Introduction

The world demand for staple crops, forages, and biofuels has augmented at an unprecedented rate and the agricultural surface has consequently increased (^{Tilman et al., 2001}; ^{Gibbs et al., 2010}; ^{Guerrero-Peña et al., 2013}). On a global scale, the highest agricultural expansion has taken place in tropical regions and it is identified as one of the main causes of deforestation, fragmentation, and degradation of ecosystems (^{Geist and Lambin, 2002}; ^{Foley et al., 2005}; ^{Green et al., 2005}). From 1980 to 1990, over 80 % of tropical forests were replaced by agricultural lands, and it is calculated that by 2050 the world demand of agricultural products could increase by 50 % and tropical countries will meet the largest demand (^{Gibbs et al., 2010}). This agricultural intensification and the expansion of monocrops have severe environmental consequences for tropical ecosystems, such as soil erosion and nutrient depletion, competition for water resources at a local scale, loss of biodiversity, and landscape deterioration (^{Tilma et al., 2001}; ^{Koh and Ghazoul, 2008}; ^{Gibbs et al., 2010}).

Palm oil and palm kernel oil are obtained from the African oil palm (Elaeis guineesis Jacq. 1897) and are required by several industries (^{Henderson and Osborne, 2000}; ^{Corley and Tinker, 2003}, ^{Tan et al., 2009}). In the early 20^th century, the demand for these oils began to exceed their production and the first plantations were established in southeastern Asia (^{Henderson and Osborne, 2000}). However, from 1960 to 2000, the world production of palm by-products underwent an increase from 1.7 to 23.8 million t (^{Corley and Tinker, 2003}). Consequently, the area allocated for these monocrop plantations has mainly increased in Malaysia and Indonesia and it has quickly expanded in certain regions of Africa, Central America, and the Amazonia. Currently, palm plantations occupy approximately 17 million ha and their potential cultivation surface could exceed 200 million ha (^{Pirker et al., 2016}). The establishment of these plantations has severe impacts on tropical systems, such as deforestation, pollution, droughts, conflicts over the land, and loss of biodiversity (^{Fitzherbert et al., 2008}; ^{Koh and Wilcove, 2008}; ^{Foster et al., 2011}; ^{Savilaakso et al., 2014}; ^{Lees et al., 2015}; ^{Mandal and Shankar, 2016}). Additionally, biofuel-oriented crops (such as oil palm), could replace food crops and created food insecurity in the countries where it is cultivated (^{Carrere, 2006}).

Mexico is a major consumer of oil palm by-products. Although there are currently palm plantations in Mexico, domestic production only meets 61.5 % of the demand, while the remainder is imported from Guatemala, Costa Rica, and Colombia (^{SAGARPA, 2017}). It is estimated there are over 8 million ha with high potential for the cultivation of palms that could meet the domestic and international demand (^{SAGARPA, 2017}), and state governments and federal programs promote its production (^{Sánchez, 2003}; ^{SAGARPA, 2017}). This oil palm expansion scenario requires evaluating the impact that this crop has in land use change and in replacing staple crop.

The objective of this study was to evaluate the increment in the oil palm cultivation area during the last three decades in Mexico, and to preliminary estimate its impact as an agent of regional land use change. In particular, identifying the trends in: change of the regional and local surface where oil palm is cultivated; replacement of natural vegetation and other crops in the municipalities where oil palm cultivation areas cover a wider area. This information will help to understand the scope of oil palm cultivation in Mexico and to guide public policies related to the environmental integrity of the tropical landscapes where this crop exist.

Materials and Methods

Crops analyses were conducted in two stages: 1) information about the national, regional, and local surface sown with oil palm was compiled; and 2) a preliminary analysis of land use change in the municipalities of the states identified during the first stage.

