Private property and Mennonites are major drivers of forest cover loss in central Yucatan Peninsula, Mexico

Land Use Policy 69 (2017) 474–484

Contents lists available at ScienceDirect

Land Use Policy journal homepage: www.elsevier.com/locate/landusepol

Private property and Mennonites are major drivers of forest cover loss in central Yucatan Peninsula, Mexico

MARK

⁎

Edward A. Ellisa,b, , José Arturo Romero Monterob, Irving Uriel Hernández Gómezb,c, Luciana Porter-Bollandd, Peter W. Ellise a

Centro de Investigaciones Tropicales, Universidad Veracruzana, José María Morelos 44, Col. Centro, Xalapa, Veracruz, C.P. 91000, México Equilibio en Conservación y Desarrollo, A.C. Calle 5 de Mayo No. 28, Colonia Portón Colorado, C.P. 91158 Xalapa, Veracruz, Mexico c Facultad de Ciencias Agropecuarias, Universidad Veracruzana. Circuito Gonzalo Aguirre Beltrán, Isleta, C.P. 91090, Xalapa, Veracruz, Mexico d Instituto de Ecologıa, A.C. Red de Ecología Funcional, Antigua Carretera a Coatepec No. 351, El Haya, C.P. 91070 Xalapa, Veracruz, Mexico e The Nature Conservancy, 4245 Fairfax Drive, Suite 100, Arlington, VA 22203, USA b

A R T I C L E I N F O

A B S T R A C T

Keywords: Deforestation Land tenure Drivers Logistic regression Mennonites Campeche Mexico

The role of land tenure and Mennonites as drivers of deforestation in the Central Yucatan Peninsula has not been empirically assessed. We evaluate different drivers and their relationship to forest cover change between 1986 and 2015 and assess how land tenure and Mennonite communities impact forest cover loss in the Municipality of Hopelchen, Campeche, Mexico. This study shows that forest cover loss has increased in the last decade (2005–2015), and that land tenure regime type is associated with this loss. Throughout the study period, statistical comparisons show rates of forest cover loss were significantly higher in private and federal property compared to forests in ejidos (communal property). Forest cover loss in Mennonite private property was also significantly higher than in non-Mennonite owned private property. The role of land tenure and the expansion of the agroindustrial production model as major drivers of forest cover loss in the region provide important insight into developing municipal land use plans and conservation strategies to reduce deforestation. Programs, incentives and policy directed towards forest conservation in the region that typically target ejido communities, will need to consider the growing trend of private property expansion within federal lands and work more closely with private property owners including Mennonite communities if deforestation reduction programs are to be successful.

1. Introduction Tropical deforestation amounts to approximately 11% of anthropogenic greenhouse gas emissions (Tyukavina et al., 2015), which represents a major challenge for global climate change mitigation. Since 2008, the United Nations Collaborative Programme1 for Reducing Emissions from Deforestation and Forest Degradation (REDD+) has become an important mechanism to drive national land use policies in developing countries and promote conservation, sustainable forest management, enhancement of carbon stocks, and sustainable rural development (Angelsen et al., 2012; Danielsen et al., 2011; UN-REDD Programme, 2010). In this context, Mexico has emerged as an international leader, with the ambitious goal of reducing deforestation and net carbon emission rates to zero by 2020 (CONAFOR, 2014; CONAFOR, 2010). Although deforestation in México has decreased over the past two decades, more than 150,000 ha of forest cover are still

⁎

1

being cleared annually (FAO, 2015). The majority of this loss occurs within tropical forest landscapes, such as the Yucatán Peninsula (Céspedes-Flores and Moreno-Sánchez, 2010; Challenger and Soberón, 2008; Velázquez et al., 2002). To develop effective strategies at local levels and sound policies at national and state levels that reduce deforestation, it is essential to identify and evaluate the proximate (direct) and underlying (indirect) causes (or drivers) of forest cover loss (Salvini et al., 2014; Kissinger et al., 2012; Geist and Lambin, 2002). In this regard, previous studies have investigated land tenure as an underlying cause of deforestation, particularly in tropical regions (Fearnside, 2001; Doherty and Heike, ́ nna and Young, 2014). For example, studies in Brazil, 2011; SantA Ecuador and Haiti have found higher rates of deforestation where land tenure is less secure (Fearnside, 2001; Cattaneo, 2001; Messina et al., ́ nna and Young, 2014). Securing tenure 2006; Dolisca et al., 2007; SantA rights and community ownership of forests have been proposed as

Corresponding author at: Centro de Investigaciones Tropicales, Universidad Veracruzana, José María Morelos 44, Col. Centro, Xalapa, Veracruz, C.P. 91000, México. E-mail addresses: [email protected], [email protected] (E.A. Ellis). Part of the United Nations Framework Convention on Climate Change (UNFCCC).

http://dx.doi.org/10.1016/j.landusepol.2017.09.048 Received 10 March 2017; Received in revised form 27 September 2017; Accepted 27 September 2017 Available online 06 October 2017 0264-8377/ © 2017 Elsevier Ltd. All rights reserved.

Land Use Policy 69 (2017) 474–484

E.A. Ellis et al.

Fig. 1. Location of study area showing land tenure regimes, Mennonite settlements and land tenure regimes in the Municipality of Hopelchén, Campeche, Mexico.

currently occupy 60% of the countrýs total forest cover (Madrid et al., 2009). However, reform of the Agrarian Law in 1992 allowed provisions for voluntary privatization of ejidos (Barnes, 2009). Although most rural ejidos have shown resilience towards privatization, there is evidence of structural changes within their institutional configuration (Barnes, 2009; DiGiano et al., 2013). This destabilization of social

mechanisms to reduce tropical deforestation (Kissinger et al., 2012), nonetheless research investigating the role of private versus communal ownership of forests in tropical regions worldwide is scarce. Mexico’s government-sanctioned communal property system (including both ejidos, and comunidades agrarias), may be playing an important role in regulating land use change, since communal properties 475

Land Use Policy 69 (2017) 474–484

E.A. Ellis et al.

2. Materials and methods

tenure towards a policy favoring private ownership and agricultural development since the 1990s raises concerns regarding its effect on forest change in the country. Previous studies have shown conflicting results regarding the relationship between land tenure regimes and forest loss in Mexico. At the national level, Deininger and Minten (1999) detected no difference in forest loss among communal and private properties. Bonilla-Moheno et al. (2013), found higher rates of forest cover loss in ejidos than in private property nationwide, although comunidades agrarias showed the lowest rates of forest loss. At the subnational level, Barsimantov and Kendall (2012) investigated eight different states within the country, showing lower rates of forest loss in areas occupied by communal property (ejidos and comunidades agrarias). Also, in Quintana Roo, DiGiano et al. (2013) found lower rates of deforestation in ejidos that had not experienced informal parcelization and privatization processes, especially when they had engaged in community forestry. In tropical forest landscapes of the central Yucatan Peninsula, the relationship between land tenure regimes (e.g. private, communal or federal property) and forest cover loss has not been assessed. Further, the dynamic nature of changes in land tenure brought about by the agrarian reform and policy shifts mentioned above, as well as changes in population configurations and production systems have not been examined in relation to forest cover loss. The effect of Mennonite settlement on forest cover change is of particular concern (Danga Pelissier, 2015). Mennonites, mostly emigrating from northern México, have settled in the region since the mid-eighties, invited and incentivized by the government to populate and agriculturally develop rural municipalities. In other tropical regions such as lowland Bolivia (Tejada et al., 2016; Killeen et al., 2008; Steininger et al., 2001) and the Paraguayan Chaco (Mereles and Rodas, 2014; Fitzgerald and Stronza, 2009), Mennonite settlements have been associated with large-scale deforestation. Also, in neighboring Belize, Trapasso (1994) reports that Mennonite immigration during the 1990s was a “major destructive force” of tropical forests. This is demonstrated by Patterson (2014) through remote sensing analysis of deforestation and agricultural land use change in Northern Belize. In particular, the state of Campeche, has been characterized by a population dynamic influenced by different migration surges, including Mennonite settlements, as well as by government policies fueling agricultural development (Schüren, 2003). The municipality of Hopelchen provides a useful case study for understanding the interaction of these dynamic and often conflicting influences. The territory is characterized by conflicting policy agendas: on the one hand, conservation policies, such as payment for environmental services and forest management, are mostly focused on ejido communities with traditional agricultural practices inherited by the centuries-old Mayan population; on the other hand, a policy of expanding industrial agriculture and an immigrating Mennonite population results in forest clearing, counteracting conservation policies. (Gómez-González, 2016; Porter-Bolland et al., 2008; Ellis and Porter-Bolland, 2008). For evaluating different deforestation drivers in the municipality of Hopelchen, and particularly the role of land-tenure, Mennonites and land use policy, we use remote sensing analysis to measure forest cover loss between 1986 and 2015 and through statistical analyses analyze different factors as possible drivers. We further compare rates of forest cover loss between different land tenure regimes (private property, communal ejido land, and federal land) and assess the impact on deforestation by Mennonite migration and agricultural expansion in the study area. We discuss these findings in the context of their implications on forest conservation and sustainable rural development planning in the central Yucatan Peninsula.

