Revista Mexicana de Ciencias Forestales Vol. 9 (46)

Marzo-Abril (2018)

Fecha de recepción/Reception date: 22 de mayo de 2017

Fecha de aceptación/Acceptance date: 21 de febrero de 2018.

_______________________________

1Programa de Investigación en Cambio Climático. Universidad Nacional Autónoma de México. México.

2Instituto Tecnológico y de Estudios Superiores de Monterrey. México.

3Facultad de Ciencias, UNAM. México.

4Servicios ambientales y cambio climático SACC, AC. México.

5Centro Nacional de Investigación Disciplinaria en Conservación y Mejoramiento de Ecosistemas Forestales. INIFAP. México.

6Centro Regional de Investigación Multidisciplinaria, UNAM. México.

*Autor por correspondencia, correo-e: jabordonez@hotmail.com

Resumen

Los estudios relacionados con los cambios progresivos en la cobertura vegetal y de uso de suelo han cobrado importancia en la investigación ambiental, ya que permiten evaluar las tendencias espacio-temporales de procesos como la deforestación y degradación ambiental, provocadas por actividades humanas. En el presente estudio, se hace una descripción de los cambios de cobertura y uso del suelo en el estado de Morelos durante el periodo 2000-2009. Se identifican 14 clases de coberturas vegetales y usos del suelo, que incluyen cuerpos de agua, los cuales no se consideran en el análisis y se registran dos coberturas que ganan terreno: agricultura, con un incremento de 1 373 ha y urbano, con 189 ha; las que más se transforman son selva baja (1 841 ha) y los bosques, con 858 ha. La dinámica del cambio es compleja y se identifican factores socioambientales y de tenencia de la tierra como las causas principales. Datos satelitales multiespectrales de LANDSAT se usaron para mapear y analizar los cambios de uso de suelo producidos en los años 2000 y 2009, mediante el software ESRI ARCMAP 10. Se empleó la técnica de clasificación supervisada, basada en el algoritmo de máxima verosimilitud. La información sobre el uso de cobertura vegetal y uso del suelo así como la identificación de las posibilidades para su uso óptimo, obtenida con percepción remota y sistemas de información geográfica, es esencial para la selección, planificación y ejecución de planes de ordenamiento territorial, desarrollo, conservación y aprovechamiento de recursos naturales, con el propósito de satisfacer la creciente demanda de las necesidades básicas del hombre, así como su bienestar.

Palabras clave: Cobertura vegetal, ordenamiento territorial, percepción remota, Sistemas de información geográfica, tenencia de la tierra, uso de suelo.

Abstract:

Studies related to the progressive changes in vegetation cover and land-use are an important component in environmental research, since they allow the evaluation of the spatio-temporal trends of such processes as deforestation and environmental degradation, caused by human activities. In the present study, land-cover and land-use change in the State of Morelos during the period 2000-2009, are described. Fourteen types of vegetation cover and land-use, including water bodies, were identified, all though the latter were no included in the analysis. Agricultural and urban land cover increased 1 373 ha and 189 ha, respectively; whereas the land-cover that undergo the greatest transformations are the forests, with 858 ha, and tropical dry forests, with 1 841 ha. The dynamics of change are complex, its main drivers being socio-environmental factors and land tenure. LANDSAT multispectral satellite data were used, with ESRI ARCMAP 10 software, to map and analyze the land-use change between the years 2000 and 2009. The supervised classification technique was used, based on the maximum likelihood algorithm. Information on the vegetation cover and land-use, and the identification of alternatives for their optimal use, obtained through the use of remote sensing and geographic information systems, are essential for the selection, planning and execution of land management plans, as well as for the development, conservation and use of natural resources required to satisfy the growing demand for basic needs and for human welfare.

Key words: Vegetation cover, land use planning, remote sensing, geographic information systems, land tenure, land use.

Introduction

Land cover and use

Land is one of the most important natural resources, on which life and various development activities depend (George et al., 2016). The changes in land-cover and land-use are the result of socio-economic activities and environmental factors (Overmars et al., 2005).

