The climate of the Venetian and North Adriatic region: Variability, trends and future change

The climate of the Venetian and North Adriatic region: Variability, trends and future change

Physics and Chemistry of the Earth 40-41 (2012) 1–8 Contents lists available at SciVerse ScienceDirect Physics and Chemistry of the Earth journal ho...

898KB Sizes 3 Downloads 72 Views

Physics and Chemistry of the Earth 40-41 (2012) 1–8

Contents lists available at SciVerse ScienceDirect

Physics and Chemistry of the Earth journal homepage: www.elsevier.com/locate/pce

The climate of the Venetian and North Adriatic region: Variability, trends and future change P. Lionello ⇑ CMCC and University of Salento, DISTEBA, Via per Monteroni, km 1-2, Plesso M, 73100 Lecce, Italy

a r t i c l e

i n f o

Article history: Available online 25 February 2012 Keywords: Venice Climate Storm surge Cyclones Waves Sea level

a b s t r a c t This editorial analyzes the evolution of the climate of the Venetian region on the basis of the contributions presented at a workshop that was organized in Venice (27–29 October 2008) by CORILA (COnsorzio RIcerche LAguna) and published in this special issue. The workshop considered past and future evolution of the regional climate, sea level, storminess, and allowed a wide discussion of important scientific results and the identification of existing gaps in the present knowledge. In the Venetian plane an unprecedented warming (3.2 °C/century) and a moderate decrease of annual precipitation ( 3%/century) are expected at the end of the 21st century, with no analogy in the past 250 years during which there was no sustained centennial trend. The understanding of past sea level evolution is in part problematic. The analysis of the Venice tide gauge time series (and its comparison with that of Trieste) suggests a centennial trend of relative sea level rise (about 1.1 mm/year) comparable to, but smaller than, the global sea level trend. However, past relative sea level in Venice has been strongly affected by tectonic motions and isostatic adjustment. If their estimated effects are subtracted from the tide gauge observation, the sea surface height in Venice would show a centennial trend (0.3 mm/year) that is much smaller than the global mean value. Unless a physical explanation for this low value is found, estimates of vertical land motions for this century need to be reconsidered. Future evolution of sea level is uncertain. Glaciers and ice sheet melting, its regional implications, regional steric effects associated with changes of temperature and salinity are all expected to be important in the future and are not adequately known. A large future halosteric contribution is peculiar for the Mediterranean Sea, where future increased salinity and consequent contraction of the water column could compensate for water mass addition and thermosteric expansion. The time series of storminess is dominated by large interannual and interdecadal variability and there is no evidence of its past or future changes on centennial time scale. Relative sea level trends are very likely to be the main cause of future changes of flood frequency and height, which will, anyway, continue to be strongly affected by large interannual and interdecadal fluctuations. Ó 2012 Elsevier Ltd. All rights reserved.

1. Introduction Venice and the environment of the Venetian lagoon are the result of a complex interaction between a peculiar morphology and human action. The natural landscape, which existed here in the Roman times has been at the same time preserved and modified by human action, which produced a peculiar civilization. The morphology of the lagoon presented a unique opportunity for the establishment of a unique city, which for its own preservation and safety controlled and maintained an environment that would have otherwise evolved differently over centuries and, eventually, disappeared. The Venetian lagoon is in fact the largest remnant of many similar water extensions that existed here during Roman times. Whether in the future it will be possible to maintain this situation and continue using the past management practice is not ⇑ Tel.: +39 0832 297085. E-mail address: [email protected] 1474-7065/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. doi:10.1016/j.pce.2012.02.002

clear. The ongoing construction of the movable barriers closing the lagoon inlets shows that new strategies have been decided. The long (and still ongoing) debate on the optimal solutions shows clearly how difficult it is to identify an effective policy on multidecadal time scale. Difficulties arise from the changing environment and climate (particularly sea level), from new societal needs, and from the consequences of previous human induced changes. Climate change, which is a main focus of this extended editorial, is a new factor, whose action is necessary to take into account. This special issue stemmed from the workshop ‘‘The climate in the Venetian and North Adriatic region: variability, trends and change’’ which was organized by CORILA (COnsorzio Ricerche Laguna) in Venice, 27–29 October 2008. The workshop aimed at understanding the factors that are important for the evolution of the Venetian environment and considered a wide set of topics: historical climatology and climate of the past, northern Adriatic sea level (its variability and trends, its extremes, land motions affecting it), storminess (cyclones, waves and storm surges), regional climate

2

P. Lionello / Physics and Chemistry of the Earth 40-41 (2012) 1–8

change. The presentations held during the workshop have evolved in the articles that are published in this special issue. The analysis of future climate scenarios at regional scale and their comparison with past variability and trends are fundamental topics. It is important to separate natural variability, which has been active during the last centuries, from the trend that most records show in the last decades, and to produce reliable scenarios of climate future evolution. Essential climate variables, such as temperature and precipitation, have fundamental effects on the environment, e.g. on the hydrographic characteristics of the lagoon and on its ecosystem, and also influence sea level, because of their effect on water density. Section 2 contains a synthesis of the articles that are published in this special issue on the evolution of temperature and precipitation and summarizes the present knowledge for the Venetian region. Relative sea level is a variable crucial for Venice, the lagoon and the whole flat coastline of the northern Adriatic. Relative sea level changes can be considered as the difference between vertical land motion (active at multiple space and time scales) and changes of sea surface height. Changes of sea surface height are associated with regional changes of mass and volume distribution of the water in the sea, which are caused by changes of sea circulation, temperature, salinity, and are, ultimately, due to the interaction with the atmosphere. The important (and negative) effect of past relative sea level rise is well known and shows how critical even small changes can be for a city whose pedestrian level is mostly about 1 m above mean sea level, with a tidal range from 50 cm (neap tide) to 100 cm (spring tide) on top of which storm surges frequently occur. A small increase of mean relative sea level would dramatically increase the frequency of floods and also change the speed of the erosion of the lagoon (see Section 4 and Plag et al., 2006). The ongoing global sea level increase and the expectation of its acceleration are causes of great concern and could imply a radical future change in the management of the lagoon. Section 3 contains a synthesis of articles that are published on this topic in this special issue and a short discussion of issues that are relevant for regional sea level evolution. Marine storminess and storm surges are ultimately the causes of present floods of Venice and nowadays the main hazard to the city. Monitoring their evolution and analyzing their future evolution is important, because increase of their frequency or changes of their extreme intensity would have important consequences for the management and planning of coastal defenses. The articles published in this special issues on this topic and a discussion of Venice flood frequency in future climate scenarios are contained in Section 4. General comments, discussion and a final synthesis are contained in Section 5.

