formato pdf, 21 977 KB)Abstract
LISBON ’S REGIONAL AND LOCAL CLIMATE.
THERMAL CONTRASTS AND RHYTHMS
The results of a regional and local climatological study are presented here. Regional research has centred on temperature and the elements which determine it and/or contribute to its spatial and temporal variations (radiation, nebulosity, fogs and winds).
In the first four chapters, a study of Lisbon’s regional climate is carried out. The study area has been extended somewhat to the North to include an «Atlantic» meteorological station (Cabo Carvoeiro); inland and southward boundaries are dependent upon climatic data availability.
Wind records are analysed in the first chapter. At 500 hPa, westerly winds prevail all year round (fig. 1.1). Wind roses referring to surface data indicate that the prevailing winds remain the same during the cold half of the year but that they mostly blow from the North, the Northeast and the Northwest in Spring and Summer (figs. 1.3 to 1.6). However, at certain meteorological stations, wind direction is strongly dependent on local relief (for instance OEIRAS). From March onwards, there is a great increase in the frequency of northerly winds, which disturb the young buds’ normal growth and gives rise to wind-shaped trees. Study of these trees reveals the direction and relative intensity of prevailing winds in detail and will be referred to in the last chapter.
As the radiation values are only measured in Lisbon (Lisboa/Geofisico, located at the Botanical Gardens), a detailed study of insolation (I) has been carried out (chapter 2). The winter month mean values are of no significance because the variable distribution has two modes (the highest frequency of very low and very high values). In Summer the mean values and the mode are high and nearly identical (fig. 2.4). In winter as well as in summer, sunny days (1≥0.80) are more frequent than very cloudy ones (I ≥0.20). The frequency of sunny days is even greater in Lisbon than in the « Cote d’Azur», which is well known for its comfortable winter conditions. The succession of sunny days (fig. 2.6) is longer than that of very cloudy ones: the longest sequence of sunny days reached 11 days in December, 9 in January, 12 in February and 9 in March, while in the same months the longest sequences of «bad» weather did not exceed 7, 7, 6 and 5 days during the 40 years studied on a daily basis. From an insolation ratio «probability calendar» (fig. 2.5, CH. P. PEGUY’s technique), a relationship between insolation values and monthly patterns of synoptical circulation has been briefly presented, and the utility of the graphic presentation mentioned above has been emphasized (for the location of solar plants, for example).
Although Monte Estoril, Lisbon and Montijo are the most favourable sites when it comes to insolation (figs. 2.8 and 2.9), no large differences were found. However, a principal component analysis clearly showed a contrast between the Tagus banks, ill-favoured by insolation in winter (on account of the frequent irradiation and advection fags), and the oceanic sites, where more sunny days are to be found during the coldest part of the year. In summer, the contrary is to be expected (compare 0/S and AZ, figs. 2.8 and 2.9). Linear regressions (valid for summer months) between insolation and fog on the one hand, and insolation and strong nebulosity on the other, permit the insolation to be determined at the meteorological stations where only fog and/or nebulosity are measured.
The chapter concerning temperature (chapter 3) has been based on long trend series from Lisboa/Geofisico (126 years) and on minimum and maximum mean temperatures from 28 climatological stations.
The long-time series have led to the conclusion that there was a global but very irregular temperature increase during the last century. The highest values of the mean maximal temperature took place during the decade embracing 1935-45, as happened in many other places in Europe. The highest minimal temperatures occurred earlier in time: between 1910 and 1920. After a slight decrease during the following decades, a recent rise has been detected. A global rise of 1.2ºC has been determined by means of a linear regression (both for minimum and maximum temperatures). For the other meteorological stations whose seasonal regime has been analysed, only short series were available at the time the data was collected (1983) and the period 1947/48-1974/75 was referred to. This was a period of rather large interannual variation, but no particular trend was discovered.
