Respiration in lakes -

“chap07” — 2004/11/8 — page 103 — #1

CHAPTER 7

Respiration in lakes

Michael L. Pace1 and Yves T. Prairie2

1 Institute of Ecosystem Studies, USA2 Département des sciences biologiques, Université du Québec à Montréal,Canada

Outline

This chapter reviews, from a quantitative perspective, estimates and models of planktonic and benthicrespiration in lakes, as derived from bottle and core incubations. Using data gleaned from the literature,empiricalmodels of planktonic andbenthic respiration are presented inwhich respiration is shown todependprimarily on lake trophic conditions and, secondarily, on temperature and other factors such as dissolvedorganic carbon loading and community structure. The possibilities and limits of alternative methods ofmeasuring lake respiration and its components are discussed. Extrapolations of the findings to the globesuggest that about 62–76 Tmol of carbon are respired annually in the world’s lakes, exceeding the estimatedgross primary production of these ecosystems. Lakes thus contribute significantly to the oxidation of land-derived organic carbon.

7.1 Introduction

Respiration in lakes recycles organic carbon arisingfrom photosynthesis back to inorganic carbon. Priorto this transformation, the organic carbon is poten-tially available to support secondary production.The efficiency of primary and secondary productionrelative to respiration is an important featureof lakesand other aquatic ecosystems. The relative rates andlocations of primary production and respiration alsoinfluence oxygen concentrations. Intensive respira-tion in sediments and isolated waters, such as thehypolimnion or in ice-capped lakes, often leads tooxygen depletion and even anoxia. Thus, the ratesand controls of respiration are of central importancein lake ecosystems.

Respiration affects net balances of carbon in lakes.Many lakes are net heterotrophic, meaning respira-tion exceeds gross primary production (del Giorgioand Peters 1994). In this case, the partial pressure ofcarbon dioxide exceeds that in the atmosphere andso carbon dioxide diffuses out of the lake while

oxygen is undersaturated and diffuses into the lake(Prairie et al. 2002). The loss of carbon dioxide tothe atmosphere, however, does not imply that stor-ageof carbon innet heterotrophic lakes is zero. Quitethe contrary, lakes accumulate organic matter insediments (Dean and Gorham 1998). This seemingcontradiction that lakes are sources of carbon to theatmosphere and sedimentary sinks for organic car-bon is reconciled by respiration of terrestrial organicinputs in lakes (del Giorgio et al. 1999). Thus, thecarbon that supports respiration in lakes is derivedfrom within-lake primary production as well asallochthonous sources in the watershed. Some lakesarenet autotrophic, wheregrossprimaryproductionexceeds respiration. These draw inorganic carbonfrom the atmosphere and store sediment organiccarbon. Thus, both net autotrophic and net het-erotrophic lakes are sinks for atmospheric carbonwhen considered on a watershed scale.

The purpose of this chapter is to summarizemeasured rates, variation in rates among lakes, andfactors regulating respiration in lakes. Methods of

103

“chap07” — 2004/11/8 — page 104 — #2

104 RE S P I RAT ION IN AQUAT I C ECOSYS T EMS

measuring respiration are not reviewed in detailbut common approaches are mentioned briefly togive the reader context for the discussion of rates.Rates of planktonic and sediment respiration arepresented from literature data on oxygen consump-tion obtained from studies using bottle and coreincubations. The data are used to develop generalempirical models of planktonic and sedimentrespiration as a function of lake trophic conditions.Thesemodels are applied to generate a global estim-ate of respiration in lakes. Anaerobic respiration isalso important in many lakes but only consideredbriefly here. Fenchel (1998) and King (Chapter 2)provide an excellent introduction to anaerobicprocesses.Wetzel (2001) reviews thegeneral signific-ance of anaerobic respiration in lakes.Alimitation oftraditionalmethods for estimating respiration is thatchanges in oxygen or carbon dioxide of enclosedcommunities are measured and extrapolated overtime and space for an entire lake. Diel cycles ofoxygen can be measured using in situ instruments.We provide an example of respiration derived fromthis “free water” approach and compare estimatesfrom in situ and enclosure approaches.

7.2 Planktonic respiration

7.2.1 Methods

Consumption of oxygen in dark bottles has beenthemain approach tomeasuring planktonic respira-tion.Alternativemethodshavebeenused, includingmeasuring activity of the electron transport system(del Giorgio 1992), changes in the partial pressureof carbon dioxide (Davies et al. 2003), and changesin oxygen isotope ratios (Bender et al. 1987). Weprimarily discuss dark bottle oxygen consumptionbecause the data synthesis provided below is basedon measurements made with this method. We alsobriefly consider the application of some of the othermethods to lake systems.

In the dark bottle method, plankton commu-nities are enclosed in bottles and replicate mea-surements of the initial and final concentrations ofoxygen are made. Because the method dependson measuring small differences in oxygen con-centrations, precision and accuracy are important.

Much effort has focused on improving standardoxygen measuring techniques to enhance accuracy(e.g. Carignan et al. 1998) and facilitate the ease ofreplicate measurements in order to improve pre-cision (e.g. Roland et al. 1999). In addition, theavailability of very sensitive instruments for oxygenmeasurement, such as dual-inlet mass spectrome-try, allows shorter incubation intervals (del Giorgioand Bouvier 2002). With improvements in methods,incubations of a few hours rather than a full dayare sufficient to estimate rates even in oligotrophiclakes (e.g. Carignan et al. 2000). Short incubationreduces artifacts associated with enclosing micro-bial communities in bottles and provides resultsover timescales commensurate with other measuresof microbial activity such as primary and bacterialproduction assays.

Differences in oxygen concentration at the begin-ning and end of a dark bottle assay represent netoxygen consumption from all processes occurringin the bottle over time. Since bottles are incubatedin the dark, it is assumed no oxygen productionoccurs. Oxygen consumption may arise from abi-otic chemical oxidation as well as biotic reactions.Chemical oxidation reactions are primarily lightdriven. The aggregate of these reactions is referredto as photooxidation, and this process is presumedto cease during dark incubations. Graneli et al.(1996) directly compared measurements of photo-oxidation and dark respiration in Swedish lakes thatvaried in dissolved organic matter and found thatphotooxidation was <10% of dark respiration.Thus, measurements of dark respiration are proba-bly not compromised by oxygen consumption dueto photooxidation reactions initiated in the lightthat might continue during some portion of thedark incubation.

Extrapolations of bottle respiration measure-ments typically make the assumption that oxygenconsumption measured in the dark is equivalentto oxygen consumption that occurs in the light.However, oxygen consumption in the light canbe higher than in the dark because of photores-piration, the Mehler reaction (a photoreduction ofO2), and light-stimulation of normal cytochromeoxidase respiration (also referred to as dark ormitochondrial respiration). In addition, a process

“chap07” — 2004/11/8 — page 105 — #3

RE S P I RAT ION IN LAKE S 105

known as the alternative oxidase respiration mayconsume oxygen in the light and the dark (Luzet al. 2002, see also Raven and Beardall, Chapter 3).Photorespiration occurs because of oxygen accu-mulation and a relative decline of carbon dioxideas photosynthesis proceeds such that the primarycarbon fixation enzyme (ribulose-1,5-bis-phosphatecarboxylase/oxygenase) no longer fixes carbondioxide but instead consumes oxygen (see alsoRaven and Beardall, Chapter 3). The Mehlerreaction involves electron tranport among photo-systems 1 and 2 with concurrent oxygen reduction(Falkowski and Raven 1997). Finally, enhancedcytochrome oxidase respiration occurs in thelight for actively growing phytoplankton in cul-tures (Grande et al. 1989; Weger et al. 1989). Theoverall effect of these mechanisms that result inlight-enhanced oxygen consumption on measure-ments of respiration is poorly understood and is anarea of active research (Luz et al. 2002). Underestim-ates of respiration in dark incubations will likelybe related to the relative importance of autotrophsand their growth rate in a given sample as wellas to contemporaneous environmental conditionssuch as light exposure and temperature (see alsoWilliams and del Giorgio, Chapter 1).

