Medium Ranged Flood Forecasting for Upper Baro Akobo River Basin: Testing the Applicability of European Flood Alert System (EFAS) Methodologies to African small - Medium Scale River Basins.
No Thumbnail Available
Date
2012-02
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Addis Ababa University
Abstract
Flooding has always been a critical and common disaster in Ethiopia taking the lives of many people and
livestock, displaced millions and destroyed a huge amount of the economy. The Baro-Akobo river basin -
a trans-national river basin in Ethiopia located between latitudes of 5 ° 31` and 10 ° 54` N, and longitudes
of 33 ° and 36 ° 17` E with an altitude range from 402m to 3106m - is one of the flood prone areas in the
country which have been hit by different floods. In every rainy season news on floods in most of the river
basins in the country is very common. Thus, establishing flood forecasting and early warning systems to
the basin and the country as well is important.
In this paper, the methods of the European Flood Alert System (EFAS) are tested and adapted. EFAS is a
medium range system to provide warning for flood hazards in European transnational rivers 3-10 days in
advance of the floods. The LISFLOOD GIS based distributed hydrological model - underlying the EFAS
system – was used with precipitation products from five different sources: TRMM (Tropical Rainfall
Measuring Mission product of NASA), PRESSIAN (Precipitation Estimation from Remotely Sensed
Information using Artificial Neural Networks product of University of Arizona), RFE (African Rainfall
Estimation Algorism product of National Oceanic and Atmospheric Administration - climate prediction
center ), ERA-interim (product of European Center of Medium-Range Weather Forecasts (ECMWF’s)
and observed precipitation (interpolated by invers distance method).
Long time observed discharge data from the Ethiopian Minister of Water Resource and Energy were used
for model calibration and verification. Additional input data are collected from 15 observed
meteorological stations data around and in the basin from Ethiopian National Meteorology Agency, static
data sets such as elevation data, leaf area index, soil data, land cover and others, and satellite precipitation
data from Joint Research Centre (JRC) which were mapped for the African continent at 0.1
o
grid scale.
The model was calibrated using the SCE-UA (genetically adaptive approach for effective and efficient
global optimization) algorithm. Both calibration and validation resulted in good scores for most
precipitation datasets. Hindcasting results using ECMWF weather forecast data indicated that the EFAS
method is capable to forecast small floods. Typical EFAS thresholds used in Europe need to be adjusted
to allow for sending alerts already for smaller floods (return periods of 1 year). Hindcasting with adjusted
thresholds yielded sound results and demonstrated the applicability of the system.
Description
Keywords
Flood forecasting and early warning, Baro-Akobo, LISFLOOD, EFAS