Can conservation agriculture improve crop water availability in an erratic tropical climate producing water stress? A simple model applied to upland rice in Madagascar

Family farms in the tropics mainly rely on rainfed agriculture. Water availability is limited to rainfall and is one of the main constraints to crop productivity. Conservation agriculture (CA) is promoted as an alternative that, among other functions, enhances water infiltration and limits evaporation from the soil thanks to a mulch of crop residues left on the soil surface. These functions are assumed to reduce the water availability constraint by limiting water stress during crop growth. But the variability of rainfall distribution combined with the wide range of agroecological conditions and the variety of crop husbandries in the tropics makes it difficult to evaluate the efficiency of mulching. The aim of this study was to capture the variability through a simple modeling approach using the crop growth model PYE-CA, which requires a limited set of parameters and a virtual experiment (VE). We applied our approach to a case study of upland rice in the Lake Alaotra region in Madagascar, where rainfall distribution is highly variable. The VE used a 17-year series of weather data with a range of soil water conditions, sowing dates, and growth and yield limitations that cover the variability of agroecological conditions and management systems in the study area. The VE revealed that variable successions of wet and dry episodes during the rainy season resulted in both water stress and an increase in deep drainage. In the majority of conditions simulated, enhancing water infiltration through CA mainly increased water loss through drainage. However, better water infiltration may also reduce the production risks involved in early sowing or crop intensification, thereby offering new opportunities to farmers. As an alternative to time consuming and labor intensive experimentation, we propose a suitable modeling approach to identify the main drivers of rainfall × crop interactions that could be extrapolated to other regions in the tropics. (Résumé d'auteur