Click for next page ( 89


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 88
7 BIOGEOCHEMICAL FEEDBACKS AND THE CARBON CYCLE SUMMARY Both the marine and terrestrial carbon cycles contain potentially important feedback processes. There are, however, major gaps in understanding. No definitive explanation has been given for the vast uptake of CO2 by the terrestrial biosphere, and no confident prediction can be given of future biological uptake or release of CO2, particularly over the long term. Few observations are available to guide the necessary scaling of vegetation- climate feedbacks from the scale of an individual leaf to a landscape mosaic of vegetation and soils. In the marine realm the strengths of a wide variety of potential feedback mechanisms involving CO2 and DMS are yet to be determined. Research into carbon uptake by the land and ocean as outlined in the U.S. Carbon Cycle Plan (Sarmiento and Wofsy, 1999) and North American Carbon Program (Wofsy and Harriss, 2002) should be undertaken to characterize and reduce the uncertainty associated with carbon uptake feedbacks. The Panel also recommends that research outlined in the Surface Ocean Lower Atmosphere Study (SOLAS) Science Plan be adopted in order to improve our understanding of DMS-climate feedbacks as well as carbon cycle feedbacks that involve air-sea transfer (such as iron-CO2 feedbacks). The U.S. Carbon Cycle Science Plan outlines a strategy to "deliver credible prediction of future atmospheric carbon dioxide levels . . . by means of approaches that can incorporate relevant interactions and feedbacks of the carbon-cycle climate system." The plan advocates strong multiagency collaboration to carry out specific program elements, which include (1) expanded, long-term observational networks in the atmosphere, ocean and terrestrial systems; (2) historical reconstructions of CO2 emissions and terrestrial carbon inventories; (3) intensive ocean and land process studies; and (4) modeling and synthesis, including the development of models that 88

OCR for page 88
BIOGEOCHEMICAL FEEDBACKS AND THE CARBON CYCLE 89 couple the carbon cycle to the rest of the climate system. SOLAS is an international research initiative designed to "achieve quantitative understanding of the key biogeochemical-physical interactions and feedbacks between the ocean and atmosphere, and of how this coupled system affects and is affected by climate and environmental change." To achieve this goal the SOLAS Science Plan recommends increased cooperation between atmospheric and marine scientists in order to develop process studies, monitoring programs, process-level models, and Earth- system models. AS greenhouse gases increase in the atmosphere and warming is produced, the net exchange of carbon between the atmosphere and reservoirs of carbon in the land and ocean may be altered. Temperature and precipitation changes may alter the uptake of carbon by plants. Increased temperature in high latitudes may change the storage of carbon by frozen soils and associated biomass. Changes in ocean temperature and circulation may alter the storage of carbon in the ocean. All these potential feedback processes will alter the amount of atmospheric carbon dioxide increase that results from fossil fuel combustion by humans. The production and uptake of other radiatively active gases in the land and ocean may also be modified as a result of climate change. The land and ocean currently exchange approximately 120 and 90 petagrams of carbon per year with the atmosphere, respectively (Prentice et al., 2001~. Although the ocean constitutes a much larger reservoir of carbon than the land biosphere, both land and ocean carbon exchanges are important for understanding the anthropogenic effect on atmospheric carbon dioxide. Both land and ocean also have the potential to produce feedbacks between climate change and uptake of anthropogenic carbon. The quantities of carbon stored as plant biomass and soil organic matter on land, or carbonate species and organic carbon in the sea, vastly exceed CO2 in the atmosphere. Analysis of long-term changes in atmospheric CO2, ~3CO2/~2Co2, and O2 show that the atmospheric increase in CO2 was less than half of the fossil fuel input between 1991 and 1997, with the remainder approximately equally partitioned among the land and ocean (Battle et al., 2000).

OCR for page 88
9o UNDERSTANDING CLIME TE CHANGE FEEDBACKS TERRESTRIAL CARBON FEEDBACKS Climat~Plant~O2 Feedbacks Atmospheric CO2 is regulated by complex processes involving terrestrial and marine plants, which fix inorganic carbon as organic matter, heterotrophic organisms that mineralize organic matter back to CO2, and a variety of geochemical and biogeochemical processes that convert CO2 to and from mineral carbonates (e.g., Cached. All of these processes are sensitive to climate. Nevertheless, assessments of climate change have long regarded feedbacks in the carbon-climate system as basically simple two- step processes, as depicted in Figure 7.1 for the terrestrial biosphere: Faster I growth I ~ + ~1 W ' ~ Conventional views of CO2-climate feedbacks. (a) Temperature(T)-respiration feedback. Terrestrial systems respond to climate . . . . . . warming Dy increasing respiration, adding CO2 to Me atmosphere from stocks of soil organic matter, increasing CO2, arid enhancing warming. | CO2 1 (a) CO=growthfeedback. Plants increase _ ~ ~ Am+ rates of photosynthesis when grown at ~ ~ _ ~ , elevated concentrations of CO2, ~, ~ Enhanced photosynthesis, reduced ET especially in dry climates or in nutrient- rich soils where other factors do not inhibit the response to CO2. Evapotrasporation (ET) is reduced, lowering the water requirement for vegetation. FIGURE 7.1 Climate-land biosphere feedback processes: Conventional view. The positive feedback loop (a) is based on the increased rate of respiration observed for almost all organisms as temperatures increase. This factor underlies the paradoxical distribution of soil organic matter with latitude. Rates of production of organic matter are slower in cold versus warm climates, but rates of decomposition decline faster than production. Huge stocks of organic carbon, several times larger than the quantity of CO2

