US: Scientists have developed a new method to assess soil salinity by using remote sensing (RS) technology. This could give land managers worldwide a regional-scale tool for measuring and inventorying soil salinity in fields where salt build-up lowers crop yields.
Dennis Corwin, soil scientist, Agricultural Research Service (ARS) led the team. He partnered with Stanford University scientist David Lobell and former University of California-Riverside statistician Scott Lesch. Other collaborators were Michael Ulmer, Keith Anderson, Dave Potts, James Doolittle, Manuel Matos and Matthew Baltes, who are soil scientists with the US Department of Agriculture (USDA) Natural Resources Conservation Service (NRCS).
Scientists used Moderate Resolution Imaging Spectroradiometer (MODIS) imagery to assess and map soil salinity across approximately 741,300 acres of North Dakota’s and Minnesota’s Red River Valley. Increased soil salinity levels in this region have been linked to higher water tables caused by management and precipitation changes over the past 20 years.
The team compared two vegetation indices: the normalised difference vegetation index (NDVI) and the enhanced vegetation index (EVI). Both indices were developed using seven years of vegetation reflectance data obtained with MODIS imagery. The team also collected soil samples from 60 fields across three counties in the Red River Valley to see how strongly field-scale soil saline levels correlated with the EVI and NDVI indices.
The researchers found that 21 to 37 percent of the variability in soil salinity levels could be correlated with EVI. Then they added another factor into their estimates: whether the land qualified for the Conservation Reserve Program (CRP), a federal programme that sets aside marginally productive land for conservation purposes. They found that 34 to 53 percent of the variability in soil salinity could be correlated to EVI and whether land was eligible for CRP inclusion.
Results from this research were published in the Journal of Environmental Quality.