Following increases in carbon dioxide is a feasiblemeans for measuring respiration but is complicatedby potential changes in the components of dissolvedinorganic carbon (DIC)—aqueous carbon dioxide,bicarbonate, and carbonate. Davies et al. (2003)describe a method applicable to lakes for measur-ing changes in the partial pressure of carbon dioxide(pCO2) in a headspace created in a sealed sample.The technique involves measuring pCO2 over timealong with at least one measurement of the totalDIC. If carbonate alkalinity is constant during sam-ple incubation, DIC can be calculated from pCO2for every time point. Respiration is then estimatedfrom the change in DIC over time. This methodhas not yet been extensively applied, but offersthe opportunity to estimate gross primary produc-tion, respiration, and net community production(Davies et al. 2003) as currently done with lightand dark bottle measurements of oxygen. Com-bining this method with oxygen techniques would,in theory, allow measurement of both carbon and

oxygen changes as a consequence of photosynthe-sis and respiration and provide direct estimates ofphotosynthetic and respiratory quotients.Another approach to production and respira-

tion measurements involves spiking samples withH2

18O. This method has been used in a few lakes(Luz et al. 2002). The advantage of this technique isthat gross primary production and respiration canbe estimated in the light allowing evaluation of theassumption necessary with other methods that res-piration in the light equals respiration in the dark.The method is technically demanding and requiresmeasuring oxygen isotopes and estimating fraction-ation for a number of processes related to biologicaloxygen consumption and production as well asabiotic exchanges of oxygen (e.g. Luz et al. 2002).

7.2.2 Data compilation

Wecompileddata from the literature for lake studiesthat reportedratesofplanktonic respirationbasedondarkbottleoxygenconsumptionalongwiththe inde-pendent variables chlorophyll a, total phosphorus,and dissolved organic carbon. Some studies usedother methods to measure respiration such as DICaccumulation indarkbottles (e.g. Graneli et al. 1996),andwe excluded these to obtain amethodologicallyconsistent dataset. We only used data from studiesmore recent than those of del Giorgio and Peters(1993) who summarized data from the literature onmean respiration in 36 lakes. We found seven stud-ies that reported mean rates for lakes along with atleast some of the independent variables—del Gior-gio and Peters (1994), Pace and Cole (2000), Duarteet al. (2000), Carignan et al. (2000), Biddanda et al.(2001), Biddanda and Cotner (2002), and Bermanet al. (2004). Our search was not exhaustive butwas extensive. Keywords in titles and abstractswere searched electronically and numerous individ-ual articles were reviewed. In addition, we visuallyreviewed all article titles and abstracts over severalyears in a few journals. A number of the studiesnoted above measured respiration in multiple lakesand so an aggregate 70 values of mean planktonicrespiration were derived from the literature.

The studies were all of north temperate lakessampled during the May–September period with

“chap07” — 2004/11/8 — page 106 — #4


the exception of Duarte et al. (2000) where lakesin northeast Spain were sampled from Decemberthrough May. Rates in the literature studies are allderived from surface, mixed-layer samples, andmeans are based on as few as one sampling time(e.g. Duarte et al. 2000), 3–4 samples from approx-imately monthly observations (e.g. del Giorgio andPeters 1994; Carignan et al. 2000), 15–17 samplesbased on weekly observations (Pace and Cole 2000).Reservoirs and saline lakes sampled by Duarte et al.(2000) were excluded from the final data, becauserespiration in these systems tended to be substan-tially lower and higher, respectively than in thefreshwater lakes. Means from Pace and Cole (2000)were derived from one reference and three lakesmanipulated with additions of nitrogen and phos-phorus over a number of years. We averaged alldata for the reference lake to generate a single mean(years 1991–7). We averaged data for 2 years prior tonutrient loading manipulations for the other threelakes (1991–2) to generate three means and thentreated each year of nutrient manipulation (loadingrates varied among years) as a separate mean. Thisgenerated 19 mean values overall representing 28lake-years of measurements.

Typically, comparisons of production, nutrientcycling, or biomass among lakes reveal rangesin variation of over a 1000-fold. Mean values ofplanktonic respiration only varied from 0.029 to6.73 mmolO2 m−3 h−1, a factor of 265. This modestrange must underrepresent actual variation amonglakes, because of the limited number of studiesand the focus on temperate lakes in summer. Amore global survey of lakes would probably revealsystems with lower respiration because of lowproductivity and low temperatures, and some withhigher respiration because hypereutrophic lakes(chlorophyll less than 75 mgm−3) are not well rep-resented in the dataset. However, the limited rangeindicates an important feature of planktonic respira-tion. Respiration is less variable than other variablesmeasured in aquatic systems.

7.2.3 Models of planktonic respiration

There are few models of planktonic respirationthat are comparable to well-developed models

of primary production based on photosynthesis—irradiance relationships (Platt and Jassby 1976) aswell as approaches that combine photosyntheticparameterswith depth distributions of primary pro-ducers to estimate rates over large scales such asthe global ocean. Models for planktonic respira-tion are largely empirical and are based on featuressuch as primary productivity as, for example, in theglobal analysis of aquatic ecosystems inclusive oflakes provided by Duarte and Agusti (1998).

TemperatureBecause lake planktonic respiration studies havefocused on surface water in summer of temperatelakes, there has been relatively little considerationof the effects of temperature on rates. Neverthe-less, Carignan et al. (2000) determined a relationshipbetween volumetric respiration and several vari-ables, including temperature. This multiple regres-sion relationship (their equation (10) of table 3) isuseful in that it allows one to isolate the influence oftemperature on respiration while statistically keep-ing all other variables constant. Over their observedtemperature range of 11–22.5◦C, the log–log slopecoefficient of temperature is 1.28, implying a Q10of about 2.3 in that temperature range. If this slopecoefficient of 1.28 remainsvalid even for lower temp-eratures, it would suggest an even higherQ10 at lowtemperatures. This prediction needs to be confirmedby additional observations across a broader temper-ature range, but the same pattern has been observedfor sediment respiration, as discussed below.

Lake trophic conditions and organic matter loadingPlankton respiration (PR) increases in concert withincreases in chlorophyll, phosphorus, and dissolvedorganic carbon (Fig. 7.1). Mean lake chlorophyll a

accounts for 71% of the variance in respiration(Fig. 7.1(a)). A similar relationship is observed withtotal phosphorus although r2 was higher (0.81) forthis relationship based on fewer observations (n =62, Fig. 7.1(b)). The equations relating PR (mmolO2 m−3 h−1) to chlorophyll a (Chl, mg m−3) andtotal phosphorus (TP mg m−3) are:

log10 PR = −0.81 + 0.56 log10 Chl (1)

log10 PR = −1.27 + 0.81 log10 TP (2)

“chap07” — 2004/11/8 — page 107 — #5


Figure 7.1 Relationships between planktonic respiration and (a)chlorophyll a (b) total phosphorus, and (c) dissolved organic carbon(DOC) for lakes. Regressions are model 1 least square plots with 95%confidence intervals for the regression and for individual predictions.

These two relationships reflect a typical patternof lakes where primary production is stronglynutrient limited and closely related to total phos-phorus inputs and concentrations. The large-scalecomparison in Fig. 7.1 indicates that planktonicrespiration is strongly coupled with variation inphytoplankton biomass and primary production.

Equation 1 compares well with the relationshipestablished by del Giorgio and Peters (1993) fromearlier literature data. They reported a regressionfor PR in units of mg Cm−3 d−1 based on chloro-phyll a with a slope of 0.65 and an intercept of 1.65.Transforming the data underlying equation (1) intocommensurable units yields the same intercept of1.65 observed by del Giorgio and Peters (1993). Thusthe two independent relationships overlap closely,differing only slightly in slope (0.65 versus 0.56) andgive similar predictions, especially in the range of0.5 to 5 mg chlorophyll m−3.

Dissolved organic carbon (DOC) in lakesis derived both from watershed sources and

within-lake primary production. If allochthonousinputs of DOC are significant to lake carboncycling and supplement ecosystem metabolism,DOC should be positively related to PR. This isthe case for 63 lakes (Fig. 7.1(c)) where there is apositive relationship between DOC (g m−3) and PR(r2 = 0.49).

log10 PR = −1.15 + 1.01 log10 DOC (3)

The positive relationship of PR with DOC reflects,in part, increased microbial metabolism associatedwith increased inputs of DOC to lakes. However,DOC is not simply the result of watershed inputsbut also arises from phytoplankton and littoral pri-mary producers, the relative contributions of whichare still not well known. Hence, the relationship iscomplex and a more interesting test of the positiveeffect of watershed-derived DOC on PR would beto compare lakeswith differingDOC concentrations(and loadings) but similar levels of total phosphorusor chlorophyll a.

