Skip to navigation – Site map
Apports des sciences historiques

The Little Ice Age in Italy from documentary proxies and early instrumental records

Le petit âge de glace en Italie, des proxies à la mesure
Dario Camuffo, Chiara Bertolin, Patrizia Schenal, Alberto Craievich and Rossella Granziero
p. 17-30

Abstracts

This paper presents and discusses sea-level rise in Venice and past temperature changes over Italy during the Little Ice Age (LIA), through the analysis of proxies and early instrumental readings. Instrumental records are available from 1872 for sea level and from 1654 onwards for temperature but with a gap from 1670 to 1716. For earlier periods, documentary proxies have been used. In Venice, the most severe winters are known from written sources since the origin of the city ; furthermore, some paintings provide a view of the frozen lagoon. From 1500 to 1758, sea level has been reconstructed using two proxies : the algal belt visible in paintings by Veronese (16th century), Canaletto and Bellotto (18th century) and the submersion of the honour stairs of the historic palaces facing the Canal Grande. The depth of the lowest steps of the water stairs is correlated with sea level and constitutes a useful proxy. The result is that sea level in Venice rose at an exponential rate after the onset of global warming. The same can be said for the storm surges flooding the city. Concerning air temperature, written sources have been gathered from public and private libraries and archives, and the information has been transformed into indexed proxies from -3 (extreme cold) to +3° C (extreme warmth), 0 being normality. In general, documentary proxies are useful to pinpoint extreme weather or unusual periods on the short time scale. Documentary proxies are also useful to determine the frequency of extreme events, but fail in determining long-term trends or cycles that are only recognizable in instrumental records. Instrumental records show that the temperature was characterized by repeated cold-warm swings, with a warm maximum culminating between 1725-30, similar or even greater than the present-day level. The multiproxy temperature reconstruction models have been tested with the proxy and instrumental data gathered in northern-central Italy, and the results are discussed.

Top of page

Author's notes

The Authors are particularly grateful to the Frogmen Team of the Italian National Police, Venice, who made the underwater measurements concerning the monumental stairs in Canal Grande, essential to this study. The kind cooperation of the colleagues Fabio Trincardi and Roberto Zonta (CNR-ISMAR, Bologna and Venice) and Francesco Di Biasio (CNR-ISAC) is also acknowledged and highly appreciated. Sincere thanks are due to Elena Fumagalli and Daniele Resini © Insula spa, Venice, for the archive research and Fig.5b. This study has been made in the framework of the EU funded project Climate for Culture (Grant 226973).

Full text

1The term ‘Little Ice Age’ (LIA) is used to describe a past climate epoch in Europe roughly from 1450 to 1850. The timing, however, varies geographically (IPCC, 2013). In Italy, climate change from the LIA to present-day global warming is presented thanks to documentary proxies and instrumental observations. This climate change produced an acceleration in sea-level rise that was, and still is, a tremendous challenge for Venice.

2Documentary proxies include man-made written or iconographic data sources, which directly or indirectly reflect the past events that affected society (STANFORD, 1986 ; BRADZIL et al., 2005) e.g. weather or natural hazards. Written sources may be narrative, administrative, daily weather logs or other types. They can be divided into ‘first class’, composed of contemporary manuscripts ; ‘second class’ reliable but not contemporary collections ; and ‘third class’ less reliable sources that should be rejected. Non-contemporary sources, i.e. second class, should not be automatically rejected because they might include precious information from early sources that has been lost or destroyed during floods, fire or wars and need validation before use. These types of documentary proxies can be transformed into indexes and used as quantitative proxy data.

3Documentary proxies constitute an extremely important source of data in the pre-instrumental period. They are quite abundant in Italy for the presence of a relevant number of well-educated people, able to report and comment extreme events, natural challenges and the impact they had on society. Many documents were produced, but most of the earliest ones were destroyed by fire, especially given that, before the 14th century, most buildings were made of wood and vulnerable to fire. The main catastrophic fires in Venice include :

976 : the Ducal Palace, including the State Archives, which was burned during popular unrest ; the fire propagated and destroyed 300 buildings in Venice.

1105 : two terrible fires destroyed 16 of the islands composing Venice.

1149 : fire destroyed 13 districts of Venice.

1509 : fire destroyed the Arsenal and 120 houses.

1686 : fire destroyed a whole district with 70 houses and several timber deposits.

1789 : fire propagated from an oil storage area and destroyed 80 houses.

1848-49 : Revolution against the Austrian occupation, with a number of fires.

1915-18, World War I : 736 bombs (112 incendiary) reached Venice.

1945, World War II : the harbour of Venice was bombed.

4Fortunately, over the times of the Venice Republic a number of local scholars and historians looked for early documents and compilations, and this activity saved the content of a number of early sources that were later destroyed by fire (especially those before 1509) and other accidents. Some documents and compilations have been published, summarized or mentioned in various books. In the obscure periods in which the original sources have been lost we are forcedly obliged to rely on late transmission of the information, making a careful choice about the reliability of the author, based on his level of education, local origin and access to original documents and verification by cross comparing his descriptions with other known events. For instance, G. TOALDO (1719‑1797) was very famous in Europe and was considered an excellent source of information, especially for his book first published in 1770 and translated in several languages on the influence of celestial bodies, especially the affect of the moon on weather and agriculture. He was an astronomer and meteorologist, and was really excellent in making meteorological observations, analyzing and commenting them. He also sought to collect past climate events from various historic sources and compilations, neglecting their quality and often not quoting the source. Therefore, he contributed positively with the contemporary data, but the historic collection needs severe revision and this is understandable for the 18th century. Similar comments can be made for other famous lists in the next two centuries, e.g. ARAGO (1858), HENNIG (1904) and EASTON (1928) who made more advanced collections with the frequent citation of sources, but without verifying them. As an example, for the year 1621 we found no mention of frost in Venice in the local, contemporary sources. However, ARAGO and EASTON report: “The Zuidersee was completely frozen over. The Venetian fleet was frozen in the Lagoon of the Adriatic Sea », relying on S. CALVISIUS. CALVISIUS published the Opus Chronologicum in 1605 and died 1615. In the third posthumous edition published in 1629, and updated by J. ZHYM, almost contemporary to the event, we read: “1621 AD. The cold was very intense for the whole month, and a part of the Baltic Sea was covered with a thick ice sheet” but there is no mention about Venice and the Adriatic Sea. The mistake probably derived from SWEENY (1837) in his article “An essay on the climate of Ireland” in answering to the question proposed by the Royal Irish Academy: “Whether we have reason to believe that a change has taken place in the climate or Ireland, and if such change has occurred through what period can we trace it, and to what causes should we assign it”, that is the crucial answer we are repeating now.

5Most of the original documents that survived fire, floods and wars are still kept in the State Archives of Venice and cover about 90 km of shelves. This is a huge source of information and we have consulted it for decades ; however, this also means that it is impossible to find and examine all relevant documents. Our efforts for decades have contributed to extend our knowledge but are far from being conclusive. In this paper we will deal with extreme winters and flooding sea surges in Venice.

6Useful iconographic sources are accurate realistic paintings that document extreme or normal events that happened in a certain region, e.g. the extreme winters in which the Venice lagoon was frozen over, or paintings made with the help of a camera obscura that document the algae belt and the sea level. They are useful in providing either qualitative or quantitative proxy data, as we will see later. The underwater depth of the lowest step on the water stairs of the palaces is used as a novel proxy for sea-level rise.

7In Italy, regular daily instrumental observations of air temperature were made in Florence and Vallombrosa from 1654 to 1670 and constitute the earliest temperature observations in the world (CAMUFFO and BERTOLIN, 2012a and b). After a gap caused by the Inquisition, regular observations started again in 1716 in Padua and Bologna, followed by other stations, the most famous being Milan (since 1763), Rome (since 1782) and Palermo (since 1791). In this paper we will consider the period since 1500 to document the LIA, as well as of the transition from the LIA to global warming.

