Abstract
Ecological site information is essential to interpreting monitoring data and guiding site-specific management of ecosystem functions and services. Ecological information includes soil properties (e.g., texture class), geomorphology characteristics (e.g., slope aspect), and ecosystem dynamics (e.g., plant succession), which are critical covariates in rangeland monitoring programs such as the Assessment, Inventory, and Monitoring (AIM) strategy conducted by the Bureau of Land Management (BLM). Based on field observations, AIM identifies ecological sites according to ecological site concepts uniquely developed within individual Major Land Resource Areas (MLRA). Here, we present and evaluate the availability of ecological site identification, soil observations, and geomorphology characteristics determined by AIM data collectors between 2012 and 2021 in 14 states of the western United States. There are 31,267 monitoring plots (79% of plots) with identified ecological sites and 29,228 plots (74% of plots) containing soil morphology descriptions of soil horizons examined in excavated pits. While soil texture class is observed in most soil horizons (98%), rock fragment volume is the soil property with the least data availability (75%). The consistency of soil data (e.g., clay content observations within the ranges of texture classes) increases as a function of time following guidance in soil profile description training for AIM data collectors. Nearly 47% of AIM plots are found on gentle slopes of 0% to 5% steepness and on Flat/Plain and Hill/Mountain landscape types. We confirmed that the AIM database is a robust source of georeferenced soil and geomorphology information that can be used for land management and research on land potential, soil geography, and assessment of soil health indicators across the western United States.
Introduction
Soil information is important to determine land potential at local and regional levels. However, land managers often lack access to critical soil data (Lobry de Bruyn and Andrews 2016). Soil information consists of measurable soil morphology properties, such as soil texture, clay content, and soil depth. Based on these properties, we can use pedotransfer functions of various soil processes (e.g., infiltration rate) to estimate regulating ecosystem services, including water storage capacity (Bouma 1989; Duniway et al. 2010; Herrick et al. 2017). For example, we can predict the rate at which water percolates into a soil profile if we have data on soil texture, structure, bulk density, and organic matter (Duniway et al. 2010). Soil texture data are also applied to manage plant growth since differences in soil texture can result in up to 10 times the difference in potential plant-available water holding capacity (Brady and Weil 2016). For this reason, soil information is needed to quantify land potential (Herrick et al. 2017) and to make decisions on where and when to apply management practices such as livestock grazing (Briske et al. 2008; Dyer et al. 2021), vegetation restoration treatments (Pilliod et al. 2017), and erosion control (Jarrah et al. 2020; Edwards et al. 2022).
Soil information is especially useful to land managers in combination with geomorphology and plant community features to identify the ecological site at a given location (Johanson and Brown 2012; Caudle 2013). An ecological site is a conceptual division of the landscape. It is defined as a distinctive kind of land based on recurring soil, landform, hydrology, geology, and climate characteristics that differs from other kinds of land in its ability to produce distinctive kinds and amounts of vegetation and in its ability to respond similarly to management actions and natural disturbances (USDA NRCS 2017). Ecological site concepts are uniquely developed within individual Major Land Resource Areas (MLRA), i.e., geographically associated land resource units delineated by the USDA Natural Resources Conservation Service (NRCS) and characterized by a particular pattern that combines soils, water, climate, vegetation, land use, type of farming, and the organization of programs and practices of the USDA NRCS (Salley et al. 2016). Threshold values of soil properties can indicate a certain process-based limitation or aptitude of a given soil to support a specific kind of plant formation and ecological site (Heller et al. 2022). For example, a Loamy Upland Savanna Ecological Site (R104XY007IA) is characterized by having a high water-holding capacity and nutrient levels that underpin a variety of productive temperate grasses (Schaetzl and Anderson 2005). On the other hand, soils of Colorado’s Deep Sand Ecological Site (R067BY015CO) often impose constraints on plant growth and primary productivity due to low water availability (Buol et al. 2011). While soils are one of the primary drivers of plant-available water content, water availability in the soils can also be influenced by vegetation composition (Lane et al. 1998). As a result, ecological site concepts are developed based on the covariance between soil morphology characteristics, plant community, and geomorphology settings (e.g., slope, aspect, and landscape type) (Herrick et al. 2018; Heller et al. 2022). Without these field-verified soil morphology characteristics (soil depth, texture class, etc.) in combination with geomorphology information, ecological sites cannot be accurately identified. The purpose of ecological site identification is to provide a consistent framework for classifying and describing rangeland and forestland soils and vegetation, thereby delineating land units that share similar capabilities to respond to management activities or disturbance (Caudle 2013; USDA NRCS 2017).
