Application of Remotely Sensed Rainfall Data in Rainfall-Runoff Modelling. A Case of Pangani River Basin, Tanzania

Abstract

Rainfall runoff modelling in a river basin is vital for number of hydrologic application
including water resources assessment. However, rainfall data from sparse gauging stations
are usually inadequate for modelling which is a major concern in Tanzania. This study
presents the results of comparison of Tropical Rainfall Measuring Mission (TRMM)
satellite rainfall products at daily and monthly time-steps with ground stations rainfall
data; and explores the possibility of using satellite rainfall data for rainfall runoff
modelling in Pangani River Basin, Tanzania. Statistical analysis was carried out to find the
correlation between the ground stations data and TRMM estimates. It was found that
TRMM estimates at monthly scale compare reasonably well with ground stations data.
Time series comparison was also done at daily and annual time scales. Monthly and annual
time series compared well with coefficient of determination of 0.68 and 0.70, respectively.
It was also found that areal rainfall comparison in the northern parts of the study area had
poor results compared to the rest of areas. On the other hand, rainfall runoff modelling
with ground stations data alone and TRMM data set alone was carried out using five Real-
Time River Flow Forecasting System models and then outputs combined by Models Outputs
Combination Techniques. The results showed that ground stations data performed better
during calibration period with coefficient of efficiency of 76.7%, 81.7% and 89.1% for
Simple Average Method, Weight Average Method and Neural Network Method respectively.
Simulation results using TRMM data were 59.8%, 73.5% and 76.8%. It can therefore be
concluded that TRMM data are adequate and promising in hydrological modelling.