Regional Assessment of Climate Change in the Mediterranean

Part I of this Report deals with the analyses of the physical aspects of climate change in the Mediterranean, aiming to assess possible changes under greenhouse concentrations scenarios, estimate uncertainties, detect and attribute past and present changes, understand mechanism and evaluate their relevance for the problem of climate change. This is realized with a vast arsenal of tools and techniques, including global and regional models, most of them specialized for the Mediterranean region, specific statistical analyses and new compilation of data sets for the region.

Mediterranean climate change during the last 60 years is based on homogenized daily temperature and quality controlled precipitation observational data and gridded products. The estimated changes indicate statistically significant Mediterranean summer temperature increase and a reduction in winter precipitation in specific areas. Reconstructions of Mediterranean sea level suggest a rise of some 150 mm since the beginning of the nineteenth century. A 20 years long reanalysis (1985–2007) was produced, showing long term temperature variability and a positive salinity trend in the ocean layers from the surface to 1,500 m depth. A prominent increase in summer temperature extremes is found in the whole Mediterranean region, while warm bias in the mid twentieth century station data is removed by homogenization. No basin-wide trends in precipitation and droughts are found for the second half of the twentieth century, while trends in extreme winds are largely negative, as are those of the related cyclones and cut-off-lows. The role of large scale pressure patterns like the NAO for variabilities and trends is discussed for the different parameters considered.

In this chapter we show results from an innovative multi-model system used to produce climate simulations with a realistic representation of the Mediterranean Sea. The models (hereafter simply referred to as the “CIRCE models”) are a set of five coupled climate models composed by a high-resolution Mediterranean Sea coupled with a relatively high-resolution atmospheric component and a global ocean, which allow, for the first time, to explore and assess the role of the Mediterranean Sea and its complex, small-scale dynamics in the climate of the region. In particular, they make it possible to investigate the influence that local air-sea feedbacks might exert on the mechanisms responsible for climate variability and change in the European continent, Middle East and Northern Africa. In many regards, they represent a new and innovative approach to the problem of regionalization of climate projections in the Mediterranean region.

The CIRCE models have been integrated from 1951 to 2050, with initial conditions obtained from a long spin-up run of the coupled systems. The simulations have been performed using observed radiative forcing (solar constant, greenhouse gases concentration and aerosol distribution) during the first half of the simulation period and the IPCC SRES A1B scenario during the second half (2001–2050).

The projections indicate that remarkable changes in the Mediterranean region climate might occur already in the next few decades. A substantial warming (about 1.5°C in winter and almost 2°C in summer) and a significant decrease of precipitation (about 5%) might affect the region in the 2021–2050 period compared to the reference period (1961–1990), in an A1B emission scenario. However, locally the changes might be even larger. In the same period, the projected surface net heat loss decreases, leading to a weaker cooling of the Mediterranean Sea by the atmosphere, whereas the water budget appears to increase, leading the basin to loose more water through its surface than in the past. The climate change projections obtained from the CIRCE models are overall consistent with the findings obtained in previous scenario simulations, such as PRUDENCE, ENSEMBLES and CMIP3. This agreement suggests that the results obtained from the climate projections are robust to substantial changes in the configuration of the models used to make the simulations.

Finally, the CIRCE models produce a 2021–2050 mean steric sea-level rise that ranges between +6.6 cm and +11.6 cm, with respect to the period of reference. Within the CIRCE project the results obtained from these models have been used to investigate the climate of the Mediterranean region and its possible response to radiative forcing. Furthermore, the data have been made available for climate change impact studies that are included in the Regional Assessment of Climate Change in the Mediterranean that has been prepared in the context of the CIRCE project.

This chapter describes the physicochemical mechanisms that formulate the air quality over the Mediterranean region and the resulted impacts on the regional climate. At first, a detailed description of the teleconnections and regional flow patterns that dominate in the region is provided. The dominant flow patterns during the different seasons of the year determine the transport paths of air pollutants and aerosols towards and across the study area. The analysis on the characteristics of the air pollution transport is separated for the different parts of the Mediterranean region (eastern, western and entire), since the sources of pollutants that reach at different points in the region vary, while specific pollutant transport paths may influence the wider Mediterranean area. Similarities and differences in patterns are discussed. The air quality over the region, as recorded from black/organic carbon, ozone, aerosol observations, is extensively discussed, along with seasonal variabilities and annual trends. There is particular discussion on the suspension of naturally-produced aerosols and especially desert dust particles in the region and their spatial influence on the aerosol levels. At the last part of the chapter, the major impacts of the transport and transformation processes (natural and anthropogenic pollutants) on the regional climate are discussed. The impacts of aerosols are distinguished in direct (the impacts on radiation budget), health (the amounts of inhaled particles and impacts on health) and indirect effects (impacts on clouds and precipitation), are discussed on qualitative and quantitative way.