Both chlorophyll and DOC are significant vari-ables in a multiple regression (r2 = 0.76, n = 63).

log10 PR = −0.92 + 0.41 log10 Chl

+ 0.30 log10 DOC (4)

The t-values associated with each variable indicatechlorophyll is highly significantwhileDOCexplainsa modest but also significant amount of the residualvariation (chlorophyll t = 7.96, p < 0.0001; DOC t =2.35, P = 0.022). Chlorophyll and DOC, however,are correlated (r = 0.71, p < 0.0001) and so thisrelationship may be less suitable for predicting PRand associated uncertainty than the single variablerelationship with chlorophyll (equation 1).

Other factors influencing planktonic respirationThere have been relatively few detailed studies onrespiration and depth in lakes. The primary rea-son is that respiration becomes difficult to meas-ure at low rates. With the declines in temperatureand primary productivity associated with increasesdepth in lakes, respiration rates probably mustalso decline. However, bacterial metabolism con-tributes significantly to respiration and partiallycompensates for the expected declines with depth.Numerous efforts have been made to partition oxy-genconsumption in thehypolimnionof lakes among

“chap07” — 2004/11/8 — page 108 — #6


water and sediments because of the importanceof hypolimnetic oxygen deficits during seasonalstratification. These studies either involve model-ing (e.g. Livingstone and Imboden 1996) or, whendirect measurements of hypolimnetic PR have beenattempted, long incubations of organism concen-trates (e.g. Cornett and Rigler 1987). Refinementsin methods for measuring respiration discussedabove provide an opportunity to now make directmeasurement over reasonable incubation times.

While planktonic respiration is strongly relatedto nutrient and organic matter loading as well thebiomass of primary producers, food web structurecan also influence respiration. Pace and Cole (2000)reported strong foodweb effects on respiration froma series of whole lake manipulation studies. Fishcommunities in three lakes were shifted to promotethe dominance of either piscivorous or planktivo-rous fish and then the lakes were fertilized withnutrients. Predator–prey interactions arising fromthe top of the food web either strongly limitedlarge crustacean grazers or alternatively tended topromote dominance of large zooplankton throughtrophic cascades (Carpenter et al. 2001). Nutrientadditions were carried out over 5 years in eachlake and dark bottle respiration was measured.Planktonic respiration differed among the fertilizedlakes at the same nutrient loading and, over the 15lake-years represented in this study, was inverselyrelated to theaverage sizeof crustaceanzooplankton(Fig. 7.2). Gross primary production, system respi-ration, and net ecosystem production as measured

Mean Crustacean Length (mm)

0.4 0.6 0.8 1.0 1.2

Re

spir

atio

n (

mm

ol m

-3d

-1)

12

16

20

24

r2= 0.61

Figure 7.2 Relationship between mean crustacean length and meanplanktonic respiration in nutrient fertilized experimental lakes.

by free water methods (discussed below) were alsostrongly affected by differences in food web struc-ture in these experimental lakes (Cole et al. 2000).

7.3 Sediment respiration

The sediment–water interface is the site of some ofthe most intense biological activity of freshwatersystems (Krumbein et al. 1994). The large exchangesurface area provided by loosely compacted sedi-ment particles, the organic nature of the sediments,and its physical stability all act synergistically toproduce an extremely active milieu. In comparisonto the water column, sediments are typically 1000-fold richer in both nutrients and carbon, and therebyprovide a highly concentrated growth medium forheterotrophic bacteria. Correspondingly, bacterialabundances andproduction rates are typically about2–3 orders greater than in the water column, aphenomenon recognized nearly half a century ago(Kuznetsov 1958). Depending on the transparencyof the water column above the sediments, the highbenthic metabolic activity is differently dividedbetween anabolic and catabolic processes. How-ever, because most of the sediments are below theeuphotic zone, respiration is by far the domin-ant metabolic process in sediments. In his treatise,Hutchinson (1957) argued that oxygen consumptionwithin lake hypolimnia can be accounted for bywater column respiration, and therefore, concludedthat sediment respiration represented only a minorportion of the hypolimnetic metabolism, a view atodds with recent literature. For example, Lund et al.(1963) estimated that about half of the hypolimneticmetabolism of LakeWindermerewas benthic, a con-clusion upheld by Cornett and Rigler (1987) whofound between 30 and 80% of the total hypolimneticoxygen consumption was attributable to sedimentrespiration in a series of Québec lakes.

7.3.1 Methods

Historically, sediment respiration has been deter-mined mostly using measurements of oxygen con-sumption (Hayes and Macaulay 1959; Pamatmat1965) although some early work also used carbon

“chap07” — 2004/11/8 — page 109 — #7


dioxide accumulation (Ohle 1952). Oxygen con-sumption and carbon dioxide release are not alwayswell coupled (Rich 1975, 1979), particularly in dys-trophic lakes and peat bogs (see Roehm, Chapter 6),and the degree of decoupling is usually attributedto anaerobic respiration pathways (methanogen-esis or use of alternate terminal electron accep-tors such as sulfates, nitrates, and metal oxides,see also King, Chapter 2). In systems whereanaerobic metabolism is negligible, oxygen con-sumption and inorganic carbon release are bothadequate measures of benthic respiration, pro-vided carbon dioxide release and consumptioninduced by CaCO3 precipitation and dissolution,respectively, are taken into account (Graneli 1979;Andersen et al. 1986, and see Hopkinson and Smith,Chapter 6, and Middelburg et al., Chapter 11). Mostreports of benthic respiration are based on oxygenconsumption.

Benthic respiration rates has been estimated byseveral direct and indirect techniques including:sediment core incubations (Hargrave 1969a; Rich1975; den Heyer and Kalff 1998; Ramlal et al. 1993;Schallenberg 1993), benthic chambers (Pamatmat1965, 1968; James 1974; Cornett 1982), verticalprofiles of decomposition products in interstitialsediment water (Carignan and Lean 1991), or asthe difference between total and water columnmetabolism in lake hypolimnia (Cornett and Rigler1987). Because these methods encompass verydifferent degrees of spatial integration and becauseof the notorious spatial heterogeneity of lake sedi-ments, it is difficult to compare rates of benthicrespiration obtained from these different methods.

James (1974) compared sediment respiration ratesobtained from cores and benthic chambers and con-cluded that cores tended to underestimate respira-tion rates by an average of 35%, although Campbell(1984) showed that core incubations can estimatewhole-lake sediment respiration within 10%. Giventhat most literature values of benthic respirationwere derived from core incubations, models of ben-thic respiration, therefore, carry the potential bias ofbeing slight underestimates.A more spatially integrative measure of sedi-

ment respiration has recently been proposed byLivingstone and Imboden (1996). The effects of

0.060

0.050

0.070

0.080

0.1 0.2 0.3 0.4 0.5sediment area:water volume

Oxy

ge

n c

on

sum

ptio

n (

mg

-O2L

-1d

-1)

Figure 7.3 Relationship between oxygen consumption as a functionof the sediment area: water volume ratio for each 1-m hypolimneticstrata in Lac Fraser, Québec (Canada).

sediment respiration can be observed through thenearly ubiquitous nonlinearities of hypolimneticoxygen profiles, with lower concentrations near thesediments. The extent of these sharper reductionsclose to the sediments is largely determined by theincreasing area of sediments in the deeper strata ofthe hypolimnion. Livingstone and Imboden (1996)proposed a simple modeling approach to estimatethe relative contributions of benthic and water col-umn metabolism to the total oxygen consumptionof the hypolimnion by relating the rate of declineof oxygen at a given depth to the sediment surfacearea: watervolumeat thatparticulardepth.Applica-tion of this method to a vertical oxygen profile fromLac Fraser (Québec) shows a very strong relation-ship, the intercept and slope of which correspondto the water column and benthic respiration rates,respectively (Fig. 7.3). If the general applicability ofthis approach can be firmly established, it may pro-vide a more useful measure of benthic respiration,by implicitly integrating differences in respirationat different depths within the hypolimnion and byavoiding potential experimental artifacts inherent toincubations in closed containers.

