Executive summary
J. M. Mellilo, D. 0. Hall, G. I. AgrenSCOPE 56 Global Change: Effects on Coniferous Forests and Grasslands, A.I. Breymeyer, D.O. Hall, J.M. Melillo, and G. Ågren eds., 1996, 459 pp.
1.1 INTRODUCTION
Human activities are rapidly changing the global environment. They are causing increases in atmospheric concentrations of carbon dioxide and other greenhouse gases; in turn, these increases are thought to be a primary force for global climate change. By the middle of the next century, the mean global temperature is expected to be higher than it has been for millions of years. In addition, changes in rainfall patterns and increased storm intensities are predicted.
A rapidly changing global climate will certainly affect the fundamental ecological processes of the planet. Exactly how ecosystems will respond to climate change is not well known. Predicting the effects of climate change on ecosystems is complicated by interactions among the effects of concurrent changes in atmospheric carbon dioxide concentration, available water, and temperature on a number of key ecosystem processes such as primary production and decomposition.
A major challenge is to understand the fate of man-induced sources of emissions of greenhouse gases. For carbon dioxide there are presently two principal sources: tropical land use changes and combustion of fossil fuels plus cement production. Some of this carbon dioxide remains in the atmosphere, some is absorbed in the oceans, and some disappears due to an increased plant production. Expanding the 1994 IPCC assessment, Schimel (1995) gives the following average balance sheet for the period 1980-1989:
Sources (1015 g . yr-1)
Tropical land use change 1.6
Fossil fuel combustion and cement production 5.5
Sinks (1015 g . yr-1)
Regrowth of Northern Hemisphere forests 0.5
Increased terrestrial plant production 1.6
(From carbon dioxide fertilization 1.0)
(From nitrogen (N) fertilization 0.6)
Net uptake in oceans 2.0
Increase in the atmosphere 3.2This budget is not balanced and there are uncertainties of 50% or more for all terms except the increase in the atmosphere and fossil fuel combustion which are reasonably accurately known. The requirement on accuracy in the terms in the terrestrial budget has to be seen against the background of the terrestrial net primary production of 60 - 1015 g . yr-1.
To improve our understanding of ecosystem responses to climate change, SCOPE, the Scientific Committee on Problems of the Environment, sponsored a project on the effects of climate change on production and decomposition in coniferous forests and grasslands. The project had three objectives: (1) a critical assessment of our understanding of the effects of climate on production and decomposition processes in coniferous forests and grasslands, temperate and tropical; (2) a review of extant forest and grassland ecosystem models, and an evaluation of their usefulness for studies of the effects of climate change on production and decay processes; and (3) the development of a foundation for the design of diagnostic and predictive models to describe the effects of climate change on forests and grasslands.
1.2 ORGANIZATION OF THE STUDY
The report contains five major sections. In the executive summary chapter we briefly review the highlights of the volume and identify future experimental and modelling needs.
Section two comprises four chapters, each of which reviews what we know about key aspects of the carbon (C) budgets in major coniferous forest and grassland biomes: boreal forests; temperate coniferous forests; tropical savannas, woodlands and grasslands; temperate grasslands. Several of these chapters discuss the usefulness of studying ecosystems along climatic gradients as part of a global change research program. All point to key research areas for a global change research agenda.
The third section of the book deals with 'ecosystem physiological responses' to global change. Each of the five chapters in this section reviews how an important ecological process might respond to changes in either climate, atmospheric carbon dioxide, or both. The five processes considered are photosynthesis, plant respiration, C allocation in plants, litter decomposition and soil organic matter dynamics.
Section four of the book summarizes the results of work done within the SCOPE project with both grassland and forest models. In a total of four chapters, grassland and forest model structures are reviewed and compared. Model runs for both biome types for various climate change scenarios are reported and analyzed. Grassland models used in this work are CENTURY and GRASS. The forest models used are BIOMASS, BIOME-BGC, CENTURY (forest version), HYBRID, MBL-GEM, PnET-CN, and Q. The focus of both the grassland and forest modelling activities is primarily site-specific. Regional and global modelling activities are reviewed in section five.