OCR for page 88
BIOGEOCHEMICAL FEEDBACKS AND THE CARBON CYCLE 91 in the atmosphere, are locked up in the soils of boreal and sub-boreal regions, and feedback (a) could thus have a major impact on future levels of CO2. The negative feedback (b), plant growth accelerated by CO2, is also a well-known biophysical process. Green plants all use the enzyme rubisco to bind CO2 during photosynthesis. Rubisco takes carbon dioxide and attaches it to ribulose bisphosphate, a small sugar with five carbon atoms; then it cuts the molecule into two identical pieces with three carbon atoms. In spite of its central role rubisco is remarkably inefficient. Typical enzymes process 1,000 molecules so, but rubisco fixes only about three carbon dioxide molecules per second. High concentrations of CO2 (roughly 260 ppm) are needed to bind with the enzyme in a cell. Plants compensate for the inefficiency by allocating substantial resources to rubisco, and most plants must allow rapid gas exchange with the interior tissues of the leaf to provide the needed high concentrations of CO2. This circulation, through opening of the stomates of the leaf, allows water loss by evaporation. Elevated CO2 thus allows plants to increase growth with fixed (or reduced) allocation to rubisco and with lower requirements for water. The CO2-growth feedback modifies the quantity of atmospheric CO2 by altering the amount of organic matter in living biomass and the inputs of fresh organic matter to soils, in contrast to the respiration feedback that alters the quantity of dead organic matter in soils, much of which is old and recalcitrant. The stocks of biomass and short-lived organic matter that may be maintained on the land impose the limit for the CO2-growth feedback. These stocks are subject to manipulation by harvesting, preservation of wood and paper, and other management. The limit on the temperature- respiration feedback is imposed by the available stores of soil organic matter, generally assumed to be larger than potential biomass stocks. Real ecosystems do not however behave just like simple organisms exposed to a single, instantaneous change in the environment. For example, some ecosystems show quite small stimulation by elevated CO2, and responses typically decline during extended studies. Several factors are at work. Stomates may remain open despite higher CO2, to restrain the rise in leaf temperature; moreover, reduced water use provides little help to plants in well-watered environments. Other resources, such as nutrients (N. P. Ca, K), often limit plant growth, inhibiting any stimulation by CO2 (Bauer et al., 2001). Some critically important feedbacks occur only on long time scales. For example, the length of the growing season has been shown to provide the dominant effect of climate on carbon sequestration by mid-latitude forests. Years with warm temperatures in spring have greater net uptake of CO2 than

OCR for page 88
92 UNDERSTANDINrG CLIMATE CHANGE FEEDBACKS cold years (Barford et al., 2001~. Greater rates of growth and carbon uptake are observed for mid-succession forests in warmer parts of the temperate zone. These effects far exceed any increase in respiration, contradicting expectations of the temperature-CO2 feedback (a). The peatlands of Alaska, Canada, and Siberia represent a very important, potentially positive, feedback between CO2 and climate (Chapin et al., 2000~. Enormous quantities of carbon have accumulated as peat since the end of the last ice age, equivalent to 200 ppm or more of atmospheric CO2 (Gorham, 1991~. Peat is preserved by being saturated with water, maintained in the low-precipitation boreal environment by very slow evaporation, or by being frozen. Peatlands that become drier are subject to fairly rapid oxidation, either by microbial activity or by natural fires (Goulden et al., 1998; Harden et al., 2000~. Evidence suggests that this process is occurring at present, and it could accelerate markedly according to some climate scenarios. The key lies in future changes in regional precipitation at least as much as with temperature. Figure 7.2 illustrates two of the feedbacks between the climate system and the terrestrial system. The same interactions viewed at the landscape scale and long times may have strong feedbacks opposite to those inferred for single organisms subjected to instantaneous perturbations of a single environmental variable (cf Fig. 7.1~. A Scientific Strategy for Terrestrial Carbon Feedbacks There is currently no definitive explanation for the vast uptake of CO2 by the terrestrial biosphere, nor is there a confident prediction of future uptake or release of CO2 from the terrestrial biosphere. The major issue is to determine the responses of whole ecosystems and landscapes to the full diversity of environmental changes attending climate change. Warming per se is likely less important than other factors, such as precipitation, evaporation, humidity, cloudiness, CO2 concentrations, land use, and land management. Physiological processes responsible for vegetation-climate feedbacks that operate at the scale of an individual leaf need to be scaled to a canopy of leaves and then to a landscape of thousands of plants. There are few observations to guide this scaling, as most studies of stomata! conductance and its response to CO2 are obtained from leaf measurements. In addition, most studies examine the short-te~ response of plants to CO2. Longer-term acclimation to high CO2 will change the short-term reduction in stomata! conductance.

