Theor. Appl. Climatol. 66, 49±60 (2000)

1

Istituto ISAO ± CNR, Bologna, Italy Osservatorio Astronomico di Brera, Milan, Italy 3 Istituto di Fisica Generale Applicata, Milan, Italy 2

Trends of Minimum and Maximum Daily Temperatures in Italy from 1865 to 1996 M. Brunetti1 , L. Buffoni2 , M. Maugeri3 , and T. Nanni1 With 4 Figures Received June 4, 1999 Revised November 2, 1999

Summary Annual and seasonal changes in maximum and minimum temperatures (Tmax and Tmin) and in daily temperature range (DTR) in Italy are investigated. Monthly average series for northern and southern Italy are analysed for evidence of trend. Tmax and Tmin show a positive trend over the period 1865±1996 which is greater in southern Italy than in northern Italy. DTR shows a positive trend, but greater in the North than in the South. There is a positive correlation between DTR and mean monthly temperature especially in spring and in summer, while there is a high signi®cant negative correlation between DTR and monthly precipitation. Analysis of temperature, precipitation and DTR during the period 1865±1996 suggests that a general relationship between the very warm last 15±20 years and an increase in the frequency of sub-tropical anticyclones over the CentralWestern Mediterranean. This relationship is based on the hypothesis that in Italy more frequent sub-tropical anticyclones could have been the most characteristic feature of the warm periods during the last 130 years.

1. Introduction Global surface temperature has increased by about 0.3 to 0.6  C since the late-19th century, and by about 0.2 to 0.3  C over the last 40 years (IPCC, 1996). The warming has not been uniform globally and some areas have cooled. In order to identify the cause of this positive trend it is very important to know the behaviour of

the diurnal temperature range (DTR). There is a general, but not global, tendency for a reduced DTR, at least since the middle of the 20th century, con®rmed with data representing more than 50% of the global land-mass (Easterling et al., 1997). The effect is more pronounced in the Northern Hemisphere than in the Southern hemisphere (Horton, 1995) and is mostly due to the faster rise of night temperatures (Karl et al., 1991; Kukla and Karl, 1993), probably related to an increase in cloud cover (Henderson-Sellers, 1986, 1989, 1992; Hansen et al., 1995, 1997a; Rebetez and Beniston, 1998) that could be linked to an increase in anthropogenic aerosols (Hansen et al., 1997a, 1997b). In some areas DTR has decreased because of depressed daily maxima (Razuvaev et al., 1995). As far as Europe is concerned, the lack of availability of databases that include long and homogeneous mean monthly maximum and minimum temperature series has forced, in recent years, climatologists to work on setting up national databases. Papers based on these databases or on single series have found a decrease of DTR that is more pronounced during the last forty years, although in some European areas the negative trend is not so evident as on the global scale and some areas actually show an increase in DTR (Beninston et al., 1994; Bohm and Auer, 1994;

50

M. Brunetti et al.

Frich, 1994; Lapin and Fasko, 1994; Dessens and Bucher, 1995; Kaas and Frich, 1995; BraÂzdil et al., 1996). In alpine located stations no long term decrease in annual average DTR has been observed (Weber et al., 1994). Easterling et al. (1997) have examined the circulation changes in the Northern Hemisphere and their relationship to the winter DTR over a region that covers all the European territory: the results suggest that strong westerly ¯ow is associated with increased DTR over the Iberian peninsula and decreased DTR over northern Europe and into Russia. For Italy, there is a general lack of information and the only database studied is that provided by the Servizio Meteorologico dell'Areonatuca Militare (1951±1980) and this does not show any signi®cant trend in DTR (Jones, 1995). The aim of this paper is to present a new database of Italian mean monthly maximum and minimum temperatures for the period 1865±1996 and to analyse it in order to identify any trend in annual

and seasonal DTR in Northern and Central-Southern Italy. 2. Data and Methods The ®rst step in our research was to set up a new database of mean monthly minimum and maximum temperatures (hereafter Tmin and Tmax) from 27 Italian stations for the period 1865±1996. The principal source was the Uf®cio Centrale di Ecologia Agraria (UCEA) database that was set up in the 1970s in the frame of a project of reconstruction and digitisation of Italian meteorological series carried out by the National Research Council (CNR). The project supported the digitisation of daily minimum and maximum temperatures for 26 secular series (Anzaldi et al., 1980). As the UCEA daily minimum and maximum temperatures dataset has many gaps, extensive work of data digitisation was necessary and a number of Tmin and Tmax values were recovered from Italian

Table 1. List of the Stations Available

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27

Station Location

Lat.

