The calligraphy of the society in the landscape

Spatial analysis can provide important information for social research, helping to verify the information collected in the field (Tucher and Ostrom, 2009). In this research an environmental model was designed to identify the trend of the landscape used before the establishment of protected areas (PAs). Through image analysis from two Brazilian state parks, we created a model of how the landscape would be if the parks did not exist, checking the effectiveness of the two PAs.
In order to obtain the land use trend, we used IDRISI Selva software (Eastman, 2012) and its Land Change Modeler module. The objective was to obtain a scenario in which the PAs did not exist, verifying if there is any measurable effectiveness of the PAs in the conservation objective. The Rio Preto State Park (RPSP) was created in 1994 and the Serra Negra State Park (SNSP) in 1998. Both are in the Espinhaço Mountain range, a transition area from Atlantic Forest to the Savanna.
According to the model generated for the RPSP, in a scenario where the park did not exist the native vegetation cover in the area would be around 47% in 2003 and 45% in 2009 (Table 1). This result indicates that the PA is fulfilling its function of conserving native vegetation. If the trend of land use prior to its creation continued, almost a quarter of the area would have been converted in 2009.
Connecting this data to the history of the region, we noted that the conservationist trend that emerged in the area where the park was installed in the late 1980s, did not have the effect captured in the changes of the landscape between 1986 and 1992. The images generated by the model indicated that, without the SNSP, native landscape coverage would be about 2% higher in 2003 and 5% higher in 2009 (Table 2). The result indicates that the SNSP is not fulfilling its role in the maintenance of the native vegetation, since the projection that does not consider the existence of the park showed that there would be a greater cover. However, another results can be noted. The amplitude of this difference shows that the land use trend of the region analysed by the model between 1986 and 1992 has a considerably lower impact on land cover, when compared to the trend recorded in the RPSP. For the SNSP, the trend would be for an increase in native vegetation, which would go from 58% in 1992 to 65% in 2003. The model resulted in a recovery landscape in the years to come.
Furthermore, the families interviewed in the field developed the idea of a close connection with their environment, and a preference in low impact agricultural practice, as reported by Galizoni (2007). Quantifying and considering trends, and verifying the existence of differences in this context, can result in model enhancement and valuable information for understanding the interactions of the local population with their environment.

The research that generated these results was funded by the European Union through the FP7 - 2010 Program, under the agreement No SSH-CT-2010-266710.