Land managers often rely on monitoring to understand the condition and trend of ecosystems and determine the effectiveness of management actions. Although it is common practice to identify the vegetation during monitoring, it is equally important to record landscape position and soil morphological characteristics that underpin ecological site identification, e.g., soil color, texture, and structure. While these soil characteristics are available through national soil characterization databases such as the National Cooperative Soil Survey (NCSS) Soil Characterization Database (NCSS 2023), soil information collected as covariates in vegetation monitoring programs are less frequently made available, with some notable exceptions such as the Land Potential Knowledge System mobile application (LandPKS) (Maynard et al. 2022), the NRCS National Resources Inventory On-Site Grazing Land study, and the Bureau of Land Management (BLM)’s Assessment, Inventory, and Monitoring (AIM) terrestrial program, which characterize key ecosystem processes following standard soil, vegetation, and geomorphological protocols (Kachergis et al. 2022; Toevs et al. 2011; Herrick et al. 2018).
The purpose of the AIM strategy is to provide a standardized approach for measuring natural resource conditions and trends on 9.9 million km2 of BLM lands (Kachergis et al. 2022). The BLM manages activities including outdoor recreation, livestock grazing, mineral development, energy production, and the conservation of natural, historical, and cultural resources. The AIM strategy provides quantitative data and tools to guide and justify policy actions, land uses, and adaptive management decisions.
Beginning in 2011, the terrestrial AIM program has collected field data in 14 states in the United States containing BLM public lands (Alaska, Arizona, California, Colorado, Idaho, Montana, Nevada, New Mexico, North Dakota, Oregon, South Dakota, Utah, Washington, and Wyoming). AIM vegetation data are widely used to support research and land management, including fire studies in sagebrush ecosystems (Barker et al. 2019), long-term vegetation responses to pinyon-juniper woodland reduction treatments in Nevada (Ernst-Brock et al. 2019), vegetation predictors for biological soil crusts (Condon and Pyke 2020), and rangeland treatment effects in New Mexico (Traynor et al. 2020). However, while the AIM soil data offer a great opportunity for soil-related research in a vast geographical area of the western United States, they have been underutilized by the research and management communities. This is in part due to data access as the AIM soil data have not been publicly released and are not under the jurisdiction of national soil databases, e.g., the National Soil Information System (NASIS) operated by the NRCS. Recently, NRCS requested and received access to AIM soil data to apply it in ecological site identification.
Here, we present a description of the AIM soils and ecological site dataset. We describe the available variables, evaluate how the data quality has improved over time, and discuss geographic variations in the data. We discuss potential errors in soil observations and present considerations for land managers and other data users when leveraging these data. We look ahead to how monitoring programs like AIM can improve their soil data collection to inform ecological site identification. Our goal is to make the soil data from monitoring programs (e.g., AIM) more visible for rangeland managers and researchers who can leverage the soil information to address a variety of topics, including the assessment of land potential, soil health indicators, and spatial patterns of soil properties across the western United States.
Materials and Methods
Structure of the AIM Soils Dataset. The soils dataset presented here is structured in the following tables: (1) plot characteristics and (2) soil horizon (tables 1 and 2). The plot characteristics table contains information on global positioning system (GPS) coordinates, ecological site identification (ID), and geomorphological settings, such as slope, aspect, and landscape type. The soil horizon table has information on soil color, texture class, effervescence, and soil structure. The soil horizon table is structured by soil horizons examined in excavated pits at AIM monitoring plots. The link between plot characteristics and soil horizon tables is made by the “Primary Key,” which is a unique identifier of each monitoring plot. We accessed the AIM dataset from the national TerrADat database in June of 2022. We obtained ecological data from 39,601 monitoring plots surveyed from 2012 to 2021 in the only 14 US states currently under the BLM AIM program (figure 1a). Plots collected in 2011 were eliminated from this evaluation because collecting soils information was not part of the protocol at that time. As of June of 2022, the AIM dataset had 29,228 monitoring plots (74% of total plots) containing soil morphology observations in 122,170 soil horizons. Although vegetation remeasurements may be repeated at a site following local objectives (e.g., every three to five years), soil pits are only measured at the first plot visit. Therefore, AIM soil observations are not typically recollected at a site. Occasionally, soil data may be recollected at the direction of the state monitoring coordinator or BLM’s local project lead to address data quality concerns.