We present a first assessment of the detection of a signal of temperature change over the Mediterranean domain, using HadCRUT3v observation dataset and model outputs from the CMIP3 climate simulations. For this study we have used two new formal detection methodologies, the ‘Regularized Optimal Fingerprint’ and the ‘Temporal Optimal Detection’, developed within the context of the CIRCE project and aiming at improving the ability to detect a climate change signal at the regional scale. We have also applied the ‘Consistency’ method that allows to answer the question whether a given forcing is a plausible explanation of an observed change. The results show the detection of a change on spatially centered temperatures, that allows to identify a regional structure of change additional to the global warming. The formal detection findings also extend to the winter and summer spatial patterns of temperature change. By applying the ‘Consistency’ method, we also detect the GS (Greenhouse gases and Sulfate aerosol) signal in observed annual and seasonal area-mean warming except in winter. Further we find that the recent trends in near-surface temperature are significantly consistent with the simulated GS patterns. Concerning precipitation, we cannot detect formally a signal of climate change on Mediterranean precipitation using 17 series of monthly precipitation from Croatian, French and Italian coastal stations and the CMIP3 climate simulations. However this may be due, at least partly, to the limited extent of the region covered with the precipitation series.

An analysis of the observed climate change in the Mediterranean region has been performed, using homogenized daily temperature, quality controlled precipitation observational data and gridded products. The estimated changes indicate statistically a significant increase Mediterranean temperature during the summer and a reduction in winter precipitation in specific areas. Several numerical simulation projections suggest that remarkable changes in the climate of the Mediterranean region might occur on the scale of a few decades. A substantial warming (about 1.5°C in winter and almost 2°C in summer) and a significant decrease of precipitation (about 5%) might affect the region in the 2021-2050 period compared to the reference period (1961-1990), in an A1B emission scenario. Physicochemical mechanisms that formulate the air quality over the Mediterranean region and the resulted impacts on the regional climate have been described. The impacts of aerosols, distinguished in direct ( impacts on radiation budget), health (amounts of inhaled particles and impacts on health) and indirect effects (impacts on clouds and precipitation), are discussed on qualitative and quantitative way. Furthermore, a first assessment of the detection of a signal of temperature change over the Mediterranean domain has been performed using HadCRUT3v observation dataset and model outputs from the CMPI3 climate change modeling exercise. For this study we have used new formal detection methodologies developed within the context of the CIRCE project, in order to try to improve the ability to detect a climate change signal at the regional scale. The application of these methods lead to the conclusion that anthropogenic forcing is a plausible explanation for the observed changes in the area mean change of near surface temperature over the Mediterranean region. A first assessment of the detection of a signal of precipitation change over the Mediterranean domain due to internal climate variability has been also given in this section.

The chapter briefly describes available water resources, water demand, and main climatic conditions in Mediterranean area; then it refers on the pathway to create a reliable chain from GCMs to hydrological model and what are the bottlenecks and the sources of uncertainties for the chain. The comparison of downscaling approaches reports that no general rule can be drawn, but these approaches must be tuned to the particular catchment and application.

This chapter discusses results of current and future-projected water cycle components over the Mediterranean region. Results are presented from an ensemble of CMIP3 multi-model simulations (here after referred to as Mariotti) and from the Meteorological Research Institute’s (MRI) 20 km grid global climate model. Referred to as CMIP3 results are surprisingly close to MRI. The projected mean annual change in the rate of precipitation (P) across the region (for sea and land), is projected to decrease by the end of the 21st century by −11% and −10%, respectively, for the MRI and Mariotti runs. Projected changes in evaporation (E) are +9.3% (sea) and −3.6% (land) for JMA runs, compared to +7.2% (sea) and −8.1% (land) in Mariotti’s study. However, no significant difference of the projected change in P–E over the sea body is found between these two studies. E over the eastern Mediterranean was projected to be higher than the western Mediterranean, but the P decrease was projected to be lower. The net moisture budget, P–E, shows that the eastern Mediterranean is projected to become even drier than the western Mediterranean. The river model projects significant decreases in water inflow to the Mediterranean of about −36% by the end of the 21st century in the MRI run (excluding the Nile). The Palmer Drought Severity Index (PDSI), which reflects the combined effects of precipitation and surface air temperature (Ts) changes, shows a progressive and substantial drying of Mediterranean land surface over this region since 1900 (−0.2 PDSI units/decade), consistent with a decrease in precipitation and an increase in Ts (not shown). The last section of this chapter reports on components of the hydrological cycle from five climate model projections for the Mediterranean region. Three of these models have an interactive Mediterranean Sea (MPI, ENEA, Météo-France), and two are versions of the Met Office Hadley Centre regional model (HadRM3-MOSES2, HadRM3-MOSES1) with different land surface schemes. The focus of this section is upon changes in evapotranspiration, and how these changes could be important in controlling available renewable water resources (runoff). These r indicate that rainfall is projected to decline across large areas by over −20% in all of the models, although in the Météo-France model the central part of the northern Mediterranean domain, ie. southern Italy and Greece, has areas of increase as well as decrease. In pockets of Turkey, the eastern Mediterranean, Italy and Spain, projections from the MPI, HadRM3-MOSES2, HadRM3-MOSES1 and ENEA models are for decreases in summer rainfall of −50% or more. Consistent with the global model projections, each of the five high-resolution models simulate increasing temperatures and decreasing evapotranspiration and precipitation for much of the Mediterranean region by the middle of this century. The strongest and most widespread reductions in precipitation are projected to occur in the spring and summer seasons, while reductions in evapotranspiration are greatest in summer.