The information about the surface sown with oil palm was obtained from the databases of the Servicio de Información Agroalimentaria y Pesquera (^{SIAP, 2017}): It covered the 1980-2016 period, cultivation cycles per agricultural year and constant cycles, in irrigated and non-irrigated lands. The municipalities from the states with larger palm cultivation areas in 2015 were selected based on the surface sown and the availability of high-resolution satellite images taken from Google Earth where oil palm plantations could be identified. Additionally, surface variation for the main crops was analyzed and it was compared with the surface variation for oil palms.

Land use change in the selected municipalities was subject to a retrospective analysis, which was carried out based on the on-screen identification and digitalization of the area in which oil palm plantations are found. For this purpose, the most recent images (2010-2016) for these municipalities available in Google Earth were used. Overall, 1266 sites were digitalized, out of which 66 % accounted for images from 2016, followed by 12 % and 11 % of images from 2012 and 2014, respectively. The remaining 11 % belonged to images from the other years. The sites were digitalized at a minimum 300-meter eye line. The average area of the sites was 9.8 ha (±0.6 SE), with a variation of 1 to 478 ha. Since palms look different with age (^{Gutiérrez-Véñez and DeFries, 2013}), only plantations that had mature palms were included, because their foliage has similar special characteristics such as texture, color, and shape (i.e., star) which set them apart from other plantations, such as mango. Unlike coconut palm plantations, oil palm plantations have a high density and a particular crop arrangement, usually in rectangular blocks, and they are usually found in the vicinity of roads (Figure 1).

Figure 1 Example of the digitalization of oil palm plantations based on Google Earth images. 

The resulting sites were transformed into vector files and they were projected over the five Land Use and Vegetation series developed by INEGI (1:250 000), to identify the land uses that these same areas were given in previous years. The Land Use and Vegetation Series include information from the last thirty years (Series 1, from 1968 to 1986; Series II, 1990; Series III, 2005; Series IV, 2009; and Series V, 2013; ^{INEGI, 2016}). Once the covers that existed within the zones in the various years were identified, cover transition was determined. The data obtained were grouped into five main classes (i.e., agricultural, pasture, rainforest, aquatic vegetation, and other kind of vegetation) and transition matrices were developed with the aim of identifying the percentage of the sites that changed from one land use to another, from one period to another. For each matrix, only the main cover classes are shown. Therefore, adding up all the covers according to their percentage does not always add up to 100 %. Although, ^{Velázquez et al. (2002)} questioned the validity of comparing vegetation series, based on methodological differences, they are used throughout Mexico to show trends in the kind of vegetation that is replaced (primary vegetation versus crops) and they are useful reference values. Nevertheless, the INEGI series have a minimum mapping unit of 50 ha (vegetation) and 25 ha (agriculture and pastureland), while Google Earth identifies even 1 ha crops. Therefore, as part of a preliminary study, this analysis included the area for oil palm accumulated in each municipality.

Results and Discussions

The area dedicated to cultivate oil palm in Mexico has increased almost seventyfold from 1318 ha (1985) to 90 118 ha (2016) (Figure 2). From 1985 to 1995, the area sown increased by 2758 ha, equivalent to 180 ha per year. From 1995, the increase rate reached 4000 ha per year. In fact, from 1985 to 2016 -based on the area cultivated in the whole country- palm move up from rank 102 to 24. As its area increased, so did domestic production: from 1600 (1985) to over 700 000 (2016) tonnes. From 2013 to 2016, the cultivated area showed a cumulative increase of 209 % (^{SAGARPA, 2017}). The rapid increase of this crop is the result of the economic income it generates, the governmental support for the cultivation of biofuel crops, and the interest of small producers in the introduction of this crop, mainly as a consequence of the failure of other productive activities (^{Castellanos-Navarrete and Jansen, 2017}).

Figure 2 Increase in the cultivation area of oil palm (Elaeis guineensis) during the 1980-2016 period in Mexico. Domestic data: dotted line; state data: continuous line. (^{SIAP, 2017}). 