OCR for page 88
BIOGEOCHEMICAL FEEDBACKS AND THE CARBON CYCLE IT sneer 1' Q $~ ~ season - ~ + L CO2 1 _/+ 1 _ 1 T rReduced ETI 93 System interactive CO2-climate feedbacks. (a) Temperature (T)-respiration feedback Terrestrial systems respond to climate warming by increasing growth due to longer growing seasons, removing CO2 from the atmosphere and storing in biomass and fresh organic matter. (b) CO=growth feedback/ Evapotranspiration (ET) is reduced, lowering latent heat fluxes and increasing climate warming through reduced cloud cover and increased sensible heat FIGURE 7.2 Climate-land biosphere feedbacks: System interactive views. The key to understanding the terrestrial biosphere's uptake of CO2 is to undertake observations and analysis at large spatial scales for extended times. These observations should integrate measurements of the carbon cycle with measurements of the energy and water cycles. The Panel supports the strategy of the U.S. Carbon Cycle Science Plan (Sarmiento and Wofsy, 1999) and the North American Carbon Program (Wofsy and Harriss, 2002) in this regard. For the purposes of this report the Panel supports the U.S. Carbon Cycle Science Plan's focus on the following two questions: 1. What has happened to the carbon dioxide that has already been emitted by human activities? 2. What will be the future atmospheric CO2 concentration trajectory resulting from both past and future emissions? These fundamental questions were articulated into six specific goals, two of which focus on the terrestrial carbon cycle. 1. Quantify and understand the Northern Hemisphere terrestrial carbon sink.

OCR for page 88
94 UNDERSTANDING CLIMA TE CHANGE FEEDBACKS 2. Determine the impacts of past and current disturbance, both natural (e.g., boreal forest and anthropogenic (e.g., land use) on the carbon budget. These goals are considered to be feasible steps over the next five or so years to address uncertainties in the carbon cycle and interactions with climate change. The focus on North America is intended as a first step to define global feedbacks involving CO2 and climate. Implementation of these goals has been laid out in plans for the North American Carbon Program (NACP) (Wofsy and Harriss, 2002~. The NACP includes radically new networks of long-term atmospheric observations and ecosystem studies. Data assimilation systems are described that for the first time would allow us to combine these data with high-resolution assimilated winds to define CO2 net exchange at landscape and continental scales. The plan also prescribes extensive manipulations and field measurements to elucidate the factors regulating CO2 uptake or release by major ecosystems. Thus, the NACP represents a systematic effort to address the carbon-climate feedbacks at the time and space scales relevant for understanding the mutual interactions of the carbon cycle and the climate system. This program, if implemented, would provide the basic information and analytical framework needed to quantify and understand climate-carbon feedbacks for North America, and it would provide the template for extension to other major land masses. In addition to the goals outlined above, the NACP will also be concerned with emissions of CO2, CH4, and CO. (Improving accounting of carbon emissions and uptake is also important for reasons other than the objectives of this report; they are vital for developing and maintaining effective greenhouse gas mitigation policies.) Previous carbon cycle research largely focused on studies of single components, such as the atmosphere or ocean, or through small-scale process studies. But carbon is exchanged continuously through the atmosphere, land biosphere, soils, and oceans. The temporal and spatial scales of the program must be appropriately large for addressing climatic issues, and data and models from all components must be brought together to develop information on global carbon balances. Results must be scaled up from process studies and inventories and rigorously compared to information gained at a regional or continental scale. These integration objectives are shared by and are embodied in the program's major elements for integration, including innovative new assimilation and data fusion systems that bring together diverse data and models, linking information at various scales to provide a consistent continental-scale carbon balance, resolved temporally by season. This coordination of science activities requires similar coordination among agencies involved in implementation.

