to provide a more standard representation of landform slopes within and eventually between valleys. Landforms were then "hung" on the distance markers made along the valley centerline. Elevations of these landforms were interpreted from the USGS topographic quadrangles.
126.96.36.199 Mapping Background
Geomorphic mapping of Landform Sediment Assemblages (LfSA’s) is a major tool that can be used for evaluating landscape evolution in Minnesota and can serve as a context for predicting the potential locations of geologically buried prehistoric cultural resources. The LfSA’s are the basic mapping unit of the Mn/Model geomorphology project. These LfSA’s are informal map units that recognize that specific landforms, of a given geomorphic position within a given geomorphic region or subregion, tend to be underlain by a sediment sequence of characteristic lithofacies. The correlation or mapping of lithofacies may be aided by both relative and absolute dating techniques. Mapping the LfSA’s provides some factual, on-the-ground basis for assigning Landscape Suitability Rankings to the map units (geomorphic surfaces and underlying deposits) based on ages and depositional/post-depositional environments. Landscape Suitability Rankings can be assigned that range from nil (i.e., too old, too young, too disturbed) to low, moderate, and high. In addition, such mapping is required for Mn/Model because maps of Holocene alluvium at the scale and level of detail required by archaeologists and planners have not been published for Minnesota or any other state.
LfSA mapping in Minnesota was designed to meet the following criteria to the extent possible within the limits of the Mn/Model project.
- Mapping must be high resolution as is practical because of the amount of detail required by the archaeology. This criterion must be balanced against the large-scale (1:24,000) and length of the valleys being mapped, and the need to be able to digitize the LfSA mapping into a working base map within the limits of the Mn/Model project.
- Mapping units must be traceable, consistently mappable, and recognizable in the field with the aid of topographic maps, aerial photography, and soil surveys. Mapping units are therefore based upon descriptive physical characteristics (including relative geomorphic positions) and not on absolute time. Absolute time had to be omitted from the original Mn/Model mapping units because few previous systematic studies have documented absolute time in Minnesota’s valleys. An "absolute time" code has been incorporated near the end of the project set-up because of Mn/Model’s 79 radiocarbon dates, but absolute time was not a factor in determining map unit boundaries.
- Mapping units must be unique, with the least amount of overlap possible.
- Mapping units must be arranged in a flexible and easily modified hierarchy of landscapes and landforms. A hierarchical scheme will allow different levels of detail to be rendered depending on the requirements of a particular user. The Mn/Model mapping is a "first cut" map product intended to be modified and refined periodically by incorporating results of future archaeological and geological investigations. As such, it must be flexible enough to easily accept such refinements, which might include new landform units at any hierarchical level and new descriptors. The mapping must be flexible enough to be applicable to all valleys in the state. Ultimately, geomorphic surfaces within a valley will be related, through correlation or relative landscape position, to geomorphic surfaces of tributary valleys. Not all map units are expected to be present in all valleys, but all map units present in the valleys must fall within the framework of the LfSA mapping hierarchy.
- Mapping units must be as archaeologically relevant as possible. This means documenting where cultural resources will not be found because of both geological age (i.e., too old or too young) and "destructional" environments (e.g., high-energy environments).
- Mapping units initially must be developed independent of known locations and ages of cultural deposits so that known cultural deposits could provide one type of check on the validity and interpretation of map units developed.
188.8.131.52 Code Key and Code Key Structure
A map unit code key was developed for the Mn/Model LfSA mapping to provide a uniform and comparable basis for assigning attributes to map polygons (see Table 12.1). The code key evolved during the construction of the 14 sets of valley, bog, and upland area maps, and consists of 33 numbered codes or fields. Each code consists of a listing of items germane to the specific code number along with respective GIS code symbols or attributes, recommended map symbols, and explanatory comments where necessary. The code is dynamic, and designed to be easily modified as the situation arises. It should be considered a guide and not a rigid template.
The 33 codes are divided into three groups with similar themes (see Section 184.108.40.206). The first group has 13 codes with a theme that describes or interprets landscapes and landforms (i.e., geomorphology). A hierarchical geomorphic classification is incorporated in this part of the code key. The landform is the basic map unit at the 1:24,000 scale. The codes or fields assigned to the landform make it a unique "species" among the map units, especially when linked to its underlying geologic materials. Genetically related landforms are grouped into landscapes. The code key includes sub-landforms to accommodate future detailed studies (i.e., larger scales). The code key includes geomorphic regions and subregions to accommodate the broader geomorphic and physiographic perspective (i.e., smaller scales). These latter two codes were incorporated so Mn/Model could use the Minnesota Department of Natural Resources’ (1998) 1:100,000 scale statewide landform coverage in conjunction with the LfSAs. A name or identifier code is provided for features at all levels in the geomorphology hierarchy, except landform subdivision.
The remaining four geomorphic codes may further differentiate the landform. These include noting stream valley order, landform surface characteristics and modifications (i.e., the presence of different channel patterns, or whether it was modified by different agents subsequent to formation), collapsed features, and various erosional aspects.
The second group of eight codes describe the material and material sequences underlying the map units mentioned in the first group above. Combining this second group with the first group of geomorphology codes mentioned above constitutes a Landform Sediment Assemblage (LfSA). These assemblages then are used in the interpretation of landscape suitability for containing cultural resources. The principal codes in this group note the texture and texture sequence of near-surface material and the thickness of surficial material over bedrock or glacial drift. Codes pertaining to overlying deposits (i.e., loess) and basement material (i.e. that which probably can not contain archaeology) are noted within this group. Two codes indicate the presence or absence of a diamicton texture and buried soils. The formal post-glacial lithostratigraphic unit is provided in one code, although few units are currently named in Minnesota.
The third group of 12 codes describe the relative, geologic, and absolute temporal ages of each Landscape Sediment Assemblage (LsSA) or Landform Sediment Assemblage (LfSA). Three codes are supplied to record chronostratigraphic information on the stage or substage of the LfSA, stage of overlying deposits, and glacial lake or glacial ice phase. Three codes allow for observations on relative ages among geomorphic units within a landform or landform subdivision, and of LfSA’s to other LsSA’s or LfSA’s. The remaining six codes allow geochronologic information, from radiocarbon ages and other means, to be assigned to an LfSA, an LfSA’s overlying deposits, and an LfSA’s basement material.
The code key may be used as a guide to develop map symbols for different map units (polygons). The map unit symbols are related to summary GIS code or attribute strings for each theme (i.e., geomorphology, materials, and temporal), and Landscape Suitability Rankings.
Landforms are the basic mapping unit, and are compatible with mapping at a scale of 1:24,000. Mn/Model landforms are those landscape features typically defined and recognized in modern geomorphology. The landforms may be further subdivided into sublandforms although this subdivision has yet to be utilized in Mn/Model mapping.
Construction of landform maps involves multiple steps for each project area. Geomorphic interpretation is involved in each step. The first step delineates different polygons, or contrasting landforms, on a USGS 7.5' topographic map (1:24,000 scale). A pencil with fine point is used so that lines are prominent enough for the digitizer to follow, yet thin enough so as to not cover too wide an area at the scale of the map. Informal working notes are marked within polygons when necessary.
Actual locations of lines are based on a combination of criteria. Foremost are geomorphic boundaries illustrated on the topographic maps by changes in patterns of contour lines. Where a hillslope of significant length coincides with a geomorphic boundary, that boundary is delimited at the base of the slope. Where hillslopes are lengthy and steep, such as some major river valley walls, they can be delineated and mapped as a landform.
Other considerations in map boundary placement include tonal contrasts on high altitude color infrared aerial photography (1:40,000 scale), materials information interpreted from NRCS soils mapping, and cross-sections interpreted from the Mn/Model geomorphology field and laboratory work (Appendix E.2). Tonal contrasts often indicate limits of alluvial fans, overbank belts, and erosional patterns, as well as different degrees of scroll bar burial by younger overbank belts in valleys. Some geomorphic boundaries evident on aerial photographs are transferred to topographic maps by tracing the geomorphic boundaries onto mylar. The mylar is scaled to 1:24,000 on a photocopy, which is in turn placed on a light table with the topographic map. The geomorphic boundaries are then traced onto the topographic map after registration of section corners and other landmarks.
The soil surveys’ interpreted sedimentary materials are often used for subdividing some of the higher levels in the geomorphology hierarchy. Some soil surveys were assembled into mosaics and scaled on a photocopying machine to match individual 7.5' topographic map sheets. These 1:24,000 photocopies are color coded for properties of interest to visually enhance property patterns for valley reaches, upland areas, or landforms in question. Lines of interest are transferred to topographic maps by tracing the lines of the colored soil map over a light table. Black and white aerial photographs often serve as the base for the NRCS county soil maps. On occasion they aid in differentiating photographic tonal contrasts as on the aerial photographs. The quality of the soil surveys can vary greatly depending on date of issue, who did the soil mapping, the quality of the base map, and other factors; therefore, soil surveys are not greatly relied upon for their genetic interpretations or for their exact placement of soil series boundaries.
Valley walls generally mark the limits of landform mapping for the valley project areas. Mapping in valleys that were created by catastrophic flooding, such as the Minnesota and St. Croix valleys, does not extend to the highest glaciofluvial or flood outburst surfaces because this would involve substantial mapping in upland areas. Mapping extends at least part way up most tributary valleys. Mapping in smaller tributary valleys is generally complete although potential large-scale map units are usually lumped together into one unit. Mapping in larger tributary valleys is arbitrarily terminated (often at section or half-section lines) at a distance that does not interfere with reasonable presentation and map production of an entire river valley. Smaller tributary valleys are generally mapped for their entire length.
For the larger and more complex main valleys, a second mapping step involves construction of a longitudinal valley section to trace and correlate terrace levels down the lengths of these valleys. This allows further geomorphic distinction between different terrace levels, particularly in the valley terrace and flood basin landforms. Longitudinal profiles are constructed by first estimating elevations of terrace levels and polygons in question, drawing a line down the center of the valley, creating a scale along the center line, determining where terrace polygons fall along the center line by constructing perpendiculars from the center line to upstream and downstream ends of terrace polygons, and scaling and plotting those surfaces along the valley center line in profile.
A third mapping step assigns a map symbol to polygons on the maps. Informal working notes may be erased or noted elsewhere for text descriptions.
The fourth mapping step is one of review and checking. Maps are systematically examined from the upstream to downstream end of the valley and north to south for uplands. They are checked for continuity and validity in geomorphic interpretation, polygon closure, polygon continuity between adjacent maps, consistency in map symbol labeling between adjacent maps, and consistency in map symbol labeling within the working models of landscape evolution. Visual field checking of the landforms occurred both before and during this step when possible. Many instances occurred where either access was denied or the visibility was obscured. Modifications and corrections, where appropriate or necessary, are made at this time. A photocopy of each map is made for backup purposes before turning maps over for digitizing.
The fifth mapping step is digitizing and labeling the original pencil-lined maps within the GIS software package. The digitizers highlighted any unclosed and unlabeled polygons as part of the Quality Control program. These digitized versions were reviewed by the Mn/Model geologists on both the ArcView software and on color plots. Mislabeled or oddly-shaped polygons relative to the original pencil-lined maps were identified by the geologists. Furthermore, lists of the polygons in the GIS database were printed out for verification against the geologists’ master lists of polygon map symbols. Corrections were noted on black-and-white plots and sent back to the digitizers with the original pencil-lined maps.
A set of short-hand notations were used for map symbols on the original pencil-lined polygons. These map symbols were digitized into the database. These short-hand notations usually sufficed rather than writing out the more complete description. A more complete short-hand description was globally substituted later for the original short-hand labels used for each polygon.
Landscapes consist of groups of genetically related landforms. Differentiation of landscapes is based on assemblages and patterns of landforms characteristic of different geologic agents (i.e. glacial ice, rivers, wind). The code key includes the following landscapes: Uplands (undifferentiated), Active Ice, Stagnant Ice, Ice Contact, Pediment, Glaciofluvial, Catastrophic Flood, Glaciolacustrine, Paleo-Valley, Peatland, Valley Terrace, Floodplain, Valley Margin, Eolian, and Lacustrine.
The Upland (undifferentiated) landscape consists of a single undivided upland unit. The Upland, undifferentiated, landscape initially was required because other mappable valley landscapes occasionally surrounded an upland remnant, and sometimes the uplands were important to the interpretations of other landscapes or landforms. The Upland (undifferentiated) landscape can and should be divided into other specific landscape units.
The Active Ice landscape consists of groups of landforms typically associated with active glacial ice. End and ground moraines formed by active ice usually lack some of the relief and features of moraines formed by stagnant ice. Underlying till is typically more uniform.
The Stagnant Ice landscape consists of groups of landforms typically associated with the dynamic supraglacial environment on stagnating ice. Typical landforms in this landscape include kames, ice-walled lake beds and other hummocks, and linked depressions, kettles and other depressions. The supraglacial drift complex underlying these landforms has textures that vary widely.
The Ice Contact landscape consists of groups of landforms that show evidence of having collapsed as the result of melting of glacial ice. Some subjectivity is used in this landscape because of scale issues and possible overlap with the Stagnant Ice landscape.
The Pediment landscape consists of erosion surface landforms on bedrock or glacial drift, with a thin mantle of colluvium or alluvium. Pediments may occur on, and therefore are distinct from, the valley margin landscape (alluvial fans and colluvial slopes).
The Glaciofluvial landscape includes outwash terraces and plains, some ice disintegration features, and some ice-contact features (i.e., kame terraces). This landscape is required because many of the study areas were created or impacted by glacial outwash events. Most river valleys are incised into or flow upon ancient outwash plains (sandurs) or outwash terraces.
The Glaciolacustrine landscape is differentiated from other glacial landscapes because some "valleys" are incised into or flow upon ancient glaciolacustrine plains. These plains have their own sets of geomorphic features, of which only some were defined for the Mn/Model project. The landscape is reserved for the exposed lake beds of extensive glacial lakes; relatively small basins, such as kettles, are not included in this landscape.
The Paleo-Valley landscape is a "catch-all" for ancient valleys that are currently occupied by wetlands or misfit streams and rivers. Paleo-Valleys may be ancient cut-offs or glaciofluvial drainageways. When distinctions were clear, a Paleo-Valley landscape could be labeled as a paleochannel within another landscape such as the Glaciofluvial landscape.
The Peatland landscape was created for the large lateral and presumably thick expanse of peat in some areas. This landscape may be lumped into other landscapes if areas of peat are recognizable and less expansive. The distinction is qualitative between when to use peatland landscape and, for example, glaciolacustrine landscape with a histosol mantle. The peatland landscape can be differentiated into several recognized peatland landforms.
The Valley Terrace landscape consists of terrace landforms that are unrelated to glaciofluvial activity and that rise above the floodplain landscape. Valley terraces are often differentiated by their relative age (i.e., higher being older).
The Floodplain landscape consists of the group of landforms commonly subject to active flooding by a river, stream, or intermittent drainage. The Floodplain landscape commonly includes the following landforms: floodplain, crevasse splay, natural levee, paleochannel, and the active river channel.
The Valley Margin landscape consists of landforms occurring at the foot of valley slopes as the result of redistribution of material from valley slopes and uplands by flowing water and gravity. The most common examples are alluvial fans and colluvial slopes.
The Eolian landscape was created for larger expanses of blowsand and loess. The Eolian landscape could be lumped into other landscapes depending on a qualitative break in lateral expanse.
The Lacustrine landscape consists of landforms related to lakes and riverine lakes that occur within the areas mapped. The Lacustrine landscape may consist of landforms such as lake, reservoir, island, wave-cut platform, terrace, peninsula, isthmus, exposed lake bed, and beach ridge.
The group of landforms that comprise a particular landscape in one of the seven different valleys, one bog, or six upland project areas may differ in another project area. Also, identically coded landforms or landscapes in two or more different project areas do not necessarily correlate. For example, the VT1 (youngest valley terrace) in the Minnesota Valley does not necessarily correlate with identically labeled landforms in other valleys; however, the code is flexible enough that when correlations among project areas is completed, the models can be adjusted accordingly. The assignment of a landscape to a map unit is not always clear-cut. One map unit could be coded for one or more landscapes, and the final decision on which landscape it was assigned rested with the mappers’ interpretation of either the last depositional environment or the dominant landform developing process. So, for example, an area may be mantled by reasonably thick outwash deposits, while the geomorphology indicates a glaciolacustrine environment was the main reason for the current landscape.
The geomorphic coding concepts have not varied throughout the two year mapping process. The mappers have always recognized that three "themes" would exist for the mapped polygons. These three themes are geomorphology, materials, and temporal assignments. The evolution of the code (attribute) key and Landform Sediment Assemblage (LfSA) mapping symbols have varied only in its organization and presentation of information, and through the addition of more information within each of these three themes. The organizational changes were made to produce a more effective research tool. Searching for specific types of data or data sets has become easier as a result of these modifications.
The first versions of the Mn/Model geomorphology code key were designed to record only the most fundamental information in as compact a manner as possible, because of the mistaken impression that it was difficult to manage and use a GIS database where polygons had numerous attributes. Landforms, material, material thickness, overlying deposits, and basement material were, and remain, the primary basis for delimiting polygons. Relative age relationships of geomorphic units within a particular landform, such as terraces (T1, T2 etc.), also were incorporated into the code key. Absolute ages were not incorporated during any of the mapping because of the dearth of radiometric dates, and the danger of misinterpretations. Mapping was based purely upon the physical characteristics mentioned above. Although landforms were always the fundamental map unit, the earliest versions of the code key were phrased in terms of landform sediment assemblages, foreshadowing the conceptual basis for interpretation of buried cultural deposits. Landform polygons of similar genetic origins were grouped into "suites," something now referred to as "landscapes" (Section 220.127.116.11). Polygons of a particular landform sediment assemblage type were defined as having different combinations of the aforementioned attributes and were listed in the first versions of the code key by their map symbol.
The landform’s underlying material was noted as to whether it was coarse, fine or peat; laterally continuous or discontinuous; and less than or greater than 2 m thick. The code was amended early on to consider material less than 1 m thick, or greater than 1 m and less than 2 m thick, etc. Overlying material is defined as material deposited on a landform without altering the landform itself, and is important to the interpretation of buried cultural deposits. The presence of loess was easy to address. In contrast, the presence of post-Euroamerican settlement alluvium, with obvious implications for archaeological survey and testing, was more problematic because the definition of map units had to be based on geomorphology and material and not absolute or stratigraphic age considerations. This issue was addressed in two ways. First, four different types of floodplains were geomorphically defined based on the degree to which they exhibited channel migration features. The types ranged from point bars through ridge and swale topography to the lack of expression of channel migration features. Although not a firm rule, there is a general increase in age with this morphological progression. Second, and more importantly, two types of overbank deposits were recognized based on the tonal contrasts they exhibited on aerial photography. Post-Euroamerican settlement alluvium typically, but not always, exhibits relatively lighter tones. Later, after radiocarbon dates were brought into the study, and patterns were recognized in the strata of the same LfSA, then temporal codes were developed that allowed a post-Euroamerican settlement age to be assigned to the LfSA or overlying deposits. Underlying material, defined as undifferentiated till or bedrock, was intended to define a "floor" beneath which buried cultural deposits would not occur.
A simple listing of map symbols followed by descriptions was becoming burdensome during mapping. Numbered codes (fields) were introduced to standardize and formalize the code key. Initially, the key contained ten codes (fields): landform sediment assemblage (LfSA) suite, basic landform, age based on geomorphic relations within a LfSA suite or basic landform unit, age based on geomorphic relationships to other LfSA suites or basic landforms, basic landform subdivision, surface characteristics, overlying deposits, texture and texture sequence of near-surface material, thickness of near-surface material over bedrock or glacial drift, and basement material. Each numbered code (fields 1-10) had a list of possible entries (attributes), including "no distinction made." A polygon would be defined by the unique combination of one entry (attribute) under each of the ten numbered codes (fields).
Listings under each numbered code also were assigned both a numeric code, intended to be entered into the GIS database and used for searching, and an alphabetic map symbol. The numbers were abandoned shortly thereafter for more descriptive characters; for example, "10" was substituted for the letter "t" (for terrace), and later for the word "TERRACE". The original basic landform code (field) was unnecessarily complicated by listing landforms (attributes) separately under a suite subheading. So, for example, in the Basic Landform code, "terrace" was listed three times under three different suites, the Valley Terrace suite, Catastrophic Flood suite, and the Glaciofluvial suite. This was driven by the early attempt to implement a numeric GIS code. The mappers recognized that there were other suites that had or could have terraces, and that researchers may have a need to search for all "terraces" regardless of their genetic origins. This paradigm shift favored a simple landform listing under the Landform code or field (Code No. 7). This shift occurred at the same time the shift from "suite" to "landscape" occurred.
The alphabetic map symbol for each entry in the code was maintained. "t" for terrace was used in mapping behind the Landscape identifier (e.g., Ot for glaciofluvial or outwash terrace). Sometime during this process the decision was made to change all lower case letters in the database and on the maps to upper case letters to eliminate the possibility of confusion between the numeral one and the lower case "L." This was in part prompted by the need for the digitizers to more easily recognize the map symbols. Where possible, map symbols were limited to one or two characters. The character map symbols in the code key are used as guides for constructing strings of letters that comprise the map symbols used to identify polygons while mapping. They are also used to create the strings of the three themes, geomorphology, materials, and temporal. These latter strings allow for an easier search by themes. There are no "rules" governing the map symbols use. Closer to the end of the mapping project when most of the possible variations were recognized across the project areas, map symbols were standardized between the mappers and globally changed for each of the delineated polygons. This helped to have a more meaningful map symbol for the digital displays between mapped areas.
Three major changes to the code key occurred during the final phase of the project. First, the three main types of information being recorded (geomorphic, material, and temporal) were formally recognized by splitting the code key into three sections to address each type of information. The tripartite division also facilitated a logical ordering of codes within each division, presumably, to some degree in the order in which they would be considered during mapping. The basis for constructing Mn/Model geomorphic maps fundamentally has always been the configuration of the landscape; however, the history of landscape evolution is particularly important to interpreting the location and potential ages of buried cultural deposits. Earlier versions of the code acknowledged the importance of temporal information by having a couple of temporal codes, but these were never fully developed or utilized. When the project focus shifted from actual mapping to interpreting the geomorphic maps in terms of Landscape Suitability Rankings, it was necessary to introduce a set of temporal codes. The temporal codes (Code Nos. 22 through 33) incorporate relative temporal information (relative age of a geomorphic unit within a landform, such as terrace sequences, and relative age of an LfSA to other LfSA’s or LsSA’s); chronostratigraphic information (stage or substage of LfSA’s and overlying deposits; glacial lake or ice phase); and geochronologic information (greater than and less than radiocarbon ages for LfSA’s, basement material, and overlying deposits).
