ECOTOXICOLOGY: A Comprehensive Treatment - Chapter 35 pdf

Here wewill focus on descriptive and experimental studies that assess effects of increased CO2, N deposition,acidification, and ultraviolet radiation UVR on the function of aquatic and t

by geopolitical boundaries InChapter 26, we described the effects of global atmospheric stressors onabundance, species diversity, community composition, and other structural characteristics Here wewill focus on descriptive and experimental studies that assess effects of increased CO2, N deposition,acidification, and ultraviolet radiation (UVR) on the function of aquatic and terrestrial ecosystems.

35.2 NITROGEN DEPOSITION AND ACIDIFICATION

move-ment of elemove-ments in aquatic and terrestrial ecosystems Significant increases in the global reservoirs

of C, N, and S as a result of combustion of fossil fuels and agricultural/land use changes disrupt thesenatural cycles and have contributed to a variety of local and global environmental concerns Globalemissions of biologically reactive N compounds (e.g., NH3, NH4, HNO3, and NO3) have increased

from about 15 teragrams (Tg) in 1860 to more than 165 Tg in 2000 (Galloway et al 2003) Althougheffects of increased N deposition have not attracted the same attention from scientists and the public

as other global atmospheric stressors such as chlorofluorocarbons (CFCs) and CO2, N poses seriousthreats to ecosystem processes Potential negative effects of excess N on forest ecosystems werefirst described by Nihlgard (1985) Biologically reactive N compounds that accumulate in the atmo-sphere are rapidly deposited on the earth’s surface where they can affect net primary productivity(NPP) (Aber et al 1995), disrupt N dynamics in soils (Gundersen et al 1998), and contribute toeutrophication (Rabalais et al 2002), acidification (Vitousek 1994), and subsequent loss of biolo-gical diversity (Stevens et al 2004) In addition, N2O is a potent greenhouse gas that contributes toglobal climate change Because the rates of production of reactive N in the biosphere greatly exceedrates of removal by denitrification, biologically active N rapidly accumulates in the environment.Predicting effects of N deposition on aquatic and terrestrial ecosystems is complicated by variation

in regional climate, hydrologic characteristics, vegetation type, and other sources of anthropogenicdisturbance (Aber et al 2003) Assessing effects of N on ecosystems is also complicated becausedeposition often co-occurs with other stressors, such as heavy metals (Gawel et al 1996) Finally,input and output of N are not necessarily coupled in all ecosystems, and leaching of nitrate willdepend on nutrient status (Gundersen et al 1998)

35.2.1 THENITROGENCASCADE

Accumulation, transfer, and denitrification of biologically reactive N compounds through thebiosphere and the changes that result have been termed the nitrogen cascade (Table 35.1)(Galloway et al 2003) Forests and grassland ecosystems, especially the soil components, are major

771

TABLE 35.1

Factors That Influence the Nitrogen Cascade in Aquatic and Terrestrial Ecosystems

Ecosystem

Accumulation Potential

Transfer Potential

Denitrification Potential

Biological Effects

sediments

Very high Moderate to high Biodiversity; altered community

structure; eutrophication Coastal marine Low to moderate;

higher in sediments

structure; algal blooms

Source: Modified from Table 1 in Galloway et al (2003).

reservoirs for N Because output of N in undisturbed forest and grassland ecosystems is generallyquite low, residence time can be many years Forest ecosystems are often N limited, and therefore

N is cycled internally with little export to surface water, groundwater, or the atmosphere As excess

N deposition increases, other environmental factors will limit NPP and unused N leaches below therooting zone, a process known as nitrogen saturation (Aber et al 1995) Land use changes in forestsand grasslands also have the potential to significantly alter internal N cycling and increase the export

of N to aquatic ecosystems and the atmosphere

Fertilizers and runoff associated with agricultural and urban areas are the primary contributors of

N to aquatic ecosystems and have been considered in previous chapters However, nitrogen oxides(NOx) from fossil fuel combustion account for about 25% of the reactive N in the environment.Although there is relatively little storage of N in overlying water, sediments represent a significantreservoir of N in aquatic ecosystems Enrichment of aquatic ecosystems by N deposition can result

in eutrophication, anoxia, and reduced biodiversity Although natural aquatic ecosystems are highlyretentive of N, this capacity for internal processing can be exceeded, especially in disturbed habitats,and downstream transport of N can contribute to eutrophication of coastal areas Transport of N fromrivers to coastal areas is generally regarded as one of the most serious threats to marine ecosystems(Rabalais et al 2002) Accumulation potential of N in estuaries is relatively low; however, as withfreshwater ecosystems coastal marine sediments may represent a significant N reservoir Because

of large amounts of organic material and low concentrations of dissolved oxygen in sediments,coastal marine ecosystems also have the greatest potential for conversion of reactive N to N2bydenitrification, a process that is often enhanced by excess reactive N In fact, denitrification in riversand estuaries greatly reduces the amount of N transported from terrestrial to coastal and offshoreareas (Galloway et al 2003)

35.2.2 EFFECTS OFN DEPOSITION ANDACIDIFICATION IN

AQUATICECOSYSTEMS

Although most of the earlier studies of atmospheric deposition focused on ecosystem effects of fur (S), attention in North America and Europe has shifted to concerns about effects of atmospheric Ndeposition Atmospheric deposition of S and N from the mid-1980s to the mid-1990s show contrast-ing temporal patterns in North America (Sirois et al 2001) While declines of SO4in precipitationwere observed, concentrations of NO3and NH4 generally remained constant or increased Thesechanges corresponded to decreased emissions of SO4and increased emissions of NOxover this sameperiod As previously N-limited forests became saturated, NO3is released to watersheds causing

sul-eutrophication, acidification and reductions in acid-neutralizing capacity (ANC; defined as the city of the watershed to withstand strong acid inputs based on the difference between total cationsand total anions) In fact, one of the most consistent indicators of N saturation in ecosystems is

capa-an increase in concentrations of NO3in stream water Detecting these trends in streams requiresaccess to long-term data that are often unavailable for many watersheds Much of the researchthat describes biogeochemical responses of streams to N deposition has been conducted in Europeand the northeastern United States In particular, experimental studies and long-term monitoring

of SO4 and NO3 in stream water at Hubbard Brook Experimental Forest documented tion effects on watersheds and significant decreases in base cation concentrations (Likens et al.1996)

acidifica-Assuming that ecosystems can only tolerate a certain level of acidification before importantprocesses become disrupted, defining the threshold point at which the capacity of an ecosystem

to withstand additional inputs is exceeded is an important exercise The concept of critical soilacidification load for a watershed is analogous to assimilative capacity in ecotoxicology The ideaassumes that once the net neutralizing capacity of an ecosystem is reached, additional atmosphericdeposition will result in soil acidification Critical soil acidification load is determined primarily by

a balance between the weathering of base cations and leaching associated with deposition of SO4

and NOx Moayeri et al (2001) developed a model to calculate critical soil acidification loads for awatershed in Ontario, Canada Model results showed that soil acidification would occur faster in aharvested watershed compared to an old-growth watershed

