Soil thickness is crucial for global hydrological and ecological processes. Topsoil (A horizon) thickness influences plant growth, carbon storage, and biogeochemical cycles. Solum thickness is determined by the balance between soil production and loss through erosion. Both A-horizon and solum thickness vary spatially and temporally. Accurately representing this variation is challenging for earth system models, which often use constant soil depth values. Mechanistic models, assuming long-term equilibrium, and empirical models, relating thickness to environmental variables, have been used but have limitations. There's a lack of national-scale assessments of temporal changes in soil thickness across diverse eco-climatic zones. This study uses a long-term, large-scale soil survey dataset to quantify spatial and temporal variations of A horizon and solum thickness across the conterminous US (CONUS) over 69 years. The hypotheses are: (1) National-scale spatial variations are mainly controlled by natural soil-forming factors, with climate having the most significant impact, followed by topography and land cover; and (2) Temporal variations are mainly driven by human activities (land cover/use change, tillage).

Existing research highlights the importance of soil thickness in various ecological and hydrological processes. Studies have attempted to quantify spatial variations using mechanistic and empirical models. Mechanistic models, based on soil production functions and sediment transport models, often assume a long-term equilibrium rarely achieved in reality. Empirical models use statistical or machine learning techniques to map soil thickness based on environmental variables, but face challenges due to high spatial variability and data limitations. While the slow rate of soil formation (approximately 2.5 cm per 1000 years) is well-established, human activities and climate change have significantly accelerated soil loss, exceeding natural formation rates by an order of magnitude. Previous studies have documented substantial A-horizon loss in specific regions of the US, particularly in agricultural areas. However, a comprehensive national-scale assessment of temporal changes across various eco-climatic zones remains absent.

This study utilized the National Cooperative Soil Survey (NCSS) Soil Characterization Database from the USDA-Natural Resources Conservation Service. Data from 1950-2018 across the CONUS were used, after applying rigorous filtering criteria to remove data points with inconsistencies or insufficient information. Two types of soil thickness were analyzed: A-horizon thickness (sum of all A horizon thicknesses in the profile) and solum thickness (thickness of all horizons above the C or R horizon). Generalized additive models (GAMs) were used to analyze the relationship between soil thickness and various environmental variables. Environmental variables included soil order, temperature and moisture regimes, parent material, land use (including land use changes), and climatic variables (precipitation, temperature, drought indices, evapotranspiration, soil moisture, etc.). Topographic variables (elevation, slope, curvature) were also included. A national-scale GAM was developed, followed by regional GAMs using Land Resource Regions (LRRs) to account for regional variations. To investigate temporal changes, chronosequences within each LRR were selected, controlling for factors like soil order, parent material, land use, and topographic characteristics. Simple linear regression models were used to estimate temporal change rates in soil thickness for each chronosequence.

The spatial distribution of A horizon and solum thickness across the CONUS showed strong longitudinal and latitudinal patterns. The A horizon was shallowest in the west (desert and Rocky Mountain regions) and deeper in the Mississippi River Basin and West Coast. Solum thickness was shallowest in the southwest desert and Nebraska Sandhills, and deeper in the southeast. At the national scale, the spatial distribution of soil thickness was strongly correlated with climatic variables. Soil moisture was primarily associated with A horizon thickness, while temperature was more strongly associated with solum thickness. Topography also played a role, with thicker soils on summits and thinner soils on slopes. Soil order and parent material exhibited more pronounced effects at the regional scale. Land use significantly impacted A horizon thickness, with cropland generally exhibiting thicker A horizons. Temporal analysis revealed varying trends across LRRs. Significant A horizon loss was observed in Mollisols of the Central Great Plains, Alfisols on steep slopes, and soils under cropping. Mollisols under cropland in the Central Great Plains showed the highest A-horizon loss rate (0.44 cm/year). In contrast, some regions, especially those with forested or pasture lands, showed an increase in A horizon thickness. The study explained about 30-40% of the variation in A horizon and solum thickness, with unexplained variation attributed to local-scale factors.

The findings demonstrate the dominant influence of climate on the national-scale spatial distribution of soil thickness, with moisture influencing A horizon formation and temperature affecting solum development. Topography and soil order play significant roles, while parent material and land use exert more localized effects. Temporal changes are primarily driven by land use and erosion processes, with severe A-horizon loss in specific regions due to agricultural practices. The interplay between natural soil formation and anthropogenic impacts is evident. The model’s relatively low explanatory power underscores the importance of local-scale factors and highlights the need for future research investigating those factors.

This study provides a comprehensive overview of the spatial and temporal variations in A horizon and solum thickness across the CONUS. Climate significantly influences spatial patterns, while land use and erosion drive temporal changes. The findings highlight the need for targeted conservation practices to mitigate topsoil loss, especially in agricultural regions. Future research should focus on investigating local-scale variations, the impact of specific management practices, and the effects of the O-horizon.

The study has several limitations. The model explained only about 30-40% of the total variation, leaving a significant portion unexplained. The analysis focuses on temporal changes since 1950, potentially underestimating losses from earlier conventional agricultural practices. Uneven sample sizes across years and regions may affect the robustness of the temporal analysis. The analysis simplifies soil thickness changes to erosion and formation, neglecting the effects of tillage practices. Finally, the O-horizon was not investigated in this study.