Modeling of vegetation index and land surface temperature to identify and compare the changing trends, using generalized estimating equations

Normalized difference vegetation index and land surface temperature data, in a sample plot each from east, center and west of Nepal, from 2000 to 2015, were analyzed to identify and compare the trends of vegetation and temperature changes during the period. The data were obtained from moderate resolutions imaging spectro-radiometer. Normalized difference vegetation index charactiszes a resolution of 250×250 m2 and a 16-day composite period while land surface temperature has 8 days frequency with resolution of 1×1 km2. The analysis was separate for normalized difference vegetation index and land surface temperature. The data were seasonally adjusted and then divided into three groups of five year period each, separate for every region. The generalized estimating equations were fitted to each period data. For all three regions, the results showed, there was a trend of  significantly rising vegetation in eastern and western sub urban parts while the central urban city had a significant decline in trend. Whereas the temperature showed statistically significant and uniform fluctuating pattern of change in all three regions. The rate of temperature rise is fastest in central region where the vegetation is continuously declining. However, the results revealed no relationship of trend of changing temperature with that of vegetation.

Keywords:

Generalized estimating equations (GEE) , Land surface temperature (LST) , Nepal , Normalized difference vegetation index (NDVI) , Satellite data