Arezzo Belluno Bologna Cagliari Cuneo Ferrara Genova L'Aquila Livorno Mantova Messina Milano Napoli Padova Palermo Parma Pavia Perugia Pesaro Piacenza Reggio C. Reggio E. Roma Rovigo Sassari Siracusa Torino

43 270 46 070 44 290 39 120 44 240 44 490 44 240 42 210 43 330 45 090 38 120 45 280 40 510 45 240 38 060 44 480 45 100 43 060 43 520 45 010 38 060 44 420 41 540 45 030 40 430 37 030 45 030

Long.

Climatic Area

Series length

Available data in the period 1865±1996 (%)

11 520 12 130 11 200 9 190 7 310 11 360 9 080 13 240 10 180 10 450 15 300 9 110 14 150 11 520 13 210 10 180 9 090 12 230 12 520 9 430 15 390 10 370 12 280 11 460 8 360 15 170 7 400

S N N S N N N S S N S N S N S N N S N N S N S N S S N

1877±1992 1873±1993 1866±1996 1879±1996 1877±1993 1879±1996 1865±1981 1874±1994 1866±1994 1866±1993 1881±1992 1763±1995 1865±1989 1866±1979 1866±1992 1866±1993 1866±1980 1866±1993 1871±1991 1872±1995 1879±1996 1876±1996 1831±1994 1879±1993 1876±1997 1878±1988 1866±1995

84.7 84.0 100 86.3 86.3 84.7 87.8 82.4 92.4 96.2 84.7 99.2 89.3 87.0 96.9 95.4 86.3 97.7 91.6 93.1 76.3 90.8 87.0 86.3 87.4 74.0 63.0

Trends of Minimum and Maximum Daily Temperatures in Italy from 1865 to 1996

meteorological year books (Ministero di Agricoltura Industria e Commercio, 1865±1878; Uf®cio Centrale di Meteorologia, 1879±1925; Servizio Idrogra®co, 1924±1998). After the new database was set up, it was tested for homogeneity and completed with the same methods described in Lo Vecchio and Nanni (1995), Maugeri and Nanni (1998), and Buffoni et al. (1999) for monthly mean temperature (T ) and monthly precipitation (P). The series features, including the percentage of missing data for each station, are shown in Table 1. To locate the stations we refer to Fig. 1. The same division of Italy into two subregions, as used in Lo Vecchio and Nanni (1995), hereafter indicated by N (Northern Italy: 15 stations) and S (Central and Southern Italy: 12 stations), was also adopted.

51

Compared to the S series, the N series were generally better correlated, on an individual basis, with the averages of the other series of that same subregion (used as references for data completion). For Tmin the median of the N monthly correlation coef®cients with the reference series was 0.84 (1st quartile: 0.77; 3rd quartile: 0.88), while for S the same value was 0.75 (1st quartile: 0.66; 3rd quartile: 0.81). For Tmax the same values were respectively 0.87 (1st quartile: 0.81; 3rd quartile: 0.90) for N and 0.75 (1st quartile: 0.71; 3rd quartile: 0.81) for S. All the correlation coef®cients have con®dence above the 99% level. The completed and homogenised Tmin and Tmax series were used to calculate the mean monthly diurnal temperature range series (DTR) de®ned as the difference between Tmax and Tmin.