Comprised of two chapters, section five deals with two aspects of ecological response to global change; biogeochemical response and biogeographical response. The biogeochemical response includes C and nutrient cycling, while the biogeographical response includes the shifting of species from place to place with climate change. One important conclusion to be drawn from the final section is the importance of coupling biogeochemistry and biogeography models for climate change research. The coupling would not only provide insights into ecological responses to climate change but also into the feedbacks between the terrestrial biosphere and the climate system.
This book reports the state of knowledge as of 1993 but no major developments have occurred during the time between the writing of this report and its final publication date that change the report's conclusions.
1.3 FOREST AND GRASSLAND ECOSYSTEMS
1.3.1 Boreal forests
The boreal forest covers approximately 12 x 106 km2 or 17% of the earth's land surface. There is now a reasonable data base at the regional and subcontinental scales of the biomass of the boreal forest, though several areas (e.g. Eurasia) still appear to be poorly represented. There is less certainty about the annual productivity of the boreal forest, owing to the more limited number of sites at which measurements have been made, potential differences in measurement techniques (e.g. increment vs. whole-ecosystem C budget) and year-to-year variability, as well as the C stored in the soil. Storage of C in boreal forest soils is highly variable and may be a major area of uncertainty in the C storage of the biome.
Fire is recognized as an important control on the structure and functioning of the boreal forest. Few quantitative data are available for changes in C storage and fluxes in the whole ecosystem at the time of burning and in post-fire succession. Water stress on shallow soils is another factor that will increase in importance under a warmer climate. Differences in response between the northern and the southern margins of the boreal forests are likely.
Attempts have been made to calculate the overall C budgets for boreal forest sites. The calculations consider both fire and human impacts. Studies from Canada and Europe suggest that boreal forests may be a substantial sink of C (5 250 Tg C yr-1). The identification of evidence for enhanced boreal tree growth in response to increasing carbon dioxide concentrations since the mid-nineteenth century deserves further attention. This is likely to appear as a positive residual when growth patterns associated with climate change, stand dynamics, etc., are taken out. However, the correlation between rising atmospheric temperatures and carbon dioxide concentrations means that the statistical separation of these two effects will be a difficult task.
Modelling of C budgets in boreal forests ranges from basic ecophysiological models, through whole-tree and ecosystem models that include nutrient availability and uptake components. These models need further verification against data from a wider range of boreal forest ecosystems. There is also a great need to couple physiological and stand dynamic models to better simulate forest dynamics at scales larger than the individual stand.
The direct effect of elevated atmospheric carbon dioxide concentrations on sapling and seedling growth has been established for several boreal forest tree species, but because it is so difficult and costly to perform environmentally relevant experiments on mature trees sapling and seedling experiments have a limited applicability to field conditions. Given the costs of manipulating whole forests, evaluation of the carbon dioxide effect of the boreal forest may have to be based on modelling and dendrological records searching for enhanced tree growth beyond that derived from changes in climate.
1.3.2 Temperate coniferous forests
The area of temperate coniferous and mixed forests is about 2.5 x 106 km2. Although the data on aboveground biomass and aboveground net primary productivity (NPP) for temperate coniferous forests are numerous, there are few complete C budgets for these ecosystems. This is largely due to the lack of information on fine root production and respiration. Estimates of fine root production range from 30 to 710 g Cm-2 yr-1. Aboveground litter inputs and their decay are among the most well studied aspects of the C budgets for temperate coniferous forests. Both climate and initial litter quality are important controls of litter decomposition. Useful predictive models of aboveground litter decay have been developed based on these controls.
Temperate coniferous forests are currently thought to be net C sinks. Many of these forests are accumulating C as they recover from disturbance. In addition, C accumulation may be stimulated by the high N deposition from the atmosphere that many of these forests are subject to, and the increase in atmospheric carbon dioxide concentration.
