Dielmethane f lux from a subtropicaleutrophic pond in November based on continuousmonitoring

Cheng Zhang·Shenggao Cheng·Yuanzheng Li·W enli Zhang·Shangbin Xiao

Abstract A f ield campaign was carried out to investigate continuous diel methane(CH4)f lux from a subtropical eutrophic pond in November 2016.The diffusivemethane f luxof asinglemeasurement hadarangefrom 2.68×10-5 to 0.028mmol·m-2·h-1 w ith an average of 0.011±0.005mmol·m-2·h-1.The diffusivemethane f lux of measurements from 9:00 to 10:30 and from 21:00 to 22:30 were very close to the average diffusive f lux of all measurements.The bubblemethane f lux at different time measurementshadmuchmore variability than the diffusive methane f lux.The bubble methane f lux of a single measurementhad a range from 0 to 0.312mmol·m-2·h-1 with anaverage of 0.024±0.054mmol·m-2·h-1.For the eutrophic pond,the average bubble and diffusive CH4 f lux were0.56±0.18and0.26±0.04mmol·m-2·day-1,respectively,and the CH4 ebullition f lux accounted for 68.23%of the total f lux.Themaximum of the bubble CH4 f lux was about 4.6 times of them inimum CH4 ebullition.Themaximum of diffusive CH4 f lux was～1.7 times of the correspondingm inimum.The diffusivemethane f luxes in daytime and nighttimewere almostequal.However,the bubblemethane f lux in daytimewas0.029mmol·m-2·h-1,which was 1.6 times of that at night.W ind speed,the?Shenggao Cheng

Keywords Methane·Eutrophic pond·Diffusive·Ebullition·Flux

1 Introduction

Thoughtheareaof verysmall ponds onlyoccupies～8.6%of lakes and ponds on a worldw ide scale,they comprise～40.6%of the diffusive CH4eff lux(Holgerson and Raymond 2016).Furthermore,the CH4ebullition f lux from shallow lakesand pondsmay bemuch larger than thatby diffusive em issions(Walter et al.2006;Maeck et al.2013;Weyhenmeyer 1999;DelSontro et al.2016;Martinez and Anderson 2013).For example,ebullition accounted for 95%of the CH4f lux from the intensively thaw lakes in North Siberia(Walteretal.2006).In a shallow eutrophic temperate lake located in Canada,the ebullitive f lux to the atmosphere accounted for＞90%of the total CH4f lux during the summer(Martinez and Anderson 2013).However,ponds have not been given seriousconsideration,especiallyfor thosethat are eutrophic.The autochthonous organic matter production via photosynthesis in eutrophic pondsmay serve as a new source formethane.

The diel biogeochemical process varies greatly in stable water environments,which results from the solar photocycle(Nim ick et al.2011).However,less attention has been paid to the detailed diel gas f luxes.To date,mostreported gas bubble f lux wasmonitored w ith upside-down funnels(Huttunen etal.2001,2003;Duchem in etal.2006;Abril et al.2005),which could represent an average f lux during themonitoring period.However,these devices also prevent the further exploration of environmental factors'inf luenceson gas f luxes.Knowing dielvariations ishelpful to disclose rapid biogeochemical processes in natural aquatic ecosystems(Nimick et al.2011).Previous measurements of diel gas f luxes were usually at intervals of 3—4 h(Zhang and Ding 2011;Xing et al.2005).In this study,we present a month-long diel CH4f lux across the water—air interfaces of a subtropic shallow eutrophic pond located at Yichang,China.We hope the continuous CH4f lux data both by ebullition and by diffusion can give some new know ledge for CH4released from small and shallow ponds.

2 M aterials and methods

2.1 Study area

The oval pond(111°18′54.82′′E,30°43′32.08N),Lianxinhu,is a very small landscape water,which is located in China Three Gorges University,Yichang,Central China.The region has a subtropicalmonsoon climate.Themean annual temperature and rainfall is16.9°C and 1215.6mm,and that is12.5°C and 47.0mm in November.The pond is about3000 m2(～125 m in length and～25 m in w idth)w ith amean andmaximum water depth of 1.45 and 1.75m respectively.The pond bottom is now covered w ith very softmuddy sediments of～10 cm in thickness.Concentrations of nitrogen and phosphorus during the month of Novemberwerearound1.567and0.481mg·L-1respectively.

