Widespread changes in 21st century vegetation cover in the Southern Cone of South America
Files
A zip file with results and reference data for the publication
Date
DOI
Authors
Stanimirova, Radost
Graesser, Jordan
Friedl, Mark A.
Version
OA Version
Citation
Abstract
South America has been an epicenter of land cover and land use change (LCLUC) for over five decades due to rapid agricultural expansion along forest frontiers, the establishment of plantations in savannas, and desertification in drylands. Most attention has focused on LCLUC in tropical forests, and so information regarding the magnitude, geography, and rate of LCLUC across the Southern Cone region (SCR) of South America is incomplete. To address this, we used Landsat to map changes in the fractional cover of bare ground, woody cover, and herbaceous vegetation at annual time steps from 1999 to 2019 over the SCR. Using field observations and Landsat imagery, we created a spectral library representative of these three cover types. We trained a machine learning model to map annual fractional cover at 30-meter spatial resolution, and used a Bayesian change point algorithm to characterize spatial and temporal trajectories of LCLUC. Our results reveal substantial changes in land cover composition over the SCR between 1999 and 2019, totaling 389,973.2 km2, equivalent to 11.6% of the study domain. Herbaceous cover in Paraguay increased by 51%, mostly because of deforestation for cattle ranching in the Dry Chaco and commodity crop agriculture in the Atlantic Forest. Uruguay showed a 62% increase in woody cover arising from the emergence and growth of pine and eucalyptus plantations. Argentina, the largest and most heterogeneous of the three countries, experienced a 38% increase in bare ground in the Patagonian Steppe due to climate and anthropogenic drivers, including reduced precipitation. Quantification of these abrupt and gradual LCLUC processes can be used to improve models of the carbon budget in the SCR and to measure carbon exchange in arid and semi-arid ecoregions, as they are increasingly understood to be important drivers of the interannual variability of the global carbon cycle.