Prof. Tarik Rezende de Azevedo, Ph.D.

Laboratory of Climatology and Biogeography

Department of Geography - FFLCH

University of São Paulo

Av. Prof. Lineu Prestes, 338. CEP: 05508-977



We have been studying the spatial distribution of the largest atmospheric precipitation of a day in Metropolitan Region of São Paulo and surroundings as a function of the relative intensity of the urban activity. It classified days with precipitation major than 20 mm and days on which occurred the 300 larger pluviometric totals in the 1990s into "workdays" and "not workdays". The conclusion, in both cases, was that the frequency, the amount precipitated and the intensity of the rain on "workdays" were significantly larger than on "not workdays" at urban area of São Paulo and some adjacent areas.


Do traffic jams in São Paulo are worst when it rains in the evening? Does it rain more in evenings when there is traffic jams? The fact is that it has been raining more in evenings of the central days of the week in Metropolitan Region of São Paulo (RMSP) since the 1990s. This curious result is, actually, just secondary product of more general investigation. Following previous papers, it was introduced new evidences that corroborate the hypothesis that the rhythm of the human activities is one of the determinant elements in the processes that occur in the lower troposphere over great urban areas (AZEVEDO, 2001a). It is introduced the anomaly in the spatial distribution of the largest daily precipitation at RMSP and surroundings as a function of the relative intensity of the urban activity. Besides that, it is showed the method used, some implications of the results yet reached and questions to be investigated in detail.

AZEVEDO (2001b) compares the conservative estimate of the amount of energy dissipated by the population of RMSP in 2000 with the global solar radiation at the ground estimated by FUNARI (1986). In both cases, he considered the accumulated energy in a year at the urban area of RMSP. The author concluded that the first is, at least, 10% of the second one. This procedure applied in regions of tempered climate, where the intensity of sun radiation is smaller and the energy consumption per capita is very greater than in Brazil, carries for estimate that is an increment up to 50% of transit energy in the boundary layer over urban areas (TOJO, 1998). Azevedo (op cit.) argues that the distribution of the energy dissipated by the urban activity is heterogeneous along the days. Being Sunday the day of minor urban activity in RMSP, Wednesday would be the activity maxim day, if it was accumulated the energy dissipated along several months or years. That is because there are several "false holidays" concentrated on Mondays and Fridays. AZEVEDO and TARIFA (2001) introduced ten evidences that there is a weekly rhythm in the own processes of the climate of RMSP, aggregating records of two meteorological stations hourly and per day of week. AZEVEDO (2001c) demonstrated that weekdays and the holiday dates are completely arbitrary regarding the astronomical rhythms and to their consequent climatic rhythms. The author introduced an algorithm to classify days in “workday” and “not workday” for RMSP. That is because, in Brazil, many weekdays become "false holiday" if they occur between holidays. It verified that, in fact, if we consider periods larger than six months, there are more "workdays" in Wednesdays. In general there are about 2,1 "workday" per each "not workday". Aggregating records of two meteorological stations at RMSP, it concludes that the averages temperature of the air, wind speed and precipitation tend to be significantly larger in "workdays", while atmospheric pressure tends to be smaller in recent years.

The increase of the urban activity in "workdays" implies in significant increase of the energy dissipated by the population and its activities. Because the portion of emitted energy that come from burning of fossil combustibles, industrial activity, commerce and services is responsible for more than ¾ of the total amount. In thesis, it should occur increase of the air temperature, local deformation in the barometric field, tendency of increasing convergence of surface winds, intensification of convective processes and increase in frequency and intensity of rainfalls (mainly the concentrated impacts in the afternoon and in the evening). However, the increase of the urban activity also implies in increment in the emission of pollutant and in maintenance of particle material in suspension by the increase of the turbulence close to the ground, as a function of the largest number of vehicles in circulation. The particle material has preponderant role in the urban radiation balance because it implies in attenuation of the sun radiation on the ground. Therefore, it is perfectly acceptable the hypothesis that, taken separately, it acts in the urban climate system reducing indirectly the air temperature near the surface. In other words, under the “domain" of particle material pollution, there should be a tendency to the “stabilization” of the boundary layer, a larger tendency to the formation of thermal inversions and significant reduction of rainfall frequency and intensity.

