Background & Summary

Rain isotope studies provide critical insights into hydrological and climatic processes, aiding in understanding moisture sources, precipitation mechanisms, and atmospheric circulation patterns. Considering the importance of rain isotope datasets, the International Atomic Energy Agency (IAEA) developed a Global Network of Isotopes in Precipitation (GNIP). Monthly composite rainwater samples are usually collected through this network and the samples are analysed for oxygen (18O/16O) and hydrogen (2H/1H) isotope ratios. The isotopic ratios are generally expressed using a δ notation, where,

$$\delta =\left[\frac{{R}_{sample}}{{R}_{VSMOW}}-1\right]\cdot 1000\textperthousand $$

R is the ratio of the heavier isotope to the lighter isotope (2H/1H or 18O/16O).

Using GNIP datasets, Dansgaard1 illustrated relationships between precipitation isotopic composition and various climatic variables such as rain amount, temperature and raindrop evaporation. The study suggested that isotopes could be used to understand modern and past atmospheric processes as reliable proxies. Aggarwal et al.2 documented a strong correlation (R2 ≈ 0.6) between rain δ18O and stratiform rain fraction and explained the underlying processes. Using the GNIP datasets, many studies have further demonstrated that isotope values are controlled regionally by the aforementioned climate parameters in South Asia2,3,4,5,6,7. The same datasets were suitably used earlier to investigate the role of moisture source variation, rain and cloud microphysical processes on isotope values for ISM3,8,9,10. The isotope-climate relationships obtained from those studies are also helpful for climate reconstruction11,12,13. Studies have also shown that many mesoscale convective events, short-term variations in moisture sources, and cloud microphysical processes strongly control rain isotope variation in the tropics14,15,16. However, the GNIP datasets (of monthly composite samples) cannot preserve the signature of those high-frequency events for obvious reasons and thus limit their application in understanding high-frequency atmospheric processes.

Indian summer monsoon (ISM) rainfall plays a crucial role in the agroeconomy of India as well as the entire South Asia. Multiple moisture sources17,18,19 and various rain mechanisms9,20 in different parts of this country make it a prototype for studying the role of atmospheric processes on isotope values. However, only four GNIP stations in India have long-term (>10 years of data) isotope records among the 37 listed stations, which are actually very limited in number. The majority of other stations have data spanning only 1 to 4 years mostly during 2003 to 2006. Considering this limitation, we initiated daily/event-based sampling in some strategic Indian locations. A part of those datasets is reported here. A brief discussion of the important findings from the earlier studies is a prerequisite for understanding the importance and interpretation of our current dataset.

Various studies suggest that mesoscale convections originated during ISM affect isotope values21,22,23. The role of cloud microphysical processes on daily rain isotope values is also investigated earlier during ISM20,24. Similar to Aggarwal et al.2, discussed earlier, significant negative correlations between stratiform rain fraction and isotope values are also obtained on a daily scale, suggesting stratiform rain affects isotope values even during convective events25,26. The role of recycled moisture on isotope variation was also studied for both monsoon and non-monsoon rain samples27,28,29. The daily/ monthly rainwater isotope datasets generated during ISM were also helpful for validating various isotope-enabled general circulation models30,31,32. These models simulate water (rain/vapour) isotopes considering several physical processes operating in the hydrological cycle. These aforementioned studies collectively underscore the importance of high-frequency isotope datasets in advancing our knowledge of atmospheric and hydrological processes and their ensuing application in the improvement of climate models. Breitenbach et al.33 demonstrated that the temporal variations of δ18O are influenced by moisture transport during ISM and vapour re-equilibrium with rain droplets. Jeelani et al.34 demonstrated that isotope values on a daily scale record a sharp transition from western disturbances to monsoon circulation from the southern foothills of the Himalayas. The transition is so sharp that the onset of monsoon can be marked by rain isotope time-series data. Sengupta et al.26 demonstrated the rain δ18O variation within active phases of ISM depends on how fast the convective cloud bands propagate and how much they precipitate over BoB. Oza et al.35 estimated the raindrop evaporation fraction from rain isotope values from certain strategic locations and showed this fraction varies from 8% to 52% over their study locations. Rahul et al.36 quantified the average fraction of recycled moisture as 13.4 ± 7.7% across latitudes over the Southern Indian Ocean. All these aforementioned studies suggest that daily/event-based rain isotope datasets are useful.

Towards this, the current manuscript presents daily datasets of rain isotopes of oxygen (δ18O) and hydrogen (δ2H) collected from 2019 to 2021 for three strategic geomorphic locations in India: Port Blair, Mahabaleshwar, and Tezpur. These three locations have unique climatic and geographical settings (Fig. 1), which may provide a comprehensive insight into variations in moisture source dynamics, rain mechanisms, and their ensuing control on rain isotope variability during ISM.

