DYNAMICS OF SPECTRAL INDICES OF REMOTE SENSING DATA ON THE EXAMPLE OF IRRIGATED LANDS OF EAST KAZAKHSTAN

Abstract

The article presents the results of field and remote monitoring works using archival satellite data. Examples of data processing from LandSat-8 satellites are shown. The purpose of the study is to identify an alternative method for monitoring irrigated areas and assessing soil salinity.

In the framework of the study, ground data of the conducted salt survey of irrigated lands of the Semey district, Abay region were used. The collected satellite images were processed and interpreted and the necessary data were extracted with subsequent entry into the geodatabase. The licensed ArcMap 10.4 software was used to process satellite images.

Vegetation indices NDVI and SAVI showed a low correlation with ground data on soil salinity. The NDSI salinity indices, SI4, and SI9, also showed a low correlation with ground data on soil salinity. NDSI values are symmetrically opposite to the vegetation index.

The PSS1 formula (Prediction Soil Salinity 1) shows a low correlation (r2=0.1) with ground data.

A good correlation (r2=0.87) with ground data was shown by PSS2 (Prediction Soil Salinity 2 - Prediction Soil Salinity 2) identified using multi-linear regression by scientists from Istanbul Technical University. Based on this formula, a salinity map of irrigated lands in the Semey district, Abay region was constructed. The area of the territory with salinity was about 1891 km2 (5%), with weak salinity of 2197 km2 (5.9%), with no salinity of 33002 km2 (88%).