7.3.2 Data compilation

The literature on sediment respiration in freshwa-ter systems is much sparser than for marine orestuarine waters, particularly in recent years. As

“chap07” — 2004/11/8 — page 110 — #8


for pelagic respiration, we synthesized the pub-lished literature and obtained values on sedimentrespiration rates from a variety of freshwater sys-tems, located mostly in North America or Europe.The lakes spanned the entire trophic range, fromthe arctic oligotrophic Char Lake (Canada, Welchand Kalff 1974) to eutrophic Lake Alserio (northernItaly, Provini 1975) and encompassed very differenttypes of sediments (organic matter content variedfrom 10 to 40% ). Sediment respiration varied only20-fold, from 1.6 to 33 mmolesO2 m−2 d−1 (aver-age is 10.9), similar to the range observed by denHeyer and Kalff (1998). For consistency, modelsdeveloped from the literature data were restrictedto measurements made in core incubations.

7.3.3 Models of benthic respiration

Existing models of benthic respiration can bebroadly categorized as those based on sedimentproperties and those based on water column orwhole system properties. In comparison with mod-els of planktonic respiration, the predictive ability ofexistingsediment respirationmodelshasbeenratherpoor and their generality tested more regionally.

Effect of temperatureThe earliest general model of benthic respirationwas originally proposed by Hargrave (1969a), whoargued that temperature was the main driving vari-able. Graneli (1978) suggested that temperature andoxygen concentration were more important factorsin regulating sediment respiration than differencesin lake trophic status and its associated lake primaryproductivity. Our compilation of sediment respira-tion rates from the literature gives, at first sight,credibility to this claim. Temperature alone has beenshown to induce changes in respiration rates ofnearly one order of magnitude (Hargrave 1969b;Graneli 1978; Ramlal et al. 1993) within the 4–25◦Ctemperature range, in sediments incubated at dif-ferent temperatures or followed through an annualtemperature cycle. Thus, temperature-induced vari-ations in sediment respiration of a given sedimentcan be nearly as large as among-lake differences (1.5and 1 orders ofmagnitude range, respectively). Sev-eral authors have shown that the temperature effect

is more pronounced at low temperatures than inwarm environments (Hargrave 1969b; Graneli 1978;den Heyer and Kalff 1998). These relationships areoften best described on a double logarithmic scale,and therefore, the relative increase in respirationrate with temperature (i.e. a Q1) decreases withincreasing temperature and can be expressed by thesimple function:

Q1 = 1 + b/t (5)

where b is the slope of the log–log relationshipbetween respiration rate and temperature (t in ◦C).An analysis of b values derived from the literatureyields an average value of 0.65± 1 and Fig. 7.4 illus-trates how the relative rate of increase in respirationrate per degree centigrade decreases with temper-ature. Most of the nonlinearity in this relationshipoccurs below 10◦C, suggesting that changes in tem-perature aremuchmore important for hypolimneticsediments than in warm littoral zones. For example,increasing the temperature of hypolimnetic sedi-ments from 4 to 5◦C has the same relative effect onbenthic respiration as increasing the temperature ofepilimnetic sediments from 20 to 25◦C. For compar-ison, Fig. 7.4 also illustrates the inferred influence oftemperatureonplanktonic respirationderived in theprevious section. It appears that, at least for warmwaters, the influence of temperature on respirationis greater for the plankton than it is for sediments.

Figure 7.4 Temperature dependence of respiration expressed asrelative increase in respiration rate per degree Celsius. Solid line is forsediment respiration, dashed line corresponds to planktonicrespiration.

“chap07” — 2004/11/8 — page 111 — #9


Assuming that the temperature relationship holdsgenerally, sediment respiration (SR) rates can benormalized to a standard temperature (e.g. 10◦C)by using

SRstd 10◦C = SR/(Temp/10)0.65 (6)

to allowmore direct comparisons among sedimentsof different systems. Application of this standard-ization to our data compilation showed that whileindividual values changed as much as threefold, itaffected the overall range of observed benthic res-piration rates very little, suggesting that the strongrelationship between temperature and benthic res-piration rates of Hargrave (1969a) may be more theresult of cross correlations between temperature andother system properties (such as lake productivity)than only the metabolic enhancement of respirationat high temperatures.

Organic matter loading, primary production,and lake trophic statusUltimately, benthic respiration below the photiczone is fueled by organic material sinking fromthe pelagic zone or transported laterally fromshallow sediments. Despite this functional coup-ling, the link between short-term organic carbonflux and sediment respiration has not been wellestablished largely because of the apparent delaybetween when the organic particles settle anddecompose. Indeed, the relative invariance ofbenthic respiration over the seasons when largechanges occur in the organic matter flux hasled several authors to argue that benthic respira-tion is the reflection of the organic matter load-ing integrated over several years rather than therapid decomposition of freshly sedimentedmaterial(Lasenby 1975; Graneli 1978; Linsey and Lasenby1985). The implications of this dampening effect ofthe accumulated surface sediments are important:reductions in lake productivity following eutroph-ication control measures may only translate into aresponse in benthic metabolism after several years(Graneli 1978; Carignan and Lean 1991). In lakeswith shallow hypolimnia, which constitute themajority of eutrophic lakes (Kalff 2002), sedimentrespiration is responsible for a large fraction of thetotal hypolimnetic oxygen consumption (Cornett

and Rigler 1987). Hypolimnetic anoxia may, there-fore, continue long after nutrient loading reductiontargets have been achieved, a phenomenon unfor-tunately observed in many lake restoration efforts(Sas 1989). Baines and Pace (1994) have shownthat organic particle flux is well correlated to laketrophic status asmeasured by epilimetic chlorophyllconcentration. In steady-state conditions, benthicrespiration should therefore be well correlated withprimary production and its determinants, particu-larly given that the fraction of the primary produc-tion carbon exported to the deeper layers appearsonlymodestly reduced inhighlyproductive systems(Baines and Pace 1994). A closely related hypothesiswas first proposed by Hargrave (1973), who sug-gested that benthic respiration should be expressedas a power function of the ratio of areal primaryproduction (PP) to mixing depth (Zm)

SR = α(PP/Zm)β (7)

Graneli (1978) further tested this pattern in a seriesof five Swedish lakes and showed that while therelationship was strong, Hargrave’s model sys-tematically overpredicted the measured sedimentrespiration. This compound variable (PP/Zm) isequivalent to the volumetric primary productionaveraged over the epilimnion, a variable known tobe well predicted from phosphorus concentration(e.g. del Giorgio and Peters 1994). Because phos-phorus concentration data are more generally avail-able compared to primary production estimates,we further examined the generality of the relation-ship between sediment respiration and variablesindicative of lake trophic status.

The analysis of our data compilation revealed thattotal phosphorus concentration was a very goodpredictor of sediment respiration (standardized to10◦C according to equation (6)) explaining 69% ofthe variance in sediment respiration and yielded thefollowing empirical equation:

log10 SRstd 10◦C = 0.17 + 0.58 log10 TP, (8)

where SR is in mmoles O2 m−2 d−1 and total phos-phorus (TP) is inmilligram permeter cube (Fig. 7.5).Because total phosphorus is also a well-known pre-dictor of chlorophyll concentrations and primaryproduction in lakes (Dillon and Rigler 1974 among

“chap07” — 2004/11/8 — page 112 — #10


log

10S

R s

td 1

0 °

C (

mm

olm

-2d

-1)

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

.5 1 1.5 2

log10 Total Phosphorus (mg m-3)

log

10S

R s

td 1

0 °

C (

mm

olm

-2d

-1)

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

.5 1 1.5 2


log

10S

R s

td 1

0 °

C (

mm

olm

-2d

-1)

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

.5 1 1.5 2


Figure 7.5 Log–log relationship between standardizedsediment respiration rates (SR, mmoles m−2 d−1) andtotal phosphorus concentrations (mg m−3). Individualdata points were obtained from Provini (1975), Ramlalet al. (1993), Schallenberg (1993), Hayes and MacAulay(1959), Graneli (1978),Welch and Kalff (1974), Lasenby(1975), and Hargrave (1972). Open circle (LakeSouthport, Hayes and MacCauley 1959) excluded fromregression line.