8Present-day global warming reproduces a situation in some way similar to what occurred at the end of the LIA. Climate changes may affect the frequency of short-term challenges for some unusual weather situations lasting for days, weeks, or a month, e.g. extreme frost in winter, or some long-term changes, e.g. sea-level rise. Both of them may have a strong impact on society. Sea-level rise is a serious problem for coastal towns, and Venice is the most famous and vulnerable city. However, short-term weather events too may have some consequences that will continue over time. For instance, unusually severe winters kill tree and cattle, with serious consequences on farming activity. However, they may trigger other events, e.g. the 1789 winter that was very severe over most of Europe. In addition to the direct consequences on people, animals and vegetation, frost made it impossible to use water mills and grind wheat. The result was food shortage, and fluvial transport of goods was impossible, accentuating the famine and in particular leaving the French people without bread. This contributed to exacerbate the situation that determined the onset of the French Revolution (1789-1799). Climate was only a trigger ; the real problem was an unstable situation due to heavy taxation and the unsustainable disparities between the declining living conditions of people and the lifestyles of the ruling class. The 1789 winter was one of the most influential winters in the history of mankind due to the synergistic combination of a series of various, independent factors. However, we cannot establish any scientific cause-effect relationship between climate and social events, although historians may take advantage of the knowledge of the climate context.

1 - Great frost in Venice

9Not all climate-related events provide homogeneous information in type (i.e. qualitatively) and level (i.e. quantitatively). The relative relevance may change with the type of event, over time and from site to site. The essential feature is that the event should be clearly identifiable, should have a strong impact on society (otherwise it can be disregarded) and could be classified in terms of severity in some objective way, e.g. depending on the consequences.

10The severe winter in Venice is an easily identifiable proxy because of the formation of ice sheets on the Lagoon. The mechanism is different from a lake where the same water body loses heat over time until a thin or thick ice slab is formed when the surface temperature drops below 0° C. A lake reacts to the cumulative effect of cold : a short period of wind blowing at very low temperature may have the same effect as a longer period of less intense cold. In the case of Venice, we should consider a dynamic balance where the tidal exchanges constitute a relevant factor. The Adriatic Sea constitutes a reservoir of mild water, e.g. 7° C (RUSSO et al., 2012). The Lagoon is characterized by shallow water and the tidal exchanges are strong in the deep canals, but weak on the marshes and near the borders. As a consequence, ice is formed when the heat removed by the atmosphere exceeds the heat supplied by the sea and the marine water temperature drops below -2° C. For instance, the presence of violent and cold Bora wind is very efficient in removing heat. Another important factor is the Moon : at neap tidal cycles (i.e. when the Moon is at first and third quarters) the tidal ranges are less than average, the exchange with the Sea is reduced and, consequently, the heat supply from the Sea.

11It is normal that some ice is formed near the internal borders, but a thick ice sheet bearing people may be formed when a cold wind (e.g. -10° C) blows from the North for two or three weeks at least, depending on the tidal exchanges. In conclusion, when the Venice Lagoon is frozen over, all the rivers and the lakes in the same region are also frozen. The opposite, however, is not proven.

12In Venice, we define a “very severe winter” (VSW) when the Lagoon was partially frozen, and the canals had ice slabs. The frost disrupted food supplies from the hinterland and the city was besieged by cold. In the hinterland, some people, animals and plants were killed by frost, and rivers and lakes were frozen over. However, the salt water of the Lagoon was partially frozen over, and the ice sheet was not thick enough to bear the load of people. The impossibility of boat transport caused severe problems to the city, with a terrible social impact.

13“Great winters” (GW) are extreme, exceptional events. During the GW, the cold lasted for a longer time or was more severe in comparison with a VSW. The extreme cold lasted for weeks, killing people, animals and trees. Wine in barrels and water in wells were frozen over. In Venice the ice sheet was thick, and could bear the load of people, animals, barrels, carts or wagons transporting food supplies. For instance, for the year 1709 we found several local sources stating that the Lagoon and all canals were frozen over by an exceptionally cold wind from the North combined with a modest tidal range, and that it was possible to safely reach on foot the Giudecca, Murano and other islands transporting food and other supplies. People walked in strict processional order to reach the hinterland and to stay on an already tested path, considered safe, without knowing that the repeated vibrations of each step were extremely dangerous for ice breaking. This means that the ice thickness exceeded 20 cm. The situation was described in a number of chronicles and other written sources, as well as in paintings that clearly show what happened (Fig. 1)

Fig. 1 – Painting of the frozen Lagoon in January 1789

Fig. 1 – Painting of the frozen Lagoon in January 1789

People are crossing the ice sheet to reach the hinterland and transporting supplies. The hinterland and the mountains in the background are covered with snow.

Crédits : Correr Museum, Venice, inv. Cl I n.2054, oil on canvas 190x375 cm).

14The list of VS and GW in Venice (Camuffo, 1987 ; Camuffo et al., 2014) since AD 1400 has been reported in Table 1.

Table 1 – List of the most severe winters in Venice

Table 1 – List of the most severe winters in Venice

15The above data are reported in Fig. 2 together with the solar irradiance reconstructed since 1615 (LEAN, 2000) and the periods of prolonged sunspot activity minima, i.e. the Spörer (1460‑1550), Maunder (1645-1715) and Dalton (1790-1830) Minima. It is clear that the GW were particularly frequent in the 15th (five GW), 16th (four GW) and 18th centuries (five GW). A remarkable attenuation in the severity of winters was found in the 17th century (two GW) followed by the 19th (no GW) and 20th centuries (one GW). The most frequent occurrence of extreme winter was not coincident with the periods of minimum solar activity, when sunspots became rare and solar irradiation was lower.

Fig. 2 – Solar activity in terms of annual mean irradiance

Fig. 2 – Solar activity in terms of annual mean irradiance

In the lower part of the graph, the Great Winters (GW, black dots) and Very Severe Winters (VSW, white circles) are reported. The Spörer, Maunder and Dalton Minima of solar activity are also indicated.

Source of solar irradiance reconstruction dataset : LEAN (2000)

16The most severe winters in Northern Italy are characterized by the meteorological situation known as “blocking”, i.e. a high pressure over Western Russia or Scandinavia that blows arctic or polar air from North. The intensity and duration of winds are the most critical factor in forming an ice sheet on the Lagoon waters. The above results show that the severity of winter is related to meteorological blocking, but this weather situation was frequent during LIA. It is unclear why winter blocking is related to solar activity and the time lag between the two makes even weaker a justification of this relationship.

17It may be useful to comment that the definitions of VSW and GW are based on some visible, objective effects of intense cold lasting for a few weeks (e.g. two to four) and are especially tailored for documentary sources and spectacular effects. However, when we have in addition the instrumental observations, VSW and GW are easily recognized on the daily series, but are often masked when monthly averages are considered. The reason is that the intense cold is a relatively short-term event that in some cases may be partitioned between two consecutive months. In addition, the end of the cold period is generally followed by a warm wind, i.e. Sirocco, blowing from Africa, which tends to raise the monthly average. The situation is accentuated with regards to seasonal averages, because the contribution of a few weeks is diluted over three months. The definition of winter severity, the averaging period or the way in which data are handled, are often crucial in defining the exceptionality of some events.

18The frozen Lagoon is an index of extremely severe cold lasting for a few weeks or a month. These two event types, i.e. VSW and GW can be easily distinguished, allowing a semi-quantitative evaluation of these extreme events during the LIA. The comparison with other water bodies or other proxies might be misleading because it is based on a different definition or responding to other additional variables that are included in the process.

19For the above reasons, we cannot make comparisons between the severity of winter and the timing of the grape-harvest as being representative of the summer and early autumn climate, or the period of growth of tree-rings considered in dendrochronology, because they refer to different parts of the calendar year.

