GAIM Plan 1998-2002

The GAIM Task Force was formed as an overarching framework activity of the IGBP for coordinating and stimulating different multi-disciplinary research components that can be synthesized to formulate an integrated view of the Earth System. The Goal of GAIM is to advance the study of the coupled dynamics of the Earth system using as tools both data and models. The challenge to GAIM is to initiate activities that will lead to the rapid development and application of a suite of Global Prognostic Biogeochemical Models. These global biogeochemical models would subsequently be linked, partly through hydrological coupling, to General Circulation Models (GCMs). These, together with socioeconomic observations and models which would address future anthropogenic perturbations, would lead to robust prognostic models of the Earth system. In order to make progress toward its goal, the Task Force must:

analyze current models and data
interpret the capability of current models and experimental programs against the demands for knowledge
and advance and synthesize our understanding of the global biogeochemical cycles and their links to the hydrological cycle and more generally to the physical-climate system as a whole

Much of the progress to date in modelling specific components within the global biogeochemical subsystem sets the context for modelling activities within the various IGBP Core Projects. GAIM recognizes, supports, and will benefit from these efforts. GAIM activity is by definition cross-cutting; therefore, the activities of GAIM intersect fundamentally with all the Core Projects as well as with the World Climate Research Program (WCRP) and International Human Dimensions Programme (IHDP).

In its initial stages, GAIM concentrated on certain key issues concerning the Carbon Cycle and aspects of the coupling between terrestrial ecosystem and climate. This start-up phase aimed to develop and test programmatic techniques and model development for tractable cross-cutting problems which could then be used as smaller-scale examples of the larger Earth System issues which GAIM must ultimately face. The theme of the carbon cycle was selected because of the relative maturity of research in the relevant individual disciplines. With the success of several model development projects and model intercomparison activities, GAIM is now poised, as the basis of this GAIM PLAN, to extend its analysis to broader issues which will be encountered in global biogeochemical models. GAIM will move now to its role as integrator of IGBP science. It will focus on Tasks which cut across the realms of the Core Projects. In addition, GAIM will examine theoretical and practical modelling techniques which will enable the effective coupling of subsystem models and the development and evaluation of Earth System Models. A few examples of these foci were articulated at a meeting of the GAIM Task Force in Barcelona in April, 1997 and are described below. The scientific plan for the next phase of GAIM represents the next step toward its ultimate goal, yet it is recognized that the necessary theoretical, technological and data resources are not yet available. The broadly cross-cutting issues and projects described in this GAIM PLAN are aimed at providing some of these resources, and are part of a continuous re-evaluation of GAIM science, implementation and organizational structure. Each of the activities in the GAIM PLAN are directed toward integration of IGBP scientific results.

Integration of IGBP Science

As Earth subsystem models develop to a more robust level, GAIM is preparing to enter an integrative phase. This will involve the establishment of techniques for coupling and integration of biogeochemical subsystem models in preparation for the construction of integrated prognostic biogeochemical models. Such integration will involve coordination with each of the IGBP Core Projects. The integration program will have three aspects, each contributing to the overall objective of developing the modelling capacity which will ensure the achievement of GAIM's goals. The first will be at the subsystem level, where GAIM activities will be designed to bring developing subsystem models into boundary compatibility. This will be done through modelling workshops involving intercomparisons of like subsystem models, and intercomparisons involving coupling between adjacent subsystem models (which must match boundary conditions and fluxes). The second segment will be at the system level, where simple Earth system models are compared to highlight differences in coupling techniques, inter-element fluxes, and sensitivity studies to reveal the differences between models of the relative importance of individual system parameters. The third segment will be at the IGBP Core Project level, where GAIM will work with Core Project modelling teams to help facilitate inter-subsystem coordination. In addition, GAIM will act as a connection between IGBP and its sister organizations, WCRP and IHDP. GAIM will organize workshops with inter-programme modelling teams to facilitate such connections.

Subsystem models

Subsystem integration involves four key linkages: land-atmosphere, ocean-atmosphere, atmospheric physics with atmospheric chemistry, and land-ocean. These subsystems have different space and time scales (which themselves depend upon what is being tracked) and are often quite stiff as linked systems and therefore difficult in perturbation experiments. There are also greatly differing degrees of parameterization with little understanding as to effects. Important linkage experiments, however, can NOW be done: a) the ocean carbon model inter-comparison project (OCMIP) has linked atmosphere GCM's (mainly as drivers) with ocean GCM's containing carbon chemistry and crude biology; b) the terrestrial carbon models (NPP Efforts) are being driven partially by GCM results; and c) GCM Atmosphere transport codes are being explored for tackling the chemistry connection.

Subsystem models are being developed at present with a variety of structures and emphases. While each model is taken to represent the processes within a biogeochemical subsystem, the analytical and numerical formulations are widely disparate, and often lead to significant differences in model results. A fundamental issue is the development of subsystem models in such a way that the boundary conditions and fluxes for each will be compatible with each of the others. This compatibility is defined in terms of the ability of each model to provide the necessary input to define the boundary conditions needed to most efficiently run the others. For example, the boundary between the terrestrial and marine biogeochemical systems involves physical and chemical conditions and fluxes which are so complex that no single subsystem model presently accounts for all. Thus, matching boundary fluxes at the boundary would be impossible unless carefully coordinated during the model development phase.

As each of the subsystems becomes better understood and models converge on realistic values of output parameters, it will be timely to couple compatible models to form a more complete Earth system model. While it is not necessary to assume (and not possible to mandate) that all models of a particular subsystem will have identical input/output parameterizations, it is essential the each model be coupled only to other subsystem models with compatible parameterizations. Thus we envision the emergence of a suite of coupled models, each with consistent coupling and interactions between model components, but each based on a different style of process formulation. The parallel development of coupled Earth system models has several advantages. The most important is that because no single model (even an integrated Earth system model based on compatibly coupled subsystem models) accounts for all processes and interactions in the Earth system, each model will necessarily result in slight differences in inter-component fluxes and sensitivities. This will set the stage for Earth system model intercomparison which will highlight the relative importance of the various processes, interactions, and feedbacks between subsystems modelled by each of the integrated models. Such intercomparisons should ultimately lead to modified integrated models which correctly account for the interactions to which the Earth system is most sensitive, while becoming unburdened from those to which it is demonstrably insensitive.

Models of the various aspects of the Earth System are each associated with some uncertainties (errors). The uncertainties arise from 1) incomplete theoretical or mechanistic understanding of Earth System processes, 2) inaccurate formulation of processes within the modelled subsystem (e.g. ocean circulation and trace gas transport), and 3) inaccurate or simplistic boundary conditions. The first two issues are being addressed through various model development and intercomparison projects (e.g. NPP, OCMIP). The latter source of uncertainty is a more complex problem and bears on subsystem model coupling. Because gas, water, and energy exchange between subsystems (e.g. ocean, atmosphere) determines their respective modelled boundary conditions, model coupling can lead to more accurate modelled fluxes across the boundary. However, coupling introduces its own uncertainties and sources of error. It is thus necessary to consider the effect of model coupling to TOTAL uncertainty. The obvious goal is to develop Earth System models in which this total uncertainty is reduced to levels less than the individual uncertainties of the subsystem models with prescribed boundary conditions.

System level

It is the ultimate mission of the GAIM Task Force to promote the development of integrated models of the Earth's biogeochemical system for eventual linking to the physical climate system studied through WCRP as well as the societal systems studied through IHDP. Simple models of the Earth system already exist, but they are not sufficiently robust to incorporate the detailed subsystem models being developed throughout the IGBP. It will nevertheless be instructive to examine such simple holistic models because some features which emerge may help identify and thus forestall potential problems in developing more comprehensive models on the basis of subsystem model coupling. Important insights can be gained from existing simple models of the Earth system, so we will build on these simple models in two ways:

1. An organized simple but total Earth System Model approach that raises difficult system dynamic issues (chaos, feedbacks, parameterization sensitivities, policy effects, etc.), and

2. An effort to collect and document existing models of key features of the Earth System (e.g. Carbon Cycle) that could run on a PC (or run over the WWW).