Second, mapping of "upland" landscapes forced the mappers to consider additional concepts and look differently at some of the more entrenched aspects of the code key that developed while mapping river valleys. A Tributary Valley landscape, a seemingly important distinction when mapping major valleys, was unworkable because of the relativity of the term "tributary." The code had to address future analyses and the unexplored relationship of prehistoric cultural deposits to streams and valleys of different order. Stream Valley Order (Code No. 10) was therefore introduced, and intended to be used with the Floodplain and Valley Terrace landscapes. At the same time, the Tributary Valley landscape was abandoned. The Minnesota Pollution Control Agency’s (MPCA) digital database of stream order was tested in the Lake Benton map area to see if it would be adequate for the needs of the Mn/Model code. The MPCA database tends to overestimate the stream order, and many first and second order streams were omitted from the database; therefore, at this time, stream valley order must be "manually" determined by analysis of drainage basins from maps. The upland areas or areas with lower order valleys are the most practical for this manual charge.
Mapping in upland areas spurred the addition of a number of upland landforms, both glacial and bedrock. The predominance of the glacial landscape, much of which is the result of ice stagnation, caused the addition of Code No. 12 to indicate whether a landscape or landform is collapsed (i.e., buried ice block melting). The mappers are still experimenting with the use of this code because of the different scales of collapsed features they have encountered. Finally, from the perspectives of soil-geomorphology and interpreting the significance of cultural deposits, we view the uplands as potentially highly dynamic, as opposed to the more traditional view of uplands as "stable"; therefore, eroded landscapes and landforms are of potentially great importance, but are not understood. Eroded areas are most often in the uplands, and may include, in addition to plow horizons, the presence of unrecognized erosion surfaces of Holocene age. The mappers wanted to be able to test and explore the unknown relationship of prehistoric cultural deposits to areas effected by moderate to severe soil erosion. The tested relationships could include site location (geographic), site integrity, and the possibility of buried "upland" cultural deposits. Code No. 13, Eroded Landscape or Landform, allows observations regarding soil erosion.
Acknowledging that uplands were and are dynamic contributed to expanding the Texture and Texture Sequence of Near-Surface Material code (Code No. 15). Possible Holocene age surficial sheetwash deposits, erosional lags, and biomantles indicated a need for a more refined system for noting texture and texture sequence. The NRCS soil series textures were incorporated into Code 15 and designed to indicate a vertical sequence of up to three textures. Code No. 16 was introduced to indicate whether a diamicton texture was present in the uppermost lithologic unit or units. Upland-related entries were added to Code No. 18, Overlying Deposits, to acknowledge the presence or likely presence of hillslope colluvium or a biomantle.
The third major change added two levels in the geomorphic hierarchy to the code key above the Landscape level (Code No. 5). These changes were spurred by requests to review the utility of the Minnesota Department of Natural Resources (DNR) Geomorphology GIS database of the state for cultural resource management, and to develop a conversion system to re-code the DNR database to make it compatible with the Mn/Model Geomorphology GIS database. The DNR database is at a 1:100,000 scale and had a mixture of incompatible attributes within most fields (e.g., both landforms and depositional environments in the Sedimentary Association/Rock Association field). Mn/Model Code Nos. 1 and 3 were added for the DNR’s Geomorphic Region and Geomorphic Subregion, respectively. Geomorphic Regions include glacial lobes, glacial lake plains, glaciofluvial valleys, and differing bedrock terrains. Geomorphic Subregions include ground and end moraines, beach levels, eolian dune fields, drumlin fields, outwash plains, and river valleys. Geomorphic regions and subregions were identified in the DNR database by their common geologic name (e.g., Lake Agassiz), rather than by a geomorphic feature (e.g., glacial lake plain). To convey the information from the DNR database, but maintain the geomorphic focus of the mapping, Code Nos. 2 and 4, identifiers for geomorphic regions and subregions, were added to the Mn/Model codes to provide the common geologic names of features. Similarly, Code Nos. 6 and 8 were added, but not used, to provide the opportunity to associate a local geographic name to a particular landscape or landform.
Although the code key is designed for geomorphic mapping and interpretation of landform sediment assemblages, the mappers recognized the detail incorporated in the key would facilitate construction of other types of maps. In particular, useful Quaternary geologic maps and stack unit maps could be constructed if lithostratigraphy was incorporated in the code key; therefore, Code No. 14 was introduced for Post-Glacial Lithostratigraphic Unit, and Code No. 21 was introduced for identifying the underlying basement material. Currently, most lithostratigraphic units of Holocene age in Minnesota are not formally defined.
A "suitable landscape" is defined as a Landform Sediment Assemblage environment that could contain cultural resources, based upon environmental considerations and appropriate age. A "Landscape Suitability Ranking" is the product of an "Age" ranking and "Depositional Environment," or a "Post-Depositional Environment" ranking. The "Age" ranking is a simple toggle switch that indicates that the prehistoric cultural resources may or may not be found within a selected time span. A "zero" ranking means that the geologic age falls outside the selected time span, and a "one" ranking means that the geologic age falls within the selected time span that could contain prehistoric cultural resources. The time span is generally determined to be less than 12,500 B.P. and greater than 200 B.P. after consultation with the many Minnesota archaeologists including the SHPO. The site-specific possibilities for these time spans are, however, discussed in each individual project section below because of the relatively late "retreat" of glacial ice and drainage of glacial lakes in northern Minnesota.
The "Depositional Environment Ranking" is an assessment of the depositional processes and energy, and the drainage conditions. The rankings range from "zero" to "three." A "zero" ranking means that the depositional environment had too high of an energy regime to contain or preserve intact cultural resources, or it was too wet to support cultural activities that would leave a detectable material record. Examples may include channel, glacial, and certain catastrophic flood and lacustrine environments. A "one" ranking means that the environment had a moderately high energy regime, or in some cases that the environment may have been normally wet. Examples include braided stream and wetland environments. A "two" ranking means that the environment had a moderate energy regime, or could be subjected to both high and low energy regimes. Examples may include environments like certain river islands, levees, and lake shorelines. A "three" ranking means that the environment had a low energy regime and was the most likely to contain and preserve cultural resources. Examples may include floodplains, certain lake shorelines, and eolian environments.
The "Post-Depositional Environment" ranking is an assessment of the apparent alterations to the land surface after the primary depositional processes creating that landform have ceased. This ranking is another "toggle" switch meaning that the environment has either been disturbed beyond having any chance for containing in situ cultural resources (equal to "zero" ranking), or there is a chance, even if remote, that the landform could contain in situ resources (equal to "one" ranking). The "zero" ranking would apply to those landforms where the land surface has been severely eroded or excavated because of mining. The "plow zone" could eventually be coded into this ranking with land use maps; however, because thick A-horizons do exist in Minnesota and the plow may not have disturbed the bottom of the topsoil, the mappers have decided not to code for farm fields.
The last piece of the Landscape Suitability model is the unit of depth. Four depths are currently coded. The uppermost unit is the land surface (0 m), which is defined in the purest sense as a two-dimensional geomorphic landform. Cultural resources may be found on this land surface and that is what is coded for under the "Surface" code (equal to zero depth in the tables). The next unit of depth ranges between 0-1 m depth. This depth interval is distinguished because the standard maximum depth at which Cultural Resource Management shovel testing is currently conducted is 1 m. The next unit of depth ranges between 1.0 and 2.0 m, and is distinguished because this is approximately the standard depth at which the NRCS soil series maps have modeled. The last unit of depth ranges between 2.0 and 5.0 m, and was arbitrarily chosen because it represents the average core depth in most of Mn/Model’s geologic project areas. Degrees of confidence in the depositional rankings, and hence Landscape Suitability Rankings, decrease with depth.
Radiocarbon-dating the geological strata of Minnesota is extremely important for the success of the Mn/Model geomorphic mapping. Both absolute (radiocarbon) and relative dating principles have helped to bracket the landforms and potential cultural components in time. This study and other recently unpublished work have indicated that organics for radiocarbon dating are available in the valley fills of Minnesota. A radiocarbon database containing 79 radiocarbon dates has been constructed for the eight original geomorphic study areas of the Mn/Model project.
Two quality control (QC) measures have been established for the radiocarbon dates in the valley and bog fills of the eight "lowland" areas that were field investigated. The first QC measure is identifying the plant matter (e.g., wood, seeds, etc.) to help in the selection of samples for radiocarbon dating. Subaerial plant parts are currently interpreted as having the best chances for a reliable radiocarbon date on their surrounding sedimentary matrix. Hence, these subaerial plant remains were usually selected first for submittals when available. Very small pieces of unidentified wood were usually selected second when available, and roots were usually selected third if an absolute time unit was needed for that stratum. Soil humates were usually selected last when available. One bone sample was submitted for a bone collagen date after removing all other organics except the protein.
The second QC measure is comparing the relatively large numbers of radiocarbon dates against each other. The aberrant dates will be recognized after enough dates have been collected from each of the important lithofacies in different landform sediment assemblages.
The following plan is designed to help eliminate errors and mistakes in the geological field and laboratory work. The plan is based on current nationally or internationally recognized and accepted methods and practices in geology. No plan is a guarantee that all variables will be controlled; however, the quality assurance plan will improve the overall scientific quality and reduce the amount of potential rework.
When practical and feasible, the following methods and practices were employed. They are listed as a "how to" manual and described as such by using future tense.
- Systematic sampling is required across the landscape to sample each of the many apparent and inconspicuous landforms (e.g., multiple and different aged terraces may appear to be a single terrace of one age because they are at the same elevation). Cores and outcrops will be described from upland to lowland in transects to help assure a sampling of the many possible aged land surfaces, their underlying geology, and their lateral relationships.
- Cores will be continuous and fully described from top to bottom.
- Cores will be wrapped in cellophane plastic and aluminum foil. The outside of the wrapped foil will have the upper and lowermost core depth measurements recorded on their respective ends, and the core box will have the range of depths marked on the outside vertical face of the box. Cores will be wrapped, boxed and labeled as soon as possible after extraction from the core barrel.
- Lubricants (except for distilled water) will not be allowed to assist the extraction of core samples from core barrels.
- Core barrels will be cleaned without the use of organic compounds prior to re-use.
- Outcrops will be cleaned of slump. The outer weathering profile created along the cut bank will be scraped until a "fresh" profile is recognized.
- Geological and pedological profiles will be documented by geologists/geomorphologists trained to describe both soils and lithologies.
- Geology descriptions may include, but not be limited to, lithology, texture, color, structure, weathering or post-depositional alterations (including soils), radiocarbon ages, and interpreted depositional environments.
- Soil descriptions will follow USDA methods and may include, but not be limited to, horizons, depth (thickness), color, mottles, structure, consistency, effervescence, lower boundary conditions, and fossils.
- Drill-crews will locate cores by measuring from a feature or features recognized on the local USGS 7.5' topographic map.
- Drill-crews will rotate assignments within each crew and between crews to help avoid "slipping" into non-value added methods.
- Experienced field crew leaders will discuss specific safety precautions regarding certain pieces of field equipment or procedures such as the drill-rigs and drilling, respectively.
- One laptop personal computer will be dedicated to the field work.
- A soils laboratory should be dedicated for the sole purpose to "house" most geological materials. These materials include topographic maps, county soil series maps, files, archival reports, etc.
- Core samples will be stored in the soil laboratory, soil storage facility, or similar soil storage facility. Cores will be stored for at least six months past the acceptance of the final geology report.
- One desktop personal computer (PC) will be dedicated to the geology project. Software will be nationally recognized, with support services available.
- Pedologic and geologic descriptions will be cross-checked by one of two Ph.D. geologists if they did not perform the original documentation. This process will also act to help standardize descriptions between geologists.
- Quality Control will include annual reviews of the geomorphology maps (final product) by a Ph.D. geomorphologist/geologist who is not the original mapper.
- Multiple radiocarbon dates on organic samples from the same stratigraphic and geomorphic context will be analyzed. Multiple analyses on different samples will help to recognize aberrant dates.
- Organic samples will be split prior to sending to the laboratory for analysis. One-half of the sample will be held in reserve as a precaution against loss in transit or contamination. Larger samples will occasionally be split three ways and sent to two different laboratories as a cross-check on the main laboratory. (Finding "large" samples in 2.5-inch diameter cores is admittedly rare.)
- Organic samples will be collected with sterile metal tools (e.g., knives, trowels, etc.) from outcrops or cores. When possible, the organic samples will be collected away from the core perimeter and core-barrel ends to avoid cross-contamination dragged into the core sample. Organic samples will be air dried and stored in aluminum foil.
- Organic material will be qualitatively identified in terms of macrobotanical constituents.
- Discreet organic samples, such as subaerial parts of wood or seeds, will be chosen for radiocarbon analyses before soils or peat, which are rich in disseminated or unknown types carbon. This preference may mean that accelerator mass spectrometer (AMS) analysis of a small organic sample is the preferred method.
Radiocarbon chronology will be cross-checked with available literature and ongoing paleoecological studies in Minnesota and the surrounding states.
The Minnesota River Valley investigation covered nearly its entire length. Mapping extended from the downstream end of Big Stone Lake at Ortonville to the mouth of the Minnesota River at the Mississippi River confluence. The Minnesota River trends southeastward for 225 km from its head at Browns Valley to Mankato where it takes an acute bend to the north. Along this reach the river is cut into the Olivia Till Plain to the north and Blue Earth Till Plain to the south. Both till plains were created by the Des Moines glacial lobe, and the Minnesota River parallels the axis of this ice advance. Downstream from Mankato, the river trends north, then gradually angles to the northeast. Through this reach it initially crosses the aforementioned till plains, then cuts through the Owatonna Moraine before joining the Mississippi River. The entire valley was formed by catastrophic flooding from the Lake Agassiz basin, and the valley is dominated by catastrophic flood landforms. Early phases of catastrophic outbursts also scoured the surrounding uplands and influenced the courses of lower ends of Minnesota tributaries. The uppermost reach of the Minnesota Valley is characterized by several valley lakes that are dammed at their lower end by alluvial fans of large tributaries. Downstream, catastrophic flood waters widened the valley walls locally where igneous and metamorphic intrusions are crossed by the valley. The valley gradually narrows as it approaches Mankato, but after angling north, it is characterized by valley walls being alternately scoured into broad arcs by catastrophic flood waters.
The river that drained Glacial Lake Agassiz was named Glacial River Warren by Upham (1883) after the Chief of Engineers of the U.S. Army, General G.K. Warren, who first recognized that the wide Minnesota Valley must have been cut by a tremendous river draining Glacial Lake Agassiz. Matsch and Wright (1967), Matsch et al. (1972), and Matsch (1983) describe a series of events during the evolution of Glacial Lake Agassiz that influenced the discharge of Glacial River Warren. They envisioned a step-wise incision of the Minnesota River valley occurring over a relatively long span of time. Matsch et al. (1972) interpreted the upland channels, with planar- and cross-stratified coarse sand and cobbly gravel, and erosional residuals as outwash deposited by the retreating Des Moines Lobe. The valley terraces with a boulder lag were interpreted as successively lower channel bottoms of Glacial River Warren. Boulder-gravel beds just above and below modern floodplain level were interpreted as the bed-load of Glacial River Warren.
More recently, Kehew and Lord (1986) developed a geomorphic model of landforms associated with catastrophic flooding. The model concludes that the full range of catastrophic flood landforms, across a wide range of elevations, can develop from a single catastrophic flood. Their model has been applied to the Minnesota Valley for the LfSA mapping.
There is a tremendous void in literature for the post-Glacial River Warren period. What little information there is comes from "gray" literature and an occasional abstract. Several geoarchaeological investigations have yielded Holocene radiocarbon ages and terminal Pleistocene radiocarbon ages from deposits underlying alluvial fans, terraces, and floodplains (Hudak 1994a, 1997; Hajic 1996b; Bower et al. 1996; Mooers et al. 1992; Bettis and Thompson 1986).
Fifty-six cores were collected from the Minnesota River valley between July 31, 1995, and October 12, 1995. All cores were taken with the Giddings hydraulic soil probe. These cores were logged in a laboratory and are presented in Appendix E.1. The methods discussed in Section 12.2 apply to this field and laboratory work.
A series of 17 cross-sections and one long-section were constructed from the soil profiles of the Minnesota River valley. Two more cross-sections showing three different core logs were modified after Hudak (1997) and included in the current report. These figures and one drawing are presented in Appendix E.2 and E.3, respectively.
18.104.22.168 Radiocarbon Dates
A variety of organic samples were collected from the various cores. A master list of organics and their identifications are presented in Appendix E.4 as Table E.1. The Mn/Model geomorphological field work yielded twelve radiocarbon dates from the Minnesota River valley. Eleven more radiocarbon dates have been included from an alluvial fan (Hudak 1997). Of these eleven dates, some have never been reported until now, while others are reported by Hudak (1997). These 23 combined radiocarbon dates and their associated data are presented in Appendix E.5 as Table E.2, and also in the cross-section profiles (Section 22.214.171.124).
One batch of radiocarbon analyses yielded some suspiciously young dates, and should be regarded with caution when viewing them in the cross-sections. Some of these dates could be explained by modern roots working their way down to as far as seven meters depth from the land surface; however, some dates from the same batch were also collected from the Root River valley, and they too were suspiciously young. One of these samples was comprised of identified seeds (clearly a subaerial part of the plant). The methods used to separate and identify these organics during the identification process may have contaminated the organic material.
126.96.36.199 Site-Specific Field Methods and Mapping
No unusual field methodologies were applied to the Minnesota River investigation. High altitude color infrared aerial photographs from the USGS NAPP program were the primary source of geomorphic information. Tonal contrasts on these air photos were sufficient to delineate features. The Big Stone, Lac Qui Parle, Swift, Chippewa, Yellow Medicine, Renville, Redwood, Brown, Nicollet, Blue Earth, Le Seuer, Sibley, Carver, Scott, Hennepin, and Dakota NRCS soil surveys were useful in supplying subsurface textural information. Mapping was done independent of previously available maps of the area for two reasons: 1) the scale of existing maps differs from the scale used here, and 2) reference to existing geologic models was avoided. Mapping was accomplished by thin pencil directly on USGS 7.5' topographic maps. Many of the geomorphic boundaries were transferred to the maps from mylar overlays of NAPP aerial photography, with appropriate adjustments for distortion of the aerial photographs.
Landform-sediment assemblages for the Minnesota River Valley and their landscape suitability rankings are summarized in Table 12.2 and discussed below.
Upland landscape. Undifferentiated upland areas (U) are isolated by Minnesota Valley LsSA’s in several places, and they are underlain by Des Moines Lobe till.
Catastrophic Flood landscape This landscape consists of erosional and depositional landforms created by, and collectively characteristic of, one or more extremely large magnitude floods, including the one(s) that might have cut the Minnesota valley. The valley itself represents a mega-landform: a Catastrophic Flood inner channel. Catastrophic flood landforms are the highest and oldest geomorphic surfaces inset below the surrounding uplands. In the upper Minnesota Valley northwest of Granite Falls, catastrophic flood landforms are abundant across the uplands beyond the valley proper. These were not mapped. Basic landforms in this landscape include bars (CB), terraces (CT), marginal channels (CMC), erosional residuals (CER), paleochannels (CPC), and erosional straths (CST). They dominate the valley landscape northwest of Montevideo where the valley is relatively wide. Southeast of Montevideo, where the valley is relatively narrow, Catastrophic Flood landforms are represented along valley margins almost continuously. Downstream of Mankato, the valley widens again as valley walls are cut in a series of alternating arcs, probably due to instability and turbulence introduced into the flood flow upon its exit from the acute angle of the valley at Mankato. Families of basic landforms occur at minimally two distinct levels within the valley; however, no temporal significance is placed on this dichotomy at this time because they could have formed at different stages during the same flood flow. Associated sediment assemblages overlie bedrock or glacial drift. Valley walls are mostly cut into tills. The valley widens in several locations as floodwaters encountered more resistant igneous and metamorphic intrusions, which caused increased turbulence and scouring. The eroded bedrock protrudes above the valley floor as irregular erosional straths. Sediment assemblages are quite variable, both among and within most basic landform types. Coarse material deposited during the flood comprises both bars and terraces, and can mantle to varying degrees the marginal channels, erosional residuals and erosional straths. Fine material may represent waning flood deposits, but in most cases represents local tributary deposits and sheetwash deposits that post-date catastrophic flooding.
Catastrophic flood bars along the Minnesota Valley are primarily streamlined mid-channel bars that occur between marginal flood channels. These LfSA’s are distributed throughout the length of the valley. The associated sediment assemblage consists of laterally continuous coarse material, greater than two meters thick, that forms the bars, and dates to the time of flooding. Bars are differentiated based on the whether bedrock or till underlying the bars is exposed beneath the bars (CBCOK) or not (CBCO).
Erosional residuals (CER) are streamlined lemniscate-shaped landforms that form by erosion of the substrate during catastrophic flooding. Clusters of erosional residuals occur northwest of Montevideo, in the New Ulm vicinity, and where igneous and metamorphic intrusions cross the valley. Hundreds of erosional residuals occur in the uplands on either side of the valley northwest of Granite Falls, but are not included in this mapping. The associated sediment assemblage in the valley is meager and consists of little or no coarse material overlying the streamlined form.
Erosional strath terraces (CST), as mapped, are formed on bedrock or glacial drift, and do not have other forms such as the streamlined erosional residuals or marginal channels. Strath terraces are overlain by little or no overlying material that may be continuous to discontinuous. Generally the overlying material, when present, is less than two meters thick. Coarse material primarily was deposited by catastrophic flooding. Fine material primarily represents accumulation of sheetwash and tributary creek deposits. On low-lying straths, fine material can also include significant amounts of Minnesota River overbank deposits, including post-Euroamerican settlement alluvium. Differentiation between strath terrace LfSA’s is based on sediment type and thickness. Coarse material may be thin and discontinuous (CST) or less than 2 m thick (CSTCO). Corresponding units that exhibit a boulder lag are CSTR and CSTRCO. Fine material may be thin and discontinuous with type "o" overbank deposits (CSTFO), or less than 2 m thick (CSTF). In just a few cases in the downstream part of the valley, silt interpreted as loess is present on strath remnants (CSTFL). Fine over coarse material may be less than two meters thick with type "o" overbank deposits (CSTFC). Where material differentiation could not be made, but type "o" overbank deposits were judged to be present, the strath terrace is mapped as CSTO.
Flood terraces are generic surfaces, formed by catastrophic flooding and elevated above the modern floodplain, that lack forms of other more recognizable Catastrophic Flood landforms (i.e. bars, channels). As mapped, flood terraces include slip-off slopes and more flat-lying pendant and alcove bars. These LfSA’s are found throughout the length of the valley. Sediment assemblages are greater than two meters thick, continuous, and dominated by sand in varying thicknesses that dates to the time of flooding. When fine material, peat, or organic muck is present, it usually overlies coarse material, and most likely post-dates flooding. Flood terraces are differentiated primarily on whether the material is coarse (CTCO) or fine (CTF). In some cases, bedrock or till is seen to outcrop beneath the flood terrace sediment assemblage (CTCOK). Rarely is loess present in the downstream parts of the valley (CTFCL).