Palm plantations were established in Campeche, Chiapas, Tabasco, and Veracruz. The first palm plantations were established in 1948, in the Soconusco region of the coast of Chiapas (^{INIFAP, 1999}) and with the first large-scale projects more than 2500 ha were established in that state in 1978, in the municipalities of Acacoyagua, Acapetahua, Mapastepec, and Villa Comaltitlán (^{SAGARPA, 2004a}; ^{ANIAME, 2006}). The first plantations in Veracruz and Campeche were established in 1997 (^{Sánchez, 2003}), while the first plantation in Tabasco was established the following year (^{SAGARPA, 2004b}). Local governments in those states have encourage the production of this crop, as a way to create permanent jobs in the industry and farm lands, maintain social stability in the zone, and prevent migration by settling producers in their smallholdings (^{SAGARPA, 2004c}).

Palm crops are currently found in 52 municipalities throughout those states. The greatest sown surface is found in Chiapas (43 443 ha), followed by Campeche (23 328 ha), Tabasco (16 195 ha), and Veracruz (7151 ha). This area will likely undergo a significant increase in the following years. The following surfaces for the potential establishment of this crop are identified: 900 000 in Chiapas (^{SAGARPA, 2004a}), 1 120 000 in Tabasco (^{SAGARPA, 2017}), and 700 000 in Campeche (^{SAGARPA, 2017}). In 1999, INIFAP determined that this crop could potentially be established in the regions of Texistepec, Jesús Carranza, Las Choapas, and Uxpanapan, in southern Veracruz. A new oil extracting plant has been planned in Campeche, where oil will be transformed into biodiesel (^{Unidad de Comunicación Social, 2014}), which could expand palm cultivation.

The percentage of the area used to cultivate palm in each municipality was: 19.0 % (10 550 ha) in Acapetahua, Chiapas; 4.7 % (1415 ha) in Mecayapan, Veracruz; 1.2 % (4571 ha) in Balancán, Tabasco; and 1.2 % (11 342 ha) in Carmen, Campeche (^{SIAP, 2017}; ^{INEGI, 2016}). The average area (±1 SE) of the digitalized plantations was different from one municipality to another. From greater to lesser, they were ranked as follows: Carmen (13.9±1.2 ha), Acapetahua (12.4±1.1 ha), Balancán (6.2±0.5 ha), and Mecayepan (3.6±0.3 ha). In some municipalities, plantations cover huge areas, while in others they are concentrated in certain regions (Figure 3).

Figure 3 Municipalities under study with the largest surface sown with oil palm (Elaeis guineensis). Oil palm sites match areas identified through high-resolution images (2009-2016) from Google Earth. 

The preliminary land use analysis in the areas where oil palm plantations are established showed they previously featured various covers (Figure 4). The land use and vegetation series showed that, in 1980, the areas currently covered by palms were used for agriculture or had other types of vegetation (rainforest, aquatic vegetation, and pastures). Pastures were the dominating vegetation (52 %). This cover was dominant in 1990 (71 %). The major transition took place between 1990 and 2005. During that period, 69 % of the pasture cover became agricultural land. Agricultural cover dominated since 2005 (˃90 %), while rainforest areas occupied the lowest rank (approximately 6 %; Figure 4).

Figure 4 Cover transition matrix (1980-2016) of the surface currently used for oil palm in the municipalities of Acapetahua (Chiapas), Balancán (Tabasco), Carmen (Campeche), and Mecayapan (Veracruz). Arrows indicate the direction of the cover change and black numbers next to the arrows indicate the percentage of change from one cover to another during that period. 

The results of this sample taken from four municipalities in southeastern Mexico point out that, starting more than a decade ago, 6 % of the areas where oil palm is cultivated was occupied by natural vegetation. These findings match the results of other studies (^{Vijay et al., 2016}; ^{Furumo and Aide, 2017}), which determined that trends in southeastern Asia and Africa are opposed to the trends in Mexico and South America, where the palm plantation areas have increased in areas that had been previously transformed. Although the figure may not seem significant, these remainders of natural vegetation provide ecosystem services -such as hydric and weather regulation- and they shelter flora and fauna (^{Foley et al., 2005}). Their transformation threatens certain species (^{Tan et al., 2009}).

