Isotopic interaction and source moisture control on the isotopic composition of rainfall over the Bay of Bengal

Atmospheric Research 235 (2020) 104760

Contents lists available at ScienceDirect

Atmospheric Research journal homepage: www.elsevier.com/locate/atmosres

Isotopic interaction and source moisture control on the isotopic composition of rainfall over the Bay of Bengal

T

⁎

Nitesh Sinhaa,b,1, , S. Chakrabortya,b a b

Center for Climate Change Center, Indian Institute of Tropical Meteorology, Pashan, Pune, Maharashtra 411008, India Department of Atmospheric and Space Science, Savitribai Phule Pune University, Pune, Maharashtra 411007, India

A R T I C LE I N FO

A B S T R A C T

Keywords: Rain isotopes Rain-vapor interaction Bay of Bengal

The isotopic interaction between ambient vapor and raindrops has been investigated for the first time at Port Blair, Andaman Islands, an environment having the minimal contribution of the continental moisture. Rainwater and ambient vapor samples were collected on a daily timescale during the Indian summer monsoon season of 2015. Oxygen (Hydrogen) heavy isotopic ratios of rainwater and ambient vapor are positively correlated and a difference in the seasonal average values was found to be 9.5‰ (65.8‰) at ~28 °C. On a daily scale, the oxygen isotope ratios of ambient vapor and rainwater are significantly correlated over a wide range of rainfall amount, but correlation weakens for rainfall exceeding ~36 mm/day. The isotopic variability of rainfall appears to be modulated by the interaction with the ambient vapor, which in turn is determined by the source moisture. In a case study, it is estimated that the vapor-rainwater isotopic exchange could lead to 18O enrichment (depletion) in raindrops (vapor) about 0.41‰ (approx. 30% of the total change), which is far above (> 4σ) of the experimental uncertainty.

1. Introduction Stable isotopes in rainwater are affected by meteorological processes give a characteristic fingerprint of their origin and appear to depend on the meteorological conditions at the time of the condensation (Dansgaard, 1964; Rozanski et al., 1992; Gat, 1996; Clark and Fritz, 1997; Kendall and McDonnell, 1998). Thus, the isotopic analysis of rainwater provides a means to determine the contributions of land and ocean-derived moisture due to their distinct isotopic ratios. Furthermore, the isotopic ratios in rain and atmospheric vapor can be used to simulate the isotopic composition of moisture sources modified by the fractionation associated with the mechanism of condensation during its transport from ocean to a continental site (Dansgaard, 1964; Rozanski et al., 1993; Gat, 2000; Araguas-Aragua et al., 2000; Yoshimura, 2015). Isotopic fractionation is believed to follow the Rayleigh distillation formulation under the equilibrium condition (Yurtsever and Gat, 1981), whereby cumulative fractionation during condensation and subsequent removal of the condensate through rainfall make the remaining vapor depleted in heavier isotopes. Under natural conditions, the thermodynamic equilibrium between the liquid and vapor phase is not always established, for example, the evaporation of an open ocean into an unsaturated atmosphere. In this

case, small differences in transferring of isotopologues of water through a boundary layer at the water-air interface impart disproportional enrichments in the water phase governed by the so-called kinetic fractionation process. The model that describes the isotope effects accompanying evaporation into an open (unsaturated) atmosphere formulated by Craig and Gordon (1965), from now on referred to as C-G Model. According to the C-G Model and assuming that there is no other water vapor source than the sea surface, the isotopic ratio of the evaporating vapor is estimated by the following Eq. (1) (Merlivat and Jouzel, 1979):

Re =

Rl α (h s (1 − αk ) + αk )

(1)

where,

h s = Rh ×

q (T ) q (SST )

(2)

Re and Rl are isotopic ratios of evaporated vapor and ocean surface water, respectively. α and αk are the equilibrium and kinetic fractionation factors respectively. hs is the relative humidity normalized to saturation at the sea surface temperature (SST) and calculated using Eq. (2); where Rh is the relative humidity near-surface air and q(T), q(SST) are specific humidity at air temperature (T) and SST respectively.

⁎

Corresponding author at: Center for Climate Change Center, Indian Institute of Tropical Meteorology, Pashan, Pune, Maharashtra 411008, India. E-mail address: [email protected] (N. Sinha). 1 Present address: IBS Center for Climate Physics, Busandaehak-ro 63beon-gil 2, (Jangjeon-dong), Geumjeong-gu, Busan 46241, Republic of Korea. https://doi.org/10.1016/j.atmosres.2019.104760 Received 5 May 2019; Received in revised form 24 September 2019; Accepted 10 November 2019 Available online 11 November 2019 0169-8095/ © 2019 Elsevier B.V. All rights reserved.

Atmospheric Research 235 (2020) 104760

N. Sinha and S. Chakraborty

during the Indian summer monsoon (ISM) of 2015 at the Andaman Islands, situated in the Bay of Bengal (BoB). Ground-level vapor (GLV) and rainwater were collected on the daily timescale in Port Blair (11.66°N, 92.73°E), the Andaman Islands. Hereinafter, we refer to the vapor at the site as GLV and generalize it as ambient vapor. The Andaman Islands are known to receive heavy rainfall during the ISM season, and as a result, humidity level stays high throughout the sampling period (Chakraborty et al., 2016). Being an island, the significant moisture would be sourced presumably from the surrounding ocean. The island has a large forest cover (~92%), which may produce some amount of transpired vapor. However, the isotopic composition of that transpiration-induced rain during high humidity is believed to be nearly the same as that of the rainwater (Chakraborty et al., 2018). Hence, the isotope study of ambient vapor and rainwater in this site is likely to provide valuable insight in understanding their interaction vis-à-vis source and ambient vapor control on the variations in the isotopic composition of regional rainfall. Based on a robust oxygen and hydrogen isotopic datasets from these samples (GLV and rainwater), we discuss (i) the isotopic relationship between ambient vapor and rainwater, (ii) the role of source moisture on the isotopic compositions of ambient vapor and rainwater, and finally, (iii) we attempt to quantify the extent of isotopic fractionation arising primarily due to exchange between the raindrops and the ambient vapor over the Andaman Islands region.

