Spatiotemporal variations of NDVI before and after implementation of Grain for Green Project in wind-water erosion crisscross region of the Loess Plateau

To propose more effective strategies and policies for ecological restoration, it is imperative to explore the variations in vegetation coverage before and after the implementation of “Grain for Green Project” and evaluate the effect of ecological restoration in the Loess Plateau. In present research, the wind-water erosion crisscross region of the Loess Plateau was selected as study area. Based on the EOT algorithm, GIMMS NDVI data and MODIS NDVI data during overlapping period were applied to establish GIMMS NDVI data (EOT NDVI) from 1982 to 2000 with the resolution of 1 km. Then, a monthly NDVI data set (EM NDVI) from 1982 to 2019 with the resolution of 1 km was integrated with the EOT NDVI data and the MODIS NDVI data from 2001 to 2019. Consequently, the spatiotemporal variations in NDVI of the study area were analyzed. The results showed that: (1) EOT algorithm was suitable for the GIMMS NDVI resampling application in the study area. Mean error (ME), mean absolute error (MAE) and root mean square error (RMSE) of EOT NDVI and MODIS NDVI in the same period were relatively low but the coefficient of determination (R²) was relatively high. (2) In 1982−2019, the average annual NDVI of the study area showed a decline trend first but then increased with the average annual NDVI growth rate of 0.007 8. (3) In 1982−2019, heterogeneity was discovered in the spatial growth rate of vegetation. The area where the average annual NDVI had used to show a downward trend before the implementation of the project (1982−2000) turned to show a significant upward trend in the average annual NDVI after the implementation (2001−2019). The results of the study can provide theoretical basis and data support for assessing the implementation effect of ecological restoration measures and future ecological environment construction.