Deforestation Linked to Agriculture

Of the seven commodities analyzed,See <a href="/gfr/data-and-methods#deforestation-forest-loss">Data and Methods</a> and Goldman et al. (2020), <a href="https://www.wri.org/publication/estimating-role-seven-commodities-agriculture-linked-deforestation">https://www.wri.org/publication/estimating-role-seven-commodities-agriculture-linked-deforestation</a> for more information about the methodology used. cattle replaced the most forest by far—cattle pasture now occupies some 45.1 million hectares (Mha) of land deforested between 2001 and 2015, accounting for 36 percent of all tree cover loss associated with agricultureAs defined by the “shifting agriculture” and “commodity-driven deforestation” classes of Curtis et al. (2018), https://doi.org/10.1126/science.aau3445 . during the time period. Oil palm ranks second (10.5 Mha), followed by soy (8.2 Mha), then cocoa, plantation rubber, coffee, and plantation wood fiber (all around 2 Mha each; though the results from rubber and wood fiber only cover a subset of countries). Together, these seven commodities account for 57 percent of all tree cover loss associated with agriculture between 2001 and 2015.This statistic has been updated from Goldman et al. 2020, https://www.wri.org/publication/estimating-role-seven-commodities-agriculture-linked-deforestation to include the latest drivers data from Curtis et al. 2018, covering the period 2001-2019.

Note: The maps show the area of forest replacement by each crop as a proportion of each second administrative level’s total land area. Maps of forests replaced by wood fiber and rubber plantations are not included here due to the limited data available. 

Globally, the conversion of forests to cattle pasture resulted in an estimated 45.1 Mha of deforestation between 2001 and 2015, five times more than for any of the other analyzed commodities.See <a href="/gfr/data-and-methods#deforestation-forest-loss">Data and Methods</a> and Goldman et al. (2020) for more information. The analysis assumes that clearing in pastures where cattle density is at least one head of cattle per square kilometer constitutes forest replacement by cattle. Cattle can be used for beef, dairy, and leather goods production, but pasture is also commonly used as a way to hold land for future use. Much of this deforestation occurred in Brazil (21.8 Mha, or 48 percent), followed by Paraguay (9 percent) and Colombia (5 percent). The states of Pará and Mato Grosso in Brazil experienced the greatest area of forest replaced by cattle, with around 5 Mha each between 2001 and 2015.

In Brazil, based on detailed pasture maps available from the Image Processing and Geoprocessing Laboratory (Laboratório de Processamento de Imagens e Geoprocessamento; LAPIG) for the year 2018,  just over half of the country’s total tree cover loss area between 2001 and 2015 (42.2 Mha) was deforestation for pasture. Seventy percent of this deforestation occurred in the Brazilian Amazon, and another 20 percent occurred in the Cerrado. In the Amazon, deforestation linked to pasture peaked in 2005, then dropped to around half that level thereafter. Deforestation in the Cerrado peaked in 2004 but was followed by a much less dramatic decline than in the Amazon. The timing of the declines in both biomes corresponds to interventions and changes in public policy to reduce deforestation within the Brazilian Legal Amazon, which also includes part of the Cerrado. 

Between 2001 and 2015, oil palm plantations expanded by 22.4 Mha globally, increasing the total footprint of oil palm by 167 percent.Based on data from Harris et al. (2019), https://www.wri.org/publication/planted-trees . More detail is available in Goldman et al. (2020), https://www.wri.org/publication/estimating-role-seven-commodities-agriculture-linked-deforestation . Nearly half (10.5 Mha) of this expansion occurred on land deforested since 2001, with 3.5 Mha expanding onto land that was primary forest in 2001.See <a href="/gfr/data-and-methods#deforestation-forest-loss">Data and Methods</a>. For more information, see Goldman et al. (2020), <a href="https://www.wri.org/publication/estimating-role-seven-commodities-agriculture-linked-deforestation">https://www.wri.org/publication/estimating-role-seven-commodities-agriculture-linked-deforestation</a> . More than two-thirds of all deforested land now used for oil palm plantations is in Indonesia (7.0 Mha), with another 2.7 Mha in Malaysia. Forests in the provinces of Central Kalimantan, East Kalimantan, and Riau in Indonesia and Sarawak in Malaysia were particularly hard-hit, with oil palm plantations occupying more than a million hectares of land deforested between 2001 and 2015 in each of these jurisdictions. In Indonesia and Malaysia, over half (53 percent and 57 percent, respectively) of new oil palm plantations replaced forests. The countries of Liberia, Solomon Islands, and Gabon all experienced much less oil palm expansion than Indonesia and Malaysia, but more than half of that expansion resulted in deforestation (74 percent, 54 percent, and 54 percent, respectively).

