Abstract
Land use/cover (LUC) and changes between 1990 and 2003 in a tropical mountainous watershed were analysed with Landsat TM images using a GIS‐RS approach. The La Antigua River upper catchment is a 1325 km2, biodiverse hydrological region in central Veracruz, Mexico. A large set of training pixels was used to optimize the representation of environmental heterogeneity. Classification accuracy was assessed with spectral and field‐checked error matrices. Overall classification accuracy for the 1990 (78.2%) and 2003 (79.7%) images was satisfactory. Ancillary data (DEM) was incorporated to improve discrimination between LUC categories. The Landsat TM sensor proved sensitive enough to separate the different spectral patterns related to the LUC classes in this complex landscape. The time interval and scale selected are suitable for strategic planning purposes. Depletion of tropical montane cloud forest, and its conversion to pasture and agriculture, was by far the most important LUC change over the period of study.
Acknowledgements
The authors thank the Comisión Nacional Forestal (CONAFOR) of Mexico and the Municipality of Coatepec in the state of Veracruz for sharing the Landsat satellite images and the 2003 Quick Bird high‐resolution image, respectively. We appreciate the valuable comments and contributions made by the reviewers. We would also like to thank Dr. Friso Holwerda for helping to improve the manuscript. Financial support for this research was provided by the Consejo Nacional de Ciencia y Tecnología (CONACyT) through a doctoral research grant (No. 149367) to the first author under The Biological Science Program of the Universidad Autónoma Metropolitana (UAM), and in part by the project (SEP‐2003‐C02‐43082). Additional funding was provided by the Departamento de Ecología Aplicada of the Instituto de Ecología, A.C. (No. 10093).