Assessing production gaps at the tree scale: definition and application to mango (Mangifera indica L.) in West Africa

Abstract

Fruit tree productivity is generally low and highly variable in the tropics. Quantifying yield gaps and their determining factors provides levers to increase production, but yield gap of fruit trees remains poorly explored at the tree scale. In order to bridge this knowledge gap, we adapted the concept of yield gap to define the production gap at the tree scale to integrate tree endogenous factors (cultivar, age, and crown dimensions). The production gap (P_g) was defined as the difference between potential tree production (P_p) and actual tree production (P_a). The concept was implemented for mango trees in West Africa. We estimated these production indicators on 280 mango trees for two years, covering a wide range of fruit load, age, cultivars, and cropping systems (extensive, diversified, and intensive) found in Senegal. Actual production was estimated yearly using ground tree image analysis. Attainable production (P_att, a locally constrained approximation of P_p) was estimated based on tree endogenous and climatic factors using stochastic frontier analysis. Our results showed that attainable production increased with tree crown area and trunk diameter, whereas the effects of tree density, temperature, and solar radiation were cultivar-dependent. On average, the actual production reached 63% of the attainable production. The production gap was higher in extensive orchards (P_g = 58% of P_att) compared to diversified (P_g = 29% of P_att) and intensive (P_g = 32% of P_att) orchards. Based on production gap estimation, we identified the drivers of mango production variation among cultivars and cropping systems in West Africa. This study demonstrates the usefulness of adapting yield gap methodology to the tree scale to analyze production gaps in diverse fruit tree-based cropping systems. Measurement of production gaps allows the integration of tree features and their variability to upscale and improve the estimation of yield gaps at the orchard scale.