Figure 7.6 Inferred relationship between the fraction of carbon flux(as % ) to sediments that is eventually respired, as a function ofchlorophyll concentration (mgm−3)

many others, Smith 1979, del Giorgio and Peters1994), sediment respiration can also be expressedas a function of chlorophyll (mgm−3) or primaryproduction:

log10 SRstd 10◦C = 0.63 + 0.40 log10 Chl, (9)

log10 SRstd 10◦C = 0.093 + 0.47 log10 PP, (10)

where primary production (PP, mgCm−3 d−1) isvolumetric primary production averaged over theeuphotic zone (del Giorgio and Peters 1994). Theshallow slopes of all three relationships imply thatSR does not increase proportionately with anymeasure of lake trophic status. The consequences

of this disproportionality can be further exploredby combining these trends with the work of Bainesand Pace (1994) who showed that the vertical fluxof carbon can be well estimated (r2 = 0.83) fromchlorophyll concentration. Assuming a respiratoryquotient of 0.9 (Graneli 1979, although the trendis not affected by the particular choice of RQ), thefraction of the vertical flux of carbon that, at steady-state, will be respired in the sediments (%Cresp)will be a declining function of lake trophic status.Expressed as a function of chlorophyll, the resultingequation is:

%Cresp = 71.4 Chl−0.22 (11)

This inferred relationship, illustrated in Fig. 7.6suggests that, in oligotrophic lakes, about 70% ofthe carbon raining down to the sediments will berespired while this fraction is considerably reducedin more eutrophic systems (to about 40% in lakeswith 20 mgm−3 chlorophyll). Long-term carbonburial is, therefore, increasingly important in moreproductive systems, a conclusion also reached byHargrave (1973).

Other factors influencing sediment respirationWhile lake productivity and temperature are clearlythe main determinants of sediment oxygenic res-piration, other factors are known to influence sed-iment respiration in lakes, albeit to lesser extents.Sample depth, sediment organic matter content,ambient oxygen concentrations are among the

“chap07” — 2004/11/8 — page 113 — #11


most cited factors influencing sediment respirationalthough general tests of their importance are rare.Lake acidification has also been shown to reducerespiration in some instances (Andersson et al. 1978).

Unlike water column respiration processes,sediment respiration is suppressed when ambientoxygen concentrations decrease (e.g. Edberg 1976;Graneli 1978; Cornett and Rigler 1984), particu-larly when they drop below 30–60 mmolO2 m−3

(Anderson et al. 1986). This is probably the resultof the supply of oxygen to the sediment/waterinterface becoming limited by the reduced diffu-sion. Whether this reduced oxygen consumption isfully compensated by increased anaerobic decom-position is unclear (Bédard and Knowles 1991).However, we suggest that the failure of empiricalmodels such as equations (4)–(6) to incorporatethis first-order concentration effect may not overlycompromise their utility.

The influence of the depth at which sedimentcores are taken has also been noted by some authors(e.g. Campbell 1984; den Heyer and Kalff 1998) butthe effect appears significant only when compar-ing sites differing in depth by orders of magnitude.Given that very deep lakes are usually oligotrophic,it may be that the depth effect is a proxy fora trophic status effect. Indeed, variations amongdepths within individual lakes are nonexistent ormodest (Lasenby 1975; Newrkla 1982). In our datacompilation, adding site depth as an additionalindependent variable was not significant (p > 0.05)once lake trophic status (as total phosphorus) wasentered in the model, providing further indicationthat the depth acted as a surrogate in earlier models.

7.4 Free water respiration

By frequent in situmeasurements of oxygen dynam-ics over diel cycles, it is possible to calculate respi-ration. This so-called “free water” approach (Odum1956) has the advantage of not requiring enclosureand incubation. The method also measures the con-tribution of all components, from bacteria to fish, tototal system respiration and oxygen consumption.Anopen questiondiscussed further below is the ver-tical and horizontal scales over which the estimateof respiration is made. This technique potentially

Figure 7.7 Diel oxygen dynamics as percent saturation in areference and experimental lake (Paul and Peter Lakes, respectively,Michigan, USA). Details in text.

provides an integrated measurement of pelagic andsediment respiration in lakes.

Respiration can be estimated from the night-timedecline in oxygen once external exchanges (e.g.with the atmosphere) are accounted for. Figure 7.7presents an example from a reference lake and anexperimental lake fertilizedwith nitrogen and phos-phorus of diel oxygen dynamics measured in situwith autonomous instruments such as rapid-pulsedoxygen electrodes (details of study in Cole et al.2000). Oxygen is undersaturated in the referencelake and diel dynamics are modest while in theexperimental lake oxygen is ovesaturated and thereis a large diel excursion (Fig. 7.7). Under these con-ditions, the decline of oxygen at night is a functionof respiration and net diffusion to or from the atmo-sphere. Oxygen diffusion at the base of the mixedlayer can be ignored (Cole and Pace 1998). Atmo-spheric gains or losses of oxygen were calculatedas the difference between measured and saturatedoxygen times a coefficient of gas exchange. The gas

“chap07” — 2004/11/8 — page 114 — #12


exchange coefficient is either modeled or measureddirectly (e.g. Clark et al. 1995; Cole and Caraco1998; Wanninkhof andMcGillis 1999). Diel cycles ofoxygen can also be used to estimate gross primaryproduction based on increases in oxygen duringthe day and the assumption that night and dayrespiration are equal.

Respiration calculated from diel oxygen dyna-mics in the mixed layer of lakes should representan integrated response of the surface mixed layerand sediments on the edge of the lake that liewithin the mixed layer. Hence, the pelagic and sedi-ment respiration are included in the estimate. Inthe small lakes (1–3 ha) illustrated in Fig. 7.8, freewater respiration is greater than dark bottle respira-tion. For example in the unfertilized, reference lake,weekly measures of dark bottle respiration weremade during the same time as diel oxygen dynam-ics were measured (Fig. 7.8). Dark bottle respirationaveraged 8.8 mmolm−3 d−1 while free water res-piration averaged 30.6 mmolm−3 d−1. While darkbottle respiration might underestimate planktonicrespiration, the free water method indicates thatbenthic respiration must be about twofold higherthan planktonic respiration in the reference lake.Thus, in small water bodies much of the overall sys-temmetabolism can occur in sediments, particularlyshallow littoral sedimentswhere benthic respirationis stimulatedbyprimaryproduction ofmacrophytesand periphyton.

Such discrepancies between littoral and pelagicrespiration can create spatial heterogeneities in gasconcentrations that must be contended with. Thus,while the free-water method offers a greater inte-gration of all the respiring components than bottlemeasurements, the approach is not without diffi-culties. For instance, continuous mapping of sur-face P co2 shows that the spatial heterogeneity inthe distribution of respiration can be substantial,even in small lakes. Figure 7.9 illustrates the spa-tial distribution of the night-time increase in sur-face P co2 in a small embayment (0.53 km2) ofLake Memphremagog (Québec). Over the bay, theincrease averaged a modest 28 µatm P co2 but somelittoral areas increased by as much as 140 µatm.As expected, the amplitude in the daily variation ismuch greater in the shallow littoral zones. Clearly,horizontal diffusivity is not sufficiently large tohomogenize gas distribution at the scale necessaryto use the free-water method without considera-tion of spatial heterogeneities. More interestingly,the night-time increase was significantly smallerin the deepest zone of the bay where samples arenormally taken or autonomous sondes deployed.Thus, significant biases may be incurred if theautonomous sondes are placed only in the deepestarea of the lake, because the metabolism originat-ing from the more active littoral areas does notcompletely homogenize even over small horizontalareas.

Figure 7.8 Comparison of free water and bottle estimatesof respiration over the summer season in Paul Lake(Michigan, USA). Details in text.

“chap07” — 2004/11/8 — page 115 — #13


0.5 0 0.5 1 Kilometers

N

EW

S

–10 – 10 10 – 30 30 – 50 50 – 60 60 – 80 80 – 100 100 – 120 120 – 140

µatm

Figure 7.9 Spatial distribution of the night-timeincrease in pCO2 in a small bay of L. Memphremagog(Québec) showing large heterogeneities.

