Beyond deforestation: Land cover transitions in Mexico
Introduction
Conversion of land cover is one of the main causes of global environmental change. Of all land conversions, tropical deforestation has received the majority of attention as this process jeopardizes the most species rich habitats on the planet and the global carbon cycle (Le Quéré et al., 2016). While it is crucial to monitor the spatial extent of deforestation, most studies have focused on the forest/no-forest dichotomy, limiting a comprehensive assessment of these complex dynamics. For example, a very basic concern is identifying what land covers replace forests. Furthermore, the no-forest class is diverse, and the frequency, time span, and sequence of land cover transitions are important factors that affect biodiversity and environmental services (Foley et al., 2005; Watson et al., 2014). In fact, land use science research has highlighted the need to identify and monitor a wider range of land classes at multiple scales to understand the transitions between natural and non-natural systems (Boillat et al., 2017; Watson et al., 2014).
The availability of frequent and fine resolution remote sensing images has facilitated the accurate identification of the extent and causes of land transformation (Song et al., 2018; Hansen et al., 2013; Feng et al., 2016). A recent study documented a net increase of tree cover at the global scale, but deforestation continues to be the dominant transition in tropical regions (Song et al., 2018). These global scale studies are invaluable, but the forest/no-forest dichotomy misses details that have important implications for biodiversity, food and commodity supply at the country and regional scales. Furthermore, spatially explicit information on the distribution of different covers classes at large scales is useful for identifying the hotspots of change because these can have important ecological, social, and economic implications.
When studies include various cover classes at broader scales and analyze time series, we are in a better position to understand the patterns and pathways of land use transitions (Rounsevell et al., 2012). But, to quantitatively describe the causal effects and mechanisms and how they related to specific drivers or policy is a challenge (Meyfroidt, 2016). This can be particularly difficult for land use studies in regions with diverse habitats where spatial determinants and local context may modify drivers of land change. Mexico is an excellent example given that it has eight biomes, with several terrestrial ecoregions, and land use patterns vary greatly among them. For example, between 2001 and 2010 woody vegetation recovery or shrubland encroachment was the dominant dynamic in coniferous forest, dry forest and desert biomes, while deforestation dominated the moist forest biome (Bonilla-Moheno et al., 2012a). However, even in habitats that had a similar outcome (e.g. an increase in woody vegetation), the proximate factors associated with this change varied from elevation and population density in the dry forest biome, to increase precipitation and land abandonment in the desert biome.
In the last decades Mexico has enacted land use policies, which have had broad impacts across the country. On the one hand, the National Forestry Commission has subsidized the reforestation and protection of forest ecosystems, as well as an increase in timber production (CNF, 2001), specifically through the National Forestry and ProArbol programs (CONAFOR, 2014). The impact of these programs varies greatly across Mexico. For example, more than half of the national timber production is located in the northern states of Durango, Chihuahua, and Michoacan. While there has been a proliferation of commercial tree plantations, mainly conifer species, in the southern states of Chiapas, Tabasco Veracruz, and Campeche (SEMARNAT, 2014; CNF, 2018). On the other hand, the National Agricultural Plan for 2017–2030 largely focuses on increasing productivity and improving access to markets, especially international, and has promoted the sectorization and intensification of agriculture. By identifying 38 strategic crops, which represented approximately 70 % of the national harvested area in 2016, and the potential regions where these crops could expand (SAGARPA, 2017), this program will likely increase the area of mechanized agriculture, particularly in southern states. Additionally, indirect land use change drivers, such as NAFTA (Galvan-Miyoshi et al., 2015), land characteristics and climate change (López-Blanco et al., 2018; Mendoza-Ponce et al., 2018; Monterroso-Rivas et al., 2018; Ureta et al., 2012), diet preferences (Ibarrola-Rivas and Granados-Ramirez, 2017), or resilience of certain agricultural sectors (i.e., maize system, Sweeney et al., 2013; Eakin et al., 2018) have influenced the national land use dynamics differentially across regions. As a result, while in some regions land protection or abandonment has facilitated the recovery of secondary forests, in others, agricultural intensification and land use policies have encouraged forest loss or shifts from pastures to crops.