Even relatively remote areas located away from sources of N can experience effects of N ition In general, water quality in alpine and subalpine lakes of the Rocky Mountains is relativelypristine, with low background concentrations of NO3(Williams and Tonnessen 2000) A survey of

depos-44 high-elevation lakes (>3000 m a.s.l.) located on both sides of the continental divide in

Color-ado indicated that higher NO3concentration and lower ANC of eastern lakes corresponded withgreater atmospheric N deposition (Baron et al 2000) Long-term changes in community composi-tion and productivity of diatoms, as revealed by paleolimnological records of lake sediments, wereconsistent with increased eutrophication These increases in N deposition and shifts in diatom floracorresponded with increases in urban, agricultural, and industrial development on the Front Range

of Colorado Williams and Tonnessen (2000) used long-term monitoring data and synoptic surveys

of 91 high-elevation lakes in the central Rocky Mountains to establish critical loads for inorganic

N deposition Episodic acidification of sensitive headwater catchments in remote Wilderness Areashas resulted from increased wet deposition of N

The concept of N saturation, originally developed to describe export of N in forest ecosystems,may also apply to watersheds Similar to the acidification effects associated with S emissions,sensitivity of watersheds to N deposition will be influenced by underlying geology, hydrologiccharacteristics, soil type, and vegetation Alpine and high-elevation watersheds, especially thoselocated above treeline, may be especially susceptible to N deposition because of their relativelynonreactive bedrock, short growing season, and limited vegetation (Fenn et al 1998) Williams et al.(1996a) reported that N deposition in the catchments of the Colorado Front Range was similar to that

in other well-studied northeastern locations, including Hubbard Brook (New Hampshire) and AcadiaNational Park (Maine) A shift in nutrient dynamics from N-limited to N-saturated conditions wasalso observed in these high-elevation watersheds as a result of increased anthropogenic N deposition.Williams et al (1996a) suggested that N saturation in high-elevation catchments may serve as anearly warning of disruption in N cycling

Ecosystem responses of oligotrophic lakes to N deposition will also depend on nutrient conditionsand the history of NO3availability Nydick et al (2004) measured effects of NO3 enrichment intwo alpine lakes with very different background levels of N Enrichment significantly increased

photosynthetic rate and chlorophyll a in a low N lake, but had no effects on a high N lake Both NO3

and PO4additions were necessary to increase productivity in the high N lake Nydick et al (2004)also reported that despite relatively little effect on benthic algal biomass, epilithon, surface sediment,

and subsurface sediment accounted for 57–92% of the NO3 uptake, indicating the importance ofbenthic processes in these lakes.

Ecosystem-level effects of SO2and NOxdeposition on acidification of aquatic ecosystems hasbeen studied extensively in Europe and North America (Hornung and Reynolds 1995, Likens et al

1996, Schindler 1988) Much of the research conducted in the Experimental Lakes Area (ELA)focused on ecosystem processes, especially primary productivity (Schindler 1987, Schindler et al.1985) Results of these and other studies of acidification showed reduced rates of primary andsecondary production, decomposition, and nutrient cycling Although the whole lake experimentalstudies conducted at Little Rock Lake in northern Wisconsin focused on community responses toacidification (Gonzalez and Frost 1994), Frost et al (1999) speculated that loss of sensitive speciesand shifts in community composition would diminish the ability of acidified lakes to maintain systemfunction

Effects of acidification on organic matter processing have been examined experimentally innatural and artificial streams Burton et al (1985) measured effects of acidification on decom-position of white birch and sugar maple in experimental stream channels (Figure 35.1) Reduceddecomposition rates in acidified stream channels were attributed to lower density of macroinver-tebrates, particularly shredder caddisflies and detritivorous isopods It is interesting to note thatsignificant effects of acidification were not observed until relatively late in the study (>80 days),

demonstrating the importance of long-term experiments However, results of long-term ments do not necessarily demonstrate significant ecological effects Smock and Gazzera (1996)

eight remaining 30

40 50 60 70 80

100

White birch

FIGURE 35.1 Decomposition rate (as percent dry weight remaining) of sugar maple and white birch in

reference (closed symbols) and acidified (open symbols) stream channels (Data from Table 2 in Burton et al.(1985).)

introduced H2SO4to a low gradient, blackwater stream in Virginia (USA) Monthly additions over

a 1-year period reduced benthic microbial respiration, but had no effect on processing rates of redmaple leaves or abundance of macroinvertebrates These results suggest that ecological effects ofacidification on ecosystem processes will likely vary with location Blackwater streams with nat-urally high levels of tannins and organic acids that typify this region are relatively insensitive toacidification

35.2.3 EFFECTS OFN DEPOSITION ANDACIDIFICATION IN

TERRESTRIALECOSYSTEMS

Although negative effects of NOx deposition on lakes and streams have been frequently observed,changes in terrestrial ecosystems, especially forests, have received the most attention Depositionrates of N to forest ecosystems range from 2 kg N/ha/year in remote reference areas to 40 kg N/ha/year

in forests downwind of industrial sources (Aber et al 1989) A large-scale survey of 68 grasslandecosystems across Great Britain showed a strong negative relationship between species richness and

N deposition (Stevens et al 2004) Although this paper focused on changes at the community level,

it served to illustrate the widespread nature of this problem At the current rate of N deposition incentral Europe (17 kg N/ha/year), these researchers estimated that species richness was reduced byapproximately 23% compared to grassland ecosystems receiving the lowest levels Similar spatialgradients in N deposition and ecological effects are also evident in the northeastern United States.For example, relatively high rates of deposition have been measured in southern New York andPennsylvania (12 kg N/ha/year), whereas low rates of deposition are measured in eastern Maine(<4 kg N/ha/year) (Aber et al 2003) Although it is well documented that alpine and other high-

elevation ecosystems are at greater risk from N pollution because of higher rates of deposition andincreased sensitivity, other landscape factors that influence N deposition are not well understood.Direct measurement of wet and dry deposition rates in forest ecosystems is difficult Weathers et al.(2000) developed a model to predict the influence of several landscape factors on N deposition inmontane ecosystems Using concentrations of Pb in forest floor soils as an index of N deposition, theseinvestigators quantified effects of forest edges, elevation, aspect, and vegetation type on N deposition

in montane forests

Fenn et al (1998) published a comprehensive review of factors that predispose ecosystems to

