Daily dynamic thresholds of different agricultural drought grades for summer maize based on the Vegetation Water Index

Effective daily drought monitoring for summer maize (Zea mays L.) is critical to prevent crop yield reduction. Most of the vegetation indicators used for drought identification have been based on changes in canopy greenness that may not change immediately under drought stress, and have relatively long-time intervals (e.g., eight or 16 days) that are limited by available remote sensing products. Moreover, drought thresholds, may be potentially different during different crop growth stages, rather than constant during the entire growing season, as has been widely used in most existing drought indicators. We used the orthorectified reflectance in the short-wave infrared (SWIR) band, extracted from MODIS MCD43A4 because of its high sensitivity to variations in vegetation water content due to this waveband's liquid water absorption characteristics. The Vegetation Water Index (VWI), calculated with the Normalized Difference Water Index (NDWI), was proposed for daily drought monitoring of summer maize in order to minimize the effects of geographical location, soil conditions, and ecological systems in different areas. Additionally, daily dynamic thresholds for drought-free, mild drought, moderate drought, and severe drought were determined with double logistic fitting functions having the highest R2 (0.79), based on the observed drought disaster and phenophase records of summer maize in the Huang-Huai-Hai region (HHH). Drought samples constructed from regional drought records and the Yearbook of Meteorological Disasters in China were used for drought validations. The results exhibited high accuracy (78.04% of total drought samples correctly showing complete correspondence with historical records regarding drought grades, and 96.90% of drought samples being within one drought grade level). Additionally, compared with the Crop Water Deficit Index (CWDI), the temporal and spatial evolution of typical drought processes were more accurately characterized by VWI. The results of this study provide a scientific basis and achieve quantitative monitoring for daily drought assessment of summer maize in HHH, and also provide a new perspective in dynamic and quantitative monitoring of agricultural disasters over large regions based on actual disaster records and remotely sensed images.