Regional‐scale Assessment of Soil Salinity in the Red River Valley Using Multi‐year MODIS EVI and NDVI

The ability to inventory and map soil salinity at regional scales remains a significant challenge to scientists concerned with the salinization of agricultural soils throughout the world. Previous attempts to use satellite or aerial imagery to assess soil salinity have found limited success in part because of the inability of methods to isolate the effects of soil salinity on vegetative growth from other factors. This study evaluated the use of Moderate Resolution Imaging Spectroradiometer (MODIS) imagery in conjunction with directed soil sampling to assess and map soil salinity at a regional scale (i.e., 10-10(5) km(2)) in a parsimonious manner. Correlations with three soil salinity ground truth datasets differing in scale were made in Kittson County within the Red River Valley (RRV) of North Dakota and Minnesota, an area where soil salinity assessment is a top priority for the Natural Resource Conservation Service (NRCS). Multi-year MODIS imagery was used to mitigate the influence of temporally dynamic factors such as weather, pests, disease, and management influences. The average of the MODIS enhanced vegetation index (EVI) for a 7-yr period exhibited a strong relationship with soil salinity in all three datasets, and outperformed the normalized difference vegetation index (NDVI). One-third to one-half of the spatial variability in soil salinity could be captured by measuring average MODIS EVI and whether the land qualified for the Conservation Reserve Program (a USDA program that sets aside marginally productive land based on conservation principles). The approach has the practical simplicity to allow broad application in areas where limited resources are available for salinity assessment.