Open-access A century and beyond: trends and implications of climate change from Agronomic Institute’s Historical Weather Dataset

ABSTRACT

The Agronomic Institute of Campinas, located in São Paulo, Brazil, is one of the country’s pioneering research centers. Since 1890, the institute has maintained daily records of air temperature and precipitation, resulting in one of the longest continuous meteorological series in South America. Based on the hypothesis that 135 years of uninterrupted meteorological data provide valuable insights into the climate evolution of this tropical region, this study aimed to detect and describe signs of climate change in the dataset and make it publicly accessible. To achieve these goals, we applied rigorous data quality assessment methods to ensure the reliability of the records, followed by parametric and nonparametric trend analyses on various climate change indices. The dataset was validated, with only a few errors identified and corrected. Trend analyses revealed a shift toward warmer conditions, while precipitation patterns showed no significant changes. However, rising air temperatures led to increased atmospheric water demand, resulting in a significant decrease in the difference between precipitation and potential evapotranspiration. This trend strongly suggests intensified regional drought conditions, linked to an extended dry season, which may elevate the risk of crop yield reductions due to heat and water stress. The complete historical dataset is now publicly available at https://rpubs.com/gabrielblain/1265900 and https://doi.org/10.5281/zenodo.14747078, along with the statistical methods and tools for downloading the records.

Key words
heat stress; drought; evapotranspiration; global warming

INTRODUCTION

The study of historical events in long-term meteorological records provides valuable insights into regional climate evolution, often revealing patterns not fully captured by climate models (Van Oldenborgh et al. 2022, Yule et al. 2023). These analyses also enable recent weather events to be understood in a historical context (MacDonald and Phillips 2006, Hawkins et al. 2022, Hegerl et al. 2019), making them a crucial step in identifying regional signals of global climate change (Hayhoe et al. 2007). Moreover, they enhance climate risk assessments (Gemmer et al. 2004) and deepen our understanding of the mechanisms driving regional weather and climate variability (MacDonald and Phillips 2006).

In addition to informing climate evolution studies, long-term datasets are indispensable for analyzing extreme weather events. These studies often focus on phenomena with long return periods (> 100 years), which typically exceed the length of available site-specific records (Goudenhoofdt et al. 2017, Blain et al. 2021). To address this limitation, parametric distributions such as the generalized extreme value, generalized pareto, and generalized logistic distributions are commonly employed (Khan et al. 2017). Guttman (1994) suggested that approximately 70 to 80 years of data are needed to ensure stable parameter estimates for the tails of precipitation frequency distributions, underscoring the importance of extended historical records.

Long-term meteorological series with at least 100 years of continuous records are relatively rare in many regions in the world. This scarcity is particularly pronounced in areas such as South America, where the availability and quality of historical records remain significant concerns. Founded in 1886 by Emperor Dom Pedro II, the Agronomic Institute of Campinas (IAC), located in the state of São Paulo, is one of Brazil’s pioneering research centers. Since 1890, the institution has maintained daily records of air temperature and precipitation, creating one of the longest meteorological series in South America (Blain et al. 2009, Blain et al. 2009). This weather station is situated at the institute’s experimental farm and has been routinely used in both academic and operational mode (Blain et al. 2017). Any missing daily record has been replaced by data extracted manually from pluviographs or from automatic weather stations situated at the same site. In 1956, due to the city’s expansion, a new weather station was installed approximately 3 km from the original site (from 22°53’S, 47°05’W, 663 m to 22°54’S, 47°05’W, 674 m). From 1956 to 1982, daily precipitation data were recorded at both locations. Air temperature data were also recorded at both sites, albeit for a shorter period (1956 to 1962). Mello et al. (1994) applied the Abbe criterion (Conrad and Pollak 1950) to these two datasets and confirmed that the simultaneously recorded precipitation and air temperature data from these two sites are homogeneous and can be regarded as a single continuous time series.

Based on the hypothesis that these 135 years of continuous meteorological records provide unique and valuable insights into the climate evolution of this tropical region, this study aimed to detect and describe signs of climate change in this historical data records and, to make it readily and public available. To achieve this, we first applied distinct data quality assessment methods to ensure the reliability of the meteorological series. We gave special attention to the above-mentioned change at the weather station’s site. We then performed parametric and nonparametric trend analyses on various climate changeindices derived from the historical data. Finally, we have made this historical dataset publicly available on https://rpubs.com/gabrielblain/1265900 (tag: Final Remarks and Data Download; Prela-Pantano et al. 2025a). Additionally, to enhance the reproducibility of our findings, we provided an R-Markdown file (in html format) containing all the statistical methods applied in this study at https://rpubs.com/gabrielblain/1265900 and https://doi.org/10.5281/zenodo.14747078 (Prela-Pantano et al. 2025b).

DATA

Campinas is a prominent city located in the state of São Paulo, Brazil (22°54’S, 47°03’W, Southeast region). Covering an area of approximately 795 km2, it is known for its strong agricultural and technological sectors and it is the third-largest city in the state. Campinas has a tropical climate (Aw, according to the Köppen classification), characterized by distinct wet and dry seasons (Alvares et al. 2014). The wet season occurs during the austral summer, with monthly precipitation totals typically exceeding potential evapotranspiration (Blain et al. 2017). December and January are the rainiest months, during which the precipitation frequency distributions follow a Gaussian (bell-shaped) pattern (Blain et al. 2007). Such distributions are common in equatorial climates (Ben-Gai et al. 1998). Conversely, the dry season takes place during the austral winter, predominantly in July and August, when monthly precipitation is generally lower than potential evapotranspiration. During this period, precipitation distributions assume an exponential shape (Blain et al. 2007), a pattern typically observed in semi-arid or arid regions (Ben-Gai et al. 1998). These variations highlight Campinas’ diverse climatic conditions.

As previously mentioned, the historical weather station of Campinas (1890–2024) is currently located at the IAC’s experimental farm and has been relocated only once, in 1956 (Mello et al. 1994). The initial location (22°53’S, 47°05’W, 663 m; 1890–1956) is approximately 3 km from its current site (22°54’S, 47°05’W, 669 m; 1957–2024). Based on the Abbe criterion, Mello et al. (1994) indicated that the meteorological records from these two sites–comprising precipitation and air temperature data–are homogeneous and can be regarded as a single continuous time series. Figure 1 depicts the two weather station’s locations.

Figure 1
The two locations of the Agronomic Institute weather station of Campinas, state of São Paulo, Brazil (yellow dots): IAC 1 (1890–1956) and IAC 2 (1957–2024). The red dotted line represents the distance between the two locations (~3 km).

As described in section “Data quality assessments”, this study also used air temperature data from the weather station Professor Jesus Marden dos Santos (Escola Superior de Agricultura Luiz de Queiroz–ESALQ, Universidade de São Paulo; 22°42’30”S, 47°38’30” W, 546 m, Piracicaba, SP, Brazil) for the year of 1931 (ESALQ 2024).