Marginal channels are broad concave to flat sluiceways. They typically occur between valley walls and flood bars or erosional residuals, but when mapped they may refer to former flood channels regardless of the relationship to valley walls. Marginal channels occur along the length of the Minnesota valley in close association with bars and erosional residuals. Northwest of Granite Falls, flood channels are also prominent in the upland landscape, but they are not mapped. Marginal channels range in elevation, but tend to be associated with the aforementioned distinct levels of Catastrophic Flood landforms. The associated sediment assemblage overlies glacial drift and bedrock. The sediment assemblage is variable in terms of thickness and sediment sequence. Coarse material in channels was primarily deposited during waning flood currents. Fine material represents local tributary overbank deposits and sheetwash deposits that post-date catastrophic flooding, but may include latest waning catastrophic flood deposits. In some cases, peat formed in bogs occupying more poorly drained parts of former channels. Northwest of Montevideo, the lowest channels are mantled by, and mapped as, younger alluvial fans and floodplain deposits. Floodplain deposits may include post-Euroamerican settlement alluvium. Some channel surfaces have boulder and cobble lags resting on them.
Differentiation is primarily based on texture and thickness of the sediment assemblage and whether basement material upon which the marginal channel is cut into is outcropping. Coarse sediment assemblages with basement material outcroppings are either thin and discontinuous (CMCO-), less than 2 m thick (CMCCO<), or greater than 2 m thick (CMCCO). Sometimes when coarse material is greater than 2 m thick the basement material was not identified (CMCCO>). Boulder and cobble lag is present in some marginal channels in the upper Minnesota valley where they co-occur with thin and discontinuous coarse material (CMCR-) or material less than 2 m thick (CMCRCO<). Fine sediment assemblages are differentiated on whether the basement material does outcrop where it is less than 2 m thick (CMCF) or doesn’t outcrop where it is greater than 2 m thick (CMCF>). The sediment assemblage may be fine over coarse material (CMCFC), a common situation where post-flood deposits are present. Peat sediment assemblages are differentiated on whether basement material outcrops where it is less than 2 m thick (CMCP) or doesn’t outcrop where it is greater than 2 m thick (CMCP>).
Paleo-Valley landscape. South of Mankato, there is an abandoned valley segment of the Blue Earth River. The valley floor is generally featureless, with the exception of tributary features, and is typically underlain by fine textured material greater than 2 m thick (YF). In some locations, type "o" overbank deposits are present (YFO). The lack of any Blue Earth River paleochannels is curious and raises the possibility that the paleo-valley was cut when the Minnesota Valley was cut during catastrophic flooding. The lowest reach of the Blue Earth River drains the basin of Glacial Lake Minnesota. If Lake Minnesota was present when catastrophic flooding carved the Minnesota Valley, the lake might have drained rapidly, cutting multiple valleys. Alternatively, Glacial Lake Minnesota itself might have been formed by overflow of catastrophic flood waters. Glacial Lake Minnesota is located at the bend of the Minnesota River at Mankato. The bend could have caused hydraulic damming, initiating overflow and temporary storage to the south. The paleo-valley and Blue Earth River Valley could have formed with ultimate drainage of these flood backwaters as the flood waned.
Valley Terrace landscape. There are two recognized valley terrace levels in the Minnesota Valley at this scale of mapping (1:24,000). Both probably represent the lowest levels of the inner channel of catastrophic flooding and probably formed during the latest stages of catastrophic flood flows. They are, however, separated from the Catastrophic Flood landscape because they usually have substantial thicknesses of younger deposits overlying the floors of these flood channels. There are no intervening paleosols, and sedimentation appears to have been more or less continuous from waning flood deposits to overbank deposits. This is particularly the case for the VT1 surfaces. Both terraces are found as intermittent remnants in the vicinity of Montevideo and downstream to the Mississippi valley confluence. LsSA’s of both terraces are variable with materials ranging from coarse to peat, and thickness ranging from little to no material to greater than 2 m. Coarser material probably represents late stage Catastrophic Flood deposits or initial tributary valley-derived deposits. Finer material may also be late stage Catastrophic Flood deposits in addition to both Minnesota River and tributary deposits. Sequences with peat are more common in the downstream reaches of the valley. Basement material consists of glacial drift or bedrock. Surface modifications include one dune field in the lowest valley reach and overbank deposits, including post-Euroamerican settlement alluvium, either from the tributaries or flooding of the Minnesota River. Substantial alluvial fans have been deposited on these terraces.
The VT2 terrace is differentiated on whether coarse material is thin and discontinuous (VT2-), less than 2 m thick (VT2CO<), less than 2 m thick with type "o" overbank deposits (VT2CO<O), or greater than 2 m thick (VT2CO). Where the sediment assemblage is fine material greater than 2 m thick (VT2F), a buried soil tends to be present. Where the sediment assemblage is peat, it is differentiated on whether it is greater than (VT2P) or less than (VT2P<) 2 m thick. On rare occasions, paleochannels with a fine texture fill are associated with the VT2 terrace. These are mapped as VPC2F.
The sediment assemblage underlying the VT1 terrace is more consistently greater than 2 m thick compared to the VT2 LfSA. Where the sediment assemblage is coarse, it is differentiated on whether it is greater than (VT1CO) or less than (VT1CO<) 2 m thick. Fine sediment assemblages are always greater than 2 m thick (VT1F) and may overlie coarse material (VT1FC) or may have type "o" overbank deposits (VT1FO). The latter case often has a buried soil associated with the sediment assemblage. Peat may be less than 2 m thick and overlie fine material (VT1PF) or it may be thicker than 2 m (VT1P). In both cases, buried soils can be expected.
Tributary valleys usually exhibit one or more series of paired or unpaired terraces. All but the lowest are usually strath terraces or have a sediment assemblage less than 2 m thick. These are not differentiated and are mapped as VT. Paleochannels are common and are undifferentiated (VPC).
Floodplain landscape. Four different types of floodplain occur in the Minnesota Valley based on morphology and expression on aerial photographs. Downstream of Montevideo, floodplains occur in a consistently narrow restricted valley, likely representing the location of the Catastrophic Flood inner channel. North of Montevideo, an inner channel is less well defined, and floodplain areas expand over a wide valley area, although valley lakes and reservoirs limit these areas. All floodplain types are found throughout the valley length south of Montevideo, although there are a number of downvalley trends in terms of relative valley area covered. There are no explicit temporal relationships in the definitions of the different floodplain types; however, given the types of processes exhibited by the Minnesota River, the age of the LfSA’s probably increases from type "w" through types "x" and "y" to the oldest, type "z," particularly expressed in the age of their respective basal deposits. The floodplain LfSA’s are all greater than two meters thick and laterally continuous over undifferentiated basement material. Type "w" and type "x" LfSA textures are either coarse or fine over coarse. Type "y" and type "z" LfSA textures are fine to depths below 5 m. Coarse material generally represents point bar and lateral accretion deposits. Fine material represents the vertical accumulation of floodplain overbank sheetwash deposits.
Although there is no difference in general sedimentation processes, the presence of overbank deposits are explicitly distinguished in the sediment assemblage when 1) they show a high reflection on aerial photographs, with or without an associated rise on topographic maps; or, 2) field observations indicate a rise in the landscape adjacent to and paralleling the river, but the rise is too broad and subtle to classify as a natural levee. Two types of overbank deposits are recognized. Type "o" deposits are assigned where relatively very light tonal contrasts on aerial photography of valley areas is interpreted as overbank deposits that are likely to include, or field evidence indicates, deposition of significant post-Euroamerican settlement alluvium. In this case "significant" means greater than the depth of plowing, or about 0.27 m. If not otherwise noted, the presence of type "o" overbank deposits is implied with the type "w" floodplain LfSA. Type "a" overbank deposits are assigned where relatively light tonal contrasts (versus "very light" for type "o") on aerial photography are interpreted as overbank deposits. They may or may not include significant post-Euroamerican settlement alluvium.
There is also a range of other types of LfSA’s recognized in the Floodplain landscape. Natural levees and crevasse splays are common features in the lowest part of the Minnesota Valley. Throughout the valley length, tributaries are responsible for meander and overbank belts. Deltas or fan deltas, sometimes complete with distributary networks, are deposited at the heads of some valley lakes and reservoirs.
The type "w" LfSA (FFW) consists of active and recently active point bars adjacent to the river. Textures are undifferentiated, but typically they are sand or a relatively thin increment of fine textured sheetwash overbank material over sand. The type "x" LfSA (FFX) consists of point bars that are no longer active and that are actively being buried by overbank deposits. Evidence of ridge and swale morphology is present. Fine textured overbank deposits overlie coarse textured lateral accretion deposits. Type "a" overbank deposits are often present (FFXA).
The type "y" LfSA (FFY) shows no surface indications of channel migration features, either because they are relatively deeply buried or were never present. Where overbank deposits are not present, drainage conditions are generally poor. These often are basinal areas, and permanent or intermittent standing bodies of water are common. When present, overbank deposits tend to be type "o" (FFYO), but type "a" overbank deposits also occur (FFYA). Type "y", type "z" floodplains (FFZ) also have no surface indications of channel migration features and occur in sheltered areas partially surrounded by valley terrace or older LfSA’s. Poor drainage conditions occur where overbank deposits are absent and basin areas tend to support standing water bodies. As with type "y", type "z" floodplains may have type "a" (FFZA) or type "o" (FFZO) overbank deposits, depending on current and former Minnesota River channel locations.
Tributary valley floodplains are undifferentiated (FF), ranging from relatively short, steep and narrow to large and wide valleys. They often include low terraces. In smaller valleys, the FF LfSA includes narrow channel belts.
The natural levee LfSA (FNL) is limited to the lowest reach of the Minnesota Valley. In this reach overbank deposits formed better-expressed ridges (i.e., narrower and slightly steeper) than farther upvalley, although the change from one to the other is gradual and subjective. Associated sediments are fine textured, laterally continuous, and thicker than two meters. Upper increments of natural levees likely consist of type "o" overbank deposits, although this is not noted and natural levees were not tested.
Crevasse splays occur in the lowest reach of the Minnesota Valley. They typically overlie and/or interfinger with floodplain LfSA’s (FCSF), but may overlie lower surfaces of the Valley Terrace LsSA (FCSV). Fine textured material representing sheetwash flows associated with the splay are laterally continuous and greater than 2 m thick. Thin increments of coarse channel sand would be expected in the lowest parts of splays.
Minnesota River paleochannels are differentiated by their relative age, which is based upon their surrounding floodplain type: FPC1 for type "w," FPC2 for type "x," and FPC3 for type "y" floodplain. If undifferentiated, they are mapped as FPC. These paleochannels occur along the length of the valley. Paleochannel fill consists primarily of thick, continuous, fine textured material that is derived from overbank flooding of the Minnesota River and some tributaries. Coarse channel material is generally deeply buried by fine material. Many of these paleochannels support lakes.
In the uppermost reaches of the Minnesota Valley, several riverine lakes and reservoirs spanning the width of the Catastrophic Flood inner channel have delta lobes (FDE) encroaching from their upstream end. These sediment-dominated delta lobes have abandoned meander belts expressed at the surface. The lobes’ sediment assemblage includes fine textured overbank deposits, which are laterally continuous both parallel with and perpendicular to the meander belts, and greater than two meters thick. In a few instances, larger tributaries also enter these water bodies, depositing a combination of alluvial fan sediments (proximal) and fan delta sediments (distal). These landforms have been mapped as alluvial fans.
Islands (FI) in the Minnesota River (FR) are underlain by fine texture material greater than 2 m thick. They are untested, but likely to have relatively thick increments of type "o" overbank deposits.
Tributary features within the Minnesota Valley include active and abandoned meander belts, overbank belts, and undifferentiated floodplains and terraces. Textures and thickness usually are undifferentiated for mapping purposes. Tributary features overlie, or are inset into, Catastrophic Flood, Valley Terrace and Floodplain landscapes that indicate a maximum relative basal age for different tributary features.
Active and abandoned meander belts (FMB) incorporate their own suite of landforms too small to map individually. These include different floodplain types, floodplain features (e.g., point bars, natural levees, etc.), and channels. All are inactive due to stream avulsion. Multiple abandoned meander belts are not uncommon for a single tributary, and the relative age relationships among them are coded (i.e. FMB1C, FMB2C). This LfSA is further differentiated by the surface upon which it was deposited and therefore maximum relative basal age (i.e., FMB1F for floodplain LfSA’s, FMB1V for valley terrace LfSA’s).
Similarly, active and abandoned overbank belts (FOB) are mapped where there is clear evidence of tributary overbank deposits being laid down across Minnesota Valley surfaces and beyond the limits of an associated tributary meander belt. This LfSA is differentiated by the surface upon which it is deposited and therefore maximum relative basal age (i.e., FOB1F for floodplain LfSA’s, FOB1V for valley terrace LfSA’s).
Valley Margin landscape. Both alluvial fans and colluvial slopes are well represented along the Minnesota Valley in all reaches. Their geomorphic position indicates they formed after the incision of the valley by catastrophic floods. Both colluvial slope and early fan deposits incorporate mass wasting products as the valley walls, freshly cut by catastrophic flooding, would have been susceptible to mass slope failures. This apparently was exaggerated in some valley reaches depending on what glacial drift stratigraphy was exhumed by valley cutting.
Although there is a continuum between small alluvial fans and colluvial slopes, alluvial fans (MAF) are distinguished by their fan shape and must be large enough to be mapped at a meaningful scale. Sedimentologically, fans exhibit better sorting than colluvial slope deposits, develop upward fining sequences, and may exhibit multiple incipient paleosols at the top of upward fining sequences. Associated sediments are variable within fans and currently are not differentiated. They vary depending on location of the contributory basin as well as within individual fans, having more fines in a down-fan direction. Fan sediments consist of interstratified coarse and fine material. Coarser textures were deposited as stream channel deposits. Finer textures primarily represent sheetflood or sheetwash deposits. The surface on which the fan rests further differentiates this LfSA by relative basal age. Fans associated with comparable tributary drainage areas tend to be larger where they overlie the catastrophic flood (MAFC) and valley terrace (MAFV) LfSA’s, compared to where they overlie and interfinger with the floodplain (MAFF) LfSA’s.
Colluvial slopes (MC) are extensive and well developed in some valley reaches. Associated sediments tend to be loams and loam diamictons, but are undifferentiated. These LfSA’s are greater than 2 m thick and laterally continuous. Depositional processes range from slumping to sheet flow.
Eolian landscape. This LfSA is represented by a single dune (EED) field with interdunal depressions (EDP) in the downstream reach of the Minnesota valley. It is the stretch of valley oriented nearly west – east where westerly winds could be effective. Associated sediment is coarse, generally greater than 2 m thick, but variable, and laterally continuous. The main part of the dune field post-dates the catastrophic flood LfSA’s upon which it was deposited (EEDC) and likely post-dates the formation of the VT2 terrace upon which some dunes are deposited (EEDV) downwind of the main dune field.
188.8.131.52 Landform Sediment Assemblage Codes
Table 12.2 provides details on each of the specific LfSA codes used for the Minnesota River Model.
If the origin of the Minnesota Valley as expressed today was by catastrophic flooding, then the full range of LsSA’s are temporally suitable for cultural deposits. Multiple radiocarbon ages in the range of 10,400 B.P. (Hajic 1996b; Bower et all. 1996; Hudak 1997), including one from this project at 10,410 ± 60 B.P. (Beta-97193), from shallowly beneath the VT1 terrace, indicate catastrophic flooding occurred prior to this time. Sedimentology of deposits overlying samples yielding these radiocarbon ages suggests no large magnitude (Catastrophic Flood) discharges occurred here after ca. 10,400 B.P. This supports the notion of Fisher and Smith (1994) that the recently recognized northwest outlet channeled discharge from Glacial Lake Agassiz during the Emerson Phase. Catastrophic flooding probably occurred around, or post-dates about 11,700 B.P., an estimated age for the Cass Phase, as suggested by the chronology of Glacial Lake Agassiz (Fenton et al. 1983). This time interval agrees well with the record in the Mississippi Valley downstream. Downcutting from the Savanna to the Kingston Terrace levels occurred sometime between about 12,200 and 10,400 B.P. (Hajic et al. 1991).
Large tracts of the valley are composed of catastrophic flood LfSA’s for which depositional environments would have been unsuitable for preservation of cultural deposits, much less habitation. Excluding post-flood deposits, it is not possible (0) for bar, terrace, marginal channel, erosional residual, and strath terrace LfSA’s to have buried cultural deposits (CBCO, CBCOK, CER, CMCCO, CMCCO-, CMCCO<, CMCCO>, CMCO-, CMCR-, CMCRCO<, CST, CSTCO, CSTR, CSTRCO, CTCO, and CTCOK).
Two LfSA’s have loess incorporated into the Catastrophic Flood LfSA’s (CSTFL, CTFCL). Where this occurs, the 0-1 m interval is ranked high (3). This ranking is warranted because eolian sedimentation of loess is ideal for burial and preservation of cultural deposits, and the loess must post-date about 10,400 B.P. Although not taken into account in the LSR’s, these LfSA’s are likely to have been subjected to localized sheetwash erosion and sedimentation and biomantle formation processes, given their ages. This applies whether or not loess is present.
Marginal channels not filled with flood sand and gravel pose a different situation. Where fine sediment is present (CMCF, CMCF>), a ranking of moderate (2) is applied to a depth of 2 m. Where the fine material is greater than 2 m thick, a ranking of low (1) is assigned to the interval greater than 2 m (CMCF>). Where fine material overlies coarse material in the less than 2 m thick range (CMCFC), the 1-2 m interval is ranked low (1). The origin of the fine material may be related to the latest stages of flood flow; however, in many cases, it is clearly related to tributary overbank deposits, localized sheetwash deposits, and upland-derived wash beyond the limits of alluvial fans and colluvial slopes. All of these environments are conducive to preservation of cultural deposits.
Peat is present in a few isolated low areas of marginal channels (CMCP, CMCP>). Cultural deposits could be buried by peat on the flanks of these low areas. The age of the peat is unknown, but as it was deposited, depression flanks were buried by colluvial processes. Within the peat, only isolated artifacts would be expected because of the very poorly drained conditions, but the preservation potential is excellent; therefore, any prehistoric cultural deposits associated with the peat are likely to be buried, and the land surface is ranked not possible (0) because of the presumed young age. A LSR of low (1) is extended to the top of material underlying peat.
The paleo-valley located where the Blue Earth River joins the Minnesota River is characterized by a fine texture fill greater than 2 m thick (YF). An extremely low energy depositional environment is possible here, particularly if the Blue Earth River never flowed through the valley. The paleo-valley could simply have received seasonal floodwaters of the Blue Earth River. A moderate (2) ranking is assigned to the depth of fine texture fill. Where type "o" overbank deposits are present (YFO), the surface is ranked not possible (0) because it is interpreted to be too young for precontact cultural resources.
Organic material from an A-horizon developed at the surface of the VT1 LfSA yielded a date of 8,500 B.P.; this A-horizon was subsequently buried by an alluvial fan (Hajic 1996b). These facts indicate that the two valley terraces and associated post-catastrophic flood sediment assemblages are terminal Pleistocene to early Holocene in age. Paleoindian through Early Archaic cultural deposits could be associated with these LfSA’s.
Coarse sediment assemblages of valley terraces are either related to catastrophic flooding, or slightly younger fluvial activity. Taking a conservative approach, these deposits are ranked low (1) to the depth to which they extend (VT2-, VT2CO, VT2CO<, VT2CO<O, VT1CO, VT1CO<). In the few instances where type "o" overbank deposits are mapped, the surface is ranked not possible (0) because of a young age. Fine-textured sediment assemblages, best represented with the VT1 LfSA’s, are ranked moderate (2) to the depth to which they occur (VT2F, VT1F, VT1FC, VT1FO). These represent lower energy depositional environments, largely floodplain overbank deposits. At the time of sedimentation, they may have been no better drained than the modern floodplain, but now are slightly better drained because of a slightly higher landscape position. Buried soils are possible, particularly for the VT1F LfSA. As with their coarse counterparts, there are instances where type "o" overbank deposits are mapped (VT1FC, VT1FO). In these cases the surface is ranked not possible (0).
There are expansive terrace remnants with peat in the downstream reach of the valley. Whether these surfaces are truly related to the VT1 and VT2 surfaces is unknown. In any case, peat has been deposited to levels comparable to the non-peaty valley terraces. Although the age of the peat is unknown, it is expected to post-date the radiocarbon-dated age of the VT1 surface from up valley. This peat is interpreted to be Historic in age, hence the surface is ranked not possible (0). Again, taking a conservative approach, the peat is ranked low (1) to its maximum depth.
Tributary terraces (VT) are numerous and occur at multiple, undifferentiated levels. They are typically underlain by coarse material, but sediment assemblages vary widely. These are cautiously ranked moderate (2) to the 2 m depth. Associated paleochannels (VPC, VPC2F) are ranked not possible (0) for all depth intervals.
The FFW LfSA is ranked not possible (0) because it is considered to be Historic in age. The FFX and FFY LfSA’s are ranked as moderate (2) to a depth of 2 m, and low (1) for depths greater than 2 m. The moderate (2) ranking is considered to be stronger where type "a" overbank deposits are present (FFXA, FFYA). Where type "o" overbank deposits are present, the most common situation, the surface is ranked not possible (0). This ranking is not carried through to the 0-1 m depth interval because the thickness of type "o" overbank deposits is not well known, but could be substantial. The 0-1 m depth increment of the FFZ LfSA is ranked low (1) because of the extremely poor drainage condition. Deeper deposits are ranked not possible (0) because of the extremely poor drainage conditions and low-lying landscape position. Where type "a" overbank deposits are present (FFZA), the low (1) ranking is extended to 2 m depth. As in other situations, the surface is ranked not possible (0) if type "o" overbank deposits are present (FFZO). Buried soils are likely to be found in the FFY and FFZ LfSA’s, particularly where overbank deposits are present.
Tributary valley floodplains will vary widely in the character of their LfSA’s (FF). They are undifferentiated by sediment type and thickness, but for the purposes of LSR’s, are considered to be greater than 2 m thick. This will vary, especially considering the fact that low terraces are incorporated in this map unit. A moderate (2) ranking is applied at this time to emphasize the great potential. This potential is expected to be greater in larger tributary valleys.
Natural levees (FNL) represent slightly better drained landscape positions and relatively low energy depositional environments. Cultural deposits can be buried and preserved within this LfSA. They are ranked moderate (2) to a depth of greater than 5 m. Crevasse splays (FCSF, FCSV) are ranked comparably for similar reasons. Although not tested, buried soils are considered to be likely occurrences in these LfSA’s. Strong possibilities exist for type "o" overbank deposits to be present on both these LfSA’s. One or more of the crevasse splays may be Historic in age.
Paleochannels are excessively poorly drained, often supporting perennial water bodies. They are ranked as not possible (0).
Delta lobes entering riverine lakes are locations of low energy sheetwash overbank sedimentation much like floodplains. With vertical accretion and progradation, intact burial of cultural deposits is likely. Deltas (FDE) are ranked moderate (2) to a depth greater than 2 m.