2.2 W ater-to-air f luxes

An automatic f loating chamber system wasused tomonitor CH4and CO2f luxes across the water—air interface.The chamber is composed of a hollow cylinder and a lid(Fig.1),which are linked by a set of electric putters in order to make the lid go up and down about 15 cm.Two fans(24 V,2.88W and 24 V,3.94W)are f ixed at the top and center of the chamber respectively.When the lid is lifted,both fans work to exchange the air inside the chamberw ith the outside air.While the lid goes down and is closely contacted w ith the hollow cylinder,only the upper fan sw itches on to properly m ix the air inside the chamber.One buoy is f ixed on the lower portion of the chamber.The volume of the chamber is around 56.70 L(diameter and height are 0.38 and 0.5m respectively).

Fig.1 The sketch of the chamber used in this study(1)alum inum alloy frame;(2)polyvinylchloride tube;(3)lid;(4)siliconegasket;(5)electric putter;(6)fan;(7)hollow cylinder;(8)buoy

The chamber employed here was hung on a steel w ire rope over the pond,in which both endpoints were f irm ly tied to pillars.The chamber was connected to a DLT-100 greenhouse gas analyzer(Los Gatos Research,USA),which was used to determ ine the CH4and CO2concentrations in the chamber online.The diel f ield campaign lasting for 28 days(from 3rd to 30th November2016)was undertaken at the same site,which is located at the center of the pond and itswater depth was～1.48m.The time span for a single f lux measurementwas 25 m in.Then the lid of the chamberwas automatically raised by the electric putters for about 5m in to m ix the air inside the chamber and the surrounding atmosphere,and the lid was putdown again.And that cycle repeated.

Herewe estimate CH4diffusive and ebullition f lux w ith the method suggested by Xiao et al.(2014).When there was no obviousmethane ebullition during ameasurement period,the CH4concentration in the chamber gradually increases and shows as a straight line(Fig.2a).Under this situation,the slope of the line was the diffusive methane rate(0.0108 ppmv/m in on Fig.2b),which was suggested by Lambert and Fre′chette(2005).The methane concentration in the chamber increased abruptly when bubbling occurs(Fig.2b).Under this situation,thediffusive methane rate(0.0091 ppmv·min-1on Fig.2b),was determined based on a relatively long and stable straight part(for example,AB on Fig.2b).The CH4concentration in the chamber produced from bubble em issions equals the concentration at the endpoint(Ct on Fig.2b)m inus the sum(Cd on Fig.2b)of the diffusive concentration and the original background value.

Fig.2 Different patterns of the CH4 concentration over time w ithout bubbles(a)and w ith bubbles(b)in the chamber during ameasurement

2.3 In-situ sam plingmeasurements

Surfacewater(0.1 m below the surface)temperature(Tw),pH,dissolved oxygen concentration in water(DO),conductivity,and Chlorophyll a(Chl a)were measured on November 3—4,November 8—9,November 15—16,and November 20—21,2016 using themulti-parameter instrument DataSonde5(Hydrolab,US).The multi-parameter instrumentwas tied to the steelw ire ropementioned above andwas2m away from the chamber.Air temperature(Ta),air pressure(Pa),air hum idity,and w ind speed(Sw i)were measured w ith a portable weather station(YGY-QXY,China).The portable weather station was f ixed to a shoreside light pole and was about 3m above the water surface.

2.4 Statistics and analysis

The raw data of gas concentrations in the chamber were averaged at an interval of 30 s prior to f lux calculating.SPSS(IBM SPSSStatistics22)was used for data analysis.Bivariate correlation was used to evaluate correlations between environmental parameters and CH4f luxes(both diffusive and bubble f luxes).Here Pearson's correlation coeff icients are presented.A signif icant difference of the diel diffusive methane f lux was obtained using one-way ANOVAs during different days.

3 Results

3.1 CH4 f lux at different time

3.1.1 Diffusive f lux

The diffusivemethane f lux of a singlemeasurementhad a range from 2.68×10-5to 0.028mmol·m-2·h-1w ith an average of 0.011 mmol·m-2·h-1.The standard deviation of all measurements is 0.005 mmol·m-2·h-1and the corresponding coeff icient of variability was 0.47.The average diffusivemethane f lux atdifferent times varied from 0.009 to 0.013mmol·m-2·h-1,and the standard deviation and the correspondingcoeff icientofvariabilitywere 0.0007mmol·m-2·h-1and 0.068 respectively.According to our data,the diffusive methane f lux of measurements from 9:00 to 10:30 and from 21:00 to 22:30 were very close to the average diffusive f lux of all measurement(Fig.3).

