Water Isotopologues Long-Term Continuous Rainfall Monitoring Contribution for Modeling Present Climate: Information Obtained from Different Time Scale Observations

Abstract

Rainfall isotopic composition has been continuously monitored at the central portion of the São Paulo state (Brazil) in different sampling time scales since 2013. The integration of different meteorological data, such as surface data from meteorological stations, HYSPLIT trajectories, reanalysis and ERA-interim data, has led to observed different conclusions based on the isotopic observation time scale. The amount effect in tropical areas is important for isotopic monthly data, explaining classical effects on monthly data, such as seasonality (high (low) isotopic composition during the dry (wet) period). Based on a daily scale, the interpretation is more complex, leading to controls on isotopic composition related to moisture source/transport and convective activity, as well as some local factors. Using microrain radar with GOES-16 imagery to identify the rainfall type, we were able to understand the cloud microphysics and sub-cloud processes responsible for rain isotope composition variation during the event. This combination of isotopic data may provide substantial subsidies and information for coupling isotopic data in GCMs. The incorporation of water isotopes into GCMs has enabled a more comprehensive evaluation of the water cycle and improvements in hydrometeorological simulations. This contribution has provided new insights into present, past, and future climate.

1. Introduction

The temporal and spatial monitoring of the atmosphere, ocean, and land surface has evolved rapidly in recent decades due to the technological revolution [1, 2], developing the World Weather Watch Global Observing System (GOS) [3]. The GOS has contributed to the understanding of climate phenomena, generating quality data that allows atmospheric scientists to measure, estimate, model, and unravel the past and project the future of the Earth’s climate. Among the range of knowledge about climate produced in recent decades, it has been highlighted that man has become a direct agent of meteorological processes, promoting climate change [4].

The perception of extreme and harmful events in our daily lives has become increasingly pronounced due to ongoing changes. It is possible to gauge the occurrence of these extreme weather events on a global scale through the maintenance of the GOS, as reported by the World Meteorological Organization (WMO) [5]. According to the WMO, between 1970 and 2019, there were more than 11,000 disasters attributed to climate and water hazards, which accounted for just over 2 million deaths and US$3.64 trillion in losses [5]. Assessments like this one become possible not only by improving the quality of meteorological data using technologies but by maintaining them over time, allowing comparison with information from the past.

Centuries-old surface weather stations (https://public.wmo.int/en/our-mandate/what-we-do/observations/centennial-observing-stations) are a clear example of the importance of long-term observations, constituting the backbone of global meteorological coverage, serving as a basis for climatology and weather forecasting studies, as they allow an incisive assessment of climate variability.

Another incisive example of long-term monitoring also encouraged by the WMO in partnership with the International Atomic Energy Agency (IAEA) was the creation of the Global Network of Isotopes in Precipitation (GNIP) (https://www.iaea.org/services/networks/gnip). The GNIP was created with the initial objective of monitoring the concentrations of tritium (³H – radioactive isotope) in the atmosphere, produced by nuclear tests during the interwar period. Later, stable isotopes of rainwater, oxygen (¹⁸O, ¹⁷O, ¹⁶O), and hydrogen (²H, ¹H) also became part of the monitoring scope, which began in the 1960s. Presently, GNIP has a monthly, online, and free database of the isotopic composition of precipitation and monthly meteorological data (precipitation and temperature) from more than 1200 stations installed in ~100 countries around the world. This database has been widely used in studies on global climate, meteorology, ecology, and hydrology, contributing to verifying and improving climate models, and hydrological models, involving surface water and aquifers at different spatial scales.

In Brazil, 28 GNIP stations were installed and rain samples were collected between 1957 and 1990 [6]. Despite having different objectives, these stations generated results from the first isotopic studies in the national territory, with emphasis on studies carried out in the Amazon, with the aim of understanding the role of the forest in generating humidity and its influence on the local hydrological cycle. Unfortunately, this effort has been discontinued, generating a large gap in isotopic data, consequently on part of the understanding of the movement of water in several hydrographic basins and its relationship with atmospheric processes. Only in 2008, a new NGIP was installed in Brazil, in the city of Belo Horizonte, followed by Rio Claro (2013), both in the southeastern region of the country. The Geological Survey of Brazil (SBG-CPRM) and the National Water Agency (ANA), with the support of the IAEA, resumed the operation of a monthly monitoring network of the isotopic composition of rainfall in several regions of Brazil, contributing to subsidize several hydrogeological studies in the future.

There is no way to project the future without assessing the past, so this present gap in monitoring the isotopic composition of rainwater, rivers, and aquifers has impaired an integrated assessment of atmospheric and hydrological processes in Brazilian watersheds, especially in the context of climate changes. Despite climate change being named as global, it is in the local context of each river basin that the main changes in the hydrological cycle are effectively felt. Understanding how these local hydrological changes are related to atmospheric processes of different temporal and spatial scales requires the use of tools capable of promoting this integration, and hence the great differential and applicability of the use of stable isotopes of water [7, 8]. For this assessment to be able to indicate what changes have been taking place over the years, long-term monitoring is essential.

In addition to the importance of stable isotopes for understanding hydrological studies, the implementation of water isotopes in General Atmospheric Circulation Models gives confidence to climate projections, as it represents physical–chemical processes without the need for large parameterizations and is perhaps one of the most useful for understanding climate processes on a global scale, involving precipitation and atmospheric circulation regimes [9].

