Compound drought and heatwave identification: daily-scale independent extreme events based on 120-year observations

Under the challenge of climate change, the extremes, especially for extreme temperature, are observed at an increasing pace and are expected to be more severe in the future. It is critical to study heatwaves concurrently with droughts because of the intensification of negative impacts, such as exacerbating water shortage, crop failure and GPP reduction, wildfire and tree mortality, etc. This research focuses on compound events of droughts and heatwaves and presents a framework for the identification of drought or heatwave events and their compounds.

While most studies only look at the summer season, we also consider compound drought and heatwave events in the winter season, as these are also important in view of their significant influence on wildfires, insect outbreaks, seed germination, etc.

We introduce the notion of "relative heatwave" as being an extreme event compared with the average of the previous 30-year temperatures for that period. Drought and heatwave events are then identified based on SPI (standardized precipitation index) and SHI (standardized heatwave index). To overcome limitations arising from the scale inconsistency (monthly drought with daily heatwave) and coarse resolution (monthly or weekly drought), we apply the daily SPI and daily SHI, bringing a more accurate measure of the start and end dates, and severity. We also propose an objective, convenient and robust method to identify the statistically extreme and independent drought and heatwave events. Thresholds for removing small-scale events and merging proximate events are found by assuming the severity of the events to follow a generalized extreme value distribution and their arrivals to follow a Poisson process. Finally, we introduce four possible ways of identifying compound events (union, conditioned on drought, conditioned on heatwave, and intersection).

To demonstrate our methodology, we made use of 120 years of daily precipitation and daily average temperature observed at the Belgian meteorological institute in Uccle, near Brussels.