Thermal seasons have been described using the mean values and standard deviation of maximum and minimum temperatures (fig. 3.6) and the probability calendars (figs. 3.7 to 3.10) referring to Lisboa (Geofisico). The warmest season extends from the second decade of July to the beginning of September. During this period, 5' of maximum temperatures exceed 32ºC and 10% are higher than 30ºC. Minimal temperature dispersion is lower than that of the maximums in Summer (see fig. 3.6) and the daily temperature range is the highest of the year (10% ≥ 14ºC).
Maximum temperature decrease in Autumn sets in earlier than the minimum temperature drop. The autumnal decline in temperature is more rapid than the rise during spring.
The reader’s attention is directed to «St.Martin’s Summer» (similar to the Indian Summer of North America), which occurs every second year (fig. 3.7). The cold season is brief: from the second half of December to the end of January low temperatures are frequent. At the beginning of January, there is a 15% probability of minimum temperatures under 4ºC occurring. The temperature increase in Spring is slow and irregular (figs. 3.7 and 3.8); the greatest variability is to be found in April. A statistical study of frost risk (based upon first and last frost dates, fig. 3.11) revealed that early and late frosts do not occur in more than 95% of the years studied.
Temperature study has been extended to the other meteorological stations. The results of a principal component analysis is shown in fig. 3.13; it is possible to separate the days according to their heat or cold intensity and also according to the particularities of seasonal temperature variation (figs. 3.13, 3.15 and 3.16). Multiple regressions have been used to quantify the effect of latitude, altitude and distance from the sea on temperature. Latitude variation is unimportant in this fairly small region. Maxi-mum summer temperatures grow higher as the distance from the sea increases (r = 0.8); minimum winter temperatures decrease towards inland, but the correlation coefficient is not very high in absolute values (-0.5). The influence of altitude is greater on winter days than on summer ones; a numerical relationship between temperature and altitude has not been found during winter nights, and this fact is most likely due to urban and topographical reasons; this assumption has been verified by calculating residues (figs. 3.17 and 3.18).
An analysis of daily data has been carried out in order to reveal phenomena hidden by the mean values (chapter four). Minimum and maximum temperatures of about 250 days and 250 nights (summer and winter) from 20 meteorological stations have been treated by means of principal analysis components. Groups of days with the same regional temperature pattern have been related to synoptical conditions, occurring at the same time (figs. 4.6, 4.7, 4.13, 4.15). At the end of the fourth chapter, detailed conclusions are presented concerning the causes of the daily temperature values and synoptical situations which correspond to the largest and slightest regional thermal contrasts.
At the end of this chapter, three regions are selected on the basis of their contrasting thermal behaviour: the coastal areas, the Tagus Valley and Lisbon. As the local climatic study of them all could not be handled here, Lisbon’s urban climate has been studied in more detail (chapters 5, 6 and 7). In the last chapter, a preliminary study of weather along the western coast is presented.
Lisbon ’s climate depends a great deal upon its position near the Tagus estuary and not far from the Atlantic Ocean; however the Serra de Monsanto (west of the city) and the Serra da Carregueira (to the Northwest) and even the more distant Serra de Sintra isolate the urban area from the maritime influence.
Within the urban area’s limits the climate also depends on relief to a great extent, especially in the southern districts (fig. 5.1). Urban morphology has been briefly presented so that some of the temperature variations may be understood.
As air pollution greatly influences the radiation balance, a brief study of smoke concentration is included here. A high concentration of this pollutant reduces solar radiation from reaching the earth’s surface (lowering daily heating) and thus leads to a decrease in terrestrial irradiation. This effect is also apparent at night when the temperature decrease is slowed down. Research was carried out on a daily basis: the concentration of smoke is by far the highest downtown in the «Baixa» (eccentric C.B.D.): there, the concentration of smoke can be more than ten times higher than that of the northern measurement points (fig. 5.4). The concentration of smoke (as well as SO 2, CO 2, NO 2, etc.) is much higher in winter than in summer during which northerly winds are fortunately frequent and strong (chapter 1).