Nitrogen is considered to be the most common limiting nutrient in temperate coniferous forests. Three major mechanisms by which enhanced N availability increases aboveground net production are: (1) increased net photosynthesis; (2) increased leaf area index (LAI); and (3) reallocation of C from fine roots and mycorrhizae to aboveground tissues.
1.3.3 Tropical savannas, woodlands and grasslands
The contribution of non forested tropical ecosystems to the global C cycle has received scant attention in the past. Tropical grasslands, savannas and drought-deciduous woodlands occupy at least 11% of the global land surface. If the definition of tropical grass-containing communities is broadened to include hot semiarid shrublands and desert grassland, the cover increases to over 20% of the global land surface. These types form a vegetation continuum with varying proportions of grasses and woody plants, which can only be separated into discrete types by applying arbitrary limits. Savannas, which in most classifications form the largest single class on this continuum, are a tropical vegetation type in which the primary production is approximately equally shared by trees and grasses. They occur in regions with a hot, wet season of 3-9 months' duration, with a warm dry period of little plant growth for the rest of the year. Fires occur every 1 to 10 years during this dry season.
Net annual C fixation in tropical tree-grass systems is about 7.6 x 1015 gC yr-1, with a possible range between 3.2 and 10.8 x 1015 g C yr-1. This is about half of the net annual C fixation attributed to tropical forests. The main controls on C fixation are water availability, nutrient availability and vegetation composition and structure. Between 0.9 and 1.7 x io~~ g C of this fixed C is returned to the atmosphere annually as the products of combustion (over 95% as carbon dioxide). Assuming steady-state conditions the rest returns via the respiration of decomposer organisms and herbivores. However, with our present state of knowledge globally it is unclear whether the C stock in savanna vegetation and soil is increasing or decreasing, since opposing processes predominate in different areas of the world.
The total C stock in tropical grasslands, savannas and woodlands is about 135 x 1015 g C, 80% in the soil and 20% in woody plants. The relatively low average soil C density (6.7 kg Cm-2) is due to the sandy nature of many savanna soils, the high soil temperatures, alternating wet-and-dry cycles and frequent fires. The tropical savannas, grasslands and woodlands have a large potential for either sequestration or release of C, depending on future land management practices, climate and atmospheric composition.
There is no direct information on the impact of rising carbon dioxide concentrations on the primary production, decomposition and herbivory in tropical grasslands and savannas. There is a need for several secure, well-documented long-term ecological research sites in tropical grasslands and savanna regions. Information about the extent, frequency and behavior of vegetation fires in this particular fire-prone system is presently inadequate, but will improve substantially within the next few years as research currently in progress is published. The dynamics and mechanisms of the interaction between woody plants and grasses in mixed tree-grass systems are being investigated in an ongoing SCOPE project. Perhaps the largest uncertainty in understanding and predicting the C budgets for the tropical woodlands, savannas and grasslands is the extent, degree and nature of land use changes, along with the human processes which drive it.
1.3.4 Temperate grasslands
Temperate grasslands are located mainly at mid-latitudes of Asia and North and South America where large changes in climate and land use have occurred or are expected to occur in the near future. They occupy approximately 8.3 x 106 km2. One of the most important characteristics of these ecosystems is that the major C store is located below ground. The size of the below ground C pool is related to temperature, whereas the size of the aboveground C pool is related to annual precipitation.
Primary production is the major C input to grassland ecosystems and its rate is controlled by annual precipitation with below ground production equalling or exceeding the aboveground production, but estimates of its magnitude are uncertain. However, the slope of the production/precipitation relationship is steeper across space than across time. This suggests that the short-term response of these ecosystems to an increase in precipitation is constrained by its structure. Changes in structure are mediated by processes which range from leaf expansion, and recruitment to migration, and which have characteristic response times. Decomposition is the major C output in these grasslands and it is primarily controlled by temperature although soil texture and litter quality play some role. Regional analysis suggests that soil organic C accumulation is more controlled by output (decomposition) than by input (primary production).
