Benthic mineralization and nutrient exchange over the inner continental shelf of western India

The western Indian continental shelf is one of the most productive coastal systems of the world ocean. This system experiences extreme changes in its oxygen regime, being normoxic from November to May and suboxic (denitrifying)/anoxic from June to October, owing to the biogeochemical response to cyclical monsoonal influence. In order to understand the impact of the seasonally varying oxygen regime on benthic mineralization, nutrient exchange and, in turn, on the shelf ecosystem, we carried out the first ever intact-core incubations during two contrasting seasons – spring intermonsoon and fall intermonsoon (late southwest monsoon) at a 28 m-deep fixed site on the inner shelf off Goa, dominated by fine-grained cohesive sediments. The results showed that incomplete sediment oxygen consumption (SOC) occurred during April as opposed to the complete SOC and subsequent sulfide flux observed in the fall intermonsoon incubations. The sediments acted as a perennial net source of DIN (dissolved inorganic nitrogen i.e. NO −3 + NO − 2 + NH + 4 ), PO 4 and SiO 4− 4 to the overlying water column. The efflux of DIN increased from 1.4 to 3.74 mmol m −2 d−1 from April to October, of which NH+4 flux comprised 59–100 %. During the oxic regime, ∼ 75 % of diffusing NH+4 appeared to be nitrified (2.55 mmol m−2 d−1), of which∼ 77 % remained coupled to benthic denitrification. Consequently, 58 % of NH +4 flux was lost in active coupled nitrification–denitrification, resulting in substantial N loss (1.98 mmol m −2 d−1) in the sediments. The continental shelf sediments switched over from being a NO−3 source during the oxic regime to a NO − 3 sink during the anoxic regime. During suboxia, benthic denitrification that is fed by NO−3 from the overlying water caused N loss at the rate of 1.04 mmol m −2 d−1. Nitrogen loss continued even under sulfidic conditions during October, possibly through the chemolithoautotrophic denitrification, at a potential rate of 3.21 mmol m −2 d−1. Phosphate flux increased more than 4-fold during October as compared to April, due to reductive dissolution of Feand Mn oxides. The SiO 4 flux increased during October apparently due to the higher availability of siliceous ooze from diatom blooms commonly occurring in the monsoon season. Slow oxidation of organic carbon (C org) under anoxia, lower temperature and reduced benthic faunal activity appeared to decrease benthic mineralization by 25 % as suggested by the drop in the C org oxidation rate from 63.8 mmol C m−2 d−1 in April to 47.8 mmol C m−2 d−1 in October. This indicated a higher preservation of C org during the late southwest monsoon. Sediment porosity, C org content and nutrients did not show significant variations from April to October. Porewaters were found to be enriched with NH +4 , PO 3− 4 and SiO 4 but depleted in NO − 3 and NO − 2 in these organic-rich sediments. Significant DIN, PO 3− 4 and SiO 4− 4 effluxes indicate the potential of benthic input in meeting nutrient demand of the phytoplankton community in this seasonally N-limited shelf system.


Introduction
Nutrients form the basis of the marine food chain, as their availability is essential to the primary production.Balance between uptake and regeneration is necessary for the maintenance of the pelagic nutrient stock and, in turn, for the functioning of the marine ecosystem.In tropical and

Published by Copernicus Publications on behalf of the European Geosciences Union.
subtropical open oceans, most of the photosynthetically produced organic matter is lost to greater depths and the entrainment of remineralized nutrients is largely prevented by the thermocline during most parts of the year, although sporadic mixing events reintroduce these nutrients (e.g., new N) to the euphotic zone (Kumar et al., 2004;Pennington et al., 2006).In contrast, shallowness of continental shelves (< 200 m) results in higher (> 80 %) downward export of plankton-derived POC (Jahnke, 1996) compared to that in open oceans.The consequent organic matter enrichment of underlying sediments triggers intense geochemical, microbial and benthic faunal activities, which control organic carbon oxidation/accumulation and regeneration of biogenic elements.The ensuing microbially mediated benthic remineralization occurring through sequential use of oxidants, i.e., O 2 > MnO 2 = NO − 3 > Fe 2 O 3 > SO 2− 4 > CO 2 (Kristensen, 2000), nutrifies the sedimentary pore water, and over time the nutrients are released to the overlying water column mainly via molecular diffusion (Devol and Christensen, 1993;Meile and Van Cappellen, 2003), macrobenthic activities (Karlson et al., 2007;Laverock et al., 2011), flowinduced advective pore water transport (Huettel et al., 2003;Cardenas and Jiang, 2011) and resuspension of surficial sediment (Johnson et al., 1999;Corbett, 2010).
Continental shelves are known to be biogeochemically active regions of the world ocean.Though they occupy just 8.6 % of the global oceanic area (Jørgensen, 1983), > 15 % of the primary productivity of the global ocean occurs here (Naqvi and Unnikrishnan, 2009), which accounts for ∼ 90 % of fishery resources (Pauley et al., 2002), thereby sustaining the livelihoods of millions of people worldwide.In the world ocean, 80 % of the organic carbon burial (Naqvi and Unnikrishnan, 2009) and 83 % of sedimentary remineralization (Jørgensen, 1983) take place in shelf sediments.Conspicuously, benthic-pelagic coupling is important in the biogeochemical cycling of C, O, N, P and Si in shelf environments.Benthic regeneration of essential macro-and micronutrients, and their fluxes to the overlying pelagic realm, play a key role in the sustenance and functioning of coastal ecosystems (Nixon et al., 1981;Devol and Christensen, 1993;Jahnke et al., 2000;Ferrón et al., 2009a).As benthic supply can control pelagic primary productivity, it has implications for CO 2 fixation and carbon burial, and ultimately impacts the global climate.
Seasonal anoxia is a characteristic biogeochemical feature of the western shelf of India (Naqvi et al., 2000), which has inspired numerous physical, geochemical and biological investigations (Naik and Naqvi, 2002;Agnihotri et al., 2008;Krishnan et al., 2008;Kurian et al., 2009;Ingole et al., 2010;Roy et al., 2011;Maya et al., 2011;Shenoy et al., 2012) over the last decade.It is considered as a biogeochemical hotspot in the world ocean due to the recent intensification of O 2 deficiency and the possible implications thereof for nitrogen cycling and feedback to global climate (Naik and Naqvi, 2006;Naqvi et al., 2009).The western Indian shelf is one of the most productive regions of the Arabian Sea, owing to nutrient enrichment through monsoonal upwelling (Shetye et al., 1990) and land/river runoff during the southwest monsoon season (June-September) (Naqvi et al., 2000(Naqvi et al., , 2006)), nitrogen fixation during the spring intermonsoon (February-April; Roy et al., 2011;Parab et al., 2012) and atmospheric DIN deposition (Naqvi et al., 2009).The system undergoes extreme change in oxygen regime over the course of a year: the shelf water column remains oxygenated from November to May and becomes suboxic/anoxic during July-October.Hypoxia (O 2 ≤ 62.5 µM; Levin et al., 2009) starts in June, followed by suboxia (denitrifying condition, O 2 ≤ 4.46 µM; see Naqvi et al., 2010, and references therein) during July-August and often by anoxia (sulfidic condition; Naqvi et al., 2010, and references therein) during September-October (Naqvi et al., 2006).However, anoxia (sulfidic condition) remains limited only to the nearshore region of the shelf and is not a regular occurrence (e.g., years 1997, 2000, 2003 and 2007).Longitudinal coexistence of hypoxia, suboxia and anoxia 1 over the shelf in the late monsoon season (Naqvi et al., 2006) makes it a unique biogeochemically active region in the world ocean.Though the western Indian shelf occupies only ∼ 8 % of the area of the Arabian Sea, it accounts for 11 % of total pelagic denitrification (Naik and Naqvi, 2002) and 59 % of the total N 2 O emission from this basin (Naqvi et al., 2000).However, benthic metabolism, sedimentary nutrient cycling and benthic-pelagic coupling, which could exert a profound influence on the shelf ecosystem, have not been assessed so far.Naqvi et al. (2006) hypothesized that anoxia during the late southwest monsoon would culminate in the dissolution of sedimentary ferric oxyhydroxide minerals and lead to the release of Fe-bound phosphate to the overlying water.The mobilized PO 3− 4 , along with other inorganic nutrients (e.g., NH + 4 ), is extensively taken up by the phytoplankton community (e.g., dinoflagellates) and later by N 2 fixers during the spring intermonsoon (February-March).This indicates the potential of benthic supply to meet the nutrient demands of the phytoplankton community, and thus benthic-pelagic coupling needs to be evaluated.The sediments over the continental shelf experience a wide range of physico-chemical conditions (temperature, dissolved O 2 concentration and nutrient loading) and receive varying quantities of organic matter from the surface over the course of a year.The rate of benthic respiration (supported by various electron acceptors (i.e., O 2 , NO − 3 , MnO 2 , Fe oxides and SO 2− 4 ) and nutrient ex-   changes are thus expected to undergo large seasonal changes, not observed in other areas of the global ocean.