OCR for page 88
BIOGEOCHEMICAL FEEDBACKS AND THE CARBON CYCLE 95 Major Program Elements of the Carbon Measurement and Analysis Strategy Long-term atmospheric measurements of the carbon budget are required from the ground, aircraft, and satellites, which should provide spatially and temporally resolved, three-dimensional atmospheric data for the major carbon gases, CO2 CH4, and CO, to enable reliable estimates for North American sources and sinks of these gases. These observations are required to obtain regional and continental sources and sinks for atmospheric CO2, CH4, and CO. The network planned by NACP extends present remote monitoring networks (Tans et al., 1996) of atmospheric observations to provide dense coverage and vertical soundings in the interior of the continent. Present networks of flux stations (Baldocchi et al., 2001) will be enhanced to provide traceable absolutely calibrated concentrations, and coverage will be extended to include many more representative regions. Intensive field programs that are planned by NACP, including large- scale airborne and field campaigns, should be launched to provide datasets to evaluate and to improve the design of atmospheric and surface measurement networks, to develop and test models, to interpret observations, and to provide atmospheric snapshots to constrain fluxes. These efforts should provide continuous feedback on uncertainties in modeling and assessment tools for carbon accounting. Inventories of carbon in major ecotones (e.g., the Forest Inventory Analysis ~Goodale et al., 20023) will need to be enhanced to encompass full carbon accounting and complemented by remote sensing and models to provide a complete carbon budget for the land. Lands (peatlands, scrub land, suburban landscapes) and carbon pools (roots, coarse woody debris, shrubs) not currently inventoried must be included. A hierarchical conceptual approach is planned in the NACP to support a multiscale interpretation, with intensive studies providing access to details and mechanisms that are extended using remote sensing, extensive inventories, and mechanistic models and join the atmospheric and ocean studies as components in a unified analysis framework. As outlined in several other disciplinary chapters of this report, the integration of models and model-data assimilation will be important. Such efforts could provide knowledge of the atmospheric concentrations of CO2 over the entire continent and adjacent waters at frequent intervals. We support the flow of information and the integration outlined by the NACP to obtain regional carbon accounting (see Figure 7.3~.

OCR for page 88
96 UNDERSTANDING CLIMATE CHANGE FEEDBACKS | Forecast winds 4-D atmospheric data: Remote sensing of CO2, CH4, CO (surface, land, oceans airbome, satellite) / ' Forecast / tower data Meteorological input data (sondes, radiances) / ~ And cover, land use, historical, inventory data, in- situ ocean data ~ . Data fusion Diagnostic models (inverse Assimilation) | Retrospective, real-time | North American sources and sinks for CO2, CH4, | FIGURE 7.3 Data flow and integration in the NACP. Complexity and level of synthesis increase down the figure. Valuable data products are delivered at each level. Note the central role played by the model-data fusion systems that combine observations Tom diverse sources, using data-driven models and advanced data assimilation arid optimization methods. A critical step will be to develop new classes of diagnostic models to determine sources and sinks of CO2 and other gases. Data-driven models of carbon dynamics in vegetation and soils will be combined in a data fusion framework with high-resolution meteorological information, surface flux data, and atmospheric concentrations to derive fluxes and a quantitative representation of the state of the atmosphere and of the carbon cycle. The Panel recommends that the NACP be implemented with major initiatives in the aforementioned key areas. We also support its plans for regular state-of-the-art assessments of carbon cycle science and carbon inventories for North America, with eventual extension of the observations and analysis framework to the entire globe. Linkage to the Global Carbon Project () of the International Geosphere-Biosphere Program me (IGBP), World Climate Research Programme fWCRP9, and International Human Dimensions Programme on Global Environmental Change (IHDP) would be useful in this regard.

OCR for page 88
BIOGEOCHEMICAL FEEDBACKS AND THE CARBON CYCLE MARINE BIOGEOCHEMICAL FEEDBACKS 97 Marine carbon feedbacks have been evaluated almost exclusively with models. This is unfortunate because the marine carbon cycle models being used for climate change studies are not capturing processes that may be key elements of feedback mechanisms. This is particularly true for the biological component of the models. The most advanced marine carbon cycle model that has been used in climate change simulations (Cox et al., 2000) does not include, for example, multiple phytoplankton species, iron limitation, nitrogen fixation, variable carbon-to-nitrogen ratios, and dissolved organic matter, all of which appear to be important features of the marine carbon cycle. Most other models used for such purposes are even simpler. The reasons for these omissions are various, but they include the lack of data for developing defendable parameterizations as well as the additional computational expense of increasing the complexity of the models. Models nevertheless can help to put rough boundaries on the strength of various feedback mechanisms. In terms of the overall feedback of the marine carbon cycle on climate on the time scale of a century, the models vary from showing almost no impact on ocean carbon uptake (Jogs et al., 1999; Maier- Reimer et al., 1996) to a reduction of about 10-15 percent (Friedlingstein et al., 2001; Matear and Hirst, 1999; Sarmiento et al., 1999;~. This overall effect represents the sum of individual feedbacks that may be considerably larger. Some of the potentially important marine biogeochemical feedbacks are described briefly below. Physical and Chemical Feedbacks on Atmospheric CO2 Solubility-Temperature Feedback The solubility of CO2 and the degree to which it reacts to form other inorganic and nonvolatile forms of carbon decreases with increasing temperature with an accurately known functionality that is described by temperature-dependent equilibrium constants. Thus, there is a positive feedback on atmospheric CO2 associated with temperature changes and the inorganic chemistry of CO2 in seawater. The few modeling studies of this feedback regard it to be of modest strength, amounting to a 10-15 percent reduction of the cumulative anthropogenic CO2 uptake by the ocean on the century time scale (Jogs et al., 1999; Matear and Hirst, 1999; Sarmiento et al., 1999).