There are various identified factors playing essential roles in the isotopic fractionation at different stages of the source moisture transforming to rainfall. Starting from the rise in the saturated air column undergoing depletion of the heavy isotopes in the vapor phase, and subsequently during the condensation phase, whereby the condensates get enriched in heavier isotopes (Dansgaard, 1964; Gat, 1996). Postcondensation processes, such as evaporation of raindrops and diffusive exchange of isotopes between the raindrops and the ambient vapors (Miyake et al., 1968; Stewart, 1975) also make the water phase enriched and ambient vapor depleted in heavy isotopes (Lawrence et al., 2004). Secondary effects, such as recycling of water vapor within the convective cells could reduce the rainwater heavy isotopic composition during the convective events (Lawrence et al., 2004; Risi et al., 2008). To interpret the isotopic variability of rainwater and to improve our understanding of their role as tracers of water movement in the atmosphere, the isotopic composition of vapor should also be considered (Dansgaard, 1964). All the factors as mentioned earlier, in turn, make the isotopic fractionations of water in the atmosphere a complex system. Under these circumstances, it is imperative to characterize the role of the various components of the hydrological cycle in modulating the isotopic composition of atmospheric vapor and rainfall. The isotopic variations on daily to seasonal timescales in vapor and rain phases can use to diagnose different moisture sources, especially the influences of the large monsoon systems and convection (Yu et al., 2015). Previous studies (e.g., Araguas-Aragua et al., 2000) have suggested that on a global scale, the monthly composite of the isotopic composition of rainfall closely reflected isotopic variability observed in near ground-level water vapor. Deshpande et al. (2010) show that the use of the isotopic composition of monthly rainfall for estimating the average monthly isotopic composition of ground-level water vapor may lead to erroneous result in low humid conditions. On the other hand, measurements over the open ocean show that the isotopic ratio of the vapor phase is controlled mainly by the ocean surface conditions, while lateral advection of vapor could alter their values significantly (Midhun et al., 2013). Dar and Ghosh (2017) estimated the source of moisture (land and sea moisture) precipitation as rain in Kolkata using a model based on isotopic compositions of source moisture, ambient vapor, and rainwater. Furthermore, recent studies on rain and vapor interactions have been used, to identify onset signals of Indian summer monsoon (Srivastava et al., 2015), to study the large-scale convective activity and low pressure/depression systems control on isotopic composition of rain and vapor (Saranya et al., 2018), and to decipher the isotopic distinctive characteristic of the convective and stratiform rainfall events (Lekshmy et al., 2018). However, despite these advances, one of the fundamental aspects, viz., the isotopic interaction between raindrops and the ambient vapor was not systematically studied. That is, a rigorous attempt of the isotopic fractionation of the falling raindrops when they interact with the ambient vapors is still missing, especially for the Indian subcontinent, which is strongly influenced by the monsoon system. Theoretical studies predicted that isotopic exchange between raindrops and water vapor could be significant, which necessitates quantification of this process causing isotopic fractionation in rainwater, preferably in a fieldbased observational framework. This information is also important to estimate the extent of isotopic fractionation in rainwater subjected to intense cyclonic activities (Lawrence and Gedzelman, 1996) and for a better understanding of the elusive nature of the amount effect (Dansgaard, 1964; Kurita et al., 2011; Conroy et al., 2016). A fieldbased observational strategy would require synchronous analysis of ambient vapor and rainwater, preferably in a site that experiences high humidity throughout the monsoon season. High humidity would help achieve near equilibration of rainwater and ambient vapor and impart a negligible effect of raindrop evaporation. Secondly, if the moisture is sourced primarily from an open ocean, but a negligible amount from evapotranspiration, then the exercise will be better constrained. With these considerations, a specific field campaign was arranged

2. Site, data, and methodology The Port Blair site is situated in the southern part of the Andaman Islands in the BoB and characterized by tropical wet climate. The site location indicated as a red dot in Fig. 1a, and vectors represent the mean wind pattern between June–September (ISM season) 2015. During the ISM season, strong southwesterly winds from the south BoB are the dominating moisture source over the Andaman Islands. It also corroborated by the back-trajectory analysis (Fig. S1), as the origin of air-masses is assumed to be approximate to the moisture source direction for the water vapor and rainfall at the study site (Draxler et al., 2010). However, a large amount of moisture over the site is expected to have come from the south BoB. As, the underlying ocean surface-water contribution to the convective systems are the major sources of moisture, particularly over the BoB region due to the stratification and higher SST (Shenoi et al., 2002). It can also be seen in the moisture parameters calculated for the BoB region during ISM season 2015 (Fig. S2). That is, due to the rainout process over the ocean surface in the course of the southwest (SW) monsoonal flow, the air-mass losses moisture and again get enriched in moisture from the underlying ocean water. In Fig. S2 (a), specific humidity is low in the air parcels over the south BoB, and the parcel gained moisture after entering the north BoB (above 8–9°N). Similarly, the spatial representations of relative humidity (Rh, %) along the SW flow for the mid of summer months show that there is a drop in Rh up to around 85°E and after that, it increases steadily (Fig. S2 (b)). A few studies also reported that the northeastern BoB contributes a significant amount of vapor to the summer monsoon rains (e.g., Achyuthan et al., 2013). Therefore, an area of 2° × 2° grid size (10°N-12°N, 90°E-92°E) in the southwest part of the Andaman Island has been considered to calculate Re using the C-G Model (red box, Fig. 1a). This moisture source area seems to offer a better constraint on the isotopic compositions of source moisture, ambient vapor, and rainwater (see Section 3.3). Seawater seasonal average oxygen (hydrogen) isotopic composition (Rl) for the ISM over the south BoB was determined to be −0.10‰ (+3.3‰) (Achyuthan et al., 2013) (Fig. 2, blue circle). The data used for the calculation of Re using the C-G Model (Eq. (1)) includes surface temperature, surface pressure, sea surface temperature, surface wind, specific and relative humidity from the Era-Interim reanalysis dataset (Dee et al., 2011). Although the global closure assumption [evaporated vapor (Re) ≈ pre-existing vapor (Ra)] may not be applicable on a 2

Atmospheric Research 235 (2020) 104760

N. Sinha and S. Chakraborty

Fig. 1. (a) shows a horizontal view of the moisture source region (red box), the site location (red circle) and mean wind pattern (vectors) for the Indian summer monsoon, 2015; (b) and (c) are vertically zoomed cross-section of location (Port Blair) and schematic representation of the conditions such as, ocean water evaporation, its advection towards the site, sampling position (three block building) for (b) rain-vapor interaction: off (non-rainy) and (c) rainvapor interaction: on; i.e. rainy conditions during vapor sampling hours. The figure is labeled with the arithmetic mean of the calculated and observed Where, oxygen isotopic values (δ18O). δ18O = −0.10‰ of ocean surface water (Achyuthan et al., 2013), and the subscripts are, e = evaporated moisture (C-G Model), v = ground level vapor (GLV), r = 24 h rain, r-v = δ18O of rain during vapor sampling hours, and v-r = δ18O of ambient vapor during rainy condition. In the lower panel (c), small left and right facing arrows represent an isotopic exchange between rain and vapor (rains were enriched and vapor got depleted supported by mean isotopic ratios). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

napc/ih/documents/other/gnip_manual_ v2.02_en_hq.pdf). The rainwater sampling kit involved a 2-l polycarbonate carboy mounted with a ~20.3 cm diameter funnel. The lower end of the funnel was connected with a PVC-tube and extended up to the bottom of the carboy. This arrangement helps reduce the exposed surface area of the collected water, thus, avoids evaporation. To validate the rain rate at the NIOT campus measured, the daily rainfall record of the Port Blair site was obtained from the India Meteorological Department (IMD) (http:// www.imd.gov.in); the rain gauge site of IMD in Port Blair was at an aerial distance of about 4 km from the NIOT campus. GLV samplings were done at the terrace of a three-storied building in the NIOT Campus from late May to the first week of October using the IAEA protocol (http://www-naweb.iaea.org/napc/ih/documents/ miba/water_vapor_protocol.pdf). The vapor inlet was installed on top of the building approx. 10 m above the ground. A cryogenic metallic trapping system maintained at ~-80 °C using a mixture of ethanol, and liquid nitrogen was used to collect the GLV. A diaphragm pump was

regional scale, the condition is assumed to be valid for an open ocean case (Merlivat and Jouzel, 1979; Uemura et al., 2008; Benetti et al., 2014). The Kalpana-1VHRR outgoing longwave radiation (OLR) data (Mahakur et al., 2013) have been used to characterize the convective activities. 2.1. Sampling and isotopic analysis Fig. 1b and c schematically show a horizontal cross-sectional view of the sampling site and illustrate the concept of rain-vapor interaction; the corresponding mean δ18O values labeled. During the summer monsoon months of 2015 (late May to early October), daily rainwater was collected at the National Institute of Ocean Technology (NIOT) campus, Port Blair. The rainwater collected in the NIOT campus was measured (volume) and converted to rain rate in a unit of mm/day using a calibration relation outlined in IAEA/GNIP precipitation sampling guide, v2.02 September 2014 (http://www-naweb.iaea.org/ 3