N saturation and the general ecosystem responses to N deposition Because of the intimate linkagesbetween forests and surrounding watersheds, research describing transport of N in terrestrial ecosys-tems has important implications for N transport to lakes and streams For example, export of basecations, increased acidification, and elevated levels of nitrate and aluminum in streams are likely to

be associated with N deposition in forests Aber et al (1989) provided a formal definition of the term

N saturation and developed a hypothesized time course describing responses of forest ecosystems to

N deposition (Figure 35.2) At a critical stage in this sequence of events, forests will likely becomenet sources of N rather than sinks Aber et al (1989) also described the limited capacity of someforest ecosystems to assimilate excess N and the potential interactions of N deposition with otheratmospheric stressors such as ozone, sulfate, and heavy metals

35.2.3.1 The NITREX Project

One of the most comprehensive and spatially extensive experimental assessments of N deposition wasconducted in several coniferous forests in northeastern Europe The NITREX (nitrogen saturationexperiments) project involved experimental additions of N to sites along a gradient of N pollution

to examine changes in structural and functional characteristics (Emmett et al 1998) Input rates of

N ranged from 13 to 59 kg/ha/year across sites Similar to the lotic intersite nitrogen experiment(LINX) experiments described inChapter 30, an important objective of the NITREX project was

to compare responses to N addition among sites Experimental treatments involved both enhanced

Deposition begins

NPP

Foliar biomass

Foliar [N]

Nitrate assimilation

FIGURE 35.2 Predicted time course for changes in NPP, biomass, foliar N concentration, and nitrate (NO3)

assimilation in a forest ecosystem in response to chronic deposition of N (Modified from Figure 1 in Aber et al.(1989).)

N input at sites with naturally low atmospheric deposition and reduced N input (using exclusion roofsand an ion exchange system) at sites with high deposition Researchers predicted that N addition inN-limited forests would have relatively little effect on leaching because of the high capacity for Nretention in these systems Most of the observed responses to N manipulations were consistent withexpectations, with reductions in N status and NO3leaching occurring at sites where N depositionwas reduced and increases occurring at sites where N deposition was enhanced (Gundersen et al.1998) Shifts in N status and cycling rates following treatments generally supported the N saturationhypothesis (Aber et al 1989)

The most consistent responses to enhanced N deposition were changes in water quality, ticularly increased NO3 A strong relationship between N leaching and N status suggested that thedistinction between N-limited and N-saturated forests could be quantitatively demonstrated (Gun-dersen et al 1998) Nitrate leaching was observed within the first year of the experiments, whereasbiological responses were often delayed The strong link between N deposition and acidificationwas also demonstrated in the NITREX project N additions caused a decrease in ANC, whereasexperimental reductions in N caused an increase in ANC (Emmett et al 1998) Ratios of carbon

par-to nitrogen (C:N) were also good predicpar-tors of the onset of NO3 leaching (Figure 35.3) becausenitrification rates are stimulated as C:N declines, resulting in a decrease in the retention efficiency

of N (Emmett et al 1998) These results suggest a simple threshold response of N export At C:Ngreater than 24, only a small proportion of nitrate leaches (approximately 10%); however, as C:Ndecreases, a rapid increase in the proportion of NO3leached was observed

Comparison of N dynamics at N-limited and N-saturated sites showed that microbial cycling

of C and N was characterized by low NH4transformation and respiration rates at N-limited sites.Surprisingly, despite a wide range of variation in N deposition rates, N transformation showed relat-ively little variation Gross mineralization and immobilization rates of NH4were highly correlatedwith respiration rates across ecosystems (Figure 35.4), indicating the important linkage between Cand N dynamics and showing that simple measurements of CO2in soils could potentially serve as ameasure of N cycling (Tietema 1998)

The potential damaging effects of atmospheric deposition on forest productivity were clearly trated by the NITREX project Experimental reduction of N and sulfur inputs to a highly N-saturatedsite resulted in a 50% increase in tree growth (Emmett et al 1998) However, in general, ecosystemresponses to experimental manipulation of N were relatively modest and required longer periods

illus-C:N ratio of forest floor

FIGURE 35.3 Relationship between N leaching and carbon:nitrogen (C:N) ratios in the forest floor Data are

from experimental results of the NITREX project in northwestern Europe (Modified from Figure 5 in Emmett

FIGURE 35.4 Relationship between gross N transformation rates and respiration in forests soils from the

NITREX experiment Solid circles= gross NH4mineralization rate; open circles= gross NH4immobilizationrate (Data from Tables 2 and 3 in Tietema (1998).)

of time Boxman et al (1998) commented that it was “remarkable that no ecosystem componentshave responded with increasing vitality to the high N levels in the initially N limited, oligotrophicforests.” Mass loss in litter bags also increased along the gradient of N status, but effects of Nmanipulation were not significant Responses of soil fauna to N treatments were generally less thaninitial differences among sites along the N gradient The modest biological responses to N treatmentsmay have resulted from the relatively short duration of these experiments Gundersen et al (1998)provided estimates of the amount of time required for several chemical and biological responses to

N treatment (Table 35.2) In general, decomposition rates and changes in NPP were relatively slowprocesses compared with NO3leaching These results again highlight the importance of conductinglong-term manipulations for assessing ecosystem effects of N deposition

35.2.3.2 Variation in Responses to N Deposition among

Ecosystems

Responses of forest ecosystems to N deposition are often variable, and factors that control differences

in N retention and export are not completely understood As described above, C:N ratios in soil are

TABLE 35.2 Predicted Timing of Selected Forest Ecosystem Responses to Changes in Chronic Additions and Reductions in N Deposition

Fast = 1 year; intermediate = 2–4 years; slow ≥ 5 years.

Note: Results are based on experimental manipulations of

nitro-gen from the NITREX project.

Source: From Table 8 in Gundersen et al (1998).

linked to the capacity of forest ecosystems to retain N, and decreased C:N may occur under conditions

of chronic N deposition (Emmett et al 1998) Surveys of wet deposition in old-growth stands ofEngelmann Spruce showed elevated levels of NO3and NH4on the eastern side of the continentaldivide in Colorado (Baron et al 2000) Greater N deposition was attributed to agricultural andatmospheric sources and was reflected in higher rates of mineralization and nitrification Subsequent

N fertilization experiments conducted in Engelmann Spruce forests on both sides of the continentaldivide showed that differences in soil conditions influenced responses to N treatments (Rueth et al.2003) Mineralization rates were unaffected by N treatment on the western side of the continentaldivide but increased by approximately two times on the eastern side

