Suitability analysis of remote sensing monitoring methods for grassland vegetation growth

The research and application of grassland vegetation growth monitoring methods have important scientific significance and application value for the sustainable use of grassland resources and the improvement of the ecological environment. Remote sensing monitoring has characteristics such as high timeliness and wide coverage, and several remote sensing monitoring methods have been increasingly used in crop growth monitoring. As the monitoring objects are all plants, these monitoring methods were tried to introduce to monitor grassland vegetation growth in this study. We applied four remote sensing monitoring methods for the crop growth: direct monitoring, vegetation growth process curve, same period comparison, and NDVI percentiles to the grassland vegetation growth monitoring of West Ujimqin County in Inner Mongolia; to clarify the suitability and limitation of these remote sensing monitoring methods for crop growth when monitoring grassland vegetation growth using MODIS Vegetation Index Products (NDVI). The monitoring results provided by the direct monitoring method and the NDVI percentile method were compared with the ground sampling data. The direct monitoring method could intuitively reflect the spatial heterogeneity of grassland vegetation growth by the grassland NDVI, and the NDVI value was significantly correlated with the dry weight of grassland yield per unit area (R²=0.5502). However, this method could not provide details on the growth of different types of grasslands owing to the limitation of the NDVI grade. The vegetation growth process curve method could only collectively reflect the changes in the overall growth of the grassland in the monitored area over time, showing that the growth was better or worse than that of the reference year. In this study, the NDVI peak values of the vegetation growth process curves for the meadow grassland and typical grassland were 0.73 and 0.55, respectively, significantly different from the whole regional lumped monitoring results (NDVI peak value was 0.60). This means that different grassland types should be monitored separately to reflect their respective growth processes. For the same period comparison method, if the selected reference year was different, the method would provide different monitoring results for grassland growth; the results from grassland growth monitoring were semi-quantitative comparative. Using the statistical analysis of NDVI data of different grassland types for 5 years, the NDVI percentile method could quantitatively evaluate the growth of corresponding grassland types, as shown in this study. The determination coefficient of the correlation between the growth score provided by the NDVI percentile method and the dry weight of grassland yield per unit area was 0.5047. To achieve a reasonable classification of these semi-quantitative monitoring results of grassland growth, the assistance of ground grassland monitoring information is required. There is an urgent need to establish a sky-air-ground integrated grassland growth monitoring platform to improve the efficiency and accuracy of grassland vegetation growth monitoring.