Appendix A. Creating historical land use reconstructions for Mato Grosso, Brazil (1901–2006).
Historical land-use reconstructions are difficult in remote locations, such as the Brazilian Amazon, due to sparse data. We used as much data as possible from historical records to constrain our annual and spatial time series of land use change. The 2001 remote sensing data sets for pasture and cropland are the starting point for creating historical land-use data sets as they provide well-calibrated spatial estimates. For the temporal trends in historical l areas of pasture and crops, we used the areas reported in state-level census estimates (Censo Agropecuário; SIDRA 2009a) and annual crop statistics (Produção Agrícola Municipal, PAM; SIDRA 2009b) to extend land-use change backwards through time. While the spatial reconstruction does not attempt to be historically and geographically precise, it accurately depicts the statewide trends in land-use transitions in pasture and croplands that we can use to estimate impacts on carbon and nitrogen cycling.
The fusion of multiple data sources for historical land-use reconstruction requires an intimate understanding of the data sets, their limitations and how to integrate them as seamlessly as possible. In this study, we had to address errors inherent in the census and annual statistics, changes in political boundaries, interpolating between different data sources, and creating a spatial domain for temporal data sets.
Between 2001 and 2006, remote sensing imagery is used to determine both the distribution and extent of pastures and croplands. For pastures, remote sensing estimates in 2001 and 2004 are used directly (Morton et al. 2006, 2009) and estimates for the other years are determined through linear interpolation. For single cropped and double cropped areas, we use newly developed remote sensing data sets for cropland estimates and cropping patterns that are well-validated (Galford et al. 2008; Galford et al. 2010). These estimates are derived from phenological analysis of a wavelet-smoothed time series of MODIS Enhanced Vegetation Index (EVI; Huete et al. 2002) at a moderately coarse resolution of 500 m, appropriate for the size of croplands in Mato Grosso (the majority over 200 ha, many > 1 000 ha; Alves 2002; Galford et al. 2008). This data set shows that croplands more than doubled from 2001–2006 to cover ~100 000 km2, with double-cropping patterns accounting for roughly 20% of croplands. We spatially combined the two estimates of pasture for 2001 and 2004; the Morton et al. (2006) data is a better estimate of pasture in forested areas, whereas the Morton et al. (2009) data is a better estimate for pasture in areas of cerrado. We randomly fill in pasture each year at a rate suggested by the census and annual crop statistics within the spatial constraints from the remote sensing data for 2001 and 2004.
Both the agricultural census and annual crop production estimates are well-correlated to the remote sensing estimates of land use for overlapping time periods (Galford et al. 2010). The Censo Agropecuário (agricultural census, 1935, 1940, 1950, 1960, 1970, 1980, 1985, 1996, and 2006) includes estimates of pasture and crop production while the PAM estimates annual crop production (1990–2006). Before using the government data sets in our historical reconstruction, we first corrected the statistics so that the total area farmed did not contain any double-accounting for areas that were double cropped with soy and maize (maize is always grown as a second-crop in a double cropping pattern in this region). The statistical estimates are well-correlated at the state level to remotely sensed estimates that are well-validated, but have high levels of uncertainty at finer resolutions (meso-, mirco-, and county-scale; Galford et al. 2010). We used the state level data because of uncertainties in the finer-scale records related to how data are collected and reported. As explained by Galford et al. (2010), annual production records are aggregated by IGBE from monthly data under the direction of the Coordenador Estadual de Pesquisas Agropecuárias (state coordinator of agricultural research) with the aid of the IBGE data collection network, other local government offices, and the producers at county, regional and state levels (IBGE 2002). For example, data is collected for each county, but the total cropland area in one county may be larger than its total area if a farm straddles two counties and the total cropland area is attributed to one county, perhaps the county where it is headquartered.
In 1978, Mato Grosso split almost in half to form two states; Mato Grosso and Mato Grosso do Sul. Prior to 1978, the Mato Grosso census records include areas that today are part of Mato Grosso do Sul, which are not part of our study area. For the late 1970s, the overlap between PAM (Mato Grosso and Mato Grosso do Sul reported together) and the Censo Agropecuário (Mato Grosso do Sul separately) data sets for crops were used to normalize the pre-1978 Mato Grosso records to reflect only the area of Mato Grosso as it exists today (Fig. A1). From 1940–1975, we relied on decadal census estimates and used linear interpolation to make an annual time series, after normalizing for the land area in Mato Grosso do Sul. The pasture area was reconstructed through annual linear interpolation of decadal census data from 1935 to 1996, after normalization to remove records accounting for Mato Grosso do Sul (Fig. A1). From 1996–2001, the pasture area was reconstructed through linear interpolation between the 1996 census data and the 2001 remote sensing pasture information. The state-level records, corrected for double-cropping and changes in political borders, were then normalized to the remote sensing information on croplands and pasture, as the remote sensing data was a slight underestimate but provided spatial detail not found in the government statistics (Fig. A1).