Experimental studies of N deposition at the ELA in Ontario (Canada) have been conducted to trast responses of different ecosystem components A 2-year N addition experiment (40 kg N/ha/year)was conducted in a boreal forest to compare effects in N-limited “forest islands” with naturally N-saturated lichen outcrops (Lamontagne and Schiff 1999) Responses to N treatments differed betweenhabitat types In contrast to forests, N-saturated lichen outcrops were highly sensitive to N depos-ition After 2 years of treatment, lichen outcrops no longer retained additional N inputs, whereas theproportion of N retained in treated and reference forest–islands was similar (Figure 35.5) These datahighlight the role that relatively small habitat patches in a landscape play in controlling ecosystemprocesses

con-Differences in abundance of dominant species will also complicate our ability to predict tem responses to N deposition in forests For example, species-specific differences in litter qualityand rates of decomposition influence N cycling and availability Surveys conducted along a gradi-ent of atmospheric N deposition in the northeastern United States showed that responses differedbetween tree species (Lovett and Rueth 1999) Rates of mineralization and nitrification of soils weresignificantly related to N deposition in maple plots but not in beech plots Thus, while increasedrates of mineralization and nitrification are typical responses to N deposition in forest ecosystems,the species composition of these forests should be considered when developing predictive models.Similarly, changes in community composition of forest ecosystems as a result of natural or anthropo-genic disturbance will affect responses to N deposition Lovett and Rueth (1999) also demonstratedthat although descriptive studies do not allow researchers to directly infer causality, comparison

ecosys-of ecosystem responses to chronic N deposition along a gradient can be a useful alternative toexperimental treatments

Community type

Retention coefficient 0.00.5

1.0

Reference watershed Treated watershed

FIGURE 35.5 Responses of lichen bedrock communities and forest communities after 2 years of N additions

in the ELA, Ontario, Canada The figure shows retention coefficients ([Total N Input – Total N Output]/Total

N Input) (Data are from Table 4 in Lamontagne and Schiff (1999).)

35.2.4 ECOSYSTEMRECOVERY FROMN DEPOSITION

Mitigating the effects of N deposition on forest ecosystems will be challenging because of thewidespread distribution of sources and tremendous variation in ecological effects However, potentialfor recovery is relatively high if N deposition is significantly reduced Experiments conducted inThe Netherlands and Germany showed rapid improvements in water quality following reductions

in N deposition to a watershed (Emmett et al 1998) Increased tree growth was also observed attwo of the three sites where N deposition was reduced (Boxman et al 1998) Soils are a major sinkfor excess N in forests, and recent evidence suggests that microbial assimilation of NO3may be animportant regulator of N retention Because major reductions in N emissions are unlikely for someareas, management options should also include strategies to enhance the incorporation of N intosoils (Fenn et al 1998)

Recovery of watersheds from the long-term effects of acidification may require many years ifbase cations are significantly depleted Reductions in atmospheric deposition of sulfate as a result

of the Clean Air Act have been associated with significant improvements in stream water chemistry.However, export of base cations, especially Ca and Mg, from acidified watersheds will likely delayrecovery (Likens et al 1996) Although the loss of base cations has been attributed primarily toprolonged exposure to acid rain and a decline in Ca in precipitation, ecological factors also influence

Ca export Hamburg et al (2003) measured levels of Ca in the forest floor, abundance of snails(organisms that require Ca for growth), and Ca export in stream water from hardwood forests ofvarious ages Results showed that Ca concentration in snails, litterfall, and the forest floor andexport of Ca in stream water increased with forest age Calcium mobilization in young stands(4.6–6.0 g Ca/m/year) was much greater than in old stands (0.4 g Ca/m/year), indicating that forestaging significantly influenced Ca dynamics

Recent amendments to the 1990 Clean Air Act in the United States are expected to have ficant effects on air quality and water chemistry across large broad geographic regions Assessingthese changes will require integrated studies of physicochemical and biological responses over largeregional areas and for relatively long periods of time Because long-term monitoring data at a regionallevel are generally lacking, researchers are often required to use trends from site-specific results toinfer regional patterns Stoddard et al (1998) analyzed trends in water chemistry from 44 Adirondackand New England (USA) lakes that were sampled from 1982 to 1994 Long-term trends in meas-ures of acidic deposition (SO4and NO3concentrations in stream water) and watershed responses

signi-to acidification (ANC and export of base cations) differed between subregions In particular, ANCincreased over time in New England lakes but decreased in Adirondack Lakes These results were not

expected and indicate that potential for recovery from acid deposition would be considerably less inthe Adirondacks The most significant finding of this research is that even with long-term data and asolid mechanistic understanding of physicochemical relationships, predicting regional trends based

on well-defined subpopulations of sentinel lakes is difficult (Stoddard et al 1998) The challenges ofmaking accurate regional predictions on the basis of a very well-understood phenomenon highlightthe potential difficulties associated with quantifying effects of poorly defined stressors such as CO2

and UV-B radiation

35.3 ULTRAVIOLET RADIATION

to really demonstrate UV-B radiation impacts at the ecosystem level requires establishing a chain of

cause and effect from molecule to ecosystem

mar-of phytoplankton from Lake Michigan by 25% Exposure to simulated UV-B reduced photosynthesis

by 40% in Georgian Bay (Furgal and Smith 1997) Remarkably, even short-term exposure to surfaceradiation (e.g., 30 min) can be sufficient to inhibit photosynthesis (Marwood et al 2000)

Because of proximity to the ozone depletion zone and the intense exposure to UVR during theearly austral spring (October–November), considerable research effort has focused on phytoplankton

in Antarctic waters, where a 50% reduction in ozone has been documented Smith et al (1992)conducted transects in the Antarctic marginal ice zone and reported a 6–12% reduction in primaryproduction associated with ozone depletion during a 6-week cruise Assuming that this reduction isrepresentative of the entire area and integrating results over the austral spring, Smith et al (1992)concluded that this change corresponded to an approximately 2% reduction in annual production ofthe Southern Ocean Similar experiments conducted in the Weddell Sea showed that productivity ofmarine phytoplankton decreased as a cumulative function of UV exposure, indicating little evidence

of photorepair (Neale et al 1998) These researchers also documented considerable variation insensitivity of primary production among sites, and attributed this variation to UV exposure beforesampling

35.3.1.1 Methodological Considerations

With respect to experimental methodology, most aquatic investigations involved the removal ofdifferent wavelengths of UVR using various filters, although some have employed experimental

enhancement of UVR using lamps In contrast, studies conducted in terrestrial ecosystems havemore commonly employed lamps to enhance UVR Because of the significantly elevated levels ofUV-B in the Southern Hemisphere, researchers have relied primarily on UV exclusion methodology.