2.1. Study area The municipality of Hopelchen is located in the northeastern portion of the state of Campeche, Mexico, in the central region of the Yucatan Peninsula known as Los Chenes, which is part of the Puuc region (Fig. 1). It encompasses an area of 7460 km2, containing flat terrain to the north and rolling hills to the south, with elevations ranging from 90 to 350 m.a.s.l. (INEGI, 2009). Mean temperature is 26 °C and mean precipitation ranges from 1100 to 1300 mm/year, with a distinct dry (December through May) and wet season. It is a region with karst topography, containing subterraneous drainage and no superficial flows of water. Soils are predominantly Redzinas, described as red or black shallow soils, low in fertility, but with good structure, drainage and organic matter (Porter-Bolland et al., 2007). The forested landscape of the municipality is made up of a mosaic of upland and lowland seasonally inundated forests, as well as patches of secondary vegetation in different successional stages (mostly due to anthropogenic disturbance) primarily because of clearing for agriculture (Porter-Bolland et al., 2008). The dominant natural vegetation type in the landscape is the tropical dry forest (89%); agriculture and pasture land uses comprise around 8 and 2% respectively, while urban and residential areas occupy less than 1% (INEGI, 2009). Hopelchen’s population includes 38,000 inhabitants (INEGI, 2009), mostly of Mayan descent (75%). Inhabitants of Mennonite descent started arriving into the area since the 1980s and currently comprise up to 14% of the total population (Danga Pelissier, 2015). Livelihood systems throughout Hopelcheńs territory differ between the north and the south, as well as between more traditional subsistence oriented communities (mostly Mayan families), and more market-oriented agroindustrial family units (mostly Menonnites). The northern and central regionis the most populous. This area has biophysical characteristics that differ from that of the south, which is a smaller area known as La Montaña. While land in the north and central parts of the territory can mostly be cultivated for agriculture, La Montaña has a more rugged topography less suitable for agriculture, thereby favoring the maintenance of a forest cover (Porter-Bolland et al., 2008). Livelihood strategies at Hopelchen are diverse (Schüren, 2003). Mechanized agriculture for producing maize, soy and sorghum for commercial markets are characteristic mostly in the north and central parts of the municipality (mostly under Mennonite cultivation). Mechanized agriculture is for growing only maize in Mayan ejidos, but small-scale slash and burn agriculture or milpa for subsistence (mostly at La Montaña) is also commonly practiced. Throughout the municipality cattle ranching is important, as well as apiculture for exporting honey to the European market. In La Montaña in particular, timber and non-timber forest products are a part of community livelihood (GómezGonzález, 2016; INEGI, 2015a; Porter-Bolland et al., 2008; Schüren, 2003). Government incentives for agricultural intensification have been in place in the region since the 1980s (Danga Pelissier, 2015), promoting agroindustrial development of commercial crops such as maize, soy and sorghum (INEGI, 2015b). While agroindustrial policy incentives have mostly fueled Mennonite production and expansion, Mayan communities have also adapted their productive systems, sometimes adopting monospecific mechanized agriculture at the expense of traditional diversified policultures. These incentives have generated tension in the municipality, as Mennonite communities rent ejido land and displace traditional agricultural plots, but also affect Mayan community honey production through agrochemical use, deforestation and the adoption of genetically modified cultivars (Villanueva-Gutiérrez et al., 2014; Rendon-von Osten and Dzul-Caamal, 2017; Gómez-González, 2016). These agroindustrial policies are also creating tension with regionwide conservation initiatives. The Puuc Chenes region surrounding Hopelchen has been established as an Early Action Area for REDD+ 476

Land Use Policy 69 (2017) 474–484

E.A. Ellis et al.

Fig. 2. Growth of private property within federal lands from 2005 to 2010 in the Municipality of Hopelchén, Campeche, Mexico (dotted areas represent private property growth).

between 2005 and 2010. The majority (88%) was converted from federal lands. The most current National Agrarian Registry (RAN, 2016) data, shows that ejido property in the municipality amounted to 508,000 ha (64%), while private property extends over 141,000 ha (18%) and federal or public land occupy 140,000 ha (18%) (RAN, 2016). Thirty-one percent of private lands were Mennonite owned (RAN, 2016). Land under ejido property has remained stable, with the exception of 7000 ha sold to Mennonites by an ejido in 2005.Private property, on the other hand increased from 91,000 ha (12% of the municipality) to 142,000 ha (18%) over the same period. Current information from interviews and maps shown at the cadastral office of the municipality in 2017 seem to indicate that this trend has continued aggressively, and private property may now occupy as much as 30% of the municipality. However, it was impossible to get reliable and more current land tenure data from the RAN in order to verify the current land tenure regime distribution in the municipality.

programs and activities in the Yucatan Peninsula due to its location within the Selva Maya, a contiguous mass of forest cover that serves as an important biological corridor as well as a major carbon stock in the region (Bezaury-Creel et al., 2014). REDD+ projects in the region seek to design and implement ecological land use planning, community forest management, and improved sustainable agricultural production in the municipality’s agriculture-dominated northern region(BezauryCreel et al., 2014). The southern portion of Hopelchen is also part of a state level reserve Balam Kin and part of its area is considered an area of influence of the Calakmul Biosphere Reserve (CBR), the largest protected area in Mexico (Acopa and Boege, 1998). Changes regarding land tenure are also part of the municipalitýs history. Private property and federal land extended in the municipality in the 19th century before much of the land was granted as ejido after the Mexican Revolution, mostly from the 1920s to 1950s (Cantún Caamal and Pat Fernández, 2011). However, there has been a growing trend in the last decade for the expansion of private property within federal lands, much of which has been occupied by Mennonites. Fig. 2 shows that 57,000 ha of land was converted to private property 477

Land Use Policy 69 (2017) 474–484

E.A. Ellis et al.

land use was also calculated for the periods 1986–1995, 1995–2005 and 2005–2015 applying the same formula.

