In collaboration with the European Centre for Medium-Range Weather Forecasts (ECMWF), the Joint Research Centre (JRC) of the European Commission has produced an upgraded hydrological reanalysis data set of the Global Flood Awareness System (GloFAS).
The system forms part of the Copernicus Emergency Management Service (CEMS). It has been created to make information available about the river discharge and flood events from 1980 to July 2022. The GloFAS v4.0 reanalysis includes daily maps of discharge around the world at a resolution of 0.05° (about 5km). It is available in the Climate Data Store of the Copernicus Climate Change Service (C3S), run by ECMWF. The reanalysis makes it possible to study flood events and droughts globally during a longer time frame than the period during which GloFAS has been operational.
Generating a high-resolution reanalysis has been made possible by ECMWF’s new computing facility in Bologna, Italy. It is a new edition of the GloFAS v3.1 reanalysis, which was introduced in 2021. The new data set uses ERA5 (ECMWF’s latest reanalysis of the atmosphere), satellite-derived data sets, and a large number of ground measurements to describe catchment physical properties and for model calibration, combined with the LISFLOOD hydrological model.
The open-source LISFLOOD hydrological model determines what happens with water that comes down as rain. It decides to what extent the water evaporates, is absorbed by the soil and plants, or runs off into rivers. For the reanalysis upgrade, ECMWF and the JRC, which manages CEMS, both worked on LISFLOOD. This included upgrades in hydrological routines and improvements in the management of large input data sets and computational performance.
ECMWF worked to improve the representation of rivers, soil and vegetation in the model, and the JRC calibrated the model and performed a parameter regionalization for ungauged catchments to ensure the best possible simulation of river flows for all catchments around the world.
The new GloFAS v4.0 hydrological reanalysis also has a finer resolution than before – it works on a 0.05° grid (about 5km) when before it used a 0.1° grid (about 10km).
Significant changes in computational performance, made possible by the parallelization of routines in the hydrological model LISFLOOD, combined with ECMWF’s new Atos HPCF, have enabled the global simulation at the increased spatial resolution.
The initial release of the new data set makes river discharge data available from 1980. In early 2023, this will be widened to other variables, such as soil moisture and snow water equivalent. From the same time, the new GloFAS v4.0 hydrological reanalysis will be updated in near real time, with a delay of about three days, to include the latest available forcing data.
The reanalysis can be used to monitor hydrological status, in particular water resources, floods and drought evolution across the world. Christel Prudhomme, the leader of the environmental forecasts team at ECMWF, who is responsible for the CEMS hydrological forecast computational center that runs GloFAS, said, “A lot of places have no data on river discharge. A country may for example have no way of knowing how much water is going to be discharged into it from rivers outside its borders. Global simulations thus close an information gap. They provide a first guess of how much water is running in rivers.”
There are also some specific applications, such as the C3S climate bulletin. In addition, the Copernicus Marine Service is investigating whether its ocean modeling can be improved by real-time estimates of river discharge. The GloFAS v4.0 hydrological reanalysis could thus provide input for an ocean model.
The organization expected the data set as a whole to be useful for identifying and understanding trends in hydrology, in particular, whether there are times during which some regions have had higher or lower river flows than usual.
The reanalysis was completed before a major upgrade of the GloFAS forecasting system to higher resolution planned for 2023. The new reanalysis will then be used for the initial conditions of GloFAS forecasts. It will thus provide input for early warnings provided by GloFAS forecasts.