Atmospheric Research 235 (2020) 104760

N. Sinha and S. Chakraborty

hours (08:00–11:00) (n2 = 27), and the collected condensed GLV samples (n3 = 107) were stored in airtight bottles and sent to the Indian Institute of Tropical Meteorology (IITM), Pune, for isotopic analysis. The isotopic analyses (δ18O and δD, Section S1.2) of rainwater and condensed GLV samples were done using the LGR's Liquid Water Isotope Analyzer (Chakraborty et al., 2016). The overall analytical precision value of the measurements obtained for δ18O (δD) was about 0.1‰ (< 1‰) and isotopes data are provided in the supplementary material. 3. Results and discussion 3.1. δ18O and δD of rain and vapor Rain δ18Or (δDr) values range from +0.20‰ (+7.59‰) to −8.99‰ (−65.68‰) with an average value of −2.86 ± 1.91‰ (−13.24 ± 15.45‰). The Local Meteoric Water Line (LMWL) is given as δDr = (7.70 ± 0.23) × δ18Or + (8.82 ± 0.79) (Fig. 2, cyan). The observed slope and intercept of the LMWL for the Port Blair site nearly the same ( ± 1sigma) as the Global Meteoric Water Line (GMWL), i.e., δD = (8.17 ± 0.07)*δ18O + (10.35 ± 0.65) (Rozanski et al. 1993). It indicates that the contribution of secondary moisture (i.e., raindrop evaporation) or evapo-transpired moisture source is negligibly small (Kumar et al., 2010). However, slightly low values, especially the slope may arise due to post-condensation effects, such as isotopic exchange and/or raindrop re-evaporation (Araguas-Aragua et al., 2000). The process of raindrop evaporation and isotopic exchange between raindrops and the ambient vapor both make the rainwater enriched in heavier isotopes, but during the high humidity condition the exchange process prevails over the raindrop evaporation (Friedman et al., 1962; Miyake et al., 1968). Our results show that δ18Or and deuterium excess (d-excess = δD- 8 × δ18O) correlation is very weak, R2 = 0.015 (Fig. S3), implying insignificant raindrop re-evaporation (Gat, 1996). It can be elaborated as: the evaporation from the falling rain results in the enrichment of the heavy isotopic species in the remnant drop along the so-called evaporation line; since the partially evaporated raindrops result in an LMWL usually having a slope < 8 in δD - δ18O space, the resulting precipitation shows a smaller d-value than at the cloud base (Gat, 1996). Hence, these two parameters (δ18O and d-excess) should show strong inverse correlation for significant raindrop evaporation. Hence, the observed isotopic variations in this site are likely to be predominantly driven by the isotopic exchange between raindrops and ambient vapor for the ISM season of 2015. In the case of GLV, the range of isotopic values, i.e., δ18Ov (δDv), varying from −9.78‰ (−64.06‰) to −17.39‰ (−113.98‰) with an average value of −12.48 ± 1.59‰ (−79.12 ± 10.23‰). The ‘vapor line’ in δ18O-δD space defined by the equation δDv = (5.69 ± 0.29) × δ18Ov -(8.43 ± 3.67) shows a strong linear correlation (r = +0.90) (Fig. 2, red). The difference between the isotopic compositions of rainwater and GLV, i.e., Δ(δr – δv) is 9.5 ± 3.5‰ for 18O and 65.8 ± 25.6‰ for D (or 2H). The observed values of Δ(δr – δv) for 18O and 2H are close to the calculated equilibrium enrichment factor (ε) at mean air-temperature of 28.4 °C; i.e., 18ε = 9.10‰ and 2 ε = 75.6‰. The calculated isotopic ratios of vapor from seawater

Fig. 2. δ18O and δD values measured Port Blair ground level vapor (red), rainwater (cyan). Isotopic compositions of vapor over the chosen source region (purple) considering kinetic fractionation (C-G Model) using mean sea waterδ18O and δD (blue, Hema et al. 2013). The equations represent the vapor and rain lines and r indicates the correlation coefficient for δ18O and δD. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

used to suck the ambient air through an inverted conical vent (at terrace ~6 m from sampling room) and passes through a PVC tube to a pre-cooled trap. The airflow rate was controlled at 800 ml/min to ensure adequate amounts (> 2 ml (ml)) of vapor was collected in three hours (hrs) between 08:00–11:00 h. To ensure the quantitative transfer of the trapped vapor, a secondary trap was used at the outlet of the primary trap, which was found to have collected a negligible amount of vapor for each collection. Temperature and relative humidity were also monitored throughout the vapor-sampling period at the site. The average air temperature was recorded as 28.4 ± 1.2 °C (Max = 31.2 °C, Min = 25.7 °C) and relative humidity 80 ± 5% (Max = 94%, Min = 65%) over the sampling site. The average amount of collected condensed vapor (GLV) was 2.9 ml, ranging from 1.4 ml to 4.5 ml in 3 h. After the end of sampling, the trap was isolated from the ambient environment and then allowed to equilibrate to room temperature. A separate arrangement was made to collect rainwater samples synchronized with the vapor sampling time to study their isotopic interaction (Section S1.1). In the text, vapor collection during ‘nonrainy’ case implies there was no rain during the vapor sampling hours (i.e., 08:00–11:00); while ‘rainy’ period means it was raining at the site during the vapor sampling hours. These two cases represent rain-vapor interaction was ‘off’ and ‘on’ respectively (Fig. 1b–c). Sampling strategy, the number of the respective samples collected, and notations used are given in Table 1. We have also collected daily rainwater samples (08:30 to 08:30 next day, 24 h); this time coincides with the IMD observation protocol. A total of 110 daily rainwater (24 h) samples were collected. Additionally, ~3 h rainwater samples were also collected during the vapor sampling

Table 1 Characterization of rainwater and ambient vapor samples, corresponding notations used for oxygen isotopic values and number of samples collected with sampling hours are reported. S. No.

Sampling strategy

Oxygen isotopic value: vapor (rain)

Sampling on a daily basis Rain and vapor for same dates Rain during vapor sampling⁎

δ Ov (δ Or) δ18O'v(δ18O'r) δ18Ov-r (δ18Or-v)

Ambient vapor

Rainwater

No. of samples collected (duration of sample collection hours) 1 2 3 ⁎

18

18

107 (3) 85 (3) 27 (3)

Rain samples were collected in a separate sampler. 4

110 (24) 85 (24)

Atmospheric Research 235 (2020) 104760

N. Sinha and S. Chakraborty

δv) for δ18O and δD respectively. Therefore, it is important to consider the non-rainy and rainy conditions separately, during the vapor sampling hours to investigate the rain and vapor interactions. Hence, the rain-vapor isotopic interaction has been investigated in the present study by way of rain and vapor samples collected simultaneously, although, the coexisting rain events during vapor sampling hours were relatively less. 3.2. Rain and vapor isotopic characteristics The rainwater and GLV samples have been categorized into two groups. The rainwater samples integrated for 24 h (termed as ‘daily scale’) and rainwater samples integrated for 3 h (synchronous with vapor sampling hours) belong to the first group. The second group consists of GLV samples (collection time 08:00–11:00) collected during the rainy and non-rainy conditions. Categorization of rain/vapor are based on sampling strategy, that is daily 3 h vapor collection and 24 h rain collection. “v” and “r” denote vapor for 3 h and integrated rain for 24 h respectively. Vapor samples were typically collected daily around 08:00–11:00 local time irrespective of the rain event. “v-r” is used to define the environmental condition when it was raining, and the concurrent vapor was collected (for 3 h). The isotopic values of such vapor were denoted as δv-r. In addition to 24 h rainwater collection, 3 h rain collection was done in synchronization with the vapor collection (08:00–11:00) and the notation “r-v” denotes such situation and corresponding isotopic values denoted as δr-v (Table 1). Thus, the first group comprises δr and δr-v values and a second group of δv and δv-r values. A comparative analysis of their isotopic characteristics is presented in this section. Fig. 4 shows the scatter plots of δ18O and δD and their corresponding regression lines for various combinations. Isotopic compositions of rainwater during vapor sampling hours show a slope, s1 = 6.90 ± 0.43 (green) that is lower than that of the daily scale rainwater slope (s2 = 7.67 ± 0.27, red) (Fig. 4a). The meteoric water line with a lower slope (s1) implies rainwater during the vapor sampling hours may have been isotopically enriched as a result of the isotopic exchange between raindrops and the ambient vapor. This is to be noted that the contribution of evapo-transpired moisture is negligible and raindrop re-evaporation process is proven to be insignificant over the site during ISM 2015. However, when the δ18O-δD relationship is examined on a daily timescale, the slope (s2 = 7.67 ± 0.27) does not differ significantly from 8 (i.e., the slope of GMWL). In addition, the isotopic compositions of rainwater collected at another site (namely Pondicherry University) in Port Blair show similar slope value (7.68 ± 0.22) for summer 2015. The site was at a few kilometers apart from the NIOT campus site (note: other results from Pondicherry University site will be discussed elsewhere). This observation indicates that the isotopic composition of rainfall on the daily scale did not show the effect of isotopic interaction. The reason for this distinct behavior due to different sampling timescales may be explained as follows. The synchronized rainwater and vapor collection (3 h) period experienced mostly rainy and high humid condition throughout this period. Whereas, the 24 h rain collection period was punctuated by no-rain condition perhaps for several hours. Hence, the isotopic interaction during the 24 h rain collection period was much less compared to the 3 h synchronized collection of vapor and rainwater. On the other hand, the slope of the vapor line for which samples were collected during the rain on the site shows a higher value (s3 = 6.33 ± 0.60, green) than that of the vapor line of the non-rainy condition (s4 = 5.61 ± 0.33, red) (Fig. 4b). The difference in slopes can be explained as, during rainfall and humid condition net diffusion from the boundary of raindrops to the atmosphere would dominate than the raindrop re-evaporation. Hence, the more diffusive HH16O species would escape from the raindrops faster than the heavy isotope species of HH18O and HD16O. The opposite phenomenon takes place in the vapor phase. We also calculated the changes in slope for the