CaTS site
In the present study, we have tried to simulate the natural conditions using ex-situ incubations at a fixed 28 m deep site on the inner shelf during two contrasting seasons.The aim of this study was to assess the impact of varying O 2 regime on benthic respiration, C org mineralization and nutrient fluxes over the inner continental shelf.Although benthic nutrient flux studies have previously been done in the eastern Arabian Sea (Pakistan margin) by Woulds et al. (2009) and Schwartz et al. (2009), to our knowledge this is the first-ever study on benthic respiration and nutrient exchange on the Indian shelf.We hypothesize that the benthic input has the potential to meet the nutrient demand of the phytoplankton community in the shelf ecosystem, especially during the nitrogen-poor, non-upwelling period (October-May).
A large portion (∼ 66 %) of the western Indian shelf (i.e., outer shelf) is covered by sandy, C org -poor and reactive permeable sediments, whereas cohesive (silty-clayey) sediments dominate the inner shelf.The benthic respiration, C org mineralization and nutrient fluxes in permeable sediments can be radically different from those observed in cohesive sediments (Janssen et al., 2005;Rao et al., 2007;Cook et al., 2007).Our results are based on the flux measurement at a fixed study site on the inner shelf where the sediments are cohesive in nature.We do not intend to extrapolate our findings to the entire western shelf considering uncertainties associated with substantial spatial heterogeneity in C org and granulometric composition across the shelf; rather, we confine our results and inferences only to the inner shelf environments.

Study site
Our study site (G5) is located at 15 • 31 N and 73 • 39 E, ∼ 10 km off Candolim, Goa, on the central west coast of India (Fig. 1) at a depth of 28 m.It lies on our regularly studied shelf transect that has been studied monthly or fortnightly since 1997 under the ongoing Candolim Time Series (CaTS) program.Physical, chemical and biological parameters (e.g., temperature, salinity, O 2 , H 2 S, N 2 O, CH 4 , pH, nutrients, chlorophyll, primary productivity, phytoplankton community and bacterial counts) are monitored under this program.
This region is strongly influenced by the monsoons, i.e., the biannual reversal of winds and surface currents.Moderate upwelling occurs along the west coast of India during June-November.Due to the unique hydrographical and biogeochemical changes induced by the southwest monsoon, a reducing condition develops over the inner shelf that gradually extends to the mid-shelf north of 12 • N (Naqvi et al., 2006).Dissolved O 2 progressively decreases to < 4 µM in July/August in subsurface and bottom waters.The system is initially buffered by high nitrate concentrations (∼ 20 µM) and it requires approximately one month before nitrate is fully denitrified and the system becomes truly anoxic (i.e., sulfate-reducing).During peak suboxia, NO − 2 accumulates up to 6 µM, coinciding with high N 2 O (180 nM).During the transition from suboxic to anoxic conditions (late August/early September), N 2 O has been found to accumulate in the water column in record high concentrations (∼ 0.8 µM; Naqvi et al., 2000).As anoxia intensifies during late September-October, O 2 falls below the detection limit, the subsurface and bottom waters turn sulfidic (7 µM H 2 S) and NO − 3 -and NO − 2 -depleted but enriched in NH + 4 (8 µM) and PO 3−  4 (2.17 µM) (Naqvi et al., 2006).However, substantial interannual variability in the duration and intensity of the O 2 deficiency has been observed (Naqvi et al., 2009).The intensity of the O 2 deficiency also varies both longitudinally and latitudinally at a given time.The western continental shelf is perhaps the only region where all three types of redox environments -i.e., normoxia, suboxia and anoxia -cooccur over the same segment of the shelf with such regularity.However, the sulfidic conditions that develop in the late monsoon are a recent phenomenon presumably caused by the recent intensification of O 2 deficiency on account of increased atmospheric DIN deposition and/or anthropogenic fertilizer runoff from land (Naqvi et al., 2006).During the northeast monsoon (November-February), the prevailing West India Coastal Current causes downwelling, which reoxygenates the water column over the shelf, thereby re-establishing the oxic regime that continues until May.Further details of seasonal variation in physical and biogeochemical features of this region have been provided by Naqvi et al. (2006Naqvi et al. ( , 2009)).
The western Indian shelf covers an area of 310 000 km 2 , of which the inner shelf (up to a depth 55-60 m) comprises A. K. Pratihary et al.: Benthic mineralization and nutrient exchange fine-grained silty clays that contain, in general, 1-4 % organic carbon (Rao and Wagle, 1997).Sediments in the northern part (north of 13 • N) of the inner shelf are composed of 61 % silt, 21 % clay and 18 % sand.The transect at 15 • N (close to the CaTS transect) in particular is dominated by fine-grained particles (∼ 72 % silt and ∼ 27 % clay) and has been found to be organic-rich (3.2 % C org ), with ∼ 22 % labile organic matter (carbohydrates and proteins) (Jacob et al., 2008).The outer shelf (at depth > 60 m) occupies around two-thirds of the total area (Mallik, 2008).Organic-poor (C org < 1 %), coarse, sandy sediments dominate the outer shelf region (Rao and Wagle, 1997;Jacob et al., 2008).The sedimentation rate over the inner shelf (at 45 m depth, off Goa) has been found to be 1.5 mm yr −1 (Kurian et al., 2009).The C : N ratios and δ 13 C values suggest that the organic carbon is of marine origin (Agnihotri et al., 2008).At a 30 m deep location (off Goa) on the inner shelf, Jayaraj et al. (2007) observed benthic fauna such as polychaetes, crustaceans, mollusks and others, of which polychaetes comprised 83 %.At a 43 m deep site (off Bhatkal) on the inner shelf, Ingole et al. (2010) also reported 43 taxa of benthic fauna such as polychaetes, crustaceans, oligochaetes, of which polychaetes comprised 85-95 %.Both these locations were dominated by silty-clayey cohesive sediments.

Incubation method
Over the last couple of decades, benthic respiration and nutrient fluxes have been measured mainly by benthic chambers/landers (in-situ) or by intact-core incubations (ex-situ) over continental shelf environments.Separating a sediment section from its natural setting may affect benthic faunal activity.The turbulence created within a chamber/lander or incubation core may differ from that found under natural conditions which may in turn change the thickness of the benthic boundary layer and O 2 penetration depth, thereby affecting the flux rates.Though in-situ incubation has been preferred over whole-core incubation in some cases (e.g., Hammond et al., 1985;Devol and Christiansen, 1993;Jahnke et al., 2000;Berelson et al., 2003), reliable flux data have been obtained through the latter method (e.g., Rysgaard et al., 1998;Hopkinson et al., 2001;Jahnke et al., 2005;Faganeli and Ogrinc, 2009;Lehrter et al., 2012) over continental shelves.Comparative studies such as those carried out by Miller-Way et al. (1994), Hammond et al. (2004) and Woulds et al. (2009) show that benthic fluxes measured through the chamber/lander and the intact-core incubation method in shallow marine environments such as continental shelves are similar in magnitude and direction with minimal systematic difference, even under anaerobic conditions, with only few exceptions.Apart from NO − 2 and PO 3− 4 fluxes, Miller-Way et al. (1994) did not observe significant variabil-ity of nutrient fluxes between the in-situ and ex-situ incubation over the Louisiana shelf.They attributed these differences to the increase in O 2 in the overlying water upon core recovery rather than to the inherent differences in two methodologies.They also argued that the differences in the in-situ and ex-situ methodologies cannot be solely attributed to the removal of sediment cores from sea bottom.Hammond et al. (2004) observed variability of fluxes between two methods over the outer shelf and near-anoxic slope on the California margin.Interestingly, they observed a similar degree of variability of fluxes within the replicate measurements in the case of both in-situ and ex-situ methods.This apparently shows the overriding effect of patchiness in causing the variability of fluxes rather than the methodological difference.Over the slope (nearly anoxic) of the Pakistan margin, Woulds et al. (2009) reported considerable differences in some nutrient fluxes through both the methods that were attributed to O 2 contamination in the whole-core incubations.However, significant differences in fluxes may appear in deeper environments (e.g., continental slope and abyssal plain), due to pressure effects on benthic microbial and faunal community metabolism.Due to logistic limitations, we chose to carry out a series of laboratory-based intact-core incubations as suggested by Hammond et al. (2004) while taking precautions to prevent O 2 contamination.

