Crop conceptual model for predicting productivity of bread wheat in semi-arid Kenya


  • Paul K Kimurto Egerton University, Njoro
  • Klaus Gottschalk Leibniz-Institut fuer Agrartechnik Potsdam-Bornim ATB
  • M.G. Kinyua , Moi University, Dept of Plant Breeding and Biotechnology P.O. Box 39000, Eldoret
  • J.B.O. Ogola University of Venda, Department of Plant Production, a, Private bag X5050
  • B. K. Towett Egerton University, Dept. of Crops, Horticulture & Soil Sciences, P.O. Box 536, Njoro



P. K. Kimurto1, K. Gottschalk2, M. G. Kinyua3, J. B. O. Ogola4, B. K. Towett 1

(1. Department of Crops, Horticulture & Soil Sciences, Egerton University, P.O. Box 536, Njoro, Kenya;

2. Leibniz-Institut für Agrartechnik Potsdam-Bornim e.V. ATB, Max-Eyth-Allee 100, 14469 Potsdam, Germany;

3. Department of Plant Breeding and Biotechnology, Moi University, P.O. Box 39000, Eldoret, Kenya;

4. Department of Plant Production, University of Venda, a, Private bag X5050, South Africa)


Abstract: Carrying out field trial-research in dryland areas is usually expensive and costly for most national breeding programmes; hence development of simple crop simulation models for predicting crop performance in actual semi-arid and arid lands (ASALS) would reduce the number of field evaluation trials.  This is especially critical in developing countries like Kenya where dry areas is approximately 83% of total land area and annual rainfall in these area is low, unreliable and highly erratic, causing frequent crop failures, food insecurity and famine.  This paper used data generated from the rain shelter by measurement of evapotranspiration together with weather variables in Katumani to predict wheat yields in that site.  Maximum yield of the wheat genotype considered for genotype Chozi under ideal conditions was 5 t/ha.  Total above-ground biomass was obtained and grain yield was to be predicted by the model.  Transpiration was estimated from the relationship between total dry matter production and normalised TE (7.8 Pa).  The results presented are based on the assumption that all agronomic conditions were optimal and drought stress was the major limiting factor.  Predicted grain yield obtained from the conceptual model compares very well with realised yields from actual field experiments with variances of 14% – 43% depending on watering regime.  This study showed that it is possible to develop simple conceptual model to predict productivity in wheat in semi-arid areas of Kenya to supplement complicated and more sophisticated models like CERES-maize and ECHAM models earlier used in Kenya.  The presence of uncontrolled factors in the simulation not accounted for in the estimation and could have contributed to decrease in observed yield need to be included in the model, hence modulation of the equations by introducing these factors may be necessary to reduce variances; thus need to be quantified.  To improve the accuracy of prediction and increase wheat production in these areas measures that conserve water and/or make more water available to the crop such as prevention or minimisation of run-off, and rain water harvesting for supplemental irrigation are necessary.

Keywords: wheat, conceptual model, drought, evapotranspiration, yield response


Citation: Kimurto P. K., K. Gottschalk, M. G. Kinyua, J. B. O. Ogola, and B. K. Towett.  Crop conceptual model for predicting productivity of bread wheat in semi-arid Kenya.  Agric Eng Int: CIGR Journal, 2010, 12(3): 25-37.


Author Biography

Klaus Gottschalk, Leibniz-Institut fuer Agrartechnik Potsdam-Bornim ATB

Post Harvest Technology





I-Land and Water Engineering