2.2. Forest cover change analysis To evaluate forest cover change, land use and land cover (LULC) in the municipality was classified for the years 1986, 1995, 2005 and 2015, using Landsat 5 TM (1986 and 1995), Landsat 7 ETM (2000) and Landsat 8 OLI (2015) images for the following dry season dates: April 4, March 22, April 26 and February 25, respectively. All images underwent geometric and atmospheric corrections using ENVI 4.7. Accuracy of geometric correction for multi-date image comparison was less than one pixel (30 × 30 m). The images were classified into four LULC types by applying an object-oriented approach with eCognition 8.0: (1) agriculture, (2) forest, (3) urban-infrastructure and (4) water. ECognition employs a multi-resolution segmentation process that extracts image objects, or regions, which are classified based on user inputs (e.g. field ground-truth coordinates, visual interpretation or thematic maps). We applied the segmentation process using RGB spectral bands set at 5, 4, 3 for Landsat 5 and 7 images, and 6, 4, 5 for Landsat 8 images using Level 1 with scale parameter 15. LULC classifications were guided by a sample of 119 ground-based observation points gathered from July to October 2015. Observation points were collected in 7 sampling transects (about 30–50 km long) located along selected secondary or tertiary roads within the municipality, stopping at intervals of 3 km and recording vegetation type and land use. To aid in classification and accuracy assessment, we also used visual interpretation of high-resolution imagery of the study area, available through Google Earth ™. We conducted a thematic accuracy assessment for the 2015 classification using a set of 150 randomly distributed points (> 2000 m apart) which were assigned one of LULC classes 1, 2 and 3 based on visual interpretation of current high resolution images from Google Earth™. We calculated confidence intervals for the 3 LULC classes and adjusted areas accordingly using the “AccurAssess” QGIS analysis tool (Olofsson et al., 2014; Mas et al., 2014). We subsequently calculated forest cover change rates for the municipality and each individual land tenure parcel, categorized as private property (PP), ejido (EJ) and federal land (FL). Forest cover change rates are based on the 2010 land tenure regime boundaries;tenure data for the full time period evaluated in our forest cover change period were unavailable or impossible to obtain from the RAN. As mentioned above, the area under ejido property appears stable while PP has been expanding within FL. By using 2010 land tenure data in our study, we therefore over-represent PP parcels and under-represent areas under FL during the 1986–1995 and 1995–2005 periods. Conversely, we underrepresent PP parcels and over-represent FL areas during the second half of the 2005–2015 analysis period. To complicate matters further, the transition and formalization of FL into PP often lack transparency, and can take up to 15 years or more, making it very difficult to get more specific or accurate forest cover change calculations for PP and FL tenure regimes over time. Although one can assume that deforested areas in FL will eventually become PP. However, since ejido coverage in the municipality is stable during our study perido, their comparison to forest cover changes in PP/FL are valid. Forest cover change rates were calculated for each PP, EJ and FL parcel (2010) for the periods 1986–1995, 1995–2005, 2005–2015 and 1986–2015, applying the formula:

2.3. Modelling deforestation drivers We applied binary logistic regression to model and evaluate a set of potential environmental and socioeconomic forest cover loss drivers in Hopelchen (Morales-Barquero et al., 2015). To model the spatial occurrence of deforestation we classified the 2015, 2005 and 1986 images into binary forest/non-forest classes. Change maps were then produced for periods 1986–2015 and 2005–2015; four categories were assigned (1) forest cover loss, (2) forest cover gain, (3) forest remaining forest, and (4) persistent non-forest. We then produced three binary response maps used as dependent variables for our logistic regression analyses: (1, current condition) non-forest/forest in 2015, (2, cumulative loss) 1986–2015 forest loss/no forest loss, and (3, recent loss) 2005–2015 forest loss/no forest loss. All non-forest and forest loss pixels were assigned a value of 1; forest and no forest loss pixels were assigned a value of 0. Explanatory spatial variables used to model the probability of forest loss events using logistic regression included: (1) elevation, (2) slope, (3) soil type, (4) distance to roads, (5) distance to settlements, (6) population, (7) socioeconomic marginalization, (8) land tenure type, (9) agricultural subsidies, (10) forest management, and (11) payment for environmental services (PES) program (Table 1). Explanatory variables were selected to capture the available set of potential environmental and socioeconomic conditions that have been associated with deforestation or forest conservation in the region and the tropics (DiGiano et al., 2013; Alix-Garcia, 2007; Dalle et al., 2006; Klepeis and Vance, 2003; Ellis and Porter-Bolland, 2008; Porter-Bolland et al., 2007). ArcGIS 10.2 (ESRI) was used for geospatial processing. 2.4. Comparison of forest cover loss rates by land tenure regimes and by Mennonite ownership We further evaluated forest cover loss in different land tenure regimes by statistically comparing forest cover change rates in PP, EJ and FL parcels in the municipality of Hopelchen for the periods 1986–1995, 1995–2005, 2005–2015 and 1986–2015. The sample of property parcels in the municipality of Hopelchen included: 41 EJ, 48 FL, and 303 PP polygons (Fig. 1). Of the PP polygons, 98 were identified as Mennonite-owned based on SAGARPA (2016) and RAN (2016). EJ parcels or territories are larger on average (11,219 ha), in contrast to PP parcels and FP land, (457 and 2910 ha, respectively). The largest EJ polygon is 75,579 ha while the largest PP and FP polygon is 10,583 ha and 33,567 ha respectively. We tested our datasets of forest cover change rates for EJ, FL and PP during the four above-mentioned periods for normality using the Shapiro-Wilk test. Due to non-normality of the data, statistical comparisons of forest cover change rates were made using the non-parametric Kruskal-Wallis test. In addition, multiple pairwise comparisons of rates of forest cover loss between each land tenure type and for each analysis period were made using the Steel-Dwass-Critchlow-Fligner procedure. Finally, forest cover loss was compared between Mennonite and non-Mennonite private property parcels applying the MannWhitney test for periods 1986–1995, 1995–2005 and 2005–2015. We used XLSTAT 2016 to perform statistical analyses.

dn = [S2/S1]1/n − 1 where dn = deforestation rate S2 = forest cover in time period two S1 = forest cover in time period one and n = number of years between time periods (Palacio-Prieto et al., 2004). Negative rate values denote forest cover loss and positive rate values indicate forest cover gain. Finally, the rate of change of agricultural

3. Results 3.1. Forest cover loss in Hopelchen The overall accuracy of our 2015 LULC classification was 90.7% (CI: 82.1–99.3%). Forest cover loss in the municipality of Hopelchen was stable between 1986 and 1995, although, since 1995, forest cover loss has been gradually increasing, particularly during the last decade 478

Land Use Policy 69 (2017) 474–484

0.5 N/A 301. N/A 99,430 −0.5 N/A 275 ha N/A 20,068 ha

Fig. 3. Forest cover loss rates and agricultural land use expansion rate in the municipality of Hopelchen, Campeche during the periods 1986–1995, 1995–2005 and 2005–2015.

(2005–2015; Fig. 3). Total net forest loss accumulated from 1986 to 2015 was 46,000 ha (6% of the municipality) of which 75% occurred within the last 10 years. Deforested area in the municipality increased from 91,302 ha in 1986 to 137,008 ha in 2015. Forest cover change rates in the municipality have gone from a nearly flat annual rate of 0.0017% for the period 1986–1995, to −0.16% between 1995 and 2005, and −0.51% for the 2005–2015 period (Fig. 3). The overall forest cover change rate for the entire analysis period (1986–2015) was −0.23%. The rate of agricultural expansion rose during all three periods (average = 2.1%), but most steeply during the most recent period (3.1%) (Fig. 3). Therefore, recent forest loss was caused mostly by the expansion of agriculture.