The three forces driving global change are related to changes in land use, atmospheric composition, and climate. A large fraction of the temperate grasslands has already been converted into croplands and the loss of C resulting from cultivation appeared to be much larger than the loss resulting from 50 years of the expected climate change. The direct response of grasslands to increases in carbon dioxide concentration is controlled by water and nutrient availability with responses being largest in nutrient-rich, and water-limited Systems. Finally, climate change will result in an increase in temperature and an increase or decrease in precipitation depending upon the region selected and the global circulation model consulted. The temperature increase will result in an increase of the decomposition rate. An increase in precipitation will result in an increase of production which may or may not offset the increase in the decomposition rate. An increase in temperature which coincides with a decrease in precipitation will certainly result in a severe C loss.
Concern over losses of soil organic matter and net C release to the atmosphere has resulted in new research on cultivation practices over the past several decades. One of the objectives has been the conservation of soil organic matter which has resulted in crop production methods that either minimize tillage or eliminate it altogether. Preliminary results suggest that these new methods have the potential to conserve or, in some cases, to increase soil C storage. A recent trend in many developed countries of converting cropland back to the original ecosystem will create a new and potentially large category of recovering or successional ecosystems. In the temperate grassland regions in the Northern Hemisphere, such conversions will have important effects on regional C budgets.
1.4 ECOSYSTEM PHYSIOLOGICAL RESPONSES
1.4.1 Photosynthesis in atmospheric and climate change
Photosynthesis provides the driving step to the biogeochemical global C cycle. While much is known about the direct effects of temperature, water vapor pressure deficit and nutrient supply on this process, until recently little has been known about the effects of elevated carbon dioxide concentration over the long term. The instantaneous effects of atmospheric carbon dioxide concentration on photosynthesis are well understood with the mechanisms being highly conserved across terrestrial vegetation; this has provided a sound platform for the development of mechanistic models that are valid at scales from the cell to the landscape. Exploration of the mechanisms shows that it is unreasonable to describe the effects of rising carbon dioxide on photosynthetic C-uptake by use of a simple multiplier, such as the fl-factor. Interactions with temperature, N and water can all markedly modify the degree and even the direction of change. This is further complicated by the effects of growth at elevated atmospheric carbon dioxide concentration on the photosynthetic apparatus. Although most of the 'down-regulation' of photosynthesis may be explained as an artifact of the pot size used or the measurements made, there are species differences and clear interactions with nutrient supply. Understanding these species differences and nutrient effects will be critical to the development of sound models for the prediction of the long-term changes of photosynthetic C-uptake by coniferous forests and grasslands.
Theoretical considerations suggest that the relative stimulation of both light-limited and light-saturated C3 photosynthesis resulting from elevated carbon dioxide concentration will increase with temperature. These are consistent with field observations. The conserved nature of the mechanism of C3 photosynthesis and its temperature relations provides a strong basis for the development of canopy and vegetation scale models across biomes. The validity of this approach will depend largely on the significance and understanding of second-order effects determined by nutrient limitations, water availability and acclimation, which may alter light-saturated photosynthesis.
1.4.2 Rising atmospheric carbon dioxide and plant respiration
There is now abundant evidence that dark respiration in foliage and possibly in other plant tissues is reduced by approximately 30% for a doubling of atmospheric carbon dioxide. The primary effect is immediate and reversible and this effect appears to be altered by changes in tissue composition in some species.
In slow growing woody perennial species in which soluble carbohydrates do not accumulate, respiration appears to be inhibited and there does not seem to be the level of engagement of the alternative oxidase that occurs in fast growing species including crops.
Although there is evidence that the inhibitory effect of elevated carbon dioxide on respiration extends to non photosynthetic microbes involved in decomposition, this would only have practical meaning for a high carbon dioxide world if these decomposers were normally exposed to atmospheric levels of carbon dioxide. It is hard to imagine that the high concentrations of carbon dioxide found in soil could be substantially altered by doubling of atmospheric carbon dioxide. Thus, the rates of soil respiration are not likely to be influenced by increasing atmospheric carbon dioxide except as related to changes in the properties of decomposing plant tissues.