Fig. 3. δD of vapor versus specific humidity (q, g/Kg) diagram, with theoretical curves of evaporation mixing (dashed black curve) with vapor in equilibrium with the ocean surface between dry end-member and initial point (chosen as 80% relative humidity over an ocean at 300 K), Rayleigh distillation (solid black curve) assuming vapor derived from an ocean source with 80% relative humidity, and rain evaporation (solid gray curve) or vapor that experienced additional isotopic change due to raindrops re-evaporation (Worden et al., 2007; Noone, 2012). Red dots represent δDv of ground-level vapor and q at the Port Blair site for summer 2015. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

using Eq. (1) (Fig. 2, purple circles) are also near to δ18Ov and δDv values at equilibrium given air-temperature. Thus, our observations suggest that GLV and rainwater maintained a near isotopic equilibrium. Further, it is known that the specific humidity (q) and δ values at the observation site in relation to theoretical ocean water mixing, Rayleigh distillation, and rain evaporation curves (Worden et al., 2007; Noone, 2012) could provide guidelines in interpreting isotopic values and provide a reliable means to delineate the moisture source (Conroy et al., 2016). Measured δDv versus q offers a conceptual basis to interpret the results as outlined in Conroy et al., (2016). In our case, Fig. 3 has been plotted following Worden et al. (2007) and Noone (2012), dashed line showing an evaporation mixing curve between a dry endmember (q = 0.56 g/kg, δ = −360‰) and vapor in equilibrium with an oceanic source at 300 K. The solid black curve shows liquid precipitation following Rayleigh distillation that assumes an original vapor derived from an oceanic source with 80% relative humidity. The solid gray line denotes a similar evolution of a vapor mass following “super Rayleigh” distillation or a vapor mass that has experienced additional isotopic exchange due to rain evaporation. The most of δDv values fall on or close to the ocean evaporation mixing line, supporting the above observations that these δDv values are primarily derived from ocean evaporation. On the other hand, lowest δDv values measured at Port Blair are above the rain evaporation line (except a few, discussed in Section 3.2), pointing to rain evaporation as an insignificant process during the measurement period. Besides, it is to note that the monsoon season also embedded by low rainfall and decreased humidity. That is for the reduced rainfall condition as well as relatively higher temperature (T) and lower humidity (RH) scenario (see Section 2.1 for variations in T and RH), the evaporation of raindrops or the land surface generated moisture may dominate. During this condition, the isotopic equilibrium between raindrops and the ambient vapor will be weakened resulting in some discrepancy in observed and theoretically calculated values (e.g., 2ε(δDr– δDv) = 75.6–65.8 = 9.8‰). It may justify a little higher amount of the uncertainty levels of ± 3.5‰ and ± 25.6‰ in the difference between the isotopic compositions of rainwater and GLV, i.e. ± Δ(δr – 5

Atmospheric Research 235 (2020) 104760

N. Sinha and S. Chakraborty

Fig. 4. Linear regression plots between δ18O and δD of rain and vapor for the two sets. (a) rain isotopic composition of the 24 h integrated rain (daily scale) and 3 h rain (during vapor-sampling hours), (b) vapor isotopic composition for the vapor collected during non-rainy and rainy conditions on site (08:00–11:00 h); (c) and (d) show the cross-correlation between δ18O of 24 h rain and 3 h vapor (green hexagon) as well as same for δD (red hexagon) respectively; (e) and (f) show the crosscorrelation between δ18O ratios of 3 h rain (during vapor sampling) and vapor while rainy during vapor sampling hours (green stars) as well as same for δD (red stars) respectively. In addition, (g) and (h) show the cross-correlation between δ18O of rain and vapor (green diamond) as well as same for δD (red diamond) respectively after segregation of dataset, those deviate from the expected ratios of rain and vapor given boundary layer temperature (Section S1.3). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) 6

Atmospheric Research 235 (2020) 104760

N. Sinha and S. Chakraborty

Fig. 5. (a) Oxygen isotopic ratios (δ18O) in ambient vapor during non-rainy condition (open black circle) and superimposed with the δ18O of ambient vapor during rainy condition (filled red circle); (b) δ18Or in rain (filled black circle) superimposed with the δ18O of ambient vapor during rainy condition (filled red circle). The thick gray line represents the three-point average of daily rain δ18O and shows a significant correlation (r) with vapor δ18O during the rainy condition. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

i.e. δDr-v vs. δDv-r, the correlation was r = 0.77 (n = 27, p < 0.01). Interestingly, coefficients of the regression equation for both non-rainy and rainy conditions are nearly the same, equations are shown in Fig. 4c and d & Fig. 4e and f respectively. Though the different number of samples in these cases may give rise to sampling artifact, however, the agreement observed in both the cases (hydrogen and oxygen) almost rules out this possibility. A close correlation also manifests the controls of pre-existing boundary layer moisture on rain isotopes. Thus, the isotopic time-series of GLV in equilibrium with rain could be used to theoretically explain the isotopic variation of rainwater for the entire monsoon season. One could expect a strong linear correlation between simultaneously collected boundary layer vapor and rain isotopic ratios (Kurita, 2013; Moore et al., 2014; Conroy et al., 2016). The observed scattering in rain-vapor isotope plots (Fig. 4a and b) in our case could give a window, which may indicate whether precipitation is in equilibrium with boundary layer vapor at boundary layer temperatures or if other processes contribute to the site's isotopic values. Therefore, an exercise has been performed to distinguish the sample data (rain and vapor measured) based on of their deviations from the expected (calculated) rain to vapor isotope ratios for a given boundary layer temperature. Fig. S4 (a) shows the measured and calculated ratios of rain to vapor oxygen isotopic values. We observe a few points deviate from the expected values of rain to vapor oxygen isotopic ratios. Thus, the segregation of the dataset those have deviated from the expected values were done, and the time series has been re-plotted (Fig. S4 (b)). A total of 17 points

