Remote Sensing Time Series to Evaluate Direct Land Use Change of Recent Expanded Sugarcane Crop in Brazil

Mitigation of global carbon emissions to prevent global warming potential using biofuels is highly dependent on direct and indirect land use change (LUC). There are still several uncertainties about how to assess the indirect LUC impacts of biofuels. However, direct LUC (dLUC) can be evaluated using remote sensing (RS). The present work has the aim to quantify the dLUC occurred during the recent sugarcane expansion for ethanol and sugar production concentrated in the South-Central region of Brazil. This region has a favorable climate for sugarcane production and also a great potential for agriculture expansion. Yearly monitoring from 2005 to 2010 using Landsat type imagery has shown that the sugarcane crop expanded during this period over 3.5 Mha in the South-Central region. To evaluate the dLUC in response to the expanded sugarcane area we used RS time series from the MODIS sensor transformed to the two-band enhanced vegetation index (EVI2), acquired from 2000 to 2009. The original sugarcane map was re-sampled to a pixel size of 250 x 250 m to be compatible with spatial resolution of the MODIS images. One percent of these pixels were systematically sampled covering 1035 pixels. Each of these pixels were carefully analyzed using a special developed web tool to visualize the entire MODIS time series overlaid with several Landsat-5 TM images acquired at key periods in order to correctly identify the land use/land cover prior to the sugarcane crop. Considering 2000 as reference year for the dLUC evaluation it was observed that: 69.8% of the sugarcane expanded on pasture land; 26.2% expanded on annual crops; 0.6% expanded on native vegetation; and 3.4 % was not sugarcane expansion but sugarcane renovation using crop rotation. It was interesting to notice that 35% of the pasture land in 2000 converted to sugarcane was first converted to annual crops. This practice is commonly adopted for one to two years on degraded pasture to improve the physical soil characteristics before introducing the sugarcane crop. It was also observed that the 0.6 % of native vegetation changed to sugarcane was previously converted to either annual crop (33%) or pasture land (67%). Although the analysis needs to be further refined the results clearly show that the dLUC observed during the recent sugarcane expansion for ethanol and sugar production in the South-Central region of Brazil is mainly occurring on pasture and agricultural land.