Islands (FI) experience similar depositional processes as floodplains and natural levees; therefore the ranking is moderate (2) to a depth of 2 m. Depths below 2 m have low (1) LSR’s because of drainage considerations. Substantial increments of type "o" overbank deposits are probable; therefore the surface has a ranking of not possible (0).
Tributary meander belts are a mosaic of landforms. Individually, they may have different rankings, but collectively they are assigned a low (1) probability to the depth of the LfSA. This depth ususally exceeds 5 m where meander belts are on or inset into the floodplain LfSA’s. Where they are inset into valley terrace or catastrophic flood LfSA’s, they can be less than 2 m thick. Buried soils can be expected, as can type "o" overbank deposits. Overbank belts are typically related to tributary meander belts occurring beyond their limits. They consist of fine-textured overbank deposits greater than 2 m thick. Buried soils can be expected, especially developed at the underlying surface the overbank deposits may bury. Moderate (2) rankings are assigned to the depth of the LfSA because of the low energy sheetwash sedimentation processes and the slightly better drained landscape position afforded by the overbank belt relative to the surface it buries. The exceptions are where overwash is deposited on catastrophic flood or valley terrace LfSA’s (FOB1C, FOB1V, FOB2C, FOB2V). In these cases, the 2-5 m depth interval is ranked low (1). Basal age of outwash belts, and therefore the age of cultural deposits they contain, will vary with the age of the surface being buried and the period of activity of the meander belt.
Radiocarbon ages from valley margin alluvial fans indicate they formed throughout the Holocene (Table E.1; Hajic 1996b; Hudak 1997). The basal age of any particular fan or colluvial slope is limited by the age of the geomorphic surface it is developed on. Because of the restricted width of the Holocene floodplain, most of the alluvial fans (MAF, MAFC, MAFF, MAFV), and nearly all colluvial slopes (MC), are deposited on the valley terrace or catastrophic flood LfSA’s. Rankings are high (3) due to the dominance of sheetwash sedimentation, moderate to well drained conditions, and paleosols. Two paleosol complexes were recognized by Hudak (1997) in the fans and were bracketed more fully in radiocarbon years during the Mn/Model geomorphic study. The youngest recognized paleosol complex ranges in age from 2,000 to 4,500 B.P., whereas the older paleosol complex ranges in age from 5,400 to 8,000 B.P. The paleosols tend to be thinner nearer the fan apex, and more dispersed vertically toward the distal end. Furthermore, Hudak (1997) interpreted that some of the near-surface fan sediments may be post-Euroamerican settlement in age, and could depend largely upon the farming practices and geology within the tributary valley and drainage basin.
The age of eolian LfSA’s is not known, but suspected to be either terminal Pleistocene, because of the surfaces upon which they are deposited, or middle Holocene, due to generally dryer conditions of the Hypsithermal climatic episode. Eolian LfSA’s (EEDC, EEDV, EDP) are ranked high (3) because of well-drained conditions and the potential of rapid burial as dunes migrate.
The Upper Mississippi River valley investigation is constrained between the cities of Bemidji and St. Cloud, Minnesota, and actually overlapped with the Bemidji Uplands Model (Section 12.12). This stretch of the Mississippi valley is divided into three different reaches based upon its landforms and shape, which indicates its complex origins. The uppermost reach is the chain-of-lakes section and runs from Lake Bemidji through Cass Lake and Lake Winnibigoshish to White Oak Lake. Another segment crosses the Glacial Lake Aitkin lakebed from downstream of the City of Grand Rapids to just downstream of the City of Aitkin. The last segment stretches from upstream of the City of Brainerd to downstream of the City of St. Cloud, and is typified by the glaciofluvial terraces and plains.
Detailed, systematic investigations concerning the geological evolution of the project area are not published. Wright (1972) has generally described the upstream area to be within the "Bemidji Area," which is generally "...heavily forested and poorly known...." Wright continues to describe the area as a pitted outwash plain from Bagley to Bemidji to Lake Winnibigoshish. These outwash streams continue eastward toward Grand Rapids and into Glacial Lake Aitkin.
The DNR (1998) GIS project indicates that the current Mississippi River project area starts across the Koochiching lobe’s outwash near Brainerd, and then courses downstream across the Wadena lobe’s Itasca moraine, the Koochiching lobe’s outwash near Lake Winnibigoshish, to Glacial Lake Upham/Aitkin, to the St. Louis lobe’s Culver ice margin, to the Rainy lobe’s and Superior lobe’s St. Croix ice margin, and the Des Moines lobe’s Bemis phase.
Hohmann-Caine and Goltz (1986) compiled geomorphic data for the uppermost stretch of the Mississippi River. These authors present a geomorphic model for the area between Wolf Lake and Cass Lake in Beltrami County. Their model discusses the Mississippi’s channel migration in absolute time, although they have no radiocarbon dates and presented no archaeological evidence to independently assess their conclusions.
Farnham et al. (1964) report on two different glacial lake stratigraphic units near the City of Aitkin that pre-date and post-date a paleosol dating close to 11,600 B.P. This average date comes from two pieces of spruce wood collected at a cut within the Aitkin diversion channel. Hobbs (1983) reports on the drainage relations for Glacial Lakes Aitkin and Upham. Both Farnham et al. (1964) and Hobbs (1983) report that the modern course of the Mississippi River developed across the Glacial Lake Aitkin plain during the waning stages of lake drainage. Hobbs (1983) contends that this occurred after 9,080 or 10,000 B.P. based upon radiocarbon and Uranium-series dates on snail shells and marl, respectively. Farnham et al. (1964) indicate that the Mississippi River occupied its modern course after the "Valders ice withdrew." These same authors also indicate that the ancient lake basin converted to a marsh about 3,100 B.P.
Unpublished dissertations by Goldstein (1985), and Mooers (1988) summarized previous investigations on the late Wisconsinan Wadena, Rainy, and Superior lobes of central Minnesota. These researchers, along with Mooers (1991), also discuss the drainage relations of these lobes. Goldstein postulated that the Des Moines lobe blocked the west side of the "Pillager gap," which was cut by westward flowing glacial waters through the St. Croix end moraine near the town of Pillager. The Des Moines lobe blocked waters from escaping to the west through the gap, and forced them back possibly to the east and then south along the east side of the St. Croix end moraine. Goldstein (1985) indicated that this late Wisconsinan event may have created the first significant Mississippi River valley between Pillager and St. Cloud. The Superior lobe, however, also had drainage relations that may have created the Mississippi River as indicated by Mooers (1988, 1991). Regardless, there are few systematic detailed studies on the Mississippi River alluvium.
Hudak (1994b) and BRW (1994) provide some geomorphic descriptions and radiocarbon dates on at least three geomorphic surfaces along the Mississippi River terraces near St. Cloud. The highest surface, which is probably an outwash terrace, must predate a 4,500 B.P. radiocarbon date from an inset sedimentary package, and an interpreted ca. 9,000 B.P. cultural resource found on the higher terrace. This higher terrace may also be mantled by a thin veneer of alluvium or blowsand. Rapp et al. (1997) also have reported a radiocarbon date from the Mississippi River floodplain near St. Cloud. This radiocarbon date is on a charcoal sample collected from a buried Ab-horizon. The charcoal yielded a date of 200± 40 B.P., and indicated to Rapp et al. (1997) that the unit overlying the buried soil was of Historic age.
Twenty-nine cores were collected from the Upper Mississippi River valley between March 18, 1997 and May 6, 1997. All cores were taken with the MnDOT Foundations Department tracked drill-rig and crews. These cores were logged in a laboratory and are presented in Appendix E.1. The methods discussed in Section 12.2 apply to this field and laboratory work.
184.108.40.206 Cross-Section Profiles
A series of 11 cross-sections and one long-section were constructed from the soil profiles of the Mississippi River valley. These figures and one drawing are presented in Appendix E.2 and E.3, respectively.
220.127.116.11 Radiocarbon Dates
A variety of organic samples were collected from the various cores. Multiple-samples were identified in general terms from the Mississippi River and eight were sent for radiocarbon dating. A master list of organics and their identifications are presented in Table E.1. The list shows the samples that have been sent and identified at the University of Minnesota-Duluth.
18.104.22.168 Site-Specific Field Methods and Mapping
Logistics were critical to ensure that this final field project was completed because of the active flooding around the Mississippi valley. There were no unusual methodologies applied to this investigation except for the use of a tracked drill-rig for coring because of the required timing and logistics. The Beltrami, Itasca, Crow Wing, Morrison, Benton, and Stearns counties’ NRCS soil series were used to help model the texture. Some of the soil maps were produced by the Chippewa National Forest staff. Advance soil series sheets were supplied by the Aitkin County NRCS. The entire valley was originally mapped as one region; however, as the code expanded from 10 to 33 GIS fields, the Mississippi River project area was divided into the headwaters, Glacial Lake Aitkin, and Brainerd to St. Cloud stretches.
High altitude color infrared aerial photographs from the USGS NAPP program were the primary source of geomorphic information. Tonal contrasts on these air photographs were adequate, in most cases, to delineate many, but not all, features at this scale. Mapping was done independent of previously available maps of the area, partly because of the difference in scale of the pre-existing maps and of the mapping being done, and partly because of the desire to map without reference to existing geologic models of the area. Mapping was accomplished by thin pencil directly on USGS 7.5' topographic maps. Many of the geomorphic boundaries were transferred to the maps from mylar overlays of NAPP aerial photography, with appropriate adjustments for distortion of the aerial photographs. The Holocene landforms were not always recognizable at this scale of mapping, mostly because the landforms were thin arcuate polygons that were smaller than the pencil line widths.
Landform sediment assemblages of the Mississippi River Valley are outlined in Table 12.3 and discussed below.
The Upland, Glaciofluvial, Glaciolacustrine, Valley Terrace, Floodplain, Valley Margin, and Lacustrine LsSA’s were used in the mapping of the Upper Mississippi River study area. The undifferentiated Upland landscape applied mostly to glacial drift and bedrock outcrops along the river course. Glaciofluvial LfSA’s dominated the mapping of the valley. Select areas such as Glacial Lake Aitkin also dominated the mapping in local areas. The Lacustrine landscape was common upstream of Grand Rapids and downstream of Brainerd on the Glaciofluvial landscape. Valley Terrace and Valley Margin LfSA’s were a relatively minor constituent at this scale of mapping. The Floodplain landscape was most prevalent across the Glaciolacustrine plains.
The Mississippi River Valley is currently the only Mn/Model-mapped valley that is divided into stretches or reaches. Other valleys will likely be divided and further subdivided into smaller reaches as more data are collected and trends are recognized at a more local scale. The Mn/Model-mapped sub-regions include the Headwaters (H-), Glacial Lake Aitkin (A-), and Brainerd to St. Cloud (B-) reaches. The landforms identified at this scale of mapping may be interpreted to be of different ages between reaches, and in some cases are probably of different ages within a reach. The landforms, when shared between reaches, have similar geomorphic traits and are therefore discussed below as landscapes and landforms across the entire mapped valley.
Upland landscape. Undifferentiated uplands were mapped at the outer edges of some of the Mississippi Valley landscapes, or between the valley and some of the ancient glaciofluvial valleys and features.
Glaciofluvial landscape. This landscape is complicated within the Upper Mississippi basin. Several events of glacial outwash have formed different parts of both the main and tributary valleys. This landscape may be correlative to other landscapes, such as Paleo-Valleys, and is only distinguished by scale. Multiple geomorphic surfaces were mapped within Mn/Model’s Glaciofluvial landscape. Some of the surfaces were not labeled with respect to relative age because of the many influences of glacial meltwaters from many ice fronts of different ages. A little more investment on time may help to decipher the geomorphic surfaces of this landscape. The Mn/Model LfSA’s relative age indicators are provided as local guides within a larger framework of glaciofluvial terraces and plains. Unlike most other Mn/Model LfSA’s, these Glaciofluvial LfSA’s are not meant to be correlative units from Bemidji to St. Cloud; however, they are meant to be correlative in a more localized area or region. The textures are mostly coarse outwash throughout the valley with some relatively thick peat mantles in areas. Braided channel patterns are apparent in many areas, especially downstream from the Brainerd area (B-OTB, B-OT1B, B-OT1BM, B-OPB, B-ORP).
The terrace LfSA’s have coarse textures, and can be further distinguished by having one of the following traits: thick glaciofluvial strata (B-OT, H-OT); thick glaciofluvial strata with braided patterns (B-OT1B); less than 2 m depth to undifferentiated glacial drift or bedrock (B-OT1); braided channel patterns with less than 2 m of coarse outwash overlying glacial drift basement material (B-OTB); braided channel patterns with natural standing water (B-OT1BM); and occasional standing water with peat (B-OTMA, H-OTMA). A paleo-rapids (B-ORP) is undifferentiated with respect to geomorphic position, although it is related in origin to the main outwash plain and tributary outwash terrace near Brainerd. This landform has very obvious braided channel patterns, and a relatively steeper gradient than the surrounding land surfaces. The outwash plain or (sandur) has greater than 2 m of coarse outwash overlying an undifferentiated basement material. These LfSA’s are further distinguished by having one of the following traits: an undifferentiated surficial pattern (B-OP); peat over coarse outwash (B-OPMA); braided channel patterns (B-OPB); or pitted plain (B-OPP). The pitted plain has many lake basins which are depressions ("pits") that may or may not reach below the local water table. Many more of the pits are currently above the local water table. The paleochannel LfSA’s all have depths of greater than 2 m to the underlying undifferentiated basement material. These LfSA’s are further distinguished by having one of the following depth and material traits: greater than 2 m of coarse outwash (A-OPC, B-OPC, H-OPC); greater than 2 m of peat over coarse outwash (A-OPCMA, B-OPCMA, H-OPCMA); or greater than 2 m of discontinuous peat over coarse outwash (B-OPCPCR).
Glaciolacustrine landscape. This landscape refers mostly to Glacial Lake Aitkin’s landforms and deposits. The amount of field work conducted in the lake plain was virtually non-existent; therefore, interpretations are skeptical until more coring is accomplished. Mapping at this larger scale also precluded the "big picture" such as the sometimes interpreted Glacial Lake Aitkin delta at the north end. The glacial lake plain LfSA’s have both a varied surficial morphology and underlying stratigraphic sequence. These LfSA’s are further distinguished by having one of the following sets of depth and/or material traits: greater than 2 m of combined peat over fine over Glaciolacustrine deposits (A-APMAPF); greater than 2 m of combined peat and coarse over Glaciolacustrine deposits (A-AMAPC); greater than 2 m of peat over undifferentiated materials (A-APMAP); greater than 2 m of peat overlying interstratified coarse and fine materials (A-APMAPQ); greater than 2 m of interstratified peat and fines (A-APMAIPF); greater than 1 m of fine materials (A-APF); greater than 1 m of fine materials with a buried soil (A-APFB); greater than 2 m of coarse over fine materials with buried soil (A-APCFB); and greater than 1 m of loess overlying Glaciolacustrine deposits (A-APFL). Another origin hypothesis for this loess-like silt is in a glaciolacustrine environment. The A-ASH LfSA is currently interpreted as a lake terrace; however, it may also be an outwash terrace that has been mantled by glaciolacustrine or alluvial sediment. Another origin hypothesis for both the A-ASH, A-APFL, and related landforms is an "ice-block kame terrace" (Hudak and Hajic, in prep.). This hypothesis would help explain the variety of textures under what would otherwise appear to be the same landform or set of landforms. The A-AD LfSA appears to be ancient "kettles" or lake basins that have been dissected by the modern Mississippi River. They are now occupied by peat with depths greater than 2 m. The A-APC LfSA is interpreted to be an ancient drainageway or paleochannel. These linear features appear to have been braided and are at the edge of the current mapping boundaries. They are currently filled with peat to depths greater than 2 m. These features may be remnants of glaciofluvial event or they may be related to the last drainage of Glacial Lake Aitkin.
Paleo-Valley landscape. This landscape is used to accomodate the abandoned tributary valleys with undifferentiated floodplains and terraces (A-YFN), and also the abandoned "v"-shaped valleys (A-YV). The A-YFN LfSA is found only on the Palisade quadrangle. Textures may vary for both LfSA’s, but these assemblages probably have finer and older overbank deposits than the active tributary valleys.
Valley Terrace landscape. This Landscape has several geomorphic surfaces. The Mississippi River valley is complex and has several reaches of different origins resulting from multiple glaciofluvial events. The post-glacial history is also different between Bemidji and St. Cloud. The terraces with the same relative age indicator are, therefore, not necessarily closely related in time and genetics across the entire project area should not be assumed. These age indicators are meant as guides to build upon as more detailed investigations are produced. One generally consistent element for these LfSA’s appears to be their surficial patterns and underlying textures. Textures are generally coarse to depths greater than 2 m. The higher and presumably older LfSA’s probably have reworked Glaciofluvial materials, and tend to be concentrated closer to the downstream portion of the mapping area. There may be laterally discontinuous eolian deposits at the surface of the higher LfSA’s, but the scale of mapping precluded modeling. Basement materials were generally undifferentiated because this LsSA was commonly too thick, and in several cases unsampled by Mn/Model geomorphic field work.
The B-VT3, A-VT3, and H-VT3 LfSA’s have fine over coarse textures to depths of greater than 2 m. This LfSA is restricted to areas close to Glacial Lake Aitkin and also to the most downstream portions of the study area. The VT2 LfSA’s may be distinguished by fine over coarse textures and meandering channel patterns (A-VT2S, B-VT2S) and, coarse textures to depths greater than 2 m (A-VT2CO, H-VT2, B-VT2). This VT2 LfSA family is commonly found close to the locations of VT3 LfSA’s. The A-VPC2, B-VPC2 and H-VPC2 are paleochannel LfSA’s that are commonly found cross-cutting their respective VT2 LfSA. VPC2 LfSA’s have coarse textures to depths greater than 2 m. The VT1 LfSA’s are found generally downstream from the glaciofluvial channel areas of the headwaters subregion. This LfSA family is further distinguished by having one of the following sets of depth and material traits: greater than 2 m of coarse texture material overlying undifferentiated basement material (A-VT1CO, B-VT1, H-VT1); greater than 1 m of fine texture material overlying undifferentiated basement material (A-VT1F); greater than 2 m of combined peat and coarse outwash overlying undifferentiated basement material (A-VT1MA; B-VT1MA); or, meander channel patterns and greater than 2 m of fine over coarse textures overlying undifferentiated basement material (A-VT1S). The VP1 LfSA’s are associated with the VT1 LfSA’s mentioned above, and are further distinguished by having one of the following sets of traits: less than 2 m thickness of coarse textures to undifferentiated glacial drift or bedrock (B-VPC1<); greater than 2 m thickness of coarse texture overlying undifferentiated basement material (B-VPC1); greater than 2 m thickness of combined peat and coarse texture overlying undifferentiated basement material (A-VPC1); or, undifferentiated materials that are greater than 2 m thick overyling an undifferentiated basement material (B-VPC). The"v"-shaped tributary valley LfSA’s are generally erosional features and therefore have less than 1 m thick (sometimes 2 m thick) alluvial sequences overlying undifferentiated glacial drift (A-VVE, B-VV, H-VV). This glacial drift is mostly glaciofluvial within the main valley confines, glacial till (possibly bedrock) outside the main valley confines, and glaciolacustrine within the Glacial Aitkin area. Sedimentary textures are variable. Some tributary valleys have undifferentiated floodplain and terraces (A-VFN, B-VFN, H-VFN). These LfSA’s have variable textures and are interpreted to be greater than 2 m thick.
Floodplain landscape. This landscape is greatly influenced along various stretches of the Mississippi River by recent man-made dams. The dams are likely responsible for some of the elevational differences between floodplain surfaces within this project area. The Floodplain landscape within the Mississippi River was difficult to sample in large numbers because of the need for tracked drill-rigs; therefore, fewer samples were collected during this investigation. The Floodplain landscape is geomorphically inset beneath the Valley Terrace landscape. The different surficial morphologies of this landscape are related to the degree of channel migration expressions, which could be interpreted with regard to relative age. The more obvious channel migration features are either still active or recently active. The least obvious are interpreted to have been covered by overbank deposits through time. Other floodplain features were also noted, such as natural levees, paleochannels (including oxbows), bars, etc. Some levees are currently river islands because of reservoir flooding; and when they were distinguishable they were mapped as levees. The A-FFW, B-FFW, and H-FFWSA LfSA’s all have meandering channel patterns with type "a" overbank deposits, which is approximately 1 m thick and is interpreted to be post-settlement age. These LfSA’s are greater than 2 m thick to an undifferentiated basement material. The type "y" floodplains have most of their surficial channel migration features masked by either overbank deposits or peat. These LfSA’s are further distinguished by having one of the following sets of material and depth traits: greater than 2 m of combined peat over coarse texture overlying undifferentiated basement material (B-FFYMA, H-FFYMA); type "a" overbank deposits of pre- and possibly post-Euroamerican settlement age, which is coarse alluvium less than 1 m thick overlying glacial drift basement material (H-FFYA); type "a" overbank, which is approximately 1 m thick (H-FFYA>). This latter LfSA has fine over coarse textures, which when combined are greater than 2 m thick to an undifferentiated basement material. The B-FNL LfSA was difficult to recognize at the 1:40,000 - 1:24,000 scale using the prescribed Mn/Model geomorphic mapping methods. Some natural levees were apparent because of the dam reservoirs creating islands out of the older levees. The levees are regarded as type "a" overbank (pre-dates Euroamerican settlement ages) and are less than 2 m thick in most cases where mapped. Cultural resources are possible underlying the levee material. The most obvious paleochannels were mapped for this project. This LfSA family is further distinguished by having one of the following sets of depth and material traits: greater than 2 m of coarse texture overlying undifferentiated basement material (A-FPC, B-FPC); greater than 2 m of combined peat over coarse texture all overlying undifferentiated basement material (B-FPCMA); type "a" overbank deposits and greater than 2 m of coarse textures overlying undifferentiated basement material (H-FPCA); and, a flood scour channel pattern (H-FPCF). This latter LfSA has greater than 2 m of coarse textures overlying undifferentiated basement material.
The delta (H-FDE) LfSA is common to the Headwaters reach where the river flows from lake basin to lake basin. The deltas are on the upstream side of several lake basins. Each delta has a faint distributary pattern, which may be flooded by the reservoir waters. Type "a" overbank deposits mantle these deltas (however the material may be very recent). Greater than 2 m of coarse textures are interpreted to overlie an undifferentiated basement material.
Many of the bars (B-FB) were neither apparent nor mappable at the 1:24,000 scale. They are certainly present but not laterally expansive. Most bars were lumped in with other Floodplain LfSA’s. They generally have greater than 2 m of coarse textures overlying an undifferentiated basement material. These features are interpreted to be very recent, and may not be in existence after a short time. The bars were not sampled as part of the Mn/Model sampling program, and may be older than what is currently interpreted.
The river islands (B-FI, H-FI) have strata greater than 2 m thick, and apparently very little type "o" overbank (see discussion in Section 12.17.2). The river channel (FR) is supposed to span the entire stretch of the project area and not be divided into the separate reaches. The channel substrate is presumed to be eroded or recently deposited.
