Predicting Key Vegetation Parameters from Multi-source Remote Sensing Data

The richness of remote sensing data has been rapidly increasing with particular new features of higher spatial and temporal resolution and more diverse bands. Recent lunch of VENμS in particular provides us with renewed promises because of its high resolution in time and space, as well as the unique bands. To explore the capability of VENμS, here in Hulunber - a test site of VENμS in the southeast Mongolia Plateau, grasslands of different grazing intensity were used to evaluate the performance of different remote sensing data for predicating key vegetation parameters (e.g., biomass, coverage, LAI, fPAR, etc.). Our initial analysis based on VENμS images in April of 2019 indicated that retrieving the parameters of grassland green vegetation compared well with other bands. For example, the red-edge showed high sensitivity to grazing intensity and time. The red-edge position (REP) was higher at high grazing plots than that at the low grazing intensity. The changes in REP with grazing intensity appeared very different between the two dates when VENμS images were available. R² between NDVI₇₀₅ of the red-edge bands (VENμS) and grazing intensities was 0.06 in April when there is no leaf, while the R² generated by using NDVI of Landsat was 0.55. This indicates a clear shortfall from Landsat for calculating NDVI during the non-growing season at our sites. Over the growing season, we will continue our exploratory endeavors with the upcoming images and complete a comprehensive assessment of VENμS advantages.