Assessment of spatial variability of tambons based on farming systems characteristics for scaling-up of diversification in Khon Kaen Province, Thailand

Data from 174 non-irrigated farms in 50 tambons (sub-districts) in Khon Kaen Province, Thailand, taken by the Office of Agricultural Economics (OAE) on land area by type of agricultural use and income source by agricultural activity were combined with data from the Soil Survey Laboratory of the Land Development Department (LDD) on predominant soil types in each tambon. Farms were assigned to agricultural activity types based on combinations of primary agricultural income source (=50% of total agricultural income) and secondary agricultural income sources (10-49% of total agricultural income). A diversification index was calculated as the sum of the number of primary and secondary income sources. Soils were characterized by texture, slope, drainage, and pH of each soil type. Six clusters of tambons were formed based on similarities among these variables. Three clusters contained 86% of the tambons: the rice-based farm cluster (31%), the field crop-based farm cluster (22%) and the cluster combining animal and vegetable-based production (33%). A fourth cluster (8%) combined field crop and animal production. The remaining two clusters (6% together) represented non-traditional and highly diversified farms, respectively. There was regional specialization in emphasis on rice (South) versus field crops (South Central, North), but tambons combining animal and vegetable production were distributed throughout all regions of the province. Tambons with heavier soils, less slope and poorer drainage had proportionally more land area in rice. These tambons also had proportionally more income from animal production. Conversely, tambons with lighter soils, more slope and better drainage had proportionally more land area in field crops and more income from field crops. Soil characteristics were not indicative of diversification. No spatial gradient either East-West or North-South was found for land use, income or diversification variables. Differences among clusters were more important than spatial differences. Stepwise regression indicated that six variables accounted for 76% of the variation in income. Three variables accounted for 65% of the variation in income: total land area (39% of variation), percentage of income from animal production (14%) and proportion of land area in field crops (13%). Three other variables contributed the remaining 11% of variation explained by the model: drainage (negative contribution); proportion of land area in other crops; and farms with their proportion of income from vegetable production greater than 50%. Animal production may be an indicator of potential for diversification-oriented research to increase the income contribution of horticultural production in farming systems.