Preprints
https://doi.org/10.5194/soil-2021-80
https://doi.org/10.5194/soil-2021-80

  13 Sep 2021

13 Sep 2021

Review status: this preprint is currently under review for the journal SOIL.

How well does Predictive Soil Mapping represent soil geography? An investigation from the USA

David G. Rossiter1,2, Laura Poggio1, Dylan Beaudette3, and Zamir Libohova4 David G. Rossiter et al.
  • 1ISRIC-World Soil Information, Postbus 353, Wageningen 6700 AJ, NL
  • 2Section of Soil & Crop Sciences, New York State College of Agriculture and Life Sciences, 233 Emerson Hall, Cornell University, Ithaca NY 14853 USA
  • 3USDA-NRCS Soil Survey Division, 19777 Greenely Rd., Sonora, CA, 95370 USA
  • 4USDA-ARS, Dale Bumpers Small Farms Research Center, 6883 South State Hwy 23, Booneville, AR 72927 USA

Abstract. We present methods to evaluate the spatial patterns of the geographic distribution of soil properties in the USA, as shown in gridded maps produced by Predictive Soil Mapping (PSM) at global (SoilGrids v2), national (Soil Properties and Class 100 m Grids of the USA), and regional (POLARIS soil properties) scales, and compare them to spatial patterns known from detailed field surveys (gSSURGO). The methods are illustrated with an example: topsoil pH for an area in central New York State. A companion report examines other areas, soil properties, and depth slices. A set of R Markdown scripts is referenced so that readers can apply the analysis for areas of their interest. For the test case we discover and discuss substan- tial discrepancies between PSM products, as well as large differences between the PSM products and legacy field surveys. These differences are in whole-map statistics, visually-identifiable landscape features, level of detail, range and strength of spatial autocorrelation, landscape metrics (Shannon diversity and evenness, shape, aggregation, mean fractal dimension, co-occurence vectors), and spatial patterns of property maps classified by histogram equalization. Histograms and variogram analysis revealed the smoothing effect of machine-learning models. Property class maps made by histogram equalization were substantially different, but there was no consistent trend in their landscape metrics. The model using only national points and covariates was not better than the global model, and in some cases introduced artefacts from a lithology covariate. Uncertainty (5–95% confidence intervals) provided by SoilGrids and POLARIS were unrealistically wide compared to gSSURGO low and high estimated values and show substantially different spatial patterns. We discuss the potential use of the PSM products as a (partial) replacement for field-based soil surveys.

David G. Rossiter et al.

Status: open (until 25 Oct 2021)

Comment types: AC – author | RC – referee | CC – community | EC – editor | CEC – chief editor | : Report abuse
  • AC1: 'Comment on soil-2021-80', David Rossiter, 16 Sep 2021 reply

David G. Rossiter et al.

David G. Rossiter et al.

Viewed

Total article views: 140 (including HTML, PDF, and XML)
HTML PDF XML Total Supplement BibTeX EndNote
101 37 2 140 10 0 2
  • HTML: 101
  • PDF: 37
  • XML: 2
  • Total: 140
  • Supplement: 10
  • BibTeX: 0
  • EndNote: 2
Views and downloads (calculated since 13 Sep 2021)
Cumulative views and downloads (calculated since 13 Sep 2021)

Viewed (geographical distribution)

Total article views: 129 (including HTML, PDF, and XML) Thereof 129 with geography defined and 0 with unknown origin.
Country # Views %
  • 1
1
 
 
 
 
Latest update: 17 Sep 2021
Download
Short summary
Maps of soil properties made by machine-learning techniques are increasingly applied in earth surface process modelling and agronomy. Maps of the same area made by different methods appear quite different, and also differ from field-based polygon soil survey maps. We explore these differences both visually and numerically, using methods that quantify the spatial patterns. Readers can apply the methods to their areas of interest in the USA with supplied R Markdown scripts.