Skip to main content
ARS Home » Research » Publications at this Location » Publication #310236

Title: Spatio-temporal analysis of multi-year Landsat 7 data for regional scale soil salinity assessment

item SCUDIERO, ELIA - University Of California
item Corwin, Dennis
item Skaggs, Todd

Submitted to: ASA-CSSA-SSSA Annual Meeting Abstracts
Publication Type: Abstract Only
Publication Acceptance Date: 7/28/2014
Publication Date: N/A
Citation: N/A

Interpretive Summary:

Technical Abstract: Despite decades of research in soil mapping, characterizing the spatial variability of soil salinity across broad regions remains a crucial challenge. This work explores the potential benefits of employing reflectance data from the six spectral bands (blue, 450-520 nrn; green, 520-600 nrn; red, 630-690 nrn; near-infrared, 770-900 nrn; infrared-I; 1550-1750 nrn, and infrared-2, 2090-2350 nrn) of the Landsat 7 (L7) satellite sensor (30x30 m resolution) for salinity assessment. Acquisitions of L 7 throughout the western San Joaquin Valley, California (ca.15000 km") were investigated over a seven-year period. Two salinity ground truth datasets were evaluated, across 23 fields farmed with various crops: 226 direct measurements (ca. 2x2 m resolution), from the 0-1.2 m soil profile; and ca. 6000 block-kriged estimations (30x30 m resolution), derived from geospatial electromagnetic induction measurements. The multi-year average of L 7 data generally provided stronger correlations (up to R2=0.41), than those observed for each single year. Slightly stronger correlations (up to R2=0.43) were observed between salinity and the multi-year temporal variability of L 7 reflectance (i.e., standard deviation at each map-cell over time). The strength of the correlations between L 7 data and soil salinity varied according to changing meteorological conditions through the seven-year period and according to soil texture at a field by field basis. Additionally, selected salinity ranges (i.e., 0-2, 2-4, 4-8, 8-16, and> 16 dS/m) were characterized by significantly different values of the blue, green, red, and near-infrared bands. The results suggest that data fusion ofthe L 7 multi-year reflectance with information on meteorological conditions, crop type, and soil texture could lead to a reliable salinity prediction model for the entire western San Joaquin Valley. Land resource managers, producers, agriculture consultants, extension specialists, and Natural Resource Conservation Service field staff are the beneficiaries of regional scale maps of soil salinity.