7.5 Global estimate of lake respiration

Although freshwater systems occupy a small arearelative to the surface area of the world’s oceans,freshwaters are comparatively more productive,leading to higher pelagic and sediment respirationrates per unit volume and area, respectively. Inaddition, lakes have a much larger relative contactarea with surrounding terrestrial ecosystems thanoceans, and substantial terrestrially derived organicmatter is oxidized during its freshwater transit(e.g. Richey et al. 2002). Theplanktonic and sedimentrespiration models presented above can be used toroughly estimate annual carbon processing in lakesat a global scale. While such calculations are neces-sarily fraught with large uncertainties and untestedassumptions, they can nevertheless provide order-of-magnitude estimates useful for casting the roleof freshwaters in the larger context of global carboncycling.

Our calculations are based on several simpleallometric relationships with lake surface area.Lake size is negatively related to lake abundance

(Meybeck 1995) and positively related to aver-age depth (Patalas 1984). Large lakes also tend tohave longer water residence times (Kalff 2002), avariable associated with reduced productivity onaverage (del Giorgio and Peters 1994). Figure 7.10depicts changes in lake abundance, depth, and pro-duction (as chlorophylla) across the logarithmic lakesize classes of Meybeck (1995). While about two-thirds of the total lake surface area is occupied bylakes less than 320 km2, more than 70% of totalvolume is in very large lakes (>320 km2). This asym-metry has large implications for the distribution ofrespiration across the lake size classes.

For shallow lakes (less than or equal to 15-mmeandepth), we calculated totalwater column respirationas theproduct ofmeandepth and thevolumetric res-piration rate estimated from the chlorophyll values(equation 2), to which we added sediment res-piration, also a function of chlorophyll (equation 9).For large and deep lakes (greater than 15-m meandepth, we integrated planktonic respiration only forthe upper 15 m, because respiration deeper in thewater column is considerably lower than respiration

“chap07” — 2004/11/8 — page 116 — #14


Figure 7.10 Global distribution of lake surface areas (km2) used in the global estimate of lake respiration. Lake surface areas are arranged inlogarithmic size classes following Meybeck (1995) and its relationships are shown, on the left axis, with lake total abundance (open circles, innumbers), and on the right axis, with lake mean depth (open squares, in m), and chlorophyll (open triangles, in mg m−3), as used in our modelestimates of global respiration. Note the line break for mean depth indicates that for large lakes unconnected mean depths points in the figurewere not used in the calculations, see text for details.

in the uppermost layers. For example, in the deep(100 m) central basin of Lake Memphremagog, epil-imnetic plankton respiration has been measured at10.5 mmolm−3 d−1 (del Giorgio and Peters 1994)and average hypolimnetic plankton respiration atabout 0.4 mmolm−3 d−1 (Cornett and Rigler 1987).Thus, our method of estimation, while empirical,has the advantage of providing a conservative esti-mate of respiration by not inflating respiration inthe vast volume ofwater contained in the deep, coldlayers of the relatively fewvery large lakes that dom-inate global lake volume (e.g. Lake Baikal). This is,however, a poor assumption in a number of waysas we discuss further below and results in what isalmost certainly an underestimate.

Integration of the estimate of respiration over lakesize classes yields a value of 62 Tmol C a−1, aboutone quarter of which occurs in the two largest lakesize classes (Fig. 7.11 and Table 7.1). Globally, sedi-ments are responsible for only about 6% of totallake R. Although our global R estimate is rather

small when compared to the total respiration of theoceans (del Giorgio and Duarte 2003), the import-ance of this figure ismore revealingwhen comparedto the gross primary production (GPP) that occurs inlakes. There are several publishedmodels that relatelake primary production to total phosphorus andchlorophyll. We used the relationship of Carignanet al. (2000), because it yields higher values foroligotrophic lakes than all other models, and thusprovides an upper estimate of pelagic primary pro-duction. Assuming then that water column GPPfollows the relationship of Carignan et al. (2000)with respect to chlorophyll, and that the depth ofthe euphotic zone is either equal to the mean depthof small lakes or 15 m for larger lake size classes,global pelagic GPP is estimated to be 51 Tmol C a−1.In all lakes there is some additional primary pro-duction from benthic and littoral communities, andthis littoral production is especially important insmall and oligotrophic lakes (Vadeboncouer et al.2003). If we assume benthic primary production is

“chap07” — 2004/11/8 — page 117 — #15


Figure 7.11 Cumulative global lake respiration (%) withincreasing lake size classes (km2).

Table 7.1 Range and averages for planktonic and sediment respiration and of global estimates derivedfrom lake size classes. For the conversion of oxygen to carbon the calculation assumes PQ = RQ = 1

Range Average(median)

Field observationsPlanktonic respiration

Volumetric (mmol O2 m−3 d−1) 0.7–162 15.7 (10.3)

Areal (mmol O2 m−2 d−1) 10–184 71.3 (61.5)

Sediment respiration (mmol O2 m−2 d−1)

Primary data 1.6–32 10.3 (9.7)Standardized at the rate of 10◦C 1.8–31 10.7 (9.2)

Global estimates

Lake surfacearea (km2)

Global surfacearea (1012 m2)

Respiration(TmolesC a−1)

GPP(TmolesC a−1)

Planktonic Sediment

0.01–0.1 0.38 1.3 0.7 1.50.1–1 0.37 2.8 0.6 2.91–10 0.35 6.2 0.6 5.810–102 0.34 12.5 0.5 11.4102–103 0.33 11.4 0.4 10.3103–104 0.32 9.7 0.4 8.9104–105 0.30 7.9 0.3 7.5105–106 0.29 6.1 0.3 6.0

Sum 2.69 58 3.8 54

“chap07” — 2004/11/8 — page 118 — #16


half of pelagic primary production in the smallestlakes and diminishes by a factor of 0.5 for successivelake size classes, then we obtain a global estimateof gross primary production of 54 Tmol C a−1. Theestimateddifference between global lake respirationand global lake production using these assumptionsis 8 Tmol C a−1. This value of excess respiration isreasonably close to the 12 Tmol a−1 of global lakeCO2 evasion estimated byCole et al. (1994) using theaverage partial pressure of carbon dioxide in lakes.This correspondence, however, is not fortuitous aswe constrainedour calculations to fall in the realmoftheCole et al. estimate. There are two aspects of com-paring our respective estimates that need to bemadeexplicit. The CO2 evasion from lakes estimated byCole et al. mayhave other sources beyond excess res-piration. This would suggest the excess respirationshould be lower than CO2 evasion. The global lakeareausedbyCole et al. (1994) excluded thevery largeinland seas (e.g. Caspian, Aral) and is, therefore,not directly comparable with our estimate. Exclud-ing our two largest lake size classes (yielding a totallake area of 2.1× 1012 m2 more commensurate withthe Cole et al. value of 2 × 1012 m2) further reducesthegapbetweenglobal lakeprimaryproduction andrespiration in our estimate to 6 Tmol C a−1.

The greatest problem with the assumptions weuse is that they exclude respiration in the deeperwaters of large lakes. This respiration is probablyquite significant. For example,McManus et al. (2003)recently estimated the annual respiration for thehypolimnion of Lake Superior as about 0.9 Tmolof C, or 30 mmolm−2 d−1. This estimate points tothe significance of large lakes in the overall calcula-tion of global lake respiration. If we apply the arealestimate for Lake Superior to all large lakes (thefour largest lake size classes), this would increaseour estimate of global lake respiration by about14 TmolC, to a total of 76 Tmol C a−1. This maybe a better estimate, but there are problems withconcluding lake respiration is much higher. Indeed,our assumptions tend to provide a realistic butupper limit to global lake GPP because the relation-ship we used yields very high estimates of pelagicproduction relative to other models. Thus, greatlyincreasing respiration would result in an unrea-sonable excess of respiration over production. One

possible explanation is that GPP is too low. We can-not, however, justify a larger value. A second andrelated problem with increasing the estimate is thatthe respiration increases mainly in the larger lakesize classes. Respiration of >75Tmol C a−1 leadsto P : R ratios for large lakes of about 0.65 on anannual scale, which may not be consistent with esti-mated fluxes and concentrations of O2 and CO2. Insummary, our calculations cannot at present be com-pletely reconciled. Better global data and models oflake respiration and primary production are neededas well as information on fluxes that can be usedto constrain these processes such as carbon dioxideevasion. Our estimates point to the critical lack ofunderstanding of respiration in large lakes and howrespiration varies with depth in lakes.

