Principal objective of the research was studying Roman landscapes with thecombinatory analysis of environmental, geomorphological data and with softcomputing applications such as GIS and algorithms, designed specifically forarchaeological purposes. The identification and the modeling of the possibleRoman cadastral systems via traditional methodological approaches comprised avital stimulus for the technological oriented approaches. Geomorphologicalanalysis of the landscape along with the implementation of the algorithmstrengthens the reliability of the output models and provides significantinformation for the palaeoenvironment. Modelling the spatial organisation of theRoman landscape in the area of Beziers (south France) pre-supposes acombinatory study of the physicogeographical, the geomorphological, thegeological and the archaeological characteristics of the area. Signifying the landboundaries and the road network of the area during Roman period indicates astraight forward human intervention on the landscape. Though, studying modernlandscape’s environmental settings may hinder conclusions about earlier periodsof use of the area, unless environmental data, to be added in the model, isevaluated in advance. Considering that the data collected and mapped nowadaysmay not be found in situ, emphasis is placed in the potential displacement of thearchaeological data to be input in the Roman cadastral models. In conjunctionwith the geomorphological analysis of the data, an erosion model, based onFractals philosophy, has been also developed assessing the consistency of dataused to derive possible models of Roman cadastres.

The evolution of landscapes as natural phenomena is rarely seen in a generation,but in volcanic regions the landscape is transformed at a tremendously fast rate.These places are physical bounded and marked by natural phenomena as well asimbued with cultural meaning. In regions where volcanic activity is dormant, it isargued that a different cultural association is formed; in this paper thepalaeoenvironmental and volcanic evolution of the sites are analysed in thevolcanically active Mývatn region (NE Iceland) and on the island of Thera (AegeanSea, Greece).The island of Thera consists of a basement of recrystallized limestones andmetamorphosed rocks. This basement is covered by a sequence of volcanicformations. The volcanic activity, which started about 1.6 million years ago, isstill active and it is the predominant natural feature of the island’s topography.The Mývatn region is a volcanically active area, straddling the American andEuropean plates along the North Atlantic ridge. The most recent lava eruptiontook place in the 1983 and land areas are constantly changing in response to newactivities. Few of the questions that the research will approach are:

• In what way have volcanoes evolved during the last millenniums and what dramatic changes have they brought to Natural resources?
• To which extent is an active long-term natural disaster a constant threat for human presence? How is human – nature balance attained?
• Is cultural meaning projected on such a natural landmark? Which aspects of human life are directly influenced (e.g. settlement patterns, land boundaries, resources and economic activities) and which aspects are more implicitly affected (e.g. mental images, religious perspective, spiritual perceptions)?

The widespread use of the GIS tools, besides the simple mapping of areas also enabled the visualization of various spatially referenced parameters for earth sciences. Parameters such as the slope or the drainage density of an area can be calculated with the GIS tools and can be depicted in maps. In many environmental problems though, the involved variables present inexact knowledge and spatial uncertainty. This is a problem appear also when we treat erosional processes, where a number of the system input variables are of fuzzy nature, such as the erodibility, the topography’s gradient, the drainage density and the vegetation - land use. The fuzzy nature of these physical parameters, if it is to be approached by mapping tools, requires the use of special methods that will assure that the geographicaldistribution of the results will be accurate. Dealing with such uncertainties and the visualization methodologies that should follow them, is the ultimate aim of this paper. In the approach that is described here, MapInfo GIS software and Matlab are used in combination to study the erosional processes on a volcanic landscape, such us Thira Island, Greece. Mapinfo software was used for the creation of the input parameters. The management of the database was created during the digitization and the creation of the input parameters and finally for the rendering of the geographical output of the study in the form of thematic maps. Matlab software wasused for the calculations within the different parameters and the gradient leveling that were applied to the logical results. The transformation of the input variables to the output ones (erosion risk) was achieved by formulating a number of fuzzy logical rules. The final step of these processes is the development of the thematic maps.Thira Island has been chosen for this case study, because its volcanic landscape and its susceptibility to erosional phenomena appeals great scientific interest and requires the introduction of different approaches for the solution of the natural problems. Besides, the output of this study can constitute an important tool for the local authorities of the island.