Human activities have a significant impact on biodiversity worldwide, on global and local climate, biogeochemical cycles, on soil quality and degradation, hydrology, and finally on food safety and human well-being (Trimble et al., 2000; Foody, 2002); furthermore, they affect the ability of biological systems to meet the increasing demand for natural resources (Tran et al., 2015). However, the effects of land-use change -are not all negative, as some are associated with an increase in food and fiber yield, having a positive impact on human health, wealth and well-being (Lambin and Geist 2006). The documentation of the land-cover and land-use changes are necessary in order to understand the underlying driving forces, their consequences, especially in the context of global climate change, rapid population growth and the ever increasing demands for environmental sustainability.

The use of Landsat Thematic Mapper (TM) images in research on land-use change in Mexico is very robust and frequent. Studies have been carried out in central Chihuahua (Alatorre et al., 2015), Hidalgo (Sierra-Soler et al., 2015), Jalisco (Morales-Barquero et al., 2015), Michoacán (Ordóñez et al., 2008; Mas and González, 2015), Yucatán (Colditz et al., 2015), the State of Mexico (Galicia and García-Romero, 2007; Zepeda et al., 2012), Morelos (Trejo and Hernández, 1996; Vega et al., 2008), Chiapas (Flamenco-Sandoval et al., 2007), and Tabasco (Guerra and Ochoa, 2006), among others.

The objectives of this research are: to develop land-cover and land-use maps in order to analyze the changes detected over a period of 9 years in Morelos, Mexico, using Landsat TM images of the years 2000 and 2009, and to analyze the changes in the vegetation cover and land-use in order to identify the areas with the highest impact.

Materials and Methods

Study area

The State of Morelos is located between the parallels 18o20´ and 19o07´ N and the meridians 98o37´ and 99o30´ W (Figure 1). It has a surface of 4 889 m2 and at the altitude ranges from 700 to over 5 000 m. The climate varies from warm subhumid to semi-cold (García, 1988). The State of Morelos contributes a total of 3 661 plant species, which renders it the sixth state with the greatest species richness, after Veracruz, Michoacán, Chiapas, Puebla and Oaxaca (Programa de Ordenamiento Ecológico Regional del Estado de Morelos, 2013).

Ubicación de Morelos en México = Location of Morelos in Mexico.

Figure 1. Location of the study area in the Mexican Republic.

Morelos has a significant amount of forests and rainforests that are increasingly endangered. In Mexico, the vegetation-type most affected by agricultural activities are the temperate forests and the low deciduous forests (Challenger, 1998; Cervantes et al., 2001); corn is the most widespread crop (Challenger, 1998).

The State of Morelos has 1 903 811 inhabitants, who amount to 1.6 % of the overall Mexican population; 84 % of the population is urban, and the remaining 16% is rural. 2.71 % of the economic activities correspond to the primary sector; 31.5 %, to the secondary sector, and 65.71 % to the tertiary sector (INEGI, 2015).

Collection of land-use data and land-cover maps

The LANDSAT satellite data sets with low cloudiness percentages (geocoded with UTM and spheroid projection, and WGS 1984 data, 14 Northern Area) with a 30 m spatial resolution for two different years (February 2000 and 2009) were downloaded from the USGS Earth Explorer website (http: // earthexplorer.usgs.gov/). Images of the years 2000 and 2009 were used because they exhibited the lowest percentages of cloudiness, compared to other images of the decade (2000-2010).

Remote sensing data sets (aerial or satellite images) are often be used to determine land-use through observations of the land-cover (Brown et al., 2000; Karl and Maurer, 2010). Information on land-use based on remote sensing data can be obtained through visual interpretation. However, this is limited to a single band or to a three-band color composition (RGB); likewise, the manual digitalization of land use patches is very tedious and subjective (Bolstad et al., 1990). Therefore, automatic remote-sensing based classification of the land use allotment in large areas is more accurate. ArcGIS Spatial Analyst features a complete set of tools to carry out supervised and unsupervised classification processes.