Sample collection
The sampling was carried out at station G5 during October 2005, when the bottom water was anoxic, and again in April 2006, when the water column was oxic.Water samples were collected at different depths for dissolved O 2 , H 2 S, N 2 O and nutrients using a 5 L Niskin sampler.Temperature and salinity were recorded by a Sea-Bird CTD profiler.The vertical profiles of the above parameters at this site are described in Naqvi et al. (2009).Bottom water (10 L) was also collected and kept in the dark at 4 • C, to be used for the core incubations.
For the incubation experiments, a sediment core was collected with minimal disturbance by divers using specially designed core liners (see Sect. 2.3).Three additional sediment cores, one each for determination of porosity, C org and pore water nutrients, were collected in 6.4 cm inner diameter (ID), 30 cm long acrylic core liners by means of a gravity corer.The core for pore water was covered with aluminum foil within 5-10 min of collection.The headspace was flushed with N 2 for 5 min (only during October), and the core was kept at 4 • C until pore water extraction.Burrows created by benthic fauna were conspicuous on the sediment surface during April, but no burrows were noticed during October.

Experimental setup
Specially fabricated acrylic core liners (ID 10.1 cm, length 48 cm, thickness 0.5 cm) were used for the experiment  (Fig. 2).The core liners were fitted with nylon caps at both the ends.The top nylon cap was attached to two 1.5 mm ID Teflon tubes through the holes, which was further attached to three-way Leur locks.The caps had nylon O rings to make the liners completely airtight.Two more Teflon tubes were attached through the holes roughly at the center of the liners.
One was used to pass air in the case of aerobic incubation and the other one was used for drawing water sample using a syringe.Two 60 mL syringes were fixed to the upper portion of the liner so as to contain the overlying water and replace the withdrawn water sample.The syringes were made impermeable to O 2 by wrapping with multiple layers of insulating tape.The syringes were tested for O 2 permeability before they were cut and fitted into the core liner.They were wrapped with tape, filled with low-oxygen water and kept in the dark for 5 days.Oxygen levels were checked 24-hourly.
No increase in O 2 was observed in the syringes; rather, O 2 decreased, confirming the impermeability of the syringes to O 2 .The average length of sediment column was ∼ 22 cm and was overlain by a water column of ∼ 1.9 L. A Teflon-coated magnetic stirrer was positioned at ∼ 10 cm above the sedi-ment surface and rotated at 60 rpm, ensuring effective mixing of the overlying water without causing resuspension of surficial sediments.The stirrer was remotely coupled with a bar magnet positioned outside the core liner and rotated by a motor.

Intact core incubations
After retrieval, the core was immediately covered with aluminum foil or black cloth and kept in a container filled with bottom water.After 2-3 h, the core was transferred to a temperature-controlled room.Later, when the suspended particles had settled (∼ 1-2 h), the overlying water was carefully siphoned off without disturbing the surficial sediment.Fresh bottom water was slowly introduced over the sediment with minimal resuspension, and the core was preincubated in the dark at in-situ temperature (±0.5 • C) for 12 h in order to acclimatize it to the laboratory conditions.During October, the preincubation was done anaerobically as the in-situ bottom water was anoxic; however, during April it was done aerobically to match the prevailing oxic condition of the bottom water at station G5.Three incubations each in October and April were carried out as listed in Table 1.
(i) Anoxic incubations in October.After preincubation of the sediment core, the overlying water was siphoned out and replaced with fresh bottom water (brought to laboratory temperature).The top was covered with the cap, the valves were closed and the incubation was started.
As explained earlier, the incubations were done anaerobically.In experiment 1, the core was kept airtight and the overlying water was purged with N 2 for 30 min to remove O 2 before the start of incubation.The incubation was carried out for 48 h.In experiment 2, the overlying water was not N 2 purged but the core was kept airtight and incubated for 56 h.Experiment 3 was done in a way similar to experiment 2 but with NO − 3spiked fresh bottom water as overlying water and incubated for 48 h.Since the bottom water at station G5 remained NO − 3 -depleted during this time, it was necessary to enrich the overlying water with NO − 3 to quantify benthic N loss.To begin with, the overlying water in experiments 2 and 3 was oxic; however, anoxia developed after 24-36 h with the O 2 concentration decreasing substantially or going below detection, followed by H 2 S accumulation.Conditions became similar in experiment 1 (after 6 h), experiment 2 (after 36 h) and experiment 3 (after 24 h), and fully anoxic incubation conditions were thereby replicated within portions of the three experiments.
(ii) Aerobic and anaerobic incubations in April.Since the bottom water during April was well oxygenated at the site, the core was successively incubated both aerobically and anaerobically.In experiment 4, the overlying water was aerated continuously by bubbling air through the tubing fixed at the center of the core.Since the air bubbles were introduced at ∼ 10 cm above the sediment-water interface, the air flow did not disturb the surficial sediment.One of the valves on the cap at the top was kept open to release the excess pressure created in the headspace due to the air influx and the core was incubated for 48 h.In experiment 5, the core was incubated aerobically during the initial 24 h, similar to experiment 4, and, later, made airtight and incubated anaerobically for another 60 h in order to study the effect of progressive shift from aerobic to anaerobic conditions on benthic exchange.Experiment 6 was conducted anaerobically up to 48 h but using NO − 3 -enriched fresh bottom water.Aerobic incubation was replicated by maintaining similar conditions in both experiment 4 and the first 24 h of experiment 5. Similarly, anaerobic incubation was also replicated when conditions became comparable during the last 60 h of experiment 5 and experiment 6.The animal burrows remained intact, and some live benthic fauna were seen throughout the aerobic incubations.

Sampling and analysis
The overlying water was sampled every 4-6 h with a 5 mL glass or plastic syringe through a tube attached to the middle portion of the core liner.As the sample was drawn, the plunger of the fixed syringe (on the upper part of the liner) was slowly pushed inward simultaneously so as to replace the withdrawn sample.Samples for dissolved O 2 were collected in a He-flushed 5 mL Hamilton gastight syringe, and fixed immediately by adding 100 µL of each of the Winkler's titration reagent solutions.After 30 min, the oxyhydroxide precipitate was dissolved by adding 200 µL of 50 % H 2 SO 4 (Grasshoff et al., 1983).The absorbance was immediately measured at 456 nm (Pai et el., 1993) with a Shimadzu UVvisible spectrophotometer with an analytical/method precision of ±0.03 µM.
Sulfide samples were collected in He-flushed 5 mL Hamilton gastight glass syringes and fixed by adding 100 µL each of the dimethyl-p-phenylene diamine and FeCl 3 .The ab-sorbance was measured at 670 nm by spectrophotometer (Grasshoff et al., 1983) with a resulting analytical precision of ±0.05 µM.
Five milliliters of sample for N 2 O analysis was collected using He-flushed 10 mL Hamilton gastight syringes in NO − 3 amendment experiments and immediately poisoned by adding 100 µL of saturated HgCl 2 solution.Nitrous oxide was analyzed with a Hewlett-Packard gas chromatograph equipped with an electron capture detector following the multiphase analysis method (McAuliff, 1971) with a precision of 4 %.

