The postmonsoon carbon biogeochemistry of estuaries under 6 different levels of anthropogenic impacts

Abstract. The different aspects of carbon biogeochemistry were studied during the postmonsoon at the Hooghly-Sundarbans estuarine system, a part of the Ganga-Brahmaputra river system located in the northeastern India. The study focused on understanding the differences in carbon biogeochemistry of estuaries undergoing different levels of anthropogenic stress by investigating anthropogenically influenced Hooghly estuary and mangrove-dominated estuaries of the Sundarbans. The salinity of well oxygenated (%DO: 91–104 %) estuaries of the Sundarbans varied over a narrow range (12.74–16.69) during postmonsoon relative to the Hooghly (0.04–10.37). Phytoplankton productivity and carbonate precipitation and/or dissolution were dominant processes controlling DIC dynamics in different parts of the Hooghly, whereas signal for mangrove derived DIC removal was observed in the Sundarbans. Influence of groundwater on estuarine DIC biogeochemistry was also observed in both the estuaries with relatively higher influence at the Hooghly than Sundarbans. In both estuarine systems, DOC behaved non-conservatively with ~ 40 % higher DOC level in the Hooghly compared to the Sundarbans. No significant evidence of phytoplankton production on DOC level was found in these estuaries, however signal of DOC input through pore-water exchange at the Sundarbans was observed. Relatively lower δ13CPOC at the Hooghly compared to the Sundarbans suggest relatively higher terrestrial influence at the Hooghly with a possibility of in situ biogeochemical modifications of POC at the Sundarbans. The freshwater run-off coupled with in situ aerobic OC mineralization controlled estuarine pCO2 level at the Hooghly, whereas the same was principally exogenous for the Sundarbans. The entire Hooghly-Sundarbans system acted as source of CO2 to the regional atmosphere with ~ 17 times higher emission from the Hooghly compared to Sundarbans. The present study clearly establishes the dominance of anthropogenically influenced estuary over relatively pristine mangrove dominated one in the regional greenhouse gas budget and climate change perspective.