METHODS

Data quality assessments

Several studies (e.g., Reek et al. 1992, Dai et al. 1999, Aguilar et al. 2003, Morice et al. 2012) highlight the importance of visually inspecting historical meteorological records for gross errors before conducting any comprehensive analysis. A visual examination of the original data revealed two significant errors. The first occurred between December 14 and 16, 1931, and the second one between November 20 and 22, 1996. In both cases, the daily maximum temperature (Tmax) values were lower than the minimum temperature (Tmin) values. Notably, the second error (November 20–22, 1996) does not appear in newer publications. For instance, the Integrated Center for Agrometeorological Information (CIIAGRO), from the Secretariat of Agriculture and Supply of the State of São Paulo, corrected these erroneous records by replacing them with data from a mechanical thermograph located at the same site. Accordingly, we replaced these three original records with the corrected data published on the CIIAGRO website. The first gross error is detailed in Table 1.

Table 1
Gross errors detected in the maximum air temperature records (Tmax) of the Agronomic Institute’s weather station. The minimum air temperature (Tmin) is also presented. Campinas, state of São Paulo, Brazil.

The analysis of Table 1 indicated that the Tmax values are physically unreliable and must be replaced. In this study, such replacement was based on a linear regression analysis (Fig. 2) that considered data from the ESALQ’s weather station. Based on the resulting equation (Fig. 2), the Tmax values of Table 1 (days 14 to 16) were, respectively, replaced by 25.4, 28.5, and 30.7°C.

Figure 2
Linear regression analysis between maximum air temperature data (Tmax; 1st November to 10th December 1931) from the weather station Professor Jesus Marden dos Santos (Escola Superior de Agricultura Luiz de Queiroz, Universidade de São Paulo; 22°42’30”S, 47°38’30” W, 546 m, Piracicaba, SP, Brazil) and Agronomic Institute’s weather station. Campinas, state of São Paulo, Brazil (1890–2024).

The second step in the data quality assessment involved verifying whether the documented change at the weather station’s site in 1956 introduced artificial shifts in the historical series. Documented changes in observing locations are generally easier to assess than undocumented ones (Wang 2008). Depending on the frequency distribution of the data (e.g., Gaussian), the impact of such changes on the mean or variance of the series can be evaluated using parametric tests such as the Student’s t-test or F-test (Wang 2008). When specific distributional assumptions cannot be made, non-parametric methods are advantageous. These methods are slightly less efficient than tests that use the correct parametric model but are often more reliable than tests based on incorrect parametric assumptions (Storch and Zwiers 1999). The Pettitt test (Pettitt 1979) is a widely used non-parametric method for detecting a single change-point in a time series. It tests the null hypothesis of no change-point against the alternative hypothesis that a change-point exists. The test statistic is based on the rank sums of observations before and after the potential change-point. Among the available tests, the Pettitt test is one of the most frequently employed in agro-hydro-meteorological studies (e.g., Wijngaard et al. 2003, Liu et al. 2012, Mallakpour and Villarini 2016). The calculations for this test are outlined in Eqs. 1 to 4.

D i j = sgn X i − X j (1)

where: Xi: data value at time i. Xj: data value at time j.

The sign function, defined as:

sgn ( Δ x ) = + 1 , Δ x > 0 0 , Δ x = 0 − 1 , Δ x < 0 (2)
U t , T = Σ i = 1 t   Σ j = i + 1 T   D i j (3)

The test statistic K is the maximum absolute value of U(t):

K T = max U t , T 1 ≤ i < T (4)

The p-value is derived using Eq. 5:

p ≈ 2 exp   − 6 K 2 n 3 + n 2 (5)

As with any change-point test, the performance of the Pettitt test in detecting non-climatic or artificial changes in data records can be influenced by true climate variations. For example, it may be affected by transitions between two phases of long quasi-periodic variations (Wang 2008) or by trends resulting from ongoing climate change (refs). To account for these potential influences, we applied the widely used Mann-Kendall’s test (Kendall and Stuart 1967) prior to using the Pettitt test on the precipitation and air temperature series (aggregated at annual and monthly scales). The Mann-Kendall’s test assesses trends in a time series by comparing all possible pairs of data points to determine their relative ranking. The calculations for the Mann-Kendall’s test are outlined in Eqs. 6 to 8.

S = Σ i = 1 n − 1   Σ j = i + 1 n   sgn x i − x j w(6)

For sample sizes n ≥ 8, the frequency-distribution of S approaches the Gaussian shape with mean E(S) = 0 and variance given by Eq. 7.

Var ( S ) = n ( n − 1 ) ( 2 n + 5 ) − Σ j = − 1 J   t j ( m − 1 ) ⋅ ( 2 m + 5 ) 18 (7)

where: tj: the number of ties of length m.

The statistic S is then standardized (Eq. 8), resulting in the Mann-Kendall’s final value. The significance of the Mann-Kendall’s statistic can be assessed using the cumulative function of the standard normal distribution.

M K = S − 1 [ Var ( S ) ] 1 / 2      S > 0 S + 1 [ Var ( S ) ] 1 / 2      S < 0 (8)

Whenever a significant trend was detected, we employed locally estimated scatterplot smoothing (LOESS; Cleveland 1979, Cleveland and Devlin 1988) to remove (decompose) long-term trends from the series before applying the Pettitt test to the detrended data. LOESS is a non-parametric regression method that uses a series of local regressions to generate a trend line, making it particularly useful for visualizing trends in large datasets. It effectively describes non-linear relationships between the dependent variable (e.g., air temperature or precipitation) and the independent variable (time) while mitigating the influence of outliers on trend slope estimation. By isolating and removing the trend component, the detrended series allows for a clearer analysis of variability and abrupt changes, such as shifts or regime changes, without interference from long-term trends. This approach aligns with prior studies that emphasize the importance of detrending in climate and hydrological analyses to detect non-linear patterns and abrupt changes (e.g., Yue et al. 2002, Khaliq et al. 2009, Storch and Zwiers 1999). In essence, this ensures that the Pettitt test focuses solely on identifying shifts in the detrended data. All statistical analyses in this study were performed at the 5 and 10% significance levels.

Climate change indices

Initiatives such as those by the World Meteorological Organization Commission for Climatology/Climate Variability (WMO-CCL/CLIVAR) emphasize the use of practical indices to assess regional indicators of climate change (e.g., Karl et al. 1999, Peterson et al. 2001). Table 2 presents the climate change indices adopted in this study, which were estimated at both annual and monthly scales. The 10th, 90th, 95th, and 99th percentiles mentioned in Table 2 were calculated using the 1961–1990 standard reference period. Additional details about these indices are available at ETCCDI (2024).

Table 2
Climate change indices.

Previous studies investigating the impacts of climate change on agrometeorological parameters in the state of São Paulo (e.g., Pereira et al. 2018) reported a widespread increase in potential evapotranspiration (PE) rates. This finding is particularly significant as agricultural drought is a complex function of soil moisture and atmospheric water demand. In this study, we also calculated monthly values for PE and the difference between precipitation and PE (PPE) using Thornthwaite’s method. As noted by Pereira et al. (2018) and Camargo and Camargo (2000), Thornthwaite’s method is well-suited for estimating PE at a monthly scale under the climatic conditions of the state of São Paulo. Furthermore, the unavailability of key meteorological data at the study sites, such as wind speed and net radiation, prevents the use of more complex methods like FAO/Penman-Monteith (Allen et al. 1998). Additional details on the Thornthwaite’s method are provided by Camargo and Camargo (2000).