Processing of additional sediment cores
For porosity determinations, the cores were sectioned at 1 cm intervals and the slices were immediately weighed.After drying at 110 • C, the dry weight was taken in order to determine the pore water content (w).Ten grams of dried homogenized sediment was fully soaked with deionized water in a 25 mL calibrated and weighed bottle after removing the entrapped air bubbles.Porosity (φ) was expressed as the ratio of pore water volume to the total volume of sediment (sediment + pore water) and was determined as follows: where G (specific gravity) is (W 2 -W 1 )/[(W 4 -W 1 )-W 3 -W 2 )], W 1 the weight of the empty bottle, W 2 = the weight of the bottle with dried sediment, W 3 = the weight of the bottle with sediment and water, and W 4 = the weight of the bottle with water.
For sedimentary C org , the core was sectioned at 1-2 cm intervals and the subsamples were freeze-dried, homogenized and subsequently decalcified using 1N HCl.C org was measured using an IR-MS (Delta V plus, ®Thermo) coupled with an elemental analyzer (EA) in a continuous flow mode.Calibrations were carried out using a laboratory standard (n-caprioic acid, C 6 H 15 NO 2 -ACA) following Higginson and Altabet (2004).Analytical precision of the analysis was found to be ≤ 2 %.The details of the analytical procedure are given in Agnihotri et al. (2008).
For pore water extraction, the core was sectioned at 1 cm intervals and collected in 50 mL centrifuge tubes inside a N 2flushed glove box.These were centrifuged at 4000 rpm for where F is the flux in mmol m −2 d −1 and ∂C/∂T h is the slope of the plot between concentration and sum of time / height.In the case of nutrients, very good linear correlation was observed between concentration and day m −1 (r 2 > 0.9) and thus the slope of linear regression fit was directly taken as the flux.However, in the case of O 2 , the slope of the polynomial fit at time zero (∂C/∂T h t = 0) was considered (Hammond et al., 2004) since curvature appeared usually after a concentration of ∼ 100 µM.Positive flux (efflux) values denote flux into overlying water from sediment and negative flux (influx) values stand for fluxes into the sediment from overlying water.Mean fluxes were calculated from replicate experiments and are presented with standard deviation in Table 2.
A concentration gradient existing between sediment pore water and bottom water results in the molecular diffusion of nutrients across the sediment-water interface.Diffusive flux can form a significant part of the total measured flux and was calculated following modified Fick's first law of diffusion appropriate for sediment (Berner, 1980): where F is the diffusive flux in mmol m −2 d −1 , φ the porosity of the surficial sediment and (∂C/∂z) z=0 is the concentration gradient (in mmol m −3 m −1 ) close to sediment-water interface.Depending on the solute profile, (∂C/∂z) z=0 was calculated differently (Dennis and Grenz, 2003).For NO − 3 and NO − 2 , the concentration difference between bottom water and pore water of the top sediment (0-1 cm) was considered as the gradient.In the case of NH + 4 , PO 3− 4 and SiO 4− 4 , the bottom water concentration and pore water concentrations up to 3 cm gave a linear fit (r 2 > 0.9) versus depth.Thus, the linear gradient over the upper 3 cm of the sediment was taken as (∂C/∂z) z=0 for these nutrients.
D s is the whole sediment diffusion coefficient (m 2 d −1 ) after correction for temperature and tortuosity (Krom and Berner, 1980): where D 0 is the molecular diffusion coefficient in seawater (Boudreau, 1997) at a particular temperature and θ is the tortuosity of the sediment and is expressed as θ 2 = 1-ln φ 2 (Boudreau, 1996).All the flux values are presented in Table 2.

Statistical analysis
Analysis of variance (i.e., ANOVA) was used to assess the variability of flux patterns between different experiments.

Oxygen consumption rates
Experiment 1 started with 23 µM of O 2 at time zero but rapidly went below the detection limit within 6 h and was not detectable until the end of the experiment (Fig. 3a).However, overlying waters in experiments 2 and 3 were oxic (O 2 > 158 µM) at the beginning but became anoxic after 36 h as O 2 decreased substantially or fell below the detection limit with subsequent sulfide accumulation (Fig. 3a, b).No significant variation (p = 0.5) was observed in the O 2 decrease patterns of experiments 2 and 3, and consistent O 2 flux was observed in experiment 2 (−48.07 mmol m −2 d −1 ) and experiment 3 (−47.57mmol m −2 d −1 ).Mean fluxes were calculated from replicate experiments and are presented with standard deviation (Table 2).
During April, the overlying water remained oxic (O 2 > 183 µM) in experiment 4 (Fig. 4a).Thus, the thin upper oxic layer of the sediment column, as expected, remained intact and did not affect the nutrient exchange.This condition was close to the naturally existing condition during the premonsoon, when the bottom water remains oxygenated (O 2 > 82 µM).In experiment 5, O 2 remained almost constant for the first 24 h, decreasing substantially thereafter to 36 µM at 48 h and remaining at that level until the end of the experiment (Fig. 4a).In experiment 6, O 2 decreased nonlinearly from 193 to 29 µM within 24 h and thereafter remained constant until 48 h (Fig. 4a).A nonlinear decrease in O 2 was also noticed in experiments 2, 3 and 5, similar to observations elsewhere by Sundby et al. (1986), Anderson et al. (1986), Malecki et al. (2004) and Skoog and Arias-Esquivel (2009).These have been attributed to (i) a decrease in the concentration gradient across the benthic boundary layer and (ii) a decrease in the thickness of surficial oxic layer and subsequently a decrease in the number of O 2 -respiring microorganisms (Sundby et al., 1986).The O 2 decrease pattern in experiment 6 did not vary significantly (p = 0.9) from that of experiment 5, and the O 2 fluxes observed in experiment 5 (−64.7 mmol m −2 d −1 ) and experiment 6 (−68.21mmol m −2 d −1 ) were consistent (Table 2).A noticeable feature was that unlike the October incubations, O 2 was not completely consumed and 20-23 µM was still present until the end of incubation.

Sulfide fluxes
Sulfide accumulated in all the incubations during October.In experiment 1, sulfide was detectable just after O 2 became zero at 6 h and increased at the rate of 4.3 mmol m −2 d −1 after 12 h (Fig. 3b).In experiment 2, S 2− was undetectable until 28 h but increased linearly at the rate of 1.95 mmol m −2 d −1 up to 48 h and then rose exponentially to 19.67 µM (Fig. 3b).Similar flux patterns and rates (2.24 mmol m −2 d −1 ) were also observed in experiment 3 after the O 2 level decreased substantially (Fig. 3b).The variation in S 2− flux patterns between these three experiments was statistically insignificant (p = 0.5).No sulfide was detected in the overlying water in experiments 4, 5 and 6 (not shown in the figures), which was to be expected, as at least 20 µM of O 2 was still present at the end of the experiment.The absence of benthic sulfide flux during normoxia, as well as sulfide release during anoxia, was also reported by Hansen and Blackburn (1991), McCarthy et al. (2008), Bartoli et al. (2009) and Faganeli and Ogrinc (2009) in other coastal systems.
During April, N species exhibited different exchange behaviors in aerobic and anaerobic incubations.In experiment 4, NO − 3 increased from 6.5 to 11.24 µM.Nitrite initially decreased until 18 h and then increased to 0.42 µM at the end (Fig. 4b) of the experiment.Similarly, NH + 4 decreased initially until 24 h and then showed an increasing trend (Fig. 4d).Both PO 3− 4 and SiO 4− 4 (Fig. 4e, f) increased linearly.In experiment 5, NO − 3 increased until 36 h and decreased thereafter (Fig. 4b).Nitrite initially increased to 0.34 µM at 60 h before registering a steep decrease (Fig. 4c).Ammonium increased in a distinct manner: a steady increase until 36 h followed by a sharp rise (Fig. 4d).In the first half of the experiment, lower NH + 4 efflux accompanied by simultaneous increases in NO − 2 and NO − 3 influx was observed.However, as O 2 decreased further, NH + 4 flux increased, while the NO − 3 and NO − 2 fluxed downward.Ammonium flux was 0.92 mmol m −2 d −1 until 36 h and increased subsequently to 3.22 mmol m −2 d −1 .Nitrate and nitrite flux initially increased at the rate of 0.57 mmol m −2 d −1 and later decreased at a rate of −0.58 mmol m −2 d −1 at 36 h.The initial (0-36 h) NO − 3 + NO − 2 efflux rate in experiment 5 was similar to the rates observed in experiment 4. In experiment 6, NO − 3 exhibited a negative flux, decreasing linearly (Fig. 4b).Nitrite remained almost constant until 36 h; thereafter it rose steeply to peak at 36 h and decreased subsequently.A prominent N 2 O peak coincided with the NO − 2 peak (Fig. 4c).Higher and faster NO − 2 buildup was observed in experiment 6 as compared to experiment 3, but rapid (subsequent) decreases were observed in both cases.A similar pattern of NO − 3 decrease and accumulation of NO − 2 and N 2 O was also reported by Naqvi et al. (2000), indicating denitrification.The highest NO − 3 + NO − (after 36 h) and experiment 6 (p = 0.8), but both varied significantly (p < 0.05) from that in experiment 4. Phosphate and SiO 4− 4 showed positive fluxes in all the experiments (Fig. 4e, f), without any significant variation (p > 0.05) in the flux pattern between the three experiments.Significant variation (p < 0.001) in NH + 4 flux was found between aerobic incubations in April and anaerobic incubations in October.Phosphate (p< 0.001) and SiO 4− 4 (p < 0.05) flux patterns in October significantly varied from those in April.