2 - Sea-level rise in Venice

20Climate changes lead to sea‑level changes, especially due to the thermal expansion of sea water. In Venice, regular tide-gauge records began in 1872 and they provide evidence for relative sea-level rise (i.e. the sea level relative to the land, as it is perceived by the local inhabitants), which is due to two main contributions : land subsidence and sea-level rise. Land subsidence is mainly due to geological factors and was relatively constant over several millennia except for a short period of underwater pumping for industrial purposes from 1920 to 1970 and culminated from 1950 to 1970 that caused 10‑12 cm of additional sinking ; later, it underwent a partial rebound (CARBOGNIN and TARONI, 1996 ; PIRAZZOLI and TOMASIN, 1999 ; CARBOGNIN et al., 2004). Land subsidence has been estimated to be in the order of 10 to 12 cm per century (COLOMBO, 1972 ; BONDESAN et al., 2001, CARMINATI et al., 2005). Locally, sea level has been affected by some other factors, e.g. a minor dynamic interference for some work that has altered the exchanges between the Adriatic Sea and the Lagoon, especially the excavation of deep channels for the passage of tanks or cruise ships. At longer time‑scales, eustasy is largely the dominant non-linear variable, the other factors being almost constant (e.g. subsidence) or having a minor impact. To reconstruct sea level over recent centuries means to identify the thermal expansion of waters in response to climate change from the LIA to present-day global warming.

21In order to extend our knowledge back in time, we studied the displacement of the algae belt on the buildings built on the side of the Venice canals. In Venice, the front of the algae belt is called “Comune Marino” (CM) and corresponds to the average level of the high tides plus the height of small local waves that regularly supply water to stone pores to feed phototrophic algae that need light and water. The CM has been evaluated 31 cm above the mean sea level (MSL) (RUSCONI, 1983) considering the tide range but disregarding waves, or a bit less (e.g. 22‑28 cm) according to other authors (D’ALPAOS, 2010) and varies over time, following the average sea-level rise. However, the above estimates are based on the measurement of tides, the average high tide being 30 cm above MSL. The contribution of waves due to wind and traffic was not included. Our measurements were repeated for four days on 80 buildings facing the Canal Grande and demonstrated that the front of the algae belt is CM = 47±2 cm above MSL, showing that waves contribute for about 17±2 cm of the sea‑level variations. However, the exact absolute value of the CM is irrelevant to our aims because we will make reference to relative changes in CM level.

22A useful source regarding the displacement of the algae front is Zendrini (1821). Describing works at the Ducal Palace in 1810 he provided some references on the position of CM that was 24.2 cm above the floor consistent with the MSL in 1897. This observation led us to conclude that the MSL in 1810 was 54.7 cm below the 2012 level.

23The Venetian painters of vedutas (Italian for “view”) Antonio Canal, nicknamed Canaletto (1697‑1768) and his pupil Bernardo Bellotto (1722‑1780) made accurate reproductions of buildings using a camera obscura on the site. The camera obscura (Fig. 3) operated like a modern reflex camera that permits viewing through the objective. In the camera obscura the light beam penetrating from the objective was projected on an inclined mirror and reflected onto a glass located on the upper surface of the box, where a sheet of paper was placed making it possible to draw precise contour lines. The visibility of the pale reflected image was improved with a mobile screen on the top that shielded against external light.

Fig. 3 – A camera obscura

Fig. 3 – A camera obscura

It operates like a modern reflex camera. The light beam (white arrow) from the objective O is projected on an inclined mirror (M, visible for having removed the left side of the box) and reflected onto a transparent glass surface on the upper surface of the box. On the top of the box, a mobile screen S shields the image from external light.

24The paintings by Canaletto and Bellotto report buildings and palaces with the exact detail of the green-brown algal band that lived on the damp belt of wall basements (Fig. 4). In historic palaces, the basement was made of big stones, decorations and several architectural features that make it easy to identify the exact level of the front of the algae. If one goes on the boat close to the buildings reproduced in the paintings and compares the position of the algal belt at the time when the painting was made and today, the upward displacement of the algal belt (and therefore that of mean sea level) is easily recognized. In practice, Canaletto’s and Bellotto’s paintings constitute a double proxy: iconographic for the view and biological for the algae level. The above methodology was applied to all the available paintings in which buildings had a clearly visible algae belt and were unaffected by renovation works. This reduced the choice to 12 out of 200 paintings reproducing Venice from 1727 to 1758, and the final result of the study was 61±12 cm since Canaletto’s time (CAMUFFO and STURARO, 2003, 2004).

Fig. 4

Fig. 4

(A) Canaletto, view of Canal Grande from Balbi Palace to Rialto (1720‑23). (B) The green-brown belt of the algae is visible as highlighted by the CM in the enlarged detail ; SL is the sea level in the painting.

Ca’ Rezzonico, Museo del Settecento Veneziano, Venice, oil on canvas, inv. Cl. I, n. 2325, 144 x 207 cm.

25Some years after the above study, another special painting was considered. In 1571 the Venetian painter Paolo Caliari, nicknamed Veronese (1528-1588), painted the Family Coccina presented to the Virgin, now kept at the Gemäldegalerie Alte Meister in Dresden. The painting included a precise representation of the family’s magnificent palace, completed ten years earlier. The palace was reproduced with extreme precision and documents have been found suggesting that Veronese probably also used a camera obscura to improve the accuracy of his rederings. An analysis of this painting and the displacement of the algae belt led us to conclude that sea level rose 82±9 cm (CAMUFFO, 2010, 2012).

26In practice, accurate pictorial documentary proxies cover the 18th century quite well, but provide only one value for 1571, and this datum needs statistical confirmation. The Venetian palaces, with their stairs, offer a novel proxy and the possibility to verify the findings from the above paintings.

27Venetian palaces do not face streets, but canals, and the most magnificent buildings face the main canal nicknamed Canal Grande. These palaces have secondary doors onto the narrow streets, but monumental doors and external water stairs face the canals (Fig. 5a), to receive supplies transported by boat and welcome guests arriving by gondola. Venice palaces were built on closely spaced wooden piles deeply implanted in sand and mud until they reached the underlying layer of hard clay (MIOZZI, 1968). Under anoxic conditions, the wood remains almost unaffected by deterioration and biological attack. The building foundations and the water stairs remain on the implanted piles. Under normal conditions, water stairs had the necessary number of steps, with the lowest step being positioned on a vertical basement (Fig. 5b) or was carried with stone corbels projecting out from the wall basement. The total run and total rise of water stairs was kept to the necessary minimum because the submerged part was useless, expensive and constituted a risk for boats. The water stairs were necessary for everyday life, and their use was possible with the assistance of two or more mooring piles that helped to avoid collision with the underwater steps and to allow boats to moor alongside while people unloaded supplies or guests jumped from the gondola to the emerging steps of the stairs. The mooring piles were implanted on the canal bottom just in front of the stair, and the emerging part was coloured like a flag for each family. Today, sea-level rise has changed the distance from the usable steps and has forced most palaces to add a wooden wharf to facilitate boat docking (Fig. 5a).

Fig. 5

Fig. 5

(A) Grassi Palace, on the Canal Grande, with most of the water stair presently under water. A wooden wharf for boat docking has been added. A Police Frogman is standing on the lowest step and is measuring the submersion depth relative to the high-tide level i.e. CM.
(B) In a secondary canal, a small five-step water stair made of Istria stone abutting a vertical brick basement, based on the bottom of the canal, during maintenance and restoration works to reinforce the damaged parts

Courtesy of Daniele Resini ©, Insula spa, Venice.

28Nowadays we are used to seeing water stairs partly or totally submerged by marine water, but the crucial question is : when a water stair was built, how was the height of the lowest step related to mean sea level and the tidal range ? The level of the lowest step of water stairs, like all the built levels in Venice, e.g. streets, floors, bridges, was related to the CM as a reference.

29From a practical point of view, a boat must emerge 40 to 60 cm above the water mark and people should pass from the boat level to the stair. The part of the stair that is permanently underwater, or is periodically submerged by the tidal water, is fully covered with slippery green algae. It is impossible to walk or stand on it. If a step was covered of algae, the problem was solved with servants unrolling a red carpet runner to make walking possible. For the above reasons, the most essential and less expensive stairs ended at the CM level. All the steps below were not usable without servants with carpets.