2. Past and future evolution of temperature and precipitation This special issues contains three manuscripts dealing with analysis of temperature and precipitation:  Projecting North Eastern Italy temperature and precipitation secular records onto a high resolution grid (here referred to as Brunetti et al., 2010).  Recovery of the Early Period of Long Instrumental Time Series of Air Temperature in Padua, Italy (1716–2007) (here refereed to as Camuffo and Bertolin, 2010).  Regional Climate change in the northern Adriatic (here referred to as Zampieri et al., 2010).  These papers consider different aspects related to the analysis of temperature and precipitation, and of their variation in space and time in the Venetian region.

Brunetti et al. (2010) discuss how to produce a sequence of high resolution (30-arc s) monthly temperature and precipitation fields for North Eastern Italy, by superimposing a highly resolved mean climate field to a coarse monthly anomaly. This resolution is much higher than that presently attainable by climate models and by model re-analysis. The background concept of this study is that the monthly anomaly fields have a spatial correlation scale sufficiently large that they can be estimated by a relatively coarse network of station data, while the fixed-in-time mean climate field (accounting for local morphological features such as level, slope, and distance from the sea) is computed using a dense network. The high resolution fields that are obtained provide information very useful for environmental applications. Camuffo and Bertolin (2010) describe the results of an accurate reanalysis of a instrumental temperature time series that was recorded over almost three centuries (1716–2007) in Padua, in the Venetian plane, about 40 km from Venice. The authors obtain a time homogeneous record, removing the discontinuities produced by the change of instrumentation, its deterioration or change of observational practice. This is an example of the precious opportunities for accurate description of the past climate that this part of the world (where the longest instrumental time series exist) offers. This careful work, when combined with the reconstruction of precipitation (Camuffo et al., 2010), allows the reconstruction of past climate evolution over almost three centuries. The analysis is suggestive of some small periodic variability of the regional climate with time scales of 35.8 year and 23.9 year, whose presence represents an interesting problem of past climate dynamics. Zampieri et al. (2010) describe a set of model experiments to examine the future evolution of regional climate conditions in a large area centered on the Venetian region. Multi-model and multi-scenario simulations are included allowing to discuss how the choice of the model affects the results and how climate change signal depends on the level of emission. These models, validated against the available information on the present climate, are the unique tool for analyzing the future evolution of climate that is caused by the increase of greenhouse gases. Results suggest a warming, which is larger in the A2 scenario (about 5.5 K in summer and 4 K in winter) than in the A1B (Nakicenovic et al., 2000). Precipitation is projected to have contrasting trends between winter and summer (positive, +0.5 mm/day and negative, 1 mm/day, over the Alps, respectively). The climate change signal is much stronger for temperature than for precipitation, and for scenario A1B in the period 2021–2050, it is significant for temperature, but not yet for precipitation. Some further analysis is shown here on the basis of the data produced by Camuffo and Bertolin (2010) and Zampieri et al. (2010). Fig. 1a and b concern the capability of models to reproduce the observed annual mean temperature and total precipitation value in Padua for the period 1961–1990. In both panels the vertical thin red and blue1 lines represent individual models of the ENSEMBLES (Hewitt and Griggs, 2004) and PRUDENCE (Christensen et al., 2007) projects, respectively. The thick black line represents the observed value. The ENSEMBLE (PRUDENCE) model results are interpolated using a Gaussian red (blue) curve in order to help visualizing their distribution with respect to the observed value, which is inside the spread of the models and not far from their mean value for both temperature and precipitation. Statistics are based on 11 simulations of the ENSEMBLES project and 27 simulations of the PRUDENCE project. Note that the Padua time series are well correlated (the value is larger than 0.8 for both annual temperature and total precipitation) with the average of the E-OBS observational dataset (Haylock et al.,

1 For interpretation of color in Figs. 1–6, the reader is referred to the web version of this article.

P. Lionello / Physics and Chemistry of the Earth 40-41 (2012) 1–8

Fig. 1. Comparison of model and observations in Padua. (a, top) Annual mean temperature for the period 1961–1990: observed value (thick black line), values of the ENSEMBLES (vertical thin red lines) and PRUDENCE (vertical thin blue lines) simulations, The probability distribution of the ENSEMBLE (PRUDENCE) model results is interpolated using a Gaussian red (blue) function. (b, bottom) Same as (a) except it shows annual accumulated precipitation.

2008) over the whole Venetian region and, therefore, this location is representative of this region. Fig. 2a and b merge the observed time series of Camuffo and Bertolin (2010) with the model simulations analyzed in Zampieri et al. (2010), referring to temperature and precipitation in Padua, respectively. In this way the evolution of climate over an almost 400 year long period is produced. Time series show the anomaly of annual values with respect to the mean for the 1961–1990

