Digital soil mapping for fire prediction and management in rangelands

Levi, Matthew R.; Bestelmeyer, Brandon T.

doi:10.1186/s42408-018-0018-4

Fire Ecology

Table 1 Spatial soil information currently available for modeling fire in the United States. ¹SS = soil survey; DSM = digital soil map; PRS = proxy remote sensing for soil properties. ²Both represents both raster and vector formats available. ³1:250 000 in continental US and 1:1 000 000 in Alaska

From: Digital soil mapping for fire prediction and management in rangelands

Dataset	Group¹	Data type	Extent	Resolution	Advantage for pre or post fire	Disadvantage for pre or post fire	Method of development	Reference
HWSD	SS	Raster	Global	1:1 000 000 to 1:5 000 000	Extensive spatial coverage	Coarse resolution; limited interpretations	Merged European Soil Database, soil map of China, regional SOTER databases, and Soil Map of the World	Wieder et al. 2014
STATSGO2	SS	Vector	Continental	1:250 000 to 1:1 000 000³	Extensive spatial coverage	Coarse resolution; limited interpretations	Soil-landscape paradigm; tacit knowledge; field and laboratory sampling and analysis	Soil Survey Staff 2018c
SSURGO and gSSURGO	SS	Both²	Continental	1:12 000 to 1:63 360	Extensive spatial coverage; numerous interpretations and properties	Variability within map units; some areas without data		Soil Survey Staff 2018a and 2018b
SoilGrids	DSM	Raster	Global	250 m and 1 km raster cells	Extensive spatial coverage; quantified model uncertainty; data gaps filled	Varying sample density in each soil type	Probabilistic based machine learning	Hengl et al. 2017
US48 SoilGrids100m+	DSM	Raster	Continental	100 m raster cells				Ramcharan et al. 2018
POLARIS	DSM	Raster	Continental	30 m raster cells				Chaney et al. 2016
Local DSM maps	DSM	Both	Regional and local	Variable (as detailed as 5 m raster cells)	Fine spatial resolution; some data gaps filled; quantified model uncertainty	Varying sample density in each soil type; difficult to locate; often specific goals	Many methods ranging from regression to machine learning	Grunwald 2009; many others
Direct soil moisture (e.g., SMAP, SMOS, GRACE)	PRS	Raster	Global	3 to 36 km raster cells	Near current data availability	Limited data record; Coarse spatial scale	Satellite remote sensing with ground validation	Entekhabi et al. 2010; Jensen et al. 2018
Indirect soil moisture	PRS	Both	Regional and local	Variable	Quantitative and process-based information	Relies on empirical relationships and subject to model uncertainty	E.g., vegetation indices, inverse process-based models, land surface models	Abatzoglou 2013; Waring 2016

Back to article page