Seawater intrusion is a pervasive problem affecting coastal aquifer, where the concentration of population and the increasing water demand creates risks of overexploitation, especially in those areas where is the only resource of drinking and irrigation water. This phenomenon is more considerable for the coastal karst aquifers, as observed in many Mediterranean countries and in some Italian regions as Friuli, Sardegna, Sicilia, Lazio, Campania and Puglia. This note aims to describe a research activity finalised to define a numerical model as management tools for groundwater resource of Salento (South Italy) to reduce the quantitative and qualitative degradation risks. The numerical codes used was MODFLOW (McDonald and Harbaught, 1988) and SEAWAT (Guo and Langevin, 2002). The active domain of the study area (active cells) covered approximately 2,300 km2 with 45,925 cells. Vertically, to allow a good lithological and hydrogeological discretization, the area was divided into 12 layers, from 214 to -350 m asl. Thickness and geometry of layers was defined on the basis of the aquifer conceptualisation based on the 3D knowledge of hydrogeological complexes. On the basis of detailed geological and hydrogeological conceptualisation, the climate change effects were considered in terms of natural recharge variations from 1930 to 1999 (Cotecchia et al., 2005; Polemio and Casarano, 2008). To take account of anthropogenic activity, mainly due to tourism and agriculture, the discharging trend was assessed, focusing on late decenniums (eighties and nineties), in which the discharge increase was mainly observed. Models representing the natural steady-state condition (using data of thirties) and transient scenarios of late decenniums were realised. The purpose of this first model implementation was, besides validated model, to assess the groundwater availability and quality in a recent period of seventy years (Polemio and Romanazzi, 2012; Romanazzi and Polemio, 2013). Results emphasize an essential decrease of piezometric levels and a worsening of seawater intrusion. On these bases, six forecasting transient scenarios were implemented, referred to future periods of about twenty years (2000-2020, 2021-2040 and 2041-2060) with the aim to predicting the evolution of piezometric level and seawater intrusion. For forecast data about precipitation and temperature, among the many models in the literature, we referred to the model developed by Giorgi and Lionello (2008), in relation to the defined scenario A1B. The model predicts temperature variations (°C) and precipitation percentage variation for the period 2001-2100. It was considered an average temperature variation form 0.9 °C (2001-2020) to 2.4 °C (2040-2060). Precipitation shows a negative percentage change (referred to 1960-80) equal to -3.9, -5.9 and -9,0% respectively for 2000-2020, 2021-2040 and 2041-2060. These climatic data are in agreement with other climate change models (Garcia- Ruiz et al., 2011). For the three future scenarios new recharge and discharge were assessed. In terms of discharge, they are mainly due to irrigation. For this kind of future utilisation two hypotheses were considered. The first assumes that type and extension of cultivations will be steady and, as an effect of climate change, the pressure on groundwater resource will further rise as necessary to satisfy irrigation demand (Dragoni and Sukhjia, 2008; Goderniaux et al., 2008). In the second hypothesis the irrigation discharge will be steady and equal to those of the 1999 due the adaption of cultivation types and irrigation practices. In both cases the scenario results show a general decrease of the piezometric head and a deterioration of water quality caused by seawater intrusion (Romanazzi et al., 2013). The results call for new land and groundwater resources management criteria. Considering the Water Framework Directive (EC, 2000) and international and regional experiences (LaMoreaux, 2010; Jiménez-Madrid, 2010; Polemio et al., 2009, Polemio et al., 2010), the study area was subdivided in three zones. To define the zone boundary, the threshold criterion was used (Polemio and Limoni, 2001; Polemio et al., 2009). The threshold between pure fresh groundwater and any type of mixing between fresh and saline groundwater was defined equal to of 0.5 g/l. In the first zone, the coastal zone, salinity was always (in the past) above the threshold, a transition zone, where salinity was variable respect to the threshold, and a third zone or inland zone where salinity value was permanently below the threshold. These three zones were implemented in the model. Different combinations of discharge criterions applied to these zones suggest the best choices to be applied for management criteria able to safely considered the future effects of climate changes.