Methodology

The satellite images were subjected to pre-processing algorithms and standard processing in order to correct geometric, atmospheric and topographic errors. The raw reflectance spectral data registered in the various bands that make up the images, as well as the combinations of different bands (spectral vegetation indices), were drawn from the images and spatially connected with the data of the available biophysical variables (figures 2 and 3).

Figure 2. View in the year 2000.

Figure 3 View in the year 2009.

Sections corresponding to the administrative boundaries of the State of Morelos were obtained for subsequent comparison (figures 4 and 5).

Figure 4. View of 2000 section with administrative boundaries of the State of Morelos.

Figure 5. View of 2009 section with administrative boundaries of the State of Morelos.

Fourteen distinct classes of vegetation cover and land-use were identified (Trejo and Hernández, 1996):

1. Temperate forests including mixed conifer associations (PF) of Pinus spp. and Abies spp.
2. Oak forests, communities made up of Quercus spp. 
3. Low deciduous forests formed by communities of Bursera spp. and Conzattia multiflora (B.L. Rob.) Standl.
4. Degraded forests grouping oak and conifer forests.
5. Rosetophile scrub species combined with Brahea dulcis (Kunth) Mart., Dodonaea viscosa (L.) Jacq.
6. Mesophilic forest: grows at an altitude of 1 800 to 1 950 m; the observed tree species were: Ardisia compressa Kunth, Eugenia crenularis Lundell, Garrya longifolia Rose, Guarea glabra Vahl, Meliosma dentata (Liebm.) Urban, Myrsine juergensenii (Mez) Ricketson & Pipoly and Fraxinus uhdei (Wenz.) Lingelsh.
7. Irrigated agriculture of sugar cane, rice and flower crops.
8. Rain-fed agriculture: subsistence agriculture, mainly corn, beans, squash.
9. Areas without apparent vegetation: areas devoid of vegetal cover.
10. Urban areas or urban cover: human settlements.
11. Riparian vegetation dominated by willows (Salix humboldtiana Willd. y Salix bonplandiana Kunth) and amate trees (Ficus cotinifolia Kunth).
12. Natural pastures: areas covered with grass or with herbaceous vegetation for grazing.
13. Secondary vegetation.
14.  Water bodies: lakes, rivers, dams. Not included in the analysis.

A first map with an unsupervised classification was generated (Richards, 1986) using Arcgis10, from an iso-cluster algorithm of the 2009 image, which served as a basis for the in-field identification of vegetation types and their corresponding land-use. A 500 × 500 grid was superimposed, each line corresponding with UTM coordinates. The field work was carried out during the months of February and March; 93 sites were sampled in order to cover all the land-use categories. Each training site was outlined in the images with polygons encompassing multiple pixels (Tran et al., 2015).

Based on the information gathered in field, the images were reclassified using the maximum likelihood algorithm in Arcgis10. Once the maps of the classified images were obtained, a confusion matrix was built for each map (2000 and 2009) in order to assess the accuracy of the grouping and the fidelity with which the classification represents the in-field observations. Once the confusion matrices were developed, the overall accuracy and the Kappa indices of both images were evaluated.

The image classification process consists in allocating the pixels of a reticulated image to predetermined land-cover classes. The basic approach to image classification is through visual interpretation, taking into account the hue, texture, size, shape and association (Lillesand et al., 2004; Serra et al., 2008; Qasim et al., 2011).

For the present study, the supervised classification utilized a maximum likelihood classification (MLC), which was selected because it has the capacity to incorporate the statistics of the training samples before allocating the land-cover to each pixel. The MLC is a parametric classifier that assumes a normal distribution of the data of the individual classes. It assesses the variance and the covariance of the spectral response patterns when classifying an unknown pixel (Nguyen et al., 2012). The MCL requires sufficient training spectral data samples for each class in order to accurately assess the statistics required by the classification algorithm. The supervised classification was carried out using the spectral signature files for the 13 classes.

