Effects of climate zone and land use on fruit biomass, morphology, and size-dependent fruiting of Diospyros mespiliformis in West Africa

---

Résumé

Wild fruit trees contribute substantially to food security, rural livelihoods, and climate adaptation across sub-Saharan Africa, yet most remain undomesticated and their fruit production is poorly quantified. Diospyros mespiliformis Hochst ex A.DC. is a key wild fruit tree in West African savannas, but quantitative information on its fruit biomass, morphological variability, and size-yield relationships across environmental gradients remains limited. This study evaluated fruit biomass, fruit and seed morphological traits, and allometric relationships across two climate zones (Sudanian and Sudano-Sahelian) and two land-use types (protected and non-protected areas) in Burkina Faso. A total of 185 fruiting trees (diameter at breast height ≥ 5 cm) were sampled across three diameter classes, with 60 trees in the Sudanian zone and 125 in the Sudano-Sahelian zone. Data were analysed using linear mixed-effects models, with the climate zone effect interpreted cautiously because the Sudanian zone was represented by a single site. Allometric equations relating fruit biomass to tree size were developed by comparing candidate functional forms and predictors. Fruit biomass increased strongly with tree size (P < 0.001) and was higher in non-protected than protected areas (P = 0.009). The effect of land use depended on climate zone (P = 0.028), being pronounced in the Sudano-Sahelian zone but negligible in the Sudanian zone. Fruit and seed morphological traits (length, width, weight, and shape) also varied with land-use type and its interaction with climate zone (P < 0.05 for most traits). In the Sudanian zone, trees from non-protected areas produced heavier and wider fruits than those from protected areas (7.97 vs 6.89 g; 23.0 vs 22.0 mm), whereas the two land-use types differed little in the Sudano-sahelian zone; seed traits showed trait-specific responses, and seed weight did not differ significantly. Power models best predicted fruit biomass, with mean crown diameter the single best predictor (R2adj = 41%), whereas compound predictors provided only marginal improvements in predictive performance. These findings provide practical non-destructive tools for estimating fruit biomass and offer insights for the sustainable management and early domestication of D. mespiliformis in West African savannas.

Mots-clés

Ebenaceae · Non-timber forestproducts · Agroforestry parklands · Fruit yield ·Domestication · Savanna ecosystems