3.1.2 Bubble f lux

The bubblemethane f lux atdifferent times hasmuchmore variability than the diffusive methane f lux(Fig.3).The bubblemethane f lux of a singlemeasurement had a range from0to0.312 mmol·m-2·h-1w ithanaverage of 0.024 mmol·m-2·h-1.The standard deviation of allmeasurementswas0.054 mmol·m-2·h-1and the corresponding coeff icient of variability was 2.28.The average bubblemethane f lux at different times varied from 0.005 to 0.072 mmol·m-2·h-1,and the standard deviation and the correspondingcoeff icientofvariabilitywas 0.013 mmol·m-2·h-1and 0.567 respectively.The average bubblemethane f lux at different time accounts for 63.97%of the totalmethane f lux w ith a span of 32.71%—86.52%.The average bubblemethane f lux ofmeasurements during 11:30—12:00,19:30—21:00,and 1:30—2:30 approximates the average bubble f lux of allmeasurements(Fig.3).

Fig.3 CH4 f lux at different times and its statistical parameters

The diffusivemethane f lux during daytime(8:00—20:00)and nighttime(20:00—8:00)were almost equal.However,thebubblemethanef luxindaytimewas 0.029mmol·m-2·h-1,whichwas1.61 timesof thatatnight(0.018 mmol·m-2·h-1)(Table 1).It may result from the fact that the bottom water temperature in daytime was higher than at nighttime.We cannot give a quantitive explanation ow ing to the absence of continuous water temperature data.The big difference of the bubblemethane between daytime and nighttime also implies that there is a potential for overestimates of the bubble methane f lux if measurementswere carried out only in daytime.

Table 1 The averagemethane f lux(mmol·m-2·h-1)in daytime and night

3.2 Daily CH4 f lux

The air temperature in November,2016 had a range from-0.5 to 19.0°C w ith an average of 11.6°C.The average bubbleanddiffusiveCH4f luxwere0.56and 0.26mmol·m-2·day-1respectively(Table 2).The maximum of thebubble CH4f lux,which occurred on November 5,2016(Fig.4),was about 4.6 times the m inimum CH4ebullition.Themaximum of diffusive CH4f lux was～1.7 times the corresponding m inimum.In general,the bubble CH4f lux appeared to decrease during the sampling period.However,there was no obvious regularity for the diffusive CH4f lux.The CH4ebullition f lux,on average,accounted for 67.15%of the total f luxes(diffusive+ebullitive)w ith a range of 43.67%—79.29%.

For the diffusivemethane f lux,its coeff icient of variation of the average atdifferent timewas lower than thatof the daily average.However,it was the opposite for the bubblemethane f lux(Table 2).

4 Discussion

4.1 CH 4 f lux

The average diffusive f luxof all measurements in November 2017 was 0.26mmol·m-2·day-1,which iswell w ithin the expected range given by Holgerson and Raymond(2016).Thismay imply that the size of a pond has a signif icant inf luence on its biogeochem ical processes.According to the standard deviation and coeff icient of variation,the amplitude of diel diffusive f lux change(0.04 mmol·m-2·h-1and 0.16)wasmuch larger than that occurring on amonthly timescale(0.02mmol·m-2·h-1and 0.07)in November2016.From thispointof view,we think that it is a better choice tomonitor gas f lux acrosswater—gas interface regularly than to just choose 1 day for the month.

ThebubbleCH4f lux accounts for68.23%of the total f lux in the Lianxinhu pond.However,no bubble was observed during a diel campaign in October,2013 in the Yezhulin pond,whichwasalso located in thesame zoneandwasalso eutrophic(Xiao et al.2014).The methane f lux from the Yezhulin pondwas0.096mmol·m-2·day-1,which ismuch smaller than the average(0.82mmol·m-2·day-1)and the minimum(0.40mmol·m-2·day-1)emitted from the Lianxinhu pond(Table 2).Because the Yezhulin pond isdredged every year,we think a possible reason responsible for the differencemay be thatmethaneem itted from the Lianxinhu pond mainly results from the degradation of old sediment organic matters.However,these orgasm ic matters were primed by the addition of algal,which resulted in amore rapid rem ineralization rate(Bianchietal.2015).

4.2 Environmental factorswhich inf luence the diel diffusive CH4 f lux

Statistical data of environmental parameters and diel CH4f luxduringNovember 8—9,November 15—16,and November20—21,2016,which are in theearly,middle,and late periods of the month respectively,are shown in Table 3.The average diffusive methane f luxduring November 8—9,2016 was very close to that during November 20—21,2016 and was lower than that during November 15—16,2016,though both the average air temperature and water temperature during November 20—21 and November 15—16,2016 were very close and higher than those during November 8—9,2016.A signif icant differencewasobtained using one-way ANOVAsbetween the diel diffusivemethane during November 15—16,2016,and that during November 8—9 or November 20—21,2016.There was no statistically signif icant difference betweenthe dieldiffusivemethane during November8—9,2016 and that during November 20—21,2016.This situation may result fromthehigher averagew indspeedduring November 15—16,2016(0.26m·s-1)than that during November8—9 and November20—21,2016(both are about 0.11m·s-1).It is well known that wind speed is a keyfactor which dom inates gases f luxes across water—air interfaces by turbulent m ixing in lakes,reservoirs,and oceans(Wanninkhof et al.2009;Gue′rin et al.2007;Upstill-Goddardet al.1990).Theaveragediffusive methane f lux during November 8—9 and November 20—21,2016 were very close(Table 3),but both air and water temperatures during November 8—9 and November 20—21,2016weremuch lower than thoseduring November20—21,2016.This situation may result from the precipitation during November 8—9,2016.