The statistical and hydrogeological analysis of the relationships between rainfall, river and piezometric level historical data can be useful to characterize the aquifers and to manage the groundwater resources. For this purpose measurements acquired every three days, relative to 1986-2009 period, concerning the Pescara river alluvial plain (Fig.1), were analyzed with several statistical methods. The alluvial bodies of the Pescara river is mainly silty-sandy. The plain aquifer is supported by Plio-Pleistocenic clayey deposits. The three wells (Fig. 2) are located in the medium-low alluvial plain. Autocorrelation and spectral univariate analysis, cross-correlation and bivariate spectral analysis have been implemented with the purpose to evaluate the memory effect, the delay of the piezometric level response to rainfall and river head/discharge impulse, and the periodical components of the time series (Mangin, 1984; Larocque et al., 1998; Polemio and Dragone, 1999).

The statistical and hydrogeological analysis of the relationships between rainfall, river and piezometric level historical data can be useful to characterize the aquifers and to manage the groundwater resources. For this purpose measurements acquired every three days, relative to 1986-2009 period, concerning the Pescara river alluvial plain (Fig.1), were analyzed with several statistical methods. The alluvial bodies of the Pescara river is mainly silty-sandy. The plain aquifer is supported by Plio-Pleistocenic clayey deposits. The three wells (Fig. 2) are located in the medium-low alluvial plain. Autocorrelation and spectral univariate analysis, cross-correlation and bivariate spectral analysis have been implemented with the purpose to evaluate the memory effect, the delay of the piezometric level response to rainfall and river head/discharge impulse, and the periodical components of the time series (Mangin, 1984; Larocque et al., 1998; Polemio and Dragone, 1999).

Second half of the 20th century was characterized by an increase of groundwater discharge. Numerous aquifers are overexploited in the world and in particular in the Mediterranean area. Problems tie to overexploitation, as piezometric decline and increase of seawater intrusion, are amplified in karst coastal aquifers where the whole effect could be a groundwater quality and quantity degradation. Focusing on Mediterranean countries, most part of coastal aquifers of Spain, France, Portugal, Slovenia, Croatia, Greece, Albania, Turkey, and Italy are karstic and affected, to different degrees, by seawater intrusion due high pumping extraction rates and low recharge. (COST, 2005; Polemio et al., 2010). Climate change may particularly aggravate these requirements, especially in the Mediterranean areas, due to the combined effects of semiarid condition climate, or reduced recharge and consequent increase of discharge (Cotecchia et al., 2003; Polemio 2005; Polemio et al., 2009). The general purpose of this paper is to prove the capability of large-scale numerical models in management of groundwater, in particular for achieve forecast scenarios to evaluate the impacts of climate change on groundwater resources of karst coastal aquifer of Salento (Southern Italy). The computer codes selected for numerical groundwater modelling were MODFLOW and SEAWAT. Three forecast transient scenarios, referred to 2001-2020, 2021-2040 and 2041-2060, were implemented, on the basis of calibrated and validated model, with the aim to predicting the evolution of piezometric level and seawater intrusion. The scenarios were discussed considering the effects of climate change, sea level rise and change of sea salinity.

Second half of the 20th century was characterized by an increase of groundwater discharge. Numerous aquifers are overexploited in the world and in particular in the Mediterranean area. Problems tie to overexploitation, as piezometric decline and increase of seawater intrusion, are amplified in karst coastal aquifers where the whole effect could be a groundwater quality and quantity degradation. Focusing on Mediterranean countries, most part of coastal aquifers of Spain, France, Portugal, Slovenia, Croatia, Greece, Albania, Turkey, and Italy are karstic and affected, to different degrees, by seawater intrusion due high pumping extraction rates and low recharge. (COST, 2005; Polemio et al., 2010). Climate change may particularly aggravate these requirements, especially in the Mediterranean areas, due to the combined effects of semiarid condition climate, or reduced recharge and consequent increase of discharge (Cotecchia et al., 2003; Polemio 2005; Polemio et al., 2009). The general purpose of this paper is to prove the capability of large-scale numerical models in management of groundwater, in particular for achieve forecast scenarios to evaluate the impacts of climate change on groundwater resources of karst coastal aquifer of Salento (Southern Italy). The computer codes selected for numerical groundwater modelling were MODFLOW and SEAWAT. Three forecast transient scenarios, referred to 2001-2020, 2021-2040 and 2041-2060, were implemented, on the basis of calibrated and validated model, with the aim to predicting the evolution of piezometric level and seawater intrusion. The scenarios were discussed considering the effects of climate change, sea level rise and change of sea salinity.

