A quantitative analysis to objectively appraise drought indicators and model drought impacts

摘要: Abstract. Drought monitoring and early warning is an important measure to enhance resilience towards drought. While there are numerous operational systems using different drought indicators, no consensus on which indicator best represents impact occurrence for any given sector. Furthermore, thresholds widely applied in these indicators but, date, little empirical evidence exists as trigger impacts society, the economy, ecosystems. The main obstacle evaluating commonly used a lack of information impacts. Our aim was therefore exploit text-based data from European Impact report Inventory (EDII) identify that meaningful region-, sector-, season-specific occurrence, empirically determine thresholds. In addition, we tested predictability based best-performing indicators. To achieve aims correlation analysis ensemble regression tree approach, Germany UK (the most data-rich countries EDII) test beds. As candidate chose two meteorological (Standardized Precipitation Index, SPI, Standardized Evaporation SPEI) hydrological (streamflow groundwater level percentiles). revealed accumulation periods SPI SPEI linked longer compared with Germany, but variability within each country, among categories and, some degree, seasons. median splitting values, regard estimates around −1 UK; distinct differences between northern/northeastern vs. southern/central regions were found Germany. Predictions approach yielded reasonable results good coverage. predictions also provided insights into EDII, particular highlighting events where missing reports may reflect recording rather than true absence Overall, presented quantitative framework proved be useful tool model occurrence. summary, this study demonstrates gain through collection analysis. It highlights role can have providing "ground truth" alongside more traditional stakeholder-led approaches.