3

period. The thin gray line shows the observed annual values, the thick black lines their running mean over a 51-year long time window. The overlapping thin colored lines represent individual model simulations of the ENSEMBLES (red) and PRUDENCE (blue) models considered in Zampieri et al. (2010). The thick colored lines show their running mean over a 21 year long period. The windows of the running means have been chosen ad hoc for obtaining a similar variability in models and observations. The time series show that what the model predict for temperature is an unprecedented behavior and that what has been observed in the last few decades is only a tiny fraction of the future warming. Note that the models (which include no periodic radiations forcing) do not reproduce the oscillations described in Camuffo and Bertolin (2010), which are, anyway, absent in the second half of the 20th century. The green lines show the spatial average of the E-OBS data (based on observations) in the area, for which temperature is actually increasing faster than in the model simulations during the 1st decade of the 21st century. It is impossible to decide from this analysis whether the larger positive trend in the E-OBS observations results from a decadal fluctuation or from a systematic undervaluation of regional warming by models. The analysis of the 21st century temperature trend in Padua in the ENSEMBLES simulations provides a value of 3.2 °C/century. Fig. 2b shows the annual precipitation, which decreases in the 21st century further extending a tendency already present in the last part of the observed record. The linear interpolation of the precipitation decrease in Padua for the ENSEMBLES model provides a rather small value, (3%/century), which is, however, the result of contrasting seasonal trends positive in winter and negative in summer (Zampieri et al., 2010). The green line represents the E-OBS values, which have stronger negative trends than the model ensemble. Obviously, also for precipitation as for temperature, the discrepancy between E-OBS and model time series does not necessarily mean that models underestimate the speed of climate change, because the larger observed trends could be produced by a decadal oscillation. Note that Fig. 2a and b shows that what has been observed so far is a small fraction of the climate change that according to models is going to happen in the 21st century.

Fig. 2. Climate evolution in the venetian region from the 18th to the 21st century. (a, left) Mean annual temperature anomaly with respect to the 1961–1990 period. The thin gray line shows the observed annual values, the thick black lines their running mean over a 51-year long time window, thin overlapping colored lines represent individual model simulations of the 11 ENSEMBLES (red) and 27 PRUDENCE (blue) models considered in Zampieri et al. (2010). The thick colored lines consider their running mean over a 21 year long period. The green line shows the E-OBS observations (the thin line shows the individual values and the thick line their running mean). (b, right) Same as (a) except it shows the ratio of annual precipitation with respect to its mean value for the 1961–1990 period.

4

P. Lionello / Physics and Chemistry of the Earth 40-41 (2012) 1–8

3. Relative sea level and its future evolution Three papers published in this special issue address important issues concerning the relative sea level:  Sea level variability and trends in the Adriatic Sea in 1993–2008 from tide gauge and satellite altimetry (here referred to as Fenoglio-Marc et al., 2011).  Recent developments in understanding sea level rise at the Adriatic coasts (here referred to as Tsimplis et al., 2009).  Resolving land subsidence within the Venice Lagoon by persistent scatterer SAR interferometry (here referred to as Teatini et al., 2010). The relative sea level, which is the variable actually relevant for coastal management in the Adriatic, results from the combination of vertical land motion (including local subsidence) and sea surface height variations. The analysis of its past and the prediction of its future are complex multidisciplinary problems, which involve several independent factors. Teatini et al. (2010) show the potential importance of variability at local scale, within different parts of the Venetian lagoon and of Venice itself. Using a technique based on SAR (Synthetic Aperture Radar) satellite interferometry, the authors have analyzed the period from 1992 and 2007, and identified values of land subsidence from less than 1 mm/year to 5 mm/year, with some points reaching also values larger than 10 mm/year in the areas around the lagoon. This technique does not show large vertical motion in the central part of the lagoon (which is now substantially stable) and, therefore, sea level observations in Venice city center should not be affected anymore by large subsidence. Fenoglio-Marc et al. (2011) compare multi-satellite altimetry observations and tide gauge records during a 16 year long period (1993–2008) and show very coherent interannual variability among the central and coastal part of the Adriatic Sea. The difference between the two datasets is used for an estimate of the vertical land motion at the coast, identifying uplift at the west coast and subsidence at the eastern coast. Particularly important for the discussion in this section is the land motion difference between Venice ( 0.2 ± 1.3) and Trieste (1.3 ± 1.1 mm/year). According to altimeter data at basin scale, the sea surface height has increased by 3.1 ± 0.3 mm/year over the time interval 1993–2008. Tsimplis et al. (2009) review recent trends of factors responsible for sea level changes in the northern Adriatic. The study shows clearly the importance of regional air–sea interaction. The mechanical action of the atmosphere accounts for 0.8 mm/year for the period 1960–2000. The magnitude of the steric trend for the same period is uncertain, because of uncertainty in the past evolution of water temperature and salinity in the northern Adriatic and it depends on the depth used for its estimation. It ranges between 0.4 mm/year (when the upper 100 m are used) to 2.4 mm/year (when the whole water column is used). The removal of mechanical action exerted by the atmosphere and steric effect suggests the presence of a residual unexplained trend, with a minimal values of about 1 mm/year and maximum of 4.4 mm/year, which the authors attribute to a combination of vertical land motion and mass exchange with the Atlantic Ocean across the Gibraltar strait (this could be a regional evidence of mass addition to the global ocean by ice melting). These results can be further discussed considering Fig. 3. Fig. 3a shows the observed time series of sea level anomaly with respect to the average value of the 1961–1990 period for the global mean (Church and White, 2011, violet line), the Trieste and Venice tide gauges (blue and red lines) and the average altimeter data (green line, after 1992) in the Adriatic. Fig. 3a suggests that the sea level

Fig. 3. Observed values of sea level anomaly with respect to the mean for the 1961– 1990 period. (a, top) Global mean (Church and White, 2011, violet line), Trieste and Venice tide gauge (blue and red lines respectively), altimeter average data in the Adriatic (green line). (b, middle) Venice tide gauge time series with the liner trends computed for the periods 1975–2000 (black line), 1950–2000 (dark gray line), 1925–2000 (light gray line). The panel includes also the global and Trieste trends (violet ad blue respectively) or the 1910–2000 period. (c, bottom) Same data except after applying linear corrections for accounting vertical land motion in Venice and Trieste. In both panels the thick lines superimposed to tide gauge time series represent a 11-year running mean. The linear trends for the period 1910–2000 for the global sea level (violet) and the Venice (red) corrected sea level are also reported.