We applied the Mann-Kendall’s trend test to the time series of each index listed in Table 2, as well as to the PE and PPE series. For cases in which the Mann-Kendall’s test identified significant trends, we used Sen’s slope to calculate the average year-to-year rate of change. The Sen’s slope, also known as the Theil-Sen estimator, is robust to outliers and non-normal data distributions. It has been widely utilized in agro-hydro-meteorological studies, including those by Jiqin et al. (2023) and Achite et al. (2024).

Trend tests based on parametric modelling: nonstationary approach

Nonstationary parametric distribution can be used as trend tests because they detect changes in the probabilistic structure of a time series (Pujol et al. 2007, Sugahara et al. 2009, Delgado et al. 2010, Blain 2011c, Kharin et al. 2013, Xavier et al. 2019, Blain et al. 2022). The underlying principle of this approach is to determine whether nonstationary versions of a parametric distribution provide a better representation of data variability compared to their stationary counterparts. If no significant improvement is observed, the time series is considered to exhibit no change in its frequency distribution, indicating no detectable signs of climate change. Conversely, if a nonstationary version is found to be the best-fitting model, the time series is assumed to exhibit significant changes in its frequency distribution, and the hypothesis of no climate change is not accepted.

Naturally, any parametric trend assessment requires an appropriate theoretical distribution to accurately describe the data’s frequency distribution. In this context, the Kolmogorov-Smirnov/Lilliefors Goodness-of-Fit test (Lilliefors 1967) is widely used in agro-hydro-meteorological studies, in which the parameters of the probability functions and the goodness-of-fit test are derived from the same data series. Unlike the original Kolmogorov-Smirnov test, the critical values for the Kolmogorov-Smirnov/Lilliefors test must be obtained through statistical simulation techniques, such as Monte Carlo simulations. Additional information on this goodness-of-fit test can be found in Blain (2014).

Regarding candidate distributions for the Pre, Tmax, and Tmin series, a review of studies such as Blain (2011a), Stagge et al. (2015), and Blain et al. (2022) suggests that the two-parameter gamma and Weibull distributions (or normal and log-normal distributions) are commonly used to model the probabilistic structure of precipitation or air temperature series. Therefore, we applied the Kolmogorov-Smirnov/Lilliefors test to assess whether these distributions could be fitted to the air temperature and precipitation series in this study (at both annual and monthly scales). In cases in which the Kolmogorov-Smirnov/Lilliefors test identified two or more suitable distributions for a particular series, we used the second-order Akaike information criterion (AICc) to rank their performance. The AICc is calculated as follows (Eqs. 9 and 10):

AIC = − 2 log   -likelihood  + 4.4 m (9)
AICc = AIC + 2 m ( m + 1 ) n − m − 1 (10)

where: AIC: the original Akaike information criterion; AICc: the second-order Akaike information criterion; m: the number of model’s parameters; n: the sample size.

After selecting the best-fitting distributions for each series, we applied the parametric trend test based on six nonstationary models that can describe a wide range of changes in air temperature and precipitation series. These models were developed within the framework of generalized linear additive models, combined with generalized additive models for distribution parameters (GAMLSS). The GAMLSS method is well-known for its flexibility in modeling the frequency distribution of response variables, as it can incorporate both linear and nonlinear functions (Hastie and Tibshirani 1986, McCullagh and Nelder 1989, Rigby and Stasinopoulos 2005). The GAMLSS models considered in this study are described in Table 3.

Table 3
Non-stationary candidate models used in the parametric trend test of this study.

The AICc was once again used to select the best-fitting GAMLSS model for each series. As described earlier, whenever the AICc selected model 1 (the stationary model), we accepted the hypothesis that the frequency distribution of the data had not changed over time. In other words, we considered the series to be free from signs of climate change. Conversely, the selection of a nonstationary model (models 2 to 6) was interpreted as evidence that the frequency distribution of the data had significantly changed over time, indicating the presence of significant signs of climate change. As outlined by Blain et al. (2022), this parametric trend analysis not only detects changes in the data’s frequency distribution but also isolates its effects on the central tendency and/or dispersion of the meteorological series. This approach provides a more comprehensive understanding of the effects of current climate change on the frequency and intensity of weather events.

RESULTS AND DISCUSSION

Data quality assessments and long-term trends: visual analysis and statistical tests

Figure 3 shows the daily values of Pre, Tmin, and Tmax, considering the correction described in section “Data”. The visual analysis of Fig. 3 reveals no outliers or additional gross errors in either the air temperature or precipitation records. As mentioned earlier, data from the IAC weather station has been extensively used for both academic (Blain et al. 2017, Litell et al. 2022, Santos Jr. et al. 2022, Pereira et al. 2023, Blain et al. 2024) and operational purposes. For example, since 1992, these data have been routinely employed in the agrometeorological governmental monitoring program of the Secretariat of Agriculture and Supply of the State of São Paulo through the CIIAGRO. Additionally, these records are published daily at the Agronomic Institute’s homepage (https://clima.iac.sp.gov.br/). However, for the air temperature series, the visual analysis suggested the presence of remarkable monotonic increasing trends. This feature is consistent with previous studies (e.g., Blain 2011b) and was further confirmed by the statistical analysis described below.

Figure 3
135 years of daily records from the Agronomic Institute’s weather station. Campinas, state of São Paulo, Brazil (1890–2024).

Table 4 presents the results of the Mann-Kendall’s and Pettitt change-point tests calculated at both annual and monthly scales. For precipitation amounts, the tests indicated no significant trends (Mann-Kendall’s test) or abrupt change points (Pettitt test) for 11 months. Only the monthly values for February showed a decreasing trend, which was significant at the 10% level, but not at the 5% level. This lack of significant trends in the precipitation series has been also observed in previous studies, such as Blain et al. (2007) and Blain (2011a). However, the Mann-Kendall’s test revealed highly significant upward trends (p < 0.001) in both minimum and maximum air temperature data over the past 135 years. This result is consistent with the widespread increase in air temperature values reported across Brazil in recent decades, as documented by several studies (Curado et al. 2023, Regoto et al., 2021). These trends align with regional climate change patterns and highlight the importance of long-term datasets in understanding the localized impacts of global warming.

Table 4
Mann-Kendall’s trend test (MK and its corresponding p-values) and Pettitt change-point test (p-values) applied to precipitation (Rain), minimum (Tmin) and maximum (Tmax) air temperature series. Campinas, state of São Paulo, Brazil (1890–2024).

As described in the “Methods” section, we applied LOESS smoothing to all air temperature series before performing the Pettitt test. The detrended series revealed no abrupt change points. In summary, the data quality assessments conducted in this study indicated no additional gross or systematic errors in the historical meteorological records of the IAC. This finding is consistent with the widespread use of this weather dataset in previous studies (Abreu et al. 2022, Blain et al. 2017, Litell et al. 2022, Santos Jr. et al. 2022, Pereira et al. 2023, Blain et al. 2024) and further supports its value in providing unique and valuable insights into the climate evolution of this tropical region. Additionally, it aligns with Mello et al. (1994), who stated that the shift in the weather station’s location (described in the introduction) did not significantly alter the frequency distribution of the meteorological records.

Non-parametric trend tests applied to the annual climate change indices

The lack of significant climate change signals in the precipitation series was also observed in the climate change indices (Table 5). This result holds true even for indices addressing extreme precipitation events, such as R95 (number of days with rainfall > 95th percentile) and R99 (number of days with rainfall > 99th percentile). At the local scale, this finding aligns with Blain (2011c), which demonstrated that the stationary version of the generalized extreme value distribution, with time-constant parameters, can be used to assess the probability of daily extreme precipitation events in Campinas. At the national scale, it is consistent with Regoto et al. (2021), who reported no strong or clear signs of change in precipitation across Southeast Brazil.