Municipality level analysis

Land use change analysis per municipality illustrates the diverse trends in land use change that this crop has created in southeastern Mexico. Since a decade ago, the municipality of Acapetahua, Chiapas, has the greatest surface sown with oil palm in the state and the country. Although grass dominates crops in this municipality, their area has decreased constantly since 2008, while oil palm crops have enjoyed a constant increase (Figure 5). This replacement is the consequence that growing oil palm -both in the municipality, as well as in the whole state- is more profitable than creating and maintaining pastures. Only in 2016, oil palm had the highest production value in the municipality and the price of its fresh fruit per tonne was more than twice the price of pasture ($1525 and $498 MXN or US$79.5 and US$25.9 dollars; ^{SIAP, 2017}), making it the second most important crop in the municipality. Additionally, this municipality has reported that staple food crops -such as corn, banana, mango, and some vegetables- were replaced and abandoned in favor of the cultivation of oil palm (^{Fletes et al., 2013}; ^{Mazariegos et al., 2014}).

Figure 5 Temporal variation of the surface (ha) sown with the major crops in the municipality of Acapetahua, Chiapas (SIAP, 2017). 

The vegetation transition analysis for the municipality of Acapetahua showed that -during the 2005-2013 period- most of the area that is currently identified as oil palm plantations, was used for agricultural purposes (99 %) and by aquatic vegetation (1 %) (Figure 6). The oil palm plantation surface (1 %) was recorded where aquatic vegetation was found. Now, these ecosystems are restricted in Mexican territory (^{Landgrave and Moreno-Casasola, 2012}). They are major carbon sinks and they provide protection against floods and storms (^{Evaluación de los Ecosistemas del Milenio, 2005}; ^{Mitsch and Gosselink, 2015}). These ecosystems were not considered attractive for agriculture in Indonesia; however, after they were drained, they became apt for this crop. Therefore, many countries have accelerated their conversion (^{Tan et al., 2009}).

Figure 6 Cover transition matrix (1980-2016) of the surface now used for oil palm in the municipality of Acapetahua (Chiapas). Arrows indicate the direction of the cover change and black numbers next to the arrows indicate the percentage of change from one cover to another during that period. 

The municipality of Carmen, Campeche, ranks second in the country, based on the surface sown with oil palm, with a 11.37 t ha^-1 output (^{SAGARPA, 2017}). In early 2000, this municipality ranked first and its potential for sowing oil palm was 140 000 ha (^{SAGARPA, 2017}). Paddy rice, grain sorghum, and grain corn occupied most of the sowing area in several periods; however, the cultivation of oil palm increased in recent years. In 2014, it increased by slightly more than 100 % and it currently overcomes the area used to grow grain corn, which was the most important crop in the municipality (Figure 7). Economic factors explain this crop replacement phenomenon. In 2016, the production value of oil palm in the municipality amounted to almost US$1.5 million dollars, while corn amounted to slightly more than US$1 million dollars (^{SIAP, 2017}). If this trend continues, oil palm could replace every other crop in the municipality.

Figure 7 Temporal variation of the surface (ha) sown with the major crops in the municipality of Carmen, Campeche (^{SIAP, 2017}). 