Fig. 1. Map of the stations distributed over the Italian territory, numbered as in Table 1

52

M. Brunetti et al.

Following the procedure of Maugeri and Nanni (1998) the series were then averaged over N and S and seasonal and annual anomalies and their 5-y running means were calculated for Tmin, Tmax and DTR. Seasonal and annual N and S average anomalies of Tmin, Tmax and DTR were analysed with the Mann-Kendall non parametric test, as described in Sneyers (1990) to look for any trend. The slopes of the linear trends were calculated by least squares ®tting. The Mann-Kendall test was also used for a progressive analysis of the DTR series, as already done in Maugeri and Nanni (1998) and described in Sneyers (1990). A correlation between seasonal and annual DTR and seasonal and annual precipitation and mean temperature was also performed. In order to better verify how the observed trends change with latitude, we also performed some analyses dividing Italy into smaller subregions. In this case, the correlation among the series of the same subregion increased, but the mean series became more sensitive to errors in the single station series. These errors were present even if the data were homogenised as none homogenisation method allows to completely eliminate the inhomogeneities that can be included in a long series of meteorological data (Easterling et al., 1996). Given that homogenisation has been performed, the effect of the errors in the station series can be minimised only by averaging over a suf®cient number of stations (Easterling et al., 1996). Unfortunately, in our situation, any attempt to increase spatial resolution causes a signi®cant reduction in the reliability of the average series and if we use climatic areas as suggested in literature (Mennella, 1967; Pinna, 1977) some areas in Central and Southern Italy would include no stations. With our data, the best classi®cation allowing reasonable average series is that obtained by dividing Italy into three subregions: North (latitude > 43.5: 15 stations), Centre (40 < latitude < 43.5: 7 stations) and South (latitude < 40: 5 stations). However this classi®cation also presents some problems as the Southern Italy average series is too much sensitive to each of the ®ve series in that region. As a consequence, even if some analyses show that the observed temperature trend increases progressively from Northern to Southern Italy, we present and discuss only the results we obtain by dividing the country into N and S subregions.

3. Results and Discussion 3.1 Trends Figure 2 shows the 5-y running means of annual and seasonal anomalies of Tmin and Tmax for N and S and their linear trends calculated by least squares ®tting. The behaviour of Tmin and Tmax is highly correlated for both N and S and, as for T (Maugeri and Nanni, 1998), all the series show a positive trend. The results of the Mann-Kendall test on seasonal and yearly Tmax, Tmin and DTR series for N and S are shown in Table 2. The trends of Tmax and Tmin are always signi®cant (signi®cance level > 95%) except for summer in N (both for Tmax and Tmin) and for spring in N (for Tmin only). The most important contribution to the positive trend is due to roughly the last 20 years for N and to roughly the last 50 years for S as indicated by the progressive application of the test. The linear regression coef®cients (Table 2) are generally greater for Tmax than for Tmin, so DTR has a positive trend with the only exception being winter in N (negative). The DTR positive trend has a signi®cance level greater than 95% for spring, summer and year in N and for autumn and year in S. For Tmax, the slopes of the annual series range between 0.44  C/100y for N and 0.58  C/100y for S resulting in increases over the period 1865±1996 of 0.58  C and 0.77  C. For Tmin the same values range between 0.27  C/100y for N and 0.49  C/ 100y for S resulting in increases of 0.36  C and 0.65  C over the study period. As a consequence, the increase in the DTR in the period 1865±1996 is very week, but signi®cant (0.22  C for N and 0.12  C for S). Comparison of these results with other studies shows that the Italian situation is anomalous because the DTR is generally characterised by a negative trend (Karl et al., 1993; Easterling et al., 1997). A detailed comparison of the results for the period 1865±1996 is, however, hampered by the lack of data. Considering only the sub-series relative to the period 1951±1996 and considering the mean anomalies of N and S annual DTR, the Mann-Kendall test gives a negative but not signi®cant value (ÿ0.7) that is not in contrast to the results of other authors for Italy and some neighbouring areas (Karl et al., 1993; Beninston et al., 1994; Bohm and Auer, 1994; Weber et al.,

Trends of Minimum and Maximum Daily Temperatures in Italy from 1865 to 1996

53

Fig. 2. 5-y running means of yearly and seasonal anomalies of monthly mean minima (continuous line) and maxima (dashed line) temperature and their linear trends (continuous and dashed straight lines respectively) calculated by least squares ®tting