OCR for page 88
98 CO2 Uptake-Ventilation Feedback UNDERSTANDING CLIMA TE CHANGE FEEDBACKS In order for substantial carbon to be taken up by the oceans, it must be first moved across the air-sea interface and then from the surface ocean to deeper in the ocean. The resistance of the air-sea interface is relatively small, and so it is largely the vertical circulation in the ocean, including the ventilation of the thermocline and the formation of intermediate and deepwaters, that regulates the uptake of anthropogenic CO2 by the ocean. As noted in Chapter 5 some models predict that the rate of overturning by the thermohaline circulation will decrease in a warmed world, which would result in a positive feedback on atmospheric CO2. The few studies on this feedback are in disagreement with regard to its strength, varying between essentially no impact on ocean carbon uptake (Maier-Reimer et al., 1996) to as much as 17 percent (Sarmiento et al., 1999~. Differences are primarily due to the sensitivity of the ocean circulation to CO2 changes. This underscores the point that marine carbon cycle models are only as good as the circulation models in which they are embedded. Stratification Mixing Feedback As discussed above, many models predict changes in ocean circulation, which can alter the CO2 balance of surface waters and therefore atmospheric CO2. For example, stratification of high-latitude waters would inhibit the upward flux of deepwaters, which are enriched in CO2 due to the decomposition of organic matter sinking from the upper ocean, resulting in a negative feedback on atmospheric CO2. This would be counteracted to some degree by a reduction in carbon export from surface waters due to the reduced upward flux of nutrients. For example, Bopp et al. (2001) found this effect to dominate the 6 percent decrease in carbon export from surface waters for a CO2 doubling in their models. The few studies on the overall feedback disagree with regard to its strength, though they generally agree that this feedback tends to have a similar magnitude (but opposite in sign) to the CO2 uptake-ventilation feedback described above (Jogs et al., 1999; Matear and Hirst, 1999; Sarmiento et al., 1999~. ENSO~O2-Upwelling Feedback A similar feedback may operate in low latitudes, as indicated by some models that predict increased frequency of El Nino events with increased

OCR for page 88
BIOGEOCHEMICAL FEEDBACKS AND THE CARBON CYCLE 99 CO2 (Timmerman et al., 1999~. Such an increase would reduce the natural marine source of CO2 to the atmosphere (due to upwelling), creating a negative feedback on atmospheric CO2. Other models show a reduction in equatorial upwelling, which would have a similar effect (Bopp et al., 2001~. This feedback has not been quantified, though at least one modeling study suggests that the equatorial Pacific does not exert a strong control on atmospheric CO2 on the century time scale (Sarmiento and Orr, 1991~. However, the relative roles of high and low latitudes in regulating atmospheric CO2 are active areas of research (Broecker et al., 1999~. Overview of Biological Feedbacks on Atmospheric CO2 Carbon Export-Temperature Feedback Phytoplankton growth rates generally increase with temperature (Eppley, 1972) and so the potential for a negative feedback exists. The fraction of photosynthetically derived material that is exported to deeper waters, however, is suggested by a recent synthesis of many field studies to decrease with increasing temperature (Laws et al., 2000), which would constitute a positive feedback. Bopp et al. (2001) found very little sensitivity of carbon export to climate warming using a simple ecosystem model. That model, however, did not include the findings of Laws et al. (2000), and so this feedback remains poorly quantified; even its sign is not known. Carbon Export-Light Feedback The exposure of phytoplankton to light depends on the surface irradiance, the opacity of the water column and the depth of the mixed layer (deeper mixed layers result in more time that phytoplankton spend in the dark). Changes in cloudiness could therefore change photosynthesis. The opacity of the water column is largely due to changes in phytoplankton abundance but also to colored dissolved organic matter, the dynamics of which are poorly understood. Finally, many climate models (e.g., Bopp et al., 2001) predict shallower mixed layers due to decreases in surface density, which could enhance light levels and therefore photosynthesis. Bopp et al. (2001), the only study to quantify this feedback, found increases in carbon export of as much as 20 percent over large regions of the high latitudes due to decreases in mixed layer depth induced by a CO2 doubling.