Introduction
Situated at the interface of land and sea, estuaries are highly susceptible to anthropogenic inputs and undergo intricate biogeochemical and hydrological processes.Estuaries play an important role in modulating global carbon (C) cycle and anthropogenic carbon dioxide (CO2) budget (Bauer et al., 2013;Regnier et al., 2013;LeQuéré et al., 2016).Atmospheric CO2 is sequestered into terrestrial systems through photosynthesis and weathering reactions and is transported to the ocean via rivers and estuaries.Tropical rivers, which constitute ~ 66% of global river water discharge, deliver ~ 0.53Pg C to the estuaries annually (Huang et al., 2012).The majority of this exported C is in dissolved form [dissolved inorganic C (DIC): 0.21PgCyr -1 and dissolved organic C (DOC): 0.14PgCyr -1 ] with some contribution as particulate [particulate organic C (POC): 0.13PgCyr -1 and particulate inorganic C (PIC): 0.05PgCyr -1 ] (Huang et al., 2012).
Although estuaries are only ~ 4% of the continental shelf regions, CO2 emission flux from estuarine surface waters is as high as CO2 uptake in continental shelf regions of the world, albeit with large uncertainty (Borges et al., 2005;Chen and Borges, 2009;Cai et al., 2006;Cai, 2011).This suggests estuaries to be not only active pathway for transport of C (Ittekkot and Laane, 1991) but also a hotspot for biogeochemical modification of labile organic matter (OM) (Frankignoulle et al., 1998).
Mangroves covering 137,760 km 2 along tropical and sub-tropical estuaries and coastlines (Giri et al. 2011) are among the most productive natural ecosystems in the world with net primary productivity of 218 ± 72Tg C yr -1 (Bouillon et al. 2008).Fine root production coupled with litter fall and wood production are primary sources of mangrove derived C to intertidal sediment (Bouillon et al., 2008).The fate of this mangrove derived C remains poorly understood.Despite taking C burial and CO2 emission flux across mangrove sedimentatmosphere interface into account, estimates of global mangrove C budget revealed a significant imbalance (~72%) between mangrove net primary productivity and its sinks (Bouillon et al., 2008).Earlier studies reported mangroves to be responsible for ~10% of the global terrestrial derived POC and DOC export to the coastal zones (Jennerjahn and Ittekkot, 2002;Dittmar et al. 2006).However, recent studies proposed DIC exchange as major C export pathway from mangrove forests, which was ~70% of the total mineralized C transport from mangrove forests to coastal waters (Maher et al., 2013;Alongi, 2014;Alongi and Mukhopadhyay, 2014).Another study reported groundwater advection from mangrove to be responsible for 93-99% of total DIC export and 89-92% of total DOC export to the coastal ocean (Maher et al., 2013).Upon extrapolating these C export fluxes to the global mangrove area, it was found that the calculated C exports were similar to the missing mangrove C sink (Sippo et al., 2016).The remaining C that escapes export gets buried in sub-surface sediment layers and participates in anaerobic processes (linked to production of biogenic trace gases like CH4) or undergoes long-term sequestration (Jennerjhan and Ittekkot 2002; Barnes et al., 2006;Kristensen and Alongi, 2006;Donato et al., 2011;Linto et al., 2014).
Apart from lateral transport of dissolved and particulate C, biogeochemical processes such as primary production, OM mineralization, CaCO3 precipitation / dissolution and wateratmosphere CO2 exchange occurring in the estuarine water column also regulate inorganic and organic C biogeochemistry of a mangrove-dominated estuary.These processes largely depend upon pH, nutrient availability, euphotic depth variability as well as planktonic and bacterial biodiversity and community compositions.The biogeochemical cycling of bioavailable elements, such as C and N, in a mangrove-dominated estuary is largely different from anthropogenically polluted estuary, where much of the OM is derived from domestic, agricultural and industrial wastes.In anthropogenically affected estuarine systems, heterotrophy generally dominates over autotrophy (Heip et al., 1995;Gattuso et al., 1998) and a substantial fraction of biologically reactive OM gets mineralized within the system (Servais et al., 1987;Ittekkot, 1988;Hopkinson et al., 1997;Moran et al., 1999).However, this is not always the case as observed in Guanabara Bay, Brazil, which acts as a strong CO2 sink enhanced by eutrophication (Cotovicz Jr. et al., 2015).Lack of ample quantitative estimation of above-mentioned biogeochemical processes in many regions of the world restrains biogeochemists from an in-depth understanding of these processes in different ecological settings.It also leads to uncertainty in estimation of C budget of coastal regions on global scale.
In India, research related to C biogeochemistry of estuarine ecosystems have been in focus since last two decades with emphasis on estuaries located in the southern India (e.g., Bouillon et al., 2003;Sarma et al., 2012;Sarma et al., 2014;Bhavya et al., 2017;Bhavya et al. 2018).During the present study, we focused on C biogeochemical differences of two adjacent estuarine systems, i.e., the estuaries of Sundarbans and Hooghly estuary, which are part of Ganga-Brahmaputra river system located in the northeastern India (Fig. 1).Characteristically, these two estuaries are very different from each other.The Hooghly estuary experiences significantly higher anthropogenic influence compared to mangrove-dominated Sundarbans as evidenced by high nutrient and freshwater input (Table 1).The anthropogenic influences largely include supply of the industrial effluents and domestic sewage on daily basis from industries and major cities (Kolkata and Howrah) located upstream (Table 1).The industries along the Hooghly is principally jute (Corchorus olitorius) based industry, which produces fabrics for packaging a wide range of agricultural and industrial commodities.
The major focus of biogeochemical studies in the Hooghly and Sundarbans has been on trace gases (Mukhopadhyay et al., 2002;Biswas et al., 2004Biswas et al., , 2007;;Ganguly et al., 2008Ganguly et al., , 2009;;Dutta et al., 2013Dutta et al., , 2015Dutta et al., , 2017) ) with exception of one comprehensive study on nutrient budget at the Hooghly estuary (Mukhopadhyay et al., 2006).Recently, attempts have been made to understand different aspects of C cycling in these two estuaries by different workers (Samanta et al., 2015;Ray et al., 2015Ray et al., , 2018;;Akhand et al., 2016).Samanta et al. (2015) have comprehensively studied DIC dynamics in the Hooghly estuary, whereas Akhand et al. (2016) focused on DIC and pCO2 at the Hooghly-Matla estuary.Different aspects of C cycling in Hooghly-Sundarbans system have been reported by Ray et al. (2015Ray et al. ( , 2018)).Barring Samanta et al. (2015), which has wider spatial and temporal coverage with respect to DIC in the Hooghly, other studies are severely limited in spatial coverage with focus on mid to lower part of the Hooghly estuary and a few locations in the Sundarbans (one location by Ray et al., 2015Ray et al., , 2018; three locations by Akhand et al., 2016).Given the vast expanse of these estuaries, extrapolation of data from these studies for the entire ecosystem may lead to overestimation/underestimation.
During the present study, we focused on understanding differences in varied aspects of C cycle (particulate organic, dissolved inorganic and organic along with gaseous form) of the Hooghly and Sundarbans during postmonsoon with relatively better spatial coverage compared to previous studies.The postmonsoon sampling was chosen because of relatively stable estuarine condition for wider spatial coverage and peak mangrove leaf litter fall during this season (Ray et al., 2011), which may have influence on estuarine C dynamics.Considering different nature and quantity of supplied OM within these two contrasting system, we hypothesize C metabolism between these two estuaries to be very different with higher CO2 exchange flux from anthropogenically influenced estuary compared to mangrove-dominated estuary.Specifically, the major aims of the present study were to: (a) investigate factors controlling DIC and DOC dynamics in the region, (b) sources of POM in these two contrasting systems, and (c) partial pressure of CO2 (pCO2) and its controlling mechanisms along with exchange across water-atmosphere interface at the Hooghly-Sundarbans during postmonsoon period.

Study area
The present study was carried out in mangrove dominated estuaries of Indian Sundarbans and anthropogenically dominated Hooghly estuary in the northeastern India.Sundarbans (21 o 32' and 22 o 40'N: 88 o 05' and 89 o E), inscribed as a UNESCO world heritage site, is the largest mangrove forest in the world situated at the land-ocean boundary of the Ganges -Brahmaputra delta and the Bay of Bengal (BOB).Out of 10,200 km 2 area of Sundarbans, 41% is in India and the rest is in Bangladesh.The Indian part of Sundarbans (or Sundarbans Biosphere Reserve) contains 4200 km 2 of mangrove reserve forest and 1800 km 2 of estuarine waterways along with reclaimed areas.The Sundarbans is crisscrossed by several rivers, such as Muriganga, Saptamukhi, Thakuran, Matla, Bidya, Gosaba and Haribhanga, forming a sprawling archipelago of 102 islands covered with thick mangroves mostly composed of Avicennia alba, Avicennia marina and Avicennia officinalis.Semidiurnal tide with mean depth ~ 6 m is general characteristic of the estuary (Dutta et al., 2015).
The second study site, the Hooghly estuary (21°31′-23°20′N and 87°45′-88°45′E), is the first deltaic offshoot of the Ganges which ultimately mixes with the northern BOB.Like estuaries of Sundarbans, tides are semidiurnal in nature in the Hooghly as well with variable depth along the channel (~ 21 m at Diamond Harbor (H6) to ~ 8 m at the mouth of the estuary; Fig 1b) (CIFRI, 2012).Before mixing with the BOB, the lower estuarine part of the Hooghly divides into two channels, one being main estuarine stream which directly mixes with the BOB and another smaller channel known as Muriganga (mean depth ~ 6 m; Sadhuram et al., 2005).
The width of the river at the mouth of the estuary is ~ 25 km (Mukhopadhyay et al., 2006).
Covering upper, middle, and lower estuarine regions, the present study was carried out during low tide condition in three major estuaries of the Indian Sundarbans [Saptamukhi (S1-S3), Thakuran (T1-T3), and Matla (M1-M3); Fig. 1a] along with its related waterways (S4 & M4).The low-tide postmonsoon sampling was preferred as it was ideal time to evaluate the effect of mangroves on the adjoining estuary due to peak mangrove litter fall (Ray et al., 2011) and groundwater (or pore-water) discharge.To compare and bring out the contrast in different components of the C cycle between mangrove-dominated and anthropogenically influenced estuaries, low-tide sampling was also performed at 13 locations (H1 -H13, Fig. 1b) in the Hooghly estuary (stretch: ~150km).
For the purpose of discussion, henceforth, both the estuarine systems will be discussed as 'Hooghly-Sundarbans system' and the estuaries of Sundarbans will be called 'Sundarbans' unless discussed individually.