30 m Pixel Land Parcel (All) Ejido Land Parcel Ejido

4. 2657 4124 N/A 435 3.2% 2016 m 5827 m N/A 317 Pixel Pixel Pixel Pixel Pixel 30 m 30 m 30 m 30 m 30 m

Digital elevation model from Instituto Nacional de Estadística y Geografía (INEGI, 2013) (INEGI, 2013) Road network vector data from INEGI (2015b) (1: 50,000) Location of settlements from INEGI (2011) Soils map form INEGI (2007) (1:250,000) Localities and 2010 Population and Household Census from INEGI (2011) Population marginalization index from CONAPO (2012) Georeferenced land tenure parcels and polygons from RAN (2016) SAGARPA (2016) CONAFOR (2016) List of authorized forest management plans from SEMARNAT (2015) 33 131 m 30 m Pixel

SD Mean Spatial Unit

Source

E.A. Ellis et al.

Percent slope derived from DEM Euclidean distance from road in meters Euclidean distance from population settlements in meters Type of soil base on FAO classification Population distribution based on surface interpolation (IDW) of 2010 population data by settlement or locality Socioeconomic marginalization index Type of tenure regime: ejido, private property or federal land Area in hectares assigned under the agricultural subsidy program PROCAMPO Presence of polygon or area set aside for the Payment for Environmental Services Area in hectares authorized for timber extraction under forest managent plan Slope Distance to Roads Distance to Settlements Soil Type Population Density

Table 2 summarizes the binary logistic regression logit best model results for the response variables of (1) current areas without forest Table 2 Standard coefficients for variables selected in binary logistic regression best models for the probability of deforestation present in 2015, deforestation occurring from 1986 to 2015 and recent deforestation from 2005 to 2015.

Marginalization Land Tenure Regime Agricultural Subsidy Payment for Environmental Services Forest Management

Elevation in meters above sea level (masl) Elevation

Description Variable

Table 1 Description of explanatory variables and spatial layers used in binary logistic regression models.

3.2. Drivers of forest cover loss in the municipality of Hopelchen

Model Variables

β

S.E.

Wald Chi2

Pr > Chi2

Deforested in 2015a DISTRD FORMAN SOIL-R SOIL-Lu SOIL-Li SOIL-G SOIL-V

−1.077 −18.630 −0.339 −0.045 −0.298 0.044 −0.085

0.349 10.132 0.170 0.111 0.146 0.131 0.107

9.511 3.381 3.973 0.166 4.182 0.112 0.627

0.002 0.066 0.046 0.684 0.041 0.738 0.428

Deforested from 1986–2015b SLOPE −0.475 DISTRD −0.952 MARGI −0.267

0.190 0.323 0.126

6.247 8.661 4.516

0.012 0.003 0.034

Deforested from 2005–2015c SLOPE −0.578 TENURE-P 0.164 TENURE-F 0.285 SOIL-R −0.017 SOIL-Lu −0.018 SOIL-Li −0.026 SOIL-G 0.213 SOIL-V 0.136

0.267 0.126 0.109 0.247 0.170 0.212 0.171 0.128

4.689 1.703 6.906 0.005 0.011 0.015 1.553 1.131

0.030 0.192 0.009 0.945 0.917 0.902 0.213 0.287

a Model: −2 Log Likelihood = 49.27, df = 7, p < 0.0001, R2 (Nagelkerke) = 0.3, AIC = 199.3. b Model: −2 Log Likelihood = 32.5, df = 3, p < 0.0001, R2 (Nagelkerke) = 0.2, AIC = 166.7. c Model: −2 Log Likelihood = 47.5, df = 8, p < 0.0001, R2 (Nagelkerke) = 0.4, AIC = 109.9.

479

Land Use Policy 69 (2017) 474–484

E.A. Ellis et al.

lower in EJ than in both PP and FL and that no differences were found between PP and FL. Similar statistical results were observed for the entire 29-year analysis period (1986–2015) showing overall trends in forest cover change processes among the three land tenure regimes. The mean forest cover change rate within EJ plots was significantly lower (−0.08%) than rate for PP (−1.12%) and FP (−0.84%) plots showing lower forest cover loss in EJ. Steel-Dwass-Critchlow-Fligner multiple pairwise comparisons show that EJ forest cover rates are lower and significantly different than those of both PP and FL, and that forest cover loss rates in PP and FL were not significantly different for the entire 1986–2015 analysis period. Deforestation rates in PP with Mennonite ownership compared to those of non-Mennonite ownership in the municipality are consistently higher during all study periods. Results show that in the initial period of 1986–1995, when Mennonite migration in the municipality basically began, forest cover loss rates were not significantly different, even though the mean rate of loss in Mennonite property was twice (−1.12%) of that in non-Mennonite PP (−0.56%). In the following period (1995–2005), forest cover loss rates double in Mennonite PP (−2.29%) and decrease in non-Mennonite PP (−0.17%), showing significant differences between both. In our most recent analysis period (2005–2015) forest cover loss increases in both Mennonite (−3.19%) and non-Mennonite PP (−0.74%) and forest loss is significantly more rapid in Mennonite PP. The overall trends in mean forest cover change rates for all three land tenure regime types (PP, FL and EJ), including Mennonite and non-Mennonite PP, during the analysis period are summarized in Fig. 5. Ejidos with communal forest land show consistently less forest cover loss (or greater change rate values) than in PP and FL. For the exception of the 1995–2005 period, forest cover loss is greater (lower negative rates of change) in PP than FL and in the most recent period, PP represents the greatest annual rate of forest cover loss among all land tenure types (−1.5%). Moreover, the most extreme rate of annual forest cover loss are observed in Mennonites-owned PP in the municipality, with the most rapid and drastic rates of forest loss observed between 2005 and 2015 (−3.19%).

cover in 2015, (2) cumulative forest cover loss from 1986 to 2015 and, (3) recent forest cover loss from 2005 to 2015. Explanatory variables were used in the model to evaluate and identify drivers related with forest cover loss. Models for all 3 response variables were significant. For the model of response variable 1 (-2 Log Likelihood = 49.27, df = 7, p < 0.0001, R2 [Nagelkerke] = 0.3), distance to roads (DISTRD) and areas under forest management (FORMAN) were drivers identified as significantly related with the probability of being cleared in 2015. Areas closer to roads had higher probabilities of not having forest cover and areas under forest management were inversely related to the probability of being cleared in 2015, or directly related to maintaining forest cover in the municipality. Soil type (SOIL) was a variable selected by the overall best model, although its significance was weak in general, showing greater probability of not having forest cover in 2015 outside Redzina soil types which cover the majority (70%) of the municipality. For response variable 2 (cumulative forest cover loss from 1986 to 2015), significant explanatory variables or drivers identified in the logistic regression model (-2 Log Likelihood = 32.5, df = 3 and p < 0.0001 goodness of fit value of R2 [Nagelkerke] = 0.2.) included proximity to roads (DISTRD), slope (SLOPE), and socioeconomic wellbeing (MARGI). Low to moderate slopes showed a higher probability of forest cover loss and lower socioeconomic marginalization indices (better livelihood conditions and greater availability of capital and wealth) were related with forest cover loss during the entire study period. Response variable 3 model (recent forest cover loss from 2005 to 2015) showed the best fit (-2 Log Likelihood = 47.5, df = 8, p < 0.0001, R2 [Nagelkerke] = 0.4) and the SLOPE variable was associated with forest loss, showing more recent forest cover loss occurring on lower percent slopes. Flat areas were mostly already deforested but recent deforestation has been encroaching on areas with gentle to moderate slopes that were previously avoided, which has been corroborated also through observations in the field. SOIL was also a selected variable for the overall best model, and although its significance was the weakest, it indicated higher probability of recent deforestation in gleysol and vertisol soil types. Land tenure type or regime (TENURE) was a significant variable, with higher probabilities of recent clearing in both federal and private lands, compared to ejido or communal forest tenure regimes. Other variables, such as population (POPDIS), agricultural subsidies (PROCAMPO), and payment for ecosystem service (PES) programs showed no discernable relation to deforestation in our logistic regression models.