OCR for page 88
100 Carbon Export-Iron Input Feedback UNDERSTANDING CLIMATE CHANGE FEEDBACKS Phytoplankton growth in many parts of the ocean is limited by the availability of iron, a substantial fraction of which is derived from wind- blown continental dust (Martin et al., 1991~. Thus, climate-induced changes in continental aridity, wind speed, and wind direction may influence phytoplankton production. Ice core data, which show higher levels of atmospheric dust during glacial times, suggests that iron may be part of a positive feedback loop (Martin, 1990~. This feedback has not been quantified using models because the incorporation of iron into marine ecosystem models is just beginning (Moore et al., 2002~. Many questions remain about how and in what form iron is delivered to the ocean, how it is made available to phytoplankton and how it is cycled in the marine ecosystem. CO2kalcification Feedback The calcification rates of coccolithophores and coral reefs have recently been shown to decrease with increasing atmospheric CO2 (Kleypas et al., 1999; Riebesell et al., 2000~. Because calcification is a source of CO2, such organisms are potentially part of a negative feedback on anthropogenic CO2. Using a simple model, Zondervan et al. (2001) suggest that this feedback is rather small for the coccolithophores, which dominate global calcification. Feedbacks Involving DimethylsulB~de Dimethylsulfide (DMS) is thought to be a major precursor of cloud condensation nuclei in unpolluted air (see Chapter 7~; therefore the release of DMS from the ocean may influence cloud albedo and climate. Phytoplankton, bacteria, and zooplankton all play important roles in marine DMS cycling, so any change to the marine ecosystem as a result of climate change is likely to affect the DMS concentration in seawater and hence its flux to the atmosphere. The turnover of DMS in the ocean mixed layer is so rapid that the flux to the atmosphere is only a small residual of much larger fluxes. Thus, modest changes in internal cycling have the potential of producing large changes in the air-sea flux. However, because the response of marine ecosystems to climate change is uncertain, the response of marine DMS emissions is also uncertain. Additional uncertainty is caused by production of the DMS precursor, dimethylsulfoniopropionate (DMSP),

OCR for page 88
BIOGEOCHEMICAL FEEDBACKS AND THE CARBON CYCLE 101 which varies greatly among phytoplankton species (Keller et al., 1989~. Zooplankton play a role in DMS cycling through grazing, which is an important mechanism for releasing DMSP from phytoplankton cells (Dacey and Wakeham, 1986~. The bacterial impact on DMS cycling is through the effect on DMS yield during DMSP consumption, as well as through the direct consumption of DMS (Kiene and Bates, 1990~. A significant sink of DMS also occurs through abiotic photochemical consumption (Kieber et al., 1996~. These points underscore the complexity of DMS cycling in seawater and the difficulty in predicting its response to climate change. It is not surprising that there is no simple relationship between DMS concentration and temperature, salinity or chlorophyll, as revealed by a recent synthesis of over 15,000 measurements by Kettle et al. (1999~. However, DMS flux tends to increase with increasing solar radiation (Bates et al., 1987), with seasonal maxima in flux and concentration occurring in the summer (Kettle et al., 1999; Kettle and Andreae, 2000~. Simo and Pedros-Alio (1999) suggest that this relationship is due to photo-inhibitory effects on bacteria (which consume DMS and reduce the DMS yield from DMSP) during conditions of high light and shallow mixed layer depth. Thus there is some support for the hypothesis of a negative feedback on the climate system involving DMS and sunlight (Charlson et al., 1987; Shaw, 1983~. However the magnitude of the feedback is not known nor is the underlying mechanism well elucidated. Ice core data provide additional insights regarding DMS-climate feedbacks. Ice core records of methanesulfonate (MSA), an atmospheric oxidation product of DMS, show that its atmospheric concentration during glacial times was substantially different compared to the present. Glacial concentrations were higher in the Southern Hemisphere (Legend et al., 1991) and lower in the Northern Hemisphere (Saltzman et al., 1997), suggesting that the sign of the feedback may vary with location. There have been a few modeling studies that have attempted to quantify DMS-climate feedbacks. The empirical model of Lawrence (1993) suggested that a CO2-induced warming could be reduced by 10 percent to 50 percent due a DMS-climate (negative) feedback. Gabric et al. (1998) applied temperature and wind speed changes from a doubled-CO2 climate model to an ecosystem model with DMS dynamics in the Southern Ocean. They found a modest (2-8 percent) increase in the flux of DMS to the atmosphere due to an increase in the gas transfer velocity and phytoplankton growth rate, both of which increase with temperature (wind speeds actually decreased slightly in the simulation). This study also supports the potential of a negative feedback, albeit a weak one. In light of the complexity of DMS

