Agricultural productivity enhancement can be due to land use and land handling changes. In this paper we propose the use of Data Envelopment Analysis models (DEA) to evaluate the spatial distribution of family farmers land use efficiency. We studied the "land productivity" evolution for a sample of family farmers from Machadinho d'Oeste (RO), during four periods of time. DEA models variables were rice, maize and coffee productions as outputs, and these crops total cultivated area as input. As a result we noticed that cultivating simultaneously rice and maize was the best performance case. Efficiency measurements had spatial dependence in the four periods of time. In 1999 and 2002 farms productive efficiency measurements were more uniformly distributed; the most efficient farmers were concentrated in the central part of the studied area.
land use; efficiency; spatial dependence