Sampling and experimental techniques
During postmonsoon (November, 2016), estuarine surface water samples were collected in duplicate at different locations of the Hooghly-Sundarbans system using Niskin bottle (Oceantest equipment; capacity: 5L).A brief description of the on and off field sampling and experimental techniques used during the present study are described below.

Sample collection and on board measurements
Water temperature and pH of the collected samples were measured onboard using thermometer (±0.1 o C) and portable pH meter (Orion Star A211) fitted with a Ross type combination electrode calibrated (as described by Frankignoulle and Borges, 2001) on the NBS scale (reproducibility: ±0.005 pH units).Salinity (±0.1) and dissolved oxygen (DO: ±0.1mgL -1 ) concentrations were measured onboard following the Mohr-Knudsen and Winkler titration methods, respectively (Grasshoff et al., 1983).For total alkalinity (TAlk), 50 ml of filtered (Whatman GF/F filter) estuarine water was titrated onboard in a closed cell using 0.1N HCl following potentiometric titration method (Bouillon et al., 2003).Uncertainty in TAlk measurements was ±1 µmolkg -1 as estimated using certified reference material (Dickson standard: CRM-131-0215).
For DIC and δ 13 CDIC measurements, estuarine surface waters were collected by gently overfilling glass vials fitted with teflon septa.Pore-water was also collected from lower littoral zone of the Lothian Island (one of the virgin island of the Indian Sundarbans) by digging a hole (~30 cm below the water table).It was not possible to collect pore-water samples from mid and upper littoral zones due to logistic problems.After purging water at least twice in the bore, sample was collected from the bottom of the bore through syringe and transferred to the glass vial (Maher et al., 2013).Twelve groundwater samples were collected from the nearby locations of the Hooghly-Sundarbans system via tube pump.After collection, all samples for DIC and δ 13 CDIC were preserved immediately by adding saturated HgCl2 solution to arrest the microbial activity.
For both DOC and SPM (suspended particulate matter) measurements, surface water samples were filtered on board through pre-weighted and pre-combusted (500 o C for 6 hours) Whatman GF/F filters (pore size: 0.7µm).Filtrate was kept for DOC analysis in brown bottles followed by immediate preservation via addition of H3PO4 (50µL/15 mL sample) (Bouillon et al., 2003), whereas the residue was kept for particulate matter analysis.Collected DIC, DOC and SPM samples were properly preserved at 4 o C during transportation to the laboratory.
Additionally, micrometeorological parameters associated with water-atmosphere CO2 exchange flux computation were continuously monitored at 10 m height over the estuary using a portable weather monitor (DAVIS -Vintage Pro2 Plus).

Laboratory measurements
The DIC were measured using Coulometer (Model: UIC.Inc. CM -5130) with analytical uncertainty of ±0.8%.The δ 13 CDIC were measured using Gas Bench attached to a continuous flow mass spectrometer (Thermo Delta V) with precision better than 0.10‰.The DOC were measured using high-temperature catalytic oxidation analyzer (Shimadzu TOC 5000), which was calibrated using potassium hydrogen phthalate (KHP) solution containing 1, 2, 5, 10, 20 mg L −1 of DOC (Ray et al., 2018).The analytical error for DOC measurement was < 2%.For SPM measurement, filter papers containing SPM were dried in hot air oven at 60 o C and final weights were noted.The SPM were calculated based on difference between final and initial weights of the filter paper and volume of water filtered.For measurement of POC and δ 13 CPOC, SPM containing filter papers were de-carbonated (by HCl fumes) and analyzed using Elemental Analyzer attached to the continuous flow mass spectrometer via conflo.The δ 13 CPOC values are reported relative to V-PDB with reproducibility better than ± 0.10‰, whereas uncertainty for POC was <10%.

Computation of air -water CO2 flux and %DO
The pCO2 was calculated based on surface water temperature, salinity, TAlk, pH and dissociation constants calculated following Millero (2013).The uncertainty for estimated pCO2 was ± 1%.The CO2 exchange fluxes (FCO2 in µmol m -2 hr -1 ) across water-atmosphere boundary of the estuary were calculated as follows: Where, KH CO2 = CO2 solubility.'k' is gas transfer velocity, which is highly variable and remains a matter of debate (Raymond and Cole, 2001).The 'k' during the present study was computed as a function of wind velocity following Liss and Merlivat (1986) parametrization.For the same wind velocity, the parametrization of Liss and Merlivat (1986) provides least 'k' value over other parametrization (Wanninkhof, 1992;Raymond and Cole, 2001;Borges et al., 2004) and therefore, the FCO2 presented during this study may be considered as the conservative estimates.The wind velocity based 'k' estimation for the Hooghly-Sundarbans system has been applied in earlier studies as well (Mukhopadhyay et al., 2002, Biswas et al., 2004).Mean global atmospheric CO2 mixing ratio in dry air during 2016 (data source: ftp://aftp.cmdl.noaa.gov/products/trends/co2/co2_annmean_gl.txt) was corrected for water vapor partial pressure to calculate pCO2(atmosphere).The fraction, "KH CO2 x [pCO2 (water) -pCO2 (atmosphere)]" is the departure of free dissolved CO2 from atmospheric equilibrium that may be termed as "excess CO2 (ECO2)" (Zhai et al., 2005).
%DO and apparent oxygen utilization (AOU) were calculated as follows: Where, [O2] Equilibrium is the equilibrium DO concentration calculated at in-situ temperature and salinity (Weiss, 1970) and [O2] Measured is the measured DO concentration of surface water.