54

M. Brunetti et al.

Table 2. Yearly and Seasonal Minimum and Maximum Temperature and Daily Temperature Range Trends for Northern and Southern Italy (Period 1865±1996) De®ned by: Linear Regression Coef®cient (b), Associated Error (b), Associated Linear Correlation Coef®cient (r) and Mann-Kendall Test Value (u); the Signi®cance Level of u is Always Greater than 95% (ˆ Signi®cance Level Greater than 99%) T min Winter

North South North South North South North South North South

Spring Summer Autumn Year

T max Winter

North South North South North South North South North South

Spring Summer Autumn Year

DTR Winter Spring Summer Autumn Year

North South North South North South North South North South

b  b ( C/100y)

r

u

0.69  0.14 0.63  0.09

0.40 0.52

4.6 5.8

0.40  0.05 ± 0.46  0.08 0.29  0.01 0.45  0.10 0.27  0.07 0.49  0.05

0.55 ± 0.45 0.24 0.38 0.34 0.65

6.8 ± 4.8 2.9 4.7 3.2 7.8

b  b ( C/100y)

r

u

0.60  0.15 0.75  0.10 0.48  0.14 0.35  0.10 ± 0.5  0.1 0.5  0.1 0.74  0.10 0.44  0.10 0.58  0.07

0.34 0.56 0.28 0.31 ± 0.34 0.35 0.52 0.38 0.57

3.4 6.3 3.6 3.5 ± 4.1 3.5 6.2 4.3 6.5

b  b ( C/100y)

r

u

ÿ0.10  0.07

0.12

ÿ2.1

0.38  0.01 ± 0.20  0.08 ±

0.33 ± 0.22 ±

3.3 ± 2.1 ±

0.29  0.06 0.17  0.07 0.09  0.06

0.20 0.22 0.15

5.4 2.2 2.4

1994; Jones, 1995; BraÂzdil et al., 1996; Easterling et al., 1997). For the whole study period, both the N and S DTR series have interesting similarities with the Prague Klementinum series (Karl et al., 1993; Brazdil et al., 1996). A more detailed analysis of the joint behaviour of Tmin and Tmax can be obtained with the progressive application of the Mann-Kendall test to

the DTR series. The results of this test are shown in Fig. 3, the graphical representation of the direct (ui) and backward …u0i † series for N and S annual and seasonal DTR. All the ui curves, with the exception of winter, have an initial decreasing trend, indicating that in the last decades of the 19th century the DTR trend has generally been negative. After the initial

Trends of Minimum and Maximum Daily Temperatures in Italy from 1865 to 1996

55

Fig. 3. Progressive values of ui (black line) and u 0i (grey line) for yearly and seasonal DTR. Straight lines indicate a con®dence level of 95%

56

M. Brunetti et al.

decrease, all the seasonal ui curves (except winter) begin to increase, intersecting the u0i curves at a point corresponding to a date within the period 1920±1940 for N and 1900±1920 for S. They both continue to increase until around 1970 in N and 1950 in S. In the last decades the curves are generally constant in N, whereas in S the DTR is constant only in autumn and decreases in spring and in summer. In winter both the N and S ui series oscillate around signi®cant and non signi®cant values with a tendency towards negative trend in N and positive trend in S. It is not easy to understand the causes of the positive trend in DTR in Italy over the period 1865± 1996. It is, however, worth noticing that an analysis of Italian cloud cover series for the period 1951± 1990 shows a decreasing trend (Piervitali et al., 1997), giving evidence of anomalous behaviour during a period for which an increasing trend in cloud cover is observed in many other areas (Henderson-Sellers, 1986, 1989, 1992; Jones and Henderson-Sellers, 1992; Karl et al., 1993). Therefore, at least in during recent decades, the Italian situation is not in contrast to the relationship between cloud cover and DTR as described in Hansen et al. (1995, 1997a). Anomalous circulation over south-western Europe, at least during winter, has also been identi®ed by Easterling et al. (1997). They examined the relationship between circulation changes in the Northern Hemisphere and winter DTR over an area that covers all of Europe. They found a dipolar pattern in the correlation between DTR and a westerly index: over the Iberian peninsula the correlation is positive whereas over northern Europe and into Russia it is negative. 3.2 Relationships Between DTR, Precipitation and Mean Temperature In a preceding paper (Brunetti et al., 1999), we observed that in Italy during the period 1865±1995, seasonal mean temperature (T ) and precipitation (P) series show negative correlation on both annual and secular time scales, with the only exception being winter in northern Italy. Here we compare the behaviour of T, P and DTR during the period 1865± 1996. Table 3 shows that the correlation between annual and seasonal DTR and P is always negative and highly signi®cant ( > 99%) whereas the one between DTR and T (Table 3) is positive