Methods

Climate of the study area

All three locations receive the majority of the rainfall during ISM. The ISM is defined as a part of an annually reversing wind system marked by large seasonal changes in rainfall amount37. Due to different onset and withdrawal dates, the duration of ISM varies across the country and the study locations as well. Port Blair receives ISM rainfall earliest among all locations (normal onset date: 22nd May), followed by Tezpur (normal onset date: 3rd June) and Mahabaleshwar (normal onset date: 10th June). Similarly, withdrawal is early in Mahabaleshwar, followed by Tezpur and Port Blair38. For better comparison, the isotope data collected during ISM for the synchronous period are only included in the present work, although some of these locations also experience sizable rainfall during other seasons. The location-specific climate information is provided in Table 1 and briefly discussed below.

Table 1 Geographical, climatological and sampling summary of the study locations.
Full size table

Port Blair (11.63°N, 92.70°E), situated on the eastern coast of South Andaman Island over the BoB, is the capital of the union territory Andaman and Nicobar Islands. The location experiences a tropical climate and warm temperatures throughout the year. Average ISM precipitation over the area is around 712 mm, with temperatures varying from 26.7 °C to 30.2 °C during the season39. Mahabaleshwar (17.92°N, 73.65°E), situated in the Western Ghat mountain (at an altitude of 1,353 meters above mean sea level), is a prototype of an Indian tropical rainforest. It receives an average rainfall of ~ 5620 mm during ISM, and temperature varies from 16.3 °C to 23.8 °C39.

Tezpur, located in the northeastern plain of Assam, India (26.64°N, 92.79°E), is situated at the bank of the Brahmaputra River and 63 km west of Kaziranga National Park. The river and the forest may be two potential sources of moisture for the rainfall over the location35. Mean rainfall and temperature (recorded in this station) are 1167 mm and 28.3 °C, respectively, during ISM39. Apart from the unique geography and climate, the locations also receive rainfall from various moisture sources and experience different rain mechanisms. Tezpur receives moisture from BoB, Mahabaleshwar from the Arabian Sea (AS) and Port Blair from multiple ocean sources. In east India, low-pressure systems over BoB provide moisture for the ISM rains37. In contrast, West India, particularly Mahabaleshwar, experiences rainfall due to orographic interception of the Arabian Sea branch of the ISM40. These differences make our datasets important for understanding the role of these processes on rainwater isotopes.

Rainwater sampling and isotope measurement

Rain samples were collected on a daily basis at 8:30 A.M (Indian Standard Time) using rain samplers made as per the guidelines of the International Atomic Energy Agency (http://www.naweb.iaea.org/napc/ih/documents/other/gnip_manual_v2.02_en_hq.pdf). Samples were collected for all rainy days (rain rate ≥ 0.1 mm/day). A 2-L carboy bottle fitted with a 20.3 cm diameter funnel was fixed ~1 meter above the ground level to avoid splashing. A tube was attached to the tip of the funnel that touched the bottom of the plastic bottle, which helped reduce the exposed surface area and minimise evaporation41. The rain samples collected in the sampler are transferred into 8 ml Polylab bottles. The name of the location, date and time of collection were labelled on the sample bottles. Samples are shipped to the Indian Institute of Tropical Meteorology (IITM), Pune, almost every 30 days interval. During the storage period, the samples were kept in dark and cold places. The samples were measured immediately after receiving them at IITM.

All rainwater samples are measured in a Laser based Liquid Water Isotope Analyzer (LWIA; Model No: TIWA-45-EP), manufactured by Los Gatos Research (LGR), formerly and now ABB Inc., with a routine analytical precision of 0.1‰ and 1‰ for δ18O and δ2H respectively using five laboratory standards (Table 2). These laboratory standards are periodically calibrated with respect to IAEA primary standards. The measurement procedure in this instrument is already available in various earlier literature41,42 and is briefly discussed below. The LWIA measures both δ18O and δ2H of liquid samples simultaneously, with every measurement taking ~ 1.5 minutes. The 1 mL aliquots of all water samples (both rain and standards) are transferred into 2 mL glass vials capped with pre-sealed silicone septa. During measurement, 1.2 μL of liquid is further sampled by an autosampler and injected into an injector block (preheated at ~70 °C). Water vapour produced in the injector block is then transferred to the isotope analyser through a Teflon tube. Between two consecutive injections, dry air is passed through the cavity, and the entire cavity is pumped to eliminate the trace of the sample memory from previous measurements. In high throughput mode, every sample measures nine injections, while the first four injections are discarded to eliminate the memory effect. The data measured in the analyser are further calibrated with respect to Vienna Standard Mean Ocean Water (VSMOW) using post-processing software.