In this context, the development, evolution, and current status of monitoring the isotopic composition of rain at the Rio Claro GNIP station (code: 8374701), operating from February 2013 to the present date, will be presented, revealing its contribution to the understanding of the atmospheric processes of the present climate. During this collection period, several atmospheric phenomena occurred and contributed to the formation of rain and changes in Rio Claro weather. Among these phenomena, some examples, such as the influence of El Niño-Southern Oscillation events during different periods in its warm phase (2014–2016, 2018–2019) and cold phase (La Niña, 2016, 2016–2017, 2020–2021, 2021–2023) according to the ONI-INDEX, extreme events related to hailstorms [10], above-average daily volume (~100 mm/day) and decreased rainfall for a period of extreme drought [11].

The observation of these and other atmospheric rainfall systems and their relationship with the isotopic composition of water was only possible due to continuous and long-term monitoring carried out with the monthly sampling (a monthly composite sample collected on the first day of every month at 12UTC) and with daily collection (e.g., a sample collected between 12:00 UTC on the 9th and 12:00 UTC on the 10th) started in February 2014. Subsequently, seeking to expand the assessment of the isotopic composition of rain, a collection was carried out at high frequency (in minutes, also named intra-event) to assess isotopic evolution during the passage of an individual rainfall event.

Thus, combining different sampling time scales with meteorological data of different spatial and temporal resolutions, the main objective of this book chapter is to share how isotope monitoring at different collection scales contributes to the understanding of the meteorological processes related to rain. Two questions serve as a basis for determining the proposed objective: (i) How has monitoring evolved over time? (ii) What was revealed by each type of collection? In order to present this contribution, this chapter is subdivided into two sections: Section 2 presents the basic concepts about stable isotopes and their application in rain; Section 3 presents the comparison of the main interpretations of meteorological controls on isotopic composition between monthly, daily, and high-frequency collection scales; Section 4, discuss how and why the monitoring of the isotopic composition of rain was carried out at different collection scales; and finally the last section, where the main conclusions are presented.

2. Deciphering the history of water with stable isotopes

The word isotopes originated from the Greek: Iso “equal,” topes “place,” means that atoms occupy the same place, that is, the same position in the periodic table, being constituted of a different number of neutrons, therefore of different masses, resulting in molecules made up of “light” and “heavy” atoms. Considering the water molecule, called isotopologues, the stable isotopes of hydrogen (H) and oxygen (O) most used in atmospheric and hydrological sciences are ¹H (light) e deuterium (D) or ²H (heavy), ¹⁶O (light) e o ¹⁸O (heavy), respectively [12, 13].

As the measurements of the amounts of these isotopes are not absolute, but represent the ratio between the least abundant isotope (heavy) over the most abundant (light), (for more details on the isotopic abundance see [12, 13]) the notation δ is used, expressed in parts per thousand (‰) and a reference standard for comparison (the δ of a sample collected in a given location is compared with this reference standard). The most used reference standard is the Vienna Standard Mean Ocean Water—VSMOW (Eq. (1)), which represents the average isotopic composition of ocean waters. Thus, positive values of δ indicate isotopic ratios that exceed the VSMOW and negative values of δ indicate ratios lower than the VSMOW [12].

$$\delta {}_{}{}^{18}Oou\delta {}_{}{}^{2}H=\left(\frac{{\frac{{}_{}{}^{18}O(ou{}_{}{}^{2}H)}{{}_{}{}^{16}O(ou{}_{}{}^{1}H)}}_{amostra}-{\frac{{}_{}{}^{18}O(ou{}_{}{}^{2}H)}{{}_{}{}^{16}O(ou{}_{}{}^{1}H)}}_{padrão}}{{\frac{{}_{}{}^{18}O(ou{}_{}{}^{2}H)}{{}_{}{}^{16}O(ou{}_{}{}^{1}H)}}_{padrão}}\right)\times 1000(‰)$$E1

Throughout the hydrological cycle, from the process of evaporation of water molecules in the ocean (the main source of moisture in the world) to the formation of rain on the continents, where water molecules reach the earth’s surface, recharging rivers, soils, and later aquifers, several water phase change processes modify the ratio of light and heavy isotopes in the water molecule. The physical-chemical process responsible for the partitioning of isotopes during water phase changes is called isotopic fractionation [12, 13, 14, 15], forming distinct water molecules, which vary spatially and temporally, registering an unique isotopic signature, functioning as a “fingerprint” about the paths taken throughout this cycle, which makes it possible for scientists to tell the story of water.

In relation to precipitation, there are two types of fractionation that directly affect the isotopic composition of rainfall: (a) equilibrium fractionation, related to preferential exchanges that different substances have for a given isotope (occurs during the process of condensation inside clouds, formation of ice and rain droplets) [7, 16, 17, 18] and (b) kinetic or nonequilibrium fractionation, related to different rates of reaction between molecules (occurs in evaporation, isotopic exchange with surrounding vapor and reevaporation of drops in the rain) [17, 19].

In isotope studies of rain, relative terms, enriched and depleted, are generally used to denote whether the heavy isotope content is higher (rich in heavy isotopes) or lower (poor in heavy isotopes), respectively. In this sense, when water evaporates from the ocean, molecules with lighter isotopes (¹H₂¹⁶O) tend to evaporate and form water vapor, depleted in heavy isotopes compared to the water that gave rise to it. Conversely, when it rains, molecules with heavier isotopes tend to precipitate (¹HD¹⁶O e or ¹H₂¹⁸O), forming an isotopic composition enriched in heavy isotopes [7, 12].