Accuracy assessment

The accuracy of a classification process refers to the degree of agreement between the classes of the image and a reference data set (Congalton and Green, 1999). In order to make a quantitative assessment, a diffuse error matrix, ―i.e. a quadratic arrangement of numbers ordered in columns and rows― was constructed (Congalton and Green, 2009; Lunetta and Lyon, 2004). The columns correspond to the reference set, and the rows are the classes generated in the course of classification process. The error matrix was constructed based on the results of the classification, and the reference data were those taken from the in-field sampling.

In addition, the total accuracy of a classification can be assessed through a Kappa agreement index (K), a discrete multivariate technique that statistically determines whether or not an error matrix is significantly different from another. It is a measure of correspondence between the classification and reference data set, in relation to the accuracy of a randomly generated classification (Congalton and Green 1999; Richards and Jia 1999; Congalton, 2004).

Change analysis

A change analysis of the “post-classification” type was performed (Chuvieco, 2000). Also, a change matrix was generated indicating the dynamics of the land-use and the vegetation cover change in the analyzed period. Furthermore, the deforestation rate was estimated for the period of interest through a variation in the formula cited by Palacio-Prieto et al. (2004):

Where:

DR = Annual deforestation rate expressed as a percentage

S2 = Forested area in the final year

S1 = Forested area in the initial year

n = Number of years of the analysis period

Results

The classified images obtained after the pre-processing and the supervised classification show the land-use and the land-cover in Morelos (figures 6 and 7). The 2000 and 2009 images represent 14 classes according to the presence of clouds in the northern area of the state. The work by Trejo and Hernández (1996) was used as a supporting criterion for defining the land use categories.

Cuerpos de agua = Water bodies; Bosque de encino = Oak forest; Bosque degradado = Degraded forests; Suelo urbano = Urban soil; Selva baja caducifolia = Low deciduos forest; Bosque mesófilo = Mesophilic forest; Agricultura de riego = Irrigated agriculture; Vegetación riparia = Riparian vegetation; Pastizal natural = Natural pastures; Vegetación secundaria = Secondary vegetation; Bosque de coníferas = Conifer forest; Sin vegetación aparente = Without apparent vegetation; Matorral rosetófilo = Rosetophile scrub; Agricultura de temporal = Rain-fed agricultura.

Figure 6. Map of land-use categories in the State of Morelos in 2000.

Cuerpos de agua = Water bodies; Bosque de encino = Oak forest; Bosque degradado = Degraded forests; Suelo urbano = Urban soil; Selva baja caducifolia = Low deciduos forest; Matorral rosetófilo = Rosetophile scrub; Bosque mesófilo = Mesophilic forest; Agricultura de riego = Irrigated agriculture; Agricultura de temporal = Rain-fed agricultura; Vegetación riparia = Riparian vegetation; Pastizal natural = Natural pastures; Vegetación secundaria = Secondary vegetation; Nubes = Clouds; Bosque de coníferas = Conifer forest

Figure 7. Map of land-use categories in the State of Morelos in 2009.

In order to validate the maps, all the classifications were assessed. There was a global accuracy of 83.5 % for the year 2000, and of 75 % for 2009. The Kappa indices (Table 1) were 0.80 and 0.71 for the 2000 and 2009 classifications, respectively. According to Viera and Garret (2005), a Kappa index with values ranging between 0.70 and 0.81 indicates an acceptable accuracy; hence, the generated maps approach come close to the actual conditions.

Table1. Validation of the classifications in the 2000 and 2009 maps.

In the year 2000, the most problematic classes, with a users’ accuracy ranging from 41.8 % to 81.7 %, were the rosetophile scrub, the mesophilic forests, the riparian vegetation, the natural pastures, the secondary vegetation and the degraded forests (Table 2). The rosetophile scrub and secondary vegetation were the categories with the largest number of omission errors and the highest error rates. The rosetophile scrub was mistaken for secondary vegetation and degraded forests; whereas the secondary vegetation was also mistaken for degraded forests (Table 2).

Table 2. Producer (PROD) and user (USER) values with assessment statistics of the accuracy of the vegetal cover and land use categories for the years 2000 and 2009.