Table 2 The laboratory water standards used in the measurements, along with their mean values and precision.
Full size table

The rainfall (mm/day) data for Mahabaleshwar and Port Blair were supplied by National Data Centre43, India Meteorological Department on request using the following link: https://dsp.imdpune.gov.in/. The rainwater sampling was carried out at the IMD observatory in Mahabaleshwar. For Port Blair, the IMD observatory is located 4 km away from the sampling station. The rainfall amount is measured at the sampling location in Tezpur. Due to logistic issues, samples could not be collected in early July 2020 in Mahabaleshwar. Other meteorological data from IND can also be obtained using the link mentioned earlier.

Fig. 1
figure 1

Map showing the sample locations considered for the current study. Topography (in contours) and land use land cover (LULC; in shades) are shown on the map. Sampling locations are marked- Mahabaleshwar (a tropical hill station), Tezpur (close to North East Indian forest) and Port Blair (a tropical island). GTOPO25 and Sentinel-2 10 m Land Use/Land Cover 2021 data are used for topography and LULC respectively. Data/Map source: GTOPO25- https://earthexplorer.usgs.gov/ & Sentinel-2 - https://www.arcgis.com/home/item.html?id=fc92d38533d440078f17678ebc20e8e2.

Full size image

Data Records

The datasets are available at figshare.com44, which includes rain isotope data at three locations from 2019 to 2021. The data are available in four coloumns- date of collection (in dd-mm-yyyy format), Rain δ18O (‰ VSMOW), Rain δ2H (‰ VSMOW), and rainfall (mm/day). Figure 2 displays the temporal variations in isotope values (δ2H) and rainfall amount during the ISM from 2019 to 2021. Local Meteoric Water Lines (LMWL) constructed on the basis of δ²H - δ¹⁸O relationships (considering all three-year data together) are shown in Fig. 3, and the values of their slopes, intercepts (along with their uncertainties) and correlation coefficients are presented in Table 3. In Port Blair, the rain isotope values vary from −10.2‰ to 0.6‰ for δ18O (Fig. 3a) and −75‰ to 8‰ for δ2H respectively (considering a total of three years of measurement; Fig. 2a). In Mahabaleshwar, rain δ2H (δ18O) values range from −92‰ to 11‰ (Fig. 2b) (−13.4‰ to 0.4‰; Fig. 3b). The station Tezpur, records maximum variation of both δ18O (−17.9‰ to 0.7‰; Fig. 3c) and δ2H (−130‰ to 16‰; Fig. 2c) values among all three locations during our study period. It is evident from Fig. 3 that the intercept value (of LMWL) of Port Blair samples is significantly lower than that of the other two locations; however, no notable difference exists among their slopes. The intercept values range from 4.3 ± 0.13‰ (Port Blair in 2019) to 12.0 ± 0.13‰ (Tezpur in 2021; Table 3).

Fig. 2
figure 2

Rain δ2H (in ‰) and rainfall amount (in mm/day) time series reported in this study for the locations – (a) Port Blair, (b) Mahabaleshwar, and (c) Tezpur for the study period (summer monsoon of 2019–2021). The x-axis represents the date of sample collection. For better clarity, only months are marked in the figure (abbreviations used for months: Jun-June, Jul-July, Aug-August, Sep-September, and Oct-October). See the figshare file, mentioned in the text, for the details of sampling dates.

Full size image
Fig. 3
figure 3

Scatter plots showing the relationship between δ¹⁸O and δ²H of the rain samples collected from (a) Port Blair, (b) Mahabaleshwar and (c) Tezpur for the entire study period (summer monsoon of 2019–2021). Mean regression lines passing through the data points are also shown in black solid lines. The equations of those lines, correlation coefficients (R2) and sample numbers (N) are also mentioned in the respective plots.

Full size image
Table 3 Slopes and intercepts of mean regression lines obtained from δ2H-δ18O relationship of rain samples collected at the study locations each year (summer monsoon period) from 2019 to 2021. Correlation coefficients (R2) and numbers of samples are also mentioned.
Full size table

Technical Validation

All laboratory analyses were conducted using Off-Axis Integrated Cavity Output Spectroscopy technique (Los Gatos Research, San Jose, CA, USA). These analyzers can deliver accurate and precise results as long as volatile organic compounds do not cause spectral interferences. Such interferences are typically absent in rainwater samples and can be easily identified using instrument software. The datasets (both isotope and rainfall) presented here are quality-controlled. The raw isotope data measured by the instrument are calibrated with respect to laboratory standards (which are also periodically calibrated with respect to IAEA primary standards), and the calibrated values are reported with respect to an international standard VSMOW.

Usage Notes

Users can freely access the daily rain isotope data of oxygen and hydrogen provided in this paper. Researchers who wish to reproduce/use the current rainfall dataset for Mahabaleshwar and Port Blair locations or perform similar studies should approach IMD directly using the following link: https://dsp.imdpune.gov.in/.