Azevedo (op cit.) argues that atmospheric pollution characterized RMSP in the 1970s on the "perceptive layer", just as proposed by Monteiro (1976). On those days came effective initiatives of accompaniment and control of the air quality. In fact, the pollution of industrial source decreased in absolute terms at RMSP during the nineties. On the other hand, since 1980s, the individual transport has been increasing a lot. In the RMSP, traffic system reached the saturation, obliging the adoption of the automobile weekly castor. There are at least two dozens of other facts and factors which would not fit argue here. The result is that the consumption of energy per capita grew more than the production of atmospheric pollutant per capita in RMSP in the last years (ESTADO DE SÃO PAULO, 2000). Azevedo (op cit.) verified that the daily extreme temperatures of "workdays" tended to be smaller than in "not workdays" in the period from 1975 to 1986. The inverse was verified in the period from 1987 to 1999. Coherently, the accumulated rain presented the same tendency. However, it was used records of only one meteorological station. The urban area of RMSP has more than two thousand Km2 in several geomorphological compartments, besides a complex distribution of heat and air pollution fonts. To evaluate if both periods, the first - “domain of the antropogenic pollution” - and the second, more recent, - “domain of the antropogenic heat” -, were characterized in the RMSP at all, AZEVEDO (2001d) used the records of 120 pluviometric posts of Waters and Power Department of São Paulo State. It selected the 19 posts, which records cover the period from 1968 to 1997 with less than 1% of lacunas and smaller deviation than 1% in the ratio beyond valid records in "workdays" and "not workdays". It totaled the precipitation of each one of the 19 posts in five consecutive periods of 6 years. To compare the two classes, on each post and in each period, he calculated the percentage deviation regarding the expected distribution. This was done because the probability of any event to occur is larger on a "workday". It totaled on each period the number of post in which the deviation of "workdays" was, with 98% confidence, smaller, equal or larger regarding the "not workdays". The results confirmed the hypothesis. From 1968 to 1979, it rained less on "workdays" on 14 of the 19 posts. From 1986 to 1997, it rained more on "workdays". From 1980 to 1985 it rained on "workdays" and "not workdays" equitably and could be considered a transition period.

AZEVEDO (2002) argues that in the case of the classification of the records per day of week, there is the inconvenience of mixing days of minor urban activity with days of larger urban activity, mainly on Mondays and Fridays ("false holiday" and "true holiday" represents more than 10% these days of week). The average values of precipitation do not show if there is an increase in the frequency and intensity of rainfalls. There is the advantage that seven classes have the same number of samples, facilitating comparisons, mainly in extreme cases. The classification into "workdays" and "not workdays" does not allow directly weekly rhythm characterization. The separation of each day, from Sunday to Saturday, into "workday" or not, results in incomparable classes because of the sampling disparity. Finally, it is possible that the increment or reduction in urban activity on a day depends on the previous day and on the next day also. It is not absurd to say that what is going to happen on the day after interferes on the day at issue, especially because one of characteristics of the human activity is teleology. For example, in RMSP there is a considerable increase on traffic jams in the afternoon and evening of a Thursday that precedes a Friday holiday. Considering this hypothesis, the author elaborated a method to disintegrate "workdays" and "not workdays" in a major number of classes, compatible with the characterization of a weekly rhythm. He proceeded calling the classification of a day as "position in week" and using the notation U for "workday" and N for "not workday". Considering the position in week, there would be eight arrangement possibilities in a three days sequence. However, just six of these sequences, or positions in week, really are significant. In this order represent the rhythm of the relative intensity of the urban activity: NNN, NNU, NUU, UUU, UUN, UNN. The advantage of this classification system is that the "true holiday" and "false holiday" are aggregate with typical days to them related. For example, in a week in which Friday is holiday. The position in week of the previous Thursday is UUN, grouped with Fridays of the normal weeks. The position of Friday in this case is UNN, grouped with Saturdays of the normal weeks. Saturday, then it grouped in a class that does not occur in the normal weeks (NNN - "not workday" preceded and followed by "not workdays"), but that is relatively frequent.