WB = Water bodies; OF = Oak forest; DF = Degradaded forest; LDF = Low deciduous forest; RS = Rosetophile scrub; MF = Mesophilic forest; IA = Irrigated agriculture; RFA = Rain-fed agriculture; WV = Without apparent vegetation;

US = Urban soil; RV = Riparian vegetation; NP = Natural pasture; SV = Secondary vegetation; CF = Conifer forest.

In the year 2009, the degraded forests, the riparian vegetation and the secondary vegetation exhibited problematic values, with the users’ accuracy ranging between 17.9 % and 68.3 %. The secondary vegetation and the degraded forest had higher error rates; the secondary vegetation was mistaken for riparian vegetation and for low deciduous forests, while the degraded forests were mistaken for conifer forests. These and the secondary vegetation exhibited the largest number of omission errors (Table 2).

Table 3 shows the estimated surfaces for each vegetation cover category and each land use category in the years 2000 and 2009.

Table 3. Surface area by vegetation cover and land use class in hectares (ha).

Abbrv = Abbreviation; OF = Oak forest; DF = Degradaded forest; LDC = Low deciduous forest; RS = Rosetophile scrub; MF = Mesophilic forest; IA = Irrigated agriculture; RFA = Rain-fed agriculture; WV = Without apparent vegetation; US = Urban soil; RV = Riparian vegetation; NP = Natural pasture;

SV = Secondary vegetation; CF = Conifer forest.

The vegetation cover that evidence a reduction in the corresponding period are the low deciduous forests with a mean annual rate of 1 841.8 ha, the degraded forest with 557.58 ha, the oak forests with 300.70 ha, and the area without apparent vegetation with 137 ha. The land uses that registered an increase are the rain-fed agriculture, with a mean annual rate of 714.41 ha, followed by the irrigation agriculture with 658.88 ha, natural pastures with 647.91 ha, rosetophile scrub with 512.49 ha, and urban land with 189.29 ha (Table 3).

In the two years, the prevalent land-use in the State of Morelos, in terms of the total percentage of surface area they occupy, corresponded to the low deciduous forests (figures 8 and 9), followed by the rain-fed agriculture, the secondary vegetation and irrigation agriculture; and lastly, by the riparian vegetation and urban land. The agriculture is the class of land use with the highest degree of transformation, with 28.97 % of the total for the year 2000, and 31.5 % for the year 2009.

Bosque de encino = Oak forest; Bosque degradado = Degradaded forest; Selva baja caducifolia = Low deciduous forest; Matorral rosetófilo = Rosetophile scrub; Bosque mesófilo = Mesophilic forest; Agricultura de riego = Irrigated agriculture; Agricultura de temporal = Rain-fed agriculture; Sin vegetación aparente = Without apparent vegetation; Suelo urbano = Urban soil; Vegetación riaparia = Riparian vegetation; Pastizal natural = Natural pasture; Vegetación secundaria = Secondary vegetation; Bosque de coníferas = Conifer forest.

Figure 8. Percentage of each land-use category in the State of Morelos in the year 2000.

Bosque de encino = Oak forest; Bosque degradado = Degradaded forest; Selva baja caducifolia = Low deciduous forest; Matorral rosetófilo = Rosetophile scrub; Bosque mesófilo = Mesophilic forest; Agricultura de riego = Irrigated agriculture; Agricultura de temporal = Rain-fed agriculture; Sin vegetación aparente = Without apparent vegetation; Suelo urbano = Urban soil; Vegetación riaparia = Riparian vegetation; Pastizal natural = Natural pasture; Vegetación secundaria = Secondary vegetation; Bosque de coníferas = Conifer forest.

Figure 9. Percentage of each land-use category in the State of Morelos in the year 2009.

The dynamics registered in the 2000-2009 period show that the most vulnerable vegetation class, in terms of surface area reduction, is the low deciduous forest. Furthermore, a significant change occurs in conifer forests and in degraded forests (figures 8 and 9). On the other hand, irrigation agriculture and rain-fed agriculture register an increase of 1 373 ha, as does urban land, with 189 ha.