Table 3. Correlation Coef®cients (r) Between Precipitation and Diurnal Temperature Range and Between Mean Temperature and Diurnal Temperature Range. The Bartlet Signi®cance Level of r is Always Greater than 95% (ˆ Signi®cance Level Greater than 99%) Winter Spring Summer Autumn Year

North South North South North South North South North South

DTR-P

DTR-T

ÿ0.69 ÿ0.61 ÿ0.71 ÿ0.61 ÿ0.58 ÿ0.52 ÿ0.69 ÿ0.44 ÿ0.61 ÿ0.37

± ± 0.44 0.38 0.42 0.46 ± 0.27 0.29 0.29

(signi®cance > 95%) in spring and summer for N and in spring, summer and autumn for S. The negative P-DTR correlation is more signi®cant in N than in S whereas the positive T-DTR relationship is comparable in both geographical areas. As in Brunetti et al. (1999), in order to obtain more detailed information about the joint behaviour of the series, we ®tted 60-y running interpolated 3rd degree polynomial curves to the series and considered both the ®tted curves and the differences between the original series and the curves (residuals). The correlation of the residuals was generally similar to that of the original series, showing that the coef®cients in Table 3 provide an indication of the relationship between T, P and DTR and their high frequency variability. However, the T, P and DTR correlation is not only a short period feature, as the analysis of the polynomial interpolating curves shows that, especially in spring and in summer (Fig. 4), the same relationship present on a annual time scales is also present on longer time scales. The negative correlation between DTR-P high frequency variability can easily be explained on the basis of the well known relationship between DTR and cloud cover (Karl et al., 1993; Hansen et al., 1995; Brazdil et al., 1996; Hansen et al., 1997a; Rebetez and Beninston, 1998), whereas the positive DTR-T relationship seems to be a more characteristic feature of Italy. This result can be explained on the basis of the atmospheric circulation over the Mediterranean region. Over Italy,

Trends of Minimum and Maximum Daily Temperatures in Italy from 1865 to 1996

57

Fig. 4. 60-y running interpolating third degree polynomial curves of temperature (thin black line), precipitation (grey line) and daily temperature range (thick black line)

58

M. Brunetti et al.

especially in spring and in summer, dry conditions are generally associated with warm sub-tropical anticyclones that cause clear skies and a higher DTR, whereas wet conditions (overcast sky and lower DTR) are usually connected with advection of cool air and low temperatures (Menella, 1967; Pinna, 1977). The relationship between warm subtropical anticyclones and clear skies can probably also explain the S autumn T-DTR positive correlation because during this season the expansion of sub-tropical anticyclones can also produce pressure patterns in S which are similar to those of summer (Rapolla et al., 1994). The interpretation of the long term results (interpolated curves) is more dif®cult, as studies on the long term evolution of atmospheric circulation patterns, for example Bardossy's and Caspary's (1990), are not available for the Mediterranean region. As far as this area is concerned, climate during recent decades is characterised by an increase in air pressure (at the surface and upper levels), mean air temperature and probably also summer heat waves, and by a reduction in cloudiness, precipitation and strong cyclogenetic events (Piervitali et al., 1997; Schonwiese and Rapp, 1997). These results seem to be related to an increase in the frequency and persistence of subtropical anticyclones over the Central-Western Mediterranean basin especially during the last 20 years and in the cold seasons (Piervitali et al., 1997). Identi®cation of a relationship between the very warm most recent 15±20 years and an increase in the frequency and persistence of sub-tropical anticyclones over the Central-Western Mediterranean basin and the analysis of the joint behaviour of T, P and DTR during the period 1865±1996, suggests that such a relationship can be generalised based on the hypothesis that in Italy more frequent sub-tropical anticyclones could have been a characteristic feature of all the warm periods during the last 130 years. This hypothesis could explain the differences in the DTR trend between Italy and Northern Europe and globally. In fact an increase in the frequency of sub-tropical anticyclones on the Central-Western Mediterranean area can be connected with stronger westerly ¯ow over Central-Northern Europe, as was observed (especially in winter) in the last 15 years (Bardossy and Caspary, 1990), when warm and very wet winters (low DTR) occurred over Germany (Caspary, 1996, 1998; SchoÈnwiese and