4. Discussion This study confirms the trend that Hopelchen is a major deforestation hotspot in the Yucatan Peninsula due to the recent expansion of commercial agriculture (Esparza-Olguín and Martínez Romero, 2011; Martínez-Romero and Esparza, 2010; Ellis et al., 2015). Our results show increasing forest cover loss rates in the municipality, particularly within the most recent period of 2005–2015 (−0.51%), with rates up to 5 times the national average for the same period (FAO, 2015). Agricultural land use grew 3% annually from 2005 to 2015 (around 3000–4000 ha per year), and most of this expansion has been for mechanized agriculture, estimated at around 3/4 of all deforested areas in 2014 (Ellis et al., 2015). The growth of mechanized agriculture in the central Yucatan Peninsula is one of the outstanding threats to tropical forests in the region (Ellis et al., 2017; Ellis et al., 2015) and globally (Laurance, 2015). These findings demonstrate the urgent need to address deforestation through adequate land use planning and effective, synergistic conservation and agricultural development strategies and policies. Identifying, quantifying and monitoring forest cover change drivers is a prerequisite for effective planning of REDD+ country strategies, and is needed for setting goals and monitoring outcomes during REDD + implementation. Our logistic regression results identify a set of important underlying drivers that are related to forest cover change dynamics in the municipality, and should be considered in strategies to reduce or stop deforestation. As in previous similar deforestation studies in the Yucatan Peninsula, proximity to roads (Bray et al., 2004; Romero-Montero, 2014) and environmental variables, such as slope and soil type (Ellis and Porter-Bolland, 2008; Porter-Bolland et al.,

3.3. Forest cover loss by land tenure regime and Mennonite property Statistical comparisons of forest cover change rates were made for PP, FL and EJ land tenure regimes in the municipality of Hopelchen (Fig. 4). During the earliest analysis period (1986–1995), no significant differences in forest cover change rates were found among land parcels of the three land tenure regime types. During this period, positive forest cover change rates (forest gain) were calculated for both EJ and FL (0.33 and 0.06), although PP had a rate of forest cover loss of −0.77%. For the following intermediate period (1995–2005) FL showed greater forest cover loss (−1.44%) followed by PP (−0.82%), while EJ had the lowest forest cover loss (−0.18%).Significant differences were found among forest cover change rates of the three land tenure regimes. SteelDwass-Critchlow-Fligner multiple pairwise comparisons indicate that forest cover change rates between EJ and FL were significant. For the most recent period (2005–2015), processes of forest cover loss increased considerably in PP (−1.48%) and EJ (−0.38%), but declined in FL (−0.95%) from the previous period. Significant differences were found among forest cover loss rates in the different land tenure regime types. Steel-Dwass-Critchlow-Fligner multiple pairwise comparisons showed forest cover loss processes were significantly 480

Land Use Policy 69 (2017) 474–484

E.A. Ellis et al.

Fig. 4. Statistical comparisons of forest cover loss rates (y-axes) in private property, federal land and ejido communal territory in the municipality of Hopelchen, Campeche during the periods 1986–1995, 1995–2005 and 2005–2015 (circles = mean, bars = maximum/minimum, box = standard deviation).

2007) are related to forest cover loss. For example, gleysols (soils located in seasonally flooded lowland forest), have increasingly been cleared in the past decades for agriculture and cattle pasture (Ellis and Porter-Bolland, 2008; Porter-Bolland et al., 2007). Forests on these soils are known as akalche, which are dense, short-statured flooded forests rich in biodiversity and important for conservation and environmental services (Romero-Montero and Ellis, 2016). Our results corroborate that flat terrain and low slopes are vulnerable to deforestation, however, they also indicate, as observed in the field, that recent deforestation has been encroaching into more moderate slopes for agricultural production. Protection and restauration of akalche forest in the region, in addition to upland forest on hills, should need to be integrated into land use and forest conservation planning. In addition, our analyses of drivers show that in the municipality of Hopelchen, a lower socioeconomic marginality index (CONAPO, 2012) was related to the increased probability of deforestation from 1986 to 2015, indicating that the ability to invest or pay for forest clearing − particularly for mechanized agricultural production − may be a positive incentive for increased forest loss, contrary to large-scale tropical

Fig. 5. Forest cover change rates in private property (total, Mennonite and nonMennonite), federal land and ejido communal property in the municipality of Hopelchen, Campeche during the periods 1986–1995, 1995–2005 and 2005–2015.

481

Land Use Policy 69 (2017) 474–484

E.A. Ellis et al.

occurring within ejidos bordering the reserve (Romero-Montero 2014; Ellis et al., 2015), demonstrating how contrasting policy and socioeconomic landscapes in these municipalities can reflect different drivers and forest cover change dynamics. Statistical and econometric analyses of broad geographical regions using coarse-scale spatial and statistical data can provide a generalized understanding of how drivers influence deforestation at regional or national levels, however, these may not be sufficient in developing and implementing strategies and policies that reduce deforestation at local and state levels. The source and scale of the environmental and socioeconomic data used (e.g. remote sensing, geographical and statistical), in addition to the availability, periodicity, and quality of this data, influences the results obtained from modelling and statistical analyses of deforestation drivers (Salvini et al., 2014; Hosonuma et al., 2012; Skole et al., 1994; Olander et al., 2008), which emphasizes the need to addressing multiple scales when planning specific intervention strategies (Kissinger et al., 2012). Our analysis provides a functional scale for municipal planning by local and state governments providing localbased results on deforestation drivers and the role of land tenure on forest cover loss rates in the region. This study also clarifies the impact of Mennonites on forest cover change processes in Central Yucatan Peninsula. We show that the increasing forest loss rates in the municipality are highly related with Mennonite private land ownership. Since their arrival in the mid 1980s, growing Mennonite communities have rapidly expanded the use of large-scale mechanized agriculture. Our results confirm the conclusions of previous studies that have shown how this trend is correlated with an increase in forest loss in the region, particularly in recent decades (Ellis and Porter-Bolland, 2008; Porter-Bolland et al., 2008). The Municipal Council of Sustainable Rural Development, as well as local inhabitants of Hopelchen (mostly of Mayan descent) have denounced this recent agricultural expansion, claiming that it impacts honey production, a major economic activity for a large proportion of the population in the municipality (Gómez-González, 2016). In 2015, a human rights and Mayan indigenous movement sued the federal government to impede an agro-industrial project in the region, promoted by the government and Monsanto Corporation, to grow genetically modified crops (GMO), claiming that increased deforestation and the use of chemicals and GMOs threatened their livelihoods from forest environments, especially to produce organic and export-grade honey. Clearly, considering recent deforestation trends and socio-environmental conflicts, the integration of conservation and emission reductions goals with sustainable rural development in this region presents a major challenge for local, national and international stakeholders involved.