This difference between light and heavy water is mainly determined by the amount of fractionation that a given portion of steam suffered, generated rain and interacted with the surface, consequently with new processes of evaporation, evapotranspiration, and subsequent condensation, from its origin (source area) to the location where rainwater was collected. Thus, the isotopic composition of rain will be more depleted (loss of heavy isotopes along the path of the steam) the farther it is from its source of water [12, 15, 20]. This so-called Rayleigh distillation concept is essential for understanding the regional processes that affect the isotopic composition of rain, as it is related to the origin of steam, moisture transport, interaction with regional atmospheric circulation, and atmospheric systems, revealing the history of rainfall.

Locally, when rainfall falls on a collection point, a stage called post-condensation processes, it is also subject to two fractionation processes: (i) rainwater composition balances with the surrounding humidity and becomes enriched. This balance depends on droplet size and relative humidity, which at lower levels is less depleted than rainfall at the cloud base, so the isotopic composition of surface rain closely resembles that of surface moisture [21]; The larger the raindrop, the greater its falling velocity and the less exchange with the surrounding vapor, resulting in δ¹⁸O depletion [9]; (ii) evaporation of the drops that fall on a layer of low humidity enriches the remaining rain, making the surface rainfall more enriched [22, 23, 24, 25, 26]. This process has been observed mainly in desert areas [15, 21, 27], although it is also seen in continental areas, mainly in light rains in a relatively dry atmosphere [20].

Thus, by combining regional and local fractionation processes, it is possible to determine the isotopic signature of rain, but what makes this identification possible is the way in which the isotopic composition of rain relates to geographic factors and climatic elements, i.e., the approach that is adopted, intends to used to interpret the history of rain. In this way, the sample collection scale determines the degree of interpretation that one intends to have on the isotopic composition of rain, so that regional processes can overlap local processes, and vice versa.

3. Evolution of sampling frequency scales and possibles interpretation of meteorological and isotopic processes

The dynamics between water phase changes and the fractionation process characterize the spatial and temporal variation of isotopic variability in different areas of the globe, since distances such as moisture source, latitude, climate, altitude, land use and land cover, and the acting atmospheric systems are entirely distinct. In this sense, isotopologues can be related to climate dynamics at different temporal and spatial scales, ranging from minutes to hundreds or thousands of years (climate variability), and spatially ranging from the micro-scale (less than 1 km), in atmospheric turbulence, to synoptic (>2000 km) involving large cloud assemblages (Figure 1) [13, 20, 28, 29].

Figure 1.

Synthesis of the relationship between atmospheric systems, isotopic sampling scale, and main meteorological interpretations for understanding climate over time and space.

Figure 1, summarizes the combination of spatial and temporal variation between sampling frequency and atmospheric systems, illustrating what spatial and temporal level a study can incorporate, resulting in different interpretations of how meteorological processes explain rainfall isotopic variability.

For the monthly sampling frequency, a composite sample is collected, in other words, the monthly isotopic signature represents the sum (of n events) or the rain-weighted average of the isotopic fractionation processes that occurred in all rainfall events in the interval of 1 month. With this isotopic signature, it is possible to interpret meteorological processes acting on a large spatial (>2000 km) and temporal (months and years) scale, such as intraseasonal, seasonal, ENSO (interannual) events, and even long-term climate variations, when isotopic monitoring occurs for at least 10 years [30]. Classical effects (such as temperature, latitude, altitude, continentality, and amount) are the main interpretations of the isotopic composition of rainfall.

The earlier research and interpretation were based on monthly samples, and it was developed as the background of concepts and interpretations of the isotopic composition of rainfall, using GNIP database. One of the main elements of the explanation of isotopic variability of meteoric waters is the Global Meteoric Water Line (GMWL), which represents a global linear relationship between δ¹⁸O-δ²H, defined by the equation: δ²H = 8 * δ¹⁸O + 10 [31]. Subsequently, using GNIP data, a new GMWL equation was computed (δ²H = 8,17 * δ¹⁸O + 11,27) [20], despite the Craig’s equation is widely used. The GMWL is a reference for local studies in different climatic regions of the globe and compares to a Local Meteoric Water Line (LMWL), which reflects the average and local relationship between δ¹⁸O-δ²H. Since, for a given temperature range and isotopic composition, deviations from the LMWL occur by equilibrium (δ¹⁸O-δ²H values around the GMWL) or kinetic processes (slopes of the line different from 8) [30, 32].

The LMWL provides an assessment of the spatial variation of rainfall isotopic composition by comparing isotopic data from different stations and identifying information on seasonal climatology [20, 31]. In addition to the comparison between rain waters, a LMWL serves as a reference to interpret the isotopic composition of other waters (soil, rivers, lakes, groundwater) and plants (stems and leaves), allowing the understanding of water movement and the interaction between these different compartments in a given watershed [7, 30, 33].

Another second-order parameter, which helps to explain the nonequilibrium (kinetic) processes of isotopic fractionation of H isotopes in relation to those of O, is the deuterium excess (d-excess), defined: d = δ²H – 8* δ¹⁸O, represented by the value of 10 in GMWL [15] and deviations of LMWL. Its variation is usually related to temperature gradients in the regions where the vapor origin is located (most of it is oceanic), isotopic exchanges during vapor transport over continents (d > 10‰) [34, 35], local evaporation processes (d < 10‰), during the falling raindrops toward the surface [15, 21, 36] and in surface and groundwater studies (aquifer recharge) [6, 7, 37].