For TEM, we needed to change the time series of land use change (Fig. A1) into a spatially explicit data set. We used the remote sensing classes from 2001 to then build backwards in time, following the changes in pasture or cropland as prescribed in Fig. A1). We assumed there was no regrowth, as secondary growth is a very minor component of Mato Grosso’s land-cover and land-use change history (Morton et al. 2006). This means that all pasture and cropland observed in 2001 was converted sometime prior. We then created an annual land use data set by reducing the area in pasture and cropland as prescribed by the reconstruction time series (Fig. A1). For each year, we used an iterative process to randomly select pixels within each land use class for reclassification as either their natural land cover or, for croplands, pasture. Future work could refine the spatial reconstruction through more complex modeling, such as including a weighting for infrastructure development, but close examination of Fig. A1 and Fig. 5 illustrates how this reconstruction reflects historical events and trends in Mato Grosso’s land-use history, including: (1) increases in agricultural development in the mid-1960s to early 1970s due to government prioritization of Amazon development, including incentives for investment (subsidies, tax holidays. and low interest rate loans) through the creation of SUDAM, the Amazon Development Agency (established in 1969), building of the Transamazon Highway (BR-320) and the Cuiabá-Santarém (BR-163), and promotion of colonization through government-directed settlement programs increased agricultural production that received social services such as health care and schools, (2) decreases in migration and agricultural production from 1985–1991 due to hyperinflation and tightening of credit, and (3) increases in agricultural development starting in the early 1990s when EMBRAPA found an agronomic solution to the persistent challenge posed by the high aluminum toxicity of the dominant soil type in the Cerrado and the tropical forest biomes (i.e., the Oxisols), along with the development of new soybean varieties made the Cerrado region’s soils productive for soybeans. The creation of the “real” or new currency in 1994 brought inflation under control and this led to a virtual explosion of deforestation in 1995 with farmers taking loans to increase area in production. Soybeans, further supported by favorable markets and world commodity prices, took off, and massive clearing of the Cerrado increased the area in soybeans several fold during the 1990’s and the early years of the 21st century.
|FIG A1. Reconstruction of croplands (A) and pasture (B) in Mato Grosso by area, plotted with estimates from government surveys, census and remote sensing used to constrain the reconstruction.|
Alves, D. S. 2002. Space-time dynamics of deforestation in Brazilian Amazonia. International Journal of Remote Sensing 23:2903–2908.
Galford, G. L., J. M. Melillo, J. F. Mustard, C. E. P. Cerri, and C. C. Cerri. 2010. The Amazon frontier of land-use change: croplands and consequences for greenhouse gas emissions. Earth Interactions, In Press.
Galford, G. L., J. F. Mustard, J. Melillo, A. Gendrin, C. C. Cerri, and C. E. P. Cerri. 2008. Wavelet analysis of MODIS time series to detect expansion and intensification of row-crop agriculture in Brazil. Remote Sensing of Environment 112:576–587.
Huete, A., K. Didan, T. Miura, E. P. Rodriguez, X. Gao, and L. G. Ferreira. 2002. Overview of the radiometric and biophysical performance of the MODIS vegetation indices. Remote Sensing of Environment 83:195–213.
IBGE. 2002. Pesquisas Agropecuárias. Série Relatórios Metodológicos 6(S). Ministério do Planejamento, Orçamento e Gestão, Rio de Janeiro, Brazil.
Morton, D. C., R. S. DeFries, Y. E. Shimbukuro, L. O. Anderson, E. Arai, F. d. B. Espirito-Santo, R. Freitas, and J. Morisette. 2006. Cropland expansion changes deforestation dynamics in the southern Brazilian Amazon. Proceedings of the National Academy of Science103.
Morton, D., R. DeFries, and Y. E. Shimabukuro. 2009. Cropland expansion in cerrado and transition forest ecosystems: Quantifying habitat loss from satellite-based vegetation phenology. IN: C. A. Klink, editor. Cerrado Land Use and Conservation, Advanced Applied Biodiversity Science. Conservation International, Washington, D.C.
SIDRA. 2009a. Censo Agropecuário. Sistema IGBE de Recuperação Automática.
SIDRA. 2009b. Produção Agrícola Municipal. Sistema IGBE de Recuperação Automática.