Plot characteristics schema in the Assessment, Inventory, and Monitoring (AIM) dataset.
Soil horizon schema in the Assessment, Inventory, and Monitoring (AIM) dataset.
(a) Ecological monitoring plot density across the western United States. The location of monitoring plots is represented using a kernel density map. (b) Types of ecological site ID in the Assessment, Inventory, and Monitoring (AIM) dataset. The NRCS standard ID (e.g., R042XB010NM) indicates that an ecological site was identified by AIM data collectors. The figure shows the proportion of plots according to the types of ecological site ID by state.
Collection Methods. AIM data are collected at random and targeted geographical locations exclusively within the BLM public lands (figure 1a). Appropriate sample designs used by AIM are identified based on the multiple management and monitoring objectives for an area. AIM sample designs are based on a Generalized Random Tessellation Stratified (GRTS) approach to define the location of monitoring plots (Stevens and Olsen 2004; McDonald 2012). GRTS approach ensures a random, spatially balanced sample within an area of interest (Stevens and Olsen 2004). With this approach, sample sizes can be expanded or reduced in response to logistical, financial, or personnel limitations and still provide a valid sample of the study area without the exclusion of specific regions, ecosystems, land use, or soil types (Kachergis et al. 2022). This allows for flexibility in monitoring while drawing inference to areas of interest with known levels of uncertainty. AIM sample designs enable rangeland condition estimates at different scales from national or ecoregional to field offices, watersheds, or smaller areas (Kachergis et al. 2022).
For soil morphology description, the AIM protocol follows the Monitoring Manual for Grassland, Shrubland, and Savanna Ecosystems (Herrick et al. 2018) as well as the Field Book for Describing and Sampling Soils, which was developed by the National Soil Survey Center, NRCS (Schoeneberger et al. 2012). Data gatherers record elevation, slope percentage, aspect, slope-shape, landscape type, and plot location at each monitoring site. A soil pit with 70 cm depth and approximately 2 to 3 sharpshooter width is dug near the plot center. AIM data collectors use the soil pit to identify the upper and lower depth boundaries (in centimeters) of major soil horizons. The AIM dataset contains observations of soil morphological properties for each soil horizon. Soil information in the AIM dataset includes soil texture, rock fragment by volume, clay percentage, and effervescence and is recorded for each soil horizon of the pit. Soil structure is an optional observation. Soils are classified at the soil series level by comparing the soil observations (recorded by AIM data gatherers) and existing soil surveys where soil series are indicated and described. Where possible, recorders use the landscape position information, slope, aspect, and elevation along with the soil pit to identify the ecological site following the NRCS ecological site classification system in Ecosystem Dynamics Interpretive Tool (EDIT, i.e., an online information system for the development and sharing of ecological site descriptions, ecosystem state and transition models, and land management knowledge) (Bestelmeyer et al. 2016).
AIM data collectors are typically organized in groups of two to three people, with hundreds of AIM data gatherers participating in field data collection annually. Each group collects data on 30 to 50 monitoring plots per year. Although the AIM strategy requires that data gatherers have previous soil knowledge, there is no minimal level of that knowledge. However, AIM data collectors are required to attend training at least every three years to ensure adhesion to the protocol. At these regional AIM trainings, data collectors learn the vegetation core methods (Kachergis et al. 2022; Toevs et al. 2011), along with soils and ecological site description and data quality assurance and quality control procedures (McCord et al. 2022). During the soil portion of AIM trainings, data gatherers learn about ecological site concepts, landscape type identification, compass and clinometer use, how to dig a soil pit, and methods for identifying soil morphological characteristics. Data gatherers calibrate soil texture class identification by hand with known soil samples and have opportunities to practice soil identification techniques under the supervision of experienced soil scientists. Calibration for soil texture observation requires AIM data collectors to practice the determination of the texture of several soils with contrasting and known textures. Although most AIM data collectors do not have extensive experience in determining soil texture by hand test, the accuracy of the exact or adjacent soil texture class has been demonstrated to be 78% in AIM data collectors and similar populations (Salley et al. 2018). After field training, AIM data collectors receive knowledge support from a network of regional and national soil scientists as questions arise. Some states such as Nevada and New Mexico have additional known soil texture samples for additional within-season practice and calibration.