waterline as well as for the vapor line as a result of isotopic distribution between ambient vapor and rainwater. The slope differences ‘s2-s1’ and ‘s4-s3’ were found more than 1σ and nearly the same for each case, i.e., 0.77 and 0.72 respectively. It clearly implies that the isotopic interaction between rain and vapor plays a dominant role, while other factors have minor control. Out of 107 GLV and 110 rainwater samples collected, 85 samples were of common dates, corresponding isotopic values are denoted as δ18O’ and δD’(Table 1). Fig. 4c and d show linear relationships between δ18O'r and δ18O'v (r = + 0.55, n = 85, p < 0.01) and δD'r and δD'v (r = + 0.65, n = 85, p < 0.01) respectively. The result implies a significant co-variation in the isotopic composition of vapor and rain phases. We have also determined the isotopic values of vapor considering isotopic equilibrium (δ18Oeq) using δ18O'r and ambient airtemperature (Majoube, 1971). Comparison of the theoretically calculated values (δ18Oeq) and the observed values δ18O'v (r = +0.57, n = 85, p < 0.01) shows that rainwater and ambient vapor are indeed consistent on the seasonal timescale. Such kind of agreement was also observed by Lekshmy et al. (2018) in the coastal region of Kerala, southwest India. Both of these exercises imply that GLV and rainwater were close to isotopic equilibrium over the site. Further, we investigated the rain-vapor interaction for the 3 h sampling interval while raining on-site, and both GLV and rainwater (in a separate rain sampler) were synchronously collected (Fig. 4e and f). The correlation between δ18Or-v and δ18Ov-r was found to be reasonably strong (r = 0.66, n = 27, p < .01). The same for the hydrogen isotopes 7

Atmospheric Research 235 (2020) 104760

N. Sinha and S. Chakraborty

of very deep convection (OLR < 180Wm−2) do not show any concurrency between the δ18O ratios of source moisture - ambient vapor rainwater, few are marked with a small circle over the rain bars (Fig. 6d, listed in Table 2). So far, we have discussed the isotopic values of source moisture, ambient vapor and rainwater only qualitatively, however, a quantitative understanding is warranted under the different atmospheric conditions (e.g. rainy versus non-rainy condition). It will give a better understanding of atmospheric controls on isotopic composition and the role of source moisture on the isotopic composition of rainfall. Henceforth, we separated the calculated as well as measured isotopic compositions (namely, δ18O) depending upon heavy rain, moderate rain, light rain, and no rain conditions during the ISM of 2015. To characterize based on the amount of rainfall, we followed the IMD criteria of heavy to light rain (http://imd.gov.in/section/nhac/ termglossary.pdf) and reported in Table 2. At Port Blair, the oxygen isotopic ratios of GLV (δ18Ov) and rainwater (δ18Or) are significantly correlated over a wide range of rainfall amount, i.e., moderate to very-light rain (Table 1, row I). The high correlation observed between δ18Ov and δ18Or is most likely arising due to isotopic interaction leading to equilibrium condition between vapor and rainwater. δ18O variations in source moisture, GLV, and rainwater also show a strong correlation with each other for the light rainfall condition (Table 2, column(c)), thus, implies an equilibrium condition between “source moisture - ambient vapor- rainwater” with a minimum contribution of secondary moisture from the other sources. On the other hand, the low correlation between source moisture and GLV oxygen isotope values during the no-rain condition (Table 1, column (e)) implies the large fraction of secondary moisture generated locally (evapotranspiration) may be due to high air temperature and low humid conditions. The lack of an agreement between the calculated and measured isotopic ratios during the heavy rainfall (Table 2, column (a)) points to the limitation of C-G Model and/or known physical mechanisms of isotopic fractionation. Earlier studies have shown that the intense rainfall includes the number of larger raindrops (Marshall and Palmer, 1948) and it takes longer for a large raindrop to reach an isotopic equilibrium (Lee and Fung, 2008). In our case, the correlation between rainfall intensity and the offset between the isotope values of vapor-rain was found to be weak (r = +0.26, p = 0.03). However, one of the cases that have not been addressed earlier and needs further investigation is the effect of a high volume of rainfall with moderate intensity on a daily scale. To elaborate further, it may so happen that a high-intensity rain for a few hours may produce a lower amount of rain than a low-intensity rain for 24 h. In conclusion, even assuming the local moisture contribution constant, it can be seen that the isotopic variation of ambient vapor, in turn, the isotopic composition of rainwater also depend on the source moisture. Moreover, for the periods of moderate to very-light rainfall amount, the crucial controlling factor for the variation in the isotopic composition over the region is rain-vapor interaction. We found a high correlation between δ18Or and δ18Oe during light rainfall and also after combining the range of moderate to very-light rainfall conditions (r = +0.48, n = 66, p < 0.0001). Similarly, for δ18Oe and δ18Ov, the correlation coefficient was found to be +0.54, n = 66, p < 0.0001. A significant positive correlation between evaporated vapor (C-G Model) and rainwater isotopic ratios is an indication of source moisture control on the isotopic composition of rainfall over the site. It also supports that the significant contributions of moisture were from the selected region in the BoB (box, Fig. 1a). Considering non-rainy and rainy conditions during vapor sampling hours, measured GLV oxygen isotopic ratios are plotted and shown by open and filled circles, respectively in Fig. 6b. Theoretically, the rainwater should be enriched while vapors are depleted in heavy isotopes due to the isotopic exchange process. Though a quantitative evaluation of the isotopic exchange process may not be possible for the individual event, however, we can address the statistics. Towards this, the

out of 85 are found to be deviated (see supplementary section S1.3 for details), which are above or below the expected ratios of rain to vapor at a given boundary layer temperature. This leads to a significant improvement in the correlations (r = +0.84, n = 68, p < .01) between rain-vapor isotopic compositions by considering only the remaining 68 data points (Fig. 4g and h). This exercise provides a conceptual basis of the factors in rain-vapor interaction in the boundary layer. Such as, when δ18Or/δ18Ov exceeds the calculated ratios for boundary layer temperatures, it may indicate additional fractionation due to rain evaporation. On the other hand, when δ18Or/δ18Ovvalues are lower than expected values, it may signify a change in the source of boundary layer moisture. Moreover, the rain falling from higher clouds in the atmosphere would have a δ18Or value that would appear more negative relative to near-surface δ18Ov, producing lower values. Furthermore, we found that the patterns of variation in δ18Ov-r and δ18Ov are quite similar and δ18Ov-r simulates the mean variations in δ18Ov (Fig. 5a). It is also corroborated in their respective regression lines as discussed above. The calculation shows that the δ18Ov-r variations are well within the domain of ± 1σ variation of δ18Ov for the whole season. Additionally, on the one hand, δ18Ov-r superimposed over δ18Or time-series illustrates an analogous variation (Fig. 5b). Thereby, the three-point average of daily δ18Or (thick gray curve, Fig. 5b) shows a significant correlation with δ18Ov-r (r = +0.64, n = 26, p < 0.01). It implies that the pre-existing ambient vapor modulates the rainwater isotope values on longer timescales. Our results corroborate the study of Chakraborty et al. (2016), which shows that the resultant moisture controls the isotopic values of the subsequent rain events. On the other hand, δ18Or-v values have not shown any significant relationship with the 24 h rainwater isotopic compositions (δ18Or). The three-point average of daily δ18Or shows a weak correlation with δ18Or-v (r = +0.42, n = 26, p = 0.03), whereas it follows a nearly similar trend (figure not shown). For the Port Blair site, it can now be elucidated that the ambient vapor and rainwater isotopic composition have nearly synchronous variations. However, one crucial factor is the isotopic composition of the source moisture, which could be one of the causes for the isotopic variations in ambient vapor and, in turn, of the rainfall. Since consideration of the daily variation in the isotopic composition of source moisture is likely to give a better understanding of the rain-vapor isotopic interaction over a specific site (viz. coastal regions), the subsequent section is focused on the controls of the source moisture in the isotopic composition of ambient vapor and regional rainfall. 3.3. Role of source moisture The meteorological parameters (namely, Rain and OLR) and isotopic variations in source moisture, ambient vapor, and rainwater over the site (Fig. 6) and their comparative evaluation on daily timescale are presented. The source moisture oxygen isotopes (i.e., Re calculated using C-G Model considered as source moisture over the site) show variations on daily timescale with a declining trend in isotopic composition between the period of June to September (−9.82‰ to −11.16‰, Avg. = −10.27‰), besides, several deep depletions were observed with standard deviation (σ) of 0.38‰. The depleted δ18O values ( ± 1σ of the average) of source moisture correspond to lower hs (Re and hs correlation, R2 = 0.86, n = 184, p < 0.01), this situation concordance with less or no rainfall conditions over the site (Black bars, Fig. 6d), also supported by high OLR (> 200Wm−2) values (dashed line, Fig. 6d). The general trend (arrows) and the deep depletion in δ18O values (filled gray bars) of the source moisture (Fig. 6a) are in concordance with the GLV (Fig. 6b). The GLV and the rainwater oxygen isotopes have shown analogous variations during the deep convection (dotted gray bars) and OLR between ~180-200Wm−2. For the periods of daily rain condition (active period of ISM), all (δ18O of source moisture -ambient vapor - rainwater) have shown a similar variation, e.g., 1 and 2 dotted bars and 2, 3, and 4 filled bars (Fig. 6). Some cases 8