AZEVEDO (2002) aggregated the hourly precipitation data of Água Funda Meteorological Station from the period 1990-2001, separating the precipitation into three classes as a function of the total accumulated per hour. The author demonstrates that the precipitation with 10 or more millimeters per hour, of 16:00 to midnight, occurred in central days of week (UUU) with more than twice the frequency of days NNN, which represents the extreme possible case of reduction of the urban activity. Besides this, the average intensity (represented preliminarily by the reason from the total precipitated by the occurrence per hour) and the daily average precipitation (if considered these more intense rainfalls) are significantly larger in direction to central days of week.

Although all the previous results corroborate the hypothesis that there is an antropogenic rhythm built-in on the RMSP climate, there are still several pertinent doubts to clarify, especially on case of atmospheric precipitation (AZEVEDO and TARIFA, 2001). This paper explores one that is important to the geographer. What would be the space distribution of this increment on intensity and frequency of the largest precipitation on "workdays"? There is a narrow relation between the phenomenon above and the location of the urban area, or do own troposphere and/or boundary layer processes cause it? It is being pointed an embryonic process of conurbation between São Paulo and Rio de Janeiro that incorporates Sorocaba, Santos, Campinas and all the urban agglomerates of Paraiba Valley. The phenomenon of alteration of precipitation structure and temporal distribution, investigated yet at the scope of RMSP, also manifest, at least on an embryonic form, across all the Megalopolis that is in formation? In the case of atmospheric pollution, the preliminary field observation allows us to infer that there is transport of pollutant among areas pointed above during at least a decade. The first try to answer these questions was testing the hypothesis that there should be an area of preferential occurrence of intense precipitation in "workdays" related to the urban area of RMSP, even though not exactly overlapped to the second.


The studied area was defined using three criteria; (a) including RMSP and its surroundings, that means here, the repetition of its extension to North, South, East and West, (b) limited by parallels and meridian in whole degrees to facilitate the search for rain-gauges and (c) it would be used only São Paulo's state posts. The combination of these three criteria resulted in São Paulo's state area defined by 22 S, 25 S, 48 W e 45 W.

The records of daily precipitation of the pluviometric net of Water and Power Department of São Paulo's Government were used. On each post of this net, the rain gauge reading occurs at 7:00 AM of the following day. In the beginnings, the sun rhythm marked the work rhythm, and it is still like this to major part of the humanity. In RMSP, the period of larger urban activity initiates before sunrise. In weekdays, it extends almost until midnight. The influence of the urban activity upon atmosphere prolongs far beyond sunset and maybe beyond midnight. As urban structure introduces high thermal inertia, heat stored during the day continues irradiating at night. The best way to separate days, considering the sun heating rhythm and urban activity simultaneously, is around 6:00 or 7:00 AM, when the effects of sun heating and human activity of previous day achieve the minimum. The best data would be continuous pluviometric records, as argued in previous papers. However, the existing net, with this temporal detail of records, results in a very small number of posts in the studied area, not allowing testing the hypothesis in investigation.

This work adopted the 1990s decade for three reasons. (a) Towards the present, the lacunas in digital records increase, because the information gathered at the posts has to be typed, checked, consisted and made available, and it takes time. (b) The number of setup and/or disabled posts grows exponentially, the bigger the period is. (c) Previous papers, summarized above, argue that the "period of domain of antropogenic heat" was characterized since the last years of the 1980s.

Records of daily precipitation (p), as a function of day position in week, were aggregated, considering the number of record days (T), the number of days with record in which p>=20 mm (N) and the total amount of precipitation on days on which p>=20 mm (P). It was calculated P/N, P/T and T/N to allow the comparison among classes. The same procedure was applied to the 300 larger daily precipitation data in each post. It was done to check if, with same number samples in each post, the space distribution obtained for p>=20 mm kept the same. As it is known, the relief affects precipitation, so there are natural rainier posts. This implies that in the case p>=20 mm, N and P vary a lot among the posts for other factor than the investigated in this work.