Valley Margin. This landscape is composed of alluvial fans and colluvial slopes. They are all interpreted to be greater than 2 m thick before reaching an undifferentiated basement material. A variety of textures can be found within this landscape, and buried soils are possible. The Valley Margin landscape is relatively small in both size and number compared to older landscapes in southern Minnesota. Some of the mapped LfSA’s in this code may be related to the coincidental fan-shapes of contours adjacent to drainages. Sampling of these LfSA’s was not logistically practical or possible because of their locations at the time of the Mn/Model geomorphic field work.
Fans are not very common at this scale of mapping, because the relief of the landscape is not high or mature enough to have generated any such landforms. Buried soils may be within this LfSA family. The presence of buried soils was not confirmed by this testing program. This LfSA family is further distinguished by having one of the following sets of traits: greater than 2 m depth of fine material overlying the Glaciofluvial Landscape (B-MAFO); or, greater than 2 m of fine material overlying the Valley Terrace Landscape (B-MAFV).
The colluvial slopes are not common at this scale of mapping because of the broad meandering of the Mississippi channel through time, the relatively lower relief, and the relatively young landscape. These slopes were not tested during the Mn/Model geomorphology field program because logistics made them difficult to reach. Textures vary depending upon the source materials. One of the following sets of traits further distinguishes this LfSA family: greater than 2 m depth of fine material overlying an undifferentiated basement material (B-MC); or, greater than 2 m depth of peat overlying an undifferentiated basement material (H-MCMA; this "slope" is found only on the Deer River quadrangle, but requires a field check because of the unusual aerial photographic signatures; these signatures may be related to the lack of surface water or near surface saturation).
Lacustrine landscape. The importance of this landscape lies in the chain of lakes through which the Mississippi River flows in the Headwaters reach. Lakes in other reaches were mapped as a demonstration of the proximity of lakes to the main channel and tributaries of the Mississippi River. Mapping these other lakes visually demonstrates the "pitted" nature of the Outwash plain around and downstream from Brainerd. Terraces, islands, peninsulas, and other lake landforms were mapped quickly. Sampling these landforms was not part of the geomorphic investigation. Future researchers may test these areas for the presence of geologically buried cultural resources.
The second youngest lake terrace (H-LT2) LfSA has greater than 2 m of variable material overlying undifferentiated glacial drift. This LfSA is recognized around the chain-of-lakes area closer to the Mississippi Headwaters and may represent an "ice-block kame terrace" (Hudak and Hajic, in prep). This LfSA deserves closer attention from a geological and archaeological perspective due to a lack of subsurface testing. The youngest lake terrace (H-LT1) LfSA has a depth greater than 2 m of variable material overlying undifferentiated glacial drift.
The beach shore (A-LSH1, B-LSH1, H-LSH1) LfSA’s are found around most of the larger lake basins. This LfSA family is further distinguished by having a depth greater than 2 m of combined discontinuous peat over fine textures, which overlies undifferentiated basement materials. The A-LLN, B-LLN, and H-LLN LfSA’s are all standing water bodies (lakes). The H-LLR and B-LLR are man-made reservoirs. The A-LPE, B-LPE, H-LPE LfSA’s are all peninsulas that have extremely variable assemblages because they were mapped entirely as a landform and not an assemblage. This experiment was to see if the archaeology would hold true to preconceived notions. Isthmus’ were also based upon the landform and not necessarily the assemblage. Paleo-lakebeds have either peat to depths greater than 2 m (A-LLB, B-LLBP), or peat over fine materials to depths greater than 2 m (B-LLBPF).
22.214.171.124 Landform Sediment Assemblage Codes
Table 12.3 provides details on each of the specific LfSA codes used for the Upper Mississippi River model.
The ages of the Glaciofluvial LfSA’s are interpreted to be barely young enough to contain cultural resources; however, the depositional environment’s energy is interpreted to be too high to preserve any in situ archaeological deposits. The probability for these LfSA’s to contain or preserve naturally buried cultural resources is therefore considered nil. Typically, the mantling strata such as peat have higher probabilities than outwash for containing cultural resources, although peat is systematically ranked a "low" probability because of it being formed in a poor drainage location.
The Glaciolacustrine strata are interpreted to be young enough to contain cultural resources. Although the ancient lake shores may have fluctuated back and forth, and ancient cultures could have followed these shorelines, the shoreline environment has been systematically lumped in with the off-shore environment, which is more inhospitable. Typically, the mantling strata such as loess and overbank deposits have higher probabilities for containing cultural resources. Peat is again considered to be a low LSR because of poor drainage.
The relative ages of the Valley Terrace LfSA’s in this landscape are young enough to contain cultural resources. Variability in depositional environments suggests caution and potential high probabilities for buried cultural resources, especially closer to the current land surface. The high probabilities could be dismissed if the valley terraces are proven to be "ice-block kame terraces." Tributary valleys to the Mississippi River are composed of both small "v"-shaped valleys and larger valleys with undifferentiated floodplains and terraces. Tributary paleochannels are also delineated. Some of the wider tributary valleys were created by glaciofluvial events. All tributary valleys are interpreted to overlie undifferentiated basement material. The age and depositional environments vary greatly within these valleys; therefore the LSR potentials listed in the GIS model are cautionary.
Prehistoric cultural resources may be possible given the age of the uppermost floodplain strata. The depositional environment is variable within the Floodplain Landscape and varies between higher energy channels to lower energy overbank floods. Buried soils are very possible in this environment, except in areas interpreted to have been eroded by younger channel migration events. The potential for cultural resources varies, but is ranked "low" at or near the surface primarily because of age; however, depositional environments also may preclude their possibility. The older floodplain types (FFY), natural levee (FNL), and river islands (FI) are currently interpreted to have the highest LSR’s.
The landform ages and depositional environments of many lacustrine environments are prime for recovering buried cultural resources. Some of these landforms are merely "uplands," which happen to be surrounded by water on most if not all sides. These uplands have a low potential for containing geologically buried cultural resources at locations distant to water; however, they still have shoreline areas, which could contain buried resources.
The Rainy River Valley project area starts at the Rainy Lake outlet just upstream from International Falls and extends to its confluence at Lake of the Woods, just downstream from Baudette. Above the outlet of Rainy Lake, the Rainy River is essentially a series of connections between closely spaced lakes. The Rainy River is cut into the glaciolacustrine plain of Lake Koochiching, the Beltrami arm of Glacial Lake Agassiz. The river flows to the west, with two jogs to the north. It is typically flanked by steep banks that ascend to narrow terrace remnants or the lake plain directly. The Rainy River potentially has a complex history. Fluctuating lake levels would have shifted positions of the lacustrine nearshore and shoreline facies. Late glacial fluctuations in ice margin positions not only resulted in changes in controlling lake outlets and levels, but could also have reversed directions of stream flow. Finally, with retreat of glacial ice and draining of glacial lakes, isostatic rebound has altered stream gradients.
Detailed investigations of the Holocene and late Pleistocene of the Rainy River are severely limited. Meyer (1993) produced a surficial geologic map of an area south of the Rainy River in Koochiching, Itasca and Beltrami counties. Eng (1980) produced a surficial geologic map of Koochiching County. Notable features relevant to evolution of the Rainy River are beach strand lines of former levels of Lake Agassiz's eastern arm.
The history of lake levels, spillway activity and ice movements in the eastern arm of Glacial Lake Agassiz is well documented by Clayton and Moran (1982), Clayton (1983), Hobbs (1983), Teller and Thorleifson (1983); however, there are only a sparse number of radiocarbon ages upon which to relate to these events. Furthermore, these radiocarbon ages relate to only a few of the above-mentioned events.
Janssen (1968) generated a pollen diagram from a core taken from Myrtle Lake in Koochiching County. It had a series of five radiocarbon ages (Stuiver 1969).
Hajic (1996a) developed a preliminary model of Holocene landscape evolution for the reach of the Rainy River in the vicinity of the mouth of the Little Fork River. Radiocarbon ages and stratigraphic information from the McKinstry archaeological site at the mouth of the Little Fork River (Yourd 1985; Arzigian et al. 1994; Thomas and Mather 1996; Hajic 1996), the Hannaford site at the mouth of the Big Fork River (Hill et al. 1995), and the Canadian side of the Rainy River (Bacj 1991) served as the database for the model. All references are in the "gray" literature, except the Bacj study, which is an unpublished dissertation, and no other detailed investigations exist.
Hajic’s (1996a) chronology highlights: 1) the final phase of Glacial Lake Agassiz drained from the area between about 9,920 and 9,610 B.P., whereas the general literature places it at 9,500 B.P.; 2) meandering stream channel and point bar facies associated with the VT2 LfSA were deposited minimally between 8,550 and 6,900 B.P., but possibly as late as about 3,000 B.P.; 3) older valley terrace LfSA’s (VT3, VT4, VT5) must date between about 9,610 and 8,550 B.P.; 4) downcutting from the VT2 to VT1 terrace level was accomplished by at least about 5,450 B.P. and probably earlier; 5) floodplain LfSA’s were actively building laterally by point bar migration and vertically by accretion of flood drapes from at least about 2,250 to approximately 700 B.P.; and 6) since about 700 B.P. vertical accretion of flood drapes has been limited on surfaces above the active floodplain.
Nineteen cores were collected from the Rainy River valley between September 5 and 7, 1996. All cores were taken with the Gidding hydraulic soil-probe. These cores were logged in a laboratory and are presented in Appendix E.1. The methods discussed in Section 12.2 apply to this field and laboratory work.
126.96.36.199 Cross-Section Profiles
A series of eight cross-sections and one long-section were constructed from the soil profiles of the Rainy River valley. These figures and one drawing are presented in Appendix E.2 and E.3, respectively.
188.8.131.52 Radiocarbon Dates
A variety of organic samples were collected from the cores. A master list of organics and their identifications are presented in Appendix E.4 as Table E.1. These samples yielded three radiocarbon dates from the Rainy River valley, which are presented with their associated data in Appendix E.5 as Table E.2. These radiocarbon dates are also displayed in the above-mentioned cross-sections. These dates are used in conjunction with other dates recovered during earlier MnDOT investigations along the Rainy River.
184.108.40.206 Site-Specific Field Methods and Mapping
There were no unusual field methods applied to the Rainy River field and laboratory investigation.
High altitude color infrared aerial photographs from the USGS NAPP program were the primary source of geomorphic information. Tonal contrasts on these air photographs were sufficient to delineate features even finer than those mapped. The Lake of the Woods Koochiching NRCS soil survey was useful in supplying subsurface textural information. The Koochiching County NRCS soil survey was not yet completed at the time of modeling. Mapping was done independent of previously available maps of the area, partly because of the difference in scale of the pre-existing maps and of the mapping being done, and partly because of the desire to map without reference to existing geologic models of the area. Mapping was accomplished by thin pencil directly on USGS 7.5-minute topographic maps. Many of the geomorphic boundaries were transferred to the maps from mylar overlays of NAPP aerial photography, with appropriate adjustments for distortion of the aerial photographs.
Landform sediment assemblages for the Rainy River are outlined in Table 12.4 and discussed below.
The Valley Terrace and Floodplain LsSA’s were used in mapping the Rainy Valley. The downstream reach also has units of the Upland, Glaciolacustrine and Peatlands LsSA’s mapped in conjunction with the Red Lake Bog model (Section 12.6). Five terraces are recognized in the Valley Terrace landscape. The oldest two are closely related to the glaciolacustrine plain and have only a thin veneer of alluvium. The second-to-oldest exhibits a series of anastomosing channels in one reach well upstream of Baudette that may represent deltaic distributaries. They occur at the same elevation as the foot of a Campbell beach ridge south of the river. The three younger terraces occur as intermittent features. The Floodplain landscape is represented by only a handful of discontinuous surfaces. Parts of a few large tributaries to the Rainy River have been included in this model.
Upland landscape. Areas referred to as "uplands" on the glacial lake plain are typically underlain by till and are undifferentiated (U). There may or may not be a very thin veneer of glaciolacustrine deposits overlying water modified till (UWA).
Glaciolacustrine landscape. The Glaciolacustrine landscape is included for the downstream part of the Rainy Valley only to provide continuity with mapping of the Red Lake Bog (Section 12.6). A glaciolacustrine plain representing the former bed of Lake Agassiz is recognized in this area, along with beach ridge remnants. Laterally continuous fine material or peat overlying fine material underlies the nearly featureless lake plain. Fine glaciolacustrine material may be greater than 2 m thick (APF), less than 2 m thick overlying till (APF<), or overlain by peat with a combined thickness greater than 2 m (APPF). There is also a beach ridge (ASH) that consists of coarse material greater than 2 m thick and is laterally continuous.
Peatland landscape. The Peatland landscape is mapped as an undifferentiated plain (BPP). It consists of peat of unknown thickness. The basement material is also unknown, but is either glaciolacustrine deposits or till.
Valley Terrace landscape Five valley terrace levels exist in the Rainy Valley. The VT4 and VT5 terraces occur as relatively continuous high-level surfaces flanking the modern channel. They are inset slightly below the surrounding lake plain. Normally there is only a gradual rise from the VT4 and VT5 terraces to the lake plain. There are two short reaches where the higher surfaces are modified into streamlined bar-like forms surrounded by multiple well-defined paleochannels. In one example, bar forms span elevations of both the VT4 and VT5 levels. In the other example, they span the VT3 and VT4 levels. These bar-like forms and paleochannels may represent former deltaic distributary channels, or large-magnitude flood channels and bars in both cases. Both examples occur in a situation where the Rainy River abruptly changes course at roughly ninety-degree angles. Sediments for both of the high-level terraces typically form a thin laterally continuous veneer less than 2 m thick over glaciolacustrine material. Some of the paleochannels (VPC5, VPC4) are partially buried by peat (VPC5P, VPC4P, VPC4PF) that post-dates the clastic veneer. Younger type "a" overbank deposits may mantle lower levels of the VT4 surface (VT4A).
The VT3 terrace primarily occurs as a series of narrow benches inset below higher surfaces and cut into the lake plain. In the vicinity of the mouth of the Little Fork River, there is a wider expanse that exhibits a broad subtle swale. Downvalley, the series of bar-like forms and paleochannels occurs immediately downstream of a ninety-degree turn of the river to the north. Typically, a short steep descent exists from higher surfaces to the VT3 terrace. Overall, the base of the VT3 LfSA probably represents a former channel of the Rainy River. The VPC3 LfSA is differentiated on whether there is a fine-over-coarse texture assemblage (VPC3), or a peat-over-coarse sediment assemblage (VPC3PC). Having a veneer of type "a" overbank sediments (VPC3A) further differentiates some downstream examples. If the paleochannel is cut into an older terrace, rather than the glaciolacustrine plain, it is noted as VPC3V. Associated sediments typically are 1 m or less thick, but may range between 1 m and 2 m thick, be laterally continuous, and overlie glaciolacustrine sediments. The sediment sequence primarily consists of fine overbank material, but there can be a basal coarse bed of channel sand present. Swales may have a thin veneer of peat.
The VT2 and VT1 terraces are more typical of alluvial aggradational terraces. They occur sporadically adjacent to the river and exhibit ridge and swale topography. Associated sediment sequences are typically up to about 5 m thick, laterally continuous, and overlie glaciolacustrine material or till. These LfSA’s consist of thinly stratified fine textured overbank deposits over a basal sand and gravel increment of channel deposits. Both surfaces are overlain by overbank deposits, with the VT2 having type "a" and VT1 having type "o" overbank deposits. Both VT2 and VT1 LfSA’s are differentiated on whether underlying deposits are fine textured (VT2FA, VT1FO) or fine over coarse textured (VT2FCA, VT1FCO).
Tributary valleys exhibit multiple terrace levels (VT) that largely remain uncorrelated. They are underlain by fine over coarse material that is undifferentiated as to thickness because it is expected to vary widely. Many undifferentiated paleochannels are present, often in the form of cutoff meanders (VPC). There are two generations of anastomosed paleochannels that flow across higher terrace levels (VPC1V, VPC2V).
Floodplain landscape. This landscape is represented by a few map polygons. Only a couple of these polygons occur upstream of the Manitou Rapids. Floodplains exhibit ridge and swale topography and have a thin (less than 1 m thick) mantle of type "o" overbank deposits (FFXO). Associated sediments consist of laterally continuous fine vertical accretion deposits that are probably at least 2-5 m thick. These associated alluvial sequences overlie glaciolacustrine deposits or glacial drift.
Several large tributaries with substantial valleys cut into the Lake Agassiz glacial lake plain within the investigated reach of the Rainy River. Meandering streams, discontinuous floodplain areas, cutoff meanders, discontinuous remnants of multiple high and low terrace levels, and several paleochannel systems characterize them. The oldest of the paleochannel systems is associated with an abandoned channel belt that represents the tributary floodplain prior to incision of the tributary into the lake plain.
Channel belts (FMB, FMBV) flank the incised meander belts of major tributaries where they are associated with a broad overbank belt and cross terraces. They are typically shallowly incised and have fills usually less than 1 m, but on occasion can be thicker than 2 m. Three distinct paleochannel systems are distinguishable along tributaries (FPC). The oldest has a very low sinuosity similar to the aforementioned abandoned channel belt, but these channels are wider. The two younger paleochannels are from meandering systems. Associated deposits are greater than 2 m thick and laterally continuous. They consist of fine over coarse material with local surface veneers of peat and/or organic muck. The overbank belt LfSA (FOBA) overlies the glaciolacustrine landscape adjacent to incised tributary valleys. It was the floodplain associated with the aforementioned abandoned channel belts prior to tributary incision. Associated deposits are less than 2 m thick. Undifferentiated tributary floodplains (FF) are discontinuous map units adjacent to channels that exhibit ridge and swale topographic relief. For tributaries of the Rainy River, this unit includes low terraces, but not high terrace levels. Deposits are fine textured, laterally continuous and greater than 2 m thick. Peat is rarely present (FFP).
220.127.116.11 Landform Sediment Assemblage Codes
Table 12.4 provides details on each of the specific LfSA codes used for the Rainy River model.
Because the high Emerson Phase of Lake Agassiz did not drain until about 9,500 B.P., the entire post-drainage landscape was young enough to support Late Paleoindian and younger cultural groups. A few undifferentiated areas of upland were surrounded by Rainy River landforms and were mapped. Whereas some are noted as wave modified, they probably all were wave modified to some degree. These uplands are underlain by glaciolacustrine clay or till. A ranking of not possible (0) is assigned to these upland areas below the land surface. This "not possible" ranking is due to the unsuitable depositional environments of glaciolacustrine clay and till.
The LSR’s are not possible (0) for glaciolacustrine clay at depth (APF, APF<, APPF) due to an unsuitable depositional environment. In contrast, sandy shoreline features (ASH) have a moderate (2) ranking for the 0-1 m depth interval and low (1) ranking below that depth. Where thin peat was deposited over glaciolacustrine clay (APPF), buried cultural deposits may be associated with the exposed lake bed clay. For this reason, the 0-1 m depth interval is ranked low (1).
The age of undifferentiated peatland that buries the glacial lake bed in this part of the basin is unknown beyond being younger than about 9,500 B.P., and could vary greatly. Drainage was probably poor across large tracts, but potential locally higher areas now buried by peat could have been utilized prehistorically. Similarly, cultural deposits could be buried by peat. Although the possibility of cultural deposits in a poorly drained setting is extremely slim, it can not be ruled out with the present state of knowledge. For these reasons, the peatland LfSA (BPP) is ranked low (1). Peat formation likely continues; therefore the surface is ranked not possible (0) because of a young age.
The valley terrace LfSA’s are the most likely locations for surface and buried archaeology in the Rainy River Valley region. Uppermost deposits are low energy floodplain overbank sheetflood deposits. Paleochannels associated with the different terraces are generally ranked not possible (0) due to the poor to ponded drainage conditions. Nevertheless, where peat is present in these paleochannels, a low (1) ranking is assigned for the 0-1 m depth interval. The VT5 and VT4 LfSA’s are old enough to have shallowly buried Late Paleoindian and Early Archaic deposits, with younger cultural deposits on the surface. The VT5 LfSA is assigned a moderate (2) rank for the 0-1 m depth interval. Because deposits are typically thin overlying the glaciolacustrine clay or till, the greater than 1 m depth interval is ranked not possible (0). The VT4 LfSA is assigned a high (3) rank for the 0-1 m interval, moderate (2) for the 1-2 m interval, and not possible (0) for greater than 2 m depth. Type "a" overbank deposits are locally present (VT4A), but the rankings remain unchanged. The VT3 LfSA is old enough to have shallowly buried Early Archaic and Middle Archaic deposits, whereas younger cultural deposits may occur on the surface. This VT3 LfSA has a high (3) ranking for the 0-1 m depth interval and a low (1) ranking for the basal 1-2 m depth interval. Some paleochannels clearly have an increment of type "a" overbank deposits (VPC3A). Where present, the surface and 0-1 m depth interval are ranked low (1) as a cautious measure.
The VT2 LfSA is assigned a high (3) rank for the 0-1 m depth interval, and moderate (2) for to the depth of at least 5 m (VT2, VT2FA). Where fine overlies coarse material (VT2FCA), the 0-1 m depth interval is ranked high (3), the 1-2 m depth interval is low (1), and the depths greater than 2 m are not possible (0). Lower rankings are due to the likely channel origin of the sand. The VT1 LfSA usually has a thin increment of type "o" overbank deposits associated with it. Where the VT1 LfSA consists of fine material, it is assigned a high (3) rank for the 0-2 m depth interval, and moderate (2) for the 2-5 m depth interval. If fine over coarse material exists, then, it is ranked as moderate (2) throughout. No paleochannels exist with the VT1 and VT2 LfSA’s, which is in contrast to the older terrace LfSA’s.
Tributary valley terraces (VT) are undifferentiated. Higher, older terraces have thinner fine material over coarse material than do the lower terraces. The VT LfSA ranking is based upon the lower terrace assemblage, which is ranked as moderate (2) for the 0-1 m depth interval, low (1) for the 1-2 m depth interval, and not possible below 2 m. Associated paleochannels (VPC) are ranked not possible (0) for all but the 2-5 m depth interval that is ranked low (1). This latter ranking is precautionary because some of the paleochannels are relatively shallow and may be cut into terrace LfSA’s that have some potential for buried cultural deposits.
The three tributary paleochannel systems that pre-date tributary incision and cross high terrace surfaces and uplands (VPC1, VPC2, VPC3) are distinguished from other tributary paleochannels. Although somewhat shallower than the VPC LfSA, they are ranked similar to the other paleochannels. Overbank belts are associated with these shallow tributary paleochannel systems. They are ranked high (3) for the 0-1 m depth interval, moderate (2) for the 1-2 m depth interval, and low (1) below the 2 m depth interval.
The floodplain LfSA (FFXO), characterized by late Holocene overbank sediments, is ranked moderate (2) for the 0-1 m depth interval, even though there are thin type "o" overbank deposits over much of the surface. The continuity and thickness of the overbank deposits is however unknown. A LSR of low (1) is assigned for the 1-2 m depth interval and not possible (0) for the 2-5 m depth interval due to poor drainage considerations. Tributary valley floodplains (FF, FFP) are assigned higher rankings because low terraces are bound to be included within this LfSA as mapped. The FF and FFP LfsA’s are ranked high (3) or moderate (2) to a depth of 2 m, and moderate (2) below the 2 m depth. The moderate rankings at shallower depths are related to the FFP LfSA where peat is present.