The purpose of the former is to highlight key scientific issues that may be lost in the large model efforts while the purpose of the latter is out-reach and education.

In order to assess the validity of Earth system models, it is critical to understand the sensitivity of the system to each of the input data. Heuristic and mathematical models are becoming developed to a point now where it is appropriate to consider model sensitivity. It will be necessary to conduct model sensitivity analyses of dynamic vegetation models, ocean carbon cycle models, GCMs, and hydrologic models as well as for simple Earth system models with respect to the various input climate and ecological data.

In the next several years, GAIM will begin to address some of the more theoretical issues involved in complex model development, coupling and evaluation. These issues are not necessarily limited to Earth System applications, nor are they being addressed primarily by Earth scientists. GAIM will explore the existing theory and focus its efforts on development of applications for Earth System Modelling. Examples of issues to be addressed in this aspect of GAIM include schemes for quantifying and comparing coupled and uncoupled model result uncertainties, determining the minimum necessary resolution for model validation data, inverse methods for applying model validation data, quantifying system responses to subsystem-level and system-level perturbations, and sensitivity studies of coupled systems.

IGBP Core Project Integration

Each of the IGBP Core Projects is developing models of the appropriate biogeochemical subsystems. Once these are completed, it will be GAIM's task to promote the coupling of the various subsystem models and the development of integrated Earth system models. Model coupling will require advance planning so that it will be possible to most effectively match boundary conditions and fluxes. Thus, input and output data sets will need to be assessed and standardized, model temporal and spatial resolutions will have to be matched or scaled where necessary, and common numerical protocols will need to be defined so that the necessary parameters will flow through one subsystem model to the next. GAIM will work closely with the Core Projects to ensure that subsystem model linkages can be made smoothly.

The development of Earth System Models is a complex problem, to which the extensive resources of various institutions will be applied. The GAIM Task Force will not compete with these efforts, but rather will encourage and complement the efforts, as the Task Force is composed of key scientists from these leading institutions world-wide. The composition of the Task Force is determined by the scientific issue being addressed, and will continue to evolve in response to the development of new Earth System Models. As such, GAIM will provide a means for planning and coordination between these various institutional efforts.

The IGBP Core Projects are organized in such a fashion to encourage interactions and collaborations between scientists specializing in each of the Earth's biogeochemical subsystems. As such, the framework is already in place for organization of collaborative and intercomparison activities which should lead most effectively toward meaningfully coupled models. GAIM will need to work closely with each of the IGBP Core Project modelling teams to help steer the modelling efforts in a direction which will result in the most efficient coupling possible.

Scientific Questions

The new GAIM PLAN is based on a set of fundamental questions. (These questions encompass the scope of the original "6 key research questions" indicated in IGBP Report 28, 1994.) These questions are for the most part too general and broad to answer directly. They are based on a set of basic observations which must be explained if we are to understand key aspects of the Earth System. The process of seeking answers to the Fundamental Questions should fill the gaps in our understanding of the connections in the "Bretherton Diagram" as well as provide the framework for constructing reliable global prognostic biogeochemical models, GAIM's ultimate goal. As such, the Fundamental Questions address the outstanding gaps in our understanding of the linkages between the various part of the Earth System as represented by the Core Projects. The fundamental questions are each composed of a set of sub-questions which are at a more tractable scope given existing scientific research programs. While even the sub-questions are too broad to be answered by the results of individual research projects, the coordinated collaboration of research groups throughout IGBP (and beyond) can effectively address them. The narrower scope of the sub-questions will provide results in the nearer-term, making related research more readily fundable in the still somewhat disciplinary funding climate of national agencies worldwide. While inroads are being made to ensure adequate funding of ambitious interdisciplinary research such as that with which GAIM has been charged, there are a number of important issues which can be addressed now in the more "traditional" framework.

The four Fundamental Questions and their sub-questions will be addressed by GAIM research directly in some cases, and by Core Projects in others. Individual research projects will generally encompass only part of each question. The results of investigations of all sub-questions together will be used to address each of the Fundamental Questions. For the next several years, this will be accomplished through workshops on selected topics which will promote the synthesis and integration of IGBP scientific results in order to address the Fundamental Questions and ultimately lead to the development of a suite of global prognostic biogeochemical models. For some subquestions, a specific research plan has been formulated and preliminary work is already underway. In contrast, some other questions should be regarded as a "call to arms" for GAIM and the rest of IGBP in conjunction with the international global change research community.

The sub-questions within the Fundamental Questions follow a general pattern including a foundation based on "paleo" (1A, 2A...), then issues related to understanding ongoing system-level processes (1B, 2B...), and finally questions aimed toward developing prognostic capabilities (1C, 2C...). Each of these sub-questions will require the integration of a number of distinct research projects. Some of these projects will be conducted by GAIM, while others will be done by the Core Projects and other programs (e.g. WCRP, IHDP). Each project will be designed to address a specific sub-question, and will require development of theory, models, and datasets throughout IGBP Core Projects and framework activities. GAIM will take the lead in coordinating project results by convening the appropriate intercomparisons, syntheses, etc. This will entail a much greater level of interaction between GAIM and the Core Projects at all levels of activity than there has been in the past. In part, the ability to address this level of issues has been made possible by scientific advances within the Core Projects over the last several years, and the successful integration of IGBP science will depend on greater interaction with and between the Core Projects.

The partitioning of carbon between the ocean, atmosphere and terrestrial ecosystems is thus related to sea level variations. It will be necessary to investigate the links between ocean-atmosphere CO2 flux and the effects of changing atmospheric CO2 concentrations on terrestrial ecosystems. While some inroads regarding atmosphere-ecosystem interactions have already been made by efforts led by GCTE as well as GAIM and WCRP, and ocean-atmosphere fluxes by JGOFS, PAGES, GAIM and WCRP, an integrated understanding of the link and quantification of the feedbacks is yet to be obtained. This will involve coordination between the various IGBP Core Projects as well as WCRP.

Data needs: Pollen records, ice cores, marine sediments, corals.

1B. What processes control horizontal transport of biogeochemcially active species (CO2, CH4, P, S, N, etc.) above, at and below the Earth's surface?

(What are the fluxes and stoichiometry of riverine biogeochemical species today and how has this changed due to human activity? - Continental Aquatic Systems)

(What is the effect of atmospheric transport of trace gases on global atmospheric composition? - Transcom)

(What is the role of ocean circulation in redistributing CO2 and other trace gases, and how does this affect ocean-atmosphere gas exchange? - OCMIP)

While vertical exchange and transport has been addressed in many projects throughout IGBP, horizontal transport studies have been more limited. In part, this is because horizontal transport involves movement from the realm of one Core Project to that of another. In addition, in introduces another dimension to the models and includes heterogeneities unlike those found in the vertical. However, some significant progress has already been made in developing the horizontal component of transport codes. While there are many examples of necessary directions for improved modelling and observations, three discrete projects are briefly described here. These may serve as example for additional horizontal transport studies.

Horizontal Biogeochemical Transport in Continental Aquatic Systems

An ultimate goal of the IGBP is to understand the Earth system at a level which makes it possible to construct prognostic biogeochemical models for coupling with physical climate and socio-economic models. While there have been considerable efforts initiated by IGBP and other organizations to investigate the vertical exchange and fluxes of biogeochemical species, relatively little attention has been paid to horizontal fluxes across terrestrial systems. These horizontal fluxes may exert a profound influence on the balance of nutrients and thus biota between upland, coastal, and marine ecosystems. The magnitude of the effect of these fluxes relative to vertical fluxes will provide an important constraint on future distributions of nutrients and other biogeochemically active species, enabling IGBP to develop more accurate prognostic models of the Earth system.

Continental Aquatic Systems are defined as all surface and subsurface water involved in the hydrologic cycle on the continents. This includes lakes, rivers, wetlands, soil moisture, and ground water from the point where precipitation reaches the Earth's surface until it reaches the sea in full marine conditions, or until it reaches some other final base level. While the "upstream" boundary of this realm is easily visualized, the "downstream" boundary is a broad zone variable in space and time wherein river water interacts with the ocean until reaching fully marine characteristics with respect to chemistry and ecology. Emphasis is placed on water, sediment, carbon, nitrogen, phosphorus, silicon and micro nutrients.