In the municipality of Carmen, the greatest cover transition took place between 1990 and 2005, when 46 % of pasture cover was transformed into agricultural land. This trend fits the perception that local land owners have about the higher profitability of oil palm cultivation versus range livestock production (^{Revel-Mouroz, 1980}; ^{Tudela, 1989}). Although most of the land, from 2005 to 2013, was identified as agricultural land (˃70 %), this municipality underwent the highest transformation of natural vegetation areas (24 %; Figure 8). Forest surface might be smaller, since the transition from agricultural cover to tropical forest recorded by INEGI over a four-year period, does not match the recovery rate foreseen for forests in the region. In this municipality since pre-Hispanic times, timber and non-timber products -such as latex, medicinal plants, and fibers, among others- were extracted from savanna forests, as well as from low- and middle-altitude tropical forests. Livestock production and crops currently generate a landscape dominated by extensive areas featuring pulses and cultivated and pasture lands (^{Noriega-Trejo and Arteaga, 2010}); however, the remaining original ecosystems house a great biological diversity and provide major environmental services.

Figure 8 Cover transition matrix (1980-2016) of the surface now used for oil palm in the municipality of Carmen, Campeche. Arrows indicate the direction of the cover change and black numbers next to the arrows indicate the percentage of change from one cover to another during that period. 

In 2016, the municipality of Balancán ranked seventh and first in the surface sown with palm in the country and the state, respectively, with a 12.61 t ha^-1 output (^{SAGARPA, 2017}). Corn crops dominate agriculture in this municipality, followed by sorghum. Since 2012, palm cultivation has increased and has almost matched sorghum (Figure 9). In Tabasco, as well as in Chiapas, the surface where staple crops are sown has undergone a significant reduction (^{Fletes et al., 2013}). In this municipality, 96 064 ha (27 % of its surface) could potentially be used to establish oil palms (^{SAGARPA, 2017}).

Figure 9 Temporal variation of the surface (ha) sown with the major crops in the municipality of Balancán, Tabasco (^{SIAP, 2017}). 

Since 2005, only agricultural (˃90 %) and rainforest (10 %) covers existed in the area currently occupied by palm plantations. Therefore, the establishment of plantations caused deforestation in the municipality (Figure 10).

Figure 10 Cover transition matrix (1980-2016) of the surface now used for oil palm in the municipality of Balancán, Tabasco. Arrows indicate the direction of the cover change and black numbers next to the arrows indicate the percentage of change from one cover to another during that period. Only major cover classes are shown; therefore, the total percentage sum for each period does not always reach 100 %. 

In the municipality of Mecayapan, Veracruz, corn crops occupy the largest surface since more than a decade ago (Figure 11). Oil palm crops have held on to the second place, with a 50 % increase from 2003 to 2004. This is the municipality where palm is sown in a smaller area. As a consequence of the low output of fruit in the state (9 %; ^{INIFAP, 1999}) -whereas Chiapas reached 18 % (^{Mata, 2014})- and the aid and credits granted by the state government, farmers keep expanding the palm crop.

Figure 11 Temporal variation of the surface (ha) sown with the major crops in the municipality of Mecayapan, Veracruz (^{SIAP, 2017}). 

As Balancán, since 2005 only agricultural (˃90 %) and rainforest (2 %) covers were found in the area where palm plantations are currently established (Figure 12).

Figure 12 Cover transition matrix (1980-2016) of the surface now used for oil palm in the municipality of Mecayapan, Veracruz. Arrows indicate the direction of the cover change and black numbers next to the arrows indicate the percentage of change from one cover to another during that period. Only major cover classes are shown; therefore, the total percentage sum for each period does not always reach 100 %. 

The polygons identified for this study make up a significant sample of oil palm cultivation in Mexico and the first effort to determine the configuration, location, and size of their plantations in the Mexican tropics. The results -complemented with the SIAP’s statistics- show the trends in land use change undergone in the regions where oil palm is grown. A subsequent step would be to specify the spatial scale of the study and include polygons with young palms, in order to evaluate their impact in the country. This could be obtained from high-resolution images that enable the development of land use maps with minimum comparable mapping areas and cover classes representing each region. The automatization in the detection of palm crop areas would help to keep the information on the national increment up to date.