Quality Assurance and Quality Control of AIM Soils Data. The AIM program actively seeks to improve quality assurance and quality control to ensure data quality (McCord et al. 2021, 2022). In addition to field training where data collectors have an opportunity to practice soil and geomorphology observations under supervision from experienced personnel, collectors may receive feedback from local soils experts throughout the season. Where possible, AIM data collectors use electronic data capture, which enables realtime data checks to reduce errors made in the field (Kachergis et al. 2022; McCord et al. 2022). For example, possible clay values are limited to boundaries of soil texture classes. The AIM dataset is annually reviewed by a centralized, national data management team that maintains quality control workflows, with contributions from many individuals across BLM offices in each state (Wilkinson et al. 2016; Kachergis et al. 2022). AIM data are also reviewed at least twice a year during the field season collection that takes place during the growing season and at the end of the field season by project leads and expert soil scientists who identify and correct errors where possible. Data review is important to identify the necessity for adjustments in regional AIM trainings that address core and contingent methods and can effectively communicate protocols and data quality procedures in one to two weeks (Kachergis et al. 2022). Although the extent of this review varies by state, efforts are ongoing to produce dashboards and other tools to support more consistent data review and correction.
Major Characteristics of the Dataset
Ecological Site in AIM Dataset. Ecological sites are identified with an NRCS Ecological Site ID in 79% (31,267) of the total monitoring plots, but it varies by state (figure 1a and table 3). While several states (Idaho, Nevada, New Mexico, etc.) contain more than 90% of monitoring plots with identified ecological sites, Colorado and California have 27% and 41% of the plots with ecological site ID, respectively (table 3, figure 1b, and Supplementary Data 1). The reason for the variability of ecological site ID is likely due to inconsistent or incomplete published concepts for those MLRAs during the data collection period. This can be attributed to ongoing ecological site classification efforts (USDA NRCS 2015) and soil survey updates (USDA NRCS 2021). Similar to other survey efforts by the US NCSS, the USDA-Ecological Site class concepts are continually being evaluated and updated. Also, there are vast areas of land that do not have ecological sites identified; in some areas, reclassification of soil maps means that AIM ecological site identification needs to be revised. As a result, at the time when AIM crews visited plots, some were unable to reference the appropriate ecological site class concept.
Summary of Assessment, Inventory, and Monitoring (AIM) monitoring plots in terms of ecological site ID and soil observations.*
As ecological site concepts are uniquely developed within individual MLRA (Salley et al. 2016), we further evaluated ecological site identification by comparing the field observed MLRA in the ecological site ID and the 2006 NRCS MLRA spatial product. MLRAs are a geographic regionalization product meant to organize the programs and practices of NRCS, including ecological site concepts. In practice, this means that site and AIM soil data described in a common MLRA would be conditioned by similar state soil forming factors, such as climate and geomorphology. In Nevada, 7,069 monitoring plots (83% of all plots) have an observed ecological site ID that matches the MLRA location. Although we analyzed the availability of ecological site ID throughout the AIM dataset (table 3), we presented here the detailed results for Nevada State since it has the greatest number of AIM plots (8,473) in which ecological sites are identified (table 3). Similar to other states (Supplementary Data 1), ecological site concepts in Nevada have been used across more than one MLRA (figure 2a). Ecological site concepts were “borrowed” from adjacent MLRAs and used in ecological site identification at monitoring plots predominantly located at the boundaries of MLRAs (figure 2a). In Nevada, the boundaries of MLRAs 24 and 25 are those that contain the highest number of monitoring plots (191 and 125 plots) that shared ecological site concepts (figure 2b). It is likely that ecological site concepts may be used in adjacent MLRAs due to (1) a lack of ecological site concepts in certain MLRAs and/or (2) the gradual ecological transition at MLRA boundaries. The comparatively uncommon instances (17% of Nevada AIM plots) where there is an MLRA mismatch in the middle of an MLRA polygon warrant further investigation prior to using the data. These mismatches may be inclusions or soil mapping errors but should be verified as accurate. The confusion matrices and maps presented here for Nevada (figure 2) are available for other states in Supplementary Data 1.