The first approach identified 415 pluviometric posts in the study area. It was applied two criteria to incorporate each post. (a) The days with record face to the number of days of the 1990s greater than 0,95. (b) The days with record for "position in week" compared to the number of days at the same "position in week" in the 1990s greater than 0,95. Therefore, the rigidity used to select the posts is a bit smaller than the one applied previously (AZEVEDO, 2001d). As described for the next stage, the absolute values in each post would not be compared, but the tendency surfaces determined with elevated variance. The deviation up to 5% in the initial interclass sampling are incorporated and become negligible. In fact, more than 89% of the remaining posts have valid records in more than 97% of the 1990s. Posts with extraordinary accumulation of precipitation on Monday, preceded by null or insignificant accumulation on Sunday and/or Saturday were also eliminated. On these posts reading on holidays did not occur. It remained 209 pluviometric posts. Although the volume of data was big, the density was small compared to the studied area and the hypothesis in investigation.

Result homogenization was done by reducing it to tendency surfaces, considering deviation and/or random errors in each post. It was applied Kriging Method with linear variogram and high variance in two steps. To the first step, it was applied the standard deviation of the values mapped. To the second, it was applied the half of variance of the same values. This procedure results in a progressive precision reduction of the rugosity position and in an increase of its significance, eliminating random noise in cartogramas. In other words, it suppresses progressively small and not significant variations, simplifying the result interpretation and increasing the probability that remaining wrinkly really represent tendencies that really exist. On the other hand, this method is not similar to linear interpolation for triangulation and manual adjustment of isolinhas. In reality, the products are similar if null error in Kriging Method is admitted. For pluviometric data, although this method is employed indiscriminately, it is not correct to consider null error (AZEVEDO, 2001e). In cartogramas, it was suppressed the null values portion of the tendency surface, using significance margins rounded for the above five multiply. They represented only remaining wrinkly, simplifying interpretation. According to the case, it was assigned neutral or null value to the dot bars in half a degree in portions of the study area that have no pluviometric posts (NE – South of Minas Gerais State, and SE – Atlantic Ocean) before the generation of the tendency surfaces. It forced the automatic exclusion of these areas hardly sampled when applied the significance margins. In synthesis, all the cautions were taken to test if doubt or lack of denser information led to the denial of the hypothesis in investigation. The results introduced ahead should be interpreted as rigorous and conservative. The data were treated in Excel Program (Microsoft) spreadsheets, with search and classification routines in Visual Basic (Microsoft). The cartogramas were generated through 6.0 Surfer Software.


The results are showed in four cartogramas. Cartogramas 1 and 2 represent the percentile deviation of the occurrence, in "workdays", of the 300 larger daily precipitations in each post, compared to the occurrence in "not workdays". In a first step (1) was applied the universe values standard deviation as error margin for trend surface generation. In the second step (2) was applied the half of the same values variance. It was suppressed each tendency surface portion between 20 and -20%. In both case negative portions did not remained. In cartograma 1, there are four areas with significant positive deviation. The most prominent positive area is at cartograma central region and is southwest prolonged. The densest urban area in RMSP is located at the same region. Cotia, Embu and Taboão da Serra municipal districts are at the middle portion of the southwest prolongation. Further on there is no occurrence of significant urban areas. Accurate observation allows realizing that are two saliencies, which, from the central region, point to Campinas and to Paraíba Valley. There is a region with subtle positive deviation at São José dos Campos, the largest urban area of Paraíba Valley. The third positive deviation area is at the north of Sorocaba. The fourth one is at Sorocaba (less significant than in the central region, but clearly characterized). There is no explanation to the southwest widespread of the first and the third area only considering urban areas overlapping. It is possible to infer that there is southwest lateral transport caused by own regional circulation processes that favor cumulus's vertical development in specific synoptic condition. Thus, the third area is better described as the southwest of Metropolitan Region of Campinas. On remaining area is not possible to affirm that major precipitation tended to occur on "workdays" in the 1990s from available pluviometric records. In the pointed areas is very probable that major precipitation occur on "workdays". Cartograma 2 shows that the central salience remain even forcing the tendency surface for null values, when admitting that the error on each post can be the half of all points variance. Although space distribution detail was lost, certainly there was a strong tendency to occur the 300 larger precipitations on "workdays" in 1990s on urban area of RMSP. The largest deviation was at least 40%.

Cartograma 3 is similar to the first one, only changing the represented information. In this case was daily precipitation larger than 20 mm on "workdays" compared with that on "not workdays". The space distribution of the areas is almost the same but the larger deviation is at southwest. The number of considered events is greater than the 300 per post of cartograma 1. However, the selected event quantity varies a lot among the posts. Its not necessary presenting the cartograma obtained forcing the tendency surface for neutral values (in this case 100) because it had similar features of cartograma 2. Therefore, the conclusions from cartogramas 1 and 2 analysis are reinforced.