In contrast, experimental designs in the Northern Hemisphere where UV increases are less dramatichave more commonly employed UV enhancement (Flint et al 2003) These are important method-ological distinctions with significant implications for how results are interpreted UVR exclusionexperiments document effects of current levels of UVR, whereas enhancement experiments attempt

to estimate effects of predicted increases Advantages and disadvantages of the different experimentaltechniques used to enhance or reduce UVR were described inChapter 26 In particular, there hasbeen considerable discussion of the artificial wavelength spectra produced by UV lamps Unreal-istic combinations of UV-A, UV-B, and photosynthetically active radiation (PAR) may artificiallyenhance effects of UV-B (Caldwell and Flint 1997) These problems may be partially addressed byeither calculating biological spectral weighting functions or by using modulated lamp systems thatmeasure incoming UV-B and adjust lamp outputs accordingly Comparisons of ecosystem processesunder ambient and UVR-excluded treatments provide important information on effects of currentlevels of UVR However, it may be difficult to extrapolate these results to conditions of enhancedUVR because it requires that we make predictions beyond those used in the experiments (Behrenfeld

et al 1995) More importantly, because of the significant influence of PAR on photosynthesis, it isessential that filtering materials that exclude and transmit UV-B allow the same amount of PAR(Flint et al 2003) Because of these issues, some combination of UV exclusion and enhancementmay be necessary to reliably estimate effects of current and increased UVR under conditions ofozone depletion

The duration of experiments designed to assess UVR effects on primary production is also animportant consideration In contrast to research conducted in terrestrial systems, most investigations

of UVR effects in aquatic ecosystems have been limited to relatively short-term experiments Watkins

et al (2001) measured effects of UVR on epilithic metabolism, pigment concentrations, nutrients,

and community composition in a boreal lake over a 4-month period Although chlorophyll a was

not affected, photosynthetic rates were increased by 37–46% and shifts in community compositionwere observed when UVR was eliminated Most of the observed response was a result of exposure

to UV-A These results strongly support the hypothesis that current levels of UVR penetratingclear-water lakes have detrimental effects on primary productivity Because the experiments wereconducted in summer and fall, investigators were able to document seasonal responses to decliningUVR Although differences among treatments were negligible in fall as a result of lower incidentUVR, differences in taxonomic composition persisted

35.3.1.2 Factors that Influence UV-B Exposure and Effects in

radi-dark Effects of UV-B were greater on phytoplankton than on benthic algal mats, which were likelyprotected by UVR-absorbing amino acids.

Habitat features, behavioral characteristics, and morphological adaptations of organisms willinfluence exposure and sensitivity to UVR Microcosm experiments conducted with artificial light toenhance UV-B showed no effects on phytoplankton, zooplankton, periphyton, or macroinvertebrates(De Lange et al 1999) Some species-specific responses were observed, but overall ecosystem char-acteristics were unaffected The lack of a response in this system was attributed to UV-B attenuationresulting from high concentrations of dissolved organic carbon (DOC), which protected organisms

from exposure Interestingly, bioassays conducted with Daphnia pulex showed higher growth of

organisms fed seston from the control microcosms than organisms consuming seston from B-treated microcosms These results suggest the intriguing possibility that energy transfer fromphytoplankton to zooplankton could be affected by UV-B

UV-Because penetration of UVR through the water column is dependent on water clarity, factors thatinfluence turbidity, trophic status, and levels of DOM will potentially influence ecosystem responses.Anthropogenic changes in water clarity, such as those resulting from acidification, climate change, orexotic species will also influence UVR exposure Invasion of exotic filter-feeding zebra and quagga

mussels (Dreissena spp.), which remove phytoplankton from the water column, has significantly

increased water clarity and UVR penetration in lakes Experiments conducted in Lake Erie (USA)showed that UVR inhibited primary production, but that effects were mediated by N availability(Hiriart et al 2002) There are also concerns over the sustainability of planktonic food webs inLake Erie as a result of removal of phytoplankton by dreissenid mussels The stability of these foodwebs will be further compromised if UVR affects phytoplankton production Allen and Smith (2002)observed that UVR significantly inhibited phosphate uptake capacity in plankton, which may result

in a potential negative feedback by diminishing P availability in this system

Vertical profiles of primary production showed that relative effects of UV-A and UV-B on tosynthesis vary with depth and water clarity (Palffy and Voros 2003) As expected, greatest effects

pho-on photosynthesis were observed near the surface, but these effects were primarily attributable toUV-A (Figure 35.6) A multiple regression model showed that UVR and vertical light attenuationaccounted for 90% of the variation in photoinhibition Effects of UV-B on ecosystem processes may

be ephemeral and change as a result of alterations in community composition and maturity Santos

et al (1997) measured successional changes in tropical marine diatoms exposed to varying UVRtreatments in the field During initial stages of colonization, primary production was reduced by

>40% when exposed to a full solar spectrum of UV-A+ UV-B + PAR These changes corresponded

to differences in composition of diatoms among treatments Effects of UVR treatments were reducedover time, suggesting that diatoms were most sensitive during the initial stages of succession

PAR only PAR+UV-A PAR+UV-A+UV-B

FIGURE 35.6 Vertical profile of phytoplankton primary production in the western basin of Lake Balaton

(Central Europe) measured on July 19, 1999 (Data from Table 1 in Palffy and Voros (2003).)

Quar tz Pyre

x (305) Mylar (323)Plastic (378)

0 2 4 6 8 Antarctic phytoplankton

Tropical phytoplankton

FIGURE 35.7 Photosynthetic rates (µg C/µg Chl a/h) of Antarctic and tropical phytoplankton when solarradiation is filtered using quartz, pyrex, mylar, and plastic film Numbers in parentheses are the wavelengths(in nanometers) corresponding to 50% transmission for each treatment (Data from Figure 9 in Helbling et al.(1992).)

As described previously, documented losses of ozone and subsequent increases in UVR havebeen greatest in Antarctic marine ecosystems Because phytoplankton in these southern oceans havehistorically been exposed to relatively low levels of UVR, it is likely that they are especially sensitive

to anthropogenic increases Helbling et al (1992) compared effects of UVR on photosynthesis oftropical and Antarctic phytoplankton populations Results showed relatively little effects of elim-inating UVR on tropical phytoplankton but dramatic effects on Antarctic organisms (Figure 35.7).These researchers also noted that most of these effects were a result of reducing UV-A, whereasUV-B had relatively minor effects on photosynthesis In contrast to these findings, Banaszak andNeale (2001) observed that photosynthesis of phytoplankton from a shallow estuarine environ-ment was more strongly inhibited by UV-B than UV-A Biological weighting functions (BWFs)that quantified effects of different wavelengths on photosynthesis were similar to those derived forAntarctic systems and showed relatively little seasonal variation, despite considerable variation inphysicochemical characteristics in this ecosystem (Banaszak and Neale 2001)

