Identification of environmental covariates that influence the formation of gullies in the Mixteca of Oaxaca.

María Lorenza Toledo, Demetrio S. Fernández Reynoso, Mario R. Martínez Menez, Erasmo Rubio Granados, José Luis García Rodríguez

Abstract


The Mixteca Oaxaqueña is a region affected by gullies, which trigger the interaction of geological, topographic and anthropogenic covariates. The latter is the result of irrational use of natural resources, which has had greater relevance since the Conquest. The present research seeks to identify the environmental covariates that favor gully formation in the Mixteca region of the Mexican state of Oaxaca, using cartography at scales lower than 1:50 000 and multivariate analysis of the main factors that, according to the literature, influence this type of erosion. The preselected covariates were subjected to statistical analysis to discriminate those that correlated highly with each other. From this process we selected the following for analysis: distance to faults, distance to peaks, terrain slope, vertical curvature, horizontal curvature, topographic position index, slope length, topographic index of moisture, population density and vegetation density. Supervised classif ication was applied to these covariates by the method of maximum likelihood, on six main components that, according to the principal component analysis (PCA), condensed 96.7% of the information contained in these ten covariables. The probabilities of gully susceptibility were obtained from the classif ication, and gully presence was verif ied in Google Earth for those pixels with probability higher than 50%. The reliability of the resulting map, based on absence and presence of gullies, was 96.2%, for a buffer of 50 m reliability was 83.4%, for a radius of 100 m it was 81.7%, and for one of 150 m it reached 77.9%. From this work, it was concluded that covariates with greater influence on gully formation, for the study area and in order of importance are distance to faults, distance to peaks, topographic moisture index, vegetation density, slope of the terrain and population density.

Keywords


maximum likelihood; multivariate analysis; monitoring points; elevation models



DOI: https://doi.org/10.28940/terra.v36i4.329

Refbacks

  • There are currently no refbacks.