In this chapter we present the results of the impact assessment on freshwater bodies in the Mediterranean region. Starting from the characterization of the general features of Mediterranean hydrology, main focus is given on large river basins discharging into the Mediterranean sea as well as to small and medium scale catchments representing almost half of the entire discharging basin. Groundwater representing a fundamental water resource for Mediterranean countries was also considered. Climate change impacts on the hydrological behavior of large river basins is investigated through the IRIS computational tool which was proved to be a versatile instrument for both climate studies and the assessment of model ability to simulate the hydrological cycle at catchment scale, taking advantage of the available observed discharge series to evaluate the reliability of future discharge projections. The results regarding some representative Mediterranean rivers using multiple climate models developed inside Circe have highlighted an open spread among twenty-first century projections. The problem of the effective information content of climate model simulations with respect to small scale impact studies is developed at the scale of medium and small catchments. Particularly at the space-time scales needed to describe the terrestrial water cycle in Mediterranean environments this is recognized among the most difficult problems facing both science and society. Therefore downscaling and bias-correction requirements have been treated in this chapter through specific methodologies which integrate dynamical downscaling with statistical downscaling always adopting ground based observation of climate variables as a powerful means to obtain more robust climate forcing for hydrological models. The assessment of climate change impacts on small and medium size catchments is developed through some representative case studies in which downscaling methodologies have been applied thanks to the availability of dense climate measurement networks. The impact assessment of water resources in the Apulia region (southern Italy) revealed a marked increase in the variability of hydrologic regimes as consequence of the increased rainfall variability predicted for the twenty-first century. Conversely only slight decreasing trends were detected in the annual water balance components. Similar results were found on a carbonate aquifer in Southern Italy in which a large Apennine spring have been selected as a significant hydrogeological systems with minimal anthropogenic pressures in the recharge areas. Finally a specific session is dedicated to the role of artificial dams in reducing the possible impacts of climate change. In particular, methodologies for the assessment of optimal dam dimensioning under climate change are presented as well as a reliability assessment based on water supply and demand imbalances.

In this chapter we present the result of two model exercises aiming at simulating the impact of climate change onto two classes of surface aquifers: lakes and rivers. Section 10.1 focuses on the impact of global warming on the thermal structure of two Italian South alpine lakes: Lake Como and Pusiano. Long term hydrodynamic simulations (1953–2050) were performed using the hydrodynamic model DYRESM (Dynamic Reservoir Simulation Model). DYRESM simulations were forced with downscaled regional climate scenarios undertaken within CIRCE. Our model simulations projected a yearly average temperature increase of 0.04°C year⁻¹ for the period 1970–2000 and 0.03°C year⁻¹ for the period 2001–2050 (A1b IPCC scenario). These results are in line with those detected in long term research studies carried out world-wide. This temperature increase is first responsible for a general increase of the water column stability and for a reduction of the mass transfer between deep and surface waters with direct implications on the oxygen and nutrient cycles. The magnitude of the temperature increase is also sufficient to impact on the growth of phytoplankton populations and it is likely one of the concurrent causes promoting the massive cyanobacteria blooms, recently detected in the two Italian case studies and in different lake environments in Europe. Section 10.2 approaches the problem of establishing a methodology to estimate the average yearly nutrient (phosphorus and nitrogen) river loads under present climate conditions and under the forcing of climate change. The case study is the Po River the largest hydrological basin in Italy and the third tributary of the Mediterranean semi-enclosed basin. The methodology developed in this study is based on a hierarchy of different numerical models which allowed to feed the MONERIS model (MOdeling Nutrient Emissions into River System) with the necessary meteorological and hydrological forcing. MONERIS was previously calibrated (1990–1995) and validated (1996–2000) under past conditions and then run under current conditions to define a control experiment (CE). Current nutrient loads have been estimated in 170,000 and 8,000 t year⁻¹ respectively for nitrogen and phosphorus. Approximately 70% of the nitrogen load is from diffuse sources while 65% of the phosphorus load originates from point sources. Nutrient loads projections at 2100 (under different IPCC scenarios) allowed to estimate that both nitrogen and phosphorus loads are strictly dependent on the resident population which is responsible of a 61 and 41% increase respectively for nitrogen and phosphorus. Projected nutrient load variations were found to be negligible when holding the resident population constant. Finally the phosphorus load is markedly influenced by the efficiency of the waste water treatment plants (WWTPs).

This part of the book aims at producing an assessment of the expected variations of the water cycle in the Mediterranean environment due to global climate changes. This is focused on main components of water cycle with particular emphasis of inland that are expected to be highly impacted by climate evolution. The items deal with different scale of investigation using various methods. The eighth chapter focuses with the influence of climate change on water cycle over the whole Mediterranean area, while the nineth and the tenth chapters deal with impact of climate change on water resources from quantitative and qualitative point of view respectively.