35.3.1.3 Comparing Direct and Indirect Effects of UVR on

Ecosystem Processes

Because UVR affects both primary producers and consumers, responses to UVR manipulationsobserved in field experiments are often a combination of direct and indirect effects Organisms

representing different trophic levels will likely show differential sensitivity to UVR Despite siderable speculation that indirect effects of UVR will be important, relatively few studies havedocumented these food web responses Research by Bothwell et al (1994) was one of the firststudies to quantify the importance indirect effects on benthic communities Although accrual rates

con-of algae were initially inhibited by UVR, changes in abundance con-of algal consumers (chironomids)mediated these responses McNamara and Hill (2000) measured effects of UV-B on photosynthesisand food resources available to grazers in experimental streams These researchers observed a dose–response relationship between UV-B irradiance and photosynthesis in both short- (4-h) and long-term(13-day) experiments

Direct or indirect UVR-induced changes in food webs can have important consequences foraquatic ecosystems If UV-B inhibits growth and nutrient uptake of primary producers or alters sizeand species composition, the quality of food resources for grazers may be affected (Hessen et al.1997) Tank et al (2003) measured direct and indirect effects of UVR in four montane lakes of varyingwater transparency in Jasper National Park, Alberta (Canada) Results of mesocosm experimentsusing filters showed that UVR altered trophic structure and function of benthic communities, butdirect and indirect effects were highly variable among lakes In contrast to expectations, exposure

to UVR generally did not reduce the quality or quantity of food resources to invertebrates UVRexposure decreased species richness and resulted in lower photosynthetic pigments in organisms fromtwo clear lakes, but other factors such as nutrient concentration and grazers were more importantthan UVR in structuring communities (Tank and Schindler 2004) Indirect effects of UVR on foodwebs in a British Columbia (Canada) stream varied among locations, but were generally weakcompared to direct effects (Kelly et al 2003) This study also failed to show effects of UVR onfood quality Vinebrooke and Leavitt (1999) manipulated UVR and density of macroinvertebrates

in an oligotrophic alpine lake to test the relative importance of direct and indirect effects of UVR.Responses of primary producers and consumers varied by species and habitat, with greatest effectsobserved on epilithic standing crop These researchers speculated that direct effects of UVR would bemore important in extreme environments, such as alpine lakes or other stressed ecosystems, whereabiotic factors regulate ecosystem processes The hypothesis that UVR will have greater effects

in stressed ecosystems has important implications for understanding potential interactions betweenUVR and contaminants and will be considered in Section 35.5.3

Most studies investigating effects of UVR in marine and lentic ecosystems have focused

on inhibition of photosynthesis However, a more comprehensive understanding of the potentialecosystem-level effects of UVR requires that other processes be considered Mesocosm experimentsusing natural assemblages of marine phytoplankton showed that exposure to enhanced UV-B signi-ficantly affected N transport rates (Mousseau et al 2000) Research conducted by Behrenfeld et al.(1995) also documented effects of UV-B on N uptake in natural plankton assemblages collected frommid-latitudes of the North Pacific Ocean Results showed that exclusion of UV-B increased uptake ofammonium and nitrate compared to ambient levels, whereas enhancement of UV-B reduced uptake.These researchers also established dose–response relationships between N uptake and UV-B dose.Results of these analyses showed that rates of N uptake were more sensitive to UV-B than C fixation,suggesting that assessment of effects based exclusively on photosynthesis may underestimate totalUV-B damage to ecosystems (Behrenfeld et al 1995)

35.3.1.4 Effects of UV-B on Ecosystem Processes in

Benthic Habitats

Effects of UVR in benthic communities have received considerably less attention than planktoniccommunities, presumably because these organisms should be protected by overlying water andbecause UVR does not penetrate into sedimentary habitats However, benthic communities occupy-ing clear, shallow water environments are likely to be exposed to intense levels of UVR In addition,Garcia-Pichel and Bebout (1996) reported that UVR penetrated a range of sediment types, with

relatively low attenuation in sandy quartz sediments where effects on photosynthetic organisms arelikely to be significant These predictions are consistent with results of experiments measuring UVReffects on benthic algal and meiofaunal communities Odmark et al (1998) exposed microbenthiccommunities collected from sandy sediments to several UVR treatments After 3 weeks of expos-ure to natural UV-B, carbon fixation rates were significantly reduced as compared to the no UV-Btreatments These researchers speculated that UV-B would have greater effects on communitiesinhabiting sandy sediments compared to sediments with a high silt and clay content Roux et al.(2002) observed reduced photosynthesis in microphytobenthic communities (primarily small diat-oms) from an intertidal mudflat exposed to UV-B; however, these effects were limited to periods ofhigh solar irradiance Finally, unlike some planktonic organisms that are able to avoid UVR in thephotic zone, behavioral avoidance in some benthic habitats is limited Because benthic algae anddiatoms can account for a significant portion of primary production in aquatic ecosystems, exposure

to UVR could have serious consequences for energy flow

35.3.2 EFFECTS OFUVRINTERRESTRIALECOSYSTEMS

While the major focus of UVR research in aquatic ecosystems has been on primary production,research in terrestrial ecosystems has documented effects on other ecosystem processes, includinglitter decomposition and biogeochemical cycles (Newsham et al 1997, Pancotto et al 2003, Zepp

et al 1995) The consensus of these investigations is that terrestrial ecosystem processes are generallyless sensitive to UVR than processes in aquatic ecosystems Caldwell and Flint (1994) predicted thatthe occurrence of UVR effects on plants from most frequent to least frequent was the following:increased production of UV-absorbing compounds> reduced growth and morphological changes 

reduced photosynthesis

35.3.2.1 Direct and Indirect Effects on Litter Decomposition

and Primary Production

Effects of UVR on litter decomposition have been described as a result of both direct and indirectprocesses Direct effects on decomposing litter are usually a result of inhibition of microbial, fungal,and other components of the soil community, which reduces decomposition rates Because theseeffects may be offset by enhanced photodegradation, which enhances decomposition rate, predictingdirect effects of UV-B on litter decomposition rates is complex Indirect effects of UV-B occur duringgrowth and senescence of plants and can result in changes in leaf chemistry (e.g., lignin content) orphysical characteristics of leaves One of the most consistent responses of plants to UV-B exposure isincreased production of protective secondary plant metabolites, including phenolics and flavonoids

If these changes in leaf chemistry influence feeding habits of other trophic levels or alter plant–herbivore interactions, there exists the possibility that higher trophic levels will be indirectly affected

by UV-B (Bassman 2004) In most instances these secondary plant compounds serve as deterrents

to herbivory and therefore are likely to mediate trophic responses to UV-B radiation Althoughaquatic ecologists routinely consider implications of cascading trophic interactions, these ideas havereceived less attention from terrestrial ecologists (Bassman 2004), perhaps because top-down control

in terrestrial ecosystems is considered relatively unimportant (Strong 1992) Nonetheless, the oftensubtle direct effects of UV-B on terrestrial plants may be less important than the indirect effects ontrophic interactions

Because effects of UV-B exposure will likely vary among locations, comparisons of plant munities across sites is a valuable approach for understanding factors that determine ecosystem-level

com-effects Moody et al (2001) measured direct effects on litter decomposition of Betula pubescens

exposed to ambient and elevated UV-B at sites in Norway, Sweden, the Netherlands, and Greece.Although the fungal community was significantly affected by UV-B, differences in mass loss andchemical composition of litter between treatments were modest Verhoef et al (2000) also reported

that litter decomposition and nutrient fluxes in a grassland ecosystem were not affected by UV-B;however, abundance of soil decomposers was significantly reduced in both UV-A and UV-B treat-ments There is also likely to be a strong seasonal component to UVR effects that will vary amongterrestrial ecosystems For example, UVR exposure to leaf litter in deciduous forests is likely to begreatest in early spring when leaf canopies are absent and incident UV-B is high Newsham et al.