If increasing carbon dioxide concentrations reduce the rates of respiration in all plant tissues and in microbes and fungi as well, then it should also be expected that there would be a significant increase in the C accumulating in ecosystems. This would result from the combined effects of stimulation of photosynthesis and reduction in respiration.
While the results of the effect of elevated carbon dioxide on respiration have been found to vary widely, even small effects would have a significant impact on global C budgets. For this reason, additional study of the role of increasing atmospheric carbon dioxide in respiration should be given a high priority for research.
1.4.3. Carbon and nutrient allocation
The term 'allocation' refers to the ways in which plants distribute element and energy resources among the various plant functions including growth, reproduction, maintenance, storage, defense and uptake of additional resources. While the basic patterns of allocation have been described in relation to the environment for many ecosystems, the physiological mechanisms controlling allocation are poorly understood. Nonetheless, a unified theory of allocation is now being developed. It can be represented as a circular set of interactions resulting in positive feedbacks between site quality, plant response, and future site quality. The power of the theory is that it unifies allocation with other ecosystem processes including photosynthesis, foliar longevity, herbivory and element cycling in a single interactive whole.
1.4.4 Elevated carbon dioxide, litter quality and decomposition
Litter from plants grown in elevated carbon dioxide often has a reduced N concentration. Sometimes the litter also has slightly increased concentrations of structural compounds such as lignin, but this is not always the case. Litter from plants grown in elevated carbon dioxide often decomposes more slowly than litter from plants grown at ambient carbon dioxide. This slower decay rate could reduce the rate of nutrient cycling and so feed back negatively to plant productivity. If warming accompanies increases in atmospheric carbon dioxide, the temperature increase will likely increase decay rate. This increase in the rate of decay and nutrient cycling may offset the negative feedback to plant productivity associated with the carbon dioxide-induced decrease in litter quality.
1.4.5 Soils, a source or sink of carbon
Detrital inputs to the soil are estimated at about 60 x 1015 g yr-1 and decomposition at 5060 x 1015 g yr-1. There is no a priori reason to suspect that litter and soil organic matter are in a steady-state condition at present, given the large changes in land use patterns that continue to occur at a global scale and the continuous input of C to the global cycle from fossil fuel burning. One study of soil C accumulation rate suggested that, on a worldwide basis, approximately 0.4 x 1015 g C yr-1, or less than 1% of primary production, escapes decomposition and is added to a soil C store of 1500 x 1015 g. This conclusion is, however, still open for question and more research is needed.
Given the uncertainties associated with C balances on a global scale, let alone on a plot scale, an input/output imbalance of 5-10% could easily escape detection, and imbalances of this magnitude would make soils either a sizeable source or sink for C. With respect to the future, it does not seem possible at this time to predict the balance between the forces that increase soil organic matter accumulation under climate change and those that decrease accumulation.
1.5 MODELLING
1.5.1 Grassland models
The SCOPE Grasslands Modelling Group developed computer models of grassland productivity and nutrient cycling. Rather than attempting to review all existing models, the group decided to concentrate on further validation and development of the CENTURY ecosystem model, a model of terrestrial biogeochemistry based on the effect of climate, human management (fire, grazing), and soil properties on plant productivity and decomposition. The grassland version of CENTURY was tested using observed data from 11 temperate and tropical grasslands around the world. In order to better simulate climatic responses of photosynthetic productivity in a range of grasslands, the CENTURY model was also compared with GRASS, a more ecophysiologically based model of grass productivity which has been applied to both tropical and temperate grasslands. The objective of this effort was twofold: (a) to establish a method to compare ecosystem dynamics across a broad range of environmental factors which influence grassland dynamics and, (b) to evaluate the sensitivity of different grassland communities to changes in climatic patterns as projected by various climate change scenarios.