Meander belt LfSA’s (FMB, FMBV) are narrow, but do have local floodplain segments. Although limited in area and undifferentiated from associated paleochannels, they are ranked high (3) for the 0-1 m depth interval and moderate (2) for depth intervals below the 2 m depth interval.
Red Lake Bog of north-central Minnesota is located in the Beltrami arm of Glacial Lake Agassiz This bog was investigated along a north-south transect, conveniently provided by Trunk Highway 72, from its northern limit at Baudette to well south of its southern limit at Blackduck. From north to south, this transect gradually ascends the glaciolacustrine plain for about 15 km where the plain is underlain by relatively thick glaciolacustrine deposits and several beach ridges. Continuing southward, the transect crosses the Rapid River and then ascends to the heart of the bog at approximately the 50 km mark. This location has a relatively thin peat unit overlying some relatively thin glaciolacustrine deposits, which in turn overlies till. The transect descends slightly into the Upper Red Lake drainage basin, centered at about the 60 km mark, and then ascends in a relatively short distance to the 90 km mark on the Erskine Moraine. During this ascent, the transect crosses a series of beach ridges and nearshore glaciolacustrine deposits. From here, the transect descends to cross a ground moraine associated with the Erskine Moraine marked by an aligned, but irregular, topography. Finally, the transect ascends in a relatively short distance to Blackduck at the 118 km mark and another moraine crest. Cores were taken along the length of this transect, but mapping extended only as far south as the foot of the northernmost moraine.
The pre-bog history of glacial lake levels is well documented by Clayton and Moran (1982), Clayton (1983), Hobbs (1983), and Teller and Thorleifson (1983); however, there are only a sparse number of radiocarbon ages that can be related to these events. Furthermore, the ages tend to be clustered and relate to only a few of these events.
Interest and concern in the bog from economic and environmental perspectives have led to the mapping and peat resource inventory of the Red Lake Bog. The program is summarized by the Minnesota Department of Natural Resources (1981, 1984a). Countywide investigations for Beltrami and Lake of the Woods counties, those crossed by Trunk Highway 72, are reported by the Minnesota Department of Natural Resources (1984b).
Descriptions of ecological aspects of the bog are provided by Heinselman (1963), Hoffstetter (1969), Griffin (1977), and Glaser et al. (1981). Modern surface vegetation was compared with surface pollen by Griffin (1975). Peat stratigraphy and radiocarbon-dated bog evolution are discussed for one local area in the bog by Janssens and Glaser (1983). They determined that peat started to form about 5,000 B.P. and the bog underwent several changes in wetland vegetation. Several radiocarbon ages from the Red Lake Bog are reported in Griffin (1977), Stuiver (1969), and Glaser et al. (1981). The geographically limited investigation by Janssens and Glaser (1983) was the only reference found on the radiocarbon dated evolution of the bogs or the surface waterways that drain this bog.
Bedrock topography underlying the bog was reported by Miller et al. (1992) as part of a hydrologic investigation of peat landforms. In the 700 m stretch of his investigation, he found that the bedrock topography varied considerably. This work was continued by Bemis et al. (1994) who found considerable stratigraphic variation beneath the peat and Lake Agassiz deposits. They also investigate the relationship of lateral changes in stratigraphy, for which they have ample evidence, and groundwater flow.
Archaeological sites are not reported for the bog, but do occur around Upper Red Lake (Minnesota Historical Society site files).
Fifty-one cores were collected from the Red Lake Bog between January 8, 1996, and April 10, 1996. All cores were taken with the MnDOT Foundations Department tracked drill-rig and crews. These cores were logged in a laboratory and are presented in Appendix E.1. The methods discussed in Section 12.2 apply to this field and laboratory work.
18.104.22.168 Cross-Section Profiles
A series of three cross-sections (which are continuous and comprise one "long"-section) were constructed from the soil profiles of the Red Lake Bog basin. These figures are presented in Appendix E.2.
22.214.171.124 Radiocarbon Dates
A variety of organic samples were collected from the various cores. A master list of organics and their identifications are presented in Appendix E.4 as Table E.1. These samples yielded six radiocarbon dates from the Red Lake Bog basin. The radiocarbon dates and their associated data are presented in Appendix E.5 as Table E.2. These radiocarbon dates are also displayed in the above-mentioned cross-sections.
126.96.36.199 Site-Specific Field Methods and Mapping
The drilling was accomplished on some of the coldest days of the year. Most of the cores were collected along the snow-packed road sides. The drillers moved as far from the highway and onto the shoulder slope as possible. Most of the cores were collected from the tops of the slopes where earth movement for highway construction of TH 71 was relatively minimal if present at all.
High altitude color infrared aerial photographs from the USGS NAPP program were the primary source of geomorphic information. Tonal contrasts on these air photographs were sufficient to delineate features even finer than those mapped. The Lake of the Woods and Beltrami NRCS soil surveys were useful in supplying subsurface textural information. Mapping was done independent of previously available maps of the area, partly because of the difference in scale of the pre-existing maps and of the mapping being done, and partly because of the desire to map without reference to existing geologic models of the area. Mapping was accomplished by thin pencil directly on USGS 7.5' topographic maps. Many of the geomorphic boundaries were transferred to the maps from mylar overlays of NAPP aerial photography, with appropriate adjustments for distortion of the aerial photographs.
Landform sediment assemblages for the Red Lake Bog are outlines in Table 12.5 and discussed below.
The Upland, Glaciolacustrine, Peatlands, Valley Terrace, and Floodplain LsSA’s were mapped for the Red Lake Bog. The Upland suite applies to slightly higher areas of the bog underlain by till with little or no overlying peat and/or lacustrine material. Where exposed, the till was wave-scoured. Till was also mapped where the incisions of tributary valleys into the glaciolacustrine plain exposed till in their banks and valley walls. Glaciolacustrine plain and largely discontinuous beach ridges are mapped in the Glaciolacustrine suite. This unit also includes areas of coarse textured material deposited in undifferentiated nearshore and shoreline environments. The Peatlands landscape is reserved for extensive and/or thick areas of the bog. Some areas are characterized by distinct peat bog patterns and landforms, but much is mapped as undifferentiated. The Valley Terrace landscape is limited to terraces along the Rapid River with only several other occurrences in other valleys. The Floodplain landscape is a minor unit used for the large streams draining the bog that are not tributaries themselves to the Rainy River. Other stream floodplains are mapped under the Tributary Valley suite, another minor unit in the bog.
Active Ice landscape. Many of the local topographic highs are underlain by a thin veneer of clay to sandy loam overlying loam to clay loam diamicton. The veneer is considered to be Lake Agassiz deposits put down at times when the topographic highs were under high lake stand waters. The underlying diamicton is interpreted as till of the Itasca Moraine. The till is mapped as wave modified (IPWA).
Glaciolacustrine landscape. The glaciolacustrine LfSA’s were the result of Glacial Lake Agassiz. The bog itself is a manifestation of poor drainage on the glacial lake plain. High areas of glaciolacustrine plain do not support peat growth. These high areas most commonly occur along the flanks of till uplands or near Upper and Lower Red Lake, and are differentiated by the texture of the sediment association and its thickness. In general, the plain is underlain by sand to sandy loam that is greater than 2 m thick (APC), less than 2 m thick (APC<), or clay that is less the 2 m thick (APF<). Where the sediment assemblage is less than 2 m thick, the underlying material is loam to clay loam diamicton interpreted as till. The clay is lacustrine in origin. The coarser material could have any number of origins, including shoreline, offshore, fluvial, and deltaic processes, all related to rising or falling shorelines. Where lake levels were more stable, shorelines with beach ridges and other features developed. Many more discontinuous shorelines exist than the classic models of Lake Agassiz’s continuous shorelines suggest. Shoreline LfSA’s are differentiated on whether they are greater than 2 m thick (ASH) or less than 2 m thick (ASH<). These features are underlain by sand to sandy loam material. Peat growth has sometimes buried parts of individual shoreline features (ASHPS).
Peatland landscape. The peatland landscape is characterized by a variety of recognized bog landform patterns. Although peat in the Red Lake Bog is often characterized as greater than 2 m thick for vast expanses, Mn/Model drilling only found thin peat. The current model takes into account the more conservative estimate of peat being greater then 2 m thick. Recognized patterns include arterial drain pattern bog (BAB) (Glaser et al. 1981; Wright and Glaser 1983; Eng 1980), ovoid-shaped bog (BOV) (Heinselman 1963; 1970; Glaser et al. 1981; Wright and Glaser 1983; Minnesota Dept. of Natural Resources 1984b; Eng 1980), raised (radial) bog (BRB) (Heinselman, 1963; 1970; Glaser et al. 1981; Wright and Glaser 1983; Minnesota Dept. of Natural Resources 1984b; Eng, 1980), and ribbed fen (BRF) (Heinselman 1963; 1970; Glaser et al. 1981; Wright and Glaser 1983; Eng 1980). Undifferentiated peatland plain is differentiated based on the best estimate of peat thickness. Peat can be greater than 2 m thick (BPP), less than 2 m thick (BPP<), or discontinuous and thin (BPP-). Where peat is less than 2 m thick, it is subdivided further if it is underlain by fine material (BPPF), coarse material (BPPS), or coarse material over loam to clay loam till (BPPS<). These areas generally occur adjacent to relatively high landscape positions within the bog, or adjacent to the bog margins.
Valley Terrace landscape. The Valley Terrace LsSA is a relatively minor component of the Red Lake Bog area. Terraces (VT) are found in the Cormorant, Sturgeon, and Black rivers within the mapping area. At least two terrace levels are present within a single valley (VT1, VT2), but these terraces are not necessarily correlative between the river valleys. The terraces were not probed, but best estimates are that depths are generally less than 2 m to the Glaciolacustrine lake bed materials.
Floodplain landscape. The majority of the mapped Floodplain LfSA’s in the mapped area are located within the aforementioned streams and river valleys draining the uplands toward the Upper and Lower Red lakes. Larger valleys have meandering streams whereas some of the shorter, steeper valleys have little floodplain area. Associated alluvium is not differentiated by texture, but is generally greater than 2 m thick. Smaller valley floodplains are undifferentiated (FF). Larger valleys tend to have type "x" floodplains (FFX). Paleochannels are also evident and document channel migration and avulsion (FPC).
188.8.131.52 Landform Sediment Assemblage Codes
Table 12.5 provides details on each of the specific LfSA codes used for the Red Lake Bog model.
From an archaeological perspective, the Red Lake Bog is an unknown quantity with tremendous potential for buried sites. Because the high Emerson Phase of Lake Agassiz did not drain until about 9,500 B.P., the entire post-drainage landscape was young enough to support Late Paleoindian and younger cultural groups.
The unsuitable depositional environment of the IPWA LfSA, associated with outliers of the Itasca Moraine, will not have buried cultural deposits; however, lower Itasca Moraine outliers have an interesting surface history of multiple inundations by Lake Agassiz. These surfaces were available for occupation by Early Paleoindian groups, which would have been subjected to both inundation and shoreline processes. Shallow burial by lake sediments or re-worked till is possible. Subsequent to the last inundation of these outliers, they would have been subject to local sheetflood erosion and sedimentation, and biomantle evolution. For these reasons, buried cultural deposits are possible in the 0-1 m depth interval; however, the LSR is low (1) largely because if cultural deposits are present, all but the youngest are likely to be disturbed (no site integrity).
A depositional environment unsuitable for occupation is represented by glaciolacustrine clay and near-shore, coarse textured LfSA’s (APC, APC<, APF<) and warrants a not possible (0) ranking below the surface. In contrast, sandy shoreline features have a moderate (2) ranking for the 0-1 m depth interval, and low (1) ranking throughout the projected thickness of the respective LfSA’s (ASH, ASH<, ASHPS). The shoreline features apply to both Glacial Lake Agassiz, as well as smaller lakes such as Upper Red Lake. Radiocarbon ages from Upper Red Lake beach and near shore sands indicate aggradation ranged from about 5,890 ± 60 (Beta-107074) to later than 2,610 ± 40 B.P. (Beta-107076). For the ASHPS LfSA, peat deposition has encroached on this former shoreline feature, potentially burying cultural deposits associated with it. For this reason, the 0-1 m depth interval is ranked low (1) whereas the 1-2 m interval is ranked moderate (2). Although the likelihood of occupation must be slim for a recently drained glacial lake basin, it is a possibility. The glaciolacustrine landforms are subject to the same possibilities as the IPWA LfSA outliers of multiple inundations, re-sedimentation of existing deposits, and deposition of new material on surfaces previously available for occupation.
After about 9,500 B.P., the Glacial Lake Agassiz basin drained for the last time. A wide expanse of terrain was available for occupation, although large tracts were probably unappealing due to poor drainage. The exposed lake bed did have some relief as indicated by the Itasca Moraine outliers and shown in Figures E-39 to E-41. These irregularities suggest variability in the inception of timing of peat growth, its location, and "sedimentation" on different paleo-landscape positions (some possibly favorable occupation locales). The basal age of peat is unknown across the basin. Basal peat radiocarbon dates in the vicinity of Upper Red Lake include 4,470 ± 50 (Beta-107080), 3,240 ± 40 (Beta-107104), 3,180 ± 40 (Beta-107084), and 180 ± 40 B.P. (Beta-107075) indicating that in this local area, relatively thin peat is late Holocene in age. These dates also indicate that a 4,290 year range exists in the basal age of the peat. Peat growth continues today, and therefore the surface of the bog is Historic in age.
Peat formation processes are conducive to burial and preservation, but poor drainage limits the prehistoric activities on the peatlands; therefore, peatland LfSA’s greater than 2 m thick (BAB, BOV, BPP, BRB, BRF) are ranked low (1) for the thickness of peat, but not possible (0) for the modern landsurface. Where peat is less than 2 m thick (BPP-, BPP<, BPPF, BPPS, BPPS<), peatland LfSA’s are again ranked low (1) only down to 2 m depth. The interface between the peat and underlying deposits is interpreted to be the most critical location in terms of buried cultural deposits.
Little is known about the valley terrace LfSA’s in terms of age and sedimentology. Much depends on deposits and configuration of paleo-landsurfaces underlying the streams. Because the streams that have terraces are essentially inset into the surface of the bog, and we are assuming the bog is still growing, it is likely that terraces, and more likely that floodplains, are at least middle to late Holocene in age. Conservatively, valley terrace LfSA’s (VT, VT1, VT2) are ranked as possible to a depth of 2 m, albeit low (1) at this time.
Floodplain LfSA’s (FF, FFX) are underlain by more than 2 m of Holocene deposits, so they are assigned a rank indicating occupation was possible. Poor drainage conditions in these "valleys", cause a low (1) LSR to a depth of 2 m. The paleochannel LfSA with sediments greater than 2 m thick is ranked not possible (0) in this poorly drained environment.
The Red River of the North (Red River) investigation spanned from the confluence of the Bois de Sioux and Otter Tail rivers near Breckenridge, Minnesota, to the Canadian border, more than 300 km (190 miles) to the north. This "valley" is different in many ways from the other river valleys investigated during the Mn/Model project. For instance, valley margins are not clearly delineated by escarpments, the alluvial strata are dominated by fine-grained sediments, and the river flows straight to the north. Most Minnesotans, North Dakotans, and Manitobans also know that flooding is practically guaranteed every year. The unique qualities of the Red River valley impacted the mapping and the assigning of "weights" to the Landscape Suitability Models. The "valley margin" ended for Mn/Model where the Red River’s alluvium ended.
The Glacial Lake Agassiz basin and its geological history have been investigated in far greater detail than any of the Red River alluvial strata. One Lake Agassiz researcher has even called the post-glacial lake strata (mostly alluvium) a "hindrance" to their studies. State-of-the-art publications on Glacial Lake Agassiz include, for example, Teller et al. (1996), Teller (1985), Teller and Thorleifson (1983), Harris et al. (1996, 1995, 1974), Fenton et al. (1983), and Clayton (1983). Reid and Olson (1996, 1994) described the post-glacial Red River alluvium (Oahe Fm.) from cores near the town of Halstad, Minnesota. These authors indicate that the Red River entrenched into the lake plain as early as 9,000 B.P. and began to aggrade. At 7,500-5,000 B.P. during the Altithermal, a depositional hiatus took place. Michlovic (1985, 1986, 1987) reports radiocarbon dates of approximately 3,400 B.P. from an unpublished buried archaeology site (less than 2 m depth) at Halstad, and between 4,000 and 3,000 B.P. from the Canning Site (approximately 1 m depth) near Hendrum, Minnesota. Other dates may exist in the "gray" literature of either cultural resource or geotechnical projects. A dearth of publications exists for the Red River strata and its geological history.
Fifty-four cores were collected from the Red River of the North "valley" between June 24, 1996, and October 19, 1996. All cores were taken with the MnDOT Foundations Department tracked drill-rig and crews. These cores were logged in a laboratory and are presented in Appendix E.1. The methods discussed in Section 12.2 apply to this field and laboratory work.
184.108.40.206 Cross-Section Profiles
A series of 21 cross-sections and one long-section were constructed from the soil profiles of the Red River of the North "valley." These figures and one drawing are presented in Appendix E.2 and E.3, respectively.
220.127.116.11 Radiocarbon Dates
A variety of organic samples were collected from the various cores. A master list of organics and their identifications are presented in Appendix E.4 as Table E.1. These samples yielded ten radiocarbon dates from the Red River of the North "valley." The radiocarbon dates and their associated data are presented in Appendix E.5 as Table E.2. These radiocarbon dates are also displayed in the above-mentioned cross-sections.
18.104.22.168 Site-Specific Field Methods and Mapping
Whereas the flooding season made logistics important both before and after the flooding seasons, there were no unusual methods applied to the Red River investigation.
Although both sides of the river had to be mapped in pencil to help with correlations, only the Minnesota side is digitized. Mapping of the outermost alluvium away from the Red River is determined by the aerial photographs and soil surveys, which vary in their own quality. Typically, where the iceberg drag lines and other linear features are distinct at the surface, the alluvium is interpreted to be thin if present at all. Where the iceberg drag lines and other lineaments are missing on the lake plain surface, the alluvium is assumed to be thicker than 1 m.
High altitude color infrared aerial photographs from the USGS NAPP program were the primary source of geomorphic information. Tonal contrasts on these air photos were sufficient to delineate features even finer than those mapped. The Wilkin, Clay, Norman, Polk, Marshall, and Kittson NRCS soil surveys were useful in supplying subsurface textural information. At the time of field survey and mapping, the Polk soil survey was in an unproofed advance sheet stage, and was provided as photocopies to the modeler. Clearly different mapping models were employed by the soil scientists of old and new, as they recognized or at least delineated different soil patterns and characteristics. Where the soil survey models did not match near county borders, the model had to rely more on the aerial photographic signatures to maintain continuity betwen counties. Mapping was done independent of previously available maps of the area, partly because of the difference in scale of the pre-existing maps and of the mapping being done, and partly because of the desire to map without reference to existing geologic models of the area. Mapping was accomplished by thin pencil directly on USGS 7.5' topographic maps. Many of the geomorphic boundaries were transferred to the maps from mylar overlays of NAPP aerial photography, with appropriate adjustments for distortion of the aerial photographs.
Landform sediment assemblages for the Red River of the North are outlined in Table 12.6 and discussed below.
The Glaciolacustrine, Paleo-Valley, Floodplain, and Valley Margin LsSA’s are used in the mapping of the Red River study area. Glaciolacustrine LfSA’s dominate the mapping of the "valley." One Paleo-Valley LfSA is mapped, which probably represents an ancient course that is correlative with a higher floodplain. Several different surfaces and their underlying strata are included/classified in the Floodplain landscape. The Floodplain could have been broken into a Valley Terrace landscape, however the strata are similar and there is a need to recognize the lateral expanse of most Red River floods. These Floodplain LfSA’s have post-Euroamerican settlement overbank material overlying older vertical accretion deposits. The older strata are often within 1 m of the land surface, which makes the Red River Floodplain LsSA different from most of the other Floodplain LsSA’s investigated during Mn/Model. The Valley Terrace landscape includes tributaries that are relatively similar to the Red River landscapes, although they are not differentiated for Mn/Model. The Valley Margin landscape is a relatively minor constituent, although there may be large slump blocks that are mapped as the Floodplain landscape (see text in Table 12.6).
Glaciolacustrine landscape. This landscape refers to Glacial Lake Agassiz landforms and deposits. Two major landforms have been mapped at this scale. They are the lake plain itself and compaction ridges. The Glaciolacustrine landscape has low relief within the project area. A coarsening upward sequence is typical within the uppermost lake plain sediments, which probably represents the waning stages of lake recession from deeper waters to shoreline. The glacial lake sediment ranges from silty clay to clay textures. Low energy, vertical accretion deposits are the dominant strata in this landscape assemblage. Coarse textures are scarce, and the distinguishable type "a" overbank deposits have unknown ages. Excavations at several archaeology sites are contributing useful temporal data. Some of these currently interpreted overbank deposits could be more closely related with the withdrawal of Glacial Lake Agassiz. Their mapped distribution, however, indicates that these deposits are more closely related to the flooding of the current Red River and its tributaries. The distinction between areas with and without overbank deposits was based upon color-infrared aerial photographs, NRCS soil series maps and photographs, and USGS topographic maps. Not all NRCS county soil series maps recognized the overbank deposits near the confluence of streams, which indicates that the changes are subtle even at their scale of mapping (typically 1:20,000). Overbank thicknesses were based upon some coring and also the ability to recognize the surficial patterns on photographs that are common to the Glaciolacustrine landscape, such as iceberg drag lines, and other lineaments. The basement material for all the LfSA’s in this landscape is interpreted to be Glaciolacustrine although there are certainly patchy areas within the project area where glacial till is recognized. The plain landscape has lower relief than the compaction ridges and is further distinguished into individual LfSA’s by having one of the following sets of traits: no significant surficial patterns and greater than 1 m thick of type "a" overbank (APA); flood scour channel patterns and greater than 1 m thick of type "a" overbank (APFA); linear, reticulated, or orbicular surface patterns and laterally discontinuous or absent overbank deposits (API); or, linear, reticulated, or orbicular surface patterns and less than 1m-thick of type "a" overbank deposits (APIA). The compaction ridge LfSA (ACRA) has relief up to 2-3 m above the Glaciolacustrine plain. This ridge has less than 1 m of type "a" overbank deposits mantling glacial lake deposits, which overlie ancient fine and coarse-grained fluvial deposits that are usually attributed to the Poplar River Formation. The beach ridge LfSA (ASH) has relief of approximately 6-10 m above the Glaciolacustrine plain. This landform is currently interpreted as a beach ridge although it was not tested during the Mn/Model field work, and may be a compaction ridge or buried end moraine. A mantle of fine glaciolacustrine sediment apparently overlies coarse-grained beach deposits.