30By contrast, the most magnificent palaces had monumental stairs with steps visible over the whole tidal range, with the lowest step close to or below the average low-tide level, or even at the level of the less frequent very low tides to provide a complete visual appearance at any time. The steps below the CM were covered with algae and were only visible at low tides. In Venice, the normal tidal range is around 60 cm (i.e. mean = 58.6 cm, median = 60 cm median) and the typical rise height of the steps is 17.3 cm, i.e. half a Venice foot. In practice, the stairs were linked to the CM, but the most magnificent buildings (herewith called outliers) had a number of submerged steps depending on the required aesthetic features. In practice, four steps covered the normal tidal range. In particular, the magnificent baroque palaces by Longhena (18th century) have huge honour stairs, that are also visible during low tide.

31The key point is that sea level has varied over centuries, and the lowest step of the newly built water stairs followed it. The underwater depth of the lowest step of the water stairs of palaces facing the Canal Grande may provide additional, independent information about the water level from the 16th to the 18th centuries. To this end, an extensive field survey was made with the cooperation of the Frogmen Team of the Italian National Police that has been essential for underwater inspections and to operate under normal water traffic conditions. The measurements consisted in : establishing the exact position of the lowest step that was often rotten or broken ; measuring the depth of this step in relation to the present water level ; measuring the distance between the simultaneous water level and the CM in a position far from the stair where the reflecting and splashing waves locally altered the green band. The Tide Prediction Center of the Municipality of Venice has kindly provided us the real time access to the simultaneous tide level readings, with transmission via Global System for Mobile Communications (GSM).

32The overview of the tide gauge record (1872‑onwards), the Canaletto and Bellotto paintings (1727‑1758), the Veronese painting (1571) and the depth of the lowest step of the monumental stairs on the Canal Grande, measured from the natural Venice reference level, i.e. the CM, is reported in Fig. 6. The height of a step (i.e. ±17.3 cm) around the best-fit line is also shown : we see that the depth of the lowest step is scattered. The upper readings are for the most essential stairs at the CM level, the lower readings include the low tide level. For these reasons, we depicted, in a different colour, the outlier buildings, i.e. stairs with two steps below the levels of other buildings of the same period or one step below the best-fit interpolation. The best-fit interpolation (i.e. the thick central line) is exponential, and has been calculated from the paintings and tide gauge observations to avoid the problem of unbalances for the wide scatter of the step proxy. The underwater survey shows that the lowest step is generally distributed within ±1 step around the above best-fit interpolation representing the average sea level.

Fig. 6

Fig. 6

Relative sea-level rise at Venice indicated by the tide gauge (1872‑onwards, black & white circles), a measurement of CM by Zendrini (green circle), the 1727-1758 Canaletto and Bellotto paintings (pink squares), the 1571 Veronese painting (blue square) and the lowest step of the water stairs on the Canal Grande (triangles : black for normal palaces, red for monumental outliers). The thick exponential line in the middle is the best fit calculated with the black dots and represents the average tide level ; the upper and lower thin lines represent the average high and low‑tide levels, respectively, and the band of the average tidal range.

33The water stair proxy follows an exponential trend showing an acceleration over the centuries from the LIA to the present-day global warming. In the 16th century the rate of sea-level rise was 11.4 cm/century ; in the 18th century it was 19.4 cm/century, today it is more than double the 16th century rate, reaching 25.4 cm/century due to the thermal expansion of sea waters.

3 - Extreme storm surges in Venice

34Sea-level rise increases the frequency of exceptional floods in Venice generated by storm surges over the Adriatic Sea. They typically occur in the cold season, when low pressure lies over the western Mediterranean, the main factor being the Sirocco wind blowing from Africa along the Adriatic Sea and dragging waters towards Venice. The heat and moisture released by the warm Mediterranean waters accentuate the low pressure and feed the local cyclonic activity. Other related factors are the atmospheric pressure pattern and the seiches, i.e. the free oscillations generated in the Adriatic Sea by the sudden change in barometric pressure. All of the above factors are determined by the atmospheric forcing. The storm surge determined by the meteorological forcing is superimposed on the lunar-solar tides, (TOMASIN, 2005) and both contribute to the total level of seawater flooding Venice. The height of the meteorological surge is added above the tidal level and the sea-level rise increases the frequency and severity of the flooding waters, locally known as “Acqua Alta” (literally : high water). At present an “Acqua Alta” is defined when the sea level exceeds 110 cm above the 1897 “Zero” reference level, and marine water starts to invade the floor of Piazza San Marco. Before instrumental records, i.e. before 1872, sea floods are known from the documentary sources since the origins of the Venice Republic and have been presented in previous papers (CAMUFFO, 1993 ; ENZI and CAMUFFO, 1995).

35The updated series of the documentary and instrumental “Acqua Alta” is shown in Fig. 7.

Fig. 7 – Frequency of sea-surge flooding in Venice

Fig. 7 – Frequency of sea-surge flooding in Venice

Vertical grey lines : floods ; thick black line : 10-yr running average of flood frequency ; thick grey line : exponential fit. The Spörer, Maunder and Dalton Minima of solar activity are also indicated with horizontal bars.

36An exponential increase is evident, following the trend of the sea-level rise, upon which some periods of increased frequency are superimposed. Solar activity is a potential cause that should be considered. Sea surges were frequent during the Spörer and the Dalton Minima of solar activity, but there is no apparent connection with the Maunder Minimum. In addition, an increase in sea-surge frequency was observed in the period between the Maunder and Dalton Minima. The mechanism linking sea surges and solar activity is complex and still unknown. The only clear factor is that the superposition of the storm surges and sea-level rise is increasing the flooding frequency at an exponential rate, which constitutes a dramatic challenge for the society and the monumental palaces of Venice. Once stone and masonry are impregnated with salt water, over the long term the dissolution-crystallization cycles around the deliquescence level (i.e. 75 % relative humidity for NaCl) will destroy the monuments.

4 - The air temperature records in Italy

37Written sources have been gathered and transformed into proxy data (CAMUFFO et al., 2010) as follows. First, the text has been analysed and classified in terms of departures from the norm, as deduced from the description of facts and consequences. The classification was made by attributing levels from + 3 to a really extreme and well-documented event (e.g. extreme heat or unusual warmth), as we can expect to occur not more than two or three times per century in that season, to - 3 in negative cases (e.g. extreme or unusual cold). The proxy zero, by definition, is the most frequently occurring class of events (i.e. the mode) of that time considered as the “normal” climate situation by the observers. Temperatures below the “normal” that have some objective consequences are perceived as negative (e.g. cold), above positive (e.g. warm) and this is reflected in the numerical values of the indices (- 3,+ 3). In the case of the winters discussed above, GW are classified as - 3 and VSW as - 2. Levels ± 2 and ±1 are used for intermediate levels of decreasing severity and level 0 is “normal”. The severity levels represent a departure the “norm” (considered the average), and may be related to the standard deviation (SD). Once the sources are indexed into severity levels they can be reported on a semi-quantitative graph.

38However, if a common period exists, in which both such indices and instrumental temperature observations are available, it is possible to pass from an arbitrary index to °C units. The methodology is simple. In the common period, we know the two distributions of the same variable (i.e. the temperature) expressed both in ±3 index levels and in °C as well. The two distributions must coincide because they are two different ways of expressing the same observed variable but in two different units. Therefore, we attribute the real °C values to the corresponding index levels. The methodology is based upon a frequency criterion, i.e. the percentile distribution of temperature readings (or indexed proxy). We know that the temperature readings follow a Gaussian distribution, centred on the average and equally departing on both sides. Geometrically, a Gaussian is defined in terms of the standard deviation (SD). Once the temperature readings (or the proxy) are classified into the above severity classes, each class will fall within ±0.5 SD, ±1 SD and ±2 SD and will contain a certain population of readings, as reported in Table 2.

Table 2 – 

Table 2 – 

Classification of temperature readings in terms of Standard Deviation (SD) classes and the population of readings falling within each class.

39At this point it is sufficient to relate the span of each class to the quantitative values of the observed instrumental readings expressed in °C, i.e. the temperature levels at the ±0.5 SD, ±1 SD and ±2 SD thresholds. Once this exercise is made over a common period, the methodology is validated over another different common period. If the result is successful, the calibration is extended to the whole proxy series. This methodology forces and fixes some homogeneity to the readings. This constitutes an advantage, but also an unavoidable distortion to the data sources.