Conclusions

In the State of Morelos, agricultural activities played a major role during the 2000-2009 period, when they exhibited a significant growth. This is understandable from the point of view of the economy, as it represents a short-term source of income for the population. However, its intensity has brought about a loss of natural cover (forests and rainforests) that may have an impact on the fertility and erosion of the land. On the other hand, there is an important transition from forests and rainforests to disturbed vegetation, and from this to pastures and agricultural lands, reflecting the ecological succession experienced by the natural vegetation.

The results show a significant increase in irrigation agriculture (11 % in 2000; 13 % in 2009) and in rain-fed agriculture (18 % in 2000, and 19 % in 2009) at the expense of low deciduous forests, which decreased from 22 % in 2000, to 18 % in 2009. The urban surface exhibited a significant growth. This suggests that the driving forces of the reduction of the area occupied by low deciduous forests in the state of Morelos are related to the economic profits generated by agricultural activities, as well as to the growth of urban areas. The results also evidence the fact that public policies in relation to the preservation and management of forests are channeled toward the north of the state (to the Chichinautzin mountain chain), having insufficient coverage of those regions where the low deciduous forests are prevalent.

The use and incorporation of new tools, such as the geographic information systems (GIS) and remote sensing, allow the estimation of the magnitude of the changes associated to differential processes of change of the vegetation cover and land-use, as well as accurately assessing the growth of anthropic activities at the expense of the existing natural vegetation.

Acknowledgements

We wish to express our gratitude for the support provided by the program for research on climate change of Universidad Nacional Autónoma de México (the National Autonomous University of Mexico). We also wish to thank Servicios Ambientales y Cambio Climático (Environmental Services and Climate Change, SACC, A.C.), RENICYT 2015/20930, for its support, as well as the Papiit IN303817 Project, “Huertos familiars y afines en México” (“Family orchards and similar agricultural systems in Mexico”), coordinated by Dr. María de Jesús Ordóñez Díaz, PhD, of CRIM, UNAM, and the local communities for the facilities provided during the study, and, particularly, to the reviewers of Revista Mexicana de Ciencias Forestales (Mexican Journal of Forestry) for their comments and suggestions.

Conflict of interests

Dr. María Cecilia del Carmen Nieto de Pascual Pola, PhD, declares that she did not participate in any of the activities involved in the editorial process of the present document.

Contribution by author

Jorge Escandón Calderón: development of the research Project, structure and in-field collation of the different vegetation covers and land uses; José Antonio Benjamín Ordóñez Díaz: analysis of the results and editing of the manuscript; María Cecilia del Carmen Nieto de Pascual Pola: revision of the manuscript; María de Jesús Ordóñez Díaz: interpretation of the results and revision of the manuscript.

References

Alatorre, L.C., E. Sánchez, J. P. Amado, L. C. Wiebe, M. E. Torres, H. L. Rojas, L. C. Bravo and E. López. 2015. Analysis of the temporal and spatial evolution of recovery and degradation processes in vegetated areas using a time series of Landsat TM images (1986-2011): Central region of Chihuahua, Mexico. Open Journal of Forestry 5(2): 162-180.

Bolstad, P. V., P. Gessler and T. M. Lillesand. 1990. Positional uncertainty in manually digitized map data. International Journal of Geographical Information Systems 4: 399-412.

Brown, D. G., B. C. Pijanowski and J. D. Duh. 2000. Modeling the relationships between land use and land cover on private lands in the Upper Midwest, USA. Journal of Environmental Management 59(4): 247–263.

Cervantes G., V., M. López G., N. Salas N. y G. Hernández C. 2001. Técnicas para propagar especies nativas de selva baja caducifolia y Criterios para establecer áreas de reforestación. Facultad de Ciencias de la UNAM. México, D.F., México. 174 p.

Challenger, A. 1998. Utilización y conservación de los Ecosistemas terrestres de México. Pasado, presente y futuro. Conabio, Instituto de Ecología y Sierra Madre. México. México, D.F., México. 847 p.

Chuvieco, E. 2000. Fundamentos de Teledetección Espacial. Ed. Rialp. Madrid, España. 568 p.