The application of these definitions has enabled a spectrum of interpretations of isotopic variability, characterized by the classic “effects” that have temperature dependence as an important element: In latitude effect, in the equatorial zone (warm) are more enriched isotopic values and in the poles (cold) more depleted isotopic values; altitude effect, occurs due to temperature decrease with increasing altitude, resulting in depletion of isotopic composition; continentality effect, associated with the transport of vapor and rainfall from the ocean to the continent, isotopic signatures in coastal areas are more enriched than continental areas (more depleted), explained by the Rayleigh distillation process; temperature effect, observed in seasonal variations in the northern hemisphere and at higher latitudes, where there is a large temperature range between summer (more enriched) and winter (more depleted). Differences in the rainfall regime over the year also characterize the amount of effect in tropical areas, defined as the negative linear relationship between δ¹⁸O and the amount of rainfall, so the greater the amount of rainfall, the greater the depletion of the monthly isotopic composition due to successive condensation processes [12, 15, 20].

In accordance with the evolution of climate monitoring technologies associated with the improvement on isotopic determination with the implementation of new techniques such as Laser Absorption Spectroscopy has increased the capacity to determine the number of samples [38], enabling the expansion of isotopic monitoring to daily sampling and increased studies of individual rainfall events with high-frequency sampling.

For the daily sampling scale, a single daily isotopic signature represents the sum of one or more events collected during 1 day or the daily rainfall-weighted average of the isotopic fractionation processes that occurred in these rainfall events over the interval of 1 day. Using the daily isotopic signature is possible to interpret the meteorological processes that operate on a large spatial and temporal scale, as demonstrated by the monthly data from continuous long-term monitoring. However, the daily scale decreases the mixing and overlapping effect between rain events and consequently the different types of isotopic fractionations that occurred during the rain formation, enabling a better resolution to understand the isotopic variability of the different types of weather, identifying usual events from extreme events. For this reason, the daily scale improves the analysis of the interannual variations and seasonal cycles, as well as the action of atmospheric systems between 1000 km and 10 km, in the range of days to a week, such as tropical cyclones, squall lines, and cloud clusters (Figure 1). The daily isotopic composition of rainfall includes evaluations of the moisture origin/transport related to large-scale circulation systems, different types of rainfall, and the influence of convective activity. The daily isotopic composition of rainfall is also related to the classical effects that are better evaluated on a daily scale.

One of the main differences between the isotopic composition of monthly and daily rainfall was observed in the assessment of the amount effect in continental tropical areas. Strong correlations between monthly δ¹⁸O-rainfall decrease at the daily scale, due to several factors still under investigation. One of the best hypotheses explaining why daily and short-term isotopic variations do not correspond to rainfall amount, suggested the importance of convective processes in modulating the isotopic content of rainfall, reflecting the integrated history of convective activity over 4 days [36]. In organized convective systems, reevaporation processes of raindrops in mesoscale subsidence updrafts form a low-level depleted vapor in the atmosphere, feeding successive convective systems in that interval of integrated convective activity [36]. Daily isotopic data provide an understanding of the evolution of this convective history, so the isotopic composition is not only related to its final rainfall amount, offering great proxy information for understanding how convection transforms water vapor into rain, after all this mechanism is essential in the energy balance and distribution of water across the globe.

At the high-frequency sampling scale, several samples are collected, hence the variation of several isotopic signatures over the evolution of an individual rain event is observed, forming an isotopic trend of variation, always mentioned in high-frequency studies [17, 21, 22, 23]. Evaluating the isotopic trend provides an interpretation of meteorological processes that operate at smaller spatial (>100 km) and temporal (hours to minutes) scales, such as cloud clusters, thunderstorms, tornadoes, and local rainfall (Figure 1). Using intra-event is possible to evaluate in detail several components, which involve the path, structure and evolution of the storm, the atmospheric system that originated it, changes in the air mass, altitude at which rain is produced (condensation level), type of rain, intensity of rainfall, and local and microphysical processes that occur inside and below the clouds, such as diffusive exchanges between low-level vapor and raindrops (main the evaporation of raindrops), and the relationship with surface meteorological data [17, 20, 21].

The application of these concepts are shown in the next chapter, based on the main results obtained during the 9-year monitoring of the isotopic composition of rainfall in Rio Claro, at different collection scales.

4. Monitoring the isotopic composition of rainfall in Rio Claro

Due to the need to understand how atmospheric processes control the formation and variability of rainfall, which is the main input in aquifer water recharge, rainfall stable isotope sampling stations were installed in several localities (Rio Claro, Brotas, and Araraquara), located in the central-southern portion of São Paulo state. This region is one of the recharge areas of the Guarani Aquifer System (SAG). In this context, the Rio Claro station was affiliated with GNIP, latitude: −22.39°S, longitude: −47.54°W and elevation of 670 m.a.g.l, starting with a monthly collection of stable isotopes of rainfall, sent to the IAEA laboratory in Vienna, along with rainfall and precipitation data.

In order to expand the monitoring of the isotopic composition of rainfall and understand the role of Amazon moisture in the formation of rainfall in the south-central portion of the state of São Paulo, southeastern Brazil, daily samples started to be collected in February 2014. Results, differences, and interpretations of meteorological controls on the isotopic composition of monthly and daily rainfall were discussed in Section 4.1.