OCR for page 88
102 UNDERSTANDING CLIMATE CHANGE FEEDBACKS cycling, these models are extreme simplifications, but they nevertheless provide a framework for attempting to quantify DMS feedbacks. Feedbacks Involving Methane and Nitrous Oxide The emission of methane and nitrous oxide from the ocean currently constitutes a very small fraction of the total greenhouse gas forcing of the atmosphere, however, there is the potential of large releases of these gases. Abundant reservoirs of methane are stored in ocean sediments in the form of clathrates, which are nonvolatile. Warming could release the methane into the water column and atmosphere, providing a positive feedback. This feedback has not been quantified, though the paleoclimate record suggests that the feedback may have been activated many times in the past (Bains et al., 1999~. The volume of methane available for release is poorly known (Gornitz and Fung, 1994~. One modeling study suggests an upper limit of 10 percent to 25 percent increase in warming over the next century due to this feedback (Harvey and Huang, 1995~. Nitrous oxide is formed in the ocean during respiration, and the rate of release appears to be a function of the dissolved oxygen concentration (Law and Owens, 1990), particularly at low oxygen levels. Because both respiration and oxygen abundance are sensitive to climate change, there is the potential for climate feedbacks involving marine N2O. This is particularly true given the fact that the amount of N2O release is only a small fraction of the total cycling of nitrogen. A Scientific Strategy for Marine Biogeochemical Feedbacks Marine biogeochemical feedbacks are to a large extent unquantified. First order questions related to even the sign of certain feedbacks exist in some cases. The rate at which the ocean takes up carbon will very likely continue to increase because of the increasing atmospheric CO2 level. Changes in ocean carbon dynamics driven by changes in circulation and biology will modulate this increase. The degree of this modulation is very poorly known due to large uncertainties in the projections of future changes in ocean circulation and of the response of ocean biota to these ocean circulation changes. While primary production is important for evaluating the overall intensity of carbon cycling in surface waters, it is the exported fraction (from surface waters) of primary production that is important to surface ocean and

OCR for page 88
BIOGEOCHEMICAL FEEDBACKS AND TlIE CARBON CYCLE 103 atmospheric CO2 levels. Our ability to quantify export and its variability on large scales is improving, but it is still poor. The rate of decomposition of organic matter exported from surface waters is also very important but even more poorly known. The feedback between marine DMS emissions and cloud albedo Is potentially very large. Over the past 15 years substantial progress has been made in evaluating the mechanisms of ocean DMS cycling, including its production, consumption, release to the atmosphere, oxidation in the atmosphere, and contribution to the cloud condensation nuclei (CCN) pool. However, the nature of the overall feedback has remained elusive. The ocean is currently a minor source of methane and nitrous oxide to the atmosphere. However, there is a poorly understood potential for a large release of these gases to the atmosphere. Marine sedimentary clathrates are a very large reservoir of methane that could be abruptly released. Large amounts of nitrogen are cycled in the marine environment and the fraction released as N2O is currently small, but the controls on this fraction are poorly understood. . Observations for Improving Understanding and Models The main areas that deserve attention in the context of marine biogeochemistry and climate are the rate of CO2 uptake by the ocean and the release of DMS from the ocean. If ocean circulation and biology do not change in the future, these rates can be projected with relatively high accuracy. Model uncertainties exist because we do not know to what extent changes in ocean physics and biology will modulate the cycling of carbon and sulfur in the sea. Thus concerted studies need to be undertaken to assess the response of the marine carbon and sulfur cycles to changes in ocean circulation and other climate variables, such as solar radiation and temperature. This will be best achieved by monitoring the ocean carbon and sulfur cycles over time scales ranging from months to decades. The annual cycle in ocean physical properties and other climate variables represents the major temporal forcing on marine biogeochemical systems and should be monitored intensively. Interannual and decadal climate variations represent another major forcing that needs to be understood in terms of feedbacks on the marine carbon and sulfur cycles. Four observing system components are selected for special attention; 1. Continued satellite-based monitoring of ocean color is needed to derive information about changes in plankton biomass and CO2fxation.