Cartograma 4 represents the average of daily precipitation larger or equal 20 mm on "workdays" using the same cartogramas 1 and 3 method and criteria. This one has impressive likeness and coherence regarding the other two. The maximum deviation area occurs exactly on RMSP central area. It is the same that average of daily precipitation larger or equal 20 mm was, at least, 40% larger on "workdays" than on "not workdays" in 1990s on RMSP central area.

The original project presupposed that other elements not considered in this paper would significantly affect deviation on largest precipitation frequency. To differentiate the deviation as a function of urban area presence, it was proposed posts separation into three classes (on periphery, in or out of urban areas). Afterwards, morphometrical indices of the frequency curve of each class obtained deviation would be applied. Finally, the mapping would have been done if frequency curves morphometrical indices were significant and coherently different. For our surprise, the calculated deviation resulted eloquent. By the simple exam of pluviometric posts names at the top of a list ordered by deviation, was possible to infer what it would be the surfaces morphology of resultant trend mapping.

To allow a better comprehension of method robustness and the reliability of results presented above, one supposes the same procedures for relief morphological analysis. Take an area which one completely ignores its relief. If pluviometric post altitudes were the values used to generate cartograma 2, the surface tendency would be almost flat with a ENE-WSW scarp. We had tested that. No significant salience remains after the same significance margin application, besides a minuscule area around the Itapeva Mountain (which has a meteorological station at its hilltop, rare case at this kind of elevation). It would be necessary about 20 neighboring posts dispersed on a 1800 meters average altitude mountainous massif at central region, to produce a previous results equivalent salience.


There is an important distinction between total pluviometric accumulated in a day and the concept of intense precipitation. We cannot assume that intense precipitation results in largest pluviometric total. However, it is the best approach to the problem investigated. Even in hourly precipitation record (which is rare in Brazil), the same problem takes place. Strictly, precipitation intensity is the rate of precipitation accumulated along time, varying according to rain duration. Thus, only pluviometric diagrams analysis, and its meticulous reduction, or automatic records in tiny intervals, allows precisely precipitation intensity determination. The common sense is enough to realize that this kind of analysis is very difficult because of research conditions and lack of records. When we assume that it has been occurring intensification of local convection precipitation, as a function of urban activity intensity, in fact, we are trying to prospect indirect evidences to support this hypothesis.

It is possible to find an indirect evidence of precipitation intensification on "workdays" in RMSP comparing cartogramas 3 and 4. It is necessary to consider a subtle fact: the frequency of precipitation with more than 20 mm has been significantly larger (about 40%) in "workdays", not exactly on RMSP central area (cartograma 3). Actually, the top deviation occurs at Southwest of RMSP and decreases to the central area, where it is still positive, but the values are lower, varying from 25 to 30%. The distribution of daily average deviation (that, in this case, is nothing else than a strategy to compensate the different number of "workdays" and "not workdays") follows an inverse tendency. The maximum occurs almost in RMSP central area, and decreases southwest. The middle way depression in both cases is due to pluviometric post absence effect. As it was adopted the principle of denying hypothesis in investigation, in cases of lack of information, the tendency surface was “lowered” at these places. Considering that (1) there is an inverse spatial relationship between frequency and total precipitation, and (2) daily average deviation is greater than frequency deviation, it is possible to conclude that, in central area, the treated precipitation kind has larger intensity. It also reduces progressively towards southwest. Another hypothesis, not very suitable, should be significant increase on the duration of each rainfall in "workdays", compared to "not workdays", towards RMSP central area. If that is true, we should not consider an intensity increase, but larger precipitation frequency and duration in "workdays", that is already very interesting for our investigation.