deforestation studies that associate poverty and marginalization with tropical deforestation (Geist and Lambin, 2002; Kissinger et al., 2012). In Hopelchen, financial credits and profits to invest in agricultural expansion are mostly obtained by the commercial farmers with mechanized agriculture, who comprise a large portion of Mennonites with private property. PROCAMPO agricultural subsidies, provided to both farmers in ejido and private property, were not found to influence deforestation in Hopelchen, opposite of the case in the neighboring municipality of Calakmul where ejidos with PROCAMPO were deforesting during the 1990s and early 2000s (Klepeis and Vance, 2003) and indicating how deforestation drivers may vary in the municipalities of the region. Strategic plans, rural development and policy aimed at reducing deforestation in agricultural development frontier zones like Hopelchen must contemplate incentives to intensify and improve production sustainably, in tandem with disincentives to clear more forest for production. The latter is a challenge given the conflicting policy environment promoting growth and development of the agroindustrial model from the agricultural government sector, at the same time, promoting forest and biodiversity conservation (Gómez-González, 2016). This challenge is being addressed now with the Mexicós implementation of its national REDD+ strategy, as well as through formal institutional alignment between federal agricultural and natural resource departments (e.g. SAGARPA, CONAFOR and SEMARNAT), which includes jointly implementing conservation agriculture, soil conservation, agroforestry, sustainable forest management, and improved, diversified production strategies, with a zero net deforestation goal in mind (CONAFOR, 2014). Our study also fills a knowledge gap by providing more clarity on the role of land tenure as a deforestation driver in the Central Yucatan Peninsula. During the most recent analysis period of our study (2005–2015), land tenure regime was a significant driver of forest cover loss in Hopelchen, with a higher probability of recent forest loss in private property and federal land than in ejido communal forests. Also, forest management was inversely related with deforested areas in 2015, although PES did not reflect as a driver of forest cover maintenance. Greater conservation of forest land and lower deforestation rates in ejidos may be a result of government incentives and programs such as community forest management and PES, however, the reformed Agrarian Law, which still mandates ejido governance to maintain forestland under common property and agricultural and urban zones separately (which ejidatarios may legally parcelize and privatize), may also play an important role. Our study demonstrates that forest cover loss rates are significantly higher within private property and federal land compared to ejido property with communal forest ownership. These trends are significant only after 1995, which coincides with the period of neoliberal economic reforms and the 1992 Agrarian Law Reform which promoted the above-mentioned parcelization and privatization of ejido land (DiGiano et al., 2013). However, in our study, we find greater forest cover loss rates in federal land during the 1995–2005 period, more likely reflecting the surge of federal lands being occupied for production purposes and the formal and informal privatization processes that have occurred supporting the arrival of Mennonites and other immigrants. As noted above, conflicting results have been reported on deforestation within different land tenure regimes in Mexico at national and regional scales (Deininger and Minten, 1999; Barsimantov and Kendall, 2012; Bonilla-Moheno et al., 2013). While some authors have observed less deforestation in ejido property compared to private property at a regional scale (Barsimantov and Kendall, 2012), others show the opposite trend for Mexico in general (Bonilla-Moheno et al., 2013), underlining the importance of scale when analyzing drivers and land tenure for local or regional level conservation and development strategies aimed at reducing deforestation. In the neighboring municipality of Calakmul, which is mostly occupied by CBR and ejido properties, there is little mechanized agriculture present, and deforestation problems are

5. Conclusions This research shows how changes in land tenure have facilitated agricultural expansion and associated deforestation in the study area. In this study, much of the recent deforestation occurring in federal lands and private property is a result of both legal and illegal land acquisition. The trend of federal lands privatization in Hopelchen has been increasing in the last decade, in conjunction with Mennonite settlement. The influence of land tenure and Mennonite agricultural expansion on forest cover loss provide important insight into developing municipal land use plans and conservation strategies to reduce deforestation. Currently the municipality has developed an Ecological Land Use Zoning Plan (Programa de Ordenamiento Ecológico Local) as part of their strategy to halt deforestation and promote sustainable agricultural development. The results of this study send a clear message for the purposes of developing effective conservation and development strategies, in addition to land use regulations to decrease deforestation in the municipality and halt the growth of the agroindustrial model; government and non-government organizations will need to support and promote the maintenance of ejido community property, particularly communal forest areas managed forfor timber or apiculture. In 482

Land Use Policy 69 (2017) 474–484

E.A. Ellis et al.

2010. Cantún Caamal, M., Pat Fernández, J.M., 2011. La reforma agraria en Campeche, ¿permanencia de una cultura indigena? Secuencia 82, 103–126. Cattaneo, A., 2001. Deforestation in the Brazilian amazon: comparing the impacts of macroeconomic shocks, land tenure, and technological change. Land Econ. 77 (2), 219–240. Challenger, A., Soberón, J., 2008. Los ecosistemas terrestres, en Capital natural de México. Conocimiento actual de la biodiversidad, vol I. Conabio, México, pp. 87–108. Dalle, S.P., De Blois, S., Caballero, J., Johns, T., 2006. Integrating analyses of local landuse regulations, cultural perceptions and land-use/land cover data for assessing the success of community-based conservation. For. Ecol. Manage. 222, 370–383. Danga Pelissier, T., 2015. Agriculturas mayas y menonitas en Hopelchén (Campeche, Península Yucatán, México. Tesis para la obtención del título de Ingeniero Especialidad Agrónomo Desarrollo agrícola y rural en países tropicales Enfoque Recursos, sistemas agrarios y desarrollo. Montpellier SupAgro, Institut des régiosn chaudes. Danielsen, F., Skutsch, M., Burgess, N.D., et al., 2011. At the heart of REDD+: a role for local people in monitoring forests? Conserv. Lett. 4, 158–167. Deininger, K., Minten, B., 1999. Poverty, policies, and deforestation: the case of Mexico. Econ. Dev. Cult. Change 47 (2), 313–344. DiGiano, M., Ellis, E., Keys, E., 2013. Changing landscapes for Forest Commons: linking land tenure with Forest Cover change following Mexico’s 1992 agrarian counter-reforms. Hum. Ecol. 41 (5), 707–723. http://dx.doi.org/10.1007/s10745-013-9581-0. Dolisca, F., McDaniel, J., Teeter, L., Jolly, C., 2007. Land tenure, population pressure, and deforestation in Haiti: the case of Forêt des Pins Reserve. J. For. Econ. 13 (4), 277–289. Doherty, E., Heike, S., 2011. Forest Tenure and Multi-Level Governance in Avoiding Deforestation under REDD. Global Environ. Politics 11 (4), 66–88. Ellis, E.A., Porter-Bolland, L., 2008. Is community-based forest management more effective than protected areas? A comparison of land use/land cover change in two neighboring study areas of the Central Yucatan Peninsula, Mexico. For. Ecol. Manage. 256, 1971–1983. Ellis, E.A., Romero Montero, A., Hernández Gómez, I.U., 2015. Evaluación y mapeo de los determinantes de deforestación en la Península Yucatán. Agencia de los Estados Unidos para el Desarrollo Internacional (USAID), The Nature Conservancy (TNC), Alianza México REDD+, México, Distrito Federal. Ellis, E.A., Hernandez Gomez, I.U., Romero-Montero, A., 2017. Los procesos y causas del cambio en la cobertura forestal de la Península Yucatán, México. Ecosistemas 26 (1), 101–111. Esparza-Olguín, L.G., Martínez Romero, E., 2011. Deforestación en Campeche: Causas y Efectos. In: Revista Fomix Campeche 3(10): 6–11FAO. (2015). Global Forest Resources Assessment 2010. FAO Forestry Paper 163. Food and Agriculture Organization. Rome, Italy. Fearnside, P., 2001. Land-tenure issues as factors in environmental destruction in brazilian amazonia: the case of southern Pará. World Dev. 29 (8), 1361–1372. Fitzgerald, Lee A., Stronza, Amanda L., 2009. Applied biodiversity science: bridging ecology, culture, and governance for effective conservation. Interciencia 34 (8), 563–570. Food and Agriculture Organization (FAO), 2015. Global Forest Resource Assessment 2015: How are the World’s Forest Changing? Rome: Food and Agriculture Organization of the United Nations, Rome, Italy. Gómez-González, I., 2016. A honey-sealed alliance: mayan Beekeepers in the Yucatan peninsula versus transgenic soybeans in Mexico’s last tropical forest. J. Agrar. Change 16 (4), 728–736. http://dx.doi.org/10.1111/joac.12160. Geist, H.J., Lambin, E.F., 2002. Proximate causes and underlying driving forces of tropical deforestation. Bioscience 52 (2), 143. http://dx.doi.org/10.1641/0006-3568(2002) 052[0143:pcaudf]2.0.co;2. Hosonuma, N., Herold, M., De Sy, V., De Fries, R.S., Brockhaus, M., Verchot, L., et al., 2012. An assessment of deforestation and forest degradation drivers in developing countries. Environ. Res. Lett. 7 (4), 044009. http://dx.doi.org/10.1088/1748-9326/ 7/4/044009. Instituto Nacional de Estadística, Geografía e Informática (INEGI), 2007. Conjunto de Datos Vectorial Edafológico, Escala 1:250 000 Serie II (Continuo Nacional). Formato digital. Instituto Nacional de Estadística y Geografía e Informática (INEGI), 2009. Prontuario de Información Geográfica Municipal de los Estados Unidos Mexicanos. (Hopelchén, Campeche. Clave geoestadística 04006. http://www3. inegi.org.mx/sistemas/ mexicocifras/datos-geograficos/04/04006.pdf). Instituto Nacional de Estadística y Geografía, 2011. Censo de Población y Vivienda 2010. Principales resultados por localidad (ITER). Instituto Nacional de Estadística y Geografía.- México. Instituto Nacional de Estadística y Geografía e Informática (INEGI), 2013. Continuo Nacional de Elevaciones Mexicano versión 3.0 (CEM 3.0). [Modelo Digital de Elevación]. Escala 1:50,000. http://www.inegi.org.mx/geo/contenidos/ datosrelieve/continental/continuoelevaciones.aspx. (3 de Marzo 2016). Instituto Nacional de Estadística y Geografía, 2015a. Anuario estadístico y geográfico de Campeche 2015/Instituto Nacional de Estadística y Geografía.- México. INEGI (c2015. 382 p.). Instituto Nacional de Estadística y Geografía, 2015b. Conjunto de datos vectoriales de información topográfica escala 1:50 000 serie III. E16A22 (Iturbide). Formato digital. Instituto Nacional de Estadística y Geografía, México. Killeen, T.J., Guerra, A., Calzada, M., Correa, L., Calderon, V., Soria, L., et al., 2008. Total historical land-use change in eastern Bolivia: who, where, when, and how much? Ecol. Soc. 13 (1). http://dx.doi.org/10.5751/es-02453-130136. Kissinger, G., Herold, M., DeSy, V., 2012. Drivers of Deforestation and Forest Degradation: A Synthesis Report for REDD+ Policymakers. Lexeme Consulting,