Atmospheric Research 235 (2020) 104760

N. Sinha and S. Chakraborty

Fig. 6. Time series of isotopic records of rain, vapor and atmospheric variables: (a) source moisture oxygen isotope ratio (C-G Model); (b) isotopic composition of the ambient vapor during the non-rainy (open circle) and rainy (closed circle) hours; the arrows indicate a declining trend in isotopic composition during the monsoon season (c) oxygen isotope record of rainfall (closed triangle); (d) the outgoing longwave radiation (OLR; dashed curve) superimposed on the daily rainfall record (bars; small circle overhead the bars are rainfall recorded during the events of deep convection). Columns (in gray and stippled) highlight the co-variation of isotopic time series of different components (source moisture - vapor- rain) during low/no rain and heavy/continuous rain conditions respectively (see text for details).

Table 2 Correlation coefficients between δ18Oe (Source moisture), δ18Ov (Ambient vapor), and δ18Or (Rain) for different rainfall (mm/d) scenarios of the year 2015. →

Scenario

(a) heavy (> 35.6 mm/d)

(b) moderate (35.5–7.6 mm/d)

(c) light (7.5–2.6 mm/d)

(d) very-light (2.5–0.1 mm/d)

(e) no-rain

n = 16

n = 28

n = 19

n = 19

n = 19

0.22 −0.14 −0.03

0.64 0.46 0.33

0.71 0.66 0.76

0.81 0.62 0.51

– 0.43 –

Correlation between

Ι. δ Ov &δ Or ΙΙ. δ18Oe&δ18Ov ΙΙΙ. δ18Or &δ18Oe 18

18

r

Numbers in bold are significant at the 95% confidence level.

oxygen isotope ratios of evaporated moisture over the ocean surface (δ18Oe) during non-rainy (no. of sample = 80) and rainy conditions (no. of samples = 27) are −10.26‰ and − 10.32‰ respectively. The values are the same within the ± 1σ uncertainty limit, implying the source moisture isotopic properties did not vary significantly on a seasonal timescale for the Andaman region. However, as mentioned earlier, they did show some variation on the daily timescale. Now, we assume that the same moisture was advected to the observation site, the δ18Ov are recalculated (GLV) as −12.32‰ and − 12.73‰ for nonrainy and rainy conditions respectively. Furthermore, non-rainy

calculation of the isotopic variations through the interaction between rain and ambient vapor has been attempted and presented in the next section.

3.4. Rain and GLV interaction: a case study We examine the mean seasonal isotopic variations in “source moisture - ambient vapor - rainwater” and attempt to quantify the isotopic interaction for two cases; GLV samples collected during (i) nonrainy and (ii) rainy conditions at the site. The arithmetic averages of 9

Atmospheric Research 235 (2020) 104760

N. Sinha and S. Chakraborty

relation with stratiform rain than with the convective rain fraction (Aggarwal et al., 2016). Since the study site is subjected to strong convective activities, this relationship (i.e., the amount effect) is always poor. But the year 2015 was probably an exception (Fig. S5), which showed a much better correlation, which was not seen in any other years (Sinha et al., 2019a, 2019b). It means that this year received more of stratiform rain and relatively less amount of convective rain. Because of this reason, rain-vapor interaction is likely to be subdued this year.

conditions can be divided into two, such as it is not raining during vapor sampling but it rained on that day (−12.47‰) and when it is was not a rainy day (−12.22‰). The observed variations for the GLV are presumably due to the preceding rainout process and moisture mixing from the surface. However, these factor is likely to be operative for all the conditions (non-rainy and rainy) for the vapor sampling, thus, a more depleted value (0.41‰, experimental uncertainty level is 0.1‰) of average δ18Ov during the rainy condition clearly indicates that the ambient vapors suffered from more isotopic fractionation due to isotopic interaction. This led to isotopic enrichment in the collected rainwater samples. It was corroborated by our observation, i.e., the rainwater samples during the vapor sampling hours were isotopically enriched (−2.65‰) than that of the 24 h cumulative rainfall (−2.77‰) collection. Although not significant in this case but it does support that the rainwater should be enriched while vapors are depleted in heavy isotopes due to the isotopic exchange process. Thus, the isotopically depleted vapors are subsequently fed to the cloud system, and the rainwater produced from this secondary moisture would also be isotopically depleted. This is one of the reasons for a relatively strong amount effect observed for the year 2015 (Rainfall vs. δ18O, r = −0.48, p < 0.01, Fig. S5) compared to the years 2012, 2013 and 2014, which were characterized by lower amount effect (Chakraborty et al., 2016; Sinha et al., 2019a, Munksgaard et al., 2019). As results show that the measured differences in the rain and vapor isotopic composition during non-rainy and rainy conditions are not very significant (as σ ~ 1‰) but came under experimental uncertainty (0.1‰). We attempted to evaluate the effect of rain-vapor interaction for this case following Rayleigh distillation and under a few assumptions. According to Rayleigh-type distillation during the rainout process,

δ18Ov ≈ δ18Oe + δ18Or

4. Conclusions For the summer months (ISM season) of 2015, the study of groundlevel vapor (GLV) and rainwater isotopic variations give a vivid picture of their interaction over the BoB region. The δ18O (or δD) of rainwater and GLV shows a significant positive correlation on the seasonal timescale, though some deviations were observed on a daily timescale. We found that the isotopic controls of the source moisture on ambient vapor could be one of the reasons for the observed discrepancy, in turn, on the isotopic variations of the rainfall. Hence, the isotopic interaction between vapor and rain can be a function of variation in the isotopic composition of the source moisture due to variability in meteorological parameters (C-G Model) on a daily timescale. Thus, isotopic variability in source moisture should also be considered for the comprehensive isotopic studies over the coastal regions. The present study investigated the isotopic relationship between the major components, i.e., “source moisture - ambient vapor - rainwater” on the basis of rainfall amount on a daily timescale. The results show that the δ18O variations in these components are positively correlated for the light rain condition. On the other hand, δ18O of ambient vapor and rainwater are significantly correlated over a wide range of rainfall amount, i.e., moderate to very-light rain. Further, no such correlations were observed for the heavy rainfall events, suggesting limitations of the Rayleigh fractionation and Craig-Gordon Model. Our current experimental setup is not adequate to overcome this limitation, but high frequency (such as sampling on hourly timescale using a laser isotope analyzer) analysis of rain and vapor samples is likely to offer a better understanding of the physical mechanisms controlling the variations of the isotopic composition during heavy rainfall conditions. In the case study, the isotopic interaction between rain and vapor over an island location has been examined, considering rainy and nonrainy conditions during vapor sampling hours. Our result shows that the isotopic exchange between ambient vapor and raindrops could lead to measurable depletion in the isotopic composition of vapor, in turn, subsequent enrichment in rainwater. Further, we estimated that the isotopic interaction might account for up to 30% in the observed variability of the isotopic composition of rain and vapor over the site. However, it may be noted that the year 2015 was characterized by higher relative humidity caused a negligible amount of raindrop evaporation. Since the rain-vapor isotopic interaction is weak in high humidity condition, the estimate of 0.41‰ is likely to be the lower limit of rain isotopic enrichment due to rain-vapor interaction process. It would have been good to repeat such exercise in another year to prove the hypothesis and quantify the rain-vapor isotopic interaction, probably using liquid water/vapor isotope analyzer. Declaration of Competing Interest None