N 2 O production
Transient buildup of nitrous oxide, coinciding with the peak in NO − 2 , was observed in both the seasons (Figs.3d and  4c) in the NO − 3 -amended experiments (experiments 3 and 6), which indicated active benthic denitrification.Nitrous oxide accumulation in October was slower relative to that in April, but the decrease was rapid after peaking at 36 h, which results from its further reduction to N 2 .No significant variation (p > 0.05) was observed in the N 2 O accumulation patterns of October and April incubations.Nitrous oxide fluxes between October (8.85 µmol m −2 d −1 ) and April (10.45 µmol m −2 d −1 ) were also comparable (Table 2).This shows that only a minor percentage of NO − 3 influx (0.2 % in October and 0.9 % in April) had resulted in N 2 O.

Discussion
In the present study we have simulated the biogeochemical conditions prevailing at station G5 through a series of intactcore incubations.In the course of the experiments, we manipulated O 2 and NO − 3 levels in order to represent the seasonal conditions during oxic and anoxic regimes.We assume that the changes in O 2 and nutrient concentrations over a few days in the experiments would mimic the natural variability occurring over the inner shelf for a longer period.Both aerobic and anaerobic incubations were duplicated during April and October except for the NO − 3 -amendment experiments.Although incubation experiments may not fully represent the natural condition (flow regime and faunal activity), recent studies reveal that ex-situ experiments can be a powerful tool for flux studies in shallow environments.Here we assume that the benthic fluxes determined through the ex-situ method reflect the combined effects of molecular diffusion, faunal activity and advective transport processes.

Sediment oxygen consumption (SOC) and sulfide oxidation
The anaerobic incubations conducted during April resulted in a consistent O 2 influx rate, giving an average SOC of 66.4 mmol m −2 d −1 (Table 2).However, in none of these experiments was O 2 completely consumed, as 20-23 µM O 2 was still present at the end.Also, the conditions were not anoxic, as sulfide was not detected at any time point.Dur-ing the first 36 h, experiment 5 was an aerobic experiment showing similar nutrient exchange patterns and flux rates as observed in experiment 4 (discussed in Sect.5.2).Although this experiment was continued for 3.5 days, sulfide was never detected at any time, which implied that the upward-diffusing S 2− was retained within the sediment, being directly or indirectly oxidized by O 2 from the overlying water.
In contrast, complete O 2 consumption followed by sulfide release was observed in the October incubations.Oxygen uptake occurred at an average rate of 47.8 mmol m −2 d −1 during this time (Table 2).Similar seasonal changes in SOC have also been observed in the Corpus Christi Bay (Mc-Carthy et al., 2008), Orbetello Lagoon (Bartoli et al., 2009) and the Gulf of Trieste (Faganeli and Ogrinc, 2009).SOC has been observed to be enhanced by macrofaunal activity (Kristensen, 2000;Nizzoli et al., 2007;Braeckman et al., 2010) and higher temperature (Hargrave, 1969;Thamdrup et al., 1998;Rowe et al., 2002;Forja et al., 2004).Average bottom water temperature at the present site varies from 28 • C during April to 22 • C during October (Naqvi et al., 2006).Thus, the observed 28 % decrease in SOC during October could be due to the prevalence of low temperatures and the absence of benthic fauna (Naqvi et al., 2006;Ingole et al., 2010).Neubacher et al. (2011) observed a 50 % decrease in SOC during hypoxia in the North Sea and Rowe et al. (2002) reported a 5-fold lower SOC under anaerobic conditions compared to aerobic conditions in the Gulf of Mexico.
Since the bottom water remains oxygenated in April (Naqvi et al., 2006), the surface sediment remains oxic (although O 2 penetration may only be to a few millimeters deep in coastal sediments; Revsbech, 1989).In fact, a ∼ 0.5 cm thick brown layer (Fe-and Mn-oxide-rich) was conspicuous at the surficial sediment during April but was absent during October.Sediments rich in Fe and Mn can be an effective trap for sulfide in coastal environments (Kristiansen et al., 2002).In such conditions, most of the upward-diffusing sulfide gets oxidized by Fe and Mn oxides in the top oxic layer which are, in turn, produced by oxidation of Fe 2+ and Mn 2+ by the downward-diffusing O 2 from the overlying water (Kristensen, 2000).The sediments at station G5 contain a high percentage of Fe (7.7 %) and Mn (0.09 %) (S.Kurian et al., unpublished data), which indicates that sulfide could be retained in the oxic sediment layer during April, culminating in no benthic S 2− flux.However, as the anoxia sets in during late September, the oxic-anoxic interface moves upwards to the water column.Consequently, Fe and Mn are reduced and flux to the overlying waters (Kristiansen et al., 2002).Thus, S 2− is no longer oxidized by Fe-and Mn oxides, although a smaller portion may be trapped as pyrite (FeS 2 ).In Corpus Christi Bay, McCarthy et al. (2008) located the O 2 -H 2 S interface below 30 mm during normoxia, but it moved up to just below the sediment surface (0-7 mm) during hypoxia.As experiment 1 was preceded by anaerobic preincubation and N 2 purging prior to time zero, the surface sediment layer must have been anoxic.Thus, the sulfide could not be trapped and escaped to the overlying water soon after O 2 fell below the detection limit at 6 h, resulting in the highest flux.However, since the overlying water was initially oxic in experiments 2 and 3, a substantial part of upward-diffusing sulfide was apparently oxidized to SO 2− 4 by Fe oxides, Mn oxides and O 2 in the top oxic sediment layer.Following a significant drop in O 2 , there was an immediate sulfide release after 6 h in experiment 1 and after 18-28 h in experiments 2 and 3.This implied that S 2− underwent substantial oxidation by available O 2 in the overlying water and was released at a lower rate (2.09 mmol m −2 d −1 ).Experiments 2 and 3 nearly represent the condition during late August-early September, where S 2− was released but was oxidized at a rate of 2.2 mmol m −2 d −1 .Experiment 1 is similar to the natural anoxic condition observed during late September-October (Fig. 5b), where S 2− was released at a higher rate without being oxidized.
Most of the O 2 is not consumed in decomposition of organic matter in marine sediments but rather in reoxidation of reduced inorganic metabolites (e.g., Mn 2+ , Fe 2+ , S 2− , NH + 4 and CH 4 ) close to the oxic-anoxic interface, being chemically or microbially mediated (Kristensen, 2000).However, reoxidation of sulfide has quantitative significance especially during anoxia, when pools of Mn oxides and Fe oxides are depleted.Up to 85 % of S 2− produced by SO 2− 4 reduction in the anoxic zone of sediment is reoxidized in the oxic zone of sediments (Thamdrup et al., 1994), and > 50 % of SOC occurs directly or indirectly by oxidation of sulfide (Jørgensen, 1989).
In experiment 3, sulfide oxidation could also be fueled by NO − 3 addition in a process known as chemolithoautotrophic denitrification (Lavik et al., 2009), which will be discussed further in Sect.5.2.Experiment 1 nearly represents the anoxic conditions prevailing during October and experiment 3 represents the biogeochemical conditions during late August-early September, when the transition from suboxia (denitrifying condition) to anoxia occurs.Sulfide accumulates up to 13 µM in the subsurface and bottom water in the shelf water off Goa during September-October (Naqvi et al., 2006) and disappears by mid-November owing to its oxidation at the oxic-anoxic interface by downwelled oxic water during November.