(1997) observed subtle and transient effects of enhanced UV-B on decomposition of oak (Quercus robur) leaf litter Lower decomposition in UV-B treatments was associated with increases in C con-

tent of leaves and reduced fungal colonization However, in a subsequent study of UV-B effects on

decomposition, Newsham et al (2001) reported that Q robur saplings exposed to a 30% increase

in UV-B (corresponding to an 18% reduction in ozone) for 2 years showed little change in chemicalcomposition These researchers concluded that recent increases in UV-B in the Northern Hemisphereare unlikely to have significant effects on organic matter pools, nutrient cycling, and decomposi-tion through alterations in litter quality Experiments conducted at high latitudes of the SouthernHemisphere where ozone depletion is greatest showed quite different results Pancotto et al (2003)employed a 2× 2 factorial experimental design to assess both direct and indirect effects of UV-B on

a native shrub community in Tierra del Fuego National Park (Argentina) Plants were grown underambient or reduced UV-B and decomposition rate of litter produced by these plants was measuredunder ambient or reduced UV-B Decomposition rate (mass loss) was significantly (14–34%) lowerunder ambient UV-B compared to reduced UV-B treatments These direct effects were found to bemore important in controlling decomposition rates than indirect effects on litter quality Pancotto

et al (2003) speculated that changes in decomposition rates have important implications for otherecosystem-level processes, including nutrient mineralization and carbon storage, in high latitudes ofthe Southern Hemisphere

Although effects of UV-B on primary production and nutrient cycling have been examined interrestrial habitats (Klironomos and Allen 1995, Gehrke 1998, Shi et al 2004), these processeshave received considerably less attention compared to aquatic ecosystems Assessing direct effects

on primary productivity is complicated because UVR can either increase or decrease physiological

processes that determine production For example, growth of Sphagnum in a subarctic bog was

significantly reduced by exposure to UV-B (Gehrke 1998) However, total production was notaffected because photosynthesis was enhanced and dark respiration was reduced Klironomos and

Allen (1995) exposed sugar maple (Acer saccharum) seedlings to enhanced UV-B and measured

shoot and root biomass Despite significant shifts in belowground carbon flow and microarthropodabundance in UV treatments, shoot and root biomass was not affected Plants inhabiting alpineecosystems are naturally exposed to greater levels of UV-B and are therefore expected to possessrepair mechanisms to reduce the damaging effects on photosynthesis In field experiments, Shi

et al (2004) exposed alpine plants to enhanced UV-B radiation that simulated a 14% reduction

in ozone depletion Photosynthesis and respiration were either similar or increased slightly undermoderate UV-B exposure These researchers speculated that alpine plants are acclimated to UV-Band that photosynthetic processes are protected by morphological adaptations such as increased leafthickness

Meta-analysis offers a quantitative approach for integrating results of multiple studies to assesscomplex relationships among variables This approach is especially appropriate for assessing ter-restrial ecosystem responses to UV-B because effects are expected to be relatively subtle and oftenindirect Searles et al (2001) conducted meta-analysis of 62 papers that investigated effects of UV-Bradiation on the concentration of UV-B-absorbing compounds, growth, morphological variables,and photosynthetic processes With the exception of UV-B-absorbing compounds, most variablesshowed relatively minor response to UV-B treatments These researchers concluded that indirecteffects in the form of alterations in herbivory are likely to be the most significant responses ofterrestrial ecosystems to elevated UV-B radiation

Finally, our understanding of effects of UV-B on terrestrial ecosystems is seriously limited by thelack of long-term investigations (Aphalo 2003) In the meta-analysis of terrestrial studies described

above (Searles et al 2001), over 80% of the studies were conducted for less than 1 year Long-termstudies are even less common in aquatic ecosystems where manipulation of UVR is more problematicbecause of experimental artifacts Although short-term experiments may help understand underlyingmechanisms, they often provide very different results than those of longer duration Experimentsconducted in Tierra del Fuego represent one of the best examples of long-term UV-B studies (Robson

et al 2003) These researchers used filters to reduce ambient levels of UV-B in a peatland ecosystemfor six field seasons It is important to note that 6 years was an insufficient time period to detectsubtle effects of UV-B for several of the responses measured

35.4 INCREASED CO2 AND GLOBAL CLIMATE

CHANGE

35.4.1 AQUATICECOSYSTEMS

Despite widespread recognition of the potential ecological effects of global climate change ated with increased levels of atmospheric CO2, research on ecosystem-level responses in aquaticsystems has been lacking Several excellent reviews describing predicted effects of climate change

associ-on distributiassoci-on and extirpatiassoci-on of species have been published (Carpenter et al 1992b, Clark et al

2001, Firth and Fisher 1992, Grimm 1992, Lodge 2001, Meyer et al 1999, Smith and Buddemeier1992); however, relatively few studies have investigated effects on ecosystem processes A recentreport published by the Pew Center on Global Climate Change (Poff et al 2002) summarized thecurrent state of knowledge on effects of climate change on aquatic ecosystems, but contained verylittle information on changes in ecosystem function (Table 35.3) It is expected that increased sur-face water temperatures associated with global climate change will affect ecosystem productivity,materials transport, nutrient dynamics and decomposition; however, little data have been collected tosupport this hypothesis Increased water temperature will likely increase rates of respiration and pho-tosynthesis, and the relative magnitude of these increases will determine overall effects on ecosystemmetabolism A survey of factors related to lake productivity along a latitudinal gradient showed thatprimary production was directly related to water temperature (Brylinsky and Mann 1973) Long-termrecords (1970–1990) indicated that increased air temperature and reduced precipitation in northwest-ern Ontario were most likely responsible for reduced discharge, increased water temperature, greaterlight penetration, and reduced concentration of DOC in boreal lakes (Schindler et al 1996) Because

TABLE 35.3

Major Conclusions of the Pew Center on Global Climate Change Report Regarding Aquatic Ecosystem Responses to Global Climate Change

• Aquatic and wetland ecosystems are very vulnerable to climate change.

• Increases in water temperature will cause a shift in the thermal suitability of aquatic habitats for resident species.