Table 5
Mann-Kendall’s trend test (MK and its corresponding p-value) and Theil-Sen slope estimator (Slope) applied to climate change indices. Campinas, state of São Paulo, Brazil (1890–2024).

Unlike the precipitation series, all indices related to air temperature showed highly significant changes (< 0.001) towards warmer conditions (Table 4). The number of nights with Tmin > 90th percentile (warm nights) and the number of days with Tmax > 90th percentile (warm days) increased at average rates of 0.48 and 0.34 days per year, respectively. The number of days with Tmax > 25°C (SU) and the number of days with Tmin > 20°C (TR) also increased at similar rates of 0.46 and 0.40, respectively. Conversely, the number of nights with Tmin < 10th percentile (cool nights) and the number of days with Tmax < 10th percentile (cool days) decreased at average rates of -0.46 and -0.20 days per year. This significant shift towards warmer conditions is consistent with Blain (2011b), who found significant upward trends in air temperature series in the state of São Paulo; with Regoto et al. (2021), who reported clear signs of warming in air temperature across Southeast Brazil; and with Dufek et al. (2008), who found widespread upward trends in minimum air temperature data across the state of São Paulo. These warming trends also align with global and regional trends reported by the Intergovernmental Panel on Climate Change (IPCC) in their Sixth Assessment Report (AR6), which indicates an increase in surface air temperatures over the past century, particularly in tropical and subtropical regions (IPCC 2021).

Non-parametric trend tests applied to the monthly climate change indices

The absence of significant climate change signals in the precipitation series and the presence of significant upward trends in air temperature were also observed at the monthly scale (Table 6). As previously mentioned, a significant trend was detected in the February series, in which the trend tests indicated a shift towards drier conditions, with an average decrease of -0.360 mm per year over the last 135 years. This result, as noted earlier, aligns with the findings of Regoto et al. (2021). Furthermore, the results in Table 5 show that the slopes (year-to-year rates of change) of indices associated with Tmin tend to be steeper than those for Tmax, which is consistent with the findings of Blain (2011b). The latter study highlighted that signs of climate change in the state of São Paulo are more evident in the Tmin series than in the Tmax series. Regarding the TXx index (the highest daily maximum temperature for each month), it is noteworthy that the largest slopes are observed in September (0.024°C·year-1) and October (0.023°C·year-1). These months have seen some of the most intense heatwaves in recent years in the state of São Paulo, as observed by Valverde and Rosa (2023).

Table 6
Mann-Kendall’s trend test (corresponding p-value) and Theil-Sen slope estimator (Slope) applied to climate change indices. Campinas, state of São Paulo, Brazil (1890–2024). The slopes for rain, potential evapotranspiration (PE) and the difference between rain and PE (PPE) and given in mm·year-1; the other indices are given in °C year-1.

Tmin: minimum air temperature; Tmax: maximum air temperature; Tavg: average air temperature; TXx: largest daily maximum temperature; TNx: largest daily minimum temperature; TXn: lowest daily maximum temperature; TNn: lowest daily minimum temperature; NC: the Mann-Kendall’s test indicated no significant trend.

From an agrometeorological perspective, the results in Table 5 reveal significant increasing trends in all monthly PE series. This suggests that atmospheric water demand (i.e., PE rates) has risen over the last 135 years in Campinas, with rates ranging from 0.092 mm per year in May to 0.263 mm per year in October. Consequently, the PPE series showed significant decreasing trends in seven months: February, June, August, September, October, November, and December. These results are consistent with those of Pereira et al. (2018), who reported that increased PE rates have intensified regional drought conditions in the state of São Paulo. Moreover, this finding further supports the transition toward drier conditions that Campinas has experienced in recent years. Notably, the significant decreases in PPE amounts from August to December align with previous studies that highlighted a recurring delay in the onset of the rainy season in São Paulo (e.g., Blain 2011a, Blain and Bardin-Camparotto 2014, Pereira et al. 2018). In São Paulo, critical plant growth stages, such as flowering, regrowth, and sprouting (e.g., for sugarcane and citrus), often occur during these months. Therefore, the observed significant decreases in PPE rates at the onset of the regional rainy season likely increase the risk of crop yield reductions due to agricultural drought. From a statistical perspective, it is noteworthy that the highest rate of change in the PPE series (in absolute terms) occurred in February. During this month, the precipitation series showed a significant decrease at a rate of 0.064 mm per year, while the PE series exhibited a significant increase at a rate of 0.191 mm per year.

Parametric trend tests

Compared to rank-based non-parametric methods (e.g., the Mann-Kendall’s test), the parametric trend test described in the “Methods” section provides a more comprehensive analysis of trends by evaluating their effects on both the central tendency (μ parameter) and the dispersion (σ parameter) of frequency distributions (Blain 2011a, Blain et al. 2021, Pereira et al. 2023). In other words, by selecting one of the six non-stationary models presented in Table 3, it becomes possible to determine whether and how the mean and/or dispersion of the frequency distributions for Pre, Tmax, Tmin have changed over time. As previously mentioned, these tests require a suitable theoretical distribution to accurately describe the data’s frequency distribution. Table 7 presents the parametric distributions selected using the Kolmogorov-Smirnov/Lilliefors goodness-of-fit test and the AICc criterion.

Table 7
Best fitting distribution selected by the Kolmogorov-Smirnov/Lilliefors goodness-of-fit test and second-order Akaike information criteria for precipitation (rain), minimum (Tmin) and maximum (Tmax) air temperature series. Campinas, state of São Paulo, Brazil (1890–2024).

The results presented in Table 7 indicate that the two most flexible distributions, Weibull and Gamma, dominated the precipitation series, while the normal and log-normal distributions were most suitable for the air temperature series. These findings for the precipitation series are consistent with the typically asymmetric nature of precipitation distributions in the state of São Paulo, which are often right-skewed, as observed in previous studies (e.g., Blain et al. 2007). Similarly, the results for both Tmin and Tmax series align with the findings of Blain et al. (2009), which suggested that annual and monthly air temperature series in the state of São Paulo typically exhibit a bell-shaped distribution. After selecting the best fitting distributions for each series, we applied the parametric trend test to all Pre, Tmin, and Tmax series (Table 8).

Table 8
Best nonstationary model for precipitation (rain), minimum (Tmin) and maximum (Tmax) air temperature series. Campinas, state of São Paulo, Brazil (1890–2024).

As expected, the results of the parametric trend test align with those of the nonparametric test, indicating no significant changes in the probabilistic structure of the Pre series. These findings hold for both annual and monthly scales and are consistent with previous studies such as Blain (2011b) and Regoto et al. (2021). Regarding the air temperature series, model 2 was the best fit for both the annual Tmin and Tmax series, as well as most of the monthly Tmin and Tmax series. This model describes a linear change in the mean parameter of the distributions fitted to the Tmin and Tmax series, while their dispersion remained constant over time. Figure 4 illustrates this change for the annual scale.

Figure 4
Year-to-year change in the mean parameter of annual air temperature frequency-distributions. Campinas, state of São Paulo, Brazil (1890–2024).