With respect to the first objective, the results show that CENTURY can simulate soil C and N levels to within 25% of the observed values for a diverse set of soils. Peak live biomass and annual plant production can be simulated within + 25% of the observed values for natural, as well as, burned, fertilized and irrigated grassland sites where annual precipitation ranged from 220 mm to over 1500 mm. Seasonal live biomass can be generally predicted to within +50% of the observed values. The model underestimated the seasonal live biomass in extremely high plant production years at two of the Russian sites. A comparison of CENTURY model results with statistical models showed that the CENTURY model had higher r~ values than the statistical models. A comparison of CENTURY with GRASS indicated that the two models behaved in a similar fashion and that no significant difference was determined between the two models in this prediction of green live biomass, annual net productivity and peak live biomass. Data and calibrated model results from this study are useful for analysis and description of grassland C dynamics across different soil, N inputs, and climatic gradients, and will be useful for validating other grassland models, including physiologically-based models.
Climate and carbon dioxide perturbations were modelled with CENTURY for 31 grassland sites, both temperate and tropical, representing 7 ecoregions of the world. The model was validated against actual field data for 16 of these sites. The GRASS model was tested under current climate conditions for a site in Colorado, USA and in Kenya. The projected climate change scenarios of the two general circulation models (GCMs) showed similar temperature changes, but contrasted in their predicted changes in precipitation for certain of these ecoregions, notably the temperate steppes and the humid savannas.
Temperature and precipitation change alone resulted in an increase in total above- and belowground production for the mesic regions (humid temperate and Mediterranean), mainly attributable to increased N mineralization, and decreased plant production in the cold desert steppe regions. Soil organic matter decreased in all the mesic and colder regions, due to increased decomposition. In line with most GCM predictions, the tropical savanna regions were affected the least. Climate-driven redistribution of grassland regions was not modelled, but the world area of grasslands is expected to increase in the future or at least remain constant.
Carbon dioxide increase alone resulted in a significant increase in production in all regions, with the greatest proportional increase in soil organic matter in tropical savanna regions. When combined with predicted climate change, carbon dioxide had an additive effect, tending to ameliorate climate change effects. The net effect of climate change and carbon dioxide was a significant increase in NPP in mesic regions (attributable to N mineralization) as well as in dry savannas, with little or no net change in cold desert steppe or humid tropical regions. Overall, soil organic matter showed a decrease, especially in temperate steppes and cold desert steppes due to stimulation of decomposition by both climate change and carbon dioxide, but tropical savanna and humid savanna regions were actually soil C sinks, regardless of GCM scenario.
Climate change alone predicts a C loss of 3-4 x io~~ g after 50 years of climate change. However, the carbon dioxide enhancement effect, amounting to 2 x 1015 g over the same time, results in a smaller net loss of 1-2 x 1015 g over 50 years. These numbers are substantially lower than previous estimates of Jenkinson and co-workers (1991).
1.5.2 Forest models
The SCOPE Forest Modelling Group compared seven different site-specific forest models under current climate and under future climate scenarios. The models used were BIOMASS, BIOME-BGC, CENTURY (forest version), HYBRID, MBL-GEM, PnET-CN, and Q. These models represent the state of the art for understanding the influence of climate on element cycling, C storage, and the interactions between plant and soil processes in temperate conifer forests. The models differ in time scale, linkages between C, N, and water, representation of heterogeneity, detail of photosynthesis, allocation, and decomposition. The models have been applied to radiata pine stands in Australia and Scots pine stands in Sweden both for validation against existing stand records and for predictions under future climate Scenarios. The ability of the models to reproduce to current conditions depends to a large extent on how constrained they are by the initial conditions.
For current and future climates there were major differences among model predictions of important storages and fluxes. Because model predictions for altered climates differed, large uncertainties still remain about the response of temperate conifer ecosystems to changes in air temperature and atmospheric carbon dioxide. Model predictions differed for: (1) the amount of C stored over a stand's lifetime; (2) the absolute response of productivity and C to increased temperature and carbon dioxide; and (3) the magnitude of the response relative to the same model's predictions under current climate. Modelling growth and C storage for the expected life of a stand using a mechanistic approach remains problematic. Therefore, our current knowledge, as incorporated into these mechanistic models, is not sufficient to provide an unambiguous answer to the question of how C storage in specific forest stands will respond to altered climate. Models predicted that growing forests are strong sinks for C. Additionally, most models did predict that productivity would increase for growing forests under elevated carbon dioxide and all models estimated that C storage would increase. Model predictions also showed that changes in fluxes and C storage that normally occur over a forest's lifetime are far greater than those expected under elevated carbon dioxide or increased temperature.