(a and b) Maps of Nevada State showing the boundaries of Major Land Resource Area (MLRA) and the location of ecological monitoring plots. (a) Monitoring plot density with ecological site identification using the concepts originated from the MLRA where the plot is located. (b) Monitoring plot where ecological site identification borrowed concept from another MLRA. (c) MLRA identification and names in Nevada. (d) Confusion matrices representing the number of plots with MLRA from plot location versus plot observation in Nevada. Most of the plots have a match between MLRA of location and observation. However, ecological site concepts may be used in adjacent MLRAs due to (1) a lack of ecological site concepts in specific regions of the MLRA and/or (2) the gradual ecological transition at MLRA boundaries. Evaluations of other states are found in Supplementary Data 1.
Soil Observation Availability. Of the total AIM monitoring plots (3,960) surveyed between 2012 and 2021, 29,228 plots (74% of total plots) contain soil morphology observations in a total of 122,170 soil horizons. The absolute number of AIM monitoring plots with excavated soil pits and consequently soil observations increase as a function of time—year (figure 3a) and state (figure 3b). The drop in the number of soils pits/plot in 2020 is likely due to plot revisits where soil pits are not sampled rather than a survey error. The number of AIM plots with soil observations varies by state (figure 3b) in which we can highlight the substantial number of plots (>2,000) with soil data in Nevada, Oregon, Wyoming, Montana, Idaho, Utah, California, Colorado, and New Mexico. Data availability in the AIM database also varies between the soil properties (table 4 and figure 4). Observations of soil effervescence, structure, and clay content are less available (75% to 82% of all soil horizons) than soil texture class and color (98% to 95% of all soil horizons). The total number of plots with soil observations increased substantially from the years 2012 (191 observations) to 2021 (19,531 observations) with the greatest number of annual soil observations (21,416 observations) in 2019 (figure 4). The availability of AIM soil data increases as a function of time, except for soil structure (figure 4). The absolute increase in the number of plots sampled and soil observations from 2015 to 2017 reflects the adoption of BLM policy that directed the use of AIM to support land use planning and greater sage grouse (Centrocercus urophasianus) habitat assessments (Bureau of Land Management 2016). In the more recent years (2020 and 2021), AIM data collectors collected more than 15,000 soil observations from >5,000 monitoring sites annually. The proportion of AIM plots decreases in 2020 due to the gradual transition of the program to focus sampling efforts on plot revisits that focus on vegetation measurements. Soil pits are not sampled upon plot revisits unless there was not a soil pit available with the first plot visit or if there were data quality concerns associated with the first soil pit.
Total Assessment, Inventory, and Monitoring (AIM) monitoring plots and those with ecological site ID and soil observations (e.g., soil texture class) as a function of (a) year and (b) state.
Level of availability of soil morphology data from >29.000 soil pits assessed between 2012 to 2021 by Assessment, Inventory, and Monitoring (AIM) strategy in monitoring plots located across the western United States.
Soil observation availability increases as a function of time (years). Data availability is lower for soil structure compared to the other soil properties ([a] soil texture class, [b] soil clay content, [c] soil structure, [d] soil effervescence, [e] soil color, and [f] rock fragment volume). Note the significant increase in clay content observations after the year 2017. The percentage indicated on the top of the bar represents the availability of data for a certain year.
Summary of Soil Observations. Soil pit depth in AIM monitoring plots averages 55 ± 21 cm across the dataset (figure 5a). The median soil pit depth is 51 cm. The protocol for AIM data gathering suggests that AIM data collectors dig a soil pit of 70 cm depth (Herrick et al. 2018). However, AIM data collectors may encounter soils with shallow bedrock or cemented horizons that limit soil pit depth (figure 5b). Among the soil texture classes defined by NASIS (figure 6b), we found that Sandy Loam is the most frequently observed (17%) in the AIM dataset, followed by Loam (13%), Clay Loam (11%), and Sandy Clay Loam (11%). Clay content of 49,542 soil horizons is, on average, 25% ± 17%, but it varies by state. Soil texture class and clay content are highly variable by state (figure 6a). For example, Arizona State has more observations of coarse soil textures (Sandy Loam and Loamy Sand) and lower clay content (21% ± 15%, n = 619) when compared to Montana State, where fine soil textures (Clay and Silty Clay) and elevated clay content (31% ± 18%, n = 3,801) are commonly observed. Using the AIM dataset, we initiated a spatial distribution analysis of soil texture classes of the upper soil horizon (0 to 16 ± 14 cm soil depth) at the monitoring plots (figure 7). AIM monitoring plots with Clay and Loam textures are predominantly in the north and northwest United States (figures 7a and 7c). Sand and Loamy Sand textures have similar spatial patterns with a great number of monitoring plots in southwest United States, such as in California, Nevada, Arizona, Colorado, Wyoming, and New Mexico (figures 7b and 7d). The most common soil texture (Sandy Loam) in the AIM dataset is geographically widespread across BLM public lands (figure 7e).