Nitrification, denitrification and net DIN flux
During April, considerably lower NH + 4 flux in the aerobic incubation (experiment 4) compared to anaerobic incubations (experiments 5 and 6) is mainly due to nitrification that occurred in the top oxic sediment layer (Fig. 5a).This is further evidenced by positive fluxes of NO − 2 and NO − 3 and positive correlation between NH + 4 and NO − 2 (r = 0.9, p < 0.001) as well as NO − 2 and NO − 3 (r = 0.96, p < 0.001).Similar lower  NH + 4 flux under oxic conditions has also been reported elsewhere by Malecki et al. (2004), McCarthy et al. (2008) and Faganeli and Ogrinc (2009).
Experiment 5 represented a transition from oxic (0-36 h) to near-suboxic (denitrifying) conditions (36-84 h), which is well reflected in nutrient flux patterns.Thus, benthic nitrification was active until 36 h, resulting in lower NH + 4 flux and positive NO − 2 and NO − 3 fluxes.As observed in experiment 4, nitrification was also evidenced by positive correlation between NH + 4 and NO − 2 (r = 0.92, p < 0.01) and NO − 2 and NO − 3 (r = 0.97, p < 0.001).This is further substantiated by the fact that the NH + 4 and NO − 3 +NO − 2 fluxes from 0 to 36 h were consistent with those from experiment 4 (Table 2).As the O 2 decreased substantially after 36 h, inhibition of nitrification led to a 3.5-fold increase in NH + through benthic nitrification from 0 to 30 h.Higher NH + 4 anaerobic fluxes in experiment 5 after 40 h (3.22 mmol m −2 d −1 ) and experiment 6 after 30 h (3.55 mmol m −2 d −1 ) were apparently due to suppressed benthic nitrification owing to prevailing low-O 2 conditions as also observed by Hansen and Blackburn (1991) and Faganeli and Ogrinc (2009) in other regions.Nitrate and nitrite showed higher influx compared to that observed in experiment 5 (after 30 h) as NO − 3 addition enhanced benthic denitrification to a rate of 1.04 mmol m −2 d −1 .
During April, NH + 4 fluxes in experiment 4 and initial NH + 4 fluxes in experiments 5 and 6 were consistent, with a mean aerobic NH + 4 flux of 0.83 mmol m −2 d −1 (Table 2).Similarly, considering experiments 4 and 5 (0-24 h), the average NO − 3 +NO − 2 efflux rate is 0.57 mmol m −2 d −1 .Mean anaerobic NH + 4 flux was 3.38 mmol m −2 d −1 considering experiments 5 and 6.It is conspicuous that the anaerobic conditions resulted in an NH + 4 flux of 3.38 mmol m −2 d −1 , but it reduced to 0.83 mmol m −2 d −1 in the aerobic incubation owing to a 75 % NH + 4 loss at a rate of 2.55 mmol m −2 d −1 through benthic nitrification (Fig. 5a).However, the lost NH + 4 did not appear as NO − 2 and NO − 3 at the same rate but rather at a lower rate (0.57 mmol m −2 d −1 ) in the aerobic incubation.This implies that 58 % of upward-diffusing NH + 4 was apparently lost in coupled nitrification-denitrification occurring at the oxic-anoxic interface in the surficial sediment (Jenkins and Kemp, 1984;Risgaard-Petersen, 2003;Granger et al., 2011) and 17 % was released to the overlying water being nitrified to NO − 2 and NO − 3 .Consequently, only 25 % N is released to the overlying water as NH + 4 .This shows that benthic denitrification coupled with nitrification causes N loss at a rate of 1.98 mmol m −2 d −1 at the sedimentary oxic-anoxic transition zone (Fig. 5a) irrespective of NO − 3 content of the overlying oxic bottom water, and 78 % of benthic nitrification was coupled with denitrification, resulting in considerable N loss.
Since experiment 4 represents the prevailing aerobic conditions (Fig. 5a) during April, the net benthic DIN release to the overlying water is 1.4 mmol m −2 d −1 , of which NH + 4 flux comprises 59 %.Experiment 6 represents the biogeochemical condition of the site during June-July, when the bottom water becomes suboxic (denitrifying) and NO − 3 -rich (∼ 12 µM).The benthic denitrification rate varied from 0.58 to 1.04 mmol N m −2 d −1 based on two anaerobic incubations (last 60 h of experiments 5 and 6).Considering methodological uncertainties, our results are consistent with Naik and Naqvi (2002) and Naqvi et al. (2006), who reported a sedimentary denitrification rate of 0.23-1.25 mmol m −2 d −1 in this region during peak suboxia.Based on NO − 3 consumption, Schwartz et al. (2009) reported a benthic denitrification rate 0.51-1.45mmol N m −2 d −1 on the Pakistan margin during April-May, which is also in agreement with our rates.Sokoll et al. (2012) also observed a benthic denitrification rate of 0.55 mmol N m −2 d −1 by intact-core incu-bation on slope sediments off Pakistan.Oxygen concentration largely determined the NO − 3 source for benthic denitrification.Our observation is similar to that reported by Rysgaard et al. (1994), who noticed that under aerobic conditions, NO − 3 generated through nitrification within the sediment fed denitrification, while, under anaerobic conditions, NO − 3 diffused from the overlying water was found to be the main source for denitrification.Dominance of denitrification and low anaerobic ammonium oxidation (anammox) activity in the suboxic (denitrifying) water column and sediments was observed by Jayakumar et al. (2009) and Sokoll et al. (2012) within the oxygen minimum zone (OMZ) of the eastern Arabian Sea margin environments.
During October, influxes of NO − 3 and NO − 2 were observed in all experiments, and NH + 4 effluxes were consistent in all incubations.Although the overlying water was initially oxic, nitrification apparently did not take place due to the inhibition by upward-diffusing free sulfide (Joye and Hollibaugh, 1995).Moreover, during October, nitrifiers were not detected in the shelf bottom water, whereas these were found in large numbers during April (Krishnan et al., 2008).Since heterotrophic denitrification and anammox are also inhibited by sulfide (Jensen et al., 2008;Aelion and Warttinger, 2010), the observed loss of NO − 3 +NO − 2 seems to be caused by other processes such as dissimilatory nitrate reduction to ammonium (DNRA) and autotrophic denitrification.As NH + 4 flux did not increase significantly in experiment 3, DNRA seems unlikely.In fact, very low DNRA activity (Sarkar et al., unpublished data) and, interestingly, significant autotrophic denitrification (S.W. A. Naqvi et al., unpublished data) have been observed in the sulfidic bottom water along this transect during late September.Nitrate is reduced to N 2 through sulfide oxidation mediated by chemolithoautotrophic bacteria such as Thiobacillus denitrificans, Thiomicrospira, Thioploca, Magnetospirillum and filamentous Beggiatoa, with transient buildup of intermediates NO − 2 , NO and N 2 O (Cordoso et al., 2006;Robertson and Kuenen, 2006).Autotrophic denitrification has been found in several coastal environments (Brettar and Rheinheimer, 1991;Brüchert et al., 2003, Dong et al., 2011;Shao et al., 2011).Chemolithoautotrophic bacteria are often found in sulfidic surficial sediments (Lam and Kuypers, 2011) and are responsible for extensive N loss in anoxic (sulfidic) shelf systems (Lavik et al., 2009).Jørgensen and Nelson, 2004) In fact, Krishnan et al. (2008) reported the presence of a large number of sulfide-oxidizing denitrifiers (Thiobacillus denitrificans like organisms) in the bottom waters of the shelf off Goa during both April and October.Moreover, concomitant buildup of NO − 2 and N 2 O in experiment 6 substantiates the fact that the N-loss pathway in October was autotrophic denitrification, in contrast to heterotrophic denitrification, which caused N loss in non-sulfidic bottom water and surface sediments during July-August.Nitrate addition in experiment 3 stimulated autotrophic denitrification to a rate of 3.21 mmol m −2 d −1 (Fig. 5b).Similar enhancement in N loss due to autotrophic denitrification was also reported by Zhang et al. (2009), Shao et al. (2011) and Dong et al. (2011) in sulfidic estuarine and coastal sediments in Southeast Asia.As Thioploca (a known chemolithotrophic S 2− oxidizer) mats have been sighted on Pakistan margin sediments (Schmaljohann et al., 2001), chemoautotrophy apparently plays a significant role in benthic biogeochemical processes there (Cowie, 2005) and also perhaps in other parts of the Arabian Sea.As experiments 1 and 2 represent the nitrate conditions typical of the late monsoon period (October), the average autotrophic denitrification rate would be 0.52 mmol N m −2 d −1 (Table 2).However, experiment 3 approximately represents the biogeochemical conditions some time during late August to early September (Fig. 5b), when the transition from suboxia to anoxia takes place.Bottom water remains close to anoxic with modest NO − 3 concentration (8-10 µM).Though sulfide was not detected during this time, it is highly possible that the released sulfide is quickly oxidized by NO − 3 and NO − 2 , creating an impression of non-sulfidic conditions (Lavik et al., 2009;Canfield et al., 2010).Naqvi et al. (2000) reported the highest N 2 O buildup (765 nM) during this transition time.Nitrous oxide accumulation and subsequent consumption in experiment 3 further indicates denitrification, which occurs under sulfidic conditions.Our results show that autotrophic denitrification can potentially cause N loss at a rate of 3.21 mmol N m −2 d −1 during this time (Table 2).As the bottom water remains NO − 3 -depleted during October, NH + 4 was the only available form of DIN being released at an average rate of 3.74 mmol m −2 d −1 (Fig. 5b).Overall, NH + 4 comprised 59-100 % of the benthic DIN release.The noticeable seasonal change was that the sediment switches from being a NO − 3 source in the oxic period to a NO − 3 sink in the anoxic period.During April, the major N loss was through heterotrophic denitrification, whereas autotrophic denitrification dominated during October.