OCR for page 88
104 UNDERSTANDING CLIME TE CHANGE FEEDBACKS 2. Expanded monitoring of the atmospheric oxygen-nitrogen ratio and atmospheric DMS concentration is needed to derive information about seasonal and interannual variations in the CO2 fixation and export to the ocean interior and sea-to-air DMSflux, respectively, on basin-wide scales. 3. High-resolution (monthly) time-series measurements are needled of the carbonate system (e.g. CO2 concentration and dissolved inorganic carbons, nutrients, oxygen, chlorophyll, dissolved" organic carbon, primary production, verticalflwres of carbon, and the main sulfur pools (particulate and dissolved DMS and DMSP) at a wide variety of ocean locations. Currently, open ocean time-series measurements are limited to the carbon cycle at a few sites, mainly in the subtropical oceans. 4. Periodic surveys of ocean chemical and physical properties are needed to evaluate the uptake and processing of carbon in the marine environment. There have been a few such surveys in the past, including the Geochemical Ocean Sections Study fGEOSECSJ of the 1970s and the WOCE Joint Global Ocean Flux Study (JGOFS) CO2 survey of the 1990s, and it is critical that they occur everyfive to ten years. These measurements should be made through a combination of autonomous buoys to derive temporally continuous time series; ship-based measurements to produce spatially extensive repeat surveys; and remote sensing. In addition to these observational strategies, increased efforts are needed to develop new technologies for measuring carbon and sulfur fluxes in the sea, particularly the air-sea flux of CO2, the sinking flux of organic carbon, the rate at which organic matter decomposes (respiration), and production and consumption of DMS and DMSP. The primary obstacle to making projections about the marine carbon and sulfur cycles is the lack of observations to inform the models. The aforementioned observations will be critical in helping to provide adequate descriptions of the relevant processes, which can lead to refined and observationally tested model representations. The transition between observation and the development and testing of corresponding model representations of the key processes should be a seamless one; we advocate facilitating this by incorporating numerical modeling into field studies during the development and execution as well as in the data synthesis phase.

OCR for page 88
BIOGEOCHEMICAL FEEDBACKS AND THE CARBON CYCLE Evaluating Progress 105 Factors critical to the uptake of carbon by the ocean that might be derivable from observations and that can be used to test models include the following. 1. The change in surface pCO2 for a given change in temperature or the change in inorganic carbon contentfor a given change in heat content. This is necessary to evaluate the solubility-temperature feedback. 2. Change in inorganic carbon inventory for a given change in ocean ventilation rate. This will allow the circulation-uptake feedback to be assessed. The ventilation rate can be estimated from various tracers of ocean circulation, such as chlorofluorocarbons. 3. Change in export production and surface nutrient concentration for a given change in stratification. This will allow the stratification-CO2-mixing and stratification-production feedbacks to be assessed. Export production can be crudely estimated on large scales from satellites and variations in atmospheric oxygen. 4. Change in export production for given changes in temperature, light, and iron dust inputs. This will allow the feedbacks between carbon export and various controls on it to be assessed. Iron dust inputs on large scales can be crudely estimated from precipitation and aerosol fields derived from satellites. 5. Change in cloudfi-action and albedo for given changes in surface ocean DMS. This will allow feedbacks involving DMS and climate to be assessed in a crude sense. Monitoring at a more detailed level (e.g., MSA, CCN densities, wind speed, SST, DMS community production) would be valuable as well. 6. Changes in concentrations of the isotopes of methane in the atmosphere and select areas of the ocean for given changes in ocean temperature. This would allow for feedbacks between warming and release of methane from clathrates to be assessed. 7. Changes in the concentrations of the isotopes of nitrous oxide in the atmosphere and select areas of the ocean for given changes in a variety of ocean physical and biological properties, including stratification, temperature, and primary production. This would allow feedbacks related to N2O release from the ocean to be evaluated.

OCR for page 88
106 UNDERSTANDING CLIME TE CHANGE FEEDBA CKS Programmatic Efforts The U.S. Global Change Research Program has developed an interagency Carbon Cycle Science Programi with a Science Plan whose goal is specifically to reduce uncertainties in understanding the carbon cycle. In addition, as part of a new international initiative the fledgling U.S. Surface Ocean Lower Atmosphere Study (SoLAS)3 has a mission to "achieve a quantitative understanding of the key biogeochemical-physical interactions between ocean and atmosphere, and of how this coupled system affects and is affected by climate and environmental change." To a large extent a successful approach toward improving understanding and modeling of biogeochemical feedbacks is directly linked to the success of these programs. We recommend that agencies work to ensure that the goals of the U.S. Carbon Cycle Science Program and SOLAS are met through adequate and sustained funding These agencies should continue to ensure that U.S. Carbon Cycle Science Program and SOLAS activities fit within the framework of international activities. ' http://www.carboncyclescience.gov 2 http://www.carboncyclescience.gov/PDF/sciplan/ccsp.pdf 3 http://w~7vw.aoml.noaa.gov/ocd/solas/