An alternative interpretation of Fig. 4 is that, in 1890, Campinas had a 50% probability of experiencing average annual Tmin (or Tmax) values equal to or lower than 14.3°C (or 26.1°C). By 2024, this probability dropped to nearly zero: 0.0000003% for Tmin and 0.002% for Tmax. This result strongly indicates a significant shift in the climate patterns of Campinas.

For the monthly Tmin series, model 2 was the best fit for all months, except for February, March, and December, in which model 4, that accounts for linear changes in both the mean and dispersion, provided the best fit. Figures 5 and 6 present, respectively, the year-to-year changes in the mean and dispersion parameters of the Tmin monthly frequency distributions.

Figure 5
Year-to-year change in the mean parameter of monthly minimum air temperature frequency distributions. Campinas, state of São Paulo, Brazil (1890–2024).

As presented in Fig. 5, the mean parameter of all monthly Tmin series exhibited a behavior similar to that of Fig. 4: a linear increase towards warmer conditions. This result aligns with previous studies (e.g., Blain 2011b), which also evaluated the frequency distribution of the Tmin series in the state of São Paulo.

As observed in Fig. 6, the only three-monthly series in which the dispersion of the Tmin frequency distribution changed over time were February, March, and December. To the authors’ best knowledge, this is the first time such change in the dispersion of these series was detected, which further confirmed the ability of parametric trend tests to provide comprehensive analysis of trends by assessing their effects on both the central tendency and the dispersion of frequency distributions. In this context, Fig. 6 indicated that the dispersion of the Tmin data of February and March increased over time. In other words, there were an increase in the variability of the Tmin values around their means. An opposite behavior was found in December, when the variability of the Tmin values around their means decreased from 1890 to 2024.

Figure 6
Year-to-year change in the dispersion parameter of monthly minimum air temperature frequency distributions. Campinas, state of São Paulo, Brazil (1890–2024).

With regards to the monthly Tmax, model 2 was again the one that best fitted all monthly series, apart from July and December, when model 5, which describes a quadratic change only in the mean parameter, was the best fitting model. In other words, as for the Tmin series, all Tmax frequency distributions presented increasing trends in their mean parameter. However, there was no change in their dispersion (see Table 8). As observed in Fig. 7, all Tmax series showed changes towards warmer conditions.

The warming trends and increased atmospheric water demand observed in this study have direct implications for the agronomic sector in the state of São Paulo, particularly for key crops such as sugarcane and citrus. During critical phenological stages like flowering, regrowth, and sprouting–commonly occurring from October to December–, high air temperatures and intensified drought conditions can significantly reduce crop yields by increasing heat and water stress (Blain 2011b, Pereira et al. 2018). For sugarcane, prolonged dry periods can reduce biomass accumulation and sugar content, while citrus trees may experience decreased fruit set and quality (Blain et al. 2022). This change toward warmer and drier conditions may also put negative pressure on crops such as vegetables (e.g., tomatoes and brassicas) and fruits (e.g., grapes and bananas). These findings emphasize the need for adaptive agricultural management practices, such as improved irrigation strategies and heat-tolerant crop varieties, to mitigate the risks posed by ongoing climate change.

Figure 7
Year-to-year change in the mean parameter of monthly maximum air temperature frequency distributions. Campinas, state of São Paulo, Brazil (1890–2024).

CONCLUSION

Based on the hypothesis that long-term continuous meteorological records provide unique and valuable insights into climate evolution, this study made publicly available one of the longest meteorological series in South America: the IAC’s historical meteorological dataset. This dataset comprises continuous daily records of precipitation and maximum and minimum air temperatures dating back to 1890. The data are accessible at https://rpubs.com/gabrielblain/1265900 and https://doi.org/10.5281/zenodo.14747078.

Before publication, rigorous data-quality tests were applied, confirming the dataset’s high reliability. These tests also verified that the 1956 weather station relocation, as detailed in the introduction, did not significantly affect the frequency distribution of the meteorological records. A few detected errors (14–16 December 1931 and 20–22 November 1996) were corrected as described in “Data” section.

Trend analyses revealed significant indicators of climate change marked by a pronounced shift toward warmer conditions in both minimum and maximum air temperature records. While no clear sign of change was detected in the probabilistic structure of annual and monthly precipitation series, the increased air temperatures resulted in significant upward trends in PE rates (a measure of atmospheric water demand). Consequently, a notable decrease in the PPE was observed, strongly suggesting intensified regional drought conditions. This drying trend was most evident from February through December, except for March, April, May, and July.

In the state of São Paulo, critical phenological stages of crop growth, such as flowering, regrowth, and sprouting (e.g., for sugarcane and citrus), often occur in October, November, and December. Therefore, the observed trends toward drier conditions and higher temperatures likely heighten the risk of crop yield reductions due to increased heat and water stress.

ACKNOWLEDGMENTS

To Conselho Nacional de Desenvolvimento Científico e Tecnológico for Fellowship for the sixth author (Process 304609/2022-6).

This study was financed in part by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - Brasil (Finance Code 001).

How to cite: Prela-Pantano, A., Brunini, O., Sobierajski, G. R., Bardin-Camparotto, L., Martins, L. L. and Blain, G. C. (2025). A century and beyond: trends and implications of climate change from Agronomic Institute’s Historical Weather Dataset. Bragantia, 84, e20250019. https://doi.org/10.1590/1678-4499.20250019

FUNDING

Conselho Nacional de Desenvolvimento Científico e Tecnológico
Grant No.: 304609/2022-6
Coordenação de Aperfeiçoamento de Pessoal de Nível Superior
Finance code 001

DATA AVAILABILITY STATEMENT

The data used in this study is available at https://rpubs.com/gabrielblain/1265900 and https://doi.org/10.5281/zenodo.14747078