The global fluxes of carbon related to lake respira-tion are relatively small when compared to rates forthe ocean and land. However, inland waters maystill be significant as sites of active carbon cyclingand carbon losses. Thus, the fluxes of carbon asso-ciated with lake respiration need to be accountedfor when estimating net ecosystem production andcarbon storage as well as evasion of the combinedterrestrial and freshwater landscape especially inregions rich in lakes.

References

Andersson, G., Fleischer, S., and Granéli, W. 1978. Influ-ence of acidification processes in lake sediments. Verh.Int. Verein. Limnol., 20: 802–807.

Anderson, G. L., Hall, P. O. J., Iverfeldt, Å., Rutgers vander Loeff, M. M., Sundby, B., and Westerland, S. F. G.1986. Benthic respiration measured by total carbonateproduction. Limnol. Oceanogr., 31: 319–329.

Baines, S. B. and Pace, M. L. 1994. Sinking fluxes along atrophic gradient: patterns and their implications for thefate of primary production. Can. J. Fish. Aquat. Sci., 51:25–36.

Bédard, C. and Knowles, R. 1991. Hypolimnetic O2consumption, denitrification, and methanogenesis in athermally stratified lake. Can. J. Fish. Aquat. Sci., 48:1048–1054.

Bender, M., Grande, K., Johnson, K., Marra, J., Hunt, C.,Donaghay, P., andHeinemman, C. 1987.Acomparisonoffour methods for the determination of planktonic com-munity metabolism. Limnol. Oceanogr., 32: 1087–1100.

“chap07” — 2004/11/8 — page 119 — #17


Berman, T., Parparov, A., and Yacobi, Y. Z. 2004. LakeKinneret: Planktonic community production and respi-ration and the impact of bacteria of photic zone carboncycling. Aquat. Microb. Ecol. 34: 43–55.

Biddanda, B. A. and Cotner, J. B. 2002. Love handlesin aquatic ecosystems: the role of dissolved organiccarbon drawdown, resuspended sediments, and terrige-nous inputs in the carbon balance of Lake Michigan.Ecosystems, 5: 431–445.

Biddanda, B., Ogdahl, M., and Cotner, J. 2001. Dominanceof bacterial metabolism in oligotrophic relative toeutrophic waters. Limnol. Oceanogr., 46: 730–739.

Campbell, P. J. 1984. Laboratory Measurement and Pre-diction of Sediment Oxygen Consumption. MSc Thesis,McGill University, Montreal, 229 pp.

Carignan, R. and Lean, D. R. S. 1991. Regeneration ofdissolved substances in a seasonally anoxic lake: therelative importance of processes occuring in the watercolumn and in the sediments. Limnol. Oceanogr., 36:683–707.

Carignan, R., Blais, A.-M., and Vis, C. 1998. Measurementof primary production and community respiration inoligotrophic lakesusing theWinklermethod.Can. J. Fish.Aquat. Sci., 55: 1078–1084.

Carignan, R., Planas, D., and Vis, C. 2000. Planktonicproduction and respiration in oligotrophic Sheild lakes.Limnol. Oceanogr., 45: 189–199.

Carpenter, S. R., Cole, J. J., Hodgson, J. R., Kitchell, J. F.,Pace, M. L., Bade, D., Cottingham, K. L., Essington, T. E.,Houser, J.N., andSchindler, D.E. 2001. Trophic cascades,nutrients, and lake productivity: experimental enrich-ment of lakes with contrasting food webs. Ecol. Monogr.,71: 163–186.

Clark, J. F., Schlosser, P., Wanninkhof, R., Simpson, H. J.,Schuster, W. S. F., andHo, D. T. 1995. Gas transfer veloci-ties for SF6 and 3He in a small pond at lowwind speeds.Geophys. Res. Lett., 22: 93–96.

Cole, J. J. and Caraco, N. F. 1998. Atmospheric exchangeof carbon dioxide in a low-wind oligotrophic lake meas-ured by the addition of SF6. Limnol. Oceanogr., 43: 647–656.

Cole, J. J. and Pace, M. L. 1998. Hydrologic variabilityof small, northern lakes measured by the addition oftracers. Ecosystems, 1: 310–320.

Cole, J. J., Caraco, N. F., Kling, G. W., and Kratz, T. K. 1994.Carbon dioxide supersaturation in the surface waters oflakes. Science, 265: 1568–1570.

Cole, J. J., Pace, M. L., Carpenter, S. R., and Kitchell, J. F.2000. Persistence of net heterotrophy in lakes duringnutrient addition and food web manipulation. Limnol.Oceanogr., 45: 1718–1730.

Cornett, R. J. 1982. Prediction and interpretation of ratesof hypolimnetic oxygen depletion. PhD Thesis, McGillUniversity, Montreal, 273 pp.

Cornett, R. J. and Rigler, F. H. 1984. Dependence ofhypolimnetic oxygen consumption on ambient oxygenconcentration: Fact or artefact? Water Resour. Res., 20:823–830.

Cornett, R. J. and Rigler, F. H. 1987. Decomposition ofseston in the hypolimnion. Can. J. Fish. Aquat. Sci., 44:146–151.

Davies, J. M., Hesslein, R. H., Kelly, C. A., and Hecky,R. E. 2003. pCO2 method for measuring photosynthesisand respiration in freshwater lakes. J. Plankton Res., 25:385–395.

Dean, W. and Gorham, E. 1998. Magnitude and signific-ance of carbon burial in lakes, reservoirs, and peatlands.Geology, 26(6): 535–538.

del Giorgio, P.A. 1992. The relationship between ETS (elec-tron transport system) activity and oxygen consumptionin lake plankton: a cross-system calibration. J. Plankton.Res., 14: 1723–1741.

del Giorgio, P.A. and Bouvier, T. C. 2002. Linking the phys-iologic and phylogenetic successions in free-living bac-terial communities along an estuarine salinity gradient.Limnol. Oceanogr., 47: 471–486.

del Giorgio, P. A. and Duarte, C. M. 2003. Total respirationin the global ocean. Nature, 420: 379–384.

del Giorgio, P. A. and Peters, R. H. 1993. Balance betweenphytoplankton production and plankton respiration inlakes. Can. J. Fish. Aquat. Sci., 50: 282–289.

del Giorgio, P. A. and Peters, R. H. 1994. Patterns inplanktonic P : R ratios in lakes: influence of lake tro-phy and dissolved organic carbon. Limnol. Oceanogr.,39: 772–787.

del Giorgio, P. A., Cole, J. J., Caraco, N. F., and Peters,R. H. 1999. Linking planktonic biomass and metabolismto net gas fluxes in northern temperate lakes. Ecology, 80:1422–1431.

den Heyer, C. and Kalff, J. 1998. Organic matter mineral-ization in sediments: a within- and among-lake study.Limnol. Oceanogr., 43: 695–705.

Dillon, P. J. and Rigler, F. H. 1974. The phosphorus–chlorophyll relationship in lakes. Limnol. Oceanogr., 19:767–773.

Duarte, C. M. and Agusti, S. 1998. The CO2 balance ofunproductive aquatic ecosystems. Science, 281: 234–236.

Duarte, C. M., Agusti, S., and Kalff, J. 2000. Particulatelight absorption and the prediction of phytoplanktonbiomass and planktonic metabolism in northeasternSpanish aquatic ecosystems. Can. J. Fish. Aquat. Sci., 57:25–33.

“chap07” — 2004/11/8 — page 120 — #18


Edberg, N. 1976. Oxygen consumption of sediment andwater in certain selected lakes. Vatten, 31: 330–340.

Falkowski, P. G. and Raven, J. A. 1997. Aquatic Photosynthe-sis. Blackwell Science, Malden, MA.

Fenchel, T. 1998. Bacterial Biogeochemistry: The Ecophysi-ology of Mineral Cycling, 2nd edition. Academic Press,307 pp.