(3)

where, δ Ov, δ Oe, and δ Or are the oxygen isotopic values for the vapor after the first rainout, evaporated moisture from the ocean surface, and rainwater respectively. However, there will be other factors, such as re-evaporation of raindrops, isotopic interaction, and evapotranspiration that may have some additional effect which may have negative or positive feedback (F) on the ambient vapor and rainwater isotopic compositions. Hence the Eq. (2) needs to be modified by an additive factor, F. 18

18

δ18Ov = δ18Oe + δ18Or ± F

18

(4)

In order to calculate the feedback factor, we assume i) The site receives the first rain from the source moisture and ii) During the rainy and humid conditions, the only factor that affects the isotopic compositions is the isotopic exchange between ambient vapor and raindrops. Using the assumption (i), the feedback factor is estimated for the Port Blair site. The calculated average isotopic values for the non-rainy condition, i.e., δ18Ov = −12.32‰, δ18Oe = −10.26‰, and δ18Or = −2.77‰ (Fig. 1b). From Eq. (3), the feedback factor (F) is 0.71‰ (notably, this value is much higher than the experimental uncertainty level of 0.1‰). On the other hand, for the rainy condition during vapor sampling hours, the average isotopic values are δ18Ov18 18 r = −12.73‰, δ Oe = −10.32‰, and δ Or-v = −2.65‰ (Fig. 1c). So, the calculated feedback factor (F) for the rainy and humid conditions is 0.24‰. Hence, it can be inferred from these calculations that the contribution of other factors to alter the isotopic composition of rain and vapor is 0.47‰ (0.71–0.24). This is approximately 70% of the maximum variation of 0.71‰, and the remaining 30% may be accounted for solely due to the isotopic interaction between vapor and rain. In general, the Port Blair region experiences high convective activities compared to other areas in the Bay of Bengal (Uma et al., 2016). However, the studied year appeared to have suffered less convective activities, as apparent in the calculated convective parameters on the daily scale for the years 2013 to 2015 (Fig. S6) with reference to climatology (2001–15). The rainwater isotope typically shows a strong

Acknowledgments The Indian Institute of Tropical Meteorology (IITM), Pune is fully supported by the Earth System Science Organization of the Ministry of Earth Sciences, Govt. of India. N. V. Vinith Kumar, National Institute of OceanTechnology, Port Blair is acknowledged for extending logistic support for sampling at their NIOT Campus, Port Blair. Amey Datye is thanked for his assistance with the laboratory work at IITM, Pune. We thank P. M.Mohan for carrying out a collection of the rainwater samples 10

Atmospheric Research 235 (2020) 104760

N. Sinha and S. Chakraborty

in the Pondicherry University, Port Blair. Era-interim datasets are obtained from http://apps.ecmwf.int/datasets/data/interim-full-daily/ levtype=sfc/.Kalpana-1 VHRR OLR data obtained from http://www. tropmet.res.in/~mahakur/Public_Data/. Partial financial support was provided by the IAEA through CRP F31004.

Geophys. Res. Lett. 23 (5), 527–530. Lawrence, R.L., Gedzelman, S.D., Dexheimer, D., Cho, H.K., Carrie, G.D., Gasparini, R., Anderson, C.R., Bowman, K.P., Biggerstaff, M.I., 2004. Stable isotope composition of water vapor in the tropics. J. Geophys. Res. Atmos. 109, D06115. Lee, J., Fung, I., 2008. “Amount effect” of water isotopes and quantitative analysis of post-condensation processes. Hydrol. Process. 22, 1–8. https://doi.org/10.1002/hyp. 6637. Lekshmy, P.R., Midhun, M., Ramesh, R., 2018. Influence of stratiform clouds on δD and δ18O of monsoon water vapour and rain at two tropical coastal stations. J. Hydrol. 563, 354–362. https://doi.org/10.1016/j.jhydrol.2018.06.001. Mahakur, M., Prabhu, A., Sharma, A.K., Rao, V.R., Senroy, S., Singh, Randhir, Goswami, B.N., 2013. A high resolution outgoing longwave radiation dataset from Kalpana-1 satellite during 2004–2012. Curr. Sci. 105 (8), 1124–1133. Majoube, M., 1971. Fractionation of oxygen 18 and of deuterium between water and its vapor. J. Chem. Phys. 68, 1423–1436. Marshall, J.S., Palmer, W.M., 1948. The distribution of raindrops with size. J. meteorol. 5, 165–166. Merlivat, L., Jouzel, J., 1979. Global climatic interpretation of the deuterium-oxygen 18 relationship for precipitation. J. Geophys. Res. 84 (C8), 5029–5033. https://doi.org/ 10.1029/JC084iC08p05029. Midhun, M., Lekshmy, P.R., Ramesh, R., 2013. Hydrogen and oxygen isotopic compositions of water vapour over the Bay of Bengal during monsoon. Geophys. Res. Lett. 40, 6324–6328. https://doi.org/10.1002/2013GL058181. Miyake, Y., Matsubaya, O., Nishihara, C., 1968. An isotopic study on meteoric precipitation. Pap. Meteorol. Geophys. 19 (2), 243–266. https://doi.org/10.2467/ mripapers1950.19.2_243. Moore, M., Kuang, Z., Blossey, P.N., 2014. A moisture budget perspective of the amount effect. Geophys. Res. Lett. 41, 1329–1335. https://doi.org/10.1002/2013gl058302. Munksgaard, N.C., Kurita, N., Sánchez-Murillo, R., et al., 2019. Data Descriptor: Daily observations of stable isotope ratios of rainfall in the tropics. Sci. Rep. 9, 14419. https://doi.org/10.1038/s41598-019-50973-9. Noone, D., 2012. Pairing measurements of the water vapor isotope ratio with humidity to deduce atmospheric moistening and dehydration in the tropical mid-troposphere. J. Clim. 25 (13), 4476–4494. https://doi.org/10.1175/jcli-d-11-00582.1. Risi, C., Bony, S., Vimeux, F., 2008. Influence of convective processes on the isotopic composition (δ18O and δD) of precipitation and water vapor in the tropics: 2. Physical interpretation of the amount effect. J. Geophys. Res. 113, 1–12. https://doi. org/10.1029/2008JD009943. Rozanski, K., Araguas-Araguas, L., Gonfiantini, 1992. Relation between long-term trends of oxygen-18 isotope composition of precipitation and climate. Science 5084 (258), 981–985. https://doi.org/10.1126/science.258.5084.981. Rozanski, K., Araguas-Araguas, L., Gonfiantini, 1993. Isotopic patterns in modern global precipitation. In: Swart, P.K. (Ed.), Climate Change in Continental Isotopic Records. vol. 78 American Geophysical Union, Washington, D.C Geophys. Monogr. Ser. Saranya, P., Gopal Krishan, Rao, Sudhir Kumar, M.S., 2018. Controls on water vapor isotopes over Roorkee, India: Impact of convective activities and depression systems. J. Hydrol. 557, 679–687. https://doi.org/10.1016/j.jhydrol.2017.12.061. Shenoi, S.S.C., Shankar, D., Shetye, S.R., 2002. Differences in heat budgets of the nearsurface Arabian Sea and Bay of Bengal: Implications for the summer monsoon. J. Geophys. Res. 107 (C6), 3052. https://doi.org/10.1029/2000JC000679. Sinha, N., Chakraborty, S., Chattopadhyay, R., et al., 2019a. Isotopic investigation of the moisture transport processes over the Bay of Bengal. J. Hydrol. 2, 100021. https:// doi.org/10.1016/j.hydroa.2019.100021. Sinha, N., Chakraborty, S., Mohan, P.M., 2019b. Modern rain-isotope data from Indian island and the mainland on the daily scale for the summer monsoon season. 23. pp. 103793. https://doi.org/10.1016/j.dib.2019.103793. Srivastava, R., Ramesh, R., Gandhi, N., Jani, R.A., Singh, A.K., 2015. Monsoon onset signal in the stable oxygen and hydrogen isotope ratios of monsoon vapor. Atmos. Environ. 108, 117–124. https://doi.org/10.1016/j.atmosenv.2015.02.062. Stewart, M.K., 1975. Stable isotope fractionation due to evaporation and isotopic exchange of falling water drops: Application to atmospheric processes and evaporation of lakes. J. Geophys. Res. 80, 1133–1146. Uemura, R., Matsui, Y., Yoshimura, K., Motoyama, H., Yoshida, N., 2008. Evidence of deuterium excess in water vapor as an indicator of ocean surface conditions. J. Geophys. Res. 113, D19114. https://doi.org/10.1029/2008JD010209. Uma, R., Lakshmi Kumar, T.V., Narayanan, M.S., 2016. Understanding convection features over Bay of Bengal using sea surface temperature and atmospheric variables. Theor. Appl. Climatol. 125, 469–478. https://doi.org/10.1007/s00704-015-1518-7. Worden, J., Noone, D., Bowman, K., Spect, T.E., 2007. Importance of rain evaporation and continental convection in the tropical water cycle. Nature 445 (7127), 528–532. Yoshimura, K., 2015. Stable water isotopes in climatology, meteorology, and hydrology: a review. J. Meteor. Soc. Jpn. 93, 513–533. Yu, W., Tian, L., Ma, Y., Xu, B., Qu, D., 2015. Simultaneous monitoring of stable oxygen isotope composition in water vapour and precipitation over the central Tibetan Plateau. Atmos. Chem. Phys. 15, 10251–10262. https://doi.org/10.5194/acp-1510251-2015. Yurtsever, Y., Gat, J.R., 1981. Atmospheric waters, Stable Isot. Hydrol. Deuterium Oxygen. Water Cycle 210, 103–142.