(a) Soil pit depth in Assessment, Inventory, and Monitoring (AIM) plots. Most of the monitoring plots (55%) contain a maximum of 51 to 71 cm soil pit depth. Average pit depth in the AIM dataset is 55 ± 21 cm (n = 29,228). The Monitoring Manual for Grassland, Shrubland, and Savanna Ecosystems (Herrick et al. 2018) suggests that AIM data collectors should dig a soil pit of 70 cm depth. (b) Data collectors may encounter soils with shallow bedrock or cemented horizons, such as the soil profile of the Colthrorp series, a Xeric Argidurids with continuous indurated duripan (Bkqm horizon) at 45 cm depth from southern Idaho. The measuring tape is in decimeters (Photo by P. McDaniel, published in Blackburn et al. [2021]).
(a) The distribution of soil texture classes observed by Assessment, Inventory, and Monitoring (AIM) crews is highly variable by state. Note that Sandy Loam (SL) is the most observed soil texture class, which represents nearly 17% of AIM soil observations in all states combined. (b) Soil texture classes and respective NASIS code.
Spatial distribution of soil texture at the topsoil (0 to 16 ± 14 cm soil depth) of Assessment, Inventory, and Monitoring (AIM) monitoring plots. Monitoring plots containing soils with (a) Clay and (c) Loam textures in the upper soil horizons are predominantly in the north and northwest United States. (b) Sand and (d) Loamy Sand texture have similar spatial pattern with great number of monitoring plots in southwest United States, such as in California, Nevada, Arizona, Colorado, Wyoming, and New Mexico. The most common soil texture in the AIM dataset is (e) Sandy Loam, which is widespread across BLM public lands. (f) USDA soil texture triangle with indicated texture classes represented on the maps.
To assess the consistency of the AIM soil information, we obtained data from the AIM soil horizon table. We determined whether clay content observation falls in the ranges for the observed soil texture class (figure 8). Data consistency improved in the last two years (2020 and 2021; see figure 8a), likely due to additional rules in the data collection using tablets, which constrain clay content ranges to the selected soil texture. The increase in agreement between clay content and texture class is also due to the improvement in training. Clay content falls within the correct range of soil texture class in more than 95% of the soil observations for all states (figure 8b). The most common soil texture classes observed by AIM data collectors (Sandy Loam and Loam) contain high levels of consistency (>98%) to the observed clay content (figure 8c).
Quality control assessment of the Assessment, Inventory, and Monitoring (AIM) soil data to determine whether clay content observation falls in or out of the ranges for the observed soil texture class. The percentage indicated next to the bar represents the proportion of clay content observations that falls in the ranges of soil texture class. (a) Data by year demonstrates that data consistency improved in the last two years (2020 and 2021). (b) Quality control assessment by state. (c) Data according to soil texture class (see NASIS code for texture class in figure 6b). Note that the most common soil texture classes observed by AIM data collectors, Sandy Loam (SL) and Loam (L), contain high levels of consistency (>98%). The data shown in this figure comes from the quality control assessment made by us using the AIM soil observations.
Summary of Geomorphology Characteristics. Most AIM plots (47%) are found in gentle slopes of 0% to 5% steepness (figure 9a). However, Colorado, Washington, and North Dakota stand out for having more AIM plots on steep terrains (>10% slope) compared to other states (figure 9b). The initial selection of points does not take topography into account. However, plots with slopes greater than 50% or unsafe slopes are rejected from the sampling; therefore, plots with steep slopes are rare. More details about rejection criteria defined by the AIM are in the Bureau of Land Management (2023).
Geomorphology characteristics of the Assessment, Inventory, and Monitoring (AIM) monitoring sites. Slope steepness by (a) range and (b) states. Slope aspect for (c) all states combined and (d) in New Mexico separately. All the 16 cardinal orientations recorded in the AIM dataset are shown in the figure. (e) Landscape type of the AIM monitoring plots.