We have been witness, during the second half of the 20th century, of an increase of groundwater discharge. Today a great number of aquifers are overexploited in the world. Problems ties to overexploitation, as piezometric decline and increase of seawater intrusion, are so more amplify in the coastal aquifers, and in particular, in karst coastal aquifers. Seawater intrusion, in fact, is a pervasive problem affecting coastal aquifer, where the concentration of population and the increasing water demand creates risks of overexploitation, especially in those areas where is the only resource of drinking and irrigation water. The whole effect could be a groundwater quality and quantity degradation. This is very often the case of coastal karst aquifers of Mediterranean countries. The general purpose of this paper is to prove the capability of large-scale numerical models in management of groundwater, in particular for achieve forecast scenarios to evaluate the impacts of climate change on groundwater resources. Study area is the karst coastal aquifer of Salento (Southern Italy), largely utilized to satisfy the agricultural demand and drinking demand with huge effects in terms of reduced availability and increasing salinity. The computer codes selected for numerical groundwater modelling were MODFLOW and SEAWAT. Groundwater flow modelling is based on the concept of a equivalent homogeneous porous medium. Three forecast transient scenarios, referred to 2001-2020, 2021-2040 and 2041-2060, were implemented, on the basis of calibrated and validated model, with the aim to predicting the evolution of piezometric level and seawater intrusion. The scenarios were discussed considering the effects of climate change, sea level rise and change of sea salinity. Some irrigation discharge scenarios were considered in the discussion . Results shows qualitative and quantitative groundwater trends from 1930 to 2060 and emphasizes an essential decrease of piezometric level and a huge worsening of the groundwater salinisation due seawater intrusion. More details on previous results of this research activity were recently published (Polemio and Romanazzi, 2012; Romanazzi and Polemio, 2013).

The aim of this work is to present a methodology, based both on the use methods of time series analyses and of geospatial analyses of monthly climatic data (rainfall, wet days, rainfall intensity, and temperature), annual maximum of short-duration rainfall (from 1 hour to 5 days), historical modification of land use, and population variations in order to characterise the effects of these variables on the occurrence of landsliding in Daunia area, located on the eastern margin of the Southern Apennines thrust belt (southern Italy). Rock strata (mainly) interbedded with clayey marls, clays and silty-clays outcrop in this area. Due to the intense strain history, these successions are found to be from stratified to deeply fractured, up to be disrupted and floating as blocks in a clayey matrix. In turn, the clay units are laminated to intensely fissured and characterised by very poor mechanical properties (Santaloia et al., 2012). The statistical analyses deal with data coming from published databases, integrated by public and private documents, referring to a wide time span. Climate data records from 1877 to 2008 were elaborated, in particular the data coming from sixteen rainfall gauges, ten of which were also thermometric. Moreover, some monthly indices of rainfall, wet days, rainfall intensity, temperature, and landslide occurrence were introduced to simplify the analysis of parameters, characterised by spatial and temporal variability. The population records are from the 19th century up to now while the time period of reference for the land use data is from 1930 up to now. As concerns the landslide events, they were collected from 1918 to 2006. The main source of these records is the AVI database, an existing Italian database that collects data about damaging floods and landslides from 1918 to 1996. This dataset was integrated up to 2006 by consulting newspapers, scientific publications, technical reports, written by the researchers of the CNR-IRPI for the Civil Protection, and also documents belonging to a research project (PS_119; Cotecchia et al. 2010). According to the landslide data collected, the landslide events resulted to be 175 in the study area. The trend analyses show that the landslide occurrence was increased with the time, despite of the rainfall and temperature data are not prone to landsliding. As a matter of fact, the trend of both the monthly rainfall and the rainfall intensity decreases, and the temperature and the wet days show a positive trend during the period of reference. The trend of the short-duration rainfall results generally to decrease. Not existing an evident relationship between climate variability and the increase of landslide occurrence, some other factors should be considered, as, for instance, the poor mechanical soil properties, the role of anthropogenic modifications and the mismanagement of risk-prone areas. In this regards, the preliminary results obtained from the data analyses of the land use and the populations could partly justify the increasing trend of landslide occurrence. More details on previous results of this research activity were recently published (Cotecchia et al., 2010; Polemio and Lonigro, 2011 and 2013; Santaloia et al., 2012).

