In this paper we attempt to classify drainage sub-basins according to their erosion risk. We have adopted a multi-step procedure to face this problem. The input variables were introduced into a GIS – platform. These variables are the vulnerability of the surface rocks to erosion, topographic variations, vegetation cover and land use and drainage basin characteristics. We then constructed a fuzzy inferencemechanism to pre-process the input variables. Next we used neural-network technology to process the input variables. The above system was trained to ‘learn’ and classify the input data. The output of this procedure was a classification of the sub-drainage basins related to their risk of erosion. This neuro-fuzzy system was applied to the island of Lefkas (Greece).

In this paper the study of geomorphological characteristics, developed because of the run off, in the Cyclades area and especially in Tinos, Mykonos and Naxos islands, has taken place. In order to explain their development, the geomorphological characteristics were correlated with the lithological, tectonical and hydrological characteristics of the study area. The study area is characterized by a wide variety of different lithological and hydrogeological factors. To achieve this study’s targets, GIS and GPS technology was used in order to geographicallyand statistically analyse the morphological, hydrological, lithological and tectonic factors of this area. The main sources were the pre-existent maps and the literature, aerial photos and satellite images, while the extensive field work resulted to the geomorphological mapping of the area. The results of this study are related to the analysis of the current morphology and are presented on maps and rodograms.

This paper describes the study of the structural control of the Meganissi island coastal zone and its geomorphological evolution; based on that it detects and presents coastal hazard. The coastal geomorphology is affected by the lithological configuration, the tectonism, the wider seismotectonic status and the sea activity. For this study, a database has been structured in G.I.S. was using all data deriving from the above characteristics as well as detailed fieldwork, aerial photos, satellite images and pre-existing maps. Data were analysed and processed in order to locate zones of high hazard risk. Finally, a hazard risk map was developed for Meganissi island.

In this paper the counteractive dynamic marine and river environments responsible for a river mouth’s and a delta’s modulation is studied. The Acheloos River (West Greece) was chosen as the case study due to the recent changes taken place in its basin. The aim of this study is to analyze the present dominant environment and the impact of human activity on this balance. A GIS database was developed and updated in order to provide access for future use and computer analysis (both spatial and chronological). The study, includes an estimation of the delta forming processes, the impact of the anthropogenic parameters on these processes, and, finally, their representation.

The Cycladic islands are located in the central Aegean Sea (Greece) in a relatively low seismicity area and are composed mainly of metamorphic and plutonic rocks. Although the Cycladic plateau is believed to be an entity, a closer morphologic study of the islands through GIS, separates them into two parts: the eastern shallower one (Andros, Tinos, Mykonos, Naxos, Paros, Syros, Ios, Sikinos, Folegandros) which formed one big island (6.978Km2) at the end of the last glacial period; the western islands (Kea, Kythnos, Serifos, Sifnos, Milos) remained separated during the same period. It is concluded that the eastern islands constitute an erosional plateau which is a product of a Neogene palaeosurface that is partially submerged due to thinning of the crust during the Quaternary.

The goal is to visualize a set of multivariate data in such a way that datavectors belonging to different classes (subgroups) appear differentiated as muchas possible. When intending such visualization, the first question should be aboutthe intrinsic dimensionality of the data. The answer may be obtained byevaluating, e.g., the fractal correlation dimension. The projection to a plane isjustified when the correlation dimension of the data is about 2. Only in such casethe performed visualization is plausible to reflect all the between group and thewithin group relationships among the data vectors. There are several recognizedmethods for mapping data to a plane. Our interest lies especially in nonlinearmethods. We consider in detail three methods: The canonical discriminantfunctions, the kernel discriminant functions and the neuroscale mapping. Weillustrate our considerations using the Kefallinia erosion data, where each datavector belongs - in a crisp way – to one of five predefined subgroups indicatingthe severity of the erosion risk. The assignments to the subgroups were performedby an expert GIS system based on logical rules established by experts.

Naxos is the biggest island of the Cyclades. It is composed primarily of mica-shists, gneiss and marbles and granodiorites in the western part. The latter extend along parts of the studied area. Detailed coastal geomorphological study and GIS analysis made possible the determination of coastal evolution of this area during Late Quaternary. A number of shallow drillings and archaeological evidence helped in the determination of the sequence of the coastal evolution of the study area. Two sets of coastal dunes were mapped reaching elevation of 4 and 8 m respectively, plus older well preserved dune fields reaching elevations up to 18m located in the south. It was found that three granodiorite capes were small islands at the end of the last transgression in Mid Holocene. Since then, parts of the coastline, prograded joining them and forming coastal dunes, beachrocks and wetlands in several locations.

The main research aim of this paper is to examine the slope-stream erosionalprocess in the Korinthos drainage basin located in the north – eastern part ofPeloponnese, Greece, by means of applying the fuzzy set theory on a GIS platform.The main steps of this procedure are: the definition of the input variables (erodibilityof rocks, slope angle, slope morphology), the construction of a fuzzy inference system based on theoretical and empirical knowledge for transforming the input to output variables (erosion – deposition) and the visualization of the output variables (Spatial distribution of the erosion-deposition processes). The earth’s surface is the result of endogenous (weathering) and exogenous (erosion) processes. Weathering destroys chemically, biologically and mechanically the structure and the cohesion of the rock. The erosional mechanism transports the sediment and depends on many factors such as: frequency and magnitude of precipitation, surface and subsurface waters, slope elements (geometry, length and angle), rocks’ susceptibility to erosion, vegetation cover, human actions and land use. The study of the erosional process demands a series of steps in order to obtain the final output maps. These steps include: 1) field-work, air-photos and satellite images interpretation, map digitization (geological, topographical, drainage), 2) definition of input and output variables, 3) formulation of the proper logical rules (Boolean or fuzzy) for the transformation of input variables to output variables, 4) analysis and visualization of the output results.The input variables used in this research are: erodibility of rocks, slope angle andslope morphology. The erodibility of rocks depends on many variables which showspatial and time variation. Firstly, it depends on physical and chemical characteristics of the rocks and the rocks’ structure and discontinuities. Secondly, it depends on the existence of a protective vegetative cover which contributes in the increase of the infiltration capacity and limits the runoff flow. Thirdly, it depends on the characteristics of the involved processes: the intensity and duration of precipitation and also its spatial and time distribution and its raindrop size.The input variables are inserted into a fuzzy inference system, constructed in orderto transform them to output variables (erosion-deposition variable). This is achievedby the formulation of the proper ‘if…then’ rules based on theoretical as well asempirical knowledge of the erosional processes.The output variable (erosion) expresses the degree of erosion and deposition ofthe various parts of the drainage basin and is presented through erosion risk maps of the area that is studied.