Appendix A. Supplementary data Supplementary data to this article can be found online at https:// doi.org/10.1016/j.atmosres.2019.104760. References Achyuthan, H., Deshpande, R.D., Rao, M.S., Kumar, Bhishm, Nallathambi, T., Shashi Kumar, K., Ramesh, R., Ramachandran, P., Maurya, A.S., Gupta, S.K., 2013. Stable isotopes and salinity in the surface waters of the Bay of Bengal: Implications for water dynamics and palaeoclimate. Mar. Chem. 149, 51–62. https://doi.org/10.1016/j. marchem.2012.12.006. Aggarwal, Pradeep K., Romatschke, U., Araguas-Araguas, L., Belachew, D., Longstaffe, Frederick J., Berg, P., Schumacher, C., Funk, A., 2016. Proportions of convective and stratiform precipitation revealed in water isotope ratios. Nat. Geosci. 9, 624. https:// doi.org/10.1038/ngeo2739. Araguas-Aragua, L., Froehlich, K., Rozanski, K., 2000. Deuterium and oxygen −18 isotope composition of precipitation and atmospheric moisture. Hydrol. Process. 14, 1341–1355. Benetti, M., Reverdin, G., Pierre, C., Merlivat, L., Risi, C., Steen-Larsen, H.C., Vimeux, F., 2014. Deuterium excess in marine water vapor: Dependency on relative humidity and surface wind speed during evaporation. J. Geophys. Res. Atmos. 119, 584–593. https://doi.org/10.1002/2013JD020535. Chakraborty, S., Sinha, N., Chattopadhyay, R., Sengupta, S., Mohan, P.M., Datye, A., 2016. Atmospheric controls on the precipitation isotopes over the Andaman Islands, Bay of Bengal. Sci. Rep. 6, 19555. https://doi.org/10.1038/srep19555. Chakraborty, S., Belekar, A.R., Datye, A., Sinha, N., 2018. Isotopic study of intraseasonal variations of plant transpiration: an alternative means to characterise the dry phases of monsoon. Sci. Rep. 8, 8647. https://doi.org/10.1038/s41598-018-26965-6. Clark, I.D., Fritz, P., 1997. Environmental Isotopes in Hydrogeology. CRC Press, Boca Raton, FL. Conroy, J.L., Noone, D., Cobb, K.M., Moerman, J.W., Konecky, B.L., 2016. Paired stable isotopologues in precipitation and vapor: a case study of the amount effect within western tropical Pacific storms. J. Geophys. Res. Atmos. 121, 3290–3303. https:// doi.org/10.1002/2015JD023844. Craig, H., Gordon, L., 1965. Deuterium and oxygen 18 variations in the ocean and the marine atmosphere. In: Stable Isotopes in Oceanographic Studies and Paleotemperatures. Lab. di Geol. Nucleate, Pisa, Italy, pp. 9–130. Dansgaard, W., 1964. Stable isotopes in precipitation. Tellus 16, 436–468. Dar, S.S., Ghosh, P., 2017. Estimates of land and sea moisture contributions to the monsoonal rain over Kolkata, deduced based on isotopic analysis of rainwater. Earth Syst. Dynam. 8, 313–321. https://doi.org/10.5194/esd-8-313-2017. Dee, D.P., et al., 2011. The ERA-Interim reanalysis: configuration and performance of the data assimilation system. Q. J. R. Meteorol. Soc. 137, 553–597. Deshpande, R.D., Maurya, A.S., Kumar, B., Sarkar, A., Gupta, S.K., 2010. Rain-vapor interaction and vapor source identification using stable isotopes from semiarid western India. J. Geophys. Res. 115, D23311. https://doi.org/10.1029/2010JD014458. Draxler, R.R., Ginoux, P., Stein, A.F., 2010. An empirically derived emission algorithm for windblown dust. J. Geophys. Res. 115, D16212. https://doi.org/10.1029/ 2009JD013167. Friedman, I., Machta, L., Soller, R., 1962. Water vapor exchange between a water droplet and its environment. J. Geophys. Res. 67, 2761–2770. Gat, J.R., 1996. Oxygen and hydrogen isotopes in the hydrologic cycle. Annu. Rev. Earth Planet. Sci. 24 (1), 225–262. https://doi.org/10.1146/annurev.earth.24.1.225. Gat, J.R., 2000. Atmospheric water balance – the isotopic perspective. Hydrol. Process. 14, 1357–1369. Kendall, C., McDonnell, J.J. (Eds.), 1998. Isotope Tracers in Catchment Hydrology. Elsevier, Amsterdam. Kumar, B., Rai, S.P., Saravana Kumar, U., Verma, S.K., Garg, P., Vijaya Kumar, S.V., Jaiswal, R., Purendra, B.K., Kumar, S.R., Pande, N.G., 2010. Isotopic characteristics of Indian precipitation. Water Resour. Res. 46, W12548. https://doi.org/10.1029/ 2009WR008532. Kurita, N., 2013. Water isotopic variability in response to mesoscale convective system over the tropical ocean. J. Geophys. Res. Atmos. 118 (10), 376–10,390. https://doi. org/10.1002/jgrd.50754. Kurita, N., Noone, D., Risi, C., Schmidt, G.A., Yamada, H., Yoneyama, K., 2011. Intraseasonal isotopic variation associated with the Madden‐Julian Oscillation. J. Geophys. Res. 116. https://doi.org/10.1029/2010JD015209. Lawrence, J.R., Gedzelman, S.D., 1996. Low stable isotope ratios of tropical cyclone rains.

11