The aim at this paper is to study the historical allouviation of the area ofThessaloniki using mainly archeological data. Such data has been found in manyplaces in this area. These findings (buildings, castles, road traces, fortifications) dateback from Neolithic up to Ottoman period. We have approached this problem bystudying the geomorphologic processes of this area and by mapping in detail theallouviation of the ancient monuments. A spatial database has been created attributing geological, geomorphological, hydrological, other environmental and archaeological data along with all the measurements undertaken during fieldwork. Apart from contemporary data, older topographical maps of the area have also been input and analysed in the GIS together with the rest datasets. A specific methodology was applied in view of recording and studying the erosion-deposition rates. The derivative results have indicated that deposition rates vary within diverse temporal and spatial conditions. This is due to the topography, the lithology and the climate changes of this area. The study, also, reported that in some places vertical depositions rates at about 0, 5 cm/year. The applied model tracing erosional processes and deposition rates can be implemented in any other location irrespective of the environmental, spatial, or archaeological data characterizing the area.

Development of accurate and efficient models on erosion risk mapping are difficult to be applied on different environmental parameters, mainly due to the complexity of the integrated parameters. In this study the developed GIS model is based on the theory of Boolean logic. The rules integrate the main parameters for erosion, e.g. vulnerability, slope, drainage density, although it is easy for more parameters to be added. As case study area Penteli mountain (North-East Attiki) was selected, that suffers from intense erosion, mainly due to the last severe fire events. With the aim of restraining erosional processes, combinatory strategies, such as assignment of tree trunks parallel to contours, have been applied. The results of this study include a model for erosion, deposition and material transfer. Complementary, a GIS based model was developed so as to define the movement of eroded material within the different drainage basins of the area. The model defines the erosional and depositional sites and the results are confirmed by the measurements on the tree trunks.

During the last years the major threat for the deforested areas of the Crete island,especially the area of Moussouron municipality, is the quick wash out of soil material. The wash out of soil material, due to logging activities or fire events, obstracts efforts for reforestation and natural reconstruction of the site. For the needs of our research, the field experiments were applied at a high slope area situated at Moussouron municipality (Crete island - South Greece), where majorfire events recently took place and caused serious damages at the vegetation. Inspecific, neighboring areas of approximately 2.6km2 and 3.2km2 have been burnt in2003 and 2004, respectively. This case study area is mainly characterized by the Mediterranean climate, having extended dry periods during summer (5% of total rainfall) and periods with intensive rainfall during winter (95% of total rainfall). The area presents high relief alterations with intense slopes (~65% of the area has slopes over 20 degrees) and lithologically is mainly characterized by two basic formations; carbonate rocks at higher and schists at the lower altitudes. The aforementioned characteristics have caused the soil to intensively wash out just after the fires occurred. The main purpose of this research is to evaluate the correspondence between the amount of eroded material and the rain intensity, at conditions of recently burnt areas. In order to approach the problem, tree trunks in a parallel to the contour lines arrangement, were installed. This way the study area was split in numerous plots with predefined dimensions and precisely measured slopes. Just after each intense rainfall, thickness measurements of accumulated material were performed at each one of the tree trunks, while the runoff material was also estimated at the lowest parts of each plot. Αt the same time, using a rain gauge, rain intensity was also measured, in order to correlate this factor to the overall soilwashout. The research, performed within a period of 20 months with a total of 18 intense rain events, concludes to the rates of denudation and soil washout at non protected deforested areas and at the same time at areas were tree trunks were put as a prevention method. Particular preservation strategies and research methodologies adjusted to the needs of specific areas, may indeed act in favour of natural resources and retain ecological balance.