in Trieste has been following the global trend with large transient departure at both interannual and interdecadal time scales. This apparently holds also for Venice, but with a further strong local subsidence in the period 1925–1970. If this transient subsidence is subtracted from the Venice time series, the average of Venice and Trieste provides a centennial trend of 1.1 mm/year. Fig. 3a confirms the agreement between tide-gauges and satellite altimeter data found by Fenoglio-Marc et al. (2011). The presence of large interannual and interdecadal fluctuations can be considered in the light of the regional air–sea interaction processes discussed by Tsimplis et al. (2009). Fig. 3b shows the annual data in Venice and the linear trends for the periods 1975–2000, 1950–2000, 1925–2000, whose values are 1.0 mm/year, 1.6 mm/year, 2.7 mm/year. The Venice time series can be compared with the global and Trieste trends for the 1910–2000 period, which are shown in Fig. 3b. The difference among these values shows the importance of local land movements in Venice and their variability in time and illustrates the difficulty of estimating sea level increase from analysis of relative sea level time series in Venice. Since in Venice a long time series is affected by large and irregular land movements, a short time series is affected by interdecadal variability, and new factors are playing an essential role in future, these past relative sea level trends cannot provide a reliable estimate of future relative sea level evolution, unless they are interpreted on the basis of a plausible physical model. However, the agreement between global and local sea level trends in Trieste is problematic. Vertical land motion can be

P. Lionello / Physics and Chemistry of the Earth 40-41 (2012) 1–8

estimated from geological evidences and models of isostatic adjustment. In Venice, geological records comparing with previous interglacial periods suggest a tectonic motion, with a vertical downward speed of about 0.7 mm/year (Antonioli et al., 2009). Glacio isostatic models show that there is still a modest upward motion (which is the consequence of the melting of the Alpine glaciers in the last ice-age), for which the best estimate is 0.1 mm/year (Spada et al., 2009). These factors produce, also in absence of any change in the sea surface height, a relative sea level rise at a rate of 0.6 mm/year. In Venice this is superimposed with a well documented local human induced land subsidence for the period 1925–1970 (Carbognin et al., 1977). In this editorial, a total loss of 15 cm from 1920 to 1970 has been adopted. Fig. 3c shows the Venice time series after subtracting a linear trend including all these processes. Therefore in Fig. 3a the red line represent the evolution of relative sea level and in Fig. 3c the estimated sea surface height, including steric, atmospheric and mass induced changes, but after having subtracted land vertical motions. The blue line represent the Trieste time series after correcting it for a steady linear trend of 0.8 mm/year (empirically estimated with the aim of overlapping it with the Venice time series), which could be explained by a slightly different vertical motion in Trieste with respect to Venice. After these corrections the sea level in Venice and Trieste are extremely coherent, both at the annual and decadal time scale, further confirming the homogeneity of mean sea level variation in the northern Adriatic. However, it should be noticed that the correction for local land subsidence in Venice does not completely agree with Carbognin et al. (1977), who estimated 11 cm, and, further, the differential vertical motion between Venice and Trieste does not agree with the estimate by Fenoglio-Marc et al. (2011), who found opposite trends with subsidence in Venice and uplift in Trieste. Therefore, the estimated land movements used in Fig. 3c are not free of problems, but if they are accepted, they would imply that the Adriatic sea level during the 20th century had a sustained trend (0.3 mm/year) much lower than the global one (1.6 mm/year). Either the vertical land movements (that have been estimated on geological time scales) have been suppressed during the 20th century or regional factors (systematic trends of atmospheric pressure, sea temperature, salinity) have determined a sea surface height trend that is quite different from the mean global value. In any case, this analysis shows the need to further analyze the possibility of a northern Adriatic trend being different from the global one on multi-decadal and even centennial time scale. A crucial point for the low coast of the Adriatic Sea and the city of Venice is the future evolution of sea level. This point is largely controversial. Uncertainties are large on the global mean future sea level, on how this will be regionally distributed, on regional factors (such as air–sea interaction) determining deviations from the global average. For the mass contribution due to melting of glaciers, ice caps, ice sheets of Greenland and Antarctica, IPCC (2007) has provided a range from 4 to 23 cm by the end of the 21st century. However, present observations of GHG emission, global temperature and sea level show that the observed rate of climate changes follows a path close to, or even possibly above, the largest rate of change that was estimated by IPCC (Rahmstorf et al., 2007; Church et al., 2011). Therefore it can be reasonably expected that the IPCC (2007) range is a conservative estimate. In fact, an acceleration of ice sheet melting, which would provide a sea level rise higher than previously estimated, has already been observed (e.g. Alley et al., 2005; Rignot et al., 2011). The global sea level rise due to temperature has been estimated in the range from 10 to 41 cm (IPCC, 2007, 5–95th percentile). This might be a conservative estimate as well, though it appears more robust than that of the mass contribution. These global values do not apply directly to the northern Adriatic Sea. First, changes of the earth gravitational field caused by ice sheet melting would produce a non-uniform geographical

