Data Lab Sandbox

umd_soy_planted_area

created_on

2023-05-04T13:11:58.897426

updated_on

2025-01-15T20:49:30.819301

spatial_resolution

resolution_description

30 x 30m

geographic_coverage

South America

update_frequency

Annual

scale

regional

citation

Source: Song et. al. 2021. “Soy Planted Area”. Accessed through Global Forest Watch on [date].

title

Soy Planted Area

subtitle

(2000-2021, S. America, Song et al. 2021)

source

Song, X.P., M.C. Hansen, P. Potopov, B. Adusei, J. Pickering, M. Adami, A. Lima, V. Zalles, S.V. Stehman, D.M. Di Bella, C.M. Cecilia, E.J. Copati, L.B. Fernandes, A. Hernandez-Serna, S.M. Jantz, A.H. Pickens, S. Turubanova, and A. Tyukavina. 2021. Massive soybean expansion in South America since 2000 and implications for conservation.

license

data_language

English

overview

This data set provides a spatially-explicit overview of the annual planted area of soybean cultivation across South America’s agricultural frontiers, as described in Song et. al. (2021). Data is available annually from 2000 to 2019 at a 30m spatial resolution for the following biomes: Amazonia, Atlantic Forest, Caatinga, Cerrado, Chaco, Chiquitania, Pampas, and Pantanal.<br><br>Song et. al. (2021) integrates long-term satellite imagery with three years of continent-wide field observations to produce annual maps of soybean cultivation. Researchers collected Landsat and MODIS imagery for November 1 through April 30 of each growing season and created 16-day time-series with phenological metrics. These metrics were combined with data on elevation, slope, and aspect and training data from 225 sampling areas to develop soybean probability maps for each year. The field observations were then used to assess the accuracy of the resulting maps, which are constrained to match the areas of field observations. <br><br>This dataset addresses the absence of long-term, consistent government statistics by expanding the spatial and temporal scope of previous maps of soybean cultivation. The maps and workflow described in Song et. al. (2021) provides key historical baseline information for tracking commodity-driven deforestation and highlights emerging hotspots of soybean expansion.

function

Provides a long-term, continental perspective of soybean expansion at a high spatial resolution

cautions

- Each yearly threshold refers to harvest year. For example, 2001 refers to the 2000-2001 agricultural season.<br>- The accuracy of the 2017, 2018, and 2019 maps were assessed as 96%, 94%, and 96%, respectively. The soy class had 8-14% rate of false positives, and a 14-17% rate of false negatives. The accuracy of the maps from previous years could not be assessed, but was found to generally match total annual harvest area statistics reported by the US Department of Agriculture.

key_restrictions

tags

why_added

learn_more

id

886f6ab7-6c8b-4228-a37f-dac817f81c6b

Is downloadable?

Yes

Versions

v1
v2
v2001
v2002
v2003
v2004
v2005
v2006
v2007
v2008
v2009
v2010
v2011
v2012
v2013
v2014
v2015
v2016
v2017
v2018
v2019
v3