REFERENCES

  • Abreu, R. C., Hallak, R. and Rocha, H. R. (2022). Effects of local vegetation and regional controls in near-surface air temperature for Southeastern Brazil. Atmosphere, 13, 1758. https://doi.org/10.3390/atmos13111758
    » https://doi.org/10.3390/atmos13111758
  • Achite, M., Cheikh, M. and Deriouche, R. (2024). Hydrology and water resources management in semiarid regions: A case study of the Wadi Cheliff basin. Hydrology, 11, 77. https://doi.org/10.3390/hydrology11060077
    » https://doi.org/10.3390/hydrology11060077
  • Aguilar, E., Auer, I., Brunet, M., Peterson, T. C. and Wieringa, J. (2003). Guidelines on Climate Metadata and Homogenization. World Meteorological Organization (WMO).
  • Allen, R. G., Pereira, L. S., Raes, D. and Smith, M. (1998). Crop Evapotranspiration: Guidelines for Computing Crop Water Requirements. FAO Irrigation and Drainage Paper 56. Rome: Food and Agriculture Organization of the United Nations.
  • Alvares, C. A., Stape, J. L., Sentelhas, P. C., Gonçalves, J. L. M. and Sparovek, G. (2014). Köppen’s climate classification map for Brazil. Meteorologische Zeitschrift, 22, 711-728. https://doi.org/10.1127/0941-2948/2013/0507
    » https://doi.org/10.1127/0941-2948/2013/0507
  • Ben-Gai, T., Bitan, A., Manes, A., Alpert, P. and Rubin, S. (1998). Temporal and spatial trends of temperature patterns in Israel. Theoretical and Applied Climatology, 61, 177-185. https://doi.org/10.1007/s007040050062
    » https://doi.org/10.1007/s007040050062
  • Blain, G. C. (2011a). Cento e vinte anos de totais extremos de precipitação pluvial máxima diária em Campinas, Estado de São Paulo: análises estatísticas, Bragantia, 70, 722-728. https://doi.org/10.1590/S0006-87052011000300031
    » https://doi.org/10.1590/S0006-87052011000300031
  • Blain, G. C. (2011b). Considerações estatísticas relativas a seis séries mensais de temperatura do ar da Secretaria de Agricultura e Abastecimento do Estado de São Paulo. Revista Brasileira de Meteorologia, 26, 279-296. https://doi.org/10.1590/S0102-77862011000200012
    » https://doi.org/10.1590/S0102-77862011000200012
  • Blain, G. C. (2011c). Incorporating climate trends in the stochastic modeling of extreme minimum air temperature series of Campinas, state of São Paulo, Brazil. Bragantia, 70, 952-957. https://doi.org/10.1590/S0006-87052011000400031
    » https://doi.org/10.1590/S0006-87052011000400031
  • Blain, G. C. (2014). Revisiting the critical values of the Lilliefors teste: towards the correct agrometeorological use of the Kolmogorov-Smirnov framework. Bragantia, 73, 192-202. https://doi.org/10.1590/brag.2014.015
    » https://doi.org/10.1590/brag.2014.015
  • Blain, G. C. and Bardin-Camparotto, L. (2014). Detecting trends in 10-day rainfall amounts at five sites in the State of São Paulo, Brazil. Acta Scientiarum Technology, 36, 685-692. https://doi.org/10.4025/actascitechnol.v36i4.20513
    » https://doi.org/10.4025/actascitechnol.v36i4.20513
  • Blain, G. C., Avila, A. M. and Pereira, V. R. (2017). Using the normality assumption to calculate probability-based standardized drought indices: selection criteria with emphases on typical events. International Journal of Climatology, 38, e418-e436. https://doi.org/10.1002/joc.5381
    » https://doi.org/10.1002/joc.5381
  • Blain, G. C., Camargo, M. B. P. and Sentelhas, P. C. (2007). Variabilidade temporal da temperatura do ar na região de Campinas, SP: Detecção de tendências de aquecimento global. Bragantia, 66, 189-196. https://doi.org/10.1590/S0006-87052007000200019
    » https://doi.org/10.1590/S0006-87052007000200019
  • Blain, G. C., Kayano, M. T., Sentelhas, P. C. and Lulu, J. (2009). Possible influences of Pacific Decadal Oscillation on the ten-day-based ratio between actual and potential evapotranspiration in the region of Campinas, São Paulo State, Brazil. Bragantia, 68, 773-782. https://doi.org/10.1590/S0006-87052009000300029
    » https://doi.org/10.1590/S0006-87052009000300029
  • Blain, G. C., Sobierajski, G. R., Pires, R. C. M., Silveira, J. M. C., Martins, L. L. and Sparks, A. H. (2024). An R-package to track soil water deficits in the root zone: when and how much to irrigate. Bragantia, 83, e2024013. https://doi.org/10.1590/1678-4499.20240138
    » https://doi.org/10.1590/1678-4499.20240138
  • Blain, G. C., Sobierajski, G. R., Weight, E., Martins, L. L. and Xavier, A. C. F. (2022). Improving the interpretation of standardized precipitation index estimates to capture drought characteristics in changing climate conditions. International Journal of Climatology, 42, 5586-5608. https://doi.org/10.1002/joc.7550
    » https://doi.org/10.1002/joc.7550
  • Blain, G. C., Sobierajski, G. R., Xavier, A. C. F. and Carvalho, J. P. (2021). Regional frequency analysis applied to extreme rainfall events: Evaluating its conceptual assumptions and constructing null distributions. Anais da Academia Brasileira de Ciências, 93, e2019406. https://doi.org/10.1590/0001-3765202120190406
    » https://doi.org/10.1590/0001-3765202120190406
  • Camargo, A. P. and Camargo, M. B. P. (2000). Uma revisão analítica da evapotranspiração potencial. Bragantia, 59, 125-137. https://doi.org/10.1590/S0006-87052000000200002
    » https://doi.org/10.1590/S0006-87052000000200002
  • Cleveland, W. S. (1979). Robust locally weighted regression and smoothing scatterplots. Journal of the American Statistical Association, 74, 829-836. https://doi.org/10.1080/01621459.1979.10481038
    » https://doi.org/10.1080/01621459.1979.10481038
  • Cleveland, W. S. and Devlin, S. J. (1988). Locally weighted regression: An approach to regression analysis by local fitting. Journal of the American Statistical Association, 83, 596-610. https://doi.org/10.1080/01621459.1988.10478639
    » https://doi.org/10.1080/01621459.1988.10478639
  • Conrad, V. and Pollak, C. (1950). Methods in Climatology (2nd ed.). London: Harvard University Press.
  • Curado, L. F. A., Paulo, S. R., Paulo, I. J. C., Maionchi, D. O., Silva, H. J. A., Costa, R. O., Silva, I. M. C. B., Marques, J. B., Lima, A. M. S. and Rodrigues, T. R. (2023). Trends and patterns of daily maximum, minimum and mean temperature in Brazil from 2000 to 2020. Climate, 11, 168. https://doi.org/10.3390/cli11080168
    » https://doi.org/10.3390/cli11080168
  • Dai, A., Trenberth, K. E. and Karl, T. R. (1999). Effects of clouds, soil moisture, precipitation, and water vapor on diurnal temperature range. Journal of Climate, 12, 2451-2473. https://doi.org/10.1175/1520-0442(1999)012<2451:EOCSMP>2.0.CO;2
    » https://doi.org/10.1175/1520-0442(1999)012<2451:EOCSMP>2.0.CO;2
  • Delgado, J. M., Apel, H. and Merz, B. (2010). Flood trends and variability in the Mekong river. Hydrology and Earth System Sciences, 14, 407-418. https://doi.org/10.5194/hess-14-407-2010
