DETECTING SALINITY AND SOIL NUTRIENT DEFICIENCIES USING SPOT SATELLITE DATA

Citation
S. Rahman et al., DETECTING SALINITY AND SOIL NUTRIENT DEFICIENCIES USING SPOT SATELLITE DATA, Soil science, 158(1), 1994, pp. 31-39
Citations number
34
Categorie Soggetti
Agriculture Soil Science
Journal title
ISSN journal
0038075X
Volume
158
Issue
1
Year of publication
1994
Pages
31 - 39
Database
ISI
SICI code
0038-075X(1994)158:1<31:DSASND>2.0.ZU;2-2
Abstract
A study was conducted to determine the predictive capability of SPOT ( Systeme Probatoire d'Observation de la Terre) satellite data for detec ting areas that are subjected to salinity encroachment and nutrient de ficiencies. The SPOT observations were related to the chemical charact eristics of soil samples collected from a corn field located south of Torrington, in east central Wyoming. On a false color infrared composi te (FCC) image, the saline areas were bright white patches; healthy ve getation was bright red to magenta in color. A nearby reservoir had a dark blue tone except where light blue tones indicate silty and/or sha llow water. From the visual interpretation of the FCC image, it was no t possible to predict soil salinity quantitatively. Statistical analys is of SPOT digital counts by bands indicated that the near infrared ba nd (X-3) was superior to the visible bands for salinity detection. Of the three SPOT bands (XS-1, XS-2, and XS-3), XS-3 was significantly co rrelated with saturated paste electrical conductivity (EC) and water s oluble Na. Brightness index (BI), the summation of digital count of th e three bands, was positively correlated with soil EC, and water solub le Na, Ca, Mg and was negatively correlated with Mn, while the normali zed difference vegetation index (NDVI = (NIR-Red)/(NIR + Red)) and rat io index (RI = NIR/Red) were negatively correlated with EC and water s oluble Na, Ca, and Mg. All the spectral bands were significantly and p ositively correlated among themselves and with BI. Analysis of varianc e indicated that the sampling points possessing high BI values had hig her EC, water soluble Na, Ca, and Mg and lower levels of Mn and Zn. Sa mples with low NDVI and RI values had high EC, water soluble Na, Ca, a nd low Mn, indicating that high salinity and nutrient deficiency can b e detected with reasonable accuracy with BI, NDVI, and RI. Among the s pectral indices, BI proved to be the best indicator of salinity and nu trient deficiency.