There is a wide range of alternative approaches to study erosion processes. In this paper the construction of a model based in the interaction of Geographical Information System (GIS) and Artificial Neural Networks (ANN) is described. The neural model uses supervised competitive learning process. The whole procedure starts with the digitization of the data and the definition of the input variables: such as slope form and gradient, lithology and vegetation - landuse. The neural modeltransforms the input variables into the erosion risk output variable. Thus, the last stage regarded the creation of an erosion risk zones map. For case study was chosen the island of Corfu (Greece). The island consists of lithologies very vulnerable to erosion and receives considerable amounts of rainfall, especially if compared to the rest of the Greek territory. Finally, the whole model was tested and the proper function of the model was confirmed by field data observations.

The post-Alpine sedimentation history in Greece is mainly characterized by thedevelopment of terrestrial, marine and lacustrine successions. During the UpperMiocene salinity crisis marine evaporites, mainly gypsum, were formed in severalbasins across Greece (i.e. Ionian Islands, Crete Island). These evaporites present simple mineralogy and they are almost insoluble. As a result, they appeared to be quite resistant in weathering process and they are easily identified in the field.On the contrary, Upper Miocene evaporites of lacustrine origin have a morecomplicated mineralogy and they are rare in field outcrops in a worldwide scale, since they are commonly easily weathered. Even though, the adjacent to Aegean Sea, Western Anatolia volcanic terrain hosts a series of lacustrine evaporites, the only similar deposit in Greece is that of Karlovassi Basin, in Samos Island.The aim of the present paper is to study the erosion risk of the Upper Mioceneevaporate - bearing formations of Karlovassi basin developing a GIS model based on Boolean logic. Some of the affecting parameters that were input into the model are: the vulnerability of the lithological formations, the run off model, the morphological slopes and the aspect, the tectonic discontinuities, the land use – land cover of the wider area and the man made intervention. The geographical distribution of the aforementioned parameters, along with their modelisation, resulted to an erosion risk map of the Karlovassi basin.

Using a given data set (the Kefallinia erosion data) with only 3 dimensions and with fractal correlation dimension rGP ¼ 1:60, we wanted to see, what really by the kernel methods is provided.We have used Gaussian kernels with various kernel width ¾. In particular we wanted to ¯nd out, whether the GDA (Generalized Discrimination Analysis) as proposed by Baudat and Anouar (2000), permits to distinguish better the high, medium and low erosion classes as compared to theclassical Fisherian discriminant analysis. The general result is that the GDA yields discriminant variates permitting for better differentiation among groups, however the calculations are more lengthy.

Kernel discriminants are greatly appreciated because 1) they may express nonlinear boundaries between classes and 2) they permit to visualize graphically the data points belonging to di®erent classes. One such method is called GDA (Generalized Discriminant Analysis), however it operates on a kernel matrix of size NxN and is for large N prohibitive. We illustrate how this method works in a real situation when dealing with relatively large data. We consider a set of predictors of erosion risk in the Kefallinia island categorized into 5 classes of erosion risk (together N=3422 data items). We argue that a proper preparation of appropriate learning samples can greatly speed up the evaluations and result in good generalization properties. Our concern is to ¯nd appropriate data for learning. This is done by a kind of sieve algorithm.

The study focuses on the application of computing elements in the field of geoarchaeology. Emphasis is placed on the meaningful combination of geomorphological and environmental data along with archaeological features within Geographic Information Systems. Principal objective was the identification and the modeling of possible Roman cadastres in the area of Béziers, in southern France. For this purpose numerous technologies have been applied; GIS, GPS, Remote Sensing techniques and Digital Image Processing methods were used for the collection, storage, digitization and analysis of geoarchaeological data. Geographical Information System (G.I.S.) was used for the processing of primary data, the production of secondary information layers, the development of the digital elevation model and its derivative files and, in essence, for the spatial and the quantitative analysis of the data. The research concluded at 3 potential Roman cadastre grids establishing orientation and distance of axes as main criteria. Further processing of the data concluded in one final Roman cadastre grid for the area of Biterrois in France.