GloFAS v3.1 uses the LISFLOOD hydrological model as core component as opposed to HTESSEL land surface model combined with LISFLOOD routing for earlier versions. This change could impact on the simulated river discharge regime and flood threshold magnitude used to generate the flood reporting points. Here we show the 2, 5, 10 and 20-year return period flood thresholds on GloFAS diagnostic points for both GloFAS v3.1 and v2.1.

GloFAS v3.1 flood thresholds are computed from GloFAS v3.1 historical (reanalysis) dataset for the period 1979-2019. This has an extra year compared with the GloFAS v2.1 historical reanalysis dataset used for GloFAS 2.1/2.2 flood thresholds for 1979-2018. Both reanalyses are forced with the same meteorological data (ERA5). This difference in reference length is not expected to have much influence on the return period magnitude estimates.

Analysis was conducted on GloFAS diagnostic points, with very dry catchments (i.e. with a 2-year return period flood magnitude of less than 1 m3/s in either v3.1 or v2.1) excluded, giving a total of 5859 catchments analysed.

For users who rely on fixed river discharge magnitudes for their decision making (and not climate relative values, computed from model simulations, such as the GloFAS return period flood threshold magnitudes), it might be important to review their protocol to accommodate for the changes between the GloFAS v3.1 and GloFAS v2.1 hydrological simulations.

20-year return period flood magnitude

Figure 1 shows the 20-year return period flood magnitude for GloFAS v3.1 (a) and GloFAS 2.1 (b) and their relative difference (c) on the analysed catchments.

There is a general increase in GloFAS 20-year flood thresholds between v2.1/2.2 and v3.1, shown by positive changes in Figure 1c: 75% of the catchments shows an increase against only 25% showing a decrease. Differences are large over most of the catchments, with only 20% of catchments showing a change within +/- 20%, against about 40% of catchments associated with at least a doubling of the 20-year flood threshold magnitude between GloFAS 2.1 and GloFAS v3.1, and 10% of catchment where the increase is greater than 5 times. The geographical pattern shows neutral changes (within +/- 20%) in the Amazonian basin and south-east Asia, decrease in magnitude in northern-western coast of south America, the Himalayas, and some parts in north-east Siberia, and largest positive changes dominate Australia, southern Africa, some areas in the Sahel, eastern areas in Europe and some some catchments in central Siberia. 

Figure 1. GloFAS v3.1 (a) and v2.1 (b) flood thresholds magnitude (m3/s) for 20-year return period for GloFAS diagnostic points with 2-year return period flood threshold magnitude of at least 1 m3/s (in either v3.1 or v2.1), and the relative difference between the two (c). The size of the dot is proportional to the catchment upstream area. Colour indicates threshold magnitude and difference.

10-year return period flood magnitude

The general pattern of changes in GloFAS 10-year flood thresholds between v2.1/2.2 and v3.1 is of an increase in the flood magnitude in the majority of GloFAS diagnostic points (Figure 2c). Neutral changes (within +/- 20%) are found in the Amazonian basin and south-east Asia. Areas of decrease are located in nothern-western coast of south America, the Himalayas, and north-east Siberia, whilst largest positive changes dominate Australia, southern Africa, some areas in the Sahel, eastern areas in Europe and some some catchments in central Siberia. 

Figure 2. GloFAS v3.1 (a) and v2.1 (b) flood thresholds magnitude (m3/s) for 10-year return period for GloFAS diagnostic points with 2-year return period flood threshold magnitude of at least 1 m3/s (in either v3.1 or v2.1), and the relative difference between the two (c). The size of the dot is proportional to the catchment upstream area. Colour indicates threshold magnitude and difference.

5-year return period flood magnitude

The general pattern of changes in GloFAS 5-year flood thresholds between v2.1/2.2 and v3.1 is of an increase in the flood magnitude in the majority of GloFAS diagnostic points (Figure 3c). Neutral changes (within +/- 20%) are found in the Amazonian basin and south-east asia. Areas of decrease are located in nothern-western coast of south America, the Himalayas, and north-east Siberia, whilst largest changes dominate Australia, southern Africa, some areas in the Sahel, eastern areas in Europe and some some catchments in central Siberia. 

Figure 3. GloFAS v3.1 (a) and v2.1 (b) flood thresholds magnitude (m3/s) for 5-year return period for GloFAS diagnostic points with 2-year return period flood threshold magnitude of at least 1 m3/s (in either v3.1 or v2.1), and the relative difference between the two (c). The size of the dot is proportional to the catchment upstream area. Colour indicates threshold magnitude and difference.

2-year return period flood magnitude

The general pattern of changes in GloFAS 2-year flood thresholds between v2.1/2.2 and v3.1 is of an increase in the flood magnitude in the majority of GloFAS diagnostic points (Figure 4c). Neutral changes (within +/- 20%) are found in the Amazonian basin and south-east asia. Areas of decrease are located in nothern-western coast of south America, the Himalayas, and north-east Siberia, whilst largest changes dominate Australia, southern Africa, some areas in the Sahel, eastern areas in Europe and some some catchments in central Siberia.

Figure 4. GloFAS v3.1 (a) and v2.1 (b) flood thresholds magnitude (m3/s) for 2-year return period for GloFAS diagnostic points with 2-year return period flood threshold magnitude of at least 1 m3/s (in either v3.1 or v2.1), and the relative difference between the two (c). The size of the dot is proportional to the catchment upstream area. Colour indicates threshold magnitude and difference.