addition, greater investment and programs aimed at improving timber and honey production are sorely needed. Furthermore, research and policy incentives should be aimed at intensifying production in the current agriculture footprint, while prioritizing protection and restoration of degraded forest fragments. Finally, programs, incentives and policies directed towards forest conservation in the region will need to work with commercial farmers and Mennonite communities if deforestation is to be addressed. Currently, The Nature Conservancy implements projects through the Yucatan Peninsula Climate Action Fund promoting land use zoning, conservation agriculture, sustainable intensification and silvopastoral systems with Mennonite farmers (TNC, 2017). Moreover, a workshop and consultation with Mennonite communities while developing the municipal land use plan in February 2017 clarified their interest in conservation, as demonstrated by a Mennonite settlement in the municipality which receives profits from hunting leases in their conservation area. Even though participation of most Mennonite settlements in conservation efforts and other municipal affairs is still poor, programs which provide economic incentives or profits from conservation areas, such as forest management, PES, apiculture and nontimber forest production (e.g. hunting and ornamentals) are more likely to be accepted. Acknowledgements Funding for this research was made possible by the W.K. Kellogg Foundation and by the United States Agency for International Development (USAID), under the terms of the Cooperation Agreement No. AID-523-A-11-00001 (Proyecto de Reducción de Emisiones por la Deforestación y la Degradación de Bosques de México) implemented by the main awardee The Nature Conservancy (TNC) and its partners (Rainforest Alliance, Woods Hole Research Center and Espacios Naturales y Desarrollo Sustentable). The authors received no financial support for the authorship, and/or publication of this article. We would like to acknowledge Sebastién Proust, TNC Regional Director for the Yucatán Península and José Trinidad Montoy, Sub-Director of Ecology of the Municipaligy of Hopelchén for supporting field visits and facilitating meetings and access to information for this study. References Acopa, D., Boege, E., 1998. The Maya forest in Campeche, Mexico: experiences in forest management at Calakmul. In: Primack, R.B., Bray, D., Galletti, H.A., Ponciano, I. (Eds.), Timber, Tourists, and Temples. Island Press, Washington, DC, pp. 81–97. Alix-Garcia, J., 2007. A spatial analysis of common property deforestation. J. Environ. Econ. Manage. 53 (2), 141–157. http://dx.doi.org/10.1016/j.jeem.2006.09.004. Angelsen, A., Brockhaus, M., Sunderlin, W.D., Verchot, L.V., 2012. Analysing REDD+: Challenges and Choices. CIFOR, Bogor, Indonesia. Barnes, G., 2009. The evolution and resilience of community-based land tenure in rural Mexico. Land Use Policy 26 (2), 393–400. Barsimantov, J., Kendall, J., 2012. Community forestry, common property, and deforestation in eight mexican states. J. Environ. Dev. 21 (4), 414–437. Bezaury-Creel, J.E., Torres, J. Fco., Proust, Sébastien, Paiz-Merino, Y., Canto-Vergara, J.M., 2014. Área De Atención Prioritaria Alianza México REDD+, Puuc-Chenes, Yucatán Y Campeche. Alianza México REDD+. Formato Cartel. (Referencias completas en: www.researchgate.net/profile/Juan_Bezaury-Creel/?ev=hdr_xprf). Bonilla-Moheno, M., Redo, D., Aide, T., Clark, M., Grau, H., 2013. Vegetation change and land tenure in Mexico: a country-wide analysis. Land Use Policy 30 (1), 355–364. Bray, D.B., Ellis, E.A., Armijo-Canto, N., Beck, C.T., 2004. The institutional drivers of sustainable landscapes: a case study of the Mayan zone in Quintana Roo, Mexico. Land Use Policy 21, 333–346. Céspedes-Flores, S., Moreno-Sánchez, E., 2010. Estimación del valor de la pérdida de recurso forestal y su relación con la reforestación en las entidades federativas de México. Investigación Ambiental 2 (2), 5–13. Comisión Nacional Forestal (CONAFOR), 2010. Visión de México Sobre REDD+: Hacia una Estrategia Nacional. CONAFOR, Zapopan, Mexico, pp. 57. Comisión Nacional Forestal (CONAFOR), 2014. Estrategia Nacional para REDD+ (ENAREDD +). Documento para consulta pública, Nov. 2014. CONAFOR, Zapopan, Mexico, pp. 107. Comisión Nacional Forestal (CONAFOR), 2016. Polígonos de PSA Nacional, PSA ATREDD +, PSA-FBP y MLPSA. Guadalajara, Jal.: Coordinación General de Producción y Productividad de la Comisión Nacional Forestal. Consejo Nacional de Población (CONAPO), 2012. Índice de marginación por localidad

483

Land Use Policy 69 (2017) 474–484

E.A. Ellis et al.

