Epikarst mapping by remote sensing

doi:10.1016/j.catena.2018.01.026

2018-06-01

PRE

Eulogio Pardo‐Igúzquiza *

Peter A. Dowd

Ana Ruiz‐Constán

Sergio Martos‐Rosillo

Juan Antonio Luque-Espinar

Víctor Rodríguez‐Galiano

Antonio Pedrera

DOI：10.1016/j.catena.2018.01.026

原文链接

原文求助

摘要

En 中文

Epikarst the shallow, surficial part of a karstic massif - has a significant influence on the spatio-temporal variability of recharge and the hydrodynamic functioning of many karst aquifers. In the Mediterranean morphoclimatic zone, the average thickness of a well-developed epikarst is around ten metres, but the spatial patterns of its degree of development are very heterogeneous due to the complex interaction of a number of different factors such as lithology, fracturing, weathering, soil and vegetation. In addition, direct field observation is difficult because good outcropping conditions are restricted to particular locations, some areas are not accessible and the size of the study area is often too large for exhaustive field surveys. Satellite-based remote sensing, however, provides a complete coverage of an entire area with spectral resolutions that detect variability in features that can define image textures related to the development of the epikarst. This paper describes a quantitative methodology for epikarst mapping using satellite images and field data. The proposed method comprises an unsupervised classification to define the spectral signature of each of three epikarst development categories in a high-resolution satellite image followed by a supervised classification of the terrain into one of the three categories on a low spatial resolution scale. The training areas in the field are assigned to the three categories by a panel of experts using the Delphi method. Geophysical data are used for validation to overcome any bias that may be introduced by the panel. The proposed methodology has been applied to the Sierra de las Nieves karstic aquifer (Malaga, southern Spain). The outcome is a map of estimated epikarst development that is an approximation to reality and which can be improved as more experimental data become available.

Keyword:

SPOT image

Preferential recharge

Spectral signature

Geophysical validation

Sierra de las Nieves