Historical data about floods represent an important tool for the comprehension of the hydrological processes, the estimation of hazard scenarios as a basis for Civil Protection purposes, as a basis of the rational land use management, especially in karstic areas, where time series of river flows are not available and the river drainage is rare. The research shows the importance of the improvement of existing flood database with an historical approach, finalized to collect past or historical floods event, in order to better assess the occurrence trend of floods, in the case for the Apulian region (south Italy). The main source of records of flood events for Apulia was the AVI (the acronym means Italian damaged areas) database, an existing Italian database that collects data concerning damaging floods from 1918 to 1996. The database was expanded consulting newspapers, publications, and technical reports from 1996 to 2006. In order to expand the temporal range further data were collected searching in the archives of regional libraries. About 700 useful news from 17 different local newspapers were found from 1876 to 1951. From a critical analysis of the 700 news collected since 1876 to 1952 only 437 were useful for the implementation of the Apulia database. The screening of these news showed the occurrence of about 122 flood events in the entire region. The district of Bari, the regional main town, represents the area in which the great number of events occurred; the historical analysis confirms this area as flood-prone. There is an overlapping period (from 1918 to 1952) between old AVI database and new historical dataset obtained by newspapers. With regard to this period, the historical research has highlighted new flood events not reported in the existing AVI database and it also allowed to add more details to the events already recorded. This study shows that the database is a dynamic instrument, which allows a continuous implementation of data, even in real time. More details on previous results of this research activity were recently published (Polemio, 2010; Basso et al., 2012; Lonigro et al., 2013)

The Damaging Hydrogeological Events (DHEs) can be defined as the occurrence of one or more simultaneous phenomena, such as droughts, windstorms, heat waves, landslides, floods and secondary floods (i.e. rapid accumulation or pounding of surface water with very low flow velocity), causing damages. They represent a serious problem, especially in DHE-prone areas with growing urbanization, where the infiltration capability is limited by buildings and where the vulnerability is higher than other areas. The paper proposes a methodology, based on both historical and time series approaches, used for describing the influence of climatic variability and urban development on the number of phenomena observed. The historical approach is finalised to collect phenomenon historical data, very important for the comprehension of the evolution of a study area. Phenomenon historical data is useful for expanding the historical period of investigation in order to assess the occurrence trend of DHEs. The historical analysis of DHEs can support decision making and land-use planning, ultimately reducing natural risks. The time series approach includes the collection and the statistical analysis of climatic data (monthly rainfall, wet days, rainfall intensity, and temperature), useful to characterise the climate variations and trends and to roughly assess the effects of these trends on river discharge and on the triggering of landslides. The time series approach is completed by tools to analyse simultaneously all data types. The study of land use variations, with a special emphasis on the urban areas, is important to understand how the modifications occurred in the territory, especially in terms of vulnerability, could influence the occurrence of DHEs. The methodology can be applied simultaneously to floods and landslides and was tested considering the municipality of Bari (southern Italy), particularly affected by flood events. Since the climate trend (decreasing trend of rainfall and rainfall intensity and an increasing trend of wet days and temperatures) does not show favourable conditions for the increase of the annual number of damaging floods, its trend is increasing. The role of anthropogenic modifications and the mismanagement of risk-prone areas should be considered to justify the increasing occurrences of floods. A validation of this hypothesis comes from the study of land use modifications, carried out comparing different temporal levels of land use (from 1959 to 2006). The analysis shows, starting from 1959 to 2006, a significant increase in urban areas (of about 50%) on the entire regional territory. The municipality of Bari, the regional main town, has undergone a remarkable development of its urban areas, from 12.45 Km2 in 1959 to 58.82 Km2 in 2006. The consequent increased vulnerability of this area has been highlighted during the recent flood event occurred in 2005, which caused six casualties, numerous injuries and damages to roads, buildings, industries, agriculture, livestock and services. More details on previous results of this research activity were recently published (Polemio, 2010; Polemio and Lonigro, 2012).