COS 191-10 - A modern soil nitrogen concentration and pool baseline for the conterminous United States

Friday, August 11, 2017: 11:10 AM
B115, Oregon Convention Center
Terry D. Loecke, 2101 Constant Ave., University of Kansas, Lawerence, KS; Kansas Biological Survey, University of Kansas, Skye Wills, USDA-NRCS and Stephen Roecker, NRCS, USDSA

Soil nitrogen (N) is a key driver of many aspects of terrestrial and aquatic ecosystems. Previous estimates of total soil N for the conterminous US (CONUS) are derived from legacy soil samples collected over an approximately 60 year period and were not intended to be scaled to aerial estimates. The Rapid Carbon Assessment Project was undertaken to ascertain the soil carbon and nitrogen stocks across CONUS within a narrow time frame (20 months), using consistent protocols, and explicitly designed for spatial interpolation. Sample locations were chosen randomly from the NRI (National Resource Inventory) sampling framework and cover all areas in CONUS with SSURGO certified maps as of Dec 2010. The project was regionalized into 17 areas for logistical reasons. Within each region, soils were grouped by official series description properties. Sites were selected by soil groups and land use/cover as indicated by NRI or NLCD (USGS National Land Cover Dataset) class so that more extensive soils groups and/or land use/covers received more points and less extensive fewer points. Each region had 375 – 400 sites, for a total of approximately 6,400 sites. At each site, basic information about land use, vegetation and management were collected as appropriate and available.


Soil N concentration across CONUS is log-normally distributed and ranges from less than 0.001% to 4.1% with a median of 0.11% (n=27,344). Vertically, soil N concentration decreases with depth on average 0.3% m-1. Soil N concentrations tends to be highest in wetlands, lowest in rangelands and intermediate in forests and grasslands (e.g., CRP and pasturelands). Soil carbon to N ratio (CN) is also log-normally distributed across CONUS and ranges from near zero to more than 400 with a median of 10.6. Organic horizons average a CN of 33 whereas mineral soils average 12.4. To calculate soil N stocks and regional pools, soil bulk density was measured from 0 to 50 cm of depth and modelled below. A Hierarchical Bayesian statistical approach was used to estimate N stocks, pools, and uncertainty for each sampling level (i.e., CONUS, region, soil group, landuse and site). N stocks were summarized by surface horizon and depth increments and mapped by linking estimates to a raster of SSURGO that includes map unit and NLCD classification. This modern soil N stock baseline data set will be useful for many application in ecosystem and climate science.