Continental Aquatic Systems play a critical role in transporting, storing and cycling nutrients in the Earth system. Their importance is highlighted by our emerging understanding of present day climate systems, our growing knowledge of the spatial and temporal patterns of natural variability, the predicted consequences of likely future climate change and above all the implications of population growth and current socio-economic projections for the coming decades. Water flowing on or below the surface of the world’s land masses is one of the main links in many biogeochemical cycles of crucial importance to the functioning of the biosphere as well as the water resource base for human populations.

River systems are linked to regional and continental-scale hydrology through interactions between soil water, evapotranspiration, and runoff in terrestrial ecosystems. As such, river systems, and more generally the water cycle itself, serve as a control on the translocation of constituents over vast distances from the continental landmass to the world's oceans and to the atmosphere. The system serves, in part, to transfer nutrients to marine biological systems and hence potentially affects oceanic productivity. With particular reference to both nutrients and sediment, (DOC), landscape disturbance greatly increases the rate of loss from the terrestrial biosphere and the consequences can be global in scope. This redistribution is important to both donor (landscape) and recipient (aquatic) ecosystems, and we need to develop tools which can quantify these phenomena and hence be able to contribute to the problem of determining the effects of changing climate, land-use and other aspects global environmental change on the redistribution of water and essential nutrients. The issue has in part been addressed in IGBP Report 39 "Modelling the Transport and the Transformation of Terrestrial Materials to Freshwater and Coastal Ecosystems" [Vorosmarty, 1997 #5410].

Continental Aquatic Systems have not been fully incorporated into our growing understanding of the Earth's biogeochemical systems because we lack the answers to some basic questions regarding the role of continental water. These basic questions are as follows:

•What are the present-day stocks, concentrations, and flux fields of fresh water-borne nutrients (from atmosphere through terrestrial ecosystems and societal systems) to the ocean?

•What is the partitioning of water and nutrient stocks and fluxes due to the influence of natural variability versus human perturbations? To what extent have humans affected global nutrient fluxes?

•How do biogeochemical processes affect the flux of nutrients through Continental Aquatic Systems?

•Which aspects of fresh water-related biogeochemical processes and fluxes are most sensitive to projected future changes in nutrient supply and water stocks/fluxes?

•To what extent do changes in global continental aquatic nutrient fluxes affect ecosystem function?

•To what extent do changes in global continental aquatic nutrient fluxes affect water resource utility and sustainability?

•To what extent do changes in global continental aquatic nutrient fluxes affect the Earth system?

The main anthropogenic inputs to continental aquatic systems are from agriculture, industrial activity, and sewage treatment plants. There are three main issues which bear on the anthropogenic effects on biogeochemical river fluxes:

1. The effect of feedbacks with changes in drainage basins on biological cycles and human society;

2. The controls of biological, physical and chemical processes (natural and anthropogenic) on fluxes of sediment, water, micronutrients, nitrogen, carbon, and phosphorous in the catchment cascade; and

3. The chemical attributes and quantities of river-borne fluxes to the ocean of sediment, water, micronutrients, nitrogen, carbon, and phosphorus.

The flux of dissolved organic carbon to the ocean has increased as a result of land use, while the flux of metals has generally increased, but by an amount that varies substantially. Human use of freshwater resources and land has also significantly increased the flux of metals, sediment, and nutrients both to the oceans and within drainage basins, especially during the period of intensive agriculture. Fluxes have doubled of Si, P, and N, total organic carbon, and sediment to the ocean. Dense human populations in Asia produce perhaps as much as 50% of the total flux to the oceans of NO3-N.

The principal issue is to model and understand how specific terrestrially-derived materials are transformed, delivered to, and mobilized along the full cascade of landscape-fluvial systems. The drainage basin serves as an essential organizing principle in this discussion. Adequate consideration must be given to interactions within the river-riparian complex, the role of wetlands, and terrestrial ecosystem dynamics. The downstream boundary is equally complex: a) nutrient and sediment trapping and recycling occurs in the estuarine and near coastal environment, and b) the actual delivery of material to the open ocean appears quite variable and our knowledge is limited by an adequate database.

From a modeling perspective, there are several aspects that need to be addressed. First there is the cycling of water between the land and the atmosphere which can produce a "residual" or runoff. This water and the associated chemical load form the basis of rivers and the recharge of aquifers. This topic is the focus of water transport models which are tied to the coupled dynamics of terrestrial ecosystem and the land-water cycle. These models transforms complex patterns of generated runoff into horizontal transport through the drainage basin (Fig. 2).

A key component in the projection of future global change is the ability to predict future concentrations of atmospheric greenhouse gases such as carbon dioxide (CO2) and methane (CH4). Unfortunately, the current state of the science cannot completely account for the growth rate and interannual variations of atmospheric CO2 and CH4 with confidence, so accurate prediction of future concentrations is difficult. One of the objectives of GAIM is to develop coupled ecosystem-atmosphere models that describe time evolution of trace gases with changing climate and changes in anthropogenic forcing. Such coupled models must include an atmospheric module which adequately describes the chemical transformations with the atmosphere, and biospheric modules which describe the emissions from different ecosystems as well as how the emissions react to climate changes. The models must be based on process-level understanding of trace gas exchanges and transformations, but can be constrained by trace gas concentrations measured by the global observing network. This is possible only with a quantitative understanding of transport processes between sources, sites of chemical activity, and observation positions. As such, the atmosphere can act as an "integrator" of biogeochemcial processes. This is particularly valuable because of the large spatial and temporal heterogenieties inherent in surface process and fluxes. The interpretation of atmospheric data using transport models can provide an independent "integral constraint" for the upscaling techniques using local data, process models and remote sensing.

Only about half of the anthropogenic CO2 remains in the atmosphere, and the fate of the other half is not completely understood. Both the ocean and terrestrial biosphere currently act as significant sinks for anthropogenic CO2, but their relative contributions are a matter of intense debate [Houghton, 1995 #5411]. The terrestrial net sink is very difficult to measure directly, even at a single location, because it results from a small imbalance between large natural uptake and efflux by photosynthesis and ecosystem respiration, neither of which can be accurately measured at large spatial scales. Until the mechanisms involved in the terrestrial uptake are more clearly elucidated, predicting the future behavior of such a sink (and therefore the atmospheric concentration) will be very difficult. A significant step toward this end was taken in the recent GCTE synthesis [Walker, 1997 #5419].

The spatial and temporal distribution of atmospheric trace gas concentrations contains a great deal of information about the distribution of sources and sinks at the surface [Conway, 1994 #4075; Francey, 1995 #4098; Keeling, 1995 #4080]. This information is key to the overall effort to understand ecosystem-atmosphere interactions because (1) the concentration field provides validation data for the testing of coupled ecosystem-atmosphere models (a "bottom-up" approach to the problem); and (2) careful analysis of the changing distribution of trace gases can yield estimates of surface fluxes on the largest spatial scales (a "top-down" or "inverse" approach). Direct observation of trace gas concentrations through flask sampling and aircraft campaigns provides the data for these calculations, but calculation of surface emissions and uptake requires a detailed understanding of the atmospheric transport and chemical transformation that occur prior to samples being collected. This requires a numerical simulation model of scalar tracer transport by the atmosphere, which may be driven by analyzed winds or from meteorological principles, and may include gas transport, reactive chemistry, or both. The "top-down" or "inversion" approach has long been used to study sources and sinks of atmospheric CO2 [Ciais, 1995 #4040; Enting, 1989 #4099; Enting, 1991 #4100; Enting, 1995 #4101; Fung, 1983 #5147; Heimann, 1989 #4102; Tans, 1989 #4107; Tans, 1990 #4108]. It has also been used to study atmospheric CH4 [Fung, 1991 #4891], chlorofluorocarbons [Hartley, 1993 #5341; Prather, 1987 #5340], and many other trace gases, both reactive and inert.

As high time-resolution global data on additional species become available (d13C and d18O of atmospheric CO2 and atmospheric O2/N2 ratio), the use of synthesis inversion techniques with atmospheric tracer transport models will result in much more reliable estimates of the changing global carbon budget of the atmosphere. Improvements in the quality and quantity of the observational data and in the mathematical formalism associated with the inversion calculation have brought us to the point where one of the greatest sources of uncertainty now lies in the transport models themselves.