When observing data from all states combined, the AIM plots are equally distributed in terms of slope aspects (figure 9c). However, the disaggregation of the data by state may show a different trend. For example, the most common slope aspect of AIM plots in New Mexico is southeast followed by southwest (figure 9d). With respect to landscape type, we noticed that Hill/Mountain and Flat/Plain are predominantly observed by AIM crews, although some states like Colorado and Nevada have fewer Flat/Plain observations than the overall trend (figure 9e).
Data Accessibility. A cleaned version of the AIM soil dataset containing only plot characteristics and soil horizon data can be accessed through the AgDataCommons DOI at https://doi.org/10.15482/USDA.ADC/1529416. The AIM dataset is regularly updated following the ongoing collection of data. The updated AIM data can be accessed through the public data portal, which can be found at https://www.blm.gov/AIM/PublicData. Raw data tables can also be requested from the BLM by emailing BLM_OC_NOC_AIM_Team@blm.gov. Although the BLM land is managed in a natural system, it is unknown how much of the coverage of the AIM survey is on land managed under specific uses (e.g., grazing). The implementation of AIM across the states is not even due to many reasons including funding, evolution of policy, timing of the adoption of AIM by each of the states, etc. The 14 US states are the only ones collecting terrestrial AIM data. By large the eastern states have very limited BLM surface lands, and the terrestrial AIM program is of limited utility on those lands.
Summary and Conclusions
We evaluated the availability of ecological site identification and soil observations determined by AIM data collectors between 2012 and 2021 in the western United States. The abundance of soil morphology observations in the AIM dataset offers a great opportunity for a variety of soil-related studies, including the quantification of land potential, the definition of patterns in soil geography, water-holding capacity predictions, and the assessment of soil health indicators across the western United States. The high availability of soil observations should encourage cooperative soil survey researchers to partner with BLM to verify the usability and accuracy of soil data. For example, researchers may analyze soil particle size distribution (percentage of clay, silt, and sand) using laboratory methods to determine, in detail, clay content and soil texture class of the same soil samples described by AIM data collectors. Moreover, laboratory determinations of organic carbon (C) content can be performed by researchers in soil samples that AIM crews observed soil color (hue, value, and chroma). Although the soil color and organic matter content relationship does not always hold true, organic matter content is one of the most important pigments that influence soil color, which is why there is a long history of relating soil color darkness to organic matter content (Konen et al. 2003; Wills et al. 2007; Aitkenhead et al. 2013). In general, a low soil color value (dark soil color) indicates a high organic C content. Likewise, soil effervescence tests performed by AIM data collectors can be correlated to carbonate contents determined in the lab. Using the AIM dataset to establish these correlations between soil morphology observations with organic matter content and salt concentration can create new opportunities for soil research. These examples can show how soil information is used for management applications and soil monitoring (Wadoux et al. 2021).
We confirmed that the AIM database is a rich source of soil information using completeness and consistency checks (McCord et al. 2022). We provided an overview about the AIM soil data, as well as a preliminary examination of the consistency in clay content and soil texture class observations. Because soil observations are highly variable by state, we suggest that future guidance to AIM crews and similar data collectors should include regional focal topics to improve the quality of soil data. Similarly, in states where ecological site identification rates are low, state-level efforts, in coordination with NRCS, will ensure that AIM vegetation data can be interpreted in the context of ecological dynamics. Future work should evaluate the level of errors in the AIM soil data to determine if those errors impact evaluations of ecosystem dynamics related to disturbance and land management through ecological site data. Since the AIM plots contain a GPS coordinate, future work will compare AIM soil observations with other geospatial data resources, such as those obtained through the LandPKS SoilID algorithm (Maynard et al. 2020, 2022). Comparing the AIM soil data against other reliable soil databases and laboratory determinations is one approach to estimate the accuracy and consistency of AIM soil observations (Rossiter et al. 2015). Regularly updating workflows to evaluate and improve the AIM soils dataset would ensure that the AIM soils data collection program and others who follow similar protocols will continue to produce relevant and useful soil data for years to come.
Supplementary Material
The supplementary material for this article is available in the online journal at https://doi.org/10.2489/jswc.2024.00068.
Acknowledgements
This research was supported by the Bureau of Land Management (BLM) under Agreement #4500104319. We thank the BLM staff for providing valuable feedback to improve our analysis.
- Received August 15, 2023.
- Revision received February 17, 2024.
- Accepted March 1, 2024.
- © 2024 by the Soil and Water Conservation Society