40The documentary proxy methodology is convenient for extreme values and short-term variability. However, it does not respond to long-term trends because the zero reference is the subjective idea of normality that the author has in mind (e.g. “the most extreme event in living memory”) and cannot be compared with previous or subsequent writers. The methodology is excellent in the frequency domain and in qualitative investigations, but limited in the purely quantitative domain. Although with the above limitations, documentary proxies are essential because they cover the early period of LIA when no instrumental records were available. In this paper, they will be used for the 1500‑1653 pre-instrumental period and the 1671‑1715 gap. The 1716‑1799 period, which is common with instrumental observations, has been used to calibrate this methodology and has not been reported in the graph (Fig. 8). In the proxy period, the arbitrary zero level has been established to be coincident with the average of the whole available period of instrumental data. This choice is justified to avoid a discontinuity in level when passing from proxy to instrumental readings.

41Concerning instrumental observations, original logs with early instrumental readings have been collected, corrected and adjusted to modern standard units of temperature and time, homogenized and made comparable to each other and provide precise information about the past climate over the last 350 years (CAMUFFO and JONES, 2002 ; CAMUFFO et al., 2010 ; 2011). Instrumental observations used in this paper are derived from the series of Florence, Vallombrosa, Padua, Bologna and Milano (see Table 3) and cover 1654-1670, and then 1716-today.

Table 3 – The longest instrumental series in Italy

Table 3 – The longest instrumental series in Italy

42In this paper, concerned with the LIA, we have reported the seasonal averages of instrumental records for reasons of consistency with the proxy series (that have been grouped per season in order to increase their density) and the multiproxy temperature reconstruction models (MTRM) to be discussed later. The series have been stopped at the end of the LIA, i.e. 1850, to improve the visual resolution of drawings and to remain within the selected time interval.

43The temperature series obtained by combining the documentary proxies and the instrumental observations are reported in Fig. 8.

Fig. 8 – Temperature anomaly for Northern-Central Italy.

Fig. 8 – Temperature anomaly for Northern-Central Italy.

The 1500-1653 and 1671-1715 periods are based on documentary proxy data (P), grey background. In this period only year-to-year variability of unusual temperatures is possible. The 1654-1670 and 1716-185 periods are based on instrumental readings (I), white background. In this period, cycles and long-term variability are detectable. The most dominant feature is the presence of repeated warm‑cold swings that characterize almost all seasons.

44The proxy period cannot discriminate cycles or trends, as already discussed, but pinpoints various extreme or unusual events. The best-documented seasons are winter and spring, thanks to documents discussing severe frost and snow. In summer, complaints concern intolerable heat and aridity ; some liturgical actions, prayers and processions to invoke rain (i.e. “pro pluvia”) are made in the hope that the heat will abate. Some mild summers are also visible. The least-documented season is autumn, because during this season temperature is not a critical factor. The instrumental period shows extreme events and repeated swings. The negative anomalies are dominant over the positive ones in winter and spring, except for the 1716-1730 warm interval.

45In order to analyze the extreme events during the LIA and compare them to present-day Global Warming, the 1500-2010 period has been subdivided into 17 sub-periods, i.e. 30-year intervals, each interval being represented by two bars (blue and red) in the histogram in Fig. 9.

Fig. 9 – Extreme events outside the 10-percentile and 90-percentiles in the 1961-1990 reference period (RP).

Fig. 9 – Extreme events outside the 10-percentile and 90-percentiles in the 1961-1990 reference period (RP).

The bars represent the number of outliers over 30-yr intervals starting from 1500. Red : warm season outliers (i.e. exceeding the 90-percentile RP) ; blue : cold season outliers (lower than 10-percentile RP). The mean of the positive blue bars increased the frequency of seasons colder than the reference level ; the negative blue bars decreased the frequency of them. Positive red bars mean increased frequency of seasons warmer than the reference level ; the negative red bars decreased frequency of them.

46For each season, the temperature levels of the 10th and 90th percentiles have been calculated for the 1961-1990 reference period. For each interval and for each season, the percentage anomaly of outlier events is reported, i.e. those falling outside of the 10th – 90th percentiles of the reference period. For each interval, blue bars represent the percentage of observed anomalous cold seasons from which the 10 values of the reference period have been subtracted. Similarly, the red bars denote warm events. This choice has been made to make the reference period 1961-90 zero, with the results being expressed results in terms of anomaly.

Winters (Fig. 9a). Positive blue bars mean increased frequency of the most severe winters; negative blue bars denote decreased frequency of the most severe winters. The positive red bars mean increased frequency of mild winters; the negative red bars decreased frequency of mild winters. The cold period peaked in frequency around 1710-1740; after 1870 a gradual decrease of cold winter frequency and an increase in mild winters is visible. The peak in present-day (i.e. 1990-2010) Global Warming is evident.

Springs (Fig. 9b). Particularly during the 16th century, chilly springs dominated and mild springs were less frequent. After 1870 the situation was inverted, with a peak in Global Warming.

Summer (Fig. 9c). Cool summers dominated during the 16th and 17th centuries. The Global Warming peak is evident.

Autumn (Fig. 9d) The 16th and 17th centuries have been quite similar to the reference period. The second half of the 18th and the 19th centuries show a slight increase in cold autumns. Global Warming is less evident.

47The most dominant feature of the instrumental readings discussed in Fig. 8 is the presence of repeated warm‑cold swings that characterize almost all seasons. Winter and spring are the most variable seasons. Summer is the most regular one. The swings are probably related to the sea surface temperature (SST) in the Mediterranean and the North Atlantic Oscillation (NAO), but for which high‑resolution data are unfortunately lacking before 1850. However, the harmonic analysis applied to SST from 1850 to 2008 shows significant peaks over the 99 % confidence limit, corresponding to 2.3, 6.3 and 73 years (MARULLO et al., 2011). The dominant NAO peaks are around 3-6, 8-10, ~20, 55–70, ~170 and ~300 years (OLSEN et al., 2012). The spectral analysis of the 1715‑2008 air temperature over the Mediterranean area revealed significant peaks at 2.2, 12.7, 26.5, 34.3 and 57.3 years (CAMUFFO et al., 2010). An accurate comparison is difficult because the swings are characterized by unstable quasi-periodicities and the records used for SST, NAO and air temperature are not of equal length. However, the air temperature swings seem to be related to both SST and NAO, as reported in Table 4 where each line reports an association of the closest values. In the SST we have also reported some multiples of the fundamental 6.3 yr period that closely fit with some of the observed cycles. However, the situation is complex and the mechanism remains unclear.

Table 4 – Significant peaks in quasi-periodicities of swings of air temperature (T) and SST in the Mediterranean and NAO.

Table 4 – Significant peaks in quasi-periodicities of swings of air temperature (T) and SST in the Mediterranean and NAO.

In parenthesis some multiples of an observed period.

48The instrumental air temperature record shows that the 1720-30 decade was particularly warm, with peaks in all seasons, at the level of the present-day global warming. The monthly averages (not reported here but presented in CAMUFFO et al., 2011), show that the highest 1720s temperatures were reached in May, September, October, November and December. Most months have no long-term trends, except for January, February and March that show a temperature increase from 1739 to present.