Atmospheric trace gas concentration is affected both by chemical and physical processes. Some trace gases such as methane are chemically reactive in the atmosphere, being lost to oxidation. They are also physically transported so that its atmospheric distribution is not directly related to its ground sources. Both mechanisms must be quantified in order to understand global atmospheric trace gas distribution, but atmospheric trace gas transport codes have been highly variable in their results, and in need of reconciliation. In order to effectively diagnose transport codes, they must be first compared in their prediction of passive trace gases, so that the physical effects can be separated from the chemical. Consequently, we have begun by considering the simpler case of chemically non-reactive CO2, and as a first step, we examine some passive tracers which have no sinks so that we can most effectively compare model results and thus promote model refinement. We will later treat reactive species such as methane separately, in preparation for ultimately incorporating these into the demonstrably realistic transport codes developed in association with the transport component of the project.

An important source of uncertainty in these calculations is the simulated transport itself, which varies among the many transport models used by the community. It is necessary to conduct a series of 3-dimensional tracer model intercomparison experiments with leading transport codes which are intended to (1) quantify the degree of uncertainty in current carbon budget estimates that results from uncertainty in model transport; (2) identify the specific sources of uncertainties in the models; and (3) identify key areas to focus future transport model development and improvements in the global observing system that will reduce the uncertainty in carbon budget inversion calculations.

The primary goal of Transcom is to improve our ability to deduce the Carbon budget of the Earth's surface from atmospheric observations. To accomplish this it will be necessary to quantify and diagnose the uncertainty in inversion calculations of the global carbon budget that result from errors in the simulated transport. The specific objectives of TransCom are (1) quantify the degree of uncertainty in current carbon budget estimates that results from uncertainty in model transport; (2) identify the specific sources of uncertainties in the models; and (3) identify key areas to focus future transport model development and improvements in the global observing system that will reduce the uncertainty in carbon budget inversion calculations. Results of an initial intercomparison of simulations of fossil fuel CO2 and the influence of seasonal vegetation were reported by [Rayner, 1995 #4106] and by [Law, 1996 #4105]. A subsequent phase involving calibration simulations of sulfur hexafluoride (SF6) has been conducted as well [Denning et al., in Review]. These preliminary results are summarized in GAIM Report #4 "Atmospheric Tracer Transport Model Intercomparison Project (TransCom)." Our initial intercomparison of global transport models used in the CO2 inversion problem revealed that inversion estimates of some carbon budget components may currently be uncertain by about a factor of two due to transport alone [Law, 1996 #4105; Rayner, 1995 #4106]. Because these models also form the dynamical core of many models of reactive chemical species, this problem is also of serious concern to those in the global atmospheric chemistry community.

Data needs: CO2, CH4 distribution, tower data.

OCMIP

One millennium from now, long after all fossil fuel supplies will have been exhausted, the ocean will have absorbed and retained about 7 out of every 8 molecules of anthropogenic CO2 ever emitted to the atmosphere. Confidence in this prediction comes from the agreement between related simulations in global ocean, carbon-cycle models, from the simplest (which divides the ocean into a few boxes) to the most complex (which contains many thousands of grid cells). Less confidence can be given to any one model's prediction about more immediate changes, e.g., over the next couple of centuries. Also less agreement exists concerning today's regional uptake patterns, as well as those during other times. Models differ substantially even without consideration of potential changes in ocean circulation and possible shifts in the composition of oceanic plankton, much less the large uncertainty in future emissions of anthropogenic CO2. Marine carbon cycle models have provided important constraints on the large-scale patterns of the marine uptake of anthropogenic CO2 [Siegenthaler, 1993 #4030; Maier-Reimer, 1996 #5007; Sarmiento, 1992 #4034], new production [Najjar, 1992 #4905; Bacastow, 1991 #4931], and remineralization [Shaffer, 1996 #5602; Lefevre, 1996 #5601], three important fluxes that have largely been elusive to direct observation over large spatial scales.

Consequently, the challenge to oceanographers is to synthesize the available data sets and incorporate into models that can be used for simulation of the partitioning of CO2 between the atmosphere and the ocean. GAIM is responding to that challenge in the Ocean Carbon-Cycle Model Intercomparison Project (OCMIP), a coordinated effort of evaluation and intercomparison of 3-D global marine carbon cycle models. The long range objective is to improve man's understanding of the ocean's carbon cycle and the crucial role of its major control -- ocean circulation. Such critical information from ocean carbon-cycle models will be necessary to help governments and international organizations make informed decisions concerning future increases in atmospheric CO2 and climate change. This understanding is also crucial to providing a firm base for proper evaluation of proposed geo-engineering solutions (e.g., ocean disposal of CO2 as a means to help limit rising atmospheric CO2).

Evaluation of marine carbon cycle models using observations of carbon-system and related parameters is necessary in order to establish the reliability of using such models for future prediction. A highly simplified "perturbation" approach, in which the natural carbon cycle and its attendant biological complexity are ignored, is feasible if ocean circulation and biogeochemistry can be assumed as invariant. [Siegenthaler, 1993 #4030]. However, there are many indications that the earth’s climate and ocean circulation are indeed changing and may change dramatically in the future (refs; [Manabe, 1994 #4999]. In that case, the potential for complex marine biogeochemical feedbacks is large [Sarmiento, 1996 #5000], and we are therefore behooved to develop the capability to model those aspects of the natural marine carbon cycle that are relevant to the air-sea partitioning of carbon dioxide. There are important data sets being compiled such as those by the Joint Global Ocean Flux Study (JGOFS), the next generation of satellite ocean color products, and a number of existing seasonal, global scale syntheses of nutrients [Conkright, 1995 #5008], dissolved oxygen [Levitus, 1994 #4913; Najjar, 1997 #5002], surface carbon dioxide [Takahashi, 1997 #5003] and chlorophyll [Yoder, 1993 #5004; Banse, 1994 #5006]. These data present an unprecedented opportunity for the evaluation of models of the natural marine carbon cycle.

Preliminary results of ocean carbon cycle model intercomparisons have identified differences between simulations of both natural and anthropogenic CO2 in various 3-D models, as well as differences between measured and simulated C-14 (for both natural and bomb components, separately) as a means to validate the model circulation fields which drive each of the four carbon-cycle models.

In order to understand the role of the ocean in the global carbon cycle, it will be necessary to improve 3-D marine carbon cycle models for their use in predicting the future partitioning of carbon dioxide between the atmosphere and the ocean, and better quantify important but poorly-known quantities that are relevant to the marine carbon cycle (e.g. new production, aphotic zone remineralization and anthropogenic CO2 fluxes and inventories). Since this will require an understanding of the natural marine carbon cycle and its anthropogenic perturbation, both of which are profoundly influenced by ocean circulation, it will necessary to first evaluate and intercompare the ability of 3-D global models to simulate the natural marine carbon cycle, the uptake of anthropogenic CO2, and response to perturbations.

Since preindustrial times, the rise in atmospheric CO2 has caused the change in the air-to-sea CO2 flux to be positive everywhere. This perturbation to the natural system, termed anthropogenic CO2, is difficult to measure in the ocean. In preliminary studies, four models have been used to estimate oceanic uptake of anthropogenic CO2. Although models agree to within 20% for global uptake of anthropogenic CO2 during the 1980's, regional uptake can differ by much more. Ocean uptake is highest in the high latitudes and at the equator (Fig. 4), i.e., in zones where deep waters uncontaminated with anthropogenic CO2 communicate readily with the surface (via upwelling and convection). The Southern Ocean dominates as the major sink also because of its large surface area. Models differ by more than 100% in their predictions of how much anthropogenic CO2 is absorbed in the Southern Ocean. Additionally, the predicted position of maximum uptake in the same region differs between models by nearly 20%.