Benthic PO 3− 4 and SiO 4− 4 release
Benthic PO 3− 4 flux is mainly controlled by the interplay of processes such as remineralization from sedimentary organic matter, adsorption to Fe oxides, co-precipitation with ferric minerals (Krom and Berner, 1981) and the formation of authigenic carbonate fluorapatite (Ruttenberg and Berner, 1993) in marine sediments.Positive fluxes of PO 3− 4 in all experiments during April and October show that the benthic PO 3− 4 regeneration rate always exceeds the rate of its adsorption and precipitation in the upper oxic layer.During April, PO 3− 4 flux did not vary significantly between the aerobic and anaerobic incubations.It is evident that O 2 in the anaerobic incubation (experiment 6) did not decrease to a level that caused anoxia in the upper sediment layer.As ∼ 20 µM O 2 was present at the end of experiment 6, the upper benthic layer was not truly anoxic, which could not cause reductive dissolution of Fe oxides and, in turn, release of Fe-bound PO 3−  4 .Ferric minerals such as FeOOH and Fe 2 O 3 are known to adsorb and bind PO 3− 4 chemically in the oxic sediment layer (Krom and Berner, 1981).Phosphate was released to overlying water at an average rate of 0.14 mmol m −2 d −1 during April (Fig. 5a).Similarly, SiO 4− 4 effluxes were similar under both aerobic and anaerobic incubations, with a mean flux of 3.83 mmol m −2 d −1 (Fig. 5a).
Phosphate fluxes increased significantly in October incubations, as the conditions became anoxic in all three incubations.Anoxia-induced reductive dissolution of FeOOH apparently released PO 3−  4 (Ingall and Jahnke, 1994; Gunnar  and Blomqvist, 1997; Slomp et al., 1998) from Fe minerals and ultimately resulted in higher PO 3− 4 fluxes.Higher benthic PO 3− 4 release was also observed by Rozan et al. (2002), Bartoli et al. (2009) and Skoog and Arias-Esquivel (2009) during anoxia in other coastal systems.Average PO 3− 4 flux in the October incubations was 0.59 mmol m −2 d −1 (Fig. 5b), which is more than 4-fold higher than that in the April incubations (Fig. 5a).Faganeli and Ogrinc (2009) observed 3-10-fold higher PO 3− 4 flux under anoxic conditions compared to oxic conditions in the Gulf of Trieste.Skoog and Arias-Esquivel (2009) also reported 26 times higher benthic PO 3− 4 release under anoxic conditions than under oxic conditions in Long Island Sound.A 36-50 % increase in benthic PO 3− 4 efflux was noticed by Bartoli et al. (2009) during anoxia in Orbetello Lagoon, Italy.The SiO 4− 4 flux during October is ∼ 1.3 times higher than that observed during April (Table 2).Diatoms remain the dominant phytoplankton in the shelf water off Goa during the southwest monsoon (Parab et al., 2006;Roy et al., 2006), due to higher NO − 3 availability through coastal upwelling.In contrast, N 2 -fixing Trichodesmium dominate the phytoplankton community, with frequent patchy blooms over the shelf, during February-March (Parab et al., 2006;Roy et al., 2011).Higher diatom growth during the southwest monsoon could result in a higher downward flux of siliceous frustules to the underlying sediments during the late monsoon.Dissolution of diatom frustules could have enriched pore water with SiO 4− 4 and led to higher fluxes in October incubations (Fig. 5b).Similar enhancement in SiO 4− 4 flux has also been reported by Grenz et al. (2000) in San Francisco Bay.

Stoichiometry of fluxes
The aerobic mineralization of planktonic organic matter can be expressed in the following equations: The flux of biogenic elements is expected to occur in the Redfield ratio, but interplay of several processes such as nitrification, denitrification, adsorption, desorption, and macrobenthic activities may cause deviation from the expected ratio (Koop et al., 1990;Cowan and Boynton, 1996;Giblin et al., 1997;Ferguson et al., 2004;Faganeli and Ogrinc, 2009) (Jenkins and Kemp, 1984) and NH + 4 adsorption onto clay minerals (Rosenfeld, 1979;Mackin and Aller, 1984).Decrease in the NH + 4 : PO 3− 4 flux ratio (5.2-7.3)during the October incubation is due to anoxia-induced higher PO 3− 4 release.Closeness of the NH + 4 : SiO 4− 4 flux ratio (0.8-1.02) to the Redfield-Brzezinski value (1.06) during the April incubations suggests that it would favor diatom growth in the water column.However, the lower the NH + 4 : SiO 4− 4 flux ratio during the October incubations (0.55-0.78) apparently would favor the growth of non-siliceous phytoplankton during the late monsoon.This is substantiated by the fact that diatoms remain the dominant community during April-May, while dinoflagellates become the major phytoplankton during September-October (Parab et al., 2006, Naqvi et al., 2009).