5 - Comparison of observations with multiproxy temperature reconstruction models

49The climate reconstruction made with the above data has been compared with the Multiproxy Temperature Reconstruction Models (MTRM) by LUTERBACHER et al. (2004), and XOPLAKI et al. (2005) that cover the period stretching back to 1500, the former for all seasons in Europe, the latter for the mid-seasons in the Mediterranean area. The large-scale multiproxy-based surface temperature reconstructions offer a quantitative assessment of large‑scale surface temperature variations. MTRM combines information from different proxy types to take advantage of the strengths of, and minimize the limitations of, individual proxies. In the above models, the 1500-1900 pre‑instrumental period has been reconstructed from multiproxy data sources ; since 1901 data are based on instrumental readings. The pre-instrumental period is based either on natural archives (i.e. ice cores, tree rings, speleothems, varved lake sediments, and subsurface temperature profiles from borehole measurements) or on multi-proxy networks. Both models reconstruct temperature for continental Europe, (25W-40E ; 35N-70N) with a seasonal resolution over the 1500-2004 time window. Europe is divided into a regular grid and the data are depicted along the X and Y axes. The X axis (longitude) is composed of 130 steps, 0.5° width each, the first point at 24.75° W and the last point at 39.75° E. The Y axis (latitude) of 70 steps, 0.50° width, the first point at 35.25° N and the last point at 69.75° N. The data are available at http://climexp.knmi.nl/​.

50The comparison between the observed data and the above MTRM calculated for the grid point over the middle point of the area where we have our documentary proxies or instrumental stations, makes it possible to evaluate how accurately variability, trends, and changes in extreme temperatures can be simulated over Northern Italy from 1500 to present.

51As previously discussed, our series are either based on written documentary sources or early instrumental observations. Wide uncertainties are associated with the early period, i.e. ±1.2°C, and the uncertainty decreases over time, when the meteorological observations had improvements in instruments and observational protocols, i.e. ±0.6°C (1654-1670 and 1716‑1726), ±0.4°C (1727‑1786), ±0.3°C (1787-1849), and ±0.2°C since 1850 (CAMUFFO et al., 2010). The difference between the proxy or instrumental observations in Northern Italy and the MTRM simulations is reported in Fig. 10.

Fig. 10

Fig. 10

Difference between observed temperature (documentary proxy data (P) and instrumental readings (I)) and MTRM calculations : LUTERBACHER et al. (2004, black line) and XOPLAKI et al. (2005, grey line).

52In the period 1500-1900, the Luterbacher et al. model suggests a winter temperature ~0.5°C below the 20th century average. In the instrumental period when a reliable quantitative comparison is possible with the observed data, the comparison of temperatures shows that the MRTM overestimated the observed winter temperature by 1°C in the 1725-1790 period. Since 1800 the simulation is good except for random departures of around 1°C. In summer, the simulation provides slightly better results. In the mid seasons, when both models operate, they provide similar results. In the proxy period some departures reach or exceed ±2°C with exceptional peaks or troughs of ±4°C. Since 1750 the band has narrowed (i.e. to within ±1°C) and remains narrow up to present. This can be explained because the early period is characterized by scarce and less accurate data, with the above limitations concerning documentary proxies ; furthermore, MTRM too have less certain tuning. After 1750, when the observational data are of better quality and more ubiquitous, the MTRM benefit from a more precise calibration and the calculated temperature fits better with the instrumental observations that constitute the true reference.

Conclusions

53Documentary proxies and instrumental observations provide a reasonably good description of what happened during the LIA. For their nature, the documentary proxies respond better to the frequency domain than to averages, and for this reason our knowledge of the early part of the LIA is necessarily limited to the frequency of extreme events and guesses about their intensity, while the second part of the LIA is better documented. The general impression is that the early part of LIA was characterized by an anomalous frequency of cold and other extreme events, especially during the 15th and 16th centuries ; the middle of the LIA was a period similar to today, both in terms of frequency and averages ; in the second half of the 18th century there was a second worsening but starting with the 19th century the situation improved with a warming trend and a reduced frequency of extreme events, except for the case of sea surges flooding Venice that are fed by Global Warming.

54In Venice, the intense frost caused spectacular effects, sometimes with dramatic impact on the society : killing people, animals and trees or leaving the city without food and beverage supplies. Very severe and great winters in which the Venice Lagoon was partially or totally frozen over decreased in frequency passing from LIA to Global Warming, with a remarkable attenuation in the 17th century. The frequency culminated in the 15th-16th century and then in the 18th century ; no GW were found in the 19th century. The frequency of the GW may be related to the occurrence of sunspot minimum activity, but the relationship is not too strict ; in particular, an anomalous number of repeated GW occurred before the Spörer solar Minimum. The relationship with the solar activity in this, and other weather situations, remains unclear.

55The intense cold periods are typically associated with blocking of high-pressure patterns located over Scandinavia or eastern Russia and lasting for a few consecutive weeks. A comparison with the instrumental records shows that the GW may be obscured when the seasonal or monthly averages are preformed because the short-term peaks of cold are distributed over a wide time interval. This suggests that the short-term resolution and the frequency analysis are essential to understand the climatic features that have characterized the LIA.

56The accuracy of the realistic veduta paintings that Canaletto and Bellotto made of Venice with the help of a camera obscura provided a quantitative evidence for the submersion of the city in the 18th century. Another painting by Veronese, also made with a camera obscura provided another information for the second half of the 16th century. The underwater depth of the lowest step of the water stairs of the Venetian palaces faced to the Canal Grande have proved to be an excellent proxy for the long-term knowledge of the relative sea level, that has now been reconstructed since the beginning of the 16th century with a large number of observations.

57The time series composed of the water stair proxy, the iconographic/biological proxy and the tide gauge records provide clear evidence of the continual sea level rise in Venice, that is characterized by an exponential trend determined by the Global Warming. In particular, in the 16th century the 11.4 cm/century submersion rate was mainly due to land subsidence. In the subsequent period we assist to an acceleration of the thermal expansion, determining an exponential sea level rise, that nowadays constitutes a terrible challenge.

58The above results are consistent with the framework of buried sediments and other archaeological remains found in the Venice Lagoon over the last two thousand years (AMMERMAN et al., 1999 ; AMMERMAN, 2005).

59In the case of storm surges determined by a cyclonic low pressure over the Western Mediterranean in the cold season, Venice is flooded by marine water. The sea level rise has increased the frequency and the level of floods. The best-fit equation of the flooding frequency is also exponential. However, the floods are not homogeneously distributed along the best-fit, but they show periods of more intense, or reduced activity. This means that the atmospheric circulation that determines cyclonic low‑pressures over the Western Mediterranean is not regular. During LIA, they were particularly frequent during the first half of the 16th century, then for almost one century, i.e. from 1730 to 1820. The correlation with the solar activity, if any, is unclear.

60The air temperature has been reconstructed with documentary proxies and instrumental observations. Documentary proxies provide excellent information in terms of frequency of extreme vents, but cannot provide average temperature levels, cycles and trends. They show that in the 16th and 17th century, periods with intense cold were more frequent than today in winter and spring, the other seasons being more regular.

61The long-term instrumental observations series show that in Italy the temperature was characterized by ceaseless decadal variability with repeated cold-warm swings, with a warm period culminated 1725-30 similar to, or even higher than, the present-day temperature.

62The multiproxy models by LUTERBACHER et al. (2004) and XOPLAKI et al. (2005) provide results very similar between them. Both models simulate within ±1°C the observed instrumental data after 1750, but the scatter is much larger in the earlier period. This can be explained with the scarcity of data to tune models, as well as with the uncertainty of proxies. This means that further research is needed to gather reliable data to improve our knowledge about LIA, especially before the 18th century.

Top of page

Bibliography

AMMERMAN A.J., (2005), The third dimension in Venice, in C.A. Fletcher and T. Spencer (eds.), Flooding and Environmental Challenges for Venice: State of Knowledge, Cambridge University Press, p. 107-115.

AMMERMAN A. J., MC CLENNEN C.E., DE MIN M., HOUSLEY R., (1999), Sea-level Change and the Archaeology of Early Venice, Antiquity, 73, p. 303–312.

ARAGO F., (1858), Œuvres, edited by J.A. Barral : Notices Scientifiques, Tome V, Sur l’état thermométrique du Globe terrestre. Chapitre 23 : Des hivers qui ont amené la congélation des grands fleuves et chapitre 24 : Les plus grandes froids observés annuellement dans les différents lieux du Globe, Gide et Baudry, Paris.

BAIADA E., (1986), Da Beccari a Ranuzzi : la meteorologia nell’Accademia bolognese nel XVIII secolo, in R. Finzi (ed) : « Le meteore e il frumento », Il Mulino, Bologna, p. 99-261.