Simulations differ most in the Southern Ocean for natural CO2, as well. OCMIP simulates the natural carbon cycle in order to properly validate ocean carbon-cycle models with ocean CO2 measurements as well as to better constrain our understanding of the partitioning of CO2 between the terrestrial biosphere and the ocean. Uncertainties in the latter are perhaps best exemplified by discrepancies between predictions from ocean models vs. those from atmospheric models which use observed distributions of CO2 and C-13 in the atmosphere to back out present-day carbon fluxes to and from the atmosphere [Tans, 1990 #4108; Ciais, 1995 #4040].

For the natural carbon cycle, it is necessary to run separate runs simulations to distinguish the effects of two major processes which along with ocean circulation control the distribution of natural CO2. The first relates to the temperature-dependent solubility of CO2. The cold waters which fill the deep ocean from the high latitudes are rich in CO2. Secondly, ocean biota act to reduce surface ocean CO2 through the combined action of planktonic uptake, rapid transport to depth of resulting particulate organic carbon, and subsequent bacterial degradation. We denote these two processes as the solubility and biological pumps, respectively [Volk, 1988 #5009].

We take the natural carbon cycle to be the state of the ocean-atmosphere carbon system prior to significant anthropogenic influence on the global carbon budget, usually considered to be the millennia leading up to the 19th century, when atmospheric CO2 concentrations were "steady" at about 280 matm. The natural marine carbon cycle plays an important role in the partitioning of carbon dioxide between the atmosphere and the ocean through two important processes: the solubility pump and the biological pump [Volk, 1988 #5009], both of which act to create a global mean increase of dissolved inorganic carbon (DIC) with depth, and therefore to maintain atmospheric CO2 at a level considerably lower by about a factor of three [Najjar, 1992 #4905] than it would otherwise be.

The solubility pump maintains a vertical DIC gradient due to the fact that cold waters, which originate in high latitudes and fill up the deep ocean, can hold more DIC than warm waters at equilibrium with a fixed atmospheric pCO2, a result of the higher solubility and dissociation of CO2 (into carbonate and bicarbonate ions) in cold water. The vertical DIC gradient depends not only on the vertical temperature gradient but also the degree to which surface waters equilibrate with the atmosphere before sinking. Unlike most gases, the equilibration of surface water CO2 with the atmosphere takes about one year, and therefore the strength of the solubility pump may depend critically on the kinetics of air-sea gas exchange.

The biological pump consists of two separate pumps: that of organic matter and calcium carbonate. The organic matter pump affects the DIC distribution through the photosynthetic formation of organic carbon in surface waters and the sinking and subsequent remineralization of this organic matter deeper in the water column. The carbonate pump affects the DIC distribution though the biogenic precipitation of calcium carbonate in surface waters and the subsequent sinking and dissolution of this material deeper in the water column. These two pumps also affect the alkalinity of seawater through the nitrate and dissolved calcium distributions.

The prominent importance of the marine carbon cycle in global carbon cycle make it essential to accurately model ocean circulation and its role in carbon uptake, release, and transport. As further models develop and are compared on the basis of their performance regarding natural and anthropogenic CO2 and various tracers, GAIM will work closely with the Core Projects, particularly JGOFS, to ensure that the models are compatible for coupling with atmospheric and terrestrial models toward the goal of integrated Earth System models.

Data needs: marine CO2, ocean color, circulation (horizontal & vertical), bomb 14C.

1C. How will changing climate and land use alter the couplings between biogeochemical cycles of different elements?

Human land use has transformed most of the Earth’s surface (i.e., its land cover) during the last 200 years with prominent influence on several subsystems of the total Earth system. The terrestrial biosphere itself is fundamentally modified by the activities of land clearing for agriculture and other purposes (industrialization, urbanization), as well as by forest and rangeland management which is taking place throughout the world although with widely differing intensities. The atmosphere is affected by these changes both through an altered energy balance over a significant part of the land surface (i.e., the more intensively managed areas) where physical properties such as albedo and roughness are modified, and through changed fluxes of H2O, CO2, CH4 and other trace gases between soils/vegetation and the troposphere [Mosier, 1991 #4049; Solomon, 1993 #4764; Woodwell, 1983 #4885]. Finally, in many coastal zones, the oceans are receiving greatly altered fluxes of carbon, nutrients and inorganic sediments from rivers due to altered land use. Aside from climate change, these changes together represent the most critical component of anthropogenic global change [Turner, 1995 #4886].

Land use changes have profound effects on the biogeochemistry of carbon, radiatively and photochemically active gases, and aerosol production (dust and biomass burning). Land use changes also affect hydrology and erosion, and by changing surface albedo and energy exchange, can have direct effects on climate. People often create highly heterogeneous landscapes which form mosaics that may encompass activities with highly divergent effects on ecological processes. The spatial arrangement of landscapes can affect exchanges of water and associated solutes and particulates in freshwater and coastal margin areas, with land cover at the land-water margins having substantial effects on water chemistry. The arrangement of landscapes also affects biological diversity, pests and pathogens, invasibility and extinctions.

The amount of forcing land use changes exert on the total Earth system is currently unknown. Sensitivity studies with altered land cover distributions in general circulation models have shown that unrealistically drastic changes, such as total deforestation of all tropical or boreal forests, may lead to feedbacks in atmospheric circulation, resulting in inadequate climatic conditions to support the vegetation which occurred prior to the perturbation [Claussen, 1996 #4887; Kutzbach, 1996 #4882]. As purely sensitivity studies of the atmospheric circulation, however, these global experiments do not attempt to mimic the land use changes which have actually occurred–they only indicate that such feedbacks indeed may be critical for the stability of the overall system. Regional climate simulations, on the other hand, have shown that at the continental scale, important teleconnections may exist through which tropical forest clearing may cause a change in climate conditions in much less disturbed areas [Salati, 1986 #4888].

A fundamental research issue from an Earth System modeling point of view, is the degree of sensitivity of the feedbacks between terrestrial biosphere, atmosphere and oceans, to the processes involved in land use change, particularly with respect to the required scale resolutions in time and space. It is possible that the required accuracy of observed or simulated data sets of changing land cover could be at a relatively coarse level. If so, then investigation of the past dynamics of such interactions, (e.g. for the last 200 years) may provide insights and should be explored in tandem with the development of predictive models that can be extrapolated into the future.

In terrestrial systems, the allocation of plant productivity among different plant parts is of fundamental importance. Although theory and experiments suggest that allocation should respond to climate and other environmental changes, very few observations exist for model validation. Litterfall, litter chemistry and decomposition rates, soil respiration, and nutrient availability are key measurements at ecosystem scale for accurate interpretation of observed changes in fluxes and storage. The lack of validation data extends to marine systems as well with regard to phytoplankton and trophic levels. Export of particulate and dissolved organic matter below the euphotic zone is a primary variable. Measurements of NPP, respiration and allocation, coupled with net CO2 fluxes determined from eddy covariance will be necessary for validation of modelled marine ecosystem changes.

Today, a consistent data base on land use change over the last 200 years at a global scale does not exist. However, little progress can be made on predicting future land use change, nor on its implications for the coupled Earth System, without testing models with observations from the past. The period of the last 200 years is particularly important for this purpose because in that time not only did the global population increase 10-fold, but population density in several regions of the world increased by more than 100 times, and agriculture replaced natural vegetation rapidly, changing the face of large parts of North America, China, India, and elsewhere. In order to study the causes and effects of land use change, it will be important to have parallel efforts to compile data regarding population changes as well as climate, soil carbon and land productivity at the same temporal and spatial scale.

Starting in 1980, global land cover data sets began to be available from satellites both at low resolution (1-4 km) and, for specific regions, at high resolution (10-30 meters). Such observations will continue to be made in the future, complemented with radar observations of land the land surface as well. These data have been classified into 17 classes including several classes of land cover defined by land use. It will be important to build a land use data base from historical and other sources up to 1980, and to ensure consistency between both types of observations. These data will make it possible to establish a time series of land cover change so that terrestrial biospheric trace gas sources and sinks can be related to future observed variations in atmospheric composition and temperature. This will lead to a better assessment of anthropogenic drivers of climate changes in the future.

Data needs: 200 year land cover/use data

2. How do changes in ecosystems interact with the physical climate system?

What processes determine how climate change affects marine and terrestrial ecosystems, and what are the potential climate feedbacks due to these processes?

