Practical assessment of soil degradation on smallholder farmers' fields in Zimbabwe: Integrating local knowledge and scientific diagnostic indicators

Crop production on smallholder farms in Southern Africa is increasingly being constrained by deteriorating land quality. Assessing soil degradation at farm-level is therefore important to enable formulation and better targeting of rehabilitation options. In this study, farmers' local indicators and scientific diagnostic parameters were combined to develop criteria for assessing soil degradation on croplands in Hwedza smallholder farming area in Eastern Zimbabwe. Farmers used common weed species, crop performance aspects and soil physical attributes to categorize arable soils into four main productivity classes: productive, moderately productive, degraded and severely degraded. Broad-leafed weeds, that included Commelina benghalensis L., Bidens pilosa L. and Leucas martinicensis (Jacq.) Ait. f., were considered as main indicators of productive soils, while grass weeds, Melinis repens (Willd.) Zizka and Eragrostis minor Host, were perceived as typifying severely degraded soils. Maize grain yields of 6 to 8 scotch carts ha⁻¹ (3 to 4 tha⁻¹) and < 2 scotch carts ha^{− 1} (< 1 tha^{− 1}) were indicative of productive and degraded soils, respectively. Consistent with farmer criteria, broad-leafed weeds contributed most of the weed biomass on productive soils, with diversity higher on productive and moderately productive than on degraded soils. Properties of soils sampled from field productivity classes differed significantly (p <0.05) and the magnitude of the differences was also influenced by farmer resource endowment. Soil organic carbon (SOC), mineralizable nitrogen (N), available phosphorus (P), exchangeable calcium (Ca) and effective cation exchange capacity (ECEC) decreased from productive to severely degraded soils, and from resource-endowed (RG1) to resource-constrained (RG3) farms. Most of the soil properties showed a decline (negative degradation indices) when compared with uncultivated lands. Exchangeable Ca and mineralizable N showed a strong relationship with SOC (R² > 0.5) across the soil productivity classes and farmer resource groups. Principal components analysis of 16 soil physico-chemical and biological properties revealed mineralizable N, microbial biomass N, exchangeable bases and available P as the most important parameters influencing soil productivity. Drawing on the relationship between weed species populations and soil properties observed in this study, a soil degradation assessment scheme is proposed for enhancing decision-making on options for rehabilitation of croplands by smallholder farmers in similar agro-ecologies in Southern Africa.