Rendon-von Osten, J., Dzul-Caamal, R., 2017. Glyphosate residues in groundwater, drinking water and urine of subsistence farmers from intensive agriculture localities: a survey in hopelchén, campeche, Mexico. Int. J. Environ. Res. Public Health 14 (6), 595. Romero-Montero, J.A., Ellis, E.A., 2016. Selva baja subperennifolia en el sureste de México. Rev. Int. Des. Reg. Sust. 1 (2), 48–56. Romero-Montero, J.A., 2014. Evaluación de los factores ambientales, socioeconómicos e institucionales que intervienen la dinámica del cambio de cobertura forestal en ejidos de Campeche y Quintana Roo, México. Universidad Veracruzana (Tesis de Maestría en Ecología 115 p). SAGARPA, 2016. Padrones de Beneficiarios. (online] Available at: http://www.sagarpa. gob.mx/agricultura/Paginas/Padrones.aspx [Accessed 3 September, 2016). Secretaría de Medio Ambiente y Recursos Naturales (SEMARNAT), 2015. Informe 2015 Aprovechamiento Forestal Maderable; Internal Report. SEMARNAT, Campeche, Mexico. Salvini, G., Herold, M., De Sy, V., Kissinger, G., Brockhaus, M., Skutsch, M., 2014. How countries link REDD+ interventions to drivers in their readiness plans: implications for monitoring systems. Environ. Res. Lett. 9, 74004. ́ nna, A.A., Young, C.E.F., 2014. Property Rights, Deforestation and Violence: SantA Problems for the Development of the Amazon. pp. 28–30 (Policy in Focus, August 2014, No. 29). Schüren, U., 2003. Reconceptualizing the post-peasantry: household strategies in mexican ejidos. Rev. Eur. Estud. Latinoam. Caribe 75, 47–63. Skole, D.L., Chomentowski, W.H., Salas, W.A., Nobre, A.D., 1994. Physical and human dimensions of deforestation in Amazonia. Bioscience 44 (5), 314–322. http://dx.doi. org/10.2307/1312381. Steininger, M.K., Tucker, C.J., Ersts, P., Killeen, T.J., Villegas, Z., Hecht, S.B., 2001. Clearance and fragmentation of tropical Deciduous forest in the Tierras Bajas, Santa Cruz, Bolivia. Conserv. Biol. 15 (4), 856–866. http://dx.doi.org/10.1046/j.15231739.2001.015004856.x. The Nature Conservany (TNC), 2017. Maya Forest. (online] Available at: http://www. nature.org/ourinitiatives/regions/northamerica/mexico/placesweprotect/mayaforest.xml [Accessed 10 March, 2017). Tejada, G., Dalla-Nora, E., Cordoba, D., Lafortezza, R., Ovando, A., Assis, T., Aguiar, A.P., 2016. Deforestation scenarios for the Bolivian lowlands. Environ. Res. 144, 49–63. http://dx.doi.org/10.1016/j.envres.2015.10.010. Trapasso, L.M., 1994. Indigenous attitudes, ecotourism, and Mennonites: recent examples in rainforest destruction/preservation. GeoJournal 33, 449. http://dx.doi.org/10. 1007/BF00806428. Tyukavina, A., Baccini, A., Hansen, M.C., Potapov, P.V., Stehman, S.V., Houghton, R.A., Krylov, A.M., Turubaniva, S., Goetz, S.J., 2015. Aboveground carbon loss in natural and managed tropical forests from 2000 to 2012. Environ. Res. Lett. 10, 1–14. UN-REDD Programme, 2010. The UN-REDD Programme Strategy 2011–2015. FAO, UNDP, UNEP, Geneva, Switzerland. Velázquez, A., Mas, J.F., Díaz, J.R., Mayorga-Saucedo, R., Alcántara, P.C., Castro, R., Fernández, T., Bocco, G., Escurra, E., Palacio, J.L., 2002. Patrones y tasas de cambio de uso del suelo en Máxico. Gaceta Ecol. 62, 21–37. Villanueva-Gutiérrez, R., Echazarreta-González, C., Roubik, D., Moguel-Ordóñez, Y., 2014. Transgenic soybean pollen (Glycine max L.) in honey from the Yucatán peninsula, Mexico. Sci. Rep. 4 (1).

Vancouver Canada. Klepeis, P., Vance, C., 2003. Neoliberal policy and deforestation in southeastern Mexico: an assessment of the PROCAMPO Program. Econ. Geogr. 79 (3), 221–328. Laurance, W.F., 2015. Emerging threats to tropical forests. Ann. Missouri Bot. Garden 100 (3), 159–169. Madrid, L., Núñez, J.M., Quiroz, G., Rodríguez, Y., 2009. La propiedad social forestal en México. Investig. Ambient. 2, 179–196. Martínez-Romero, E., Esparza, O., 2010. Estudio de caso: deforestación en el estado de Campeche. Causas directas e indirectas de la principal amenaza sobre la biodiversidad. In: Villalobos-Zapata, G.J., Mendoza Vega, J. (Eds.), 2010. La Biodiversidad en Campeche: Estudio de Estado. Comisión Nacional para el Conocimiento y Uso de la Biodiversidad (conabio). Gobierno del Estado de Campeche, Universidad Autónoma de Campeche, El Colegio de la Frontera Sur. México, pp. 730. Mas, J.-F., Pérez-Vega, A., Ghilardi, A., Martínez, S., Loya-Carrillo, J.O., Vega, E., 2014. A suite of tools for assessing thematic map accuracy. Geogr. J. 2014, 1–10. http://dx. doi.org/10.1155/2014/372349. Mereles, M.F., Rodas, O., 2014. Assessment of rates of deforestation classes in the Paraguayan Chaco (great south American Chaco) with comments on the vulnerability of forests fragments to climate change. Clim. Change 127 (1), 55–71. http://dx.doi. org/10.1007/s10584-014-1256-3. Messina, J., Walsh, S., Mena, C., Delamater, P., 2006. Land tenure and deforestation patterns in the Ecuadorian Amazon: conflicts in land conservation in frontier settings. Appl. Geogr. 26 (2), 113–128. Morales-Barquero, L., Borrego, A., Skutsch, M., Kleinn, C., Healey, J., 2015. Identification and quantification of drivers of forest degradation in tropical dry forests: a case study in Western Mexico. Land Use Policy 49, 296–309. Olander, L.P., Gibbs, H.K., Steininger, M., Swenson, J.J., Murray, B.C., 2008. Reference scenarios for deforestation and forest degradation in support of REDD: A review of data and methods. Environ. Res. Lett. 3 (2), 025011. http://dx.doi.org/10.1088/ 1748-9326/3/2/025011. Olofsson, P., Foody, G.M., Herold, M., Stehman, S.V., Woodcock, C.E., Wulder, M.A., 2014. Good practices for estimating area and assessing accuracy of land change. Remote Sens. Environ. 148, 42–57. http://dx.doi.org/10.1016/j.rse.2014.02.015. Palacio-Prieto, J.L., Sánchez-Salazar, M.T., Casado Izquierdo, J.M., Propin Frejomil, E., Delgado Campos, J., Veĺasquez Montes, A., Chias Becerril, L., Ortiz Álvarez, M.L., González Sánchez, J., Negrete Fernández, G., Gabriel Morales, J., Márquez Huitzil, R., 2004. Indicadores para la caracterización y ordenamiento del territorio. SEMARNATINE-UNAM. Patterson, C., 2014. Deforestation, Agricultural Intensification, and Farm Resilience in Northern Belize: 1980–2010. A Thesis Submitted for the Degree of Doctor of Philosophy at the University of Otago. Dunedin New Zealand. Porter-Bolland, L., Ellis, E.A., Gholz, H.L., 2007. Land use dynamics and landscape history in la Montaña, Campeche, Mexico. Landscape Urban Plann. 82 (4), 198–207. http:// dx.doi.org/10.1016/j.landurbplan.2007.02.008. Porter-Bolland, L., Sánchez-González, M.C., Ellis, E.A., 2008. La conformación del paisaje y el aprovechamiento de los recursos naturales por las comunidades mayas de La Montaña, Hopelchén, Campeche. Investigaciones Geográficas 66, 65–80. Registro Agrario Nacional (RAN), 2016. Catálogo de Núcleos Agrarios. (catalogo.datos.gob.mx. [online] Available at: https://datos.gob.mx/busca/dataset/catalogo-denucleos-agrarios [Accessed 3 March, 2016]).

484