Table 2 Statistics of the averaged daily CH4 f lux and the average CH4 f lux at different times in November,2016

Fig.4 Daily CH4 f lux and its statistical parameters

Table 3 Statistical data of environmental parameters and diel CH4 f lux for different days

The average methane ebullition f lux during November 8—9,2016 was much higher than that during November 15—16 and November 20—21,2016 though the average air and water temperatures during the latter two periodswere higher than those during the previous periods(Table 3).We cannot give a reasonable cause for this phenomenon.

Signif icant correlations betweenthe diel diffusive methane f lux and main environmental factors were only observed during November 20—21,2016(Table 4).

Diel diffusion CH4f lux was signif icantly positively correlated to the surface water temperature and air temperature,and negatively correlated to DO and Chl-a when allmeasurements of 3 days were taken into account together(Table 5).This phenomenon was also observed in the Xiangxi Bay in August,which is also eutrophic,and an algae bloom occurred during the season(Xiao etal.2013).As we know,alga bloom resulted in the increasing DO(Xiao etal.2013;Boto and Bunt1981).DO increasesw ith strong photosynthesis of phytoplankton.The metabolism(respiration alternating photosynthesis)of phytoplankton inf luences the diel gas f luxes greatly(Xiao et al.2013).Xing et al.(2005)also thought that phytoplankton m ight dom inate C dynam ics in eutrophic water bodies.Thew ind speed is a key factor to dom inate gas exchanges across water—air interfaces in aquatic ecosystems(Xiao et al.2014;Gue′rin et al.2007;Liss and Merlivat 1986;Wanninkhof 1992;Wanninkhof et al.1985;Cole and Caraco 1998).However,there was no signif icant correlation between the diel methane f lux and the w ind speed.This may result from the situation thatallw ind speedsmeasured were very low(Table 3).

Both CH4production and em issions were highly temperature dependent(Zimov et al.1997;Yvon-Durocher 2014;Westermann etal.1989).Thissituation occurs in the pond according to ourdata(Table 5).Asmentioned earlier,the bubblemethane f lux in daytimewas 1.61 timesof that at nighttime.We think this resulted from the higherwater temperature during the daytime,which caused more active methanogenesis.Duringnighttime,dissolvedmethane concentration in waterm ightdecreasew ith lowering water temperature.The phenomenon that the diffusive methane f lux in daytime and nighttimewere almostequalmay also be induced by the lowering air temperature in night.The cooling effect of water surface m ight cause turbulent mixing in the water column,which enhances the gas transfer velocity(Xiao etal.2014;Macintyre etal.2001).This remains to be verif ied because we had no dissolved methane concentration data.However,unexpectedly,therewas no obvious correlation between the daily diffusive methane and the air temperature(Table 6).

Table 4 Pearson's correlation coeff icients between environmental parameters and diel CH4 f lux for different dates

Table 5 Pearson's correlation coeff icients between diel diffusive CH4 f lux and environmental parameters(N=151)

Table 6 Pearson's correlation coeff icients between daily diffusive CH4 f lux and weather factors(N=28)

5 Conclusions

For the eutrophic pond in this study,the average bubble anddiffusiveCH4f luxeswere0.56±0.18and 0.26±0.04mmol·m-2·day-1,respectively,and the CH4ebullition f lux accounted for 68.23%of the total f lux.The diffusive methane f lux in daytime and nighttime were almostequal.However,thebubblemethane f lux in daytime was 0.029 mmol·m-2·h-1,which was 1.61 times that that at nighttime.The maximum of the bubble CH4f lux was about 4.6 times of the m inimum CH4ebullition.The maximum of diffusive CH4f lux was～1.7 times of the corresponding minimum.W ind speed,the surface water temperature,and DO dominatemethane eff luxes from the pond,and the latter is in nature subjected to themetabolism of algae in the pond.However,key environmental factors which dom inate gas f lux processes vary w ith different weather conditions.W ind speedmay be unimportantwhen it is extremely low.

Acknow ledgementsThis work was f inancially supported by the NationalScienceofChina(Nos.41273110,91647207,and 51509086)and Natural Science Foundation of HubeiProvince,China(2014CFB672).