The high-frequency rain sampling (details in Section 4.2) was important to understand the isotopic variability during the passage and evolution of different types of rainfall, convective and stratiform, since monthly and daily assessments did not help to understand how these rainfall types control the isotopic composition in Rio Claro.

The Rio Claro region has an average annual rainfall of around 1500 mm, characterized by two distinct seasons. The first season is a rainy and warm spring-summer period occurring between October and March. The second season, which is cooler and less rainy, occurs in the autumn-winter period between April and September. Among the primary regional weather systems, the Cold Fronts (polar air masses), active throughout the year. The South Atlantic Convergence Zone (SACZ) is prevalent in summer and the South Atlantic Subtropical High (SASH), which inhibits rain formation during winter and contributes to the transport of moisture from the Atlantic Ocean to the continent in the other seasons of the year. The Atlantic Ocean serves as the primary moisture source in Brazil, with additional contribution from the Amazon rainforest’s evapotranspiration. The moisture from the Amazon is crucial to the formation of rainfall. It is transported across South America by low-level jets during the rainy season [39, 40, 41, 42].

4.1 Monthly and daily

Between February 2013 and December 2022, 109 monthly samples were collected in Rio Claro station, resulting in a variation from −12.65‰ to 3.07‰, arithmetic mean and standard deviation of −4.73 ± 3.27‰ for δ¹⁸O, and from −94.90‰ to 23.70‰ (−23.46 ± 4.22‰) for δ²H, and from −0.86‰ to 24.12‰ (14.39 ± 4.22‰) for d-excess. Considering the mean δ¹⁸O value, 56% of the samples were isotopic composition higher-4.73‰, while the remaining 44% can be considered depleted. For d-excess, lower values than <10‰ were observed in only 14% of the dataset, while 86% of these values were greater than >10%. These results characterize the dispersion of the isotopic composition values and help to visualize an overview of this variation, with slight distribution between enriched and depleted values, and lower influence of local processes (kinetic fractionation) that decrease d-excess values.

The δ¹⁸O and δ²H values were aligned around GMWL, characterized by the e Monthly LMWL (δ²H = 8.12 + δ¹⁸O * 15.24) with intercept close to the GMWL value (8) and higher slope (10) (Figure 2A), indicating a predominance of processes related to the continental moisture recycling, which explain the higher d-excess values observed.

Figure 2.

Overview of the isotopic composition variation in Rio Claro station. (A) Classic δ¹⁸O-δ²H relationship by monthly (2013–2022) and daily (2014–2022) dataset, global meteoric water line (GMWL) and monthly (green)/daily (orange) local meteoric water lines, (B) monthly average isotopic composition weighted by the amount of rainfall and monthly mean of rainfall, and (C) daily isotopic composition and rainfall.

One of the common practices in isotopic hydrology studies is the use of the average isotopic composition weighted by the amount of rainfall. This weighted average can be used for a seasonal, annual comparison, or for the full data set. In Figure 2B, the bars in blue are the monthly averages over the monitoring period and the points and lines in green are the average weighted by the amount of rainfall in each month (δ¹⁸O_wgd = ∑ (δ¹⁸O_month-i * P_month-i) / ∑ P_month-i). A clear seasonal distribution was observed, characterized by depleted values (δ¹⁸O < −6.54‰) during the rainy period (October–March) and enriched values (δ¹⁸O > −5.43‰) in the less rainy months (April–September). December (δ¹⁸O -7.39‰) and February (δ¹⁸O -7.38‰) were the most depleted months, while September was the most enriched (δ¹⁸O -0.59‰) (Figure 2B).

The weighted averages presented more depleted values for δ¹⁸O (−6.69‰) and δ²H (−24.12‰) compared to the arithmetic mean, and lower for d-excess (13.14‰), illustrating that the amount of rainfall is more relevant for isotopes than d-excess. The inverse relationship between rainfall amount and monthly isotopic composition characterizes the amount effect, of strong negative and significant (p-valor <0.0001) correlations, r = −0.58 (δ¹⁸O) and − 0.60 (δ²H). In contrast, the correlations between monthly temperature and isotopic composition was weak, r = −0.34 (δ¹⁸O) and − 0.37 (δ²H), despite the significance (p < 0.0002).

The moisture-recycled transport by air masses (LMWL) and local amount effect controlled the monthly isotopic composition of rainfall by the condensation-related mechanism (Rayleigh distillation). The seasonal variations in isotopic compositing were distinct due to the different sources of moisture, transport, available moisture, and the performance of different atmospheric systems [38]. However, the dynamics of these main synoptic features change from day to day, since the monthly isotopic composition overlaps the observed variability, making the evaluation very focused on months of high and low rainfall [38].

The daily isotopic composition (represented as orange dots in Figure 2A and C), was evaluated based on 674 samples collected between February 2014 and December 2022. Despite the daily values varied aligned to monthly data, greater range of values were observed for δ¹⁸O, δ²H and d-excess, −21.74‰ a 4.89‰ (−4.77 ± 4.34‰), −158.50‰ a 43.40‰ (−25.59 ± 34.45‰) e − 1.81‰ a 32.51‰ (12.62 ± 5.29‰), respectively. As observed in the monthly analysis, 57% of samples were more enriched (43% more depleted) in relation to the mean arithmetic δ¹⁸O values, which is also very similar to the monthly arithmetic mean. Values of d-excess lower than <10‰ were observed in 27% of the dataset, while 73% of these values were greater than >10‰. These results reinforce the good distribution between enriched and depleted rainfall, with a greater influence of local processes (kinetic fractionation that decreasing d-excess) being observed with daily data in relation to monthly data.