Mixing model calculation
Considering salinity as a conservative tracer and an ideal indicator for estuarine mixing mechanism (Fry, 2002), conservative mixing model was applied to the Hooghly estuary to understand addition/removal of dissolved and particulate C by in situ biogeochemical processes.Concentrations and stable isotopic compositions of dissolved or particulate C (presented as C) during conservative mixing (CCM and δ 13 CCM) were computed as follows (Carpenter et al., 1975, Mook andTan, 1991): Here, 'S' denotes salinity, the suffixes CM, F, M and S denote conservative mixing, freshwater end member, marine end member and sample, respectively.FF = freshwater fraction = 1 -(SS / SM) and FM = marine water fraction = (1-FF).CSample > CCM indicates C addition, whereas reverse indicates removal.For model calculation, mean salinity, concentrations of C and δ 13 C of samples collected at salinity ≤ 0.3 at the Hooghly estuary were considered as end member values for freshwater, whereas respective values for marine end member were taken from Dutta et al. (2010) and Akhand et al. (2012).Quantitative deviations (∆C and ∆ 13 C) of measured C concentrations and δ 13 C from the respective conservative mixing values were estimated as follows (Alling et al., 2012): Plots between ∆C and ∆ 13 C for DIC and POC have been used to understand processes influencing DIC and POC in the Hooghly-Sundarbans system.However, the above model could not be applied to DOC due to unavailability of  13 CDOC during the present study.
Unlike Hooghly, direct application of above-mentioned conservative mixing model was not justified for mangrove-dominated Sundarbans due to narrow salinity gradient (see later).
However, assuming that apart from conservative mixing only mangrove derived C (∆CMangrove) contributes to estuarine C pool, an approach can be taken to quantify ∆CMangrove.Two different mass balance equations as used by Miyajima et al. (2009) for estimating ∆DICMangrove was extended to calculate ∆CMangrove during the present study: For model calculation, δ 13 CMangrove was taken as -28.4‰ for Sundarbans (Ray et al., 2015) and end members were taken as same as the Hooghly as estuaries of Sundarbans are offshoot of lower Hooghly estuary.

Environmental parameters
During the present study, water temperature did not show any distinct spatial trend and varied from 28 -29 o C and 30.5 -33 o C for the Sundarbans (Table 2) and Hooghly (Table 3), respectively.Salinity of the estuaries of Sundarbans varied over a narrow range (12.74 -16.69; Table 2) with minimum at the upper estuarine location throughout.A relatively sharp salinity gradient was noticed at the Hooghly estuary (0.04 -10.37;Table 3).Surface water DO concentrations were marginally higher in the Sundarbans (6.46 -7.46 mgL -1 ) than the Hooghly (5.24-7.40mgL -1 ).Both pH and TAlk in the Hooghly estuary (pH: 7.31 to 8.29, TAlk: 1797 to 2862 µeqL -1 ) showed relatively wider variation compared to the estuaries of Sundarbans (pH: 8.01 to 8.13, TAlk: 2009 to 2289 µeqL -1 ; Table 2 & 3).

Discussion
Based on the observed salinity gradient, the Hooghly estuary can be divided into two major salinity regimes: (a) fresh-water zone (H1-H6) and (b) mixing zone (H7 -H13; Fig. 1b).Due to narrow salinity range, no such classification was possible for the estuaries of Sundarbans.% DO calculations showed relatively well-oxygenated estuarine environment in the Sundarbans (91 -104%) compared to the Hooghly (71 -104%; Fig. 2).Based on the results obtained during the present study, below we discuss different components of C cycle within Hooghly-Sundarbans system.