Field measurements led to the recognition of some urban thermal patterns; their presentation in chapter 6 does not claim to be exhaustive and the different types of temperature distribution may not occur most frequently. However, it was confirmed that air temperature from the city-centre’s boundary layer (or canopy boundary layer) is frequently higher than that of the countryside, especially during the night. Heat island intensity seldom exceeds 4ºC. Nevertheless, on particular occasions, such as in the presence of fog over the Tagus estuary, temperature differences reach 10ºC; although in this case the warmer areas are Monsanto and the NW districts of Lisbon and not the Centre. Field data was assembled according to the season (Summer and Winter) and the time of day (middle of the day and dawn).
During very windy summer days, there is no real urban effect on temperature. The city acts as an obstacle to the wind: the highest temperatures are found in the more protected localities of the city. The highest temperatures were measured in the southern districts, which coincide with the city-centre (fig. 6.11). When the atmosphere is calm during the morning, a slight breeze manages to reach the southern Lisbon districts. However, the urban thermal pattern varies continuously from the early after-noon to the evening. A kind of «battle» takes place between the estuary and sea-breeze (blowing weakly from the SE in the morning and then from the South and SW) and the gale-force North winds, whose .speed increases gradually. The North winds win the «battle» and sooner or later during the afternoon they blow throughout the urban areas. When the estuarine or maritime air reaches the southern districts of Lisbon the temperature may be 2 to 5ºC lower than that at the Airport (meteorological station used as a «standard» to calculate temperature deviations). But the positive temperature deviations seldom exceed 1ºC (figs. 6.4, 6.6, 6.7) The core of the heat island is then situated over the northern parts of Lisbon. These results were verified by Constant Level Balloons in the Summer of 1987; on that occasion, the thickness of heated urban air reached 500 m over the northern districts of Lisbon (figs. 6.19 and 6.20). The first thermal pattern to be described occurs on at least 43% of the occasions and the second on 39%. As is pointed out in the text, these frequencies will be confirmed with further field work.
During the summer nights, there is nearly always a temperature increase towards the centre. On windy nights (11.5%) the southern and southeasterly districts are the warmest; there are large differences among the different places located nearby, but the highest temperatures were found in the down-town valleys (Baixa, Restauradores, etc., protected from the wind). On calm, humid nights (e.g. 2/8/85, fig. 6.31, circa 8.5% a of nightly occasions), the highest temperatures were measured at the top of the downtown hills (Camões). It was found that the estuarine air which invaded the city during the day accumulated along the valley floors. On other occasions, this cooler and more humid air accumulated along the banks of the Tagus causing a temperature decrease towards the centre (e.g. 29/6/83, 15' of the occasions). During the rare (4.5%a) very warm and dry nights, heat island intensity exceeded 4 or even 5ºC in some places (e.g. 27/7/81, fig. 6.28). Urban downtown valleys were the warmest places.
The mean thermal pattern is shown in figs. 6.33 and 6.34. The tentacle-like form of Lisbon’s heat island (along the main streets) should be noted. The easterly part of Lisbon is warmer than the westerly part, as it is more protected from the maritime air.
During winter days, two different thermal patterns were detected (as on summer days). When the wind blew from the North and Northwest, there was generally a slight increase in temperature towards the Centre (15/1/83 and 24/1/83, fig. 6.39, 28Vo of winter days). On foggy days (or when an advection of humid air took place), the waterside areas (in which the centre was included) were the coldest and most humid; temperature increased with altitude and distance from the river (fig. 6.42, 6% of winter days).
Night-time winter field measurements revealed that there was always an air temperature increase towards the Centre. However, the warmer areas were not always the same. When the wind was strong, the temperature became higher over the central valley beds either on very cold nights (11/2/83 and 21/1/83, which represent 9% of the population) or cold ones (8 and 10/1/83, 12%).
