Regression Analysis of Normalized Difference Vegetation Index (NDVI) to Compare Seasonal Patterns and 15 Year Trend of Vegetation from East to West of Nepal

Understanding the changing patterns and trend of vegetation is essential for its socio-environmental values. Normalized difference vegetation index (NDVI), a satellite based data obtained from Moderate Resolutions Imaging Spectro-radiometer (MODIS) were analysed. The data have a characteristic resolution of 250 × 250 m2 and a 16-day composite period. They were ordered separately for each sample plot from east, centre and west of Nepal, for 15 years period, 2000 to 2015. MODIS, Land Surface Temperature (LST) data were used to identify unreliable NDVI values and hence eliminated. Also, the unusually fluctuating NDVI values during the rainy season were removed. A cubic spline function (for seasonal patterns), linear regression model (for NDVI trend) and generalized estimating equations (GEE for comparison of the changing trends) were the models used. The results showed a patterned annual seasonal vegetation and significant trends during the 15 years. Seasonal growth showed a peak in rainy season and trough in the winter season, with slight temporal variation among the areas with a characteristic shift of seasonal greening (start of greening) and browning (end of greening) from east to west of Nepal. The NDVI trend was significantly rising in eastern and western suburban areas while the central urban city had a significant decline. The temporal shift of start and end of the season from east to west can be of value to agriculturalists.