Understanding the Vegetation Dependence of Flash Drought Detection and Performance of Early Warning

Flash droughts can cause devastating impact on both natural and managed ecosystems. Planning and mitigation efforts will benefit from impact-based flash drought detection and early warnings. However, the performance of the latter often varies depending on vegetation type and climate regimes. Here we seek to identify the source(s) of the performance dependence for the rapid changes index of solar induced fluorescence (SIF-RCI) as an early warning for flash drought events detected based on the soil water deficit index (SWDI). To account for data dependence and uncertainties, we derived the SIF-RCI from three different sources, GOSIF and CSIF during 2000-2022 and GOME-2 during 2007- 2018, and derived the SWDI based on three different sources including GLDAS, GLEAM, and ERA5-Land. Sensitivity analyses were conducted to assess the impact of vegetation-relevant parameters on both the SWDI-based flash drought frequency and the SIF-RCI-based early warnings; the performance statistics of the early warmings were then derived for different combination of vegetation parameters. This allowed us to quantify the contribution of each parameter to the uncertainty and vegetation dependence of the flash drought early warning performance, and to propose solutions for advancement. While our study was conducted for the global land, this presentation will place a special focus on the contiguous Unites States where data from other sources are available as supplementary supporting evidence.