BONDESAN M., GATTI M., RUSSO P., (2001), Vertical Ground Movements Obtained from I. G. M. Levelling Surveys’, in Castiglioni, G. B. and Pellegrini, G. B. (eds.), Illustrative Notes of the Geomorphological Map of the Po Plain, Suppl. Geogr. Fis. Dinam. Quat., 4, p. 141–148.

BRÁZDIL R. PFISTER C., WANNER H. Von STORCH H., LUTERBACHER J., (2005), Historical Climatology in Europe – The State of the Art, Climatic Change, 70, 3, p. 363-430.

BRUNETTI M., BUFFONI L., Lo VECCHIO G., MAUGERI M., NANNI T., (2001) Tre secoli di Meteorologia a Bologna. ISAO-CNR, Bologna ; Istituto di Fisica Generale Applicata-Università di Milano ; Osservatorio Astronomico di Milano-Brera. CUSL, Milano, p. 1-95.

CALVISIUS S., 1629 (posthumous) : Sethi Calvisii Opus Chronologicum ex Authoritate Potissimum Sacrae Scripturae et Historicorum Fide Dignissimorum, ad Motum Luminarium Coelestium.... Editio Tertia multis in locis emendata & ad praesentem 1629 usque annum continuata (updated by J. Zhym). Thymii, Frankfurt.

CAMUFFO D., (1987), Freezing of the Venetian Lagoon since the 6th Century AD, in Comparison to the Climate of Western Europe and England, Climatic Change, 10, 43-66.

CAMUFFO D., (1993), Analysis of the Sea Surges at Venice from A.D. 782 to 1990, Theor. Appl. Climatol., 47, p. 1–14.

CAMUFFO D., (2010), Le niveau de la mer à Venise d’après l’œuvre picturale de Véronèse, Canaletto et Bellotto, Revue d’Histoire Moderne et Contemporaine, 57, 3, p. 92-110.

CAMUFFO D., (2012), La Camera Oscura : il nostro occhio nel passato, in Fondazione Bracco : Il vedutismo veneziano : una nuova visione, Pinacoteca di Brera, Italgraf, Rubiera, p. 53-143.

CAMUFFO D., BERTOLIN C., (2012a), The earliest temperature observations in the World: the Medici Network (1654-1670), Climatic Change, 111, 2, p. 335‑363.

CAMUFFO D., BERTOLIN C., (2012b), Recovery of the early period of long instrumental time series of air temperature in Padua, Italy (1716-2007), J.‑Phys.‑ and Chem.‑Earth, 40-41, p. 23-31.

CAMUFFO D., JONES P. (Eds.), (2002), Improved understanding of past climatic variability from early daily european instrumental sources. Kluwer Academic Publishers, Dordrecht, 392 p.

CAMUFFO D., STURARO G., (2003), Sixty-cm submersion of Venice discovered thanks to Canaletto’s paintings, Climatic Change, 58, p. 333‑343.

CAMUFFO D., STURARO G. (2004), Use of proxy-documentary and instrumental data to assess the risk factors leading to sea flooding in Venice, Global and Planetary Change, 40, p. 93­103.

CAMUFFO D., BERTOLIN C., BARRIENDOS et al., (2010), 500‑year temperature reconstruction in the Mediterranean basin by means of documentary data and instrumental observations, Climatic Change, 101, p. 169-199.

CAMUFFO D., BERTOLIN C., DIODATO N. et al., (2011), Climate change in the Mediterranean over the last five hundred years. Chapter 1, in Elias Carayannis (Ed.) Planet Earth 2011 - Global warming challenges and opportunities for policy and practice, Intech, Rijeka, Croatia, p. 1-26.

CAMUFFO D., BERTOLIN C., CRAIEVICH A., GRANZIERO R., (2014), When the Lagoon was frozen over in Venice: documentary, pictorial and instrumental evidence, in: R. Gertwagen and T. Bekker-Nielsen (Eds.), The Inland Seas: towards an Ecohistory of the Mediterranean. Franz Steiner Publishers, Geographica Historica, Stuttgart (in press).

CARBOGNIN L., TARONI G., (1996), Eustatismo a Venezia e Trieste nell’ultimo secolo, Atti Istituto Veneto di Scienze Lettere ed Arti, 154, p. 281–298.

CARBOGNIN L., TEATINI P., TOSI L. (2004), Eustasy and land subsidence in the Venice lagoon at the beginning of the new millennium, Journal of Marine Systems, 51, p. 345-353.

CARMINATI E., DOGLIONI C., SCROCCA D., (2005), Magnitude and causes of long-term subsidence of the Po Plain and Venetian region, in Fletcher C.A., Spencer T. (eds.), Flooding and Environmental Challenges for Venice and its Lagoon: State of Knowledge. Cambridge University Press, Cambridge, p. 21‑28.

COLOMBO P., (1972), Problemi relativi alla difesa della città di Venezia, Rivista Italiana Geotecnica, 1, p. 7–30.

D’ALPAOS L., (2010), Fatti e misfatti di idraulica lagunare. La laguna di Venezia dalla diversione dei fiumi alle nuove opere delle bocche di porto. Istituto Veneto di Scienze Lettere ed Arti, Venice, 329 p.

EASTON, C., (1928), Les Hivers dans l’Europe Occidentale, Brill, Leiden, 210 p.

ENZI S., CAMUFFO D. (1995), Documentary Sources of Sea Surges in Venice from A.D. 787 to 1867, Natural Hazards, 12, p. 225–287.

HENNIG R., (1904), Katalog bemerkenswerter Witterungsereignisse von den ältesten Zeiten bis zum Jahre 1800. Abhdl. Kgl. Preuß. Met. Inst., Bd. II, 4, 93 p.

IPCC, (2013), Working Group 1 contribution to the ipcc fifth assessment report Climate Change 2013 : the Physical Science Basis.

LEAN J., (2000), Evolution of the Sun’s Spectral Irradiance Since the Maunder Minimum, Geophysical Research Letters, 27, 16, p. 2425-2428.

LUTERBACHER J., DIETRICH D., XOPLAKI E., GROSJEAN M., WANNER H., (2004), European seasonal and annual temperature variability, trends and extremes since 1500, Science, 303, p. 1499–1503.

MARULLO S., ARTALE V., SANTOLERI R., (2011), The SST Multidecadal Variability in the Atlantic–Mediterranean Region and Its Relation to AMO, Journal of Climate, 24, 16, p. 4385-4401.

MAUGERI M., BUFFONI L., DELMONTE B., FASSINA A., (2002), Daily Milan temperature and pressure series (1763–1998): Completing and homogenising the data, Climatic Change, 53, p. 119–149.

MIOZZI E, (1968), Venezia nei secoli. La Laguna, vol. III, Trevisan, Castelfranco Veneto, Venice, p. 403-464.

OLSEN J., ANDERSON N.J., KNUDSEN M.F., (2012), Variability of the North Atlantic Oscillation over the past 5200 years, Nature Geoscience, 5, p. 808–812.

PIRAZZOLI P.A., TOMASIN A., (1999), L’evoluzione recente delle cause meteorologiche dell’Acqua Alta, Atti Istituto Veneto di Scienze Lettere ed Arti, 157, p. 317–344.

RUSCONI A., (1983), Il Comune Marino a Venezia : ricerche e ipotesi sulle sue variazioni altimetriche e sui fenomeni naturali che le determinano, Uff. Idrogr. Magistrato Acque, Tech. Rep. 157, Venice, 39 p.

RUSSO A., SANDRO CARNIEL S., SCLAVO M., KRZELJ M., (2012), Climatology of the Northern-Central Adriatic Sea, Chapter 7, 37 p., edited by Shih-Yu (Simon) Wang and Robert R. Gillies, Modern Climatology, InTech, Open Access Company, ISBN978-953-51-0095-9, http://www.intechopen.com

STANFORD M., (1986), The nature of historical knowledge. Blackwell, Oxford, 256 p.

SWEENY J., (1830), An essay on the climate of Ireland. The Transactions of the Irish Academy, vol. XVII, Dixon Hardly, Dublin, p. 179-233.