Major drivers of DIC dynamics
In the Hooghly, DIC concentrations during the present study were relatively higher compared to that reported by Samanta et al. (2015) for the same season, whereas δ 13 CDIC values were within the same range (DIC: 1700 -2250µM; δ 13 CDIC: -11.4 to -4.0‰).Statistically significant correlations between DIC -salinity (r 2 = 0.43, p = 0.015) and δ 13 CDIC -salinity (r 2 = 0.58, p = 0.003) in the Hooghly suggested potential influence of marine and freshwater mixing on DIC and δ 13 CDIC in the estuary (Fig. 3a & 3b).The above-mentioned significant relationships during the present study coupled with earlier δ 18 O -salinity (Ghosh et al., 2013) and DIC dynamics (Samanta et al., 2015) studies in the Hooghly rationalize application of two end member mixing model in this estuary to decipher in situ processes influencing DIC chemistry.
Based on the methodology discussed earlier, calculated ∆C for DIC (∆DIC ~ -0.27 to 0.17) predicted dominance of DIC addition (n = 4) over removal (n = 2) in the freshwater region of the Hooghly, whereas only removal was evident in the mixing zone.In case of ∆δ 13 C for DIC (∆δ 13 CDIC), values were mostly positive (n = 9), i.e., measured δ 13 CDIC was higher compared to estimated δ 13 CDIC due to conservative mixing.Deviation plot (∆DIC vs. ∆δ 13 CDIC; Based on these calculations, both organic and inorganic processes (productivity, carbonate precipitation and dissolution) along with physical processes (CO2 outgassing across water-atmosphere interface) appeared to regulate DIC chemistry in the Hooghly estuary.
Spatially, PP and CO2 OG appeared to regulate DIC in the mixing zone (n = 5 out of 7) of the Hooghly.Earlier studies have advocated high phytoplankton productivity in non-limiting nutrient condition during postmonsoon in the Hooghly (Mukhopadhyay et al., 2002;Mukhopadhyay et al., 2006).However, based on the present data, particularly due to lack of direct PP measurements, it was difficult to spatially decouple PP and CO2 outgassing in the mixing zone.In contrast to the mixing zone, CP and CD appeared to be dominant processes affecting estuarine DIC chemistry in the freshwater region of the Hooghly.
In mangrove-dominated estuaries of Sundarbans, our measured δ 13 CDIC values were within the range of that reported by Ray et al. (2018), whereas DIC concentrations were comparatively lower (DIC: 2130 ± 100 µmolkg -1 , δ 13 CDIC: -4.7 ± 0.7‰).Our data also showed Although data for primary productivity is not available for the study period, earlier studies have reported postmonsoon as peak season for phytoplankton productivity (Biswas et al., 2007;Dutta et al., 2015).Given the evidences for presence of DIC removal processes in the Sundarbans, a comprehensive study focused on rate measurements of these processes with higher spatial and temporal coverage is desirable to understand the balance between influx and out-flux of DIC in the Sundarbans.
Other than biogeochemical processes, factors such as groundwater and pore-water exchange to the estuary might also play significant role in estuarine DIC chemistry (Tait et al., 2016).High pCO2 and DIC along with low pH and TAlk/DIC are general characteristics of groundwater, specially within carbonate aquifer region (Cai et al., 2003).Although all the parameters of groundwater inorganic C system (like pH, TAlk and pCO2) were not measured during the present study, groundwater DIC were ~5.57and ~3.61 times higher compared to mean surface water DIC in the Sundarbans and Hooghly, respectively.The markedly higher DIC in groundwater as well as similarity in its isotopic composition with estuarine DIC may stand as a signal for influence of groundwater on estuarine DIC, with possibly higher influence at the Sundarbans than Hooghly as evident from the slope of the TAlk -DIC relationships (Hooghly: 0.98, Sundarbans: 0.03).In the Sundarbans, to the best of our knowledge, no report exists regarding groundwater discharge.Contradictory reports exist for the Hooghly, where Samanta et al. (2015) indicated groundwater contribution at low salinity regime (salinity < 10, same as our salinity range) based on 'Ca' measurement, which was not observed based on 'Ra' isotope measurement in an earlier study (Somayajulu et al., 2002).Pore-water DIC in the Sundarbans was ~7.63 times higher than the estuarine water, indicating possibility of DIC input from the adjoining mangrove system to the estuary through pore-water exchange depending upon changes in hypsometric gradient during tidal fluctuation.A first-time baseline value for advective DIC influx from mangrove sediment to the estuary (FDIC) via pore-water exchange was estimated during the present study using the following expression (Reay et al., 1995): FDIC = Sediment porosity x Mean linear velocity x Mean pore water DIC conc.
Mean linear velocity = Pore water specific discharge / Sediment porosity Using pore-water specific discharge and porosity as 0.008 cm min -1 and 0.58 (Dutta et al., 2013, Dutta et al., 2015), respectively during postmonsoon and extrapolating the flux value over daily basis (i.e., for 12 hours as tides are semidiurnal in nature), mean FDIC during postmonsoon was calculated as ~ 770.4 mmol m -2 d -1 .However, significant impact of porewater to estuarine DIC may be limited only in mangrove creek water (samples not collected) as evident from narrow variability of estuarine TAlk and DIC as well as no significant correlation between them (p = 0.93).A comprehensive investigation on ground and pore waters are needed to thoroughly understand their importance in controlling DIC chemistry of the Hooghly-Sundarbans system.
From the above discussion it appears that on an average ~ 327 µM higher DIC in the Hooghly compared to the Sundarbans may be due to cumulative interaction between freshwater content to the individual estuaries as well as degree of biogeochemical and hydrological processes.Relatively higher freshwater contribution in the Hooghly compared to the Sundarbans (as evident from salinity) as well as significant negative relationship between DIC -salinity proved significant impact of freshwater on DIC pool in the Hooghly.However, detailed quantification of other biogeochemical and hydrological processes is needed to decipher dominant processes affecting DIC dynamics in the Hooghly-Sundarbans system.