Paleo-Valley landscape. This landscape is an "ancient" channel of the Red River that may still be used during high water stages. Less than 2 m of both type "a" overbank deposits and fine alluvium exist within this channel, which in turn overlies the Glaciolacustrine basement materials. The "v"-shaped valley (YV) LfSA’s are small abandoned tributaries to the Red River or its larger tributaries. These "v"-shaped tributaries are generally erosional features and therefore have less than a 2 m thick alluvial sequence overlying undifferentiated glacial drift. Sedimentary textures may be variable, but they are commonly fine over interstratified fine and coarse deposits. The RYV LfSA’s are a tributary’s "v"-shaped paleochannels that are incised into the Red River’s high Floodplain surface. Because the stream order codes (Code No. 10) for the entire state have not yet been implemented at this stage of work, a distinction needs to be made that these paleochannels were related to the tributaries and were not from the Red River; therefore, the implementation of the Paleo-valley code helps distinguish between the trunk stream and the tributary, and also between the recently active tributary paleochannels and these higher "v"-shaped tributary paleochannels. The YFN LfSA’s are considered former valleys of the Red River or its major tributaries, and have a greater than 2 m thick combined fine over interstratified fine and coarse deposits, which overlie an undifferentiated glacial drift basement.
Valley Terrace landscape. The Mn/Model LfSA code is designed to eventually incorporate the stream orders so that streams can be differentiated relative to each other. For now, the stream orders have not been calculated, which makes labelling the tributary valleys and tributary paleo-valleys awkward. The coding system maintains flexibility to convey the important information. The "v"-shaped valley LfSA (VVE), although not a terrace per se, is lumped with the Valley Terrace landscape for classification purposes. Generally these "v"-shaped LfSA’s are erosional features and therefore have less than a 2 m thick alluvial sequence overlying undifferentiated glacial drift. Sedimentary textures may be variable but are mostly fine over interstratified fine and coarse deposits. The valley with undifferentiated floodplain and terraces (VFN) have a greater than 2 m thick sequence of combined fine over interstratified fine and coarse deposits, which overlie an undifferentiated glacial drift basement.
Floodplain landscape. This landscape has at least three different aged LfSA’s, two of which could have been split off and incorporated with the Valley Terrace landscape. The decision to group these with the Floodplain LsSA was based upon the extreme expanse and frequency of floods along the Red River. The different aged terraces within this landscape do not always line up neatly in elevation, which indicates that some of the "terraces" are slump blocks. Regardless, the slump blocks have received flood sedimentation through time and may contain sediments of the proper age to preserve cultural resources. This LsSA includes types "x" and "y" basic landforms, which translates to "inactive channel migration features present," and "channel migration features not present," respectively. Paleochannels are, however, found on most of the geomorphic surfaces. Some of the topographically higher paleochannels may be caused by flood water scouring, and might actually be younger than topographically lower surfaces and paleochannels. All of the Floodplain LfSA’s have type "a" overbank deposits, which includes alluvium that is older than post-settlement age (for the Red River project area). Buried soils are common to all LfSA’s in this landscape. The strata are mostly fine-grained vertical accretion deposits, although lenses of coarser-grained materials are occasionally found. The basement materials are all undifferentiated glacial drift, which is usually of glaciolacustrine origin, and less frequently ice-contact origin. The Red River project area is probably the only Mn/Model geomorphology river project that consistently has strata older than post-Euroamerican settlement age within one meter of the current floodplain surface. The dominant vertical accretion processes caused by the relatively low energy during flooding, and the relatively young age of the Red River Valley make the Red River Floodplain landscape a prime area for deeply buried cultural resources.
The tributaries to the Red River are composed of both small "v"-shaped valleys and larger valleys with undifferentiated floodplains and terraces. All the tributary valleys are interpreted to overlie glacial drift, which is mostly glaciolacustrine; however, glacial till may be present in areas. The tributary valleys have relatively high potentials for geologically buried cultural resources for the same reasons as the Floodplain landscape mentioned above. Conditions are similar for both landscapes except for erosion in "v"-shaped valleys.
The LfSA’s are distinguished as follows: The second next to youngest surface, type "y" Floodplain (FFY3) has less than 2 m depth to undifferentiated glacial drift basement. The next to youngest surface, type "y" Floodplain, does not have evident channel migration features. This LfSA is further distinguished by having one of the following sets of traits: less than 2 m depth to undifferentiated glacial drift basement (FFYA2); or less than 1 m depth to undifferentiated glacial drift basement (FFYA2<). The next to youngest, type "x" Floodplain has evident channel migration features that are not regularly active. This LfSA is further distinguished by having one of the following traits: meandering channel pattern and less than 2 m depth to undifferentiated glacial drift basement (FFXSA2); or, meandering channel pattern, and less than 1 m depth to undifferentiated glacial drift basement (FFXSA2<). The next to youngest paleochannel (FPC2) is at or near the same level as the other FFX2 or FFY2 landforms and has greater than 2 m depth to the glacial drift basement. These paleochannel deposits may include patches of peaty soils. The youngest floodplain surface may be distinguished by 1) having channel migration features that are not evident with less than 1 m depth to the glacial drift basement (FFY1); or 2) if channel migration features are evident but not regularly active, then this LfSA is further distinguished by having one of the following sets of traits: meandering channel pattern, and less than 2 m depth to undifferentiated glacial drift basement (FFXSA1); or, meandering channel pattern, and less than 1m depth to undifferentiated glacial drift basement (FFXSA1<). The youngest paleochannel (FPC1) has greater than 2 m depth to the undifferentiated glacial drift basement. These paleochannel deposits may include patches of peaty soils. The last of the LfSA’s is the active river channel (FR).
Valley Margin landscape. This LfSA was neither tested nor examined in the field and may be a product of coincidence with the contour lines near some drainageways. If accurate, then this LfSA is rare at the 1:24,000 scale in the Red River Valley. The potentials for deeply buried cultural resources is high because of the depositional environment and its Holocene age.
22.214.171.124 Landform Sediment Assemblage Codes
Table 12.6 provides details on each of the specific LfSA codes used for the Red River valley model.
Although ancient cultures could have followed ancient shorelines back and forth across the landscape, this model has systematically chosen to include the less distinguishable shorelines with the more common offshore deposits, which means the majority of the Glaciolacustrine landscape is regarded as inhospitable to human life. Although the age of the glacial lake deposits are correct for human occupation, the habitat remains unsuitable. The overbank deposits, alluvium, and other sedimentary materials that may mantle the Glaciolacustrine landscape, do have potentials for containing or burying cultural resources. All but the API LfSA has received some sort of type "a" overbank. Moderate (2) LSR’s are assigned to 1 m depth for the APA, APIA, ACRA, and ASH LfSA’s, which is interpreted to be the maximum depth of overbank sediment. The APA and APFA are interpreted to have greater than 1 m thickness of type "a" overbank deposits based upon the absence of lineaments at the surface; therefore these LfSA’s have been assigned a low (1) LSR for the 1-2 m depth interval.
The Paleo-Valley LsSA was created within the time span of potential cultural resources for Minnesota. The RYV and YV LfSA’s, however, have low potentials for geologically buried cultural resources because of the depositional environment being relatively "wet." The location of "sites" may be more probable immediately next to these abandoned cuts, depending on the interpreted adjacent depositional environment. The YFN LfSA has a greater potential because of conservative estimates regarding undifferentiated terraces and floodplain.
The Valley Terrace LfSA’s also have a conservative estimate for their undifferentiated terraces and floodplain (VFN). The "v"-shaped valleys do have some peaty growth and fill materials within them; however, they are ranked low (1) because of the poor drainage conditions.
The Red River project area is probably the only Mn/Model geomorphology river project that consistently has strata older than post-Euroamerican settlement age within 1 m of the current floodplain surface. The dominant vertical accretion processes, caused by the relatively low energy during flooding, and the relatively young age of the Red River Valley make the Red River Floodplain landscape a prime area for deeply buried cultural resources. The surfaces with less obvious recent channel migration activity are assigned higher LSR’s (all FFY LfSA’s), and their depths of high rankings depends on the interpreted depths of the overbank and other potential fill materials. The FFX LfSA’s all have been assigned a LSR of moderate (2) because of the potential for more recent disturbances caused by channel activities to the geomorphic surface. Their depth potentials also depend upon the interpreted depths of overbank deposits.
The MAFA LfSA is questionable as an alluvial fan; however, if it exists then the age and depositional environment causes the LSR to be high (3) to a depth of 2 m and moderate (2) to a depth of 1 m.
The Rock River mapping project spanned more than 80 km (50 miles) from north to south within the southwestern corner of Minnesota. The Rock is the only major river of Minnesota belonging to the Missouri River drainage system. The headwaters of the river start at the crest of the Coteau des Prairies (Coteau), which is a wedge-shaped plateau that stands above the surrounding plains of southwestern Minnesota. Glacial end moraines are situated at this crest and are responsible for the outwash that initially carved the Rock Valley. The Rock River is now an underfit stream. The general region appears to be well drained, as indicated by the absence of natural lakes.
Wright (1972:576) describes this section of Minnesota as the "Coteau des Prairies, inner part." He mentions that the area is covered with loess that mantles tills of both pre-Wisconsin and Wisconsin age. The loess thickens to the southwest and probably has a source from the outwash deposits of the Sioux River. The Flandreau, Rock, and Kanaranzi river valleys carried meltwaters and sediment from the margin of the Des Moines lobe across the region. More detailed publications and "gray" literature regarding the evolution of the Rock River alluvial sequences are apparently scarce, if present at all. The Minnesota DNR (1998) and Patterson (1997) have mapped the Rock valley as an outwash channel emanating from the Bemis moraine.
Forty-five cores were collected from the Rock River valley between June 26 and July 2, 1996. All cores were taken with the Giddings hydraulic soil probe. These cores were logged in a laboratory and are presented in Appendix E.1. The methods discussed in Section 12.2 apply to this field and laboratory work.
126.96.36.199 Cross-Section Profiles
A series of seven cross-sections and one long-section were constructed from the soil profiles of the Rock River valley. These figures and one drawing are presented in Appendix E.2 and E.3, respectively.
188.8.131.52 Radiocarbon Dates
A variety of organic samples were collected from the various cores. A master list of organics and their identifications are presented in Appendix E.4 as Table E.1. These samples yielded five radiocarbon dates from the Rock River valley. The radiocarbon dates and their associated data are presented in Appendix E.5 as Table E.2. These radiocarbon dates are also displayed in the above-mentioned cross-sections.
184.108.40.206 Site-Specific Field Methods and Mapping
There were no deviations from the normal operating procedures within the Rock River investigation. High altitude color infrared aerial photographs from the USGS NAPP program were the primary source of geomorphic information. Tonal contrasts on these air photographs were sufficient to delineate even features finer than those mapped. The Rock and Pipestone NRCS soil surveys were useful in supplying subsurface textural information. Mapping was done independent of previously available maps of the area, partly because of the difference in scale of the pre-existing maps and of the mapping being done, and partly because of the desire to map without reference to existing geologic models of the area. Mapping was accomplished by thin pencil directly on USGS 7.5' topographic maps. Many of the geomorphic boundaries were transferred to the maps from mylar overlays of NAPP aerial photography with appropriate adjustments for distortion of the aerial photographs.
Landform sediment assemblages of the Rock River Valley are outlined in Table 12.7 and described below.
The Upland, Pediment, Glaciofluvial, Paleo-Valley, Valley Terrace, Floodplain, and Valley Margin LsSA’s are used for the Rock River mapping project. The Upland landscape is undifferentiated glacial drift and bedrock, the latter of which is mostly Sioux Quartzite in this region. The Pediment landscape is unique to the mapping within the Mn/Model geomorphology project. The Pediments often appear to be fans, however, they are typically only several meters thick before reaching the basement material. The Pediment LsSA’s were recognized mostly from road-cut outcrops near the Iowa border. Pediments have also been independently and concurrently recognized in the Floyd River valley of northwestern Iowa (Mandel, 1997). The Glaciofluvial landscape dominates the mapping of the Rock River valley. Two geomorphic surfaces are recognized within the mapping area. The Paleo-Valley LsSA could have been lumped with the Glaciofluvial LsSA, but was broken apart in case of different origins. The Valley Terrace LsSA is geomorphically higher than the Floodplain LsSA and inset into the Glaciofluvial LsSA. Sometimes these LsSA’s are very close in elevation. The Valley Terrace landscape probably receives occasional flooding, but less frequently than the Floodplain landscape. The Floodplain landscape is relatively small in area compared to the Valley Terrace and Glaciofluvial landscapes. The Valley Margin landscape is again relatively small in area, but is dominated by more colluvial slopes than by alluvial fans. Most of the colluvium is loess and reworked loess.
Upland landscape. The Upland landscape is comprised of undifferentiated highlands outside of the valley flanks.
Pediment landscape. The Pediment LsSA is inset beneath the uplands and is therefore younger than the local tills. A Pediment LsSA is a long slope composed of a relatively thin unit of transported sediment that overlies an erosional surface from which the sediment was derived. Pediment and Valley Margin LsSA’s may be confused or misinterpreted in a smaller-scaled investigation. The initial distinguishing factor that worked in the Rock Valley was comparing the apparent volume of the "fan" or slope relative to the volume of missing sediment removed from the feeder valley or gullies. These two volumes should either balance, or the "fan’s" volume should be smaller (assuming that the fan has been eroded by the trunk stream through time). The Pediment slope’s (or "fan’s") volume in this study area appeared greater than what the upstream feeder valleys and gullies had lost. Thus, the initial suspicions were raised about how this LfSA developed.
Parts of the Pediment slope LfSA’s are cut into the tills and possibly bedrock at the proximal ends of the pediments. The pediments are also cutting into the higher outwash terrace near the Iowa border, and grade onto the youngest outwash terrace. The Pediment Slope LfSA’s are shaped like an alluvial fan or colluvial slope. These LfSA’s are mantled by fine alluvium, colluvium, or loess. The textural sequence is typically fine over coarse, and is usually less than 2 m thick before reaching the basement outwash, till, or bedrock. This LfSA is further distinguished by having one of the additional traits: type "a" overbank (possibly colluvial) deposits that may or may not be post-settlement in age (PPDDA); or, a loess mantle that is less than 1 m thick (PPDDL). The overbank deposits appear to be reworked loess. Distributary patterns are common to the Pediment surface, and were not mapped as separate LfSA’s at the 1:24,000 scale. Most of this Pediment LsSA was identified in 1997 by examining road-ditch excavation work south of Interstate 94 and west of the Rock River. All the recognized Pediments were on the west side of the Rock River Valley.
Glaciofluvial landscape. The glaciofluvial landscape has two different geomorphic surfaces that are traceable throughout the Rock River valley. The higher of these two surfaces includes both plains (OPL2, OPL2<) and terraces (OTL2<), whereas the lower surface includes both terraces (OTA1, OTL1, OTL1<) and paleochannels (OPCL1, OPCOL1). The higher terrace and plain are subjectively distinguished by their lateral expanse. Some plains are dissected by younger valleys making them appear small in size. The Pediment slopes apparently post-date the higher and older glaciofluvial surface in the Rock Valley because it truncates the higher outwash surface, and grades to the lower outwash surface. The older glaciofluvial surface grades up to an end moraine on the Coteau des Prairies. The younger outwash surface is inset beneath the higher and older plain at the end moraine’s front and was not recognized within the channel breaching the end moraine, although the channel had to be formed or forming. Other glaciofluvial channels (now occupied by underfit tributary streams similar to the Rock River itself) enter the Rock River valley mostly from the east off the Coteau. The glaciofluvial textures are all coarse with a fine mantle of either loess (all LfSA’s but OTA1) or alluvial overbank sediments (OTA1) derived from a tributary. The loess apparently fines from silt loam to silty clay loam in the upstream direction and is typically less than 1 m thick. OTL1 and OPL2 LfSA’s have less than 2 m combined thickness of loess and outwash overlying bedrock or glacial drift. Braided channel patterns are apparent on many surfaces. Tributary paleochannels may also mark the broader outwash surfaces.
Paleo-Valley landscape. The main Paleo-Valley mapped is apparently related to the glaciofluvial events derived from the end moraine on the Coteau des Prairies, and for now this Paleo-Valley is correlated to the youngest outwash surface. The Paleo-Valley landscape may have younger landscapes overlying it such as the Floodplain landscape, as is the case here; however, the Paleo-Valley LsSA was more practical to help distinguish from the Rock River valley floodplain. This Paleo-Valley landscape has equal or greater chances than the main valley of the Rock River to contain suitable landscapes for preserving cultural resources. The depositional environments and ages have been documented through coring and radiocarbon dating (Appendix E). The main reason for these equal or better chances is because this landscape has been inset by younger tributary deposits. This LfSA group is further distinguished by having one of the following sets of traits: less than 1 m thick loess over coarse outwash (YTL1<); greater than 2 m thick loess over coarse outwash (YTL1); or, meandering channel patterns related to post-glacial tributaries to the Rock River, and greater than 2 m thick fine over interstratified fine- and coarse-textured strata (YTSA1). The 2+ m thick fine textures at the surface are overbank deposits that may be older than post-Euroamerican settlement age.
Valley Terrace landscape. This LsSA is primarily inset beneath the youngest outwash terrace, and is overlying part of the Glaciofluvial landscape. The Valley Terrace LsSA may actually have either a steep escarpment between the lowest outwash terrace and itself, or a very gentle slope. The textural sequence of the entire Valley Terrace LsSA is composed of fine alluvium over interstratified fine and coarse alluvium. Buried soils were rarely recognized and when they were recognized they were typically within 1.5 m of the surface. The best chances for deeply buried sites are within the uppermost 2 m of depth because the relatively higher energy environments apparently created the lowermost strata below 2 m depth.
The tributary valleys include both small "v"-shaped valleys (VV) and larger valleys with undifferentiated floodplains and terraces (VFN). Tributary overbank belts are mapped as part of the Floodplain LsSA and were distinguished where the expanse and thickness are sufficient to map. Tributary paleochannels are also delineated and mapped as part of the Floodplain LsSA. Some of the wider tributary valleys were created by a glaciofluvial event(s) related to the end moraine(s) on the Coteau des Prairies. All tributary valleys are interpreted to overlie glacial drift or reworked glacial drift where mapped.
The Valley Terrace LsSA is divided into four LfSA groups based upon their Landscape (Code 5), Landform (Code 7), and Relative Age (Code 25). The "v"-shaped tributaries (VV) are generally erosional features and therefore have less than 1 m thick (sometimes 2 m thick) alluvial sequences overlying undifferentiated glacial drift. The glacial drift is mostly glaciofluvial within the main valley confines, and glacial till (possibly bedrock) outside the main valley confines. Sedimentary textures are variable. The undifferentiated valleys and floodplain LfSA’s have greater than 2 m of combined fines over interstratified fine and coarse textures, which overlies an undifferentiated glacial drift basement. This LfSA group is further divided into LfSA’s that have one of the following sets of traits: an undifferentiated surficial pattern and mantling deposit (VFN); or, a meander channel belt with type "o" overbank deposits, which are mostly post-Euroamerican settlement in age (VFNSO). The youngest terrace LfSA (VTA) has less than 2 m depth to outwash. The youngest paleochannel (VPC) is inset into the VTA LfSA mentioned above, and has a greater than 2 m thick fine over interstratified fine- and coarse-textured stratigraphic sequence. Type "a" overbank deposits mantle this LfSA, and is interpreted to be older than post-Euroamerican settlement age.
Floodplain landscape. The entire Floodplain landscape was relatively small in square area within the Rock River valley, which made distinguishing the different floodplain types difficult at the 1:24,000 scale. The floodplain types were small but dominated by type "w," and were therefore lumped together with the type "w" floodplain. Tributary overbank belts are mapped as part of the Floodplain LsSA and were distinguished where the expanse and thickness are sufficient to map. Cut-off tributary paleochannels are also delineated and mapped as part of the Floodplain LsSA.
The main characters of the type "w" floodplain (FFW) includes active or recently active point bars and channel migration features. These features include meander belts (FFWSO) and oxbow lakes. The landscape may be inset or overlie the Valley Terrace landscape. Type "o" overbank deposits are common and are mostly of post-Euroamerican settlement age. Fine textures (i.e., loam or finer as defined for the Rock River Valley) mantle interstratified fine- and coarse-textures, which overlies coarse glacial outwash. The Rock River alluvium is commonly greater than 2 m thick. The depositional environment is of mixed energies, which means that there are mixed potentials for containing preserved cultural resources within the Floodplain landscape. The overbank LfSA is distinguished by type "a" overbank deposits which may or may not be post-Euroamerican settlement age, fine over interstratified fine and coarse textures, and an undifferentiated glacial drift basement. This LfSA group is further divided into LfSA’s by having alluvium greater than 2 m thick (FOBA); or an abandoned overbank belt with a thickness less than 2 m (FOBA<). The paleochannel LfSA (FPC) has variable textures and is interpreted to be mostly greater than 2 m thick.
Valley Margin landscape. This LsSA is composed of both alluvial fan and colluvial slope LfSA’s. Both of these landforms overlie the Glaciofluvial LsSA. All LfSA’s in the Valley Margin LsSA have fine textures greater than 2 m thick, and commonly have buried soils. The fine textures result from the downslope movement of reworked loess. This landscape can be confused with the Pediment landscape if not examined in outcrop or by coring.
Alluvial fans (MAF) are not common in the Rock River valley at the 1:24,000 scale of mapping and those that were recognized, were not "available" for field investigations during the Mn/Model geomorphology fieldwork. Most of the data were therefore gathered from NRCS soil series. The fans that were recognized were typically greater than 2 m thick and found on or interfingered with the youngest outwash terrace. Colluvial slopes are the more common Valley Margin landform. This LfSA is further distinguished by having one of the following sets of traits: undifferentiated mantle (MC); or, a loess or reworked loess mantle (MCL).
220.127.116.11 Landform Sediment Assemblage Codes
Table 12.7 provides details on each of the specific LfSA codes used for the Rock River valley model.
