Dheed : a global database of dry and hot extreme events

Method

Hot and dry extreme events are detected using ECMWF climate reanalysis (ERA5) hourly data. Hot conditions are assessed through daily maximum 2m temperature (Tmax). Dry conditions are assessed through daily differences between total precipitation and evapotranspiration averaged over the preceding 30, 90 and 180 days (PEI_30, PEI_90, PEI_180). Conditions at a specific location (spatial grid cell) are considered extreme when Tmax is greater than the 99th percentile and a drought indicator is less than the 1st percentile of their respective distributions over time for that specific spatial grid cell. These Discrete Extreme Occurrences (DEO) are stored in the EventCube. Through a connected component analysis, connected DEOs are assigned unique labels when they are contiguous in space or time. These labels are stored in the labelcube. Summary statistcs for all uniquely labelled extreme events are stored in table EventStats. 

For a graphical overview of the method, see the workflow.

Limitations

Heatwaves and drought conditions are assessed on the empirical cumulative distribution functions of the full time series, that is,

there is no reference period

there is a unique threshold for all days (no seasonality)

Compound hot and dry conditions are assessed from the univariate temporal distribution of temperature and drought indicators. Multivariate distributions are not considered.

Changes in Dheed 3.0

The analysis includes year 2022.

Data from ERA5 preliminary back-extension from 1950 to 1978 was updated to ERA5 final product for those years.

A minumum duration of three consecutive days is set for heatwaves.

Changes in Dheed 4.0

The analysis includes year 2023.

Data from ERA5 has been completely reprocessed. An error in the processing of the ERA5 data introduced a monthly bias in previous versions. These should not be used.