DOC in the Hooghly-Sundarbans
During the present study, DOC concentrations in the Hooghly estuary were higher compared previously reported by Ray et al. (2018) (226.9 ± 26.2 to 324 ± 27µM), whereas DOC in the Sundarbans were comparable with Ray et al. (2018) (262.5 ± 48.2µM).The marine and freshwater mixing did not appear to exert major control over DOC in the Hooghly-Sundarbans system as evident from lack of significant correlations between DOC and salinity (Hooghly freshwater: r 2 = 0.33, p = 0.23; Hooghly mixing region: r 2 = 0.10, p = 0.50; Sundarbans: r 2 = 0.27, p = 0.10, Fig. 4a).Our observations showed similarity with other Indian estuaries (Bouillon et al., 2003) with opposite reports from elsewhere (Raymond andBauer, 2001a, Abril et al., 2002).This indicates that DOC in this sub-tropical estuarine system is principally controlled by processes other than mixing of two water masses.
Although it is difficult to accurately decipher processes influencing DOC without δ 13 CDOC data, some insights may be obtained from estimated ∆C of DOC (∆DOC).The estimated ∆DOC in the Hooghly indicated both net addition (n = 3) and removal (n = 3) of DOC in the freshwater zone (∆DOC = -0.16 to 0.11); whereas, only net addition was evident throughout the mixing zone (∆DOC = 0.08 to 1.74).In the Sundarbans, except lower Thakuran (St.T3, ∆ DOCM1 = -20µM), net addition of mangrove derived DOC was estimated throughout (∆DOCM1 = 2 -134µM).
In an estuary, DOC can be added through in situ production (by benthic and pelagic primary producers), lysis of halophobic freshwater phytoplankton cells and POC dissolution.
DOC can be removed through bacterial mineralization, flocculation as POC, and photooxidation (Bouillon et al., 2006).At the Hooghly -Sundarbans system, no evidence for freshwater phytoplankton (δ 13 C: -33 to -40‰; Freitas et al., 2001) was found from δ 13 CPOC, ruling out its potential effect on DOC.Although an indirect signal for phytoplankton productivity was observed in the freshwater region from δ 13 CDIC and POC relationship (r 2 = 0.68, p = 0.05), further evaluation of its impact on DOC was not possible due to lack of direct primary productivity measurements.Contradictory results exist regarding influence of phytoplankton productivity on DOC.Some studies did not find direct link between DOC and primary productivity (Boto and Wellington, 1988), whereas primary productivity mediated significant DOC formation (~ 8 -40%) has been reported by others (Dittmar & Lara 2001a, Kristensen & Suraswadi 2002).
The DOC -pCO2 relationship suggested inefficient bacterial DOC mineralization in the Hooghly (freshwater zone: p = 0.69, mixing zone: p = 0.67, Fig. 4b).However, significant positive relationship between these two in the Sundarbans (r 2 = 0.45, p = 0.02, Fig. 4c) indicated increase in aerobic bacterial activity with increasing DOC.In mangrove ecosystems, leaching of mangrove leaf litter as DOC is fast as ~ 30% of mangrove leaf litter leaching as DOC is reported within initial 9 days of degradation (Camilleri and Ribi, 1986).In the Sundarbans, mangrove litter fall peaks during postmonsoon (Ray et al. 2011) and its subsequent significant leaching as DOC was evident during the present study from comparatively higher DOC compared to POC (DOC:POC = 0.50 -3.39, mean: 1.79 ± 0.94%).Our interpretation for Sundarbans corroborated with that reported by Ray et al. (2018) for the same system as well as Bouillon et al. (2003) for the Godavari estuary, South India.
Despite high water residence time in the Hooghly (~ 40 days during postmonsoon, Samanta et al., 2015) and in mangrove ecosystem like Sundarbans (Alongi et al., 2005, Singh et al., 2016), DOC photo-oxidation may not be so potent due to unstable estuarine condition in the Hooghly-Sundarbans system (Richardson number < 0.14) with intensive vertical mixing and longitudinal dispersion coefficients of 784 m 2 s −1 (Goutam et al., 2015, Sadhuram et al., 2005).The unstable condition may not favor DOC -POC interconversion as well but mediated by charged complexes and repulsion -attraction interactions, the interconversion partly depends upon variation in salinity.More specifically, the interconversion is efficient during initial mixing of fresh (river) and seawater and the coagulation is mostly complete within salinity range 2 -3.This appeared to be the case in the Hooghly, where DOC and POC was negatively correlated in the freshwater region (r 2 = 0.86, p = 0.007, Fig. 4d), which was missing in the mixing region (p = 0.43) and in the Sundarbans (p = 0.84).
Although estimated ∆DOC indicated largely net DOC addition to the Hooghly-Sundarbans system, except leaf litter leaching in the Sundarbans, no significant evidence for other internal sources was found.This suggested potential contribution from external sources that may include industrial effluents and municipal wastewater discharge (i.e., surface runoff) in the freshwater region of the Hooghly (Table 1).However, there is no direct DOC influx data to corroborate the same.Relatively higher DOC compared to POC (DOC/POC > 1) at some locations (H2, H5, H6) may stand as a signal for higher DOC contribution at those locations, but it is not prudent to pinpoint its sources due to lack of isotopic data.Although anthropogenic inputs are mostly confined to freshwater region, relatively higher DOC in the mixing zone of the Hooghly compared to freshwater region suggested DOC input via some additional pathway, possibly groundwater discharge.The contribution of groundwater to the Hooghly estuary within the salinity range observed during the present study has been reported (Samanta et al., 2015).However, there is no report of groundwater mediated DOC influx to the estuary.For mangrove-dominated ecosystems like Sundarbans, a recent study by Maher et al. (2013) estimated ~ 89 -92% of the total DOC export to be driven by groundwater advection.To understand spatial variability of DOC chemistry in the Hooghly-Sundarbans system, a thorough investigation related to groundwater and surface runoff mediated DOC flux is warranted.
Overall, on an average ~ 40% higher DOC in the Hooghly compared to the Sundarbans appeared to be due to cumulative effect of freshwater contributions, higher anthropogenic inputs, influence of biogeochemical processes and groundwater contribution.However, DOC inputs via other pathways may be dominant over freshwater mediated input as evident from insignificant DOC -salinity relationship during the present study.To quantitatively understand the relative control of the above-mentioned contributors to the DOC pool in the Hooghly-Sundarbans system, the individual components need to be studied in detail.