TOALDO G., (1770), Della vera influenza degli astri : delle stagioni e delle mutazioni di tempo, Saggio Meteorologico. 1st edition, Stamperia del Seminario, Padua.

TOMASIN A., (2005), Forecasting the water level in Venice: physical background and perspectives, p. 71-78, in C.A. Fletcher and T. Spencer (eds), Flooding and Environmental Challenges for Venice: State of Knowledge. Cambridge University Press, Cambridge.

XOPLAKI E., LUTERBACHER J., PAETH H. et al., (2005), European spring and autumn temperature variability and change of extremes over the last half millennium, Geophys. Res. Lett., 32, 15, p. 1-4, L15713.

ZENDRINI A., (1821), Nuove ricerche sull’andamento del livello del mare. Istituto Scienze Lettere e Arti, Milano, p. 155-164.

Top of page

List of illustrations

Title Fig. 1 – Painting of the frozen Lagoon in January 1789
Caption People are crossing the ice sheet to reach the hinterland and transporting supplies. The hinterland and the mountains in the background are covered with snow.
Credits Crédits : Correr Museum, Venice, inv. Cl I n.2054, oil on canvas 190x375 cm).
URL http://journals.openedition.org/mediterranee/docannexe/image/7005/img-1.png
File image/png, 887k
Title Table 1 – List of the most severe winters in Venice
URL http://journals.openedition.org/mediterranee/docannexe/image/7005/img-2.jpg
File image/jpeg, 44k
Title Fig. 2 – Solar activity in terms of annual mean irradiance
Caption In the lower part of the graph, the Great Winters (GW, black dots) and Very Severe Winters (VSW, white circles) are reported. The Spörer, Maunder and Dalton Minima of solar activity are also indicated.
Credits Source of solar irradiance reconstruction dataset : LEAN (2000)
URL http://journals.openedition.org/mediterranee/docannexe/image/7005/img-3.png
File image/png, 40k
Title Fig. 3 – A camera obscura
Caption It operates like a modern reflex camera. The light beam (white arrow) from the objective O is projected on an inclined mirror (M, visible for having removed the left side of the box) and reflected onto a transparent glass surface on the upper surface of the box. On the top of the box, a mobile screen S shields the image from external light.
URL http://journals.openedition.org/mediterranee/docannexe/image/7005/img-4.png
File image/png, 412k
Title Fig. 4
Caption (A) Canaletto, view of Canal Grande from Balbi Palace to Rialto (1720‑23). (B) The green-brown belt of the algae is visible as highlighted by the CM in the enlarged detail ; SL is the sea level in the painting.
Credits Ca’ Rezzonico, Museo del Settecento Veneziano, Venice, oil on canvas, inv. Cl. I, n. 2325, 144 x 207 cm.
URL http://journals.openedition.org/mediterranee/docannexe/image/7005/img-5.png
File image/png, 980k
Title Fig. 5
Caption (A) Grassi Palace, on the Canal Grande, with most of the water stair presently under water. A wooden wharf for boat docking has been added. A Police Frogman is standing on the lowest step and is measuring the submersion depth relative to the high-tide level i.e. CM.(B) In a secondary canal, a small five-step water stair made of Istria stone abutting a vertical brick basement, based on the bottom of the canal, during maintenance and restoration works to reinforce the damaged parts
Credits Courtesy of Daniele Resini ©, Insula spa, Venice.
URL http://journals.openedition.org/mediterranee/docannexe/image/7005/img-6.png
File image/png, 860k
Title Fig. 6
Caption Relative sea-level rise at Venice indicated by the tide gauge (1872‑onwards, black & white circles), a measurement of CM by Zendrini (green circle), the 1727-1758 Canaletto and Bellotto paintings (pink squares), the 1571 Veronese painting (blue square) and the lowest step of the water stairs on the Canal Grande (triangles : black for normal palaces, red for monumental outliers). The thick exponential line in the middle is the best fit calculated with the black dots and represents the average tide level ; the upper and lower thin lines represent the average high and low‑tide levels, respectively, and the band of the average tidal range.
URL http://journals.openedition.org/mediterranee/docannexe/image/7005/img-7.png
File image/png, 137k
Title Fig. 7 – Frequency of sea-surge flooding in Venice
Caption Vertical grey lines : floods ; thick black line : 10-yr running average of flood frequency ; thick grey line : exponential fit. The Spörer, Maunder and Dalton Minima of solar activity are also indicated with horizontal bars.
URL http://journals.openedition.org/mediterranee/docannexe/image/7005/img-8.png
File image/png, 104k
Title Table 2 – 
Caption Classification of temperature readings in terms of Standard Deviation (SD) classes and the population of readings falling within each class.
URL http://journals.openedition.org/mediterranee/docannexe/image/7005/img-9.png
File image/png, 7.9k
Title Table 3 – The longest instrumental series in Italy
URL http://journals.openedition.org/mediterranee/docannexe/image/7005/img-10.png
File image/png, 18k
Title Fig. 8 – Temperature anomaly for Northern-Central Italy.
Caption The 1500-1653 and 1671-1715 periods are based on documentary proxy data (P), grey background. In this period only year-to-year variability of unusual temperatures is possible. The 1654-1670 and 1716-185 periods are based on instrumental readings (I), white background. In this period, cycles and long-term variability are detectable. The most dominant feature is the presence of repeated warm‑cold swings that characterize almost all seasons.
URL http://journals.openedition.org/mediterranee/docannexe/image/7005/img-11.png
File image/png, 78k
Title Fig. 9 – Extreme events outside the 10-percentile and 90-percentiles in the 1961-1990 reference period (RP).
Caption The bars represent the number of outliers over 30-yr intervals starting from 1500. Red : warm season outliers (i.e. exceeding the 90-percentile RP) ; blue : cold season outliers (lower than 10-percentile RP). The mean of the positive blue bars increased the frequency of seasons colder than the reference level ; the negative blue bars decreased the frequency of them. Positive red bars mean increased frequency of seasons warmer than the reference level ; the negative red bars decreased frequency of them.
URL http://journals.openedition.org/mediterranee/docannexe/image/7005/img-12.png
File image/png, 180k
Title Table 4 – Significant peaks in quasi-periodicities of swings of air temperature (T) and SST in the Mediterranean and NAO.
Caption In parenthesis some multiples of an observed period.
URL http://journals.openedition.org/mediterranee/docannexe/image/7005/img-13.png
File image/png, 4.3k
Title Fig. 10
Caption Difference between observed temperature (documentary proxy data (P) and instrumental readings (I)) and MTRM calculations : LUTERBACHER et al. (2004, black line) and XOPLAKI et al. (2005, grey line).
URL http://journals.openedition.org/mediterranee/docannexe/image/7005/img-14.png
File image/png, 146k
Top of page

References

Electronic reference

Dario Camuffo, Chiara Bertolin, Patrizia Schenal, Alberto Craievich and Rossella Granziero, « The Little Ice Age in Italy from documentary proxies and early instrumental records », Méditerranée [Online], 122 | 2014, Online since 19 June 2016, connection on 22 January 2018. URL : http://journals.openedition.org/mediterranee/7005 ; DOI : 10.4000/mediterranee.7005

Top of page

About the authors

Dario Camuffo

Institute of Atmospheric Science and Climate, Italian National Research Council (ISAC-CNR), Padova Italy, d.camuffo@isac.cnr.it

By this author

Chiara Bertolin

Institute of Atmospheric Science and Climate, Italian National Research Council (ISAC-CNR), Padova Italy, c.bertolin@isac.cnr.it

By this author

Patrizia Schenal

Free Lance Architect, litasch31@vodafone.it

Alberto Craievich

CA REZZONICO, FONDAZIONE MUSEI VENEZIANI, VENICE, ITALY, alberto.craievich@fmcvenezia.it

By this author

Rossella Granziero

Correr Museum, Fondazione Musei Veneziani, Venice, Italy, rossella.granziero@fmcvenezia.it

By this author

Top of page

Copyright

Tous droits réservés

Top of page