The major objectives of this study are to identify the regions in Mexico where there have been sustained trends of expansion or reduction of pasture and croplands and determine the agricultural products likely associated with these changes. We expect that these dynamics will vary greatly given the high diversity of environmental conditions across Mexico (e.g. seven biomes and 40 major terrestrial ecoregions), and thus a better understanding of these changes and where they are occurring will be useful for allocating resources for regional land use planning and conservation efforts. To understand these dynamics, we analyzed the changes in woody vegetation, grasslands (i.e. pastures), and croplands classes at the ecoregions scale from 2001 to 2014 using annual land use/land cover maps based on MODIS 250 m imagery. Where there were extensive area of crop loss or gain, we use agricultural census data to determine the identity of the major crops. If the dynamics in Mexico are similar to overall land-use patterns described for all Latin America (Graesser et al., 2015), we expect that increases in pasture will come at the expense of forest, whereas increases in crops will mainly replace pasturelands.
Section snippets
Methods
Land cover classification followed Graesser et al. (2015) methodology. We created annual land classifications using MODIS MOD13Q1 (16-day L3 Global 250 m) imagery (pixel 250 m), which includes 23 samples per year. For each pixel, we calculated the statistics mean, standard deviation, minimum, maximum, and range for EVI, and red, NIR, and MIR reflectance values from calendar years 2001–2014. These statistics were calculated for all years and these data were associated with reference data from 2005
Results
At the country scale, between 2001 and 2014 there were dramatic changes in the extent and spatial distribution of the three main land cover classes (Fig. 1). The woody (i.e. trees and shrubs) vegetation class increased by more than 2.8 million ha, croplands increased by 291,000 ha and the pastureland class declined by approximately 1.5 million ha. The remaining area (∼1.5 million ha) not explained by these conversions was mostly due to the transitioning of areas classified as bare ground into
Discussion
Between 2001 and 2014, the distribution of land cover classes has changed dramatically across Mexico. At the country scale, the woody vegetation and croplands class increased while pastures decreased; however, contrary to our original expectations, cropland expansion was not a proximate driver of pastures reduction. In fact, the extent and the spatial aggregation of major changes in crop and pasture cover varied greatly across ecoregions, and hotspots of crop expansion did not correspond with
Conclusion
As an ongoing debate on how to effectively conserve biodiversity and environmental services, while obtaining goods from nature is taking place (e.g., Phalan et al., 2011; Grau et al., 2013; Ceddia et al., 2014; Mehrabi et al., 2018), one component that all would agree on is the urgent need for accurate information on land dynamics to inform planning and management of sustainable land systems.
Here we have shown how a diversity of drivers and conditions produce contrasting land use dynamics
Declaration of Competing Interest
We declare no conflict of interest.
Acknowledgements
We thank Jordan Graesser for helping with data processing. This research was supported by the National Science Foundation (Dynamics of Coupled Natural and Human Systems CNH grant; award # 0709598). We thank two anonymous reviewers for their comments, which helped improve the manuscript.
References (65)
- et al.
Land system science in Latin America: challenges and perspectives
Curr. Opin. Environ. Sustain.
(2017) - et al.
Vegetation change and land tenure in Mexico: a country-wide analysis
Land Use Policy
(2013) - et al.
Woody vegetation expansion in a desert grassland: Prehistoric human impact?
J. Arid Environ.
(2007) - et al.
Land use biodiversity impacts embodied in international food trade
Glob. Environ. Change
(2016) Earth science data records of global forest cover and change: assessment of accuracy in 1990, 2000, and 2005 epochs
Remote Sens. Environ.
(2016)- et al.
Beyond “land sparing versus land sharing”: environmental heterogeneity, globalization and the balance between agricultural production and nature conservation
Curr. Opin. Environ. Sustain.
(2013) - et al.
Diversity of Mexican diets and agricultural systems and their impact on the land requirements for food
Land Use Policy
(2017) - et al.
The decline of the itinerant milpa and the maintenance of traditional agrobiodiversity: crops and weeds coexistence in a tropical cloud forest area in Oaxaca, Mexico
Agric. Ecosyst. Environ.
(2016) - et al.
Challenges for land system science
Land Use Policy
(2012) - et al.
Mexican maize production: evolving organizational and spatial structures since 1980
Appl. Geogr.
(2013)