The physical climate system is comprised of atmospheric, oceanic, and land (and ice) components. Their changes affect marine and terrestrial ecosystems, in general through controls on temperatures, solar radiation, moisture, and other environmental factors, and through nutrient addition and removal by transport. Over land, nutrients are largely added by atmospheric transport, accompanied by wet and dry removal processes. Removal is largely accomplished through runoff but in some cases, e.g. ammonium ions and N2O, the atmosphere also removes nutrients. Similarly, nutrients are removed and added to the oceanic euphotic zones through climate related transport processes, including turbulent mixing, large scale circulations and upwelling, and gravitational settling. The large changes in the nutrient availibilities during El Niño is a good example of climate coupling to oceanic ecosystems.

Ecosystems are strongly controlled by land and oceanic temperatures, at a cellular level through the temperature dependencies of enzymatic processes, and cellular sensitivities to temperatures outside the organisms' range of acclimation. Temperatures in turn depend on solar and thermal radiation and surface sensible and latent heat fluxes. These have both atmospheric and especially over land ecosystem controls. The atmospheric controls largely depend on the atmospheric hydrological cycle in providing clouds, precipitation, and surface moisture fields. On the ground, soil moisture is needed to maintain the evapotranspiration of vegetation and hence allow the uptake of CO2 from the air, and N and P from the soil. High water tables can reduce the oxidation state of the soil, switching biological carbon and nitrogen oxidation processes into other pathways such as methanogenesis.

Ecosystems in turn feedback on climate through their controls on surface temperatures and exchanges of water, CO2, and other climatalogically important gases. Temperatures are particularly affected by ecosystem controls of albedo and surface roughness. Shading of snow surfaces by forests, especially in the springtime, has a major effect on high latitude temperatures relative to the effect of poleward tundra or equatorward grasslands. Tropical forest cool relative to pastures even though they absorb more energy through lower abedos. Their cooling is affected by greater roughness, hence more turbulent mixing, and greater evapotranspiration, especially during the dry season, where soil layers reached by roots become dry.

The carbon cycle is dominated by the exchange of CO2 between the land, atmosphere and ocean. As a greenhouse gas, CO2 affects climate by altering the atmospheric energy balance. Exchanges between the atmosphere and other reservoirs determine how atmospheric CO2 concentration varies with climate change or in response to perturbations such as fossil fuel combustion or land use. Hence the globally averaged CO2 concentration constitutes a direct link between the physical climate system and the terrestrial and marine biogeochemical cycles. There are also other geochemical cycles that may contribute to a direct link between biology and climate. Two such examples are the effect of DMS production by some species of marine phytoplankton and the fertilizing effect of atmospheric iron deposition on marine primary productivity [Bates, 1997 #5208; Hegg, 1991 #5209; Platt, 1997 #5207].

Marine ecosystem models often simulate the coupled dynamics of carbon and nutrients, but differ from terrestrial ecosystem models in that the focus is much more on trophic interactions. The understanding of processes that determine abundance, fluctuations, and production of marine animals must necessarily involve coupled physical-biological models, linking performance of the individual organism to local physical processes, and linking both the biology and local physics to basin-scale changes in global climate. Modelling is expected to play an important role at several levels. The explicit incorporation of physical variables and processes in biological population models should lead to significant strides forward. Appropriately constructed models of both physical and biological processes should guide the choice of field experiments and observations, while result of those field exercises should feed back interactively into the models.

2A. What have been the impacts of climate changes on marine and terrestrial ecosystems during the past 200 ka, and what have been the feedback effects on the physical atmosphere/ocean system?

Sedimentary archives recording the past composition of biotic assemblages on land and in the ocean are providing an ever-increasing body of knowledge about the nature of climatically forced variations in ecosystem structure on decadal/centennial to multimillenial time scales. Important examples of such data sources are pollen assemblages in peat and lake sediments, plant macrofossil assemblages in middens, fossil vertebrate and beetle remains in a variety of terrestrial deposits, and microfossil assemblages of planktonic and benthonic marine organisms in deep-sea sediments. Some of these data sources enable us to make inferences regarding climate (for example, assemblages of planktonic foraminifera in deep-sea cores are commonly interpreted in terms of sea-surface temperatures). Other data from the same archives give additional paleoclimatic information, e.g. alkenone ratios in marine sediments (a geochemical proxy for sea-surface temperatures) and a variety of biological and geomorphic indicators of changing lake levels (a proxy for lake and catchment water balance).

Although the role of natural climate change on glacial-interglacial time scales in forcing changes in ecosystems has long been understood, it has only more recently been appreciated that such large-scale changes in the distribution of terrestrial ecosystem types (biomes) could have major feedback effects on the physical climate system. Such feedbacks, acting through the modification of land-atmosphere exchanges by vegetation properties such as surface albedo and conductance to water vapor are known as biogeophysical feedbacks. They are distinct from biogeochemical feedbacks (involving the exchange of CO2 and other trace gases). Biogeophysical feedbacks act rapidly and directly on the atmosphere, whereas biogeochemical feedbacks act indirectly through modifying the atmospheric burden of radiatively active gases and aerosols. However, large-scale structural changes in ecosystems occur naturally at slower rates than modifications to CO2 or trace gas exchange. Thus, in order to investigate the working of biogeophysical feedbacks under natural conditions, it is necessary to describe and model conditions farther back in the past than it is provided by direct observational records.

It will ultimately be necessary to quantify the role of biogeophysical feedbacks in determining the change in mean climate and biome distribution between various times in the geologic past and the present. Preliminary sensitivity experiments with AGCMs where the land surface has been artificially changed to match the known distribution of biomes of the past (e.g. 6000 yr BP) have shown that both the modelled monsoon expansion and the high-latitude warming are greatly amplified by the change in biome distribution [Foley, 1994 #4881; Kutzbach, 1996 #4882]. AGCM simulations that do not allow for these land-surface changes greatly underestimate both climate changes- the qualitative nature of the changes is generally simulated, but the magnitudes fall well short of those indicated by the validation data such as those synthesized by the BOIME 6000 project.

Similar feedbacks may have been involved at other times, too. The climate during the last glacial-maximum may have been considerably modified by the great expansion of tropical grasslands [Crowley, 1997 #4883], itself a consequence of cold ocean conditions (generating less precipitation) and possibly also of low CO2 (favoring C4 over C3-type photosynthesis in the tropics). The start of the last glaciation may also have been aided by the effects of biogeophysical feedbacks: low northern-hemisphere summer insulation led to cold summers in the north, a retreat of forest, and an increase in spring albedo that set off a "chain reaction" helping to start snow accumulation in the areas where ice sheets were later to develop. On the other hand, experiments with current DGVMs have shown that many vegetation types exhibit considerable "inertia" (on time scales of a century or more) in their response to scenarios of future climate change. In some cases this inertia simply buffers the system against rapid change but in other cases where tolerance levels of existing vegetation (e.g. with respect to drought) are exceeded, there can be an opposite effect such that the existing vegetation is not immediately replaced by a climatically adapted type but rather by a generalist or opportunist assemblage. The situation is further complicated by land use changes which can serve as a trigger. The potential importance of such effects is not well known, yet is controversial because these two types of ecosystem response have radically different implications for the feedback of climate-induced vegetation changes on both the physical climate system and the global carbon cycle. These issues all point to the potentially large role of vegetation changes in the climate system, prompting the development of asynchronously coupled AGCM/biome models (as used e.g. by de Noblet et al. 1996) and ultimately, GCMs that include a dynamical biosphere component.

Although these various modelling studies focused on land-atmosphere interactions, there is ample evidence that the actual past climate change involved the ocean as well. In the model of Foley et al. (1994), sea-ice albedo feedback was crucial in amplifying the effect of a relatively small change in total forest area around the Arctic. In northern Africa during the Holocene, coupled model experiments have generally been unable to simulate the full northward migration of the monsoon as shown by the paleodata and there reasons to suspect that changed North Atlantic SST patterns might have played a synergistic role. Indded, recent model experiments with a simplified coupled ocean-atmosphere-vegetation model have successfully predicted both the general features of the mid-Holocene climate and the observed abrupt collapse of the northern African monsoon a little over 5000 years ago. Glacial initiation is certainly not limited to orbital forcing and atmosphere-land interactions; major changes in the thermohaline circulation also occurred towards the end of the last interglacial and presumably contributed to the eventual global cooling. Full understanding of such atmosphere-biosphere-ocean interactions may also require consideration of how vegetation properties influence freshwater fluxes, an important boundary condition for ocean circulation.

