Shortwave Infrared Water Stress Research Articles

Grass curing-driven fire danger index in a protected mountainous grassland using fused MODIS and Sentinel-2

ABSTRACT This study assessed the spatio-temporal variation of the degree of curing (DoC) in the Golden Gate Highlands National Park (GGHNP) from 2016 to 2020. The estimation of the degree of grass curing was carried out with the high spatial resolution fused remotely sensed data of MODIS and Sentinel-2 employing Index-then-Blend (IB) data fusion methods on Google Earth Engine (GEE) to compute Grassland Curing Index (GCI) using soil and vegetation moisture content indices-based grassland curing algorithm. Grassland curing indices were computed on Google Earth Engine and converted into Grassland Curing Maps (GCMs), which were then synthesized into fire danger maps. The study showed that all selected indices have the highest DoC in the month of September, with the minor time variation observed in MODIS and Sentinel 2-derived GCI estimated using Shortwave Infrared Water Stress (GCI_SIWSI) in the months of August and March, respectively. The analysis further revealed that GCI_SIWSI (44%) derived from Sentinel-2 data yielded the highest area of extremely high fire danger, followed by GCI estimated using Normalized Difference Moisture Index (GCI_NDMI) (25%) derived from fused data, MODIS-derived GCI_SIWSI and fused-derived GCI estimated using Global Vegetation Monitoring Index (GCI_GVMI) (both 16%). The GCI_SIWSI derived from Sentinel-2 data also showed that more than 95% of the entire landmass of the study area was within a high to extremely high fire danger zone, making the park’s vegetation susceptible to fire. Among all the four selected indices correlated against fire point, only SIWSI revealed a positive relationship across MODIS, Sentinel-2 and Fused data. It is also evident from the study that GCMs derived from Fused data outperformed MODIS and Sentinel-2 with the highest R2 and F values of 0.65 and 380, respectively. These results indicate that GCI derived from fused remotely sensed data is a promising GCI for accurately assessing the DoC over mountainous grassland environments. The study paved the way for increasing the spatial resolution for estimating the DoC for fire and fuel management of parks and other fire agencies within mountainous grassland environments.