5

distribution of added mass of water and imply regional differences in sea level change (Church et al., 2011; Plag, 2008). Second, steric changes have a strong regional connotation. According to the 4th IPCC report (Meehl et al., 2007) and considering only thermal expansion, the median spatial standard deviation of thermal expansion in model projections is 0.08 m, which is about 25% of the central estimate of global average sea level rise during the 21st century under A1B. Moreover, the estimate of steric effects for the Mediterranean Sea is uncertain, because its poor representation in global simulations adds to the usual uncertainties of projections (note that the Mediterranean Sea in not included in the sea level maps of the 4th IPCC report). Marcos and Tsimplis analyzed a set of 12 global model simulations for the AIB and A2 scenarios and obtained a thermosteric contribution in the range from 1 to 38 cm at the end of the 21st century, which is very similar to the range of global values proposed by IPCC (2007). Future evolution of salinity is a major source of uncertainty. While it is clear that salinity will increase, because of the increased water deficit due to increased evaporation and decreased precipitation (Giorgi and Lionello, 2008), it is not clear how much it will change the sea surface height in the Mediterranean. While thermosteric effects could occur without any change of mass, halosteric effects require a change of the water deficit in the basin and, therefore, could alter the present water exchange across the Gibraltar strait. If the increased water deficit would be exactly compensated by the flow across the Gibraltar strait and the water mass in the Mediterranean Sea will remain constant, increased salinity would imply a sea level decrease between 0 and 36 cm, overcompensating the thermosteric expansion. However, the functioning of the Gibraltar strait and the baroclinic exchange across it in future climate conditions are to a large extent unknown. Accelerated ice melting, increased water deficit, regional warming (and new water mass characteristics) are new factors that are likely to bring the northern Adriatic Sea to unprecedented conditions. In this changing scenario extrapolating to the future the present sea level is arbitrary, very doubtful and it can be used just as a reference against which to evaluate future tendencies. Fig. 4 shows how the expected values and uncertainty of the Mediterranean Sea surface height at the end of the 21st grow as different contributions are taken into account. The linear trend for the period 1910–2000 shown in Fig. 3c (3b) would imply a sea level rise of 3 cm (11 cm) in 100 year, which is marked with the gray (black) horizontal line. Note that here only the sea surface height is considered and for the analysis of impacts the relative sea level must be considered, including vertical land movements. The uncertainty due to the glaciers and ice-sheet melting contribution is from 4 cm to 23 cm. It increases to a range from 17 to 79 cm if thermosteric effects at average Mediterranean scale are accounted for. If also halosteric effects are included, then the uncertainty bar grows from 19 to 79 cm, making even a decrease of future sea level possible (Marcos and Tsimplis, 2008). However, it must be noted that the assumption of a constant Mediterranean total water mass is controversial and is just a reasonable over-simplification of future conditions. The evolution of the water masses and total mass without accounting properly for the mass exchange across the Gibraltar strait is not accurate. Recent estimate based on a high resolution model of the Mediterranean Sea ( 48 cm, Sevault et al., 2009) and with a coupled (Artale et al., 2009) atmosphere–ocean model (1.4 cm for the period 1951–2050, reported in Planton et al., 2011) are not conclusive, but they confirm the importance of salinity changes for future Mediterranean Sea level. The mechanical action of the atmosphere, which has been responsible for partially compensating the sea level increase during the 20th century (Tsimplis et al., 2009) is not included in Fig. 4, and it might imply a further decrease of sea surface height. Finally, as this discussion

6

P. Lionello / Physics and Chemistry of the Earth 40-41 (2012) 1–8

Fig. 4. Conceptual scheme showing the uncertainty of future sea level increase (yaxis, cm) in the Mediterranean Sea. The red bar shows the uncertainty range of global sea level rise due to Glaciers and Ice Sheets melting (GIC). The blue line (GIC + T) the uncertainty obtained adding the regional thermosteric contribution and the yellow line (GIC + T + S) the uncertainty obtained further adding the halosteric contribution (Marcos and Tsimplis, 2008). The horizontal black/gray lines represent the sea level change that would be obtained by extrapolating the present trend shown in Fig. 3b and c until the end of the 21st century. Uncertainties due to changes of the Earth gravitational field, to the regional mechanical action of the atmosphere and to a large melting of Greenland and Antarctic ice sheets are not included.

has concerned the mean level of the Mediterranean Sea and not of the northern Adriatic Sea, there is a further source of uncertainty. Fig. 4 does not mean to describe the most extreme relative sea level rise scenarios in the northern Adriatic. An even wider uncertainty has been computed leading to estimates in the range from 47 cm to 148 cm at the end of the 21st century (Plag et al., 2006) when extreme conditions, such as accelerated melting of Greenland and West Antarctica are considered. The meaning of Fig. 4 is essentially to stress uncertainties on future evolution, to show how much the future is likely to be different from the past, and how unreliable a prediction of the future based on direct and unquestioned extrapolation of observed trends would be. 4. Storminess characteristics, its future evolution and storm surges

Matulla et al. (2011) show that cold season (October–March) storminess in northern Italy and the northern Adriatic Sea since 1760 is characterized by pronounced interannual and interdecadal variability, but no sustained long-term trend. The analysis is based on the geostrophic wind derived from homogenized daily mean pressure series in Genoa, Milan, Padua, Turin, and Hvar. Lionello et al. (2010a) confirm the absence of important sustained trends of severe marine storminess in the northern Adriatic during the second half of the 20th century by the analysis of hourly sea level time series and significant wave height records. There are some small negative trends, but the time series are dominated by large interannual variability. Further, this study shows that cyclones producing extreme storm surges differ from those producing high waves, and both have specific characteristics that distinguish them with respect to other cyclones passing over northern Italy. Lionello et al. (2010b) show scenario climate projections for storm surges and wind waves in the northern Adriatic Sea. In general extreme storms are stronger in future scenarios, but differences are not statistically significant and this study does not provide a clear evidence for an effect of climate change on Mediterranean extreme storminess. Fig. 5 shows the time series of the storminess indexes described in Matulla et al. (2011) and the value of the 99.9 percentile (99.9p) of observed hourly sea level (It considers the residual after subtracting the astronomical tide and the annual mean sea level.). Note that the 99.9p is a threshold that is surpassed on average only for about 9 h in a year and therefore it represents an extreme value of sea level. It is evident that there is a very poor correlation between the 99.9p and the indices in Matulla et al. (2011). Therefore, although the Matulla et al. (2011) analysis represents well the evolution of storminess over northern Italy, it is uncertain whether it can be associated with the storm surge frequency in Venice. Lionello et al. (2010a) show that the systems producing the flood of Venice have peculiar tracks and area of cyclogenesis. This result suggests that it may be difficult to relate an extreme storm surge in Venice to general regional indices and it provides a background for the lack of correlation between 99.9p and the Matulla indices. Note that the lack of large multi-decadal oscillations and trend in the 99.9p is consistent with the behavior of the time series of extreme surges shown in Lionello et al. (2010a) and other past studies (Lionello, 2005; Barriopedro et al., in press), which eventually show a link with the 11-year sun spot cycle, but no other large multi-decal variability or trend. Scientific literature shows no convincing evidence that future storminess will have characteristics much different from the present one (Lionello et al., 2003, 2007, 2008, 2010b). Moreover, sea