These case studies all indicate the necessity of modelling atmospheric, oceanic and biospheric dynamics in a single framework. They cast some doubt on current models of the physical atmosphere-ocean system, as used to predict possible future climate changes, due to the omission of biogeophysical (and hydrological?) feedbacks associated with changes in terrestrial ecosystems. In particular, current models omit a potential mechanism for amplifying warming in high latitudes. Paleodata can provide benchmarks for the evaluation of more closely coupled earth system models, when forced by known boundary conditions in the past. Yet the task of predicting the future is more complex still, because the land surface no longer evolves under the control of the physical environment alone. Instead, it is greatly (and increasingly) modified by land use changes, including deforestation, which themselves would be expected to alter climate through biogeophysical mechanisms. This fundamental question thus relates to developing a functional understanding of the role of the biosphere as an interactive component of the physical climate system. A subsequent section deals in part with how these same biogeophysical processes may act as an additional, so far neglected component of how human activities may be modifying the global climate.

Data needs: Pollen records, ice cores, marine sediments, lake records, corals

2B. What is the role of ecosystem level processes (growth, competition, disturbance, mortality, decomposition, soil organic matter dynamics, migration) on the broad-scale structure of the biosphere? To what extent may plant population processes accelerate or delay climate- or CO2-related changes in the distribution of vegetation?

The importance of climate-vegetation interactions in governing Earth's surface energy and moisture fluxes has been recognized for many years with regard to the tropics [Dickinson, 1988 #5236; Henderson-Sellers, 1984 #3672] and more recently for boreal forests [Bonan, 1992 #5594]. Plant metabolic processes move carbon, nutrients and water through plants and soil on rapid as well as intermediate time scales. This cycling affects the energy balance and provides key controls on biogenic emissions. Some of the carbon fixed by photosynthesis is incorporated into plant tissue and is thus delayed from returning to the atmosphere until it is oxidized by fire or decomposition. The structure of terrestrial ecosystems, which responds on longer time scales, is the integration of the intermediate time scale processes, and responds to climate changes with alterations of species composition as well as migration of biomes. The carbon loop is ultimately closed back to the climate system, since it is the structure of ecosystems that controls the terrestrial boundary conditions for carbon and water exchange, surface roughness, albedo, and latent heat exchange. It is possible to treat the hierarchy of time scales by using a nested approach. For example, the metabolic activities of terrestrial plants associated with growth and maintenance constitute the fastest interactions (seconds to days) and determine latent heat, energy water and CO2 exchange through gross photosynthesis and respiration. Intermediate processes (days to weeks) include the development of leaf area, soil water balance, trace gas exchange, and decomposition of organic soil materials. Longer term (annual) time steps encompass net primary productivity, ecosystem production, and long-term changes in carbon and nutrient pools in plants and soils. At the longest time scales (decades to millennia), biome distributions respond to changes in climate and atmospheric chemistry.

Vegetation density and net primary productivity react to changes in atmospheric composition as well as climate (temperature, precipitation) [Myneni, 1997 #4819]. The exercise of predicting changes in vegetation as a result of various global change scenarios should pave the way for accurate ecosystem-atmosphere interaction modelling capabilities within GAIM and IGBP in general. Recent research has shown that the distribution of vegetation has a significant influence on climate, particularly over the continents.

In climate models, changing vegetation corresponding to land use change modifies the climate significantly (because of changes to albedo, roughness and the Bowen ratio). The models indicate that expansion of boreal forests in response to atmospheric greenhouse warming would enhance warming through reduced albedo, particularly in times of snow cover. This feedback between ecosystem migration and climate is one example of the acceleration of global change by ecosystem level processes. Exploration of the details of these processes requires further development of Dynamic Global Vegetation Models (DGVMs). These models treat successional dynamics as well as ecosystem redistribution. For example, following the abandonment of agricultural land, fluxes and pools of C, N, and P in secondary vegetation often do not attain the same levels as found in the "undisturbed" natural vegetation. The recovery of natural vegetation in abandoned areas depends on the intensity and duration of agricultural activity and the amount of soil organic matter at the time of abandonment. To simulate the biogeochemistry of secondary vegetation, models must capture patterns of plant growth during secondary succession. These patterns depend substantially on the status of nutrient pools inherited from the previous stage. Changes in hydrology are also involved because plants that experience water stress will alter the allocation of carbon toward the roots. Presses such as reproduction, establishment and light competition are included in DGVMs and interact with C, N and H2O cycles. Disturbance regimes such as fire and land use are also incorporated. These forcing functions may themselves be altered in response to ecosystem changes exhibited by the terrestrial system.

In paleoexperiments using climate models, the simulation of reconstructed environments appears to require correct specification of land cover. On the basis of models, we can thus infer how particular ecosystem types partition the surface energy budget. Two main lines of research follow from this observation. First is the coupling of DGVMs to physical climate systems to allow for reciprocal influences as climate and vegetation change over time. Second is a concerted effort to test models of land-atmosphere coupling against observations. While this is difficult using traditional predicted-observed relationships (because of the spatial continuity of the atmosphere making spatial comparisons only weakly appropriate, and because of the lack of ability to replicate). Replications in time using data assimilation and forecast-validation cycles may be a new avenue for testing land-atmosphere coupling, as they have been for other aspects of atmospheric models.

The coupling of ecological models with climate models needs to be considerably developed before providing robust results. However, some preliminary interpretations are possible based on present models. The most fundamental is that climate-vegetation interactions at the ecosystem level can substantially alter the sensitivity of climate models. In addition, correlations have been found between CO2 concentrations and stomatal parameters; C4 plants may have proliferated in the Miocene due to H2O and CO2 stress. If these stresses are reduced in response to changes in atmospheric chemistry and precipitation patterns, there may be fundamental consequences in terrestrial vegetation and thus to the relationship between climate and ecosystems. In addition, several lines of evidence (carbon isotopes, pollen records, ecosystem model results) point to a significant role of CO2 concentration in influencing the changes in vegetation distribution between the last glacial maximum and the Holocene. AGCM experiments have indicated that the effect of changing CO2 on stomatal conductance alone (and possible further compensatory effects on leaf area index) could have significant and complex regional effects on climate. These issues must be resolved before we can hope to understand the effects of anthropogenically-induced changes in atmospheric chemistry and climate.

Data needs: H2O CO2 heat momentum fluxes, leaf area, ET, 200 year land cover/use data

2C. How will future natural and anthropogenic changes in ecosystems and their interactions with climate affect the Earth System? In particular, what are the likely consequences of future land-use changes for the climate of the next 200 years?

It is well known that even though the continental surface represents only one third of the planet and that not all of this is vegetated, the area of foliage exceeds the total planetary surface. Anthropogenic alterations in land cover can thus have a profound effect on the exchange of energy, water and nutrients between the soil, terrestiral ecosystems, and the atmosphere.

Land plants are critical conduits of energy, water and trace gas exchanges. Despite these facts, global climate models have only recently begun to incorporate sophisticated parameterizations of the land/vegetation surface. Most coupled ocean-atmosphere GCMs use a "bucket" model for continental processes. The simple 15cm "bucket" model is a much poorer representation of land/vegetation processes than many of the SVATS (Soil-Vegetation-atmosphere transfer schemes) now available.

Critical gaps between models have resulted from the different scientific and technical issues confronting climate modellers and the ecosystem modellers. The former group focus on expediency in computational process and adequacy in delivering energy, water and momentum exchanges and budgets. The latter group has developed its models from two different, and complementary directions: 1) energy, carbon and other trace element exchanges and budgets, and 2) vegetation dynamics and relations to climate. In addition, there are significant differences in the time and space scales of these model types: "land-surface scheme" - minutes to days; "biogeochemical" - days to years; and "biome model" - years to millennia. Some of these are included in Dynamic Global Vegetation Models (DGVMs), and are even now beginning to be addressed in atmospheric GCMs.