This influence of the kinetic fractionation is characterized by the daily LMWL (δ²H = 7.83 + δ¹⁸O * 11.84), of a lower intercept and slope compared to the monthly LMWL, despite the close values in relation to GMWL. Continental moisture recycling processes also influence the isotopic composition of daily rainfall, since higher d-excess values are observed in most samples analyzed.

A clear seasonal variation in “V-shaped” (Figure 2C, black arrows) was observed for the daily isotopic composition of the rainfall. Enriched δ¹⁸O values were predominant in the dry period, from April to September (δ¹⁸O > −0.40‰ and daily average rainfall of 7.2 mm/day), while depleted δ¹⁸O values (δ¹⁸O < −9.0‰ and daily average rainfall of 19.8 mm/day) during the rainy season, between November and February (black circles highlighted in Figure 2C). The mean and standard deviation of daily rainfall during the complete monitoring period was 12.39 ± 13.62 mm/day.

The daily weighted average isotopic composition for the entire dataset was −6.15‰, −35.52‰ e 13.67‰ for δ¹⁸O, δ²H and d-excess, respectively. Despite the weighting by quantify decreasing the values of δ¹⁸O, δ²H and increasing those of d-excess in relation to the daily arithmetic mean, the amount effect was not observed for the daily scale. The weak and nonsignificant (p-valor >0.05) Spearman correlations were observed between daily rainfall and δ¹⁸O (r − 0.24) and δ²H (r − 0.22). In addition, the same occurred for daily temperature correlations of δ¹⁸O (r − 0.10) and δ²H (r − 0.13).

The role of the moisture source/transport mentioned in the evaluation of the monthly data became even more relevant on the daily analysis. Since, it is possible to observe the change in the transport of moisture every day, improving the understanding of the formation of rainfall in Rio Claro, from the origin to the successive condensation processes that occur along the history of rain and its relationship with the atmospheric systems over the course of the days. One of the most used tools in rain isotope studies is the Hybrid Single-Particle Lagrangian Integrated Trajectory (HYSPLIT) model, a mathematical system that calculates trajectories and simulates the dispersion and deposition of particles in the atmosphere (http://ready.arl.noaa.gov/HYSPLIT_traj.php) [43]. Meteorological parameters (e.g., rainfall rate (mm.h⁻¹), temperature (K), humidity (%), and trajectory height (meters) from different databases can be used to compose the meteorological information of the trajectory. Trajectories can be determined in backward mode (refers to pastime trajectories) or forward (future time trajectories), calculated by time determined by the user, as well as the coordinates and initial altitude.

In Rio Claro, back-trajectories were determined in different studies [44, 45, 46]. The Atlantic Ocean is the main moisture source, followed by the Amazon Forest and, in lower events from South Brazil. Between these source regions and Rio Claro station, along pathways of moisture interact with regional circulation, atmospheric systems, and convective activity, resulting in the moisture recycling process, illustrated by the monthly and daily LMWL. This mechanism occurs in different conditions between the seasons of the year, characterizing the observed seasonal variability.

During the rainy season, the moisture from the Amazon interacts with the pressure gradients formed by the increase in temperature, favoring the formation of the SACZ, resulting in more depleted isotopic values observed on the monthly and daily scale. In opposite, during winter, the moisture from the Amazon decreases, being more associated with the Atlantic Ocean, hence rainfall occurs when the FF is strong enough to overcome the circulation of a high-pressure system, such as the South Atlantic Subtropical High (SASH), resulting in a more enriched isotopic composition [44, 45, 47].

The comparison of the isotopic composition rainfall in Rio Claro, with other GNIP stations was also carried out, with the objective of spatially extending the Rio Claro analysis, confirming the role of these mentioned regional processes. This influence of available moisture and different moisture transport conditions was also observed for Belo Horizonte, Brasília, and Rio de Janeiro stations [47]. In this work, the seasonal, continental, and amount effects on the monthly isotopic composition of rainfall in all locations were confirmed. For the daily isotopic composition, one of the main meteorological controls on the variability in δ¹⁸O and d-excess values was the strong convective activity (the deeper and more organized the convection, the greater depletion in the isotopic composition) during the summer. Meteorological parameters by reanalysis, such as Precipitable Water (kg m⁻²), Outgoing Longwave Radiation (OLR - W.m⁻²), 500 hPa vertical velocity field or Omega (Pa s⁻¹) were used to indicate the convective activity, and they were associated with daily isotopic composition.

A comparison of the isotopic composition of rainfall in different El Niño-Southern Oscillation (ENOS) events, 1997–1998 (ENOS 1) and 2014–2016 (ENOS 2), were carried out. The daily isotopic data from Rio Claro, Bragança Paulista, Campinas, Piracicaba e Santa Maria da Serra in the central-east portion of São Paulo state were used [48]. The same seasonal effect was identified, and mainly the influence of different available moisture conditions, which was higher in the dry season of ENSO 1, generating impoverishment in the isotopic composition of rainfall (δ¹⁸O = <−4.60‰) compared to the dry season of ENSO 2, whose available moisture was lower and, consequently, the most enriched rainfall (δ¹⁸O = <−2.80‰).

In addition, statistical tests presented in previous studies [45, 46] confirmed the influence of regional parameters on isotopic variability. In these studies, linear regression models were applied with significant results, explaining part of the isotopic variability, mainly values around δ¹⁸O = −4.0‰ e − 5.0‰. The regression models, resulting in the need to investigate these rainfall events, did not explain strong negative and positive δ¹⁸O values (Figure 2C).