Rapp, 1997), whereas in Italy (both N and S) temperatures were high and precipitation low (clear sky ± high DTR). The hypothesis is also in agreement with the dipolar pattern described in Easterling et al. (1997) suggesting that strong westerly ¯ow over Central-Northern Europe is associated with increased winter DTR over the Iberian peninsula and decreased DTR over Northern Europe and into Russia. Moreover, it is in agreement with the results of Hurrel (1995) that show a negative winter correlation between the NAO and precipitation in Southern Europe (high NAO index ± frequent anticyclones ± dry weather) and a positive one in Central and Northern Europe (high NAO index ± strong westerlies ± wet weather ± high Tmin ± low DTR). The in¯uence of an increase in the frequency of the anticyclonic conditions over the CentralWestern Mediterranean on the distribution of DTR anomalies in Europe probably depends on both the season in which it occurs and the spatial extension of the phenomenon: winter anticyclonic conditions similar to those observed during recent decades seem to be associated with a decrease in DTR over Central-Northern Europe (stronger westerlies ± increase in cloud cover ± increase in minimum temperature). Summer conditions are probably more connected with an increase in DTR over South-Western Europe (clear sky ± increase in maximum temperature). 4. Conclusions The results contained in this work are as follows: 1. annual and seasonal Tmax and Tmin show a signi®cant ( > 99%) positive trend during the period 1865±1996, greater for Central-Southern (S) than for Northern Italy (N); 2. annual and seasonal DTR shows a signi®cant ( > 95%) positive trend over the period 1865± 1996. The trend is higher for N than for S; it is negative for northern winter and not signi®cant for southern spring and summer; 3. the progressive application of the MannKendall test to Tmax and Tmin shows that the positive trend is due to roughly the last 20 years for N and to the last 50 years for S; 4. the progressive application of the MannKendall test to the DTR series shows that both for N and S the trend is initially negative. For

Trends of Minimum and Maximum Daily Temperatures in Italy from 1865 to 1996

N the trend reaches positive values between 1950 and 1970, then remains constant while for S it reaches positive values between 1930 and 1950, it then decreases in spring and summer and remains constant in autumn. In winter the DTR trend is negative in N and not signi®cant in S; 5. annual and seasonal DTR series show strong negative correlation (signi®cance > 99%) with precipitation (P) series and an interesting correlation (signi®cance > 95% in spring and in summer both for N and S and in autumn for S) with mean temperature (T ) series; 6. the correlation between T, P and DTR is not only a high frequency feature, as especially in spring and in summer the same relationship, present on an annual time scale, is also present on longer time scales. The analysis of the joint behaviour of T, P and DTR over the period 1865±1996 suggests that a relationship between the very warm most recent 15±20 years and an increase in the frequency of sub-tropical anticyclones over the Central-Western Mediterranean area can be generalised. This is suggested by making the hypothesis that in Italy more frequent sub-tropical anticyclones could have been the most characteristic feature of the warm periods during the last 130 years. References Anzaldi, C., Mirri, L., Trevisan, V., 1980: Archivio Storico delle osservazioni meteorologiche, Pubblicazione CNR AQ/5/27, Rome. Bardossy, A., Caspary, H. J., 1990: Detection of climate change in Europe by analysing European atmospheric circulation patterns from 1881 to 1989. Theor. Appl. Climatol., 42, 155±167. Beniston, M., Rebetez, M., Giorgi, F., Marinucci, M. R., 1994: An analysis of regional climate change in Switzerland. Theor. Appl. Climatol., 49, 135±159. B ohm, R., Auer, I., 1994: A search for greenhouse signal using daytime and nighttime temperature series. In: Heino, R. (ed.) Climate Variation in Europe. Helsinki: Painatuskeskus, pp. 9±29. BraÂzdil, R., BudõÂkovaÂ, M., Auer, I., Bohm, R., Cegnar, T., Fasko, P., Lapin, M., Gajic-Capka, M., Zaninovic, K., Koleva, E., Niedzwiedz, T., Ustrnul, Z., Szlai, S., Weber, R. O., 1996: Trends of maximum and minimum daily temperatures in central and southeastern Europe. Int. J. Climatol., 16, 765±782. Brunetti, M., Maugeri, M., Nanni, T., 1999: Variations of temperature and precipitation in Italy from 1866 to 1995. Theor. Appl. Climatol. (in press).

