Participatory data collation and standardized hydrometeorological indicators improve understanding of the extent and drivers of flood and drought impacts at the catchment scale

Long-term (>150 years) records of the impacts of droughts and floods at the catchment scale are rare. Here we present a novel approach to improve understanding of historical flood and drought impacts and their hydrometeorological drivers using bottom-up participatory data collation and logistic regression with generalized additive models. Applied to a groundwater-dominated catchment in Southeast England (UK), a database of historical flood and drought impacts was collated based on detailed searches by stakeholders of digital and paper literature covering 1800 to 2023. The spatiotemporal distribution of flood and drought impact reports was evaluated, and drivers of impacts assessed by development of generalized additive models using standardized hydrometeorological indices. The spatiotemporal extent of reported impacts varies between droughts and floods, and between impact classes (e.g. surface water, agriculture). When combined with month and year as co-variates, the standardized precipitation index with a 12 month accumulation period (SPI-12) is the strongest predictor (lowest Akaike Information Criterion and highest deviance explained) of reported flood and drought impacts. The performance of the generalized additive models is better (higher deviance explained, adjusted R2 and area under curve) than previous studies. There are no clear threshold values of SPI-12 associated with reported flood and drought impacts. For months where SPI-12 = −2/+2, the modelled weighted probability of a reported drought/flood impact is 0.46/0.26. We demonstrate the strengths and limitations of participatory approaches to understanding floods and droughts. Application of the developed protocol can improve understanding of relationships between indicators and impacts in catchments worldwide.