Selection of Model Homegardens: Does the District Heterogeneity Classifies the Homegardens?

Among numerous studies done with homegardens (HGs), few efforts have been drawn up to examine biodiversity, input availability and the influence of the complex environment of households and HGs on its produce. Interventions to improve HGs are generally done within the administrative boundaries of the region and the success of such attempts should be evaluated using appropriate tools. To address this research gap a study was formulated aiming at grouping selected HGs and identifying the variables contributing to grouping; and thereby suggesting the characteristics of a model HG. A HG survey was conducted focusing Ratnapura and Hambantota districts of Sri Lanka. Pooled and separated samples of HGs in the two districts were employed in K-means cluster analysis and the groups obtained were subjected to discriminant function analysis to derive the important variables in discriminating HGs. The HGs in Ratnapura were grouped into three categories and key variables contributing to differentiate those were annual expenditure for HG, labor usage, above ground biomass, disease occurrence, species density of trees, and tree density. There were four groups of HGs in Hambantota and key variables that contributed to their differences were; annual expenditure for HG, pest occurrence, annual income from HG, species density of trees, the total share of produce in HG used for consumption, technical knowledge on farming and land extent. Accordingly, the income-driven and diversity-enhanced components of a HG would make it a model HG achieving maximum productivity. However, due to the heterogeneity of HGs, distinct groups within a district could not be identified, suggesting the inappropriateness of making administrative boundaries-based decisions on the interventions in HGs.