POLARIS: A 30-meter probabilistic soil series map of the contiguous United States

Abstract

Chaney, Nathaniel W.; Wood, Eric F.; McBratney, Alexander B.; Hempel, Jonathan W.; Nauman, Travis W.; Brungard, Colby W.; and Odgers, Nathan P., "POLARIS: A 30-meter probabilistic soil series map of the contiguous United States" (2016 A new complete map of soil series probabilities has been produced for the contiguous United States at a 30 m spatial resolution. This innovative database, named POLARIS, is constructed using available high-resolution geospatial environmental data and a state-of-the-art machine learning algorithm (DSMART-HPC) to remap the Soil Survey Geographic (SSURGO) database. This 9 billion grid cell database is possible using available high performance computing resources. POLARIS provides a spatially continuous, internally consistent, quantitative prediction of soil series. It offers potential solutions to the primary weaknesses in SSURGO: 1) unmapped areas are gap-filled using survey data from the surrounding regions, 2) the artificial discontinuities at political boundaries are removed, and 3) the use of high resolution environmental covariate data leads to a spatial disaggregation of the coarse polygons. The geospatial environmental covariates that have the largest role in assembling POLARIS over the contiguous United States (CONUS) are fine-scale (30 m) elevation data and coarse-scale (~2 km) estimates of the geographic distribution of uranium, thorium, and potassium. A preliminary validation of POLARIS using the NRCS National Soil Information System (NASIS) database shows variable performance over CONUS. In general, the best performance is obtained at grid cells where DSMART-HPC is most able to reduce the chance of misclassification. The important role of environmental covariates in limiting prediction uncertainty suggests including additional covariates is pivotal to improving POLARIS' accuracy. This database has the potential to improve the modeling of biogeochemical, water, and energy cycles in environmental models; enhance availability of data for precision agriculture; and assist hydrologic monitoring and forecasting to ensure food and water security.

Toward seamless hydrologic predictions across spatial scales

“…The predictions should have small uncertainties to be useful for the end users. These grand challenges also imply that the next generation of land surface and hydrologic models must incorporate probabilistic descriptions of the subgrid variability of geophysical land surface properties -such as POLARIS (Chaney et al, 2016b) and SoilGrids (Hengl et al, 2017) -to cope with the large uncertainties that characterize the related process below the representative elementary area (REA) scale. Consequently, great efforts should be made in hyper-resolution monitoring at the global scale in improving the computational efficiency of LSMs/HMs and in the development of scale-invariant parameterizations for these models.…”

The CAMELS data set: catchment attributes and meteorology for large-sample studies

“…The STATSGO data set is subject to several critical limitations, many of them being overcome by the recently released POLARIS data set (Chaney et al, 2016) and SoilGrids (Hengl et al, 2017). A key advantage of these two data sets is that they describe soil attributes at high resolution, using machine learning algorithms to estimate uncertainty.…”

The evolution of process-based hydrologic models: historical challenges and the collective quest for physical realism

“…We see three specific paths forward. First, there are numerous opportunities to improve information on geophysical properties, including estimates of vegetation structure (Simard et al, 2011), soil depth (Pelletier et al, 2016), soil properties (Chaney et al, 2016b), bedrock depth and permeability (Fan et al, 2015), and the physical characteristics of rivers (Gleason and Smith, 2014).…”

“…The challenge is to estimate spatial variations in the storage and transmission properties of the landscape. Advances are possible through the development of new data sources on geophysical attributes (Simard et al, 2011;Gleason and Smith, 2014;Fan et al, 2015;Chaney et al, 2016b;Pelletier et al, 2016;De Graaf et al, 2017), new approaches to link geophysical attributes to model parameters (Samaniego et al, 2010;Kumar et al, 2013;Rakovec et al, 2015), and new diagnostics to infer model parameters Yilmaz et al, 2008;Pokhrel et al, 2012). Such focus will give the parameter estimation problem the scientific attention that it deserves, rather than the far-too-common approach where parameter estimation is relegated to a "tuning exercise" in model applications.…”