Uncertainty analysis has become the standard approach to hydrological modelling, but
has yet to be effectively used in practical water resources assessment. This study of the Great
Ruaha River basin in Tanzania is based on the use of regional estimates of mean runoff,
groundwater recharge and three flow points on flow duration curves (FDCs) to constrain ensemble
outputs from the Pitman monthly model using Monte Carlo parameter sampling. The constraint
bounds were quantified from gauged data available for 26 sub-basins together with assumptions
about the spatial variations in hydrological response using limited physical sub-basin property and
climate data. The results are encouraging in that the simulated FDC ranges bracket the observed
curves at two gauging stations downstream of many ungauged sub-basins that are important sites
for water resources development decision making. Further work is required to refine the constraint
bounds, the input parameter sets and climate data and achieve a more robust model of the basin.