Long-term trend analysis of vegetation changes using MODIS-NDVI time series during 2000-2017 (Case study: Kurdistan province)

Over time, environmental and human factors have caused positive and negative changes in the quantity and quality of vegetation, and these changes will continue. Vegetation greenness trend may be either increasing (Greening) or decreasing (Browning). Understanding the factors creating these changes and detection of trends in the past and future, would be useful for managers and decision makers. This study investigates the long-term changes in Kurdistan vegetation greenness by using the 16-day composite MODIS NDVI time series for 2000–2017, and 414 images with 250-m pixel size. The coordinate system of MODIS data was converted from the sinusoidal to the geographic map projection and a time series of data was created. After seasonality removal in the time series data, the long-term trend analyses were performed using parametric ordinary least squares (OLS) regression and non-parametric Theil-Sen and Mann-Kendall methods at pixel by pixel basis. Both OLS and Theil-Sen represented the similar result and indicated a slight increase in NDVI i.e., the poor greenness during the study period. According to the Mann-Kendall trend analysis, about 97 % of the province experienced a slight increasing and 2.46 % showed decreasing trend. However, based on the Kendall significance test, only 12 % of the province had a significant increasing or decreasing trend at 1 % confidence level and the rest of the area showed a slight and negligible increasing trend.