Three papers in this special issue analyze storminess and storm surges.  Storminess in northern Italy and the Adriatic Sea reaching back to 1760. (Matulla et al., 2011).  Extreme storm surge and wind wave climate scenario simulations at the Venetian littoral (Lionello et al., 2010a).  Marine storminess in the northern Adriatic: characteristics and trends (Lionello et al., 2010b). Two papers concerns the past (Matulla et al., 2011) and the future (Lionello et al., 2010b) of storminess. A third analysis present evolution and the link of storm surges to cyclones (Lionello et al., 2010a). These papers suggest that intense storminess has not shown centennial trends in the past, it will not change in the future and that interannual and interdecadal variability is the main feature to be accounted for while describing its time series.

Fig. 5. Time series of the storminess indices ‘‘GPT’’ (thin black line) and ‘‘PHG’’ (thin yellow line) as in Fig. 2 of Matulla et al. (2011) and of the 99.9p (thin red line) of observed hourly sea level (after subtracting the astronomical tide and the annual mean sea level). The 99.9p has been normalized with the standard deviation of the hourly values. The thick lines show the corresponding 11-year running means.

P. Lionello / Physics and Chemistry of the Earth 40-41 (2012) 1–8

7

Fig. 6. Frequency of hourly sea in Venice. (a, left) Hourly sea level probability density function (black line, black labels on x-axis) and cumulative probability distribution (pink line and pink labels on y-axis) for the period 1940–2010. The probability density function has been computed using 1 cm bins and averaging values on seven contiguous bins. (b, right) Annual average total duration as function of the threshold hourly sea level (y-axis, cm) and of sea level rise (x-axis, cm) using the distribution shown in (a). Label on contour lines denotes hours (h), days (d), weeks (w) and months (m).

level increase will have a negligible effect on Adriatic astronomical tides (Lionello et al., 2005). Therefore, a tentative estimate of future hourly sea level frequency can be based on applying the effect of relative sea level rise to the present hourly distribution, following the reasoning first applied in Plag et al. (2006). Fig. 6a shows the statistical distribution of hourly sea level (after subtracting the astronomical tide) with respect to the annual mean value for the period 1940–2010. Fig. 6b shows the average number of hours per year above a given threshold (y-axis) as function of sea level rise (x-axis). For a fixed frequency, the corresponding threshold level grows linearly with sea level. Consider the 95 cm and 105 cm sea level, representing values at which approximately 29% and 54%, respectively, of Venice is flooded in the first decade of the 21st century. These values are now reached for about 1 and 3 h, respectively, every year. A 25 cm relative sea level rise would imply that they are reached for more than 0.5 days and 1.5 days, respectively. A 50 cm relative sea level rise would imply that they are reached for about 6 days and 2 weeks, respectively.

5. Conclusions This extended editorial discusses the evolution of the climate of the Venetian region (with a focus on sea level) on the basis of the manuscripts published in this special issue and of the discussion during the workshop organized by CORILA in Venice, October 2008. The analysis of annual temperature and precipitation provides a consolidated perspective: climate models compare satisfactorily with observed annual mean temperature and accumulated precipitation values; the recent warming is unprecedented during the last three centuries; climate change will lead to completely different temperature conditions during the 21st century. Precipitation on annual basis will change only marginally in the Venetian plane, but contrasting positive/negative winter/summer trends are large (Zampieri et al., 2010). These new climate conditions are very likely to produce large changes on the environment (with consequences on society), on the temperature and salinity of the lagoon (with important effects on its ecosystem) and on water masses of the North Adriatic (with important effects on sea level). Observations (Camuffo and Bertolin, 2010) show past oscillations with

periods of 35.8 and 23.9 years, whose source is unexplained yet. Models do not contain a mechanisms for reproducing these past oscillations, which, anyway, are small with respect to recent trends and future climate change. Precipitation is more problematic and its interseasonal variability is not adequately reproduced by models, which falsely attribute to the Venetian region a Mediterranean climate with a low summer minimum (Zampieri et al., 2010). Recent observations provide a weak indication that models underestimate the pace of climate change and both warming and precipitation decrease are progressing more quickly than projected. The understanding of sea level evolution is less consolidated than for temperature and its future is uncertain. It is, however, clear that unprecedented changes are likely, caused by both regional (steric effects) and global remote factors (accelerated ice-sheet melting in the polar regions). Uncertainty in future evolution is large and very relevant for the vulnerable condition of Venice and its lagoon. A large and dramatic increase cannot be excluded and it is considered likely by some scientists (Rahmsdorf et al., 2007, Plag et al., 2010). In the Mediterranean Sea the prediction of future sea level is particularly complex, because the effects of thermosteric expansion, ice-sheet and glacier ice melting could be compensated by halosteric contraction due to increased sea salinity. Continuous monitoring during the next decades and improved modeling of sea level are very important for timely adopting the right coastal management strategies. Considering the large uncertainties of century-scale predictions, the need for short-term forecasting on decadal time scales has been emphasized (Plag et al., 2010). Also understanding of past evolution is not free of problems. The analysis in this editorial suggests that during the 20th century either the sea surface height in the North Adriatic has been rising much more slowly than the global value or current estimate of land movements need to be revised and tectonic motion has not contributed significantly for the last 100 years. Though air–sea interaction can play a substantial role (Tsimplis et al., 2009) the possibility that it has substantially reduced the centennial sea surface height trend to the extent of solving this problem is not documented in any published study. Uncertainty is large not only for future but also for past steric effects, particularly for the role of salinity. High resolution modeling of Mediterranean Sea, accurate

