New version of the African Flood and Drought Monitor live…
African Flood and Drought Monitor (AFDM)
Access the system here: http://hydrology.soton.ac.uk/apps/afdm/
The African Flood and Drought Monitor (AFDM) is an operational system for early warning of flood and drought conditions across the country. It has been developed by Princeton Climate Institute (PCI) in collaboration with University of Southampton and Princeton University, with funding support from UNESCO Intergovernmental Hydrology Programme (IHP) and the International Center for Integrated Water Resrouces Management (ICIWaRM). The system is based on a set of ground, satellite and modeled datasets, which are combined to provide a consistent picture of hydrological conditions close to real-time, as well as forecasts out to 10-days for floods and out to 6 months for drought.
The system is operational and is updated every day, about 1/2 day behind real-time. It runs a hydrological model at 5km resolution that is forced by a hybrid reanalysis-satellite dataset of precipitation and temperature. The model runoff is routed through a vector river model to produce estimates of streamflow at 1000’s of river reaches across Africa. The model outputs are used to estimate drought indices, which are also updated every day. Short-term forecasts of flood and drought variables are generated daily, based on the NOAA Global Ensemble Forecast System (GEFS). The GEFS provides an ensemble of precipitation and temperature forecasts, which are downscaled to 5km resolution and bias-corrected to remove any biases from the GEFS model. These forecasts are used to drive the hydrological model to generate an ensemble of hydrological forecasts and flood/drought variables for the next 10 days.
Once a month an ensemble of seasonal forecasts is made of precipitation and temperature, which are used to drive the hydrological model out to 6 months into the future. Currently the system uses the ECMWF climate forecast system to produce the precipitation and temperature forecasts, which is one of the best performing models for the region, and has 51 ensemble members. The precipitation and temperature data are downscaled and bias-corrected. The hydrological model outputs are used to calculate seasonal forecasts of drought indices and other statistics such as probability of drought.