The greatest increase in the surface of palm plantations at regional scale has not led to changes of native vegetation. This is similar to trends described for Malaysia (^{Lam et al., 2009}), Colombia (^{Pardo et al., 2015}), and other Latin American regions (^{Furumo and Aide, 2017}). However, in some municipalities, plantations were established in areas where forest was found. These patterns agree with warnings that oil palm plantations indeed cause deforestation (^{Feintrenie et al., 2010}; ^{Corley, 2009}) and loss of biodiversity (^{Koh and Wilcove, 2008}; ^{Tan et al., 2009}).

The replacement of staple crops by oil palm at the local level is a sign of the struggle between staple and industrial crops for land use, and shows the potential problems for local food security (^{Carrere, 2006}). There is scarce information about the effect of oil palm in food security: in one Colombian municipality, it was estimated that palm plantations took the place of almost 5000 ha of agri-food crops (^{Herrera and Cumplido, 2015}). Brazil, Colombia, and Mexico are examples that the biofuel monocrops (including oil palm) pose challenges to food security and local life strategies (^{Selfa et al., 2015}).

The economic income generated by the oil palm industry is the main reason that its expansive trends remain strong (^{Feintrenie et al., 2010}), even for Mexican small farmers. For example, small rural producers from Chiapas -whether they have low, middle, or high incomes- report this crop as their main source of income (^{Castellanos-Navarrete and Jansen, 2016}). Governmental programs encourage the production and sowing of these crops; besides, they provide access to agrochemicals and infrastructure credits. Therefore, current Mexican development policies lead to the consolidation of strategies that increase palm cultivation, maximize its production, expand its plantations in states where they already exist, and introduce them in Oaxaca, Puebla, Quintana Roo, and Yucatán (^{SAGARPA, 2017}). Additionally, there are plans to maximize domestic and international commercialization, which will be accompanied by more agrochemicals, herbicides, and fertilizers.

Given the worldwide significance and the exponential growth of this crop, comprehensive management plans and certification processes for sustainable palm plantations must be developed (^{Ivacic et al., 2016}). This would minimize its impact on ecosystems and maximize its benefits for local population. However, despite their importance, current global initiatives aimed at certifying sustainable oil palm plantations (Roundtable on Sustaniable Palm Oil) are insufficient and not very effective (^{Laurance et al., 2010}; ^{Cattau et al., 2016}; ^{Azhar et al., 2017}). The problem can be partially explained because these programs are only accessible to production companies with large-scale monocrop plantations which provide very few benefits for the local communities. Therefore, developing alternative certification programs for small producers with diversified smallholdings would be more sustainable -both in environmental and social terms (^{Azhar et al., 2017}). In the case of Mexico, SAGARPA suggests that producers and cooperatives consider participating in sustainability certification programs that provide an added value to the product and certainty for the consumers. This certification includes a clause that determines that new plantations cannot be established in natural areas, which represents a first step towards an improvement (^{SAGARPA, 2017}). The feasibility of these strategies must be analyzed in the light of national reality. Two fundamental aspects regarding the preservation of the environmental integrity and local livelihoods in tropical landscapes should be subjected to detailed studies: 1) regulating the establishment of oil palms without increasing the agricultural border or substituting native vegetation, not even outside protected natural areas; and 2) maintaining and increasing local and regional food security. This preliminary diagnosis can act as a model for these studies.

Conclusions

Oil palm plantations were established in southeastern Mexican states and based on national agricultural development programs, this trend is expected to continue. Most of the surface where these plantations were established, comes at the expense of other agricultural systems, mainly those dedicated to staple crops. A smaller ratio of the surface currently used to grow oil palm comes from the transformation of primary vegetation, which suggest that for some municipalities it can have a significative impact on biodiversity, as some of these ecosystems have restricted distributions.

Financial support drives the growth of oil palm plantations in Mexico, but it is not complemented by environmental or socio-economical impact assessments. Management plans and certification programs for sustainable production which maximize environmental integrity and local livelihoods, can be implemented.