Runoff harvesting potential for crop production in Kitui, Kenya
A 24 year rainfall data study was done for Mutomo, southern Kitui. Seasonal rainfall was found inadequate in 71% and 57% of the dry "long" and "short" rains, respectively. be tween years, Mon th Iy ra in fa II with coefficient was highly variable of variation often above 100%; there were also many small storms compared to big ones as evidenced by large data skewness. The "long" rains last shorter than the "short" rains, and by virtue of seasonal total, the former was classified as unfavourable for a crop like maize. On average, 58% and 67% of rainfall in April and November were found to be potentially runoff-producing. These results justified water harvesting for supplementing rainfall to improve maize production in the area. A water harvesting experiment for maize production was conducted at the Mutomo site in the "short" rains of 1990 on a field of 5% slope. The experimental area consisted of twelve plots in three blocks of four plots each. Treatments were 0 (control), 13.6, 19.6 and 27.2 m2 of sloping catchment areas above each 16 m2 level cropped area. The soil was deep with no apparent restriction within 150 cm. The soil texture was sandy clay for the top 60 cm and clay below this depth. Bulk densities varied from 1490 for the top 15 cm to 1540 kgm-3 from 60-105 cm. Infiltration tests conducted with double ring infiltration xv infiltrometer gave a range rat es 0 f 14 .2 t0 75 .5 mm h-1• of final The mean organic matter content was 3.68%. The soil was high in base saturation and low in macro nutrients especially ni trogen and phosphorus, but showed no hazards of salinity or sodicity. Soil sampling for moisture determination was done to 105 cm depth every ten days. A water balance was done for each treatment for 82 days of the rainfall season. Runoff simulation was done using a portable rainfall simulator. Computer generated daily runoff values were determined for each rainfall event. The rainfall was 606.3 mm and pan evaporation was 341.9 mm. High intensity rainfall resulted in a high mean runoff of 61%. Computer generated runoff ranged from 8.5% to 84% early in the dry and in the wettest periods of the season, respectively, showing high sensitivity to "antecedent mois~ure. The total runoff was estimated at 287.0, 259.7, 185.9 and 0 mm and the percolation below root zone at 368.5, 341.9, 268.6 and 84.4 mm for the 27.2, 19.6, 13.6 and 0 m2 catchments, respectively. Soil moisture increased to saturation in the wet period, and reduced to 183 mm at the end of the season. Yields of Katumani maize grown at the experimental site were not significantly different at P=0.05, a fact attributed to the above average rainfall In the season.