8

P. Lionello / Physics and Chemistry of the Earth 40-41 (2012) 1–8

representation of its interaction with the atmosphere, understanding of the role of Gibraltar strait on connecting global to regional sea level change are all important for predicting future sea level. There is an evident need of society for reducing the large uncertainty of science on this issue. Venice and the lagoon are very sensitive to sea level extremes. Storminess has been always characterized by large interannual and interdecadal variability and long term sustained trends have been not relevant in the past (Matulla et al., 2011; Lionello et al., 2010a). Climate change simulations do not provide evidence of neither intensification nor attenuation of marine storms (Lionello et al., 2010b). The analysis of the results of this special issue suggests that storm surge in the northern Adriatic is a nonlinear process that is difficult to relate to general regional indicators. However, considering the present statistics of storm surges to be valid also in future, even a small relative sea level rise would be capable of a large change in the frequency of floods, with important consequences for the management of the Venetian coastlines and lagoon. Acknowledgements I am grateful to M. Zampieri, C. Bertolin and D. Camuffo for the data that have been used in Figs. 1 and 2. L. Fenoglio-Marc, the center for Permanent Mean Sea Level, Centro Previsioni Maree and A. Tomasin for the time series used in Fig. 3, O. Krueger for the data used in Fig. 5. I thank indeed F. Antonioli, D. Gomis, H.-P. Plag and M. Tsimplis for important comments on the content of this editorial. CORILA is acknowledged for having organized the workshop that has constituted the premise for this special issue and the MedCLIVAR project for the support provided to the initiative. References Alley, R.B., Clark, P.U., Huybrechts, P., Joughin, I., 2005. Ice-sheet and sea-level changes. Science 310, 456–460. Antonioli, F., Amorosi, A., Correggiari, A., Doglioni, C., Fontana, A., Fontolan, G., Furlani, S., Ruggieri, G., Spada, G., 2009. Relative sea-level rise and asymmetric subsidence in the northern Adriatic. Rendiconti Soc. Geol. It. 9, 2–5. Artale, V., Calmanti, S., Carillo, A., Dell’Aquila, A., Hermann, M., Pisacane, G., Ruti, P., Sannino, G., Struglia, M., Giorgi, F., Bi, X., Pal, J., Rauscher, S., 2009. An atmosphere–ocean regional climate model for the Mediterranean area: assessment of a present climate simulation. Clim. Dyn. 35 (5), 721–740. doi:10.1007/s00382-009-0691-8. Barriopedro, D., Garcia-Herrera, R., Lionello, P., Pino, C., in press. A discussion of the links between solar variability and high storm surge events in Venice. J. Geophys. Res. doi:10.1029/2009JD013114. Brunetti, M., Lentini, G., Maugeri, M., Nanni, T., Simolo, C., Spinoni, J., 2010. Projecting north eastern Italy temperature and precipitation secular records onto a high-resolution grid. Phys. Chem. Earth (this special issue). Camuffo, D., Bertolin, C., 2010. Recovery of the early period of long instrumental time series of air temperature in Padua, Italy (1716–2007). Phys. Chem. Earth. Camuffo, D., Bertolin, C., Diodato, N., Barriendos, M., Domínguez-Castro, F., Cocheo, C., Valle, A., Garnier, E., Alcoforado, M.J., 2010. The western Mediterranean climate: how will it respond to global warming? Clim. Change 101 (1–2), 137– 142. doi:10.1007/s10584-010-9817-6. Carbognin, L., Gatto, P., Mozzi, G., Gambolati, G., Ricceri, G., 1977. New trend in the subsidence of Venice. In: Rodda, J.C. (Ed.). Land Subsidence. IAHS Publ. No. 121, Washington, DC, USA, pp. 65–81. Christensen, J.H., Carter, T.R., Rummukainen, M., Amanatis, G., 2007. Prediction of regional scenarios and uncertainties for defining European climate change risk and effects: the PRUDENCE project. Clim. Change 81. Church, J.A., White, N.J., 2011. Sea-level rise from the late 19th to the early 21st Century. Surv. Geophys.. doi:10.1007/s10712-011-9119-1. Church, J.A., Gregory, J.M., White, N.J., Platten, S.M., Mitrovica, J.X., 2011. Understanding and projecting sea level change. Oceanography 24 (2), 130–143. Fenoglio-Marc, L., Braitenberg, C., Tunini, L., 2011. Sea level variability and trends in the Adriatic sea in 1993–2008 from tide gauges and satellite altimetry. Phys. Chem. Earth (this special issue). Giorgi, F., Lionello, P., 2008. Climate change projections for the Mediterranean region. Global Planet Change 63, 90–104. Hewitt, C.D., Griggs, D., 2004. Ensembles-based predictions of climate changes and their impacts. EOS AGU Trans. 85 (52). doi:10.1029/2004EO520005. Haylock, M., Hofstra, N., Klein Tank, A.M.G., Klok, E.J., Jones, P., New, M., 2008. A European daily high-resolution gridded data set of surface temperature and precipitation for 1950–2006. J. Geophys. Res. doi:10.1029/2008JD010201.