The amount of deforestation that occurred previously on land now used for oil palm plantations increased markedly between 2001 and 2012—with peaks in 2009 and 2012–then declined to a record low in 2015 as part of a more than sixfold decline in overall deforestation rates in Indonesia and Malaysia. Studies have cited lower prices for crude palm oil (with reductions in deforestation rates generally lagging price drops by one year), Indonesia’s moratorium on new licenses to convert primary forests and peatlands (implemented in 2011), and corporate zero-deforestation commitments as possible explanations for the decline.Austin et al. 2019, https://doi.org/10.1088/1748-9326/aaf6db; Gaveau et al. 2019, https://doi.org/10.1111/conl.12622 .

Globally, soy farms occupy 8.2 Mha of land deforested between 2001 and 2015.See <a href="/gfr/data-and-methods#deforestation-forest-loss">Data and Methods</a>; for more information, see also Goldman et al. (2020), <a href="https://www.wri.org/publication/estimating-role-seven-commodities-agriculture-linked-deforestation">https://www.wri.org/publication/estimating-role-seven-commodities-agriculture-linked-deforestation</a> . Almost all (97 percent) of this deforestation occurred in South America, with 61 percent in Brazil, 21 percent in Argentina, 9 percent in Bolivia, and 5 percent in Paraguay. The Brazilian state of Mato Grosso had by far the most forest replacement for soy, accounting for over one-third of the total. Within Brazil, 48 percent of forest area replaced by soy occurred in the Amazon, with an additional 45 percent in the Cerrado.

In South America, deforestation on land currently used for soy peaked in 2004 and then dropped to around half of that rate thereafter, in line with overall deforestation declines in the Brazilian Amazon, the Brazilian Cerrado, and Paraguay. There was also a second decline after 2012, most notably in Argentina.

Between 2001 and 2015, cocoa farms replaced an estimated 2.3 Mha of forest globally.See <a href="/gfr/data-and-methods#deforestation-forest-loss">Data and Methods</a>; for more information, also see Goldman et al. (2020), <a href="https://www.wri.org/publication/estimating-role-seven-commodities-agriculture-linked-deforestation">https://www.wri.org/publication/estimating-role-seven-commodities-agriculture-linked-deforestation</a> . Indonesia and Côte d’Ivoire had the most forest area replaced by cocoa (25 percent and 22 percent of the global total, respectively), followed Brazil (19 percent), Ghana (10 percent), and Cameroon (6 percent). Although West African countries such as Ghana did not have the largest area of forest replaced by cocoa, the area represented a much larger share of the country’s total tree cover loss (one-third for Ghana; just over a quarter for Côte d’Ivoire). Deforestation on land later used for cocoa remained relatively steady from 2001 to 2013 but experienced a sharp increase after 2014. Meanwhile, cocoa production has steadily increased since the turn of the century, increasing by 44 percent between 2001 and 2015.See FAOSTAT, http://www.fao.org/faostat/en/#data .

In the seven countries where rubber plantation data are available (Brazil, Cambodia, Cameroon, Democratic Republic of the Congo, India, Indonesia, and Malaysia), rubber plantations replaced 2.1 Mha of forest between 2001 and 2015.See <a href="/gfr/data-and-methods#deforestation-forest-loss">Data and Methods</a>; for more information, see also Goldman et al. (2020), <a href="https://www.wri.org/publication/estimating-role-seven-commodities-agriculture-linked-deforestation">https://www.wri.org/publication/estimating-role-seven-commodities-agriculture-linked-deforestation</a> . However, the analyzed countries account for only 40 percent of total rubber production, with important producing countries such as China, Thailand, and Vietnam missing due to a lack of data.See FAOSTAT, http://www.fao.org/faostat/en/#home . Of the countries analyzed, Indonesia experienced the most forest replacement by rubber plantations (1.0 Mha), followed by Malaysia (0.7 Mha). Whereas rubber was traditionally cultivated as part of intermixed agroforestry operations, monoculture plantations now dominate rubber production.See Feintrenie and Levang 2009, https://doi.org/10.1007/s11842-009-9086-2 . Deforestation on land later used for rubber plantations markedly increased from 2001 to 2012, then declined through 2015, following a similar pattern to the price for rubber, which dropped off after 2011.See World Bank Commodity Markets, https://www.worldbank.org/en/research/commodity-markets .

Nearly 2 Mha of forest were replaced by coffee plantations between 2001 and 2015, of which 1.1 Mha were for robusta coffee and 0.8 Mha were for arabica coffee.See <a href="/gfr/data-and-methods#deforestation-forest-loss">Data and Methods</a>; for more information, see also Goldman et al. (2020), <a href="https://www.wri.org/publication/estimating-role-seven-commodities-agriculture-linked-deforestation">https://www.wri.org/publication/estimating-role-seven-commodities-agriculture-linked-deforestation</a> . Forest replacement by robusta coffee was most prevalent in Indonesia (33 percent of the global total), followed by Brazil (16 percent), Madagascar (14 percent), and Vietnam (12 percent). Forest replacement by arabica coffee was most prevalent in Brazil (36 percent), Peru (20 percent), and Colombia (8 percent).

Deforestation on land later used for robusta coffee increased markedly over the analysis period, but deforestation on land later used for arabica coffee remained steady. 