Finally, the isotopic composition of daily rainfall was also associated with the distinction of different types of rain, convective (high δ¹⁸O values) and stratiform (low δ¹⁸O values) [49], relationship that has been widely investigated in different parts of the world. For Rio Claro, no good correlation was observed between δ¹⁸O and convective and stratiform rainfall from two different databases, classification from the Global Precipitation Measurement (GPM) [50] and the ERA-interim (convective and large-scale precipitation converted in mm) provided by the European Center for Medium-Range Weather Forecasts (ECMRWF) [47].

4.2 High-frequency

For understanding the evolution of strong depleted and enriched rainfall events observed on a daily scale, identifying, and classifying rainfall types (convective and stratiform) and local processes related to the falling raindrops in Rio Claro, high-frequency collection was implemented.

A total of 312 samples were collected between 5, 10, and 30 minutes, from the beginning to the end of 18 rainfall events, between September 2019 and February 2021. As the collection was carried out manually, it was very difficult to be in the university facilities since the beginning of the rain. Therefore, the collection of events was performed randomly. Even so, in all seasons of the year, some intra-event was collected, covering a diverse range of atmospheric systems (frontal systems, prefrontal and postfrontal atmospheric instability, atmospheric instability thermal atmospheric during the summer, trough, and SACZ) and rainfall types (convective, stratiform, mixed (mixture between convective-stratiform during the same event), and localized rain) [51].

The isotopic composition of intra-event was combined with the meteorological data of high temporal resolution (minutes per hour), such as surface data (rainfall (mm.min⁻¹), temperature (°C), relative air humidity (%) and pressure (kPa), 1-minute interval) from meteorological automatic station (METER - Em50); vertical profile of the atmosphere over the collection point by micro rain radar (METEK MRR-2 operates at a frequency of 24.230 GHz with a modulation of 0.5–15 MHz) provide the reflectivity (Zc - dBZ), fall velocity (w – m.s⁻¹), rainfall rate (mm.min⁻¹) and liquid water content (g.m⁻³), in 1-minute interval and height resolutions in a range bin of 31 measurement heights; ERA-5 vertical column of (humidity (%), temperature (°C), liquid water (kg.m⁻³) and ice (kg.m⁻³) content; imageries from the GOES-16 satellite (to identify convective nuclei and monitoring the formation and evolution of clouds with the brightness temperature (°C) of the images); determination of trajectories by the HYSPLIT model and regional data from ERA-5, easterly vapor flux (kg.m⁻³), latent heat flux (W.m⁻²) and OLR (W.m⁻²) [51].

Rainfall types were classified employing a micro rain radar, GOES-16 imagery, and surface meteorological station. Radar images depicted distinct vertical structures for rainfall. Convective rainfall displayed a vertical structure, whereas the stratiform exhibited a horizontal structure. The horizontal structure’s feature is the melting layer (or bright band in radar images) which can be quantified by a difference of approximately 4 dBZ in radar reflectivity between radar-measured heights. The identification of the melting layer determines the incidence of stratiform rainfall, and this analysis was conducted for all rainfall events evaluated in this study. The vertical structure is inadequate to confirm the occurrence of convective rainfall. The GOES-16 image was used to identify convective nuclei by analyzing a set of 40 pixels over Rio Claro with a brightness temperature lower than −38°C. The rainfall intensity was computed considering at least 10 mm per hour to determine convective rainfall. During the event, the mixed rainfall was a combination of convective and stratiform rainfall, while local rainfall was defined by the absence of a melting layer and convective nuclei [51].

This isotopic and meteorological dataset on a short temporal scale of minutes resulted in interesting explanations about the evolution of rainfall in Rio Claro, indicating different meteorological controls for convective, stratiform, mixed, and local rainfall.

Figure 3 illustrates an intra-event classified as mixed rainfall. This event occurred during the summer and was formed by predominant moisture from the Amazon Forest, according to HYSPLIT trajectories, where over the state of São Paulo it interacted with the passage of a cold front, characterizing a frontal system (warm mass from Amazon and cold mass from the polar) that produced cloud systems observed in GOES-16 image (Figure 3A). These systems generated rainfall over Rio Claro station, starting at 12:48 pm until 2:28 pm (local time, -3UTC).

Figure 3.

Summary of local information used in the intra-event analysis. (A) GOES-16 image over the Rio Claro station, included convective nuclei around the station and in São Paulo state, (B) vertical profile of fall velocity (w) of micro rain radar for convective (vertical structure) and stratiform part (horizontal structure marked by the melting layer), (C) 18O (green), d-excess (red) and rain rates (blue), and (D) relative humidity (dark blue) and temperature (dark red). In B, C, and D, the vertical dot line separates the convective and stratiform parts of the event.

In the first part of this event, the convective fraction was observed, characterized by a convective core in the GOES-16 image with a brightness temperature < −38°C in at least 40 pixels around the collection point (Figure 3A), structure vertical of the fall velocity by micro rain radar, without the presence of the melting layer (Figure 3B). In the second part of this event, a change in the vertical structure of the fall velocity was observed, with the presence of the melting layer (the values of w and reflectivity increase when this layer occurs), characterizing the stratiform fraction of this rain event (Figure 3B).

