GEO-SPATIAL PREDICTIVE MODELING OF LAND SUITABILITY FOR THE DEVELOPMENT OF INLAND VALLEYS RICE-BASED SYSTEMS IN TOGO AND BENIN

ABSTRACT

Rice has become one of the main staple foods in Sub-Saharan Africa. Yet, domestic rice production never satisfied demand. Inland valleys in West Africa are important landscapes for rice cultivation and are targeted by national governments to attain self-sufficiency. However, spatial studies that assess the suitability of inland valleys (IVs) for rice cultivation are limited. Agricultural land suitability analysis for crop production is one of the key tools for ensuring sustainable agriculture and for attaining the current global food security goal in line with the Sustainability Development Goals. We developed an ensemble model approach to map rainfed lowland rice suitability using four machine learning algorithms, namely Boosted Regression Trees (BRT), Generalized Linear Model (GLM), Maximum Entropy (MAXENT) and Random Forest (RF) based on presence-only environmental niche models (ENMs). Our modeling shows that RF and GLM produced better generalizability compared to MAXENT and BRT. The topographical variables, climate covariates as well as soil water content parameters are major predictors of suitable IV condition for rainfed rice cultivation. Besides, the proximity of IVs to roads and urban centers provide an additional suitable condition for rice production. This study demonstrated that Togo and Benin have enough IVs suitable to meet their domestic rice production. Our modeling results in 155,000-225,000 Ha of the suitable IV in Togo; corresponding to 34.44-50% of the total IVs area. We estimate that 53.8% of the suitable IV area (121133.4 Ha) is needed for Togo to attain self-sufficiency in rice while the remaining 103,866.6 Ha could be safeguarded for other purposes. Benin suitable inland valleys represent 351,000-406,000 Ha corresponding to 40.81-47.21% of the total inland valley area. We estimate that 60.1% of the Benin suitable IV area (244149.7 Ha) is needed to attain self-sufficiency in rice while 161,850.27 could be used for other purposes. Climate change will play a significant role in attaining food security and rice self-sufficiency in West Africa especially in a context whereby most of the crop production is based on the rainfed system. However, its effects on rainfed IVs suitability for rice production in terms of direction and magnitude of change are not fully understood. We quantitively modeled future climate change impact on rainfed lowland rice suitability based on the previously mentioned methods. We considered four time periods (the 2030s, 2050s, 2070s, and 2080s) of four Representative Concentration Pathways (RCP2.6, RCP4.5, RCP6.0, and RCP8) from an ensemble set of thirty-two spatial downscaled GCMs climate data. Our modeling showed a significant reduction in potentially suitable IVs areas as early as the years 2030s up to more than 50% losses by the end of the century in Togo and Benin. The direction of change in IVs suitability is a decrease in maximum suitability levels and statistically significant losses of suitable IVs areas. Areas losses are linked to changes in temperature regime (annual mean temperatures and relative oscillation of night to day temperatures). Areas that remain stable (no loss) or become suitable (gain) are linked to topography (cooling effect) and more favorable rainfall conditions.  With current rice production far from meeting national demands coupled with rapid population growth and dietary shifts, strong adaptation measures along with technological advancement and adoption are needed to cope with the adverse effects of climate change on wetland rice areas.