Four critical areas for future research were identified. First, the model predictions differed for the increased temperature scenario, because the models made different assumptions about how temperature affects productivity. In some models the major controls were on gas exchange and in some models on nutrient mineralization. Second, because the models made different assumptions about the long-term controls over productivity, model predictions for stand development disagreed. These factors are in general poorly understood. Third, the biogeochemical feedbacks constraining plant response to elevated carbon dioxide are difficult to handle. Finally, there is a critical need for flux measurements made at the ecosystem level, and an understanding of the acclimation processes.
1.5.3 Global-scale biogeochemistry modelling
The terrestrial ecosystem model (TEM) was used to estimate NPP, vegetation C, and soil C for the potential global distribution of grasslands and conifer forests under conditions of contemporary carbon dioxide and climate. For grasslands, the TEM estimates an area average NPP of 267 g C m~2 yr ~, vegetation C of 413 g Cm -2, and soil C to 1 m of 9160 g Cm -2 The estimates for conifer forests
~ 800 g Cm - 2 TEM was also used to investigate potential responses of NPP and C pools to doubled carbon dioxide and associated climate changes predicted by GCMs. For high- and low-temperature scenarios applied to grasslands, TFM predicts NPP increases of 27 and 18%, vegetation C increases of 36 and 27%, and soil C increases of 4~0 and 6.8%, respectively. For high- and low-temperature scenarios applied to conifer forests, TEM predicts NPP increases of 33 and 27% and vegetation C increases of 34 and 23%, respectively. Soil C in conifer forests decreased by 0.1% for the high-temperature scenario, but increased 2.9% for the low-temperature scenario.
The absolute responses of soil C in grasslands and conifer forests are less when TEM is calibrated for soil C to 20cm depth rather than I m; the response of NPP, vegetation C, and the proportional response of soil C are insensitive to the calibration depth. These results demonstrate that it is important to identify at the calibration site the actively decomposing soil C that is appropriate to the time frame of interest. The responses of TEM for climate change are generally consistent with the results of other models that have examined potential responses of C cycling to climate change, but there is little information available for comparison to changes in both carbon dioxide and climate.
1.5.4 Global-scale biogeography modelling
Assessments of C cycling in coniferous forests and grasslands should be based on a description of land cover with realistic geography and adequate resolution. Such description can be derived from a global geographic data base of existing land cover types. Some global data bases for land cover are described in terms of their underlying assumptions and classification principles. These data bases differ from each other for many regions. We conclude that there is an urgent need for a new assessment of current global land cover for future assessments of global change on ecosystems.
Global ecosystem and land cover distribution is largely driven by climate and human land use. The climatic component is represented by a variety of equilibrium models or climate classifications, differing from each other in terms of inclusion of biophysical processes. Classifications based on annual means of climatic parameters generally have a poor fit. The performance improves considerably if seasonality and an adequate soil moisture balance are taken into account. Recent plant functional type models represent the most realistic and robust patterns of climate-driven ecosystem distribution, and they are therefore valuable tools to determine current and future ecosystem distributions for global change studies. Different applications of global vegetation models are reviewed and discussed.
Global land cover bases and bioclimatic classifications or models have different shortcomings. Nevertheless, a pairwise comparison between data bases from either of these groups may lead to insights into the limitations of either the data base or the model. If so, we can start to derive some general principles which should be included in a model used for global change studies. The discrepancy between different data sets can perhaps be explained by the large differences in the actual classifications. The conclusion is that we have only very approximate comprehension of actual global land cover. This conclusion is further strengthened by the fact that even in the statistical, tabular-based country censuses such as FAO-agricultural and forestry yearbooks, large differences occur between the data.