There are two explanations to southwest deviation of more intense precipitation:

(a) In regional scale, the local convective process intensification depends on troposphere vertical structure and boundary layer, but mainly on secondary circulation. Azevedo (2002) analyzed in detail hourly precipitation at Água Funda meteorological station. Precipitation has been more frequent and intense, resulting in larger amount on central days of the week, mostly in afternoon end and evening, at least at this post. This fact allows us to infer that it really is local convective process precipitation. Most events occur in spring and summer. As the increasing of antropogenic heating effect should occur when low troposphere specific humidity is larger, and considering that there is a deviation towards southwest, part of the events should occur in Tropical Atlantic Mass borders. That is true mainly when its action center is southerly, with NE Winds upon RMSP. In the other hand, part of the events could occur near the instability line between Tropical Continental and Tropical Oceanic, and/or induced instability inside Tropical Atlantic. In unusual cases, these events could occur when the Equatorial Continental dominates the Continent interior. In contact with the Tropical Atlantic, it forms a “corridor” directing the Aliseos into Brazilian southeast. It could occur as well, in current terminology, in presence of ZCAS. Finally, these are just conjectures.

(b) The active element that intensifies most local convective processes is the incoming of oceanic breeze in the Atlantic Plateau in the afternoon and/or the evening. The largest breeze penetration into the continent occurs exactly at São Paulo's Atlantic Plateau area. The breeze frequently reaches the Peripheral Depression border and even advances. Depression border is the limit to the occurrence of stratus and strato-cúmulus associated to breeze by Föhn Effect. Unfortunately, there is no literature about the details of dynamic aspects of breeze incoming in this area. It is probable that breeze fluxes suffer direction changes as a function of regional relief peculiarities. The calm that precedes breeze inversion is the favorable period for exacerbation of city effect on temperature and pressure field. Although not studied yet, there is breeze “rotation” (instead of simple direction inversion), just as registered at Meteorological Station of Água Funda. The incoming of breeze usually occurs as a flux, starting with low velocity, almost parallel to Serra do Mar scarp, in morning end and/or afternoon beginning (from ENE-NE). It gradually “rotates" towards typical direction of ESE-SE. It is possible that precipitation and convective process intensification, treated in this paper, occurs on the period that precedes oceanic breeze incoming.

One has to remind that continental air starts moving before fresher and humid air arrives. When it happen, the process of the breeze is already acting. In addition, it is necessary to consider that RMSP urban area is very widespread. Serra do Mar adjacent areas receive freshest air before NW extreme. Therefore, there is no reference to systematic investigation on this fact in literature. We do not know yet how long the freshest air delay “scanning” the urban area from one side to the other.


A second process, not systematically studied yet, is the low level breeze between Paraíba Valley and São Paulo Basin. Field observation allows us to infer that there is a sensible flux from the Valley to east urban area, in the period of ocean-continent breeze inversion, mostly at morning end and/or afternoon beginning. In this metropolis portion, it results in NE Wind. A full strato-cumulus carpet on low altitude frequently covers the RMSP extreme northeast municipal districts. Sometimes it occurs close to the ground. It derived from air elevation that spills from the Valley in direction to São Paulo Basin (or perhaps is “sucked” from São Paulo Basin, when occurs intensification of the urban breeze in the central days of week (AZEVEDO and TARIFA, op cit.)). It is also possible that this phenomenon is related to the Tropical Atlantic Anticyclone border, or even is an aspect of the continental breeze.

Besides the precipitation kind mentioned above, the occurrence of coldfront incoming generated precipitation must be considered in the universe of analysis. Many coldfront incoming results in intense precipitation and/or elevated totals, but not all extreme events are associated to them (MONTEIRO, 1973). A secondary hypothesis is that treated events occur during the apex of pre-frontal instability. However, there is no explanation for the southwest displacement of phenomenon occurrence area. The most plausible would be a southeast displacement. This displacement to southwest could not be considered, unless the "reflected polar front" occurred with exceptional frequency in the 1990s. On the other hand, maybe the oceanic breeze has an important role in pre-frontal thunderstorms genesis, hypothesis at least not attempted on literature, but plausible. Anyway, coldfront tends to spread precipitation usually in a succession from SW to NE in this region. This should result on different precipitation distribution, regarding the cases mentioned previously.