The Pediment LfSA surfaces (PPDDA, PPDDL) grade to the youngest outwash terrace surfaces (OTL1, OTL1<); therefore, the deepest underlying strata are probably too old to contain cultural resources since the younger outwash surface is probably equal in time to the Bemis or Altamont moraines. The age of the stratigraphically higher (mantle) sediments may, however, be young enough to contain in situ cultural resources. The depositional environment could be analogous to an alluvial fan, but is probably more like an erosional surface from the appearances of the outcrops. A low (1) landscape suitability ranking (LSR) was assigned for this model to depths of the underlying basement material (<2 m). The glaciofluvial depositional environments are interprted to be too high in energy to preserve cultural resources. Furthermore, the age of the outwash is interpreted to be too old for cultural resources here at the perimeter of the Des Moines lobe. The best chances for geologically buried cultural resources are within the strata mantling the glaciofluvial sediment, which were deposited under lower energy environments on top of the Glaciofluvial landscape. Loess, in particular, and any overbank may be young enough for Paleoindian and possibly younger cultures. Whereas terrace ages are suitable for containing cultural material, the potential for preserving in situ cultural resources is low due to the range of depositional environments (energies) found on the terraces. The paleochannels (VPCA) have low to nil potentials for suitable landcapes because even the type "a" overbank has varying textures (i.e., depositional energies). The floodplain depositional environments are of mixed energies, which means that there are mixed potentials for containing preserved cultural resources within the Floodplain LfSA. The floodplain overbank is a type "a" deposit, which may or may not be post-Euroamerican settlement age. Because of the variable energies and the potential for variable time spans on the type "a" overbank deposits, a conservative estimate is presented herein, and a high (3) LSR is assigned to those LfSA’s lacking recent channel migration activities and type "a" overbank deposits (FOBA, FOBA<); whereas the the more recently active channeled landforms have a moderate (2) LSR at depths below the Historic overbank deposits (FFWSO), and a low (1) LSR for the actual channel itself (FPC, FR). The Valley Margin LfSA’s are interpreted to be of the right age and depositional energy to contain geologically buried cultural resources (i.e, to have favorable Landscape Suitability Rankings). The alluvial fans tend to be ranked high (3) and the colluvial slopes tend to be ranked moderate (2) to depths greater than 5 m. The difference between the two lies in the fact that the slopes are being prone to more slump from the valley walls versus the fans receiving more alluvium.
The entire length of the Root River valley in southeast Minnesota was mapped, from Chatfield to the valley mouth at the Mississippi River valley. Mapping was extended further to incorporate the North Branch from the dam impounding Lake Florence at Stewartville to Chatfield, the Middle Branch from the R13W-R12W range line to Chatfield, and from the Mower-Fillmore County line to the Root River at Lanesboro. The Root River and its branches drain an area of dissected uplands not covered by late Wisconsin glacial ice. The valleys of the Root River and its branches upstream of Peterson are narrow, meandering, and cut into bedrock. Insides of meander bends typically exhibit one or more terrace levels. The valley widens locally where rock-core cutoff meanders and high erosional straths are present. Downstream of Peterson, the valley and floodplain widen. Rock-core meanders and valley wall meander arcs are common features that cause further local widening and the preservation of broad terrace remnants. The Root River is channelized from Rushford to the valley mouth. The floodplain is wide and exhibits a range of abandoned channels that increase in frequency downstream.
Wright (1972) labeled the southeastern portion of Minnesota as the Rochester Till Plain. This includes the Root River valley and its tributaries. Hobbs (1995) has mapped the uplands and lowlands of this area at a smaller scale and for a different purpose than the Mn/Model geomorphic mapping project. Other workers have attempted to interpret both the earliest and deepest incisions of the southeastern Minnesota river valleys (e.g., Lively and Olsen 1986; Lively and Alexander 1985; Hobbs 1985). Few have given attention to the valley fills at a scale of mapping they warrant. Mason (1995) and Mason and Knox (1997) report on radiocarbon dates from the colluvial slopes within the Root River valley and its lowermost tributaries. Their dates indicate that mass wasting was prevalent in the Root valley between 29,000 and 12,500 B.P. Numerous "gray" literature must exist on the archaeology of the Root River valley. These data were, however, not accessible for this report.
Eighty-nine (89) cores were collected from the Root River valley between July 8 and August 23, 1996. All cores were taken with the Giddings hydraulic soil probe. These cores were logged in a laboratory and are presented in Appendix E.1. The methods discussed in Section 12.2 apply to this field and laboratory work.
18.104.22.168 Cross-Section Profiles
22.214.171.124 Radiocarbon Dates
A variety of organic samples were collected from the various cores. A master list of organics and their identifications are presented in Appendix E.4 as Table E.1. These samples yielded thirteen radiocarbon dates from the Root River valley. The radiocarbon dates and their associated data are presented in Appendix E.5 as Table E.2. These radiocarbon dates are also displayed in the above-mentioned cross-sections.
126.96.36.199 Site-Specific Field Methods and Mapping
There were no unusual field or laboratory methods applied to the Root River investigation. High altitude color infrared aerial photographs from the USGS NAPP program were the primary source of geomorphic information. Tonal contrasts on these air photos were sufficient to delineate features even finer than those mapped. The Houston, Fillmore, and Olmstead NRCS soil surveys were useful in supplying subsurface textural information. Mapping was done independent of previously available maps of the area, partly because of the difference in scale of the pre-existing maps and of the mapping being done, and partly because of the desire to map without reference to existing geologic models of the area. Mapping was accomplished by thin pencil directly on USGS 7.5' topographic maps. Many of the geomorphic boundaries were transferred to the maps from mylar overlays of NAPP aerial photography, with appropriate adjustments for distortion of the aerial photographs.
Landform sediment assemblages for the Root River are outlined in Table 12.8 and discussed below.
The Upland, Valley Terrace, Floodplain, Valley Margin, and Eolian LsSA’s were mapped in the Root River valley.
Upland landscape. Several undifferentiated upland areas are completely surrounded by Root Valley LsSA’s. These upland areas are typically underlain by eroded bedrock terrain that may have been exposed to pre-Wisconsinan glaciation.
Valley Terrace landscape. This landscape constitutes the majority of area within the Root River valley. The Root River is confined by local Paleozoic bedrock formations, and in places in its middle to upper reaches is gorge-like. The Valley Terrace landscape has five main geomorphic surfaces, of which the highest seem to "hang" on, or cut into the bedrock valley walls. The highest and oldest Valley Terrace LfSA’s (VT5-VT3) are commonly mantled by loess, which indicates a relatively older age for their sediment assemblage. The VT5 and VT4 LfSA’s have informally been called "rock-core" meanders in tributaries of the Upper Mississippi River valley. The VT5 LfSA is the top of the actual core of these incised meanders. These cores are traceable throughout the valley, but are more common in lower valley reaches. The VT5 sediment assemblage commonly consists of loess over bedrock or loess over coarse sand over bedrock. The VT5 surface is differentiated on the basis of thickness and texture of associated sediment. This surface can be underlain by material that is thin, discontinuous, and undifferentiated by texture (VT5-), coarse and less than 2 m thick (VT5CO<), coarse and greater than 2 m thick (VT5CO), and fine over coarse and greater than 2 m thick (VT5FC). The VT5 LfSA is further differentiated if loess is present (VT5L, VT5L<, VT5CO<L, VT5FCL). The VT4 LfSA is the incised meander, and is sometimes a bedrock bench. VT4 LfSA’s are traceable throughout the valley and are differentiated along lines similar to that for the VT5 LfSA. The VT4 LfSA’s can be underlain by material that is coarse and less than 2 m thick (VT4CO<), coarse and greater than 2 m thick (VT4CO), and fine over coarse and greater than 2 m thick (VT4FC). Coarse material is often coarse sand and gravel. The LfSA’s are differentiated further if loess is present (VT4L, VT4L<, VT4FCL).
The VT3 LfSA correlates to the "Savanna Terrace" and its underlying strata within the Upper Mississippi River valley. Terrace remnants can be traced nearly the length of the valley. Some of these VT3 LfSA’s occupy old "rock-cored" meanders. The VT3 LfSA may or may not have a loess mantle depending on the degree of mass wasting and soil erosion. Medium to coarse sands and gravels are generally underlying the VT3 and VPC3 LfSA’s. The geomorphically lower LfSA’s in this landscape typically have fine overbank deposits of variable thickness overlying coarse channel and bar sands and gravels. As with the preceding terraces, the VT3 sediment association can consist of material that is thin, discontinuous, and undifferentiated by texture (VT3-), coarse and less than 2 m thick with underlying bedrock exposed (VT3CO), coarse and greater than 2 m thick (VT3C), and coarse over fine and greater than 2 m thick (VT3CF). In fewer cases, the VT3 terrace can be underlain by material that is fine and less than 2 m thick (VT3F<) and fine over coarse and greater than 2 m thick. These LfSA’s are differentiated further if loess is present (VT3L, VT3L<, VT3CL, VT3FCL). Finally, there are cases where a buried soil is present within the sediment assemblage (VT3COB). Few paleochannels are associated with the VT3 terrace (VPC3). Where they were noted in the subsurface, material was fine textured and greater than 2 m thick (VPC3F).
The V2T terrace is close to the VT1 terrace in elevation. They appear to be one merged terrace in the upstream portions of the valleys and diverge in the downstream direction. The VT2 LfSA has the highest recognizable geomorphic surface that is not loess-mantled within the Root River valley. The VT2 and VT1 LfSA’s are restricted in width relative to the VT3 and geomorphically higher LfSA’s. This width decreases downstream of the midpoint of the valley. The VT2 sediment assemblage apparently is always greater than 2 m thick. It ranges from coarse (VT2CO) to fine (VT2F) to fine over coarse (VT2FC). Occasionally, a buried soil is present (VT2COB). The buried soil is probably the pre-Euroamerican settlement surface soil. Straight to meandering paleochannels associated with the VT2 terrace are evident (VPC2).
The City of Houston sits on one of the largest VT1 terrace remnants. The VT1 rises on the order of 1-3 m above the floodplain surface. As with the VT2 terrace, the VT1 sediment assemblage is apparently always greater than 2 m thick. This sediment assemblage ranges from coarse (VT1CO) to fine over coarse (VT1FC) to, occasionally, peat or organic muck (VT1P). Buried soils are present in both fine and coarse sediment assemblages (VT1FB, VT1COB). The buried soil is possibly the pre-Euroamerican settlement surface soil. Paleochannel segments associated with the VT1 terrace occur in the lower half of the valley (VPC1).
Tributary valleys also exhibit a suite of terraces. These were not traced in detail and are simply mapped as undifferentiated VT.
Floodplain landscape. Four different types of floodplain, including "undifferentiated," occur in the Root River Valley based on morphology and expression on aerial photographs. All floodplain types are found throughout the length of the valley, although there are a number of downvalley trends in terms of relative valley area covered. There are no explicit temporal relationships in the definitions of the different floodplain types; however, given the types of processes exhibited by the Root River, a general increase in age of the basal increments of the LfSA’s is likely from type "w" through types "x" and "y." The younger deposits tend to thicken downstream, and abandoned channel belts and individual paleochannels become more prevalent downstream. The lowermost part of the Root River floodplain appears more like a delta near the confluence with the Mississippi River. The floodplain LsSA is greater than two meters thick and laterally continuous. Coarse material generally represents point bar lateral accretion deposits or channel sand and gravel. Fine material represents the accumulation of floodplain overbank sheetflood deposits. Type "o" overbank deposits are nearly ubiquitous and deeply bury a soil that is laterally extensive. These overbank deposits tend to thicken downstream as do the underlying floodplain deposits. This buried soil is the pre-Euroamerican settlement surface soil in the valley and serves as a prominent marker.
Type "w" floodplain LfSA’s consist of active and recently active point bars adjacent to the river. Overbank and point bar deposits of historic age can be in excess of 4 m thick in some of lower stretches of the Root River valley. Much of this is very recent in age. Textures are typically sand (FFWCO) or a relatively thin increment of fine textured material over sand (FFWFCO). The buried soil is present only occasionally (FFWCOB).
Type "x" floodplain LfSA’s consist of inactive point bars that are actively being buried by overbank deposits. Evidence of ridge and swale morphology is present. Type "x" floodplain LfSA’s are dominated by fine textures, but have textures that range from coarse (FFXCO) to fine (FFXFO) to fine over coarse (FFXFCO). All three of this LfSA’s have counterparts with the buried soil (FFXCOB, FFXFOB, FFXFCOB). Where the sediment assemblage is unknown, FFXO is assigned. Tributary valleys exhibit a similar type of floodplain with similar facies (FFXT).
The type "y" LfSA shows no surface indications of channel or point bar migration features, either because they are relatively deeply buried or were never present. Drainage conditions are generally poor. Type "y" LfSA’s are dominated by thick increments of fine textured floodplain overbank sheetflood deposits. As with type "x" LfSA’s, underlying material ranges from coarse (FFYCO) to fine (FFYFO) to fine over coarse (FFYFCO) to a few instances of peat or organic muck (FFYP). The buried soil is usually present (FFYCOB). Occasionally, the texture of the sediment assemblage could not be estimated (FFYO).
Crevasse splay LfSA’s are best defined in the lower reaches of the Root River valley. These splays overlie and/or interfinger with floodplain LfSA’s. The crevasse splay LfSA’s are greater than 2 m thick. As would be expected given that splays fine rapidly away from the channel, the sediment assemblages range from fine to coarse (FCSFOB, FCSCOB). The buried soil is usually present, but on occasion it is absent (FCSFC) suggesting the possibility of multiple episodes of splaying with the younger being Historic in age.
Natural levee LfSA’s are present in the lower reaches of the Root River Valley. They overlie and/or interfinger with floodplain LfSA’s and are greater than 2 m thick. Textures range from coarse (FNLCO, FNLCOB) to fine (FNLFOB) to fine over coarse (FNLFC). Most examples have type "o" overbank deposits and the buried soil (FNLCOB, FNLFOB). Similar to the splays, there are examples that apparently lack these features suggesting a younger age for a second episode of levee construction. FNL was mapped where the underlying sediment assemblage could not be determined.
Floodplain paleochannels are common in the lowest part of the valley. At least two generations are evident, but are not associated with a particular floodplain type yet. All are filled with fine texture material (FPCFO1, FPCFO2) and undoubtedly have an increment of type "o" overbank deposit.
An undifferentiated delta is present at and beyond the mouth of the Root Valley. Distinct lobes and distributary paleochannels make it typical of deltas. Low levees are present but indistinguishable in this area. The underlying sediment assemblage is either fine material (FDEFOB) with type "o" overbank deposits and the buried soil, or fine over coarse material (FDEFC) apparently lacking the buried soil. FDE is mapped where the sediment assemblage is not determined.
Tributary floodplains (FF) occupy both small "v"-shaped valleys and larger valleys with undifferentiated floodplains and terraces. The sediment assemblage is greater than 2 m thick, ranging from coarse (FFCO) to fine (FFFB) to fine over coarse (FFFC). A buried soil is present in the FFFB LfSA.
Two generations of meander belts within the Root River valley are currently recognized and delineated. Some are inset into the Valley Terrace LsSA. In all cases, the associated sediment assemblage is greater than 2 m thick. Where texture information is available, the older sequence has a fine texture (FMB2FBF) and exhibits the buried soil. Elsewhere, texture is not confident and differentiation is made based on the LsSA the meander belt is inset into (FMB2F, FMB2V). The other meander belt LsSA’s range from coarse texture (FMBCOF, FMBCOV) to fine (FMBFBF, FMBFBV, FMBFF) to fine over coarse (FMBFCF, FMBFCV). The sediment assemblage could not be determined confidently (FMBF, FMBV) in some places. The other major differentiating factor in the LfSA’s was whether the meander belt was inset into the Floodplain LsSA (FMBF) or Valley Terrace LsSA (FMBV).
Some of the tributary meander belts have associated outwash belts (FOBV). This LfSA is greater than 2 m thick and is composed of fine texture material to that depth.
Valley Margin landscape. Both alluvial fans and colluvial slopes are well represented within the Root River valley. Fans and slopes are often above the floodplain of the main channel, which means they are relatively high and dry. The Valley Margin landforms are composed of mass wasting, sheetwash, and channel deposits that are often in excess of 2 m thick. This means they range from diamictons to laminated deposits. Sheetwash sedimentation apparently dominates. Redeposited loess can provide these LfSA’s with a silty texture. Fans primarily are differentiated on the basis of texture and the LsSA they rest upon. Textures range from coarse (MAFCOF) to fine (MAFFBF) to fine over coarse (MAFFCBF). Multiple buried soils are common (MAFFBV). Fans were deposited on either the Floodplain LsSA (MAFF) or Valley Terrace LsSA (MAFV). This provides information on the maximum possible basal relative age of the fan. Colluvium (MC) typically is greater than 2 m thick and is undifferentiated to sediment assemblage.
Eolian landscape. This LsSA is currently recognized only on the VT3 terrace of the Valley Terrace LsSA in several locations. Dunes, sheetsand, and interdunal depressions and ponds are the three main landforms, although the former two are not distinguished. All have sandy textures. Dunes are labeled EEDV whereas interdunal depressions are labeled EDPV. Dunes post-date or are coeval with the latest phases of aggradation to the VT3 level. There is likely some overlap with youngest loess sedimentation.
188.8.131.52 Landform Sediment Assemblage Codes
Table 12.8 provides details on each of the specific LfSA codes used for the Root River valley model.
The upland landscape of eroded bedrock terrain consists of a single undifferentiated LfSA. Because of age considerations, it is ranked not possible (0) for all depth intervals below the land surface.
The three oldest valley terrace LfSA families, VT5, VT4, and VT3, all predate known prehistoric occupation in the upper Midwest. They are ranked not possible (0) for all depth intervals, with a number of exceptions. Where the material is fine (VT3F, VT3F<), or fine over coarse (VT3FC, VT4FC, VT5FC), or where loess is present (VT3FCL, VT3FL, VT3L, VT3L<, VT4FCL, VT4L, VT4L<, VT5CO<L, VT5FCL, VT5L, VT5L<), the 0-1 m depth interval is ranked high (3). When the fine material is greater than 2 m thick (VT3FCL, VT3L, VT4L, VT5L), the high (3) ranking is extended to the 1-2 m depth interval. Fine sediments may be loess, colluvium, or various types of slopewash. All these depositional environments are conducive to burial of cultural deposits with a relative minimum of disturbance. Where material is thin, discontinuous, and undifferentiated by material texture (VT3-, VT5-), a low (1) ranking is assigned the 0-1 m depth interval because the origin of the material is unknown.
Little is known about the younger VT2 and VT1 terrace LfSA’s. Based on radiocarbon ages from floodplain LfSA’s (Table E.1) and tentative regional correlation to the Kingston Terrace of the Mississippi Valley, both terraces are considerably older than 2,800 B.P. and probably post-date about 11,000 B.P. Regardless of exact age, they fall comfortably with the temporal span of prehistoric human occupation. Because of the lack of knowledge of these LfSA’s, rankings may be somewhat inflated to err on the side of caution. In nearly all LfSA’s for the two terrace levels, the 0-1 m depth interval has a high (3) ranking because of relatively moderate drainage conditions, particularly in comparison to adjacent floodplain LfSA’s. The 2-5 m depth interval has a ranking of low (1), even where coarse channel or bar sand is present, largely because it has yet to be investigated. Rarely, thick peat is present (VT1P). The surface of the peat is considered modern with a ranking of not possible (0) and the peat is assigned a ranking of low (1) throughout because of wet conditions. Where type "o" overbank deposits are present (VT1COB), the 0-1 m depth interval is ranked not possible (0) because of the Historic age of the overbank deposits. The buried soil in this LfSA is the pre-Euroamerican settlement surface soil. In some instances, the 1-2 m depth increment of VT1 and VT2 LfSA’s is ranked high (3) or moderate (2). This was done where fine or fine over coarse material is greater than 2 m thick (VT1FB, VT1FC, VT2, VT2F, VT2FC). The ranking reflects the potential thickness of fine material, although the origin of the material is not known. The VT1FB LfSA indicates the possible presence of type "o" overbank deposits.
Paleochannels associated with the valley terrace LfSA’s are recognized for the three youngest only (VPC1, VPC2, VPC3, VPC3F). Fill is usually undifferentiated by texture. The paleochannels are poorly to excessively poorly drained and ranked not possible (0) for all depth intervals.
Floodplain LfSA’s are middle Holocene and younger in age. Preliminary radiocarbon dating suggests the FFY LfSA’s, excluding younger overlying deposits, are younger than about 4,950 B.P. but older than about 600 B.P.; FFX LfSA’s, are younger than about 2,800 B.P. but older than about 150 B.P.; and, FFW LfSA’s are younger than about 150 B.P. Nearly all floodplain LfSA’s are overlain by type "o" overbank deposits that post-date about 150 B.P. and thicken downstream. This ubiquitous deposit renders the surface survey technique for locating cultural deposits useless for floodplains in the Root Valley. For these LfSA’s, the 0-1 m depth interval is ranked not possible (0). There are two exceptions. The FFX LfSA, undifferentiated by texture, is assigned a moderate (2) ranking for the 0-1 m depth interval. The FFXT LfSA, present in tributary valleys, is assigned a low (1) ranking throughout its thickness. Floodplain sediment assemblages underlying type "o" overbank deposits are assigned a ranking of moderate (2) for the 1-2 m depth interval. That assignment is extended to the 2-5 m depth interval for the FFY LfSA’s, and for the FFX LfSA’s that exhibit greater than 2 m of fine material (FFXFOB, FFXO) or are undifferentiated by texture (FFXO, FFXT). Fine material is interpreted as overbank sheetflood deposits conducive to burial and preservation of prehistoric cultural deposits. The FFW LfSA’s are ranked not possible (0) for all depth intervals by virtue of their Historic age.
Other floodplain LfSA’s, including crevasse splays, deltas, tributary floodplains, meander belts, overbank belts, and natural levees, are habitable depositional environments dominated by processes conducive to burial and preservation of prehistoric cultural deposits. Paleochannels are the lone exception due to their poor drainage conditions. In these floodplain LfSA’s, a ranking of moderate (2) is assigned to all depth intervals with some exception. The most extensive exception is where type "o" overbank deposits are present, the 0-1 m depth interval is assigned not possible (0) due to age considerations (FCSCOB, FCSFOB, FDEFOB, FFFB, FMB2FBF, FMBFBF, FMBFBV, FNLCOB, FNLFOB). Two floodplain LfSA’s that are inset into Valley Terrace LfSA’s (FMBCOV, FMBFCV) are ranked low (1), between 1-5 m in depth, because of the combination of coarse material overlying valley terrace LfSA’s. Similarly, in some situations where fine over coarse material is present, the 2-5 m depth interval is ranked low (1) (FDEFC, FFFC, FMBFCF). Several LfSA’s, including several meander belt LfSA’s representing a range of landforms, are ranked low (1) for their entire thickness down to 5 m (FFCO, FLR, FMB2F, FMB2V, FMBCOF, FMBF). Finally, floodplain paleochannels are ranked not possible (0) for all depth intervals.
The valley margin and eolian landscapes represent depositional environments highly conducive to burial and preservation of prehistoric cultural deposits. Sheetwash sedimentation aided by an ample supply of loess dominates the former. Dune growth and migration dominate the latter and could preserve cultural deposits in various degrees of integrity. LfSA’s representing these environments are ranked high (3) for all depth intervals. The only two LfSA exceptions (MAFFBF, MAFFBV) are where buried soils interpreted as the pre-Euroamerican settlement surface soil are present. The overlying deposits are assumed to be type "o" overbank or equivalent colluvial deposits.
The Mn/Model Final Report (Phases 1-3) is available on CD-ROM. Copies may be requested by e-mail: email@example.com
Mn/Model was financed by the Minnesota Department of Transportation using funds set aside by the Federal Highway Administration's Intermodal Surface Transportation Efficiency Act.
The Mn/Model process and the predictive models it produced are copyrighted by the Minnesota Department of Transportation (MnDOT), 2000. They may not be used without MnDOT's Consent.