Robusta, generally grown in full sun at lower latitudes, is often used in instant coffee and has seen an increase in demand from consumers in Asia and Brazil.Kurose 2019, https://asia.nikkei.com/Business/Markets/Commodities/Asians-love-for-instant-coffee-fuels-Robusta-cultivation2 . Robusta has grown from 36 percent of global coffee production in 2000 to 44 percent in 2015.See the U.S. Department of Agriculture’s Foreign Agricultural Service, https://apps.fas.usda.gov/psdonline/app/index.html#/app/advQuery . Arabica, on the other hand, typically grows in agroforestry systems at higher latitudes. Previous research has suggested that, in some cases, arabica coffee has slowed down the rate of deforestation because trees in coffee cultivation areas are spared from clearing because arabica coffee can grow in shade.Blackman et al. 2005, https://www.semanticscholar.org/paper/Deforestation-and-Shade-Coffee-in-Oaxaca%2C-Mexico%3A-Blackman-Albers/b5ca9cac55b3f1113b24fc1b5dd13a0d8301d5b6?p2df; Getahun et al. 2013, https://doi.org/10.1016/j.foreco.2013.05.001 .
 

For the 10 tropical countries where wood fiber plantation data are available (Argentina, Brazil, Cambodia, China, India, Indonesia, Malaysia, Rwanda, South Africa, and Vietnam), wood fiber plantations replaced an estimated 1.8 Mha of forest between 2000 and 2015.See <a href="/gfr/data-and-methods#deforestation-forest-loss">Data and Methods</a>; for more information, see also Goldman et al. (2020), <a href="https://www.wri.org/publication/estimating-role-seven-commodities-agriculture-linked-deforestation">https://www.wri.org/publication/estimating-role-seven-commodities-agriculture-linked-deforestation</a> . By far the most forest area replaced by wood fiber plantations occurred in Indonesia, accounting for 90 percent of the measured deforestation, followed by Malaysia (5 percent) and China (4 percent). 

The forest extent analysis is limited by various data and attribution issues and methodological assumptions, including the following:

• Commodity data sets have limited coverage and quality. Only oil palm has recent, detailed maps of extent at a global level, and even these maps likely exclude many small-scale plantations. The analysis also uses detailed data on South American soy, Brazilian pasture, and wood fiber and rubber plantations for select countries. Outside of these regions and commodities, the analysis relies on global 10-kilometer resolution data on crop and pasture extent.For more information, see Goldman et al. (2020), https://www.wri.org/publication/estimating-role-seven-commodities-agriculture-linked-deforestation . Due to the coarse resolution of these data, and the assumptions the coarseness necessitates, the results are likely to be less accurate. Furthermore, these data are from 2010 (2000 for pasture), so the amount of forest replaced by a specific commodity is assumed to be proportional to its area during that year and may be misrepresented if significant expansion or contraction of that commodity has occurred since then. 
• The data cannot capture complex land-use change transitions. The analyses presented here do not take into account other possible land uses between the deforestation event and the establishment of the commodity, except in the cases of oil palm and soy, where we include estimates of direct deforestation. The analysis also does not consider any forms of indirect land-use change (e.g., the target commodity displacing other activities that may, in turn, expand into forested areas). Including indirect land-use change could change the amount of deforestation attributed to each commodity, especially for oil palm, soy, and cattle.
• The data measure tree cover loss rather than deforestation directly. All tree cover loss in an area later used for one of the target commodities is assumed to be deforestation because replacement of a forest with a crop or pasture represents a permanent land-use change. Historical data from Indonesia and Malaysia are used to filter out oil palm plantations established prior to 2000 from the analysis to avoid counting the felling of old, unproductive oil palm trees as tree cover loss.For more information about the data sets used in this analysis, see Goldman et al. (2020), https://www.wri.org/publication/estimating-role-seven-commodities-agriculture-linked-deforestation . However, it is possible that some plantation dynamics in oil palm plantations outside Indonesia and Malaysia, or for other types of plantations, are included in the figures and result in an overestimation of deforestation. 
• The data may miss some forms of tree cover loss. The tree cover loss data may not detect all changes related to commodity production. Much of the production of cocoa and coffee occurs on very small farms (less than one hectare) that may not be captured by the Hansen et al. (2013)Hansen et al. 2013, https://doi.org/10.1126/science.1244693 . tree cover loss data, resulting in an underestimation of deforestation for these commodities. The analysis may also underestimate the conversion of dry forest and woody savanna areas, which are not well represented in the Hansen et al. (2013) tree cover loss data. For the detailed soy and pasture analyses, we define tree cover as any woody vegetation with a minimum of 10 percent canopy cover (analyses for other commodities use 30 percent) to minimize underestimations in South American biomes such as the Cerrado and the Chaco, where these commodities play a big role.For more information about the methodology, see Goldman et al. (2020), https://www.wri.org/publication/estimating-role-seven-commodities-agriculture-linked-deforestation .

Further discussion about the methods, assumptions, and limitations of this analysis is available in Goldman et al. (2020).