Classification of Trichonis Lake graben (Western Greece) alluvial fans and catchments using geomorphometry and artificial intelligence

In this paper morphometric variables describing 35 alluvial fans and their contributing catchments in the Trichonis lake graben are examined. Trichonis graben, located in western continental Greece, is a Plio-Quaternary depression bounded by E-W and NW-SE trending active normal faults. The morphology of the fans and their corresponding catchments was quantitatively expressed through 15 selected morphometric parameters estimated using Geographical Information System (GIS) techniques. The relationships among the geomorphometric features of the fans and their catchments were examined, while the Self-Organizing Map (SOM) method, an unsupervised classification type of Artificial Neural Network (ANN), was used to investigate the clustering tendency of fans based on morphometric variables describing both the fans and their catchment areas. Very strong positive relations were found between the size of fans and size of catchments, as well as between the Melton’s Ruggedness Number and the slopes of fans, while the correlation between size of catchments and fan slopes was negative. Geomorphometry coupled with artificial intelligence methods led to the identification of four groups of alluvial fans and catchment areas with different geomorphic characteristics. These groups have a clear physical explanation and reflect the processes (fluvial and/or debris flow) related to the development of the fans. The geographical distribution of the fan-catchment systems is primarily controlled by the local tectonic patterns which are the main control on the accommodation space for the development and accretion of the fans.