MODELING EFFECTS OF LAND MANAGEMENT IN THE BRAZILIAN AMAZONIAN SETTLEMENT OF RONDONIA

The socioeconomic and ecological aspects of land-use change are interr elated, especially in the Brazilian Amazon, where immigrants are rapid ly cutting the forest to establish farms. A computer simulation has be en developed that projects land-use changes, carbon release, and the t ime a family can remain on a farm lot as a function of initial soil an d vegetation conditions, market and road infrastructure, and decision variables. The model simulates extremes of land-use practices and typi cal land-use conditions for central Rondonia, Brazil. Typical land-use practices in Rondonia are based on published accounts of farmers' act ivities. The most severe practices are defined based on extreme land-m anagement conditions along the Transamazon Highway. The best land-use practices are innovative farming techniques that use a diversity of pe rennial crops. Model projections using the typical land-use scenario p roduce changes in land cleared, carbon release, and farmer turnover ra tes representative of central Rondonia, based on interviews with 87 fa rmers. Results from both the model projections and the interview data show that farmers who have been on the lots a decade have cleared abou t half of the lot. The most striking result is the similarity in the t ypical and worst-case scenarios over the 40-year projection period in the percentage of land cleared and carbon released and in the number o f families on the farm lots and abandoning the lots over time. Model r esults from the best-case scenario compare well to the three farmers o ut of 87 interviewed who practice innovative farming. In both cases, f armers who have been on their land an average of eight years have defo rested less than 20% of the lot. Spatial indices of fractal dimension and contagion are used to quantify the effects of the different land-m anagement practices. Changes in the fractal index over the 40 years of projected land management demonstrate the decrease in complexity as a forested region becomes dominated by agriculture for the typical and worst-case scenarios. Low values of the contagion index indicates that a heterogeneous mix of land uses is retained with the best-case scena rio. These results suggest that crop diversity and nontraditional tech niques provide both a social and environmental improvement over the ot her scenarios: the people are able to maintain themselves on the land, less carbon is released, and the land maintains a mix of habitat type s. The model results illustrate that both social and environmental eff ects of land-management practices need to be considered.