10th International Rainwater Catchment Systems Conference
"Rainwater International 2001"
Mannheim, Germany - September 2001

Section 2: Rainwater Catchment in Humid and Arid Regions

Paper 2.30

Investigating the Feasibility of RWH Under Semiarid Conditions: Benefits of Computer Simulation in Adding Value to Field Experiments

Omari Mzirai, J.W. Gowing, N. Hatibu, F. B. Rwehumbiza, M.D.B Young
Soil-Water Management Research Group
Faculty of Agriculture
Sokoine University of Agriculture
Morogoro
Tanzania
email swmrg@suanet.ac.tz orswmrg@yahoo.co.uk

Agricultural production in semi- arid areas depends on direct rainfall, which is highly variable with respect to both time and space. Significant variation in the temporal and spatial distribution of rainfall usually results in serious shortage of soil-water available to plants and thus poor crop and livestock production. To overcome this problem, feasibility of RWH was investigated in representative areas of semi- arid lands in Tanzania. Analysis of rainfall characteristics was done in these areas to reveal different rainfall variability parameters (i.e. start and cessation of rainy season, duration of season, total rain, and rainy days). The results revealed that rainfall was low and highly variable with CV varying from 20% to 50% during long rainy season and 42 -50% during the short rainy season. It was also found that in some areas, dry spells of 15 days may occur during both the short and the long rainy season with probability 30%.

Though efforts have been made through experimental research to investigate RWH options, meaningful extrapolation of results has been a big problem. A twin-track approach was used in this project, where experimental work was linked to the development of simulation model (PARCH-THIRST) designed to assess how any particular RWH innovation is likely to perform. The potential of the model to overcome problems of spatially and temporally extrapolating experimental results and delivering results of direct relevance to farmers and extension services is presented in this paper. The results of simulation, shows that the model proved to be reliable in prediction with high accuracy. Performance of cropping systems with and without RWH is demonstrated on the basis of scenario simulations. The twin-track approach introduced additional requirements into the experimental effort in the short-term in order to provide all the data necessary for developing the model, but it can be concluded that this burden was worthwhile due to added value for the work as it makes research better and more efficient. It provides more complete understanding and more accurate predictions than it would be the case for field research alone unless it was continued over a long period and replicated on a large number of sites.

PDF of full document available to members (1pp, 8kb)