    » https://doi.org/10.5194/hess-14-407-2010
  • Dufek, A. S., Ambrizzi, T. and Rocha, R. P. (2008). Are Reanalysis data useful to calculate Climate Indices over South America? Annals of the New York Academy of Sciences, 1146, 87-104. https://doi.org/10.1196/annals.1446.010
    » https://doi.org/10.1196/annals.1446.010
  • Escola Superior de Agricultura “Luiz de Queiroz” (ESALQ) (2024). Available at: http://www.leb.esalq.usp.br/leb/exceldados/ROB31.TXT Accessed on: Jan 1, 2025.
    » http://www.leb.esalq.usp.br/leb/exceldados/ROB31.TXT
  • Expert Team on Climate Change Detection and Indices (ETCCDI) (2024). ETCCDI Climate Change Indices. ETCCDI. Avaliable at: https://etccdi.pacificclimate.org/list_27_indices.shtml Accessed on: Jan 1, 2025
    » https://etccdi.pacificclimate.org/list_27_indices.shtml
  • Gemmer, M., Becker, S., and Jiang, T. (2004). Observed monthly precipitation trends in China 1951-2002. Theoretical and Applied Climatology, 77, 39-45. https://doi.org/10.1007/s00704-003-0018-3
    » https://doi.org/10.1007/s00704-003-0018-3
  • Google Earth (2024). Instituto Agronômico de Campinas. Available at: http://earth.google.com/ Accessed on: Jan 1, 2025
    » http://earth.google.com/
  • Goudenhoofdt, E., Delobbe, L. and Willems, P. (2017). Regional frequency analysis of extreme rainfall in Belgium based on radar estimates. Hydrological and Earth System Sciences, 21, 5385-5399. https://doi.org/10.5194/hess-21-5385-2017
    » https://doi.org/10.5194/hess-21-5385-2017
  • Guttman, N. B. (1994). On the sensitivity of sample L moments to sample size. Journal of Climatology, 7, 1026-1029. Available at: https://www.jstor.org/stable/26198664 Accessed on: Jan. 1, 2025.
    » https://www.jstor.org/stable/26198664
  • Hastie, T. and Tibshirani, R. (1986). Generalized Additive Models. Statistical Science, 1, 297-310. https://doi.org/10.1214/ss/1177013604
    » https://doi.org/10.1214/ss/1177013604
  • Hawkins, E., Brohan, P., Burgess, S., Burt, S., Compo, G., Gray, S., Haigh, I., Hersbach, H., Kuijjer, K., Martinez-Alvarado, O., McColl, C., Schurer, A., Slivinski, L. and Williams, J. (2022). Rescuing historical weather observations improves quantification of severe windstorm risks. EGUsphere. https://doi.org/10.5194/egusphere-2022-1045
    » https://doi.org/10.5194/egusphere-2022-1045
  • Hayhoe, K., Wake, C. P., Huntington, T. G., Luo, L., Schwartz, M. D., Sheffield, J., Wood, E., Anderson, B., Bradbury, J., DeGaetano, A., Troy, T. J. and Wolfe, D. (2007). Past and future changes in climate and hydrological indicators in the U.S. Northeast. Climate Dynamics, 28, 381-407. https://doi.org/10.1007/s00382-006-0187-8
    » https://doi.org/10.1007/s00382-006-0187-8
  • Hegerl, G. C., Brönnimann, S., Cowan, T., Friedman, A. R., Hawkins, E., Iles, C., Müller, W., Schurer A. and Undorf, S. (2019). Causes of climate change over the historical record. Environmental Research Letters, 14, 123006. https://doi.org/10.1088/1748-9326/ab4557
    » https://doi.org/10.1088/1748-9326/ab4557
  • [IPCC] Intergovernmental Panel on Climate Change (2021). Climate Change 2021: The Physical Science Basis. Contribution of Working Group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press. Available at: https://www.ipcc.ch/report/ar6/wg1/ Accessed on: Jan. 1, 2025.
    » https://www.ipcc.ch/report/ar6/wg1/
  • Jiqin, X., Zhang, T. and Li, H. (2023). Climate change impacts on hydrological cycles in arid regions: A review. Water and Climate Change, 14, 977-988. https://doi.org/10.2166/wcc.2023.508
    » https://doi.org/10.2166/wcc.2023.508
  • Karl, T. R., Nicholls, N. and Ghazi, A. (1999). CLIVAR/GCOS/WMO workshop on indices and indicators for climate extremes: Workshop summary. In T. R. Karl, N. Nicholls and A. Ghazi (Eds.) Weather and Climate Extremes (p. 3-7). Dordrecht: Springer.
  • Kendall, M. G. and Stuart, A. (1967). The advanced theory of statistics. London: Charles Griffin.
  • Khaliq, M. N., Ouarda, T. B. M. J., Gachon, P. and Bobée, B. (2009). Identification of hydrological trends in the presence of serial and cross-correlation: A review of selected methods and their application to annual flow regimes of Canadian rivers. Journal of Hydrology, 368, 117-130. https://doi.org/10.1016/j.jhydrol.2009.01.035
    » https://doi.org/10.1016/j.jhydrol.2009.01.035
  • Khan, S. A, Hussain, I., Hussain T., Faisal, M., Muhammad, Y. S. and Shoukry, A. M. (2017). Regional frequency analysis of extremes precipitation using L-moments and partial L-moments. Advances in Meteorology, 2017, 1-20. https://doi.org/10.1155/2017/6954902
    » https://doi.org/10.1155/2017/6954902
  • Kharin, V. V., Zwiers, F. W., Zhang, X. and Wehner, M. (2013). Changes in temperature and rainfall extremes in the CMIP5 ensemble. Climatic Change, 119, 345-357. https://doi.org/10.1007/s10584-013-0705-8
    » https://doi.org/10.1007/s10584-013-0705-8
  • Lilliefors, H. W. (1967). On the Kolmogorov-Smirnov test for normality with mean and variance unknown. Journal of the American Statistical Association, 62, 399-402. https://doi.org/10.1080/01621459.1967.10482916
    » https://doi.org/10.1080/01621459.1967.10482916
  • Litell, M. F., Xavier, A. C. F., and Blain, G. C. (2022). Evaluation of nested climate change models in the study of extreme events. Journal of Earth System Sciences, 131, 120. https://doi.org/10.1007/s12040-022-01853-w
    » https://doi.org/10.1007/s12040-022-01853-w
  • Liu, L., Xu, Z.-X. and Huang, J.-X. (2012). Spatio-temporal variation and abrupt changes for major climate variables in the Taihu Basin, China. Stochastic Environmental Research and Risk Assessment, 26, 777-791. https://doi.org/10.1007/s00477-011-0547-8
    » https://doi.org/10.1007/s00477-011-0547-8
  • MacDonald, N. and Phillips, I. D. (2006). Reconstructed annual precipitation series for Scotland (1861–1991): Spatial and temporal variations, and links to the atmospheric circulation. Scottish Geographical Journal, 122, 1-18. https://doi.org/10.1080/00369220600830771
    » https://doi.org/10.1080/00369220600830771
  • Mallakpour, I. and Villarini, G. (2016). Investigating the relationship between the frequency of flooding over the central United States and large-scale climate. Hydrological Sciences Journal, 61, 245-254. https://doi.org/10.1080/02626667.2015.1008482
    » https://doi.org/10.1080/02626667.2015.1008482
  • McCullagh, P. and Nelder, J. A. (1989). Generalized linear models. London and New York: Chapman and Hall.
  • Mello, M. H. A., Pedro Júnior, M. J., Ortolani, A. A. and Alfonsi, R. R. (1994). Chuva e temperatura: cem anos de observações em Campinas. Campinas: Instituto Agronômico.
  • Morice, C. P., Kennedy, J. J., Rayner, N. A. and Jones, P. D. (2012). Quantifying uncertainties in global and regional temperature change using an ensemble of observational estimates: The HadCRUT4 Dataset. Journal of Geophysical Research: Atmospheres, 117. https://doi.org/10.1029/2011JD017187