The credibility at large scales of current models remains essentially unknown because of a chronic lack of validation data. The data void is accentuated by the diversity of quantities being predicted: energy; carbon and nitrogen; and ecosystem type distributions. In the absence of adequate observational information, many of the inputs for one model type are derived from predictions of other models which masquerade as "data". There is an urgent need to assemble and document data (observed and derived from data assimilation) of direct use for evaluating vegetation/biome models.

There is also an urgent need for a well organized suite of model intercomparisons. GAIM is moving into a position to establish a nested series of model intercomparison projects aimed at establishing the importance of land/vegetation for climate simulations. Each level of the nested set should be defined in terms of the best available data for both input and evaluation so that model performance becomes the primary discriminant. Nesting is crucial because much richer data sets are available at the site and watershed level, thus enabling testing many of the internal dynamics. However, testing overall performance across large regional to global environmental gradients is a sine qua non for credibility at large scales. Thus, evaluation at multiple scales is a key part of our strategy.

Human land use is expected to continue to intensify in the future, likely at accelerated rates in large areas, due to

1) agricultural responses to a growing food demand;

2) changes in forest exploitation (clearing, wood production);

3) afforestation measures attempting to sequester carbon that otherwise would reside longer in the atmosphere; and

4) introduction of additional exotic species into disturbed and undisturbed ecosystems.

Predictive models of the Earth System need to account for past and present extent and intensity of human land modification, and the possible changes OF these in the future. To predict future changes in land use and land cover at the global scale is an unprecedented challenge in global change research, since human decision making at the local scale is one of the most important drivers. Projections therefore carry uncertainties of a different kind, as compared to physical or biological models, reflecting socio-economic constraints as well as non-monetary influences on land use systems. This represents one of the most critical links between IGBP and IHDP, and is a focus of Fundamental Question #4.

Data needs: 200 year land cover/use data, GPPDI, 100 year climate data

3. How do changes in the radiatively active gas composition of the atmosphere interact with the biosphere and physical climate system?

What controls atmospheric composition and what feedbacks exist between trace gases and terrestrial/marine sources and sinks?

Terrestrial and marine ecosystems are a primary source/sink of atmospheric trace gases. However, because ecosystems interact with both climate and atmospheric CO2, changes in CO2 concentrations and their associated climatic effects may have a significant influence on trace gas exchange. Methane for example is a product of anaerobic respiration in wetlands. The rate at which it is produced depends on temperature and inundation such that increases in temperature and precipitation would lead to greater methane emissions. Drying leads to reduction in methane production, but it has a further effect of allowing oxidation (or even fire), thus producing CO2. The large areas of high latitude peat may be vulnerable to a shift from reducing to oxidizing conditions, potentially leading to a massive discharge of carbon into the atmosphere. Methane production in wetlands is also linked to atmospheric CO2 concentration, and methanotrophs in dry strata of wetlands can act as methane sinks.

To predict the variation of CO2 as a function of the environment, both in an "off-line mode", i.e. with specified environmental parameters (such as temperature, precipitation), or as component in a coupled earth system model will require a description of 1) the oceanic carbon system including the carbonate chemistry, nutrient cycles, marine biota and sedimentation processes as embedded in the physical marine environment; 2) the carbon system on land including living vegetation and soils and its coupling to both the nutrient cycles of nitrogen and phosphorus, and to the exchange fluxes of water, energy and momentum; and 3) the atmospheric chemistry of the nutrients or trace elements that affect terrestrial and marine primary production.

Changes in the radiative forcing of the climate system by variations of CO2 are documented through observations from air bubbles in ice cores and direct measurements [Bruno, 1997 #5210; Leuenberger, 1992 #5211] over at least four different timescales:

1. Parallel to the glacial-interglacial cycles, significant CO2 concentration variations occurred which reflect concomitant changes in altered distributions of ecosystems on land and changes in ocean circulation, chemistry and marine biology.

2. Following the last ice age, Holocene atmospheric CO2 first underwent a slow increase but was otherwise characterized by relatively constant CO2 levels until the start of the industrial era.

3. During the industrial era, fossil fuel CO2 emissions and changes in land use induced an almost exponential CO2 increase.

4. Beginning in 1958, direct observations document with increasing detail the temporal and spatial patterns of the industrial era CO2 increase.

There are a few particularly conspicuous and intriguing features in these records that require an in-depth assessment by comprehensive models of the global carbon and nutrient cycles. Atmospheric CO2 concentration fell by almost one third during glacial times as compared to interglacial periods and was accompanied by some changes in oceanic nutrient distribution and primary productivity [Mortlock, 1991 #5213; Murray, 1993 #5212]. In addition there has been a substantial increase observed in the amplitude of the seasonal cycle of atmospheric CO2 concentration over the time of direct measurements (more than 15% over 35 years) [Dianovklokov, 1989 #5215; Keeling, 1996 #5058; Randerson, 1997 #5214]. There have also been interannual variations of the atmospheric CO2 growth rate which are believed to reflect transient fluctuations induced by climate fluctuations in the terrestrial and oceanic carbon systems [Braswell, 1997 #5113]; Kindermann, 1996 #4541]. Finally, excess CO2 from anthropogenic emissions (emissions from fossil fuel burning and changes in land use) has been partitioned between the atmosphere, ocean and terrestrial biosphere over the industrial era in ways that must ultimately reflect biogeochemical flux mechanisms.

A host of additional observations exist which allow an assessment of the relative roles of the various components of the earth system that affect changes in atmospheric CO2, and which provide constraints on comprehensive models of the global carbon and nutrient cycles. These include:

• the high resolution spatial and temporal distribution of atmospheric CO2 and some other atmospheric components as observed by the global monitoring networks

• measurements of the air-sea difference in partial pressure in CO2 indicating air sea carbon exchanges

• stable isotopes of carbon and oxygen in atmospheric CO2 (13C, 18O),

• atmospheric oxygen/nitrogen ratio

• the distribution and evolution of natural and bomb radiocarbon in the different reservoirs

• time series of forest inventory data and dendrochronology

• remote sensing data from satellites: the

	Climate system: CO2 - O2
OBSERVATIONAL RECORD	ADDITIONAL OBSERVATIONS TO BE EXPLAINED	FORCINGS (changes in 1-way; 2-way)
Keeling (D) 1957 + DCO2 seasonal cycle	10% change in NDVI	temperature/rainfall
Keeling (D) 1957 + DCO2 seasonal cycle	ocean color as biotic index, pCO2 and S'CO2	seasonal change (spring)
Flask network CO2	N deposition	SST
O2/N2 (last 5 years)	-	UVB
Natural 13C, 14C, 18O	-	sealevel
Bomb 14C	-	radiation/cloud
280-200-280 ice core (150 ka)	-	land use
Mauna Loa record	-	energy CO2 - N

	Climate system: CH4
OBSERVATIONAL RECORD	ADDITIONAL OBSERVATIONS	FORCINGS
Ice core 1700-	13CH4, other isotopes	climate (T,P, water vapor)
atmospheric record - historical	wetlands	wetlands
100 year firn/snow record	recent time/space CH4	rice, agriculture
long-term a) Holocene CH4 record; b) glacial	-	energy

	Climate system: S
OBSERVATIONAL RECORD	ADDITIONAL OBSERVATIONS	FORCINGS
Greenland & other ice cores	-	industrial emissions
DMS distributions	Aerosols, etc.	coal, oil, etc.
-	-	natural cycles (DMS, etc.)

	Vegetation-Climate interaction
OBSERVATIONAL RECORD	ADDITIONAL OBSERVATIONS	FORCINGS
distribution of vegetation- Current, Past	Soil moisture	Atmospheric trace gas composition- CO2, etc.
NDVI: patterns of seasonal activity	-	Changes in hydrology
Time series of phenology	-	-

	Land cover changes
OBSERVATIONAL RECORD	ADDITIONAL OBSERVATIONS	FORCINGS
Land cover	N deposition	economics
DSoil carbon	changes in erosion & sedimentation	population & demographics
-	-	international social programs
-	-	climate variability