Semiautomatic recognition of karstic depressions by analysis of aggregates in their morphometric distributions. Application to the karst of the State of Yucatán, Mexico

  • Carlos Paredes Departamento de Ingeniería Geológica y Minera, Escuela de Minas y Energía, Universidad Politécnica de Madrid, Ríos Rosas 21, 28003 Madrid, Spain.
  • Hector Matos Departamento de Ingeniería Geológica y Minera, Escuela de Minas y Energía, Universidad Politécnica de Madrid, Ríos Rosas 21, 28003 Madrid, Spain.
  • Rogelio de la Vega Departamento de Ingeniería Geológica y Minera, Escuela de Minas y Energía, Universidad Politécnica de Madrid, Ríos Rosas 21, 28003 Madrid, España.
  • Eulogio Pardo Departamento de Investigación en Recursos Geológicos, Instituto Geológico y Minero de España, Ríos Rosas 23, 28003 Madrid, Spain.
Keywords: karst, doline, cenote, uvala, polje, clustering, k-means, Yucatán, Mexico

Abstract

The Yucatan Peninsula, one of the five regions in which the Republic of Mexico is physiographically divided, is mainly constituted by carbonate rocks of sedimentary origin, which, due to its geological, structural, climatic and hydrogeological conditions, hosts important dissolution processes, which configure surface and underground karst landscapes. Among these, karst depressions in the form of cenotes, sinkholes, uvalas and poljes, have been the subject of many studies, because they constitute important groundwater and historic sociocultural resources, very sensitive to the environmental alteration. Currently, there are about 3000 collected karst depressions in the State of Yucatan (EY), however, the thick vegetation cover usually hinders its location and / or classification, reducing the completeness of the inventory. This paper presents the procedure used to build a catalog in the EY, grouped by types of karstic depressions, using a k-means semi-automatic classification system, using the most relevant morphometric parameters of the depressions, identified by a hierarchical clustering. The inventories used here are: cenotes (from SEDUMA), water bodies (from INEGI), and hypsometric level curves from 1:50 000 topographical maps (INEGI) which have been complemented with the depressions obtained, by two algorithms: DEM (15 m resolution, CEM 3.0, INEGI) filling and hierarchical, implemented in ArcGis, For each inventory, once each has been corrected and rectified, the morphometric parameters have been calculated: area, perimeter, circularity, shape, width, length, and elongation, the first two (logarithmically transformed) and the third and fourth ones, have turned out to be more relevant, after the cluster hierarchical analysis. The k-means clustering method provide, after synthesis of each grouping on inventories, a total of 6587 small depressions, geographically around to the Chicxulub impact structure, 4428 uvalas, 89 large depressions, all of them outside to the Chicxulub edge, and 5176 cockpit karst depression, mos o the spreaded over Sierra de Ticul, Finally, probabilistic distributions of most representative morphometric parameters, have been determined, characterizing each karstic sink tipology resulting clusters.

Published
2019-07-28
Section
Articles