    » https://doi.org/10.1029/2011JD017187
  • Pereira, L. B., Martins, L. L., Rodrigues, I. C. A., Sobierajski, G. R. and Blain, G. C. (2023). Changes in extreme air temperatures in one of the South America’s longest meteorological records: Campinas, Brazil (1890-2022). Bragantia, 82, e20230128. https://doi.org/10.1590/1678-4499.20230128
    » https://doi.org/10.1590/1678-4499.20230128
  • Pereira, V. R., Blain, G. C., Avila, A. M. H., Pires, R. C. M. and Pinto, H. S. (2018). Impacts of climate change on drought: changes to drier conditions at the beginning of the crop growing season in southern Brazil. Bragantia, 77, 201-211. https://doi.org/10.1590/1678-4499.2017007
    » https://doi.org/10.1590/1678-4499.2017007
  • Peterson, T. C., Folland, C., Gruza, G., Hogg, W., Mokssit, A. and Plummer, N. (2001). Report on the activities of the Working Group on Climate Change Detection and Related Rapporteurs, 1998–2001. World Meteorological Organization, 1-143.
  • Pettitt, A. N. (1979). A non-parametric approach to the change-point problem. Applied Statistics, 28, 126-135. https://doi.org/10.2307/2346729
    » https://doi.org/10.2307/2346729
  • Prela-Pantano, A., Brunini, O., Sobierajski, G. R., Bardin-Camparotto, L., Lopes, L. M. and Blain, G. C. (2025a). 135 Years of Meteorological Records: IAC’s Historical Weather Station. Rpubs. Available at: https://rpubs.com/gabrielblain/1265900 Accessed on: Jan. 7, 2025.
    » https://rpubs.com/gabrielblain/1265900
  • Prela-Pantano, A., Brunini, O., Sobierajski, G. R., Bardin-Camparotto, L., Lopes, L. M. and Blain, G. C. (2025b). 135 Years of Meteorological Records: IAC’s Historical Weather Station. Zenodo. https://doi.org/10.5281/zenodo.14747078. Accessed on: Jan. 7, 2025.
    » https://doi.org/10.5281/zenodo.14747078
  • Pujol, N., Neppel, L. and Sabatier, R. (2007). Regional tests for trend detection in maximum precipitation series in the French Mediterranean region. Hydrological Sciences Journal, 52, 956-973. https://doi.org/10.1623/hysj.52.5.956
    » https://doi.org/10.1623/hysj.52.5.956
  • Reek, T., Doty, S. R. and Owen, T. W. (1992). A deterministic approach to the validation of historical daily temperature and precipitation data from the cooperative network. Bulletin of the American Meteorological Society, 73, 753-762. https://doi.org/10.1175/1520-0477(1992)073<0753:ADATTV>2.0.CO;2
    » https://doi.org/10.1175/1520-0477(1992)073<0753:ADATTV>2.0.CO;2
  • Regoto, P., Dereczynski, C., Chou, S. C. and Bazzanela, A. C. (2021). Observed changes in air temperature and precipitation extremes over Brazil. International Journal of Climatology, 41, 5125-5142. https://doi.org/10.1002/joc.7119
    » https://doi.org/10.1002/joc.7119
  • Rigby, R. A. and Stasinopoulos, D. M. (2005). Generalized additive models for location, scale and shape. Journal of the Royal statistical Society: Series C (Applied Statistics), 54, 507-554. https://doi.org/10.1111/j.1467-9876.2005.00510.x
    » https://doi.org/10.1111/j.1467-9876.2005.00510.x
  • Santos Jr., E. P., Xavier, A. C. F., Martins, L. L. and Sobierajski, G. R., Blain, G. C. (2022). Using a regional Frequency analysis approach for calculating the Standardized Precipitation Index: an operational approach based on the two-parameter gamma distribution. Theoretical and Applied Climatology, 148, 1199-1216. https://doi.org/10.1007/s00704-022-03989-7
    » https://doi.org/10.1007/s00704-022-03989-7
  • Stagge, J. H., Tallaksen, L. M., Gudmundsson, L., Van Loon, A. F. and Stahl, K. (2015). Candidate distribution for climatological drought indices (SPI and SPEI). International Journal of Climatology, 35, 4027-4040. https://doi.org/10.1002/joc.4267
    » https://doi.org/10.1002/joc.4267
  • Storch, H. V. and Zwiers, F. W. (1999). Statistical analysis in climate research. Cambridge: Cambridge University Press.
  • Sugahara, S., Da Rocha, R. P. and Silveira, R. (2009). Non-stationary frequency analysis of extreme daily rainfall in Sao Paulo, Brazil. International Journal of Climatology, 29, 1339-1349. https://doi.org/10.1002/joc.1760
    » https://doi.org/10.1002/joc.1760
  • Valverde, M. C. and Rosa, M. B. (2023). Heat waves in São Paulo State, Brazil: Intensity, duration, spatial scope, and atmospheric characteristics. International Journal of Climatology, 43, 3782-3798. https://doi.org/10.1002/joc.8058
    » https://doi.org/10.1002/joc.8058
  • Van Oldenborgh, G. J., Wehner, M. F., Vautard, R., Otto, F. E. L., Seneviratne, S. I., Stott, P. A., Hegerl, G. C., Philip, S. Y. and Kew, S. F. (2022). Attributing and projecting heatwaves is hard: we can do better. Earth’s Future, 10, e2021EF002271. https://doi.org/10.1029/2021EF002271
    » https://doi.org/10.1029/2021EF002271
  • Wang, X. L. (2008). Accounting for autocorrelation in detecting mean shifts in climate data series using the penalized maximal t or F test. Journal of Applied Meteorology and Climatology, 47, 2423-2444. https://doi.org/10.1175/2008JAMC1741.1
    » https://doi.org/10.1175/2008JAMC1741.1
  • Wijngaard, J. B., Klein Tank, A. M. G. and Können, G. P. (2003). Homogeneity of 20th century European daily temperature and precipitation series. International Journal of Climatology, 23, 679-692. https://doi.org/10.1002/joc.906
    » https://doi.org/10.1002/joc.906
  • Xavier, A. C. F., Morais, M. V. and Blain, G. C. (2019). Selecting the best nonstationary generalized extreme value (GEV) distribution: on the influence of different numbers of GEV-models. Bragantia, 78, 606-621. https://doi.org/10.1590/1678-4499.20180408
    » https://doi.org/10.1590/1678-4499.20180408
  • Yue, S., Pilon, P. and Cavadias, G. (2002). Power of the Mann-Kendall and Spearman’s rho tests for detecting monotonic trends in hydrological series. Journal of Hydrology, 259, 254-271. https://doi.org/10.1016/S0022-1694(01)00594-7
    » https://doi.org/10.1016/S0022-1694(01)00594-7
  • Yule, E. L., Hegerl, G., Schurer, A. and Hawkins, E. (2023). Using early extremes to place the 2022 UK heat waves into historical context. Atmospheric Science Letters, 24, e1159. https://doi.org/10.1002/asl.1159
    » https://doi.org/10.1002/asl.1159

*

Corresponding author: gabriel.blain@sp.gov.br

Section Editor:

CONFLICT OF INTEREST

Nothing to declare.

Publication Dates

  • Publication in this collection
    30 May 2025
  • Date of issue
    2025

History

  • Received
    14 Jan 2025
  • Accepted
    01 Apr 2025
location_on
Instituto Agronômico de Campinas Avenida Barão de Itapura, 1481, 13020-902, Tel.: +55 19 2137-0653, Fax: +55 19 2137-0666 - Campinas - SP - Brazil
E-mail: bragantia@iac.sp.gov.br
rss_feed Acompanhe os números deste periódico no seu leitor de RSS
Ir para o topo