Flood return-period thresholds in GloFAS are derived from annual maximum discharge (AMAX) series extracted from the long-term river discharge reanalysis. The generation of the GloFAS flood thresholds for GloFAS v4 and earlier versions (such as the v1, v2 and v3 families) follows the below workflow:
- Extract the annual maxima time series in the long reanalysis period
- Fit a Gumbel extreme value distribution using the first two L-moments (λ1, λ2)
- The Gumbel parameters are:
...
Adjusting for non-stationarity
The standard approach described above assumes that the annual maxima are stationary over the reanalysis period. However, long-term changes in the meteorological forcing or other components of the modelling system can introduce systematic trends in the simulated annual maxima. Large trends in GloFAS river discharge have previously been documented, for example by Zsótér et al. (2020), particularly in the Congo basin and the southwestern USA. If such trends are sufficiently large, fitting a stationary extreme-value distribution directly to the complete historical sample may result in flood thresholds
...
that are not representative of the more recent discharge regime and, consequently, of the real-time forecasts.
Up to GloFAS version 4, including the v1, v2 and v3 model families
...
, flood thresholds were generated directly from the unadjusted annual maxima.
With the introduction of GloFAS v5 in 2026, an annual maxima adjustment methodology was introduced
...
to address the
...
impact of non-stationarity,
...
particularly large long-term trends in river discharge that are present in some regions of the GloFAS reanalysis.
The adjustment methodology
...
used
...
from GloFAS version 5
...
onwards is described
...