On the other hand, when there was no wind, the tops of urban hills (Largo de Camões, Campo de Ourique) were the warmest places. The valley beds were the coldest places, (on account of cold air drainage and concentration), particularly if the construction density was low. This pattern occurred under anticyclonic conditions, on cold nights (17 and 12 January 1983 and 2 March 1984 for example, figs. 6.48 and 6.52, 27Vo) as well as on abnormally warm ones (1/2/85). The analysis of an infra-red image of the eastern part of Lisbon confirmed some of these results (fig. 6.58 and M. J. ALCOFORADO, 1986). The importance of urban build-up structure on air temperature was pointed out: in some parts of the suburban area with high building density, the night-time temperature could be as high as in some city districts, although it was lower than the core of the urban heat island (fig 6.57).
In order to provide continuous information, two thermo-hygrographs were placed in urban districts. One of them in a very narrow canyon in a relatively old quarter of the city (Bairro Alto) and the other in the yard of a more recent building dating from the nineteen fifties and situated north of the Centre (Av. de Berna). Temperature deviations (city station temperature - Airport temperature at the same time) were calculated and analysed in chapter 7. The relationship of the value and the signal of the deviations and the meteorological parameters have been established by visual comparison of thermograph records and by calculating the correlation values between the two above-mentioned variables (temperature deviations every 4 hour and meteorological data). As an illustration of the results, the importance of nebulosity in Winter in the Bairro Alto has been briefly referred to. During winter days, the Bairro Alto thermograph receives very few hours of direct solar radiation (from 11 a.m. to 1 p.m. in January): as a consequence, temperature differences on cloudy winter days are smaller than on sunny days (negative correlation coefficients between temperature differences and nebulosity) except for the 4 p.m. observations (positive correlation); at 4 p.m. the sun is still shining at the Airport, while the Bairro Alto temperature has decreased since 1 p.m. (mask effect). These facts help to explain why temperature deviations were high and negative on sunny days at 4 p.m. On cloudy days, the cooling rate is the same and the temperature differences are smaller. The importance of wind direction is greater during the summer afternoons: temperature differences are greater when the winds blow from the W, NW and N and when their speed is higher (Bairro Alto is protected from cooler winds). We find the same effect during winter nights: temperature differences are larger (Bairro Alto is warmer) during easterly wind conditions. Tables with the correlation coefficients and the regression equations have been published elsewhere (M. J. ALCOFORADO, 1988)
These observations confirm the result of field measurements: night-time heat islands are more frequent and their intensity reaches 3 or 4ºC in winter as well as in summer.
Day-time thermal patterns are more unsteady. The northern districts are frequently warmer than the Airport. Diurnal temperature increase towards the citycentre is generally due to protection from North winds.
In the last chapter, the importance of local climatic study for tourism is emphasized. A coastal area west from Lisbon was chosen as a sample study-area. Local inquiries involving tourists clearly showed the enormous influence of the wind on human comfort at Guincho (fig. 8.1).
The conclusions of a previous study of the prevailing winds (M. J. ALCOFORADO, 1984) have subsequently been completed by providing field measurements on temperature, humidity and wind, both on pleasant days and on uncomfortable ones, where not many people are able to stand the strong wind, which can reach 80km/h. At the Airport, there are small temperature and humidity differences between «good» and «bad» days at the beach. The wind generally blows from the NW both in the morning and the afternoon during the most unpleasant occasions; when it comes from the east (surface and 850 hPa records), fine weather can be expected at Guincho. The atmospheric pressure difference between a coastal and an inland meteorological station was circa 2 hPa lower on occasions where it was calm at Guincho. The field measurements (fig. 8.9) also proved that at Praia Grande (leeward of the mountain), the weather was cooler but nearly always calmer than at Guincho (as was shown by wind-shaped trees). On windy days at Guincho, this practical experience causes a sort of South-North migration on the western edge of the Serra de Sintra. It is our aim to go further into this research in order to draw up a «real-time» forecast that could be useful to tourists.
Finally, future research openings have been suggested: either going deeper into certain points which have been insufficiently developed here, or undertaking a study of a related field: the complex relations-hip between climate and humans (comfort, architecture, tourism, health, farming). However, further work should be based on more modern documents (as digital terrain models), better meteorological instruments and more advanced treatment of statistical data.