The rebuilding of hourly synoptic conditions and/or rhythm of secondary circulation, as well as breeze mechanism of a whole decade, is an impractical task. However, it is possible to take a shortcut. Instead of analyzing the whole decade, it could be selected the dates on which occurred the major precipitation. It is still not possible to tell how many days they represent, because the 300 days with larger precipitation are not the same in all posts. The synoptic characterization of secondary circulation rhythm moments that favor more intense precipitation in "workdays" would allow a first approach to the processes from direct object investigation. Until now, it is possible to upload conjectures and sketch explanations and/or hypothetical causality, although coherent and plausible they are not enough. They must be confirmed by solid empirical facts.


During the 1990s, in central portion of RMSP, the frequency, intensity and total amount of precipitation in the 300 days with larger pluviometric totals was, at least, 40% larger in the "workdays" compared to "not workdays". The same occurred on days with pluviometric totals equal or greater than 20 mm. In borders, this percentage is smaller - 20% approximately - considering that "workdays" are more frequent than "not workdays". There are evidences that the phenomenon is not limited to RMSP, extending to other areas that compose the urban complex in formation, pointed as Megalopolis Rio de Janeiro - São Paulo. On the other hand, there is evidence of lateral displacement of the phenomenon towards southwest. This displacement may be influenced by preferential occurrence of considered precipitation type in a specific arrangement of secondary circulation that favors the vertical development of cumulus. Other hypothesis is that this kind of precipitation is related to the continent-ocean breeze dynamics in this region.


AZEVEDO, T. R. de (2001a). Derivação antrópica do clima na Região Metropolitana de São Paulo abordada como função do ritmo semanal das atividades humanas. Tese de Doutorado. FFLCH, Universidade de São Paulo, São Paulo.

AZEVEDO, T. R. de (2001b). “O fluxo de calor gerado pelas atividades humanas no estudo climático da Região Metropolitana de São Paulo” in TARIFA, J. R. & AZEVEDO, T. R. de, orgs., Os climas da cidade de São Paulo: teoria e prática, Coleção Novos Caminhos, n. 4, Departamento de Geografia, FFLCH, Universidade de São Paulo, São Paulo.

AZEVEDO, T. R. de (2001c). “Dramas climáticos e responsabilidade social: a relação entre o ritmo das atividades econômicas - dias de trabalho e descanso - e o clima urbano” in Política & Trabalho, n.17. UFPB, João Pessoa.

AZEVEDO, T. R. de (2001d). “A atividade urbana faz com que os dias ‘úteis’ sejam mais chuvosos na Região Metropolitana de São Paulo” in anais do IX Simpósio Brasileiro de Geografia Física Aplicada. Universidade Federal do Recife, Recife.

AZEVEDO, T. R. de (2001e). “Distribuição espacial da chuva: um ensaio metodológico” in TARIFA, J. R. & AZEVEDO, T. R. de, orgs., Os climas da cidade de São Paulo: teoria e prática, Coleção Novos Caminhos, n. 4, Departamento de Geografia, FFLCH, Universidade de São Paulo, São Paulo.

AZEVEDO, T. R. de (2002). “Precipitação na Cidade de São Paulo em função da atividade urbana” in Anais do XII Congresso Brasileiro de Meteorologia. Foz do Iguaçu, Sociedade Brasileira de Meteorologia.

AZEVEDO, T. R. de & TARIFA, J. R. (2001). “O fluxo de calor gerado pelas atividades humanas no estudo climático da Região Metropolitana de São Paulo” in GEOUSP n.9. Departamento de Geografia, FFLCH, Universidade de São Paulo, São Paulo.

ESTADO DE SÃO PAULO (2000). Balanço Energético do Estado de São Paulo. Secretaria de Minas e Energia do Governo do estado de São Paulo.

FUNARI, F. L. (1983). Insolação, radiação solar global e radiação líquida no Brasil. Dissertação de Mestrado, Faculdade de Filosofia, Letras e Ciências Humanas, Universidade de São Paulo, São Paulo.

MONTEIRO, C. A. F. (1973). A dinâmica climática e as chuvas no Estado de São Paulo. Instituto de Geografia, Universidade de São Paulo, São Paulo.

MONTEIRO, C. A. F. (1976). “Teoria e Clima Urbano” in Teses e Monografias n.1. Instituto de Geografia, Universidade de São Paulo, São Paulo.

TOJO, J. F. (1998). La ciudad y el médio natural. Akal Editora, Madrid.


(Submitted to Editorial Commission of Geousp in April 2002)