Originally, crop growth models were focused on rendering and summarizing insight in the complex interaction of different physiological processes with the environment. With the increasing emphasis on sustainable soil policies, crop growth models have also become a tool for monitoring agricultural systems, for evaluating soil quality and for making rational land use decisions. In the past three decades, comprehensive models had very limited operational value in many developing countries because of the massive data needs. With the recent establishment of geographic information institutes in several developing countries and the organization of the available climatic and edaphic resources into centralized databases, the minimum data requirements for crop growth modeling will be met much more easily in the future. A major challenge is therefore to design a crop growth model capable of revealing farmer-oriented land use constraints and sustainable management policies, making optimal use of the increased availability of multi-temporal climatic records and spatially variable edaphic characteristics.

This research resulted in the development of DAICROS, a transparent, functional, and daily crop growth model, incorporting the soil-water-plant-environment interactions affecting soil water availability and crop growth. The interaction between meteorological, edaphic and crop specific factors is analysed through a daily multi-layered water balance, describing the cyclic movement of water in the cropped field. The process-based model describes evaporation, transpiration, surface storage and run-off, infiltration and percolation. Water uptake by a one-dimensional root system is differentiated with the rooting depth, and the transpiration rate is affected by both water and oxygen availability. The actual transpiration rate is feeded into a crop growth model estimating the corresponding biomass production, under the reported radiation and temperature regime. This biomass production model simulates the photosynthesis rate, respiration rate and expansion of the crop canopy.

In the absence of experimental plot data, validation of the modeling results does not comprehend a straightforward comparison of modeled and actual water dynamics and crop yields. Sensitivity analysis of the model parameters and comparison of the model behaviour with the yields reported in agricultural statistics of the country, proved to be equally valuable. The behavioral analysis of the key variables illustrates the reliability of DAICROS under different sets of environmental conditions. Successful validation was based on natural resources databases ranging from semi-arid to humid tropical and subtropical climates, restricting the actual domain of relevance to these specific environments.

Performance of DAICROS is evaluated by simulating the soil water dynamics and crop production in several agricultural land use systems of Rwanda. Different land use scenarios illustrate the impact of temperature, rainfall distribution, slope gradient, soil depth, soil water holding capacity and management practices - the crop and sowing date selection - on the productivity of cereals, tubers, legumes, and oil crops when grown in the very different agro-ecological zones of Rwanda. Due to the large interannual variability in rainfall patterns, farmers face a dilemma when determining the optimal sowing date for the crops of the main agricultural seasons. The crop growth model illustrates the delicate equilibrium between sowing date, crop cycle length, and the length of the rainy seasons, limiting options for intensification. It proves successful in identifying the different driving forces affecting land use choices made by the Rwandan farmers.