59

Buffoni, L., Maugeri, M., Nanni, T., 1999: Precipitation in Italy from 1833 to 1996. Theor. Appl. Climatol., 63, 33±40. Caspary, H. J., 1996: Recent winter ¯oods in Germany caused by changes in the atmospheric circulation across Europe. Physics and Chemistry of the Earth, 20, 459±462. Caspary, H. J., 1998: Atmospheric circulation changes across Europe as a driving force for an increasing ¯ood risk in southern Germany. In: Lemmela, R. (ed.) Proceedings of The Second International Conference on Climate and Water. Espoo, Finland. Dessens, J., B ucher, A., 1995: Changes in minimum and maximum temperatures at the Pic du Midi in relation with humidity and cloudiness, 1882±1984. Atmos. Res., 37, 147±162. Easterling, D. R., Peterson, T. C., Karl, T. R., 1996: On the development and use of homogenised climate datasets. J. Climate, 9, 1429±1434. Easterling, D. R., Horton, B., Jones, P. D., Peterson, T. C., Karl, T. R., Parker, D. E., Salinger, M. J., Razuvayev, V., Plummer, N., Jamason, P., Folland, C. K., 1997: Maximum and Minimum trends for the globe. Science, 277, 364±367. Frich, P., 1994: Observed trends in the diurnal temperature range in the North Atlantic Region. In: Kukla, G., Karl, T. R., Riches, M. R. (eds.) Asymmetric Change of Daily Temperature Range. College Park, Maryland: U.S. DOE, pp. 109±111. Hansen, J., Sato, M., Ruedy, R., 1995: Long-term changes of the diurnal temperature cycle: implications about mechanism of global climatic change. J. Atmos. Res., 37, 175± 209. Hansen, J., Sato, M., Lacis, A., Ruedy, R., 1997a: The missing climate forcing. Phil. Trans. R. Soc. London B., 352, 231±240. Hansen, J., Sato, M., Ruedy, R., 1997b: Radiative forcing and climate response. J. Geophys. Res., 102/ D6, 6831± 6864. Henderson-Sellers, A., 1986: Cloud changes in a warmer Europe. Clim. Change, 8, 25±52. Henderson-Sellers, A., 1989: North American total cloud amount variations this century. Glob. and Plan. Change, 1, 175±194. Henderson-Sellers, A., 1992: Continental cloudiness changes this century. Geo. Journal, 27.3, 255±262. Horton, E. B., 1995: Geographical distribution of changes in maximum and minimum temperatures. Atmos. Res., 37, 102±117. Hurrel, J. W., 1995: Decadal trends in the North Atlantic Oscillation: regional temperatures and precipitation. Science, 269, 676±679. IPCC, 1996: Climate Change. The IPCC Second Assessment Report, Houghton, J. T., Meira Filho, L. G., Callander, B. A., Harris, N., Kattenberg, A., Maskell, K. (eds.), Cambridge University Press, N.Y. 572 pp. Jones, P. D., 1995: Maximum and minimum temperatures trends in Ireland, Italy, Thailand, Turkey and Bangladesh. Atmos. Res., 37, 67±78. Jones, P. D., Henderson-Sellers, A., 1992: Historical records of cloudiness and sunshine in Australia. J. Climate, 5, 260±267.