1.6 FUTURE RESEARCH
This SCOPE project has pointed to several areas of future research.
1.6.1 Reduction of uncertainties in data bases
For all biomes investigated it is recognized that there is a lack of reliable data bases of the extent and rate of production. In some cases well-documented studies of a few sites exist but their representativeness can be questioned. Long-term studies are required of representative sites where trends in behavior can be detected. Short-term flux studies of carbon dioxide and water for tests of our mechanistic interpretations of system properties are also required.
Those mechanisms which are not well understood at present are respiration and allocation. Belowground storage and processes which result in C sequestration and losses require special attention because of the increasing awareness of their importance and their sensitivity to external manipulations.
1.6.2 Reduction of uncertainties regarding key processes
The SCOPE study shows that key areas of divergence between the forest models arise from the formulation of the effects of carbon dioxide on water and nutrient use, on allocation of NPP to different plant components (explicitly or implicitly), and on long-term coupling of C and N storage. The models also differ in the degree to which canopy processes are coupled to the atmosphere and the feedbacks which result from this coupling. It is crucial that modellers and experimentalists communicate new results suggesting mechanisms not included in extant models.
1.6.3 Modularization of models
From the model comparison activity, we concluded that converting all of the models to a modular structure will facilitate model comparisons. A modular structure would make it possible to exchange modules describing key ecological processes such as photosynthesis among models. By doing this we could pinpoint the conceptual differences among models and we could explore the system-level consequences of these differences.
1.6.4 Validation of models
One way to begin to narrow the range of possible responses and to determine which models are most accurate is to collect appropriate measurements from natural ecosystems that can confirm or validate the projections of the models. Certain measurements can and should be made in the near future. In particular, these include flux measurements of C and water exchange over large areas of major ecosystem types. Measurements over the seasonal cycle and, periodically, over several years to determine fluxes during years with different temperatures and precipitation amounts are essential. These data can be used to determine whether measured evapotranspiration and net C exchange are consistent with values predicted by particular models. It is also important to recognize that the ability of a model to simulate current ecological conditions does not validate its ability to accurately simulate responses to future climate change. Model sensitivity to different climates may be examined by simulating ecological responses to Holocene climates or more distant past warm periods for which paleoecological data exist.
1.6.5 Development of models of transient ecological responses
From both a scientific and policy perspective it is critical that we develop models that incorporate transient dynamics and make real-time predictions about the patterns of ecological change. This is not a trivial task, in part because of the numerous aspects of ecological response (e.g. vegetation dieback, migration, succession, soil development) that must be incorporated and because the time constants for different responses can vary widely. Several modelling groups are working on transient models, and we simply want to convey to others the critical importance of this work.
1.7 CONCLUSION
At the start of this SCOPE project there was some doubt as to whether we could usefully catalyze cooperation between empiricists and modellers in a way which would progress research in the field of climate change. Fortunately the research was ahead of its time and initiated research which has contributed significantly to the role of coniferous forests and grasslands in global change. During the project continuous inputs were made by the participants into numerous IGBP and IPCC studies which provided a more solid scientific basis to these studies than would otherwise have been the case.
The review of the forest and grassland ecosystem models provides a unique comparison of models and sites around the world. While the outcome was not as perfect as we initially wished, the exercise has shown the limitations of model comparisons and the problems of using a single model to compare imperfect data from many different sites. Nevertheless, the conclusions with regard to terrestrial C sources and sinks and the role which various physiological processes in plants and soils play in such C fluxes are of considerable importance in predicting global change effects.
As a result of the research undertaken in the project we feel confident in highlighting those areas of research which should receive priority in the future. There is little doubt that such future research should enhance active collaboration between experimentalists and modellers so as to reduce the distinction between such disciplines. This is the way forward to help solve global change research problems.
© Unless indicated otherwise, all material on this website is under SCOPE's copyright
Last updated: 26.03.01