Human activities are changing the Planet, inducing high rates of extinction and a worldwide depletion of biological diversity at genetic, species, and ecosystem level. Biodiversity not only has an ethical and cultural value, but also plays a role in ecosystem function and, thus, ecosystem services, which are essential to civilization, economic production, and human wellbeing. The functional role of biodiversity is still poorly known; a minimum level of biodiversity is required for sustainable preservation of ecosystem functions, and as an insurance for future environmental changes. A large part of the biodiversity of the Alps is linked to an interaction between the natural environment and traditional human practices. At present, the change in land-use, with both intensification and abandonment, and other environmental and socioeconomic processes at different scales (urbanization, tourism, pol- lution, global change, etc.) are important forces of environmental change. Mowing and livestock grazing are prima- ry factors inhibiting woody plant succession in many areas of the Alps. Abandonment and fragmentation has result- ed in an expansion of ecotones and edge, with increase in tick-hosts and possibly changes in host-parasite interac- tions resulting from species concentration. The abandonment of mountain fields and meadows with a consequent expansion of shrubs and forests has caused a decrease of several grassland species, such as rock partridge Alectoris graeca; some arthropod communities of grassland have also been affected. Many forest species should find new opportunities, but in several cases the forests have become too dense for some species, such as for capercaille Tetrao urogallus. In the low altitude belts, a high species diversity co-occurs with human disturbance. Biodiversity studies require an interdisciplinary approach by the life sciences, and an interface to socioeconomic sciences. Preservation of species and landscape diversity cannot prescind from a dialogue between different actors and interests.

We discuss how sophisticated machine learning methods may be rapidly integrated within a GIS for the development of new approaches in landscape epidemiology. A multitemporal predictive map is obtained by modeling in R, analyzing geodata and digital maps in GRASS, and managing biodata samples and weather data in PostgreSQL. In particular, we present a risk mapping system for tick-borne diseases, applied to model the risk of exposure to Lyme borreliosis and tick-borne encephalitis (TBE) in Trentino, Italian Alps.

Abstract Blood-sucking arthropods can provide efficient transport of infectious agents from one host to another. The large variety of combinations of host, vector, and pathogen locked into eternal triangles is testimony to the advantage to parasites of utilizing this form of transmission. We consider in detail those infections transmitted by ticks, since these are almost all zoonoses, with wildlife species playing a central role in maintaining endemic cycles of infection. Ticks are relatively long-lived and many species are catholic in their blood-sucking habits, so they can act as the principal disease reservoir and often carry a community of pathogens. Humans or their livestock are frequently infected, often with fatal consequences, as they accidentally intrude on these natural cycles.

Diversity patterns of plants, fungi and arthropods were studied in 4 forest sites in the Provinces of Bolzano and Trento (Italian Alps). Two decidous mixed forest at low altitude (560 and 680 m), and two spruce forests close to the timber line (1750 and 1780 m) were sampled; 2351 species were identified. Data on climate, land use and stress in the forests were correlated with species diversity. Species richness was higher in the decidous forests, and a great turnover of species occurred between the decidous and coniferous stands. Similar patterns in some diversity parameters were recorded.

We applied the novel bootstrap 632+ rule to choose tree-based classifiers trained for modeling the risk of parasite presence in a host population of ungulates. The method is designed to control overfitting: compact classification trees (CART) are selected using a nonlinear combination of the resubstitution error and the standard bootstrap error estimate. Model selection based on the 632+ rule offers a gain over cross-validation for CART models. The tree classifier selected by the new rule for this application favourably compared with standard multivariate GLIM models.

A tree-based classifier of tick presence, developed to estimate the risk of tick-borne diseases, is currently being interfaced and applied with a Geographical Information System (GIS) based on GRASS (USA-CERL Geographic Resources Analysis Support System). Environmental factors (altitude, substratum, vegetation, exposition, etc.) and tick sampling are used to predict occurrence of the parasite Ixodes ricinus on a target territory. A Tcl/Tk interface to GRASS has been developed for regional data from the Sistema Information Ambiente Territorio (S.I.A.T.) of the Autonomous Province of Trento. The system predictions are consistent with the current knowledge of tick ecology and can be used for an effective control of tick-borne diseases.

The most important tick-deer system potentially supporting the epidemiology of Lyme disease in the Italian Alps is that regarding Ixodes ricinus (L.) and roe deer (Capreolus capreolus L.). In this study, the pattern of tick infestation on 562 male roe deer harvested in September 1994 in 56 game districts of Trentino, Northern Italy, was assessed. The prevalence and density of infestation by I. ricinus were analyzed by a model based on classification and regression trees (CART), using both discrete and continuous variables concerning environmental and host parameters. The model discriminated attitude and host density as the 2 variables having the greatest effect on the prevalence and density of infestation of deer; the levels of infestation were higher at an altitude below 1125 m or at roe deer densities over 8.5 head per 100 ha. The density of tick infestation tended to be higher in older roe deer.

As a contribution to the Terrestrial Ecosystem Research Initiative (TERI) within Framework IV of the EU, ECOMONT aims at investigating ecological effects of land-use changes in European terrestrial mountain ecosystems.ECOMONT is coordinated by Prof Cernusca (University of Innsbruck) and is carried out by eight European partner teams in the Eastern Alps, the Swiss Alps, the Spanish Pyrennees and the Scottish Highlands.ECOMONT focuses on an analysis of structures and processes in the context of land-use changes, scaling from the leaf to the landscape level.The following research topics are being investigated: Spatial distribution of vegetation and soil in the composite experimental sites; physical and chemical soil properties, SOM status and turnover; canopy structure, primary production, and litter decomposition; water relations of ecosystems and hydrology of catchment areas; microclimate and energy budget of ecosystems; gas exchange of single plants and ecosystems; gas exchange between the composite experimental sites and the atmosphere, population and plant biology of keyspecies, plant-animal interactions, potential risks through land-use changes; GIS; remote sensing -environmental mapping; modelling activities integrating from plant to ecosystem and landscape level.First results of ECOMONT show that land-use changes have strong impacts on vegetation composition, structure and processes, on soil physics and chemistry, and therefore strongly affect exchange processes with the atmosphere and biogeochemical cycles.Abandonment of traditional 145

Cases of Lyme disease and tick-borne encephalitis were recognized recently in the Province of Trento, Italian Alps. Assessment of areas of potential risk for these tick-borne diseases is carried out by a model based on classification and regression trees (CART), using both discrete and continuous variables. Data on Ixodes ricinus (L.) occurrence resulted from extensive sampling carried out by standard methods in 99 sites over an area of approximately 2,700 km2 in the Province of Trento. A series of environmental parameters were recorded from each site and population densities of roe deer, Capreolus capreolus (L.), were considered. The CART model discriminates 2 variables that appear to have the greatest effect on the mesoscale occurrence of ticks: altitude and geological substratum, with a drastic decrease of tick frequency above an altitude of approximately 1,100 m and on volcanic substrata. The model is effective in identifying the mesoscale areas at greater potential risk, with a relatively low sampling effort.