Colditz, R., R. M. Llamas and R. A. Ressl. 2015. Land cover change analysis in Mexico using 30m Landsat and 250m Modis Data. In: 36th International Symposium on Remote Sensing of Environment. The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. 11–15 May 2015. Berlin, Germany. https://admin.egal2017.bo/static/archivos_publicos/2168.pdf   (12 de febrero de 2018).

Congalton, R. G. and K. Green. 1999. Assessing the accuracy of remotely sensed data: Principles and practices. Lewis Publisher. Boca Raton, FL, USA. 137 p.

Congalton, R. G. and K. Green. 2009. Assessing the Accuracy of Remotely Sensed Data: Principles and Practices. 2nd ed. CRC/Taylor & Francis. Boca Raton, FL, USA. 178 p.

Congalton, R. G. 2004. Putting the map back in map accuracy assessment. In: Lunetta, R. S. and J. G. Lyon (eds.). Remote sensing and GIS accuracy assessment. CRC Press. Boca Raton, FL, USA. pp. 1-11.

Flamenco-Sandoval, A., M. Martínez R. and O. R. Masera. 2007. Assessing implications of land-use and land-cover change dynamics for conservation of a highly diverse tropical rain forest. Biological Conservation 138(1-2): 131 –145.

Galicia, L. and A. García-Romero. 2007. Land Use and Land Cover Change in Highland Temperate Forest in the Izta-Popo National Park, Central Mexico. Mountain Research and Development 27(1):48-57.

García, E. 1988. Modificaciones al sistema de clasificación climática de Köppen. Instituto de Geografía, UNAM. México, D.F., México. 98 p.

George, J., L. Baby, A. P. Arickal and J. Dev Vattoly. 2016. Land Use/Land cover Mapping With Change Detection Analysis of Aluva Taluk Using Remote Sensing and GIS. International Journal of Science, Engineering and Technology 4(2): 383-389.

Guerra M., V. y S. Ochoa G. 2006. Evaluación espacio-temporal de la vegetación y uso del suelo en la Reserva de la Biósfera Pantanos de Centla, Tabasco (1990-2000). Investigaciones Geográficas 59: 7-25.

Foody, G. M. 2002. Status of land cover classification accuracy assessment. Remote Sensing of Environment 80(1): 185–201.

Instituto Nacional de Estadística, Geografía e Informática (INEGI). 2015. México en cifras. http://www3.inegi.org.mx/sistemas/mexicocifras/default.aspx?e=17    (5 de mayo de 2016).

Karl, J. W. and B. A. Maurer. 2010. Multivariate correlations between imagery and field measurements across scales: comparing pixel aggregation and image segmentation. Landscape Ecology 25: 591–605.

Lambin, E. F., H. Geist and R.R. Rindfuss. 2006. Introduction: local processes with global impacts. In: Lambien, E.F. and H. J. Geist (ed.). Land-use and land-cover change. Springer. Berlin, Heidelberg. Germany. pp- 1-8.

Lillesand, T. M., R. W. Kiefer and J. W. Chipman. 2004. Remote sensing and image interpretation. 5th ed. Wiley. Great Britain, New York. 763 p.

Lunetta, R. S. and J. G. Lyon. 2004. Remote Sensing and GIS Accuracy Assessment. Lewis Publishers. Boca Raton, FL, USA. 292 p.

Mas, J. F. and R. González. 2015. Change detection and land use / land cover database updating using image segmentation, GIS analysis and visual interpretation. In: ISPRS Geospatial Week. The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. 28 September–3 October. La Grande Motte, France. pp. 61-65.

Morales-Barquero, L., A. Borrego, M. Skutsch, C. Kleinn and J. R. Healey. 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.

Nguyen, L. T., W. Pang, W. Chan, W. Peng and Y. Zhang. 2012. Predicting collective sentiment dynamics from time-series social media. In: Cambria, E., Y. Zhang, Y. Xia and N. Howard. (comps.). Proceedings of the First International Workshop on Issues of Sentiment Discovery and Opinion Mining (Wisdom ‘12). ACM. New York, NY, USA. doi: 10.1145/2346676.2346682.