Major drivers of particulate organic matter
The average POC during the present study was considerably higher compared to that reported by Ray et al. (2018) for the Hooghly-Sundarbans (Hooghly: 40.3 ± 1.1 to 129.7 ± 6.7µM, Sundarbans: 45.4 ± 7.5µM).However, the present POC values were within the range reported for a large set of Indian estuaries (POC: 51 -750 µM; Sarma et al., 2014).No significant SPMsalinity or POC-salinity relationship was observed during the present study (Fig. 5a & 5b), except for a moderate negative correlation between POC and salinity (r 2 = 0.62, p = 0.06) in the freshwater region of the Hooghly.This inverse relationship may be linked to freshwater mediated POC addition.Also, as described earlier, contribution of POC via surface-runoff is also a possibility in this region due to presence of several industries and large urban population (St: H2: Megacity Kolkata) that discharge industrial effluent and municipal wastewater to the estuary on regular basis (Table 1).Primary signal for surface runoff mediated POC addition was evident in the freshwater zone where ~ 61% and ~ 43% higher POC at 'H3' and 'H4' compared to an upstream location (St.H2) was observed.However, based on the present data, it is not possible to decouple freshwater and surface runoff mediated POC input to the Hooghly estuary.Relatively lower contribution of POC to the SPM pool of the Sundarbans (0.66 -1.23%) compared to the Hooghly (0.96 -4.22%; Fig. 5c) may be due to low primary production owing to high SPM load (Ittekkot and Laane, 1991) as observed in the mangrove-dominated Godavari estuary in the southern India (Bouillon et al., 2003).
In the estuaries of Sundarbans, isotopic signatures of POC showed similarity with terrestrial C3 plants.Interestingly, despite being mangrove-dominated estuary (salinity: 12.74 -16.55) no clear signature of either freshwater or mangrove (δ 13 C: mangrove leaf ~ -28.4‰, soil ~ -24.3‰,Ray et al., 2015Ray et al., , 2018) ) borne POC was evident from δ 13 CPOC values, suggesting towards the possibility of significant POC modification within the system.Modification of POC within the estuaries of Indian sub-continent have been reported earlier (Sarma et al., 2014).Inter-estuary comparison revealed relatively lower average δ 13 CPOC at the Hooghly (mean δ 13 CPOC: -24.87 ± 0.89‰) compared to the Sundarbans (mean δ 13 CPOC: -23.36 ± 0.32‰), which appeared to be due to differences in degree of freshwater contribution, anthropogenic inputs (high in Hooghly vs. little/no in Sundarbans), nature of terrestrial C3 plant material (mangrove in the Sundarbans vs. others in Hooghly) as well as responsible processes for POC modification within the system.
To decipher processes involved in POC modification, estimated ∆C for POC (∆POC) in the Hooghly indicated both net addition (n = 3) and removal (n = 3) of POC in the freshwater region (∆POC = -0.45 to 0.48), whereas removal (n = 6) dominated over addition (n = 1) in the mixing region (∆POC = -0.39 to 0.07).In an estuary, POC may be added through freshwater and surface runoff mediated inputs, phytoplankton productivity, and DOC flocculation.The removal of POC is likely due to settling at subtidal sediment, export to adjacent continental shelf region, modification via conversion to DOC and mineralization in case of oxygenated estuary.
The plot between ∆δ 13 C for POC (∆δ 13 CPOC) and ∆POC (Fig. 5d) indicated different processes to be active in different regions of the Hooghly estuary.Decrease in ∆POC with increase in ∆δ 13 CPOC (RR; n = 4 for mixing region and n = 1 for freshwater region) suggested modification of POC due to aerobic respiration (or mineralization).This process did not appear to significantly impact estuarine CO2 pool as evident from the POC -pCO2 relationship (freshwater region: p = 0.29, mixing region: p = 0.50; Fig. 5e).Decrease in both ∆POC and ∆δ 13 CPOC (SD; n = 2 for mixing region and n = 2 for freshwater region) supported settling of POC to sub-tidal sediment.Despite high water residence time (~ 40 days during postmonsoon, Samanta et al., 2015), this process may not be effective in the Hooghly due to unstable estuarine condition (described earlier).Increase in ∆POC with decrease in ∆δ 13 CPOC (SR, FR & PP; n = 2 for freshwater region) indicated increase of POC via surface and freshwater runoff as well as phytoplankton productivity.Increase in both ∆POC and ∆δ 13 CPOC (n = 1 for mixing region and n = 1 for freshwater region) may be linked to DOC to POC conversion by flocculation.
In the Sundarbans, negative and lower ∆POCM2 (-209 to -28µM) compared to ∆POCM1 (-35 to 327µM) suggested DIC like behavior, i.e., simultaneous removal or modification along with addition of mangrove derived POC.No evidence for in situ POC-DOC exchange was obvious based on POC-DOC relationship; however, signal for POC mineralization was evident in the Sundarbans from POC -pCO2 relationship (r 2 = 0.37, p = 0.05, Fig. 5f).Similar to the Hooghly, despite high water residence time in mangroves (Alongi et al., 2005, Singh et al., 2016), unstable estuarine condition may not favor efficient settlement of POC at sub-tidal sediment.The export of POC from the Hooghly-Sundarbans system to the northern BOB, without significant in situ modification, is also a possibility.This export has been estimated to be ~0.02-0.07Tg and ~ 0.58Tg annually for the Hooghly and Sundarbans, respectively (Ray et al. 2018).

pCO2 and FCO2 in the Hooghly-Sundarbans
The estimated pCO2 for the Hooghly-Sundarbans system were in the range reported for other tidal estuaries of India (Cochin estuary: 150-3800µatm, Gupta et al., 2009;Mandovi -Zuari estuary: 500-3500µatm, Sarma et al., 2001).In the Sundarbans, barring three locations (S3, T3 and M2), a significant negative correlation between pCO2 and %DO (r 2 = 0.76, p = 0.005; Figure not given) suggested presence of processes, such as OM mineralization, responsible for controlling both CO2 production and O2 consumption in the surface estuarine water.
Although based on our own dataset, it is not possible to confirm the same.However, relatively higher pCO2 levels during low-tide compared to high-tide at Matla estuary in the Sundarbans (Akhand et al. 2016) as well as in other mangrove systems worldwide (Rosentreter et al., 2018, Call et al., 2015, Bouillon et al., 2007) suggested groundwater (or pore water) exchange to be a potential CO2 source in such systems.
Unlike Sundarbans, ECO2 -AOU relationship did not confirm significant impact of OM respiration on CO2 in either freshwater (p = 0.50) or mixing regions (p = 0.75) of the Hooghly (Fig. 6c).Overall, pCO2 in the freshwater region of the Hooghly was significantly higher compared to the mixing zone (Table 3), which may be linked to CO2 supply in the freshwater region through freshwater or surface runoff from adjoining areas (Table -1).Interestuary comparison of pCO2 also revealed ~1291 µatm higher pCO2 in the Hooghly compared to the Sundarbans, which was largely due to significantly higher pCO2 in freshwater region of the Hooghly (Table 2 & 3).Lack of negative correlation between pCO2 -salinity in freshwater region (Fig. 6d) of the Hooghly suggested limited contribution of CO2 due to freshwater inputs.
Therefore, CO2 supply via surface runoff may be primary reason for higher pCO2 in the Hooghly estuary.
Positive mean FCO2 clearly suggested the Hooghly-Sundarbans system to be a net source of CO2 to the regional atmosphere during postmonsoon (Fig. 6e & 6f).Specifically, from regional climate and environmental change perspective, anthropogenically influenced Hooghly estuary was a relatively greater source of CO2 to the regional atmosphere compared to the mangrove-dominated Sundarbans as evident from significantly higher CO2 emission flux from the Hooghly ([FCO2] Hooghly: [FCO2] Sundarbans = 17).However, despite being a CO2 source, FCO2 measured for the estuaries of Sundarbans were considerably lower compared to global mean FCO2 reported for mangrove-dominated estuaries (~ 43-59 mmol C m -2 d -1 ; Call et al., 2015).Similarly, FCO2 measured for the Hooghly estuary were relatively lower compared to some Chinese estuarine systems (Pearl River inner estuary: 46 mmol m -2 d -1 , Guo et al., 2009; Yangtze River estuary: 41 mmol m -2 d -1 , Zhai et al., 2007).
The difference in FCO2 between Hooghly and Sundarbans may be due to variability in pCO2 level as well as micrometeorological and physicochemical parameters controlling gas transfer velocity across water-atmosphere interface.Quantitatively, the difference in 'k' values for the Hoogly and Sundarbans were not large (k Sundarbansk Hooghly ~ 0.031 cmhr -1 ).Therefore, large difference in FCO2 between these two estuarine systems may be due to difference in pCO2.Taken together, supporting our hypothesis, it appears that differences in land use and degrees of anthropogenic influence have the potential to alter the C biogeochemistry of aquatic ecosystems with anthropogenically stressed aquatic systems acting as a relatively greater source of CO2 to the regional atmosphere than mangrove-dominated ones.