The isotopic composition of the rainfall responded to the evolution of the mixed event, with greater variation in δ¹⁸O values (−10.29‰ ~ −11.61‰) during the convective phase and constant δ¹⁸O values in the stratiform phase (−10.07‰ ~ −10.97‰) (Figure 3C). For d-excess the variation was inverse to the variation of δ¹⁸O. Lower d-excess values were observed during the transition from convective to stratiform phase (7.06‰) and at the end part of event (5.90‰ e 4.30‰) (Figure 3C). Relating the δ¹⁸O values with rainfall rate, temperature, and relative humidity, it was observed that the lowest δ¹⁸O value occurred in the interval of greatest rainfall intensity (Figure 3C), while temperature and humidity accompanied the small variation in δ¹⁸O values during the entire period of the event (Figure 3D).

The event in Figure 3 illustrated how the change in the rainfall types shifted the δ¹⁸O and d-excess shape of variation, modulating the kinds of fractionation, in equilibrium (lower values of δ¹⁸O in the highest rainfall intensity) and or kinetic (lower values of d-excess due to a change in the vertical structure of the rain, reflecting lower rain rates, hence the isotopic exchange between raindrops and low-level vapor) during the stratiform phase [51].

A comparison between intra-events of stratiform rainfall indicated that the isotopic composition represents the life cycle of the rain system. During the passage of the stratiform cloud over Rio Claro was observed that the rainfall events in the development phase were enriched in relation to events of the mature and dissipating phase (very depleted isotopic composition) [51].

High-frequency sampling not only provides an assessment of intra-event isotopic composition, but it can also facilitate an understanding of processes occurring over the course of a day, providing more than one isotopic signature. Thus, convective rainfall events reflected diurnal differences in isotopic composition, producing strong negative δ¹⁸O values and higher d-excess during the night due to higher humidity conditions between the cloud base and the surface in relation to the daytime conditions, δ¹⁸O less negative and smaller d-excess.

5. Conclusions

The continuous monitoring of the isotopic composition of rainfall in the GNIP Rio Claro station, Brazil, produced an interpretation of the present climate in the central-southern portion of São Paulo state, on different sampling scale. Monthly, daily, and high-frequency sampling scales were used like the vision instrument to see the dynamic between atmospheric systems linked to the rainfall and the water stable isotopes.

The monthly scale generated an overview of the classic effects (seasonal, rainfall amount, and continentality) in the isotopic composition of rainfall. These interpretations and LMWL are relevant references of isotopic signatures to relate the regional atmospheric dynamics with the local hydrological in the watersheds localized in the Rio Claro region.

The daily isotopic composition offered a more detailed interpretation viewfinder, characterized by monitoring the successive kinds of rainy weather (from synoptic to local spatial scales) day by day (from day to a week), resulting in temporal isotopic observations with clear seasonal patterns linked to stronger convective activity and moisture transport from Amazon during spring-summer (low isotopic composition) and cold fronts activity combined with moisture transport from Atlantic ocean during autumn-winter (high isotopic composition).

For high-frequency or intra-event scale, the interpretation view is enlarged like a magnifying glass zoom, about the formation and evolution of a rain system locally. The rainfall types and their influences on the isotopic composition were improved, clarifying the strong δ¹⁸O-δ²H values and revealing new meteorological approaches like the diurnal isotopic composition differences linked to the day-night convection, and life cycle of stratiform rainfall.

Therefore, the history of the isotopic composition of rainfall in Rio Claro is characterized by the Atlantic Ocean and Amazon Forest moisture origins, transported over the long continental pathways resulting in moisture recycled. This mechanism was observed in all sampling scales, with a major emphasis on the daily scale. During these pathways, the moisture interacted with regional circulation, atmospheric systems, air masses, and convective activity removing the heavy isotopes in rainfall and forming the subsequently depleted vapor and next rainfall until arrived at the collection point. Locally, the convective and stratiform rainfall change the variations of δ¹⁸O and d-excess values that reflect their different formation, including the microphysical processes into cloud and the processes in the below cloud base during falling raindrops (vertical structure of rainfall) at the surface (meteorological data).

Our results provide meteorological observation data and key insights for isotopic interpretation in tropical areas. Based on the acquired knowledge, it will be possible to use isotopic information and establish its relationship with the climate and the occurrence of extreme rainfall events to improve the understanding of past atmospheric processes (recorded in groundwater, glaciers, and spelotems), recent (plants, river water, rain) and future (indicative of water scarcity and increase in extreme events). By studying past events, it is possible to improve our management of water usage and availability considering climate change scenarios. Furthermore, implementing the use of isotopic methods in researching lesser-known hydrological systems can provide solutions to the challenges faced in managing water and climate resources, specifically those impacting tropical regions.

Acknowledgments

The research that produced the isotopic monitoring program in monthly, daily, and high-frequency scales were funded by grants from the São Paulo Research Foundation (FAPESP) under Processes 2015/15749-2 and 2018/06666-4, and by the International Atomic Energy Agency grant BRA-17984 under the initiative CRP-F31004 “Stable isotopes in precipitation and paleoclimatic archives in tropical areas to improve regional hydrological and climatic impact models,” and BRA-23531 under the initiative CRP-F31006 “Isotope Variability of Rain for Assessing Climate Change Impacts.” VS thanks FAPESP for the scholarship provided under the Processes 2013/06704-0, 2016/18735-5, 2019/03467-3 and 2021/10538-4. This study also was financed in part by the Coordination of Superior Level Staff Improvement (CAPES).

Conflict of interest

The authors declare no conflict of interest.