• Seasonal shifts in stream runoff will have significant negative effects on many aquatic ecosystems.

• Wetland loss in boreal regions of Alaska and Canada is likely to result in additional releases of CO 2 into the atmosphere.

• Coastal wetlands are particularly vulnerable to sea level rise associated with increasing global temperatures.

• Most specific ecosystem responses to climate change cannot be predicted because new combinations of native and nonnative species will interact in novel situations.

• Increased water temperatures and seasonally reduced streamflows will alter many ecosystem processes with potential direct societal costs.

• The manner in which humans adapt to a changing climate will greatly influence the future status of inland freshwater and coastal wetland ecosystems.

Source: Poff et al (2002).

of cascading trophic-level interactions in many lake ecosystems, alterations of one trophic level willlikely have consequences for both upper and lower trophic levels Results of microcosm experimentswith aquatic microbes showed that warming increased primary production and decomposition by bothdirect effects on temperature-dependent physiological processes and indirect effects on trophic struc-ture Finally, climate-induced alterations in the composition of riparian canopies may have significanteffects on the quality and quantity of allochthonous detritus delivered to lakes and streams.

35.4.1.1 Linking Model Results with Monitoring Studies in

Aquatic Ecosystems

Much of the research in lakes and streams documenting potential effects of climate change has beenlimited to hydrologic models that predict modifications in discharge resulting from altered precip-itation patterns Alterations in the flow regime of aquatic ecosystems are likely to be significant,especially in western U.S watersheds where modest changes in precipitation are expected to res-ult in dramatic reductions in stream runoff (Carpenter et al 1992a) Long-term records of streamdischarge are available for many watersheds (e.g., U.S Geological Survey); therefore, predictivemodels that relate regional changes in climate to altered flow regimes within a watershed can bedeveloped Associating local weather patterns and stream discharge over the past several decadesmay also provide useful insights into potential trends associated with global climate change How-ever, extrapolation from global and regional models to local conditions may not be appropriate forsome areas For example, general circulation models (GCMs) for the Rocky Mountains predict thattemperature should increase under a two times CO2scenario However, long-term monitoring inthis region showed a decline in mean annual temperature and an increase in precipitation (Williams

et al 1996b), demonstrating that climate in alpine areas may be controlled more by local conditionsthan by regional trends

The oceans have long been recognized as an important sink for excess CO2 released to theatmosphere Recent evidence indicates that approximately 48% of the anthropogenic C releasedbetween 1800 and 1994 was sequestered by oceans and that, without this oceanic uptake, atmospheric

CO2levels would be about 55 ppm greater than current levels (Sabine et al 2004) These authorsalso suggest that the strength of the oceans as a sink for atmospheric CO2has diminished and thatthe fraction of CO2currently stored in the oceans is approximately one-third of the total potentialstorage Relatively complex feedback mechanisms will determine the effects of increased CO2andglobal temperatures on oceanic ecosystems Using results of long-term (1958–2002) surveys ofmarine phytoplankton in the Northeast Atlantic Ocean, Richardson and Schoeman (2004) associatedincreased sea surface temperatures with increased primary production in cooler areas and decreasedproduction in warmer areas Changes in primary production were also related to production ofgrazers, which propagated to other trophic levels The effects of changes in primary production onglobal carbon flux are difficult to predict because other factors, especially nutrient availability, willdetermine CO2uptake In one of the more ambitious attempts to understand the relationship amongnutrients, primary production, and CO2flux, Coale et al (2004) performed multiple iron injectionexperiments in large areas (15 km2) of the Southern Ocean Rates of photosynthesis increased

from 0.29 to 6.9 mmol C/m3/day and nitrate concentrations decreased by approximately 2 µMfollowing treatment, indicating that iron may play an important role in controlling CO2uptake in thisregion

Unlike research in marine ecosystems, the role of freshwater ecosystems as a source or sinkfor atmospheric carbon has received little attention Community metabolism varies greatly amongaquatic ecosystems, with the highest productivity observed in marshes (Figure 35.8) Although it

is generally assumed that freshwater ecosystems export CO2to the atmosphere, the importance ofaquatic biota as sources or sinks depend on overall ecosystem productivity (Duarte and Agusti 1998).Factors such as nutrient enrichment and trophic structure also regulate primary production and CO2

flux in freshwater ecosystems Whole ecosystem experiments conducted in Wisconsin (USA) showed

Production Respiration

FIGURE 35.8 Median gross primary production and respiration of freshwater and marine ecosystems Results

were compiled from five decades of studies that reported O2evolution as a surrogate for carbon flux (Datafrom Table 1 in Duarte and Agusti (1998).)

that shifts in top predators in experimentally enriched lakes regulated the flow of C and determined

if a lake was a sink or source of CO2(Schindler et al 1997)

The relative lack of information concerning potential effects of climate change on primary duction, nutrient cycling, decomposition, and other processes in aquatic ecosystems is surprisingand in sharp contrast to tremendous research efforts currently underway in terrestrial habitats.Considerably more research effort is necessary to understand responses of aquatic ecosystems

pro-to elevated CO2 and associated climate change Climate-induced changes in water ure, hydrology, and physicochemical characteristics in aquatic ecosystems are likely to influencecontaminant transport, bioavailability, uptake, and toxicity Aquatic ecotoxicologists need todevelop a better appreciation for how these processes will be affected in a warmer, CO2-enrichedworld

temperat-35.4.2 TERRESTRIALECOSYSTEMS

In contrast to the relatively limited research on effects of climate change in freshwater ecosystems,studies conducted in terrestrial ecosystems have been extensive These studies have considered sev-eral facets of the CO2 problem (Figure 35.9): direct influences of CO2 enrichment and indirecteffects of increased temperature (Körner 2000) and shifts in terrestrial vegetation (Wolters et al.2000) Much of this research has focused on assessing the role of forests, grasslands, and otherecosystems in C sequestration, a problem that requires a better understanding of the complex rela-tionship between CO2and C storage and the role of soil nutrients, especially N, in regulating storage.Quantifying global C sequestration is a difficult problem because of variation among ecosystems andbecause of complex feedback processes For example, boreal peatlands occupy only about 2% of theearth’s surface but sequester about 33% of the global soil C Increased soil temperature associatedwith climate change could result in positive feedback by increasing decomposition rate of peatlandsthereby releasing large quantities of stored C to the atmosphere Alternatively, warmer temperaturescould cause negative feedback by enhancing productivity and C storage in these areas Storage of

C in terrestrial ecosystems will be largely dependent on the rate of turnover within C pools and theavailability of N Carbon allocated to pools with relatively fast turnover such as leaves and rootswill likely result in little long-term storage In contrast, C allocated to pools with slow turnover willresult in a large increase in soil C (Allen et al 2000) Finally, if elevated CO2significantly increases