Organic matter mineralization
Sediment oxygen consumption is usually considered to be one of the principal determinants of the oxidation rate of total C org in marine sediments (Canfield et al., 1993;Glud, 2008).Dissolved inorganic carbon (T CO 2 ) flux has been observed to be fairly linear with SOC (Giblin et al., 1997;Hopkinson et al., 2001;Ferrón et al., 2009b), which indicates that O 2 is directly or indirectly responsible for benthic respiration.Oxygen directly oxidizes C org in aerobic mineralization but reoxidizes only the reduced inorganic products of anaerobic mineralization, e.g., Fe 2+ , Mn 2+ and H 2 S fluxing into the top oxic zone (Canfield et al., 1993).However, since reduction of Mn oxides, Fe oxides and SO 2− 4 and subsequent reoxidation of Fe 2+ , Mn 2+ and H 2 S by O 2 have the same net stoichiometry as in C org oxidation by O 2 (Jørgensen, 1977;Canfield et al., 1993), SOC can be used to calculate the C org mineralization rate.The total C org oxidation rate is quantified by taking O 2 consumption in aerobic nitrification and carbon oxidation through denitrification into consideration (Canfield et al., 1993) as follows: where T C oxid is the depth-integrated rate of total C org oxidation, O 2 NH4-oxid is the O 2 consumed in NH + 4 oxidation to NO − 3 and C denit is the C org oxidized through denitrification.In aerobic mineralization, C : O 2 is 1, O 2 : NH + 4 ratio in complete nitrification is 2, and C : N in heterotrophic denitrification is 1.25 (Canfield et al., 1993).
During April, considering the nitrification rate of 2.55 mmol m −2 d −1 , the O 2 consumption for NH + 4 oxidation will be 5.1 mmol m −2 d −1 , while the benthic denitrification (coupled to nitrification) of 1.98 mmol N m −2 d −1 would oxidize C org at the rate of 2.47 mmol m −2 d −1 under the oxic conditions.This leads to a net C org oxidation rate of 63.8 mmol C m −2 d −1 .However, as both nitrification and heterotrophic denitrification are inhibited by sulfide during October, O 2 flux can be directly considered as the C org oxidation rate (Canfield et al., 1993).Considering experiments 1 and 3, the mean C org oxidation rate would be 47.8 mmol C m −2 d −1 during the sulfidic event.Organic carbon mineralization has been found to be faster under aerobic and compared to anaerobic conditions (Benner et al., 1984;Henrichs and Reeburgh, 1987;Lee, 1992;Dauwe et al., 2001), especially where the sediment has more refractory organic matter, although aerobic and anaerobic degradation rates can be similar in some cases (Kristensen et al., 1995;Hulthe et al., 1998).Thus, during April, 63.8 mmol of C org were mineralized aerobically square meter in a day, while 47.8 mmol of C org was anaerobically mineralized during October.It appears that the 25 % decrease in the C org mineralization rate from oxic to anoxic regime arises due to a slower oxidation of C org under a stratified water column during October.Neubacher et al. (2011) observed a 46 % decrease in N remineralization in the North Sea under hypoxic conditions.Similarly, Dauwe et al. (2001) reported 50 % less anoxic C org mineralization than oxic mineralization in the North Sea.Higher temperatures (Kristensen et al., 1992;Middelburg, 1996;Sanz-Lázaro et al., 2011) and benthic faunal activity (Rysgaard et al., 2000;Heilskov and Holmer, 2001;Nascimento et al., 2012) are known to stimulate remineralization of C org in marine sediments.As porosity and C org did not change significantly between April and October (see Sect. 4.6 and Fig. 6), the prevailing low temperature (22 • C) and the absence of benthic macrofaunal activity (Naqvi et al., 2006;Ingole et al., 2010) apparently slowed down C org mineralization at the study site during anoxia.

Porewater nutrients and diffusive fluxes
The porosity of surficial sediment did not vary significantly (p = 0.2) between April and October (Fig. 6a).Vertical distribution of C org did not vary significantly (p = 0.6) between the oxic and anoxic regimes (Fig. 6b).Porewater at the study site was enriched with NH + 4 , PO 3− 4 and SiO 4− 4 compared to overlying water column during both the seasons (Fig. 6d, e and f).Vertical distribution of pore water NH + 4 and SiO 4− 4 did not vary significantly (p > 0.05) between April and October.However, NO − 3 + NO − 2 and PO 3− 4 varied significantly (p < 0.05) between two seasons.As these profiles were from single core observations, the degree of spatial heterogeneity could not be ascertained.However, the effect of bioirrigation and PO 3− 4 immobilization/desorption by/from Feand Mn-oxide minerals also cannot be ruled out.Macrobenthic activity reduces due to sulfidic conditions (Rowe et al., 2002) during October, and anoxia causes reductive dissolution of Fe-and Mn oxides and release of PO 3− 4 to pore water.Similar seasonal variation in the pore water PO 3− 4 profile has been observed by Woulds et al. (2009) in the sediments on the Pakistan margin.Low NO − 3 and NO − 2 were observed in the pore water, which usually occurs because of N loss through benthic denitrification, as O 2 penetration is limited to the upper few millimeters of the sediment.Bioirrigation was conspicuous in the NO − 3 +NO − 2 profile in April (Fig. 6c), which apparently caused the penetration of the oxic bottom water to the deeper anoxic layers, thereby nitrifying the pore water NH + 4 to NO − 2 and NO − 3 .In contrast, low NO − 3 + NO − 2 in pore water was observed in October, due to anoxia-related decrease of benthic faunal activity as well as enhanced benthic denitrification.
The calculated diffusive fluxes of nutrients did not vary significantly between April and October except in the case of NO − 3 and NH + 4 .The diffusive NH + 4 flux was especially high, comprising 73-92 % of total benthic NH + 4 flux, because of a higher concentration gradient across the sediment-water interface and higher diffusivity of NH + 4 .The diffusion coefficient (D s ) of NH + 4 is 3.4 and 1.9 times higher than that of PO 3− 4 and SiO 4− 4 , respectively (Schulz, 2006).Flux enhancement factors (the ratio of measured flux to calculated flux) during the oxic period were 14.7, 1.3, 4.7 and 6.6 for NO − 3 + NO − 2 , NH + 4 , PO 3− 4 and SiO 4− 4 , respectively, and 11.6, 1.03, 9.2 and 8.2, respectively, during the anoxic period.There was clear evidence of benthic faunal activity in the aerobic incubations during April, as some burrows and live benthic organisms were conspicuous throughout the experiments.Interestingly, recent observations (October 2011 and2012;Badesab et al., unpublished data) showed the presence of live benthic macrofauna (e.g., polychaeta, bivalves and crustaceans) in the G5 sediments during October under sulfidic conditions.These adaptable species survive the anoxia and proliferate afterwards.Polychaetes proliferate even during anoxia.Thus, the benthic faunal activity apparently did not stop during October incubations.Stirred chambers can also cause advective pore water transport (Huettel and Gust, 1992), although its magnitude may differ from that in natural environments.As advective pore water transport depends on the bottom current flow and surface roughness (mounds, ripples etc), it is not necessary that it would occur at the same magnitude during both seasons.Although its role in pore water exchange may be smaller compared to benthic faunal activity (bioirrigation and bioturbation) in cohesive sediments, it may not be negligible.The stirring rate was kept the same during April and October, which possibly created a constant bottom flow.However, surface roughness could have increased during October, thereby increasing the advective transport.Thus, benthic faunal activity and advective pore water transport apparently enhanced the benthic nutrient exchange rate.

Benthic turnover time
The number of days required to replenish the pelagic nutrient inventory of a system by the benthic nutrient supply under steady state is referred to as the benthic turnover time.It is calculated by dividing the depth-integrated pelagic nutrient stock by the benthic nutrient efflux rate (Warnken et al., 2000).Considering experiment 4 and the aerobic part of experiment 5, the turnover time of NH + 4 is 16 days (Table 3) during April.However, when nitrification is taken into consideration, the turnover time of DIN is enhanced (∼ 67 days) during this period.In experiment 6, the turnover time of DIN (40 days) decreased due to inhibition of nitrification at low O 2 .As experiment 6 nearly represents the biogeochemical condition of the system during July-August, it is evident that prevailing anaerobic conditions during July and August will cause a quicker replenishment of the pelagic DIN stock by the benthic flux.Apparently, the turnover time of PO 3− 4 (165 days) and SiO 4− 4 (29 days) did not change significantly be-and inferences based on incubation experiments from the inner shelf may not represent the entire western shelf.Future studies focusing on both the inner and outer shelf might improve the understanding of the benthic mineralization and nutrient fluxes in this region.
coincided with negative fluxes of NO − 3 and NO − 2 due to benthic denitrification.This resulted in N loss at a rate of 0.58 mmol m −2 d −1 and NH + 4 fluxed out at a rate of 3.22 mmol m −2 d −1 without being nitrified.

Table 1 .
Details of the sediment-water flux incubations done in October 2005 and April 2006 at station G5, off Goa.

Table 2 .
Sediment-water flux rates of dissolved oxygen, nutrients, sulfide and nitrous oxide at station G5, off Goa, during April and October.Mean flux rate was calculated from replicate experiments and the standard deviation is expressed in parentheses.Diffusive fluxes of O 2 , H 2 S and N 2 O are not available as their pore water profiles could not be measured.Positive fluxes denote flux into overlying water from sediment and negative fluxes denote flux into the sediment from overlying water.All values are in mmol m −2 d −1 .