Informations et ressources scientifiques
sur le développement des zones arides et semi-arides

Accueil du site → Master → Kenya → 1998 → A HYDROLOGIC STUDY OF WATER HARVESTING MICRO CATCHMENTS IN A SEMI ARID AREA

Moi University (1998)

A HYDROLOGIC STUDY OF WATER HARVESTING MICRO CATCHMENTS IN A SEMI ARID AREA

Katana, Samuel J. Sirya

Titre : A HYDROLOGIC STUDY OF WATER HARVESTING MICRO CATCHMENTS IN A SEMI ARID AREA

Auteur : Katana, Samuel J. Sirya

Université de soutenance : Moi University

Grade : Master of Philosophy 1998

Résumé
Micro catchment water harvesting is a method of collecting surface run off from a catchment area over a flow distance less than 100 m. Soil and rainfall characteristics have important role to playas they influence the process of infiltration and thus the runoff generation in the catchment area. The study, therefore, focussed on evaluating the rainfall parameters describing the temporal variation in intensity and the soil parameters namely hydraulic conductivity, the wetting front suction and fillable porosity influencing the infiltration process. The research study was carried out on bare runoff plots of both rectangular and triangular shapes of area 10 m2 each constructed on a natural slope of 5 % on a typical loamy sand soil in Chepareria Division of West Pokot District, Kenya. The study involved the measurement of rainfall, runoff and soil moisture. Rainfall was measured using both automatic and standard raingauges. The autographic data was used for storm characterization. Runoff was collected in plastic buckets located at the lower end of each runoff plot and measured using a graduated cylinder and runoff efficiency computed. Soil samples were taken both before and after each storm event and the soil moisture content determined using oven drying method. In addition, the spirit burning method of soil moisture determination was evaluated against the oven drying method by taking soil samples in pairs. Computer programs in Basic were developed based on the original Green and Ampt model (GA) for immediate ponding, Green-Ampt-Mein-Larson infiltration model (GAML) for ’steady rain and modified GAML for the case of unsteady rain (GAMLM) to predict runoff. The parameters of the models namely hydraulic conductivity, K, and average wetting front suction, Say, were estimated using texture based regression equations emanating from United States Department of Agriculture. The parameters were also evaluated by calibration involving actual rainfall and runoff data. The curve number (CN) method was also investigated for applicability in predicting runoff from bare micro catchments through calibration. The runoff efficiencies of the storms collected varied from as low as 23.33 % to as high as 95.71 % with 75 % of the storms having efficiencies greater than 50 %. A comparison of the runoff depths from the rectangular and triangular plots indicated that shape had negligible influence on the runoff generation from the micro catchments. The storm characterization indicated that the storms could satisfactorily be described by triangular hyetographs whose empirical equations both for the rising and the falling limbs were of linear form. The values of hydraulic conductivity, K, and average wetting front suction, Say, obtained by calibration were 1.5 mm/h and 12 mm while from texture based regression equations were 36.8 mm/h and 27.6 mm respectively. The regression equation based values tended to under predict surface runoff while predictions based on calibrated values were satisfactory. A comparison of the original Green and Ampt (GA) , Green-Ampt-Mein-Larson (GAML) and modified version of GAML (GAMLM) infiltration models indicated that GAMLM predicted runoff with greatest accuracy ( R2 = .983 about 1:1 line) followed by GAML (R2 = .974) and least by GA (R2 = .969). A comparison of the spirit burning and oven drying methods of moisture determination indicated good agreement ( coefficient of determination = 0.989) . Both models showed low sensitivity to soil moisture deficit. However, the models appeared most sensitive to K and moderately to S which is in agreement with studies reported in most hydrologic literature. An investigation on the applicability of eN method for runoff prediction from bare micro catchments indicated that it could be applied satisfactorily if the catchments were treated equivalent to paved surfaces (R2 = 0.981).

Présentation

Page publiée le 30 novembre 2013, mise à jour le 13 mars 2018