60

M. Brunetti et al.: Trends of Minimum and Maximum Daily Temperatures in Italy from 1865 to 1996

Kaas, E., Frich, P., 1995: Diurnal temperature range and cloudcover in the Nordic countries: observed trends and estimates for the future. Atmos. Res., 37, 211±228. Karl, T. R., Kukla, G., Razuvayev, V., Changery, M. J., Quayle, R. G., Heim, R., Jr., Easterling, D. R., Fu, C. B., 1991: Global warming: evidence for asymmetric diurnal temperature change. Geophys. Res. Lett., 18, 2253±2256. Karl, T. R., Jones, P. D., Knight, R. W., Kukla, G., Plummer, N., Razuvayev, V., Gallo, K. P., Lindseay, J., Charlson, R. J., Peterson, T. C., 1993: A new perspective on recent global warming: Asymmetric trends of daily maximum and minimum temperature. Bull. Amer. Meteor. Soc., 74, 1007±1023. Kukla, G., Karl, T. R., 1993: Nighttime warming and the greenhouse effect. Environmental Science and Technology, 27, 1468±1474. Lapin, M., Fasko, P., 1994: Daily air temperature range and relative air humidity correlation at Hurbanovo. In: BraÂzdil, R., KolaÂr, M. (eds.) Contemporary Climatology. Brno, pp. 359±363. Lo Vecchio, G., Nanni, T., 1995: The variation of the atmospheric temperature in Italy during the last one hundred years and its relationship with Solar Output. Theor. Appl. Climatol., 51(3), 159±165. Maugeri, M., Nanni, T., 1998: Italian surface air temperature variations: recent trends and an update to 1993. Theor. Appl. Climatol., 61, 191±196. Mennella, C., 1967: Il Clima d'Italia. Napoli: Fratelli Conti Editori, 724 pp. Ministero d'Agricoltura, Industria e Commercio - Direzione di Statistica 1865±1878: Meteorologia Italiana, Rome. Piervitali, E., Colacino, M., Conte, M., 1997: Signals of climatic change in the Central-Western Mediterranean basin. Theor. Appl. Climatol., 58, 211±219.

Pinna, M., 1977: Climatologia. Torino: UTET, 442 pp. Rapolla, R., Conte, M., Spezie, G., 1994: La Ni~ na phenomenon and some cases of drought in the central Mediterranean. In: Proceedings of the Conference Atmospheric hysics and dynamics in the analysis and prognosis of precipitation ®elds, Universita La Sapienza, Rome. Razuvaev, V. N., Apasova, E. G., Bulygina, O. N., Martuganov, R. A., 1995: Variations in the diurnal temperature range in the European region of the former USSR during the cold season. Atmos. Res., 37, 45±51. Rebetez, M., Beniston, M., 1998: Changes in sunshine duration are correlated with changes in daily temperature range this century: An analysis of Swiss climatological data. Geophys. Res. Lett., 25, 3611±3613. Sch onwiese, C. D., Rapp, J., 1997: Climate Trend Atlas of Europe based on observations 1891±1990. Dordrecht: Kluwer Academic Publisher, 228 pp. Servizio Idrogra®co, 1924±1998: Annali. Rome: Ist. Poligra®co dello Stato. Sneyers, R., 1990: On the Statistical Analysis of Series of Observation. WMO, Technical Note N. 143, Geneve, 192 pp. Uf®cio Centrale di Meteorologia, 1879±1925: Annali, Rome. Weber, R. O., Talkner, P., Stefanicki, G., 1994: Asymmetric diurnal temperature change in the Alpine region. Geophys. Res. Lett., 21, 673±676. Authors' addresses: Michele Brunetti, Teresa Nanni, Istituto ISAO±CNR, via Gobetti 101, I-40129 Bologna, Italy; Letizia Buffoni, Osservatorio Astronomico di Brera, via Brera, 28, I-20121 Milan, Italy; Maurizio Maugeri, Istituto die Fisica Genereale Applicata, via Brera 28, I-20121 Milan, Italy.