Ordóñez, J. A. B., B. H. J. de Jong, F. García-Oliva, F. L. Aviña, J. V. Pérez, G. Guerrero, R. Martínez and O. Masera. 2008. Carbon content in vegetation, litter, and soil under 10 different land-use and land-cover classes in the Central Highlands of Michoacan, Mexico. Forest Ecology and Management 255(7):2074-2084.

Overmars, K. P. and P. H. Verburg. 2005. Analysis of land use drivers at the watershed and household level: Linking two paradigms at the Philippine forest fringe. International Journal of Geographical Information 19: 125–152.

Palacio-Prieto, J. L., M. T. Sánchez-Salazar, J. M. Casado I., E. Propin F., J. Delgado C., A. Velásquez M., L. Chias B., M. I. Ortiz Á., J. González S., G. Negrete F., J. Gabriel M. y R. Márquez H. 2004. Indicadores para la caracterización y ordenamiento del territorio. Semarnat-INE-UNAM. México, D.F., México. 161 p.

Programa de Ordenamiento Ecológico Regional del Estado de Morelos. 2013. Secretaría de Desarrollo Sustentable y Universidad Autónoma del Estado de Morelos. http://obum.zmcuernavaca.morelos.gob.mx/MORELOS/OET/POEREM_DIAGNOSTICO.pdf  (12 de febrero de 2018).

Qasim, M., K. Hubacek, M. Termansen and A. Khan. 2011. Spatial and temporal dynamics of land use pattern in District Swat, Hindu Kush Himalayan region of Pakistan. Applied Geography 31(2): 820-828.

Richards, J. A. 1986. Remote Sensing Digital Image analysis. An Introduction, Springer. Berlin, Germany. 454 p.

Richards, J. A. and X. Jia. 1999. Remote sensing digital image analysis. Springer. Berlin, Germany. 363 p.

Serra, P., X. Pons and D. Saurí. 2008. Land-cover and land-use change in a Mediterranean landscape: A spatial analysis of driving forces integrating biophysical and human factors. Applied Geography 28(3): 189-209.

Sierra-Soler A., J. Adamowski, Z. Qi, H. Saadat, and S. Pingale. 2015. High accuracy Land Use Land Cover (LULC) maps for detecting agricultural drought effects in rainfed agro-ecosystems in central Mexico. Journal of Water and Land Development 26 (7-9): 19-35.

Tran, H., T. Tran and M. Kervyn. 2015. Dynamics of Land Cover/Land Use Changes in the Mekong Delta, 1973–2011: A Remote Sensing Analysis of the Tran Van Thoi District, Ca Mau Province, Vietnam. Remote Sensing 7(3):2899-2925.

Trejo, I. y J. Hernández. 1996. Identificación de la selva baja caducifolia en el estado de Morelos, México, mediante imágenes de satélite. Investigaciones Geográficas (Es5): 11-18.

Trimble, S. W. and P. Crosson. 2000. U.S. soil erosion rates. Myth and reality. Science 289(5477): 248–250.

Vega G., A., J. López-García y L. Manzo Delgado. 2008. Análisis espectral y visual de vegetación y uso del suelo con imágenes Landsat ETM+ con apoyo de fotografías aéreas digitales en el Corredor Biológico Chichinautzin, Morelos, México. Investigaciones Geográficas 67:59-75.

Viera, A. J. and J. M. Garrett. 2005. Understanding interobserver agreement: the kappa statistic. Family Medicine 37(5): 360-363.

Zepeda G., C., X. A. Nemiga, A. Lot H. y D. Madrigal U. 2012. Análisis del cambio del uso del suelo en las ciénegas de Lerma (1973-2008) y su impacto en la vegetación acuática. Investigaciones Geográficas 78:48-61.

All the texts published by Revista Mexicana de Ciencias Forestales–with no exception– are distributed under a Creative Commons License Attribution-NonCommercial 4.0 International (CC BY-NC 4.0), which allows third parties to use the publication as long as the work’s authorship and its first publication in this journal are mentioned.