IPCC, 2007. Climate change, 2007: the physical science basis. Contribution of working group I to the fourth assessment report of the intergovernmental panel on climate change. In: Solomon, S., Qin, D., Manning, M., Chen, Z., Marquis, M., Averyt, K.B., Tignor, M., Miller, H.L. (Eds.). Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA. Lionello, P., 2005. Extreme surges in the gulf of Venice. Present and future climate. In: Fletcher, C., Spencer, T. (Eds.), Venice and its Lagoon, State of Knowledge. Cambridge University Press, Cambridge UK, pp. 59–65. Lionello, P., Cavaleri, L., Nissen, K.M., Pino, C., Raicich, F., Ulbrich, U., 2010a. Severe marine storms in the northern Adriatic: characteristics and trends. Phys. Chem. Earth (this special issue) Lionello, P., Elvini, E., Nizzero, A., 2003. A procedure for estimating wind waves and storm-surge climate scenarios in a regional basin: the Adriatic sea case. Clim. Res. 23, 217–231. Lionello, P., Mufato, R., Tomasin, A., 2005. Sensitivity of free and forced oscillations of the Adriatic sea to sea level rise. Clim. Res. 29, 23–39. Lionello, P., Boldrin, U., Giorgi, F., 2007. Future changes in cyclone climatology over Europe as inferred from a regional climate simulation. Clim. Dyn.. doi:10.1007/ s00382-007-0315-0. Lionello, P., Cogo, S., Galati, M.B., Sanna, A., 2008. The Mediterranean surface wave climate inferred from future scenario simulations. Global Planet. Change. doi:10.1016/j.gloplacha.2008.03.004. Lionello, P., Galati, M.B., Elvini, E., 2010b. Extreme storm surge and wind wave climate scenario simulations at the Venetian littoral. Phys. Chem. Earth (this special issue). Marcos, M., Tsimplis, M.N., 2008. Comparison of results of AOGCMs in the Mediterranenan Sea during the 21st century. J. Geophys. Res. 113, C12028. doi:10.1029/2008LC004820. Matulla, C., Hofstätter, M., Auer, I., Böhm, R., Maugeri, M., von Storch, H., Krueger, O., 2011. Storminess in northern Italy and the Adriatic sea reaching back to 1760. Phys. Chem. Earth (this special issue). Meehl, G.A., Stocker, T.F., Collins, W.D., Friedlingstein, P., Gaye, A.T., Gregory, J.M., Kitoh, A., Knutti, R., Murphy, J.M., Noda, A., Raper, S.C.B., Watterson, I.G., Weaver, A.J., Zhao, Z.-C., 2007. Global climate projections. In: Solomon, S., Qin, D., Manning, M., Chen, Z., Marquis, M., Averyt, K.B., Tignor, M., Miller, H.L. (Eds.), Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA. Nakicenovic, N., Alcamo, J., Davis, G., de Vries, B., Fenhann, J., Gan, S., Gregory, K., Grbler, A., Jung, TY., Kram, T., La Rovere, E.L., Michaelis, L., Mori, S., Morita, T., Pepper, W., Pitcher, H., Price, L., Raihi, K., Roehrl, A., Rogner, H.-H., Sankovski, A., Schlesinger, M., Shukla, P., Smith, S., Swart, R., van Rooijen, S., Victor, N., Dadi, Z., 2000. IPCC Special Report on Emissions Scenarios. Cambridge University Press, Cambridge, UK. Plag, H.-P., Tsimplis, M. N., Hammond, W., Pugh, D., 2006. Appraisal of the relative sea level rise scenario for Venice. Final Project Report, Nevada Bureau of Mines and Geology, University of Nevada, Reno, Report NBMG-RP-VEN3, p. 134. Plag, H.-P., 2008. Sea level rise around the Dutch coast: past changes and future scenarios. Report provided to the Committee for the Assessment of Future Sea Level Rise at the Dutch Coast, Delta Commission of the Dutch Parliament, February 2008, p. 20. Plag, H.-P., Adegoke, J., Bruno, M., Christian, R., Digiacomo, P., McManus, L., Nicholls, R., van de Wal, V., 2010. Observations as decision support for coastal management in response to local sea level changes. In: Hall, J., Harrison, D.E., Stammer, D. (Eds.). Proceedings of OceanObs’09: Sustained Ocean Observations and Information for Society, vol. 2, Venice, Italy, 21–25 September 2009, ESA Publication WPP-306, doi:10.5270/OceanObs09.cwp, p. 69. Planton, S., Lionell P., Artale, V., Aznar, R., Carillo A., Colin J., Congedi, L., Duboisa, A., Elizalde, C., Gualdi, S., Hertig, E., Jacobeit, J., Jordà, G., Li, L., Mariotti, A., Pian, C., Ruti, P., Sanchez-Gomez, E., Sannino, G., Sevault, F., Somot, S., Tsimplis, M.N., 2012. The climate of the mediterranean region in future climate projections. In: Lionello, P. (Ed.), The climate of the Mediterranean region: Form the past to the future. Elsevier, Amsterdam. doi:10.1016/B978-0-12-416042-2.00008-2. Rahmstorf, S., Cazenave, A., Church, J.A., Hansen, J.E., Keeling, R.F., Parker, D.E., Somerville, R.C.J., 2007. Recent climate observations compared to projections. Science 316, 709. doi:10.1126/science.1136843. Rignot, E., Velicogna, I., van den Broeke, M.R., Monaghan, A., Lenaerts, J., 2011. Acceleration of the contribution of the Greenland and Antarctic ice sheets to sea level rise. Geophys. Res. Lett. 38, L05503–L05508. doi:10.1029/2011GL046583. Sevault, F., Somot, S., Beuvier, J., 2009. A regional version of the NEMO ocean engine on the Mediterranean sea. NEMOMED8 User’s Guide. Technical Note 107, CNRM, Météo-France, Toulouse. Spada, G., Stocchi, P., Colleoni, F., 2009. Glacio-isostatic adjustment in the Po plain and in the northern Adriatic region. Pure Appl. Geophys.. doi:10.1007/s00024004-0498-9. Teatini, P., Tosi, L., Strozzi, T., Carbognin, L., Cecconi, G., Rosselli, R., Libardo, S., 2010. Resolving land subsidence within the Venice lagoon by persistent scatterer SAR interferometry. Phys. Chem. Earth (this special issue). Tsimplis, M.N., Raicich, F., Fenoglio-Marc, L., Shaw, A.G.P., Marcos, M., Somot, S., Bergamasco, A., 2009. Recent developments in understanding sea level rise at the Adriatic coasts. Phys. Chem. Earth (this special issue). Zampieri, M., Giorgi, F., Lionello, P., Nikulin, G., 2010. Regional climate change in the northern Adriatic. Phys. Chem. Earth (this special issue).