Grande, K., Marra, D. J., Langdon, C., Heinemann, K.,and Bender, M. L. 1989. Rates of respiration in thelight measured in marine phytoplankton using an 18Oisotope-labelling technique. J. Exp. Mar. Biol. Ecol., 129:95–120.

Granéli, W. 1978. Sediment oxygen uptake in southSwedish lakes. Oikos, 30: 7–16.

Granéli, W. 1979. A comparison of carbon dioxide pro-duction and oxygen uptake in sediment cores from foursouth Swedish lakes. Holarct. Ecol., 2: 51–57.

Graneli, W., Lindell, M., and Tranvik, L. 1996. Photo-oxidative production of dissolved inorganic carbon inlakes of different humic content. Limnol. Oceanogr., 41:698–706.

Hargrave, B. T. 1969a. Similarity of oxygen uptake bybenthic communities. Limnol. Oceanogr., 14: 801–805.

Hargrave, B. T. 1969b. Epibenthic algal production andcommunity respiration in the sediments ofMarion Lake.J. Fish. Res. Board Can., 26: 2003–2026.

Hargrave, B. T. 1972. A comparison of sediment oxygenuptake, hypolimnetic oxygen deficit and primary pro-duction in Lake Esrom, Denmark. Verh. Int. Verein.Limnol., 18: 134–139.

Hargrave, B. T. 1973. Coupling carbon flow through somepelagic and benthic communities. J. Fish. Res. Board Can.,30: 1317–1326.

Hayes, F. R. and MacAulay, M. A. 1959. Lake water andsediment. V. Oxygen consumed in water over sedimentcores. Limnol. Oceanogr., 4: 291–298.

Hutchinson, G. E. 1957. Treatise of Limnology. Vol I: Geo-graphy, Physics and Chemistry. John Wiley and Sons,New York.

James, A. 1974. The measurement of benthal respiration.Water Res., 8: 955–959.

Kalff, J. 2002. Limnology. Prentice-Hall, USA.Krumbein, W. E., Paterson, D. M., and Stal, L. J. (eds).

1994. Biostabilization of Sediments. BISVerlag, Oldenburg,526 pp.

Kuznetsov, S. I. 1958. A study of the size of bacterial pop-ulations and of organic matter formation due to photo-and chemosynthesis in water bodies of different types.Verh. Int. Verein. Limnol., 13: 156–169.

Lasenby, D. C. 1975. Development of oxygen deficits in 14southern Ontario lakes. Limnol. Oceanogr., 20: 993–999.

Linsey, G. A. and Lasenby, D. C. 1985. Comparison ofsummer and winter oxygen consumption rates in atemperate dimictic lake. Can. J. Fish. Aquat. Sci., 42:1634–1639.

Livingstone, D. M. and Imboden, D. M. 1996. The pre-diction of hypolimnetic oxygen profiles: a plea for adeductive approach. Can. J. Fish. Aquat. Sci., 53: 924–932.

Lund, J. W. G., MacKereth, F. J. H., and Mortimer,C. H. 1963. Changes in depth and time of certainchemical and physical conditions and of the standingcrop of Asterionella formosa Hass. in the north basin ofWindermere in 1947. Philos. Trans. R. Soc. Lond. Ser. B,246: 255–290.

Luz B., Barkan, E., Sagi, Y., and Yacobi, Y. Z. 2002.Evaluation of community respiratory mechanisms withoxygen isotopes: a case study in Lake Kinneret. Limnol.Oceanogr., 47: 33–42.

McManus, J., Heinen, E. A., and Baehr, M. H. 2003.Hypolimnetic oxidation rates in Lake Superior: role ofdissolved organic material on the lake’s carbon budget.Limnol. Oceanogr., 48: 1624–1632.

Meybeck, M. 1995. Global distribution of lakes. InA. Lerman,D.M. Imboden, and J. R.Gat (eds)Physics andChemistry of Lakes, 2nd edition, Springer-Verlag, Berlin,pp. 1–35.

Newrkla, P. 1982. Annual cycles of benthic communityoxygen uptake in a deep oligotrophic lake (Attersee,Austria). Hydrobiologia, 94: 139–147.

Odum, H. T. 1956. Primary production in flowing waters.Limnol. Oceanogr., 1: 103–117.

Ohle, W. 1952. Die hypolimnische Kohlen-dioxid-Akkumulation als produktions-biologischer Indikator.Arch. Hydrobiol., 46: 153–285.

Pace, M. L. and Cole, J. J. 2000. Effects of whole-lakemanipulations of nutrient loading and food web struc-ture on planktonic respiration. Can. J. Fish. Aquat. Sci.,57: 487–496.

Pamatmat, M. M. 1965. A continuous-flow apparatus formeasuring metabolism of benthic communities. Limnol.Oceanogr., 10: 486–489.

Pamatmat, M. M. 1968. An instrument for measuring sub-tidal benthic metabolism in situ. Limnol. Oceanogr., 13:537–540.

Patalas, K. 1984. Mid-summer mixing depths of lakes ofdifferent latitudes. Verh. Int. Verein. Limnol., 22: 97–102.

Platt, T. and Jassby, A. D. 1976. The relationship betweenphotosynthesis and light for natural assemblages ofcoastal marine phytoplankton. J. Phycol., 12: 421–430.

Prairie, Y. T., Bird, D. F., and Cole, J. J. 2002. The sum-mermetabolic balance in the epilimnion of southeasternQuebec lakes. Limnol. Oceanogr., 47: 316–321.

“chap07” — 2004/11/8 — page 121 — #19


Provini, A. 1975. Sediment respiration in six Italian lakesin different trophic conditions. Verh. Int. Verein. Limnol.,19: 1313–1318.

Ramlal, P. S., Kelly, C. A., Rudd, J. W. M., and Furitani, A.1993. Sites of methyl mercury production in remoteCanadian shield lakes. Can. J. Fish. Aquat. Sci., 50:972–979.

Rich, P. H. 1975. Benthic metabolism in a soft-water lake.Verh. Int. Verein. Limnol., 19: 1023–1028.

Rich, P. H. 1979. Differential CO2 and O2 benthic com-munity metabolism in a soft-water lake. J. Fish. Res.Board Can., 36: 1377–1389.

Richey, J. E., Melack, J. M., Aufdenkampe, A. K., Ballester,V.M., andHess, L. L. 2002. Outgassing fromAmazonianrivers and wetlands as a large tropical source of atmo-spheric CO2 . Nature, 416: 617–620.

Roland, F., Caraco, N. F., Cole, J. J., and del Giorgio,P.A. 1999. Rapid and precise determination of dissolvedoxygen by spectrophotometry: evaluation of interfer-ence from color and turbidity. Limnol. Oceanogr., 44:1148–1154.

Sas, H. 1989. Lake Restoration by Reduction of Nutrient Load-ing: Expectations, Experiences, Extrapolations. AcademiaVerlag Richarz, St Augustin, 407 pp.

Schallenberg, M. 1993. The ecology of sediment bac-teria and hypolimnetic catabolism in lakes: The relativeimportanceof autochthonousandallochthonousorganicmatter. PhDThesis, McGill University, Montreal, 259 pp.

Smith, V. H. 1979. Nutrient dependence of primaryproductivity in lakes. Limnol. Oceanogr., 24:1051–1064.

Vadeboncoeur, Y., Jeppesen, E., Vander Zanden, M. J.,Schierup, H.-H., Christoffersen, K.C., and Lodge, D. M.2003. From Greenland to green lakes: cultural eutrophi-cation and the loss of benthic pathways in lakes. Limnol.Oceanogr., 48: 1408–1418.

Wanninkhof, R. and McGillis, W. R. 1999. A cubic rela-tionship between air-sea CO2 exchange andwind speed.Geophys. Res. Lett., 26: 1889–1892.

Weger, H.G., Herzig, R., Falkowski, P.G., andTurpin, D.H.1989. Respiratory losses in the light in a marine diatom:measurements by short-term mass spectrophometry.Limnol. Oceanogr., 34: 1153–1161.

Welch, H. E. and Kalff, J. 1974. Benthic photosynthesisand respiration in Char Lake. J. Fish. Res. Board Can.,31: 609–620.

Wetzel, R. G. 2001. Limnology: Lake and River Ecosystems,3rd edition. Academic Press, San Diego.

Documents

Respiration in lakes -