Conclusions
The present study focused on investigating different aspects of C biogeochemistry of the anthropogenically affected Hooghly estuary and mangrove dominated estuaries of the Sundarbans during postmonsoon.Following conclusions were deduced from the study:  With the exception of SPM, physicochemical parameters of the Hooghly estuary varied over a relatively wider range compared to the Sundarbans.
 Coupled with freshwater contribution, inorganic and organic C metabolism appeared to be dominant processes affecting DIC in the Hooghly.However, in the Sundarbans, significant DIC removal over addition was noticed.Influence of groundwater on estuarine DIC biogeochemistry was also observed with relatively higher influence at the Sundarbans.
 Higher DOC level in the Hooghly appeared to be regulated by coupled interactions among anthropogenic inputs, biogeochemical processes and groundwater contribution rather than freshwater mediated inputs.
 Signatures of freshwater runoff, terrestrial C3 plants, and anthropogenic discharge were found in POC of the Hooghly, whereas evidence for only C3 plants were noticed at the Sundarbans with possible POC modification.
 Organic matter mineralization and surface run-off from adjoining areas appeared to be dominant controlling factor for pCO2 in the Sundarbans and Hooghly, respectively, with higher average pCO2 in the Hooghly compared to the Sundarbans.
 The entire Hooghly-Sundarbans system acted as source of CO2 to the regional atmosphere with ~17 times higher emission from the Hooghly compared to Sundarbans, suggesting dominance of anthropogenically stressed estuarine system over mangrovedominated one from regional climate change perspective.
similarity with Khura and Trang river, two mangrove-dominated rivers of peninsular Thailand flowing towards Andaman sea, although from hydrological prospective these two systems are contrasting in nature [Sundarbans: narrow salinity gradient (12.74 -16.69) vs. Khura and Trang river: sharp salinity gradient (~ 0 -35);Miyajima et al., 2009].Like Hooghly, δ 13 CDIC -salinity relationship was statistically significant (r 2 = 0.55, p = 0.009) for the Sundarbans, but DICsalinity relationship remained insignificant (p = 0.18) (Fig.3d & 3e).Given the dominance of mangrove in the Sundarbans, the role of mangrove derived OC mineralization may be important in regulating DIC chemistry in this ecosystem.Theoretically, ∆CMangrove for DIC (∆DICMangrove) estimated based on DIC (∆DICM1) and δ 13 CDIC (∆DICM2) should be equal.The negative and unequal values of ∆DICM2 (-41 to 62 µM) and ∆DICM1 (-186 to 11 µM) indicate large DIC out-flux over influx through mangrove derived OC mineralization in this tropical mangrove system.The removal mechanisms of DIC include CO2 outgassing across estuarine water-atmosphere boundary, phytoplankton uptake and export to adjacent continental shelf region (northern BOB,Ray et al., 2018).The evidence for CO2 outgassing was found at almost all locations covered during the present study (10 out of 11 locations covered; see section 4.4).Also, a recent study byRay et al. (2018) estimated DIC export (~ 3.69Tg C yr -1 ) from the estuaries of Sundarbans as dominant form of C export.

Fig. 2 :
Fig. 2: %DO -salinity at the Hooghly-Sundarbans systems (Green, grey and blue colors indicate freshwater region of the Hooghly, mixing region of the Hooghly and Sundarbans, respectively).

Fig. 4 :
Fig.4: (a) DOC -salinity in the Hooghly-Sundarbans, (b) DOC -pCO2 in the Hooghly, (c) DOC -pCO2 in the Sundarbans, and (d) DOC -POC in the Hooghly-Sundarbans (Green, grey and blue color indicates freshwater region of the Hooghly, mixing region of the Hooghly and the Sundarbans, respectively).

Fig. 6 :
Fig. 6: (a) ECO2 -AOU in the Sundarbans (b) pCO2 -salinity in the Sundarbans (c) ECO2 -AOU in the Hooghly (d) pCO2 -salinity in the Hooghly (e) FCO2 -salinity in the Hooghly, and (f) FCO2 -salinity in the Sundarbans (Green, grey and blue color indicate freshwater region of the Hooghly, mixing region of the Hooghly and the Sundarbans, respectively).