ARCEME-DHP: Database of atmospheric drought events interrupted by heavy precipitation

Events are defined based on time series of ERA5-Land [1] variables, aggregated daily by time zone.

Daily indicators of evaporative balance

Input data are daily aggregates of ERA5-Land [1] total precipitation and reference evapotranspiration (from Singer et al. 2020 [2]).

$PE = Precipitation - ReferenceEvapotranspiration$

Accumulation period $p$: 30, 90, 180, 360 days

At time $t$ (in local time):

$PE_{p,t} = \frac{\sum_{i=0}^{p-1} PE_{t-i}}{p}$

Daily indicator of heavy precipitation

For the detection of heavy precipitation, we use the daily maximum of a moving sum over 24h (MP) instead of the total daily precipitation (TP) to ensure that heavy precipitation spreading over two days is flagged as extreme and that the detected event starts on the day in which the heavy precipitation effectively starts.

Quantiles

Drought

For each grid cell $(x,y)$, day-of-year (as Month-Day) quantiles are computed for a reference period (1991-2020), with 31 days sliding windows centered on day-of-year $d$. 

$q_{d,x,y,p,i}$ for $i \in \left\{0.01, 0.05, 0.1, 0.15, 0.25, 0.5, 0.75, 0.85, 0.9, 0.95, 0.99\right\}$

For February 29, less values are used to compute the quantiles. 

Heavy precipitation

For precipitation, we do not deseasonalize the time series. The quantiles are computed on all values of daily total precipitation ($TP$) greater than 0.0001 meter (rain days) over the reference period (1991-2020).

$q_{x,y,TP,i}$ for $i \in \left\{0.01, 0.05, 0.1, 0.15, 0.25, 0.5, 0.75, 0.85, 0.9, 0.95, 0.99\right\}$

Event detection

Drought

A 0.05 threshold is applied: all $PE$ values below the threshold are flagged as extreme. 

$DEO_{t,x,y,p} = $    \begin{cases}
        1 & \text{if } PE_{t,x,y,p} < q_{d,x,y,p,0.05}\\
        0 & \text{otherwise}
    \end{cases}

To ensure that current conditions are dry, we set an additional condition for $PE_{30}$ in the previous 5 days. 

$DEO_{t,x,y,30} = $
    \begin{cases}
        1 & \text{if } PE_{t-i,x,y,30} < q_{d-i,x,y,30,0.05} \text{ for } i \in [1,5]\\
        0 & \text{otherwise}
    \end{cases}

The Discrete Event Occurrences (DEO) are saved in a data cube, encoding the different indicators on different bits.

Heavy precipitation

We use a local threshold (daily total precipitation greater than 95th percentile of daily total precipitaion of raindays over reference period) combined it with a global threshold. The global threshold is the latitude weighted mean of the 75th percentile of daily total precipitation for all land grid cells.

$DEO_{t,x,y,TP} =$ 
    \begin{cases}
        1 & \text{if } MP_{t,x,y} > q_{x,y,TP,0.95} \text{ AND } MP_{t,x,y} > q_{0.75,TP|\text{land}}\\
        0 & \text{otherwise}
    \end{cases}

Events labelling

DEOs are combined and labelled as individual events through a connected component analysis. The current combination follows the rule: any $DEO_{t,x,y,p}$ for $p \in \{90,180,360\}$ AND $DEO_{t,x,y, 30}$ AND $DEO_{t,x,y,TP} $. 

Database

The database consists of
- a zarr Event Cube storing the DEOs of the individual indicators of drought and heavy precipitation from 1991 to 2023 
- a zarr Label Cube, with the labelled connected events from 2015 to 2023, obtained from a connected component analysis.
- a csv summary table with statistics describing the labelled events.

Checksum

7e0d40779e908ebf1f38c0c8ddf28d8b5003a89adaad6dcc6d7ad4855aaedafe  labelcube_dhp_era5land_qdoy0.05_ref1991_2020.zarr.zip
6cc4616645e6e157c091680245495e5e5191ee03ba3bcbfa316bd35ac01fbbef  dhp_era5land_eventcube_qdoy0.05_ref1991_2020.zarr.zip
76cbf56199426241ea625d69fd9b31e8defff715ce5e692177858c3bad6a9410  dhpEventStats_2015_2023_qdoy0.05_ref1991_2020.csv.zip

References

[1] Muñoz-Sabater, J., Dutra, E., Agustí-Panareda, A., Albergel, C., Arduini, G., Balsamo, G., Boussetta, S., Choulga, M., Harrigan, S., Hersbach, H., Martens, B., Miralles, D. G., Piles, M., Rodríguez-Fernández, N. J., Zsoter, E., Buontempo, C., & Thépaut, J.-N. (2021). ERA5-Land: A state-of-the-art global reanalysis dataset for land applications. Earth System Science Data, 13(9), 4349–4383. https://doi.org/10.5194/essd-13-4349-2021

[2] Singer, M. B., Asfaw, D. T., Rosolem, R., Cuthbert, M. O., Miralles, D. G., MacLeod, D., Quichimbo, E. A., & Michaelides, K. (2021). Hourly potential evapotranspiration at 0.1° resolution for the global land surface from 1981-present. Scientific Data, 8(1), 224. https://doi.org/10.1038/s41597-021-01003-9

Citation

Melanie Weynants, Khalil Teber, Miguel D. Mahecha, Fabian Gans. 2025. ARCEME-DHP Databse of drought events interrupted by heavy precipitation. https://doi.org/10.5281/zenodo.15705050

Funding

This work was funded by the European Space Agency in response to ESA CfP/5-50033/23/I-KE.