Ecophysiology, Climate Change,
Thomas J. Goreau
Pernetta, J.C.; Leemans, R.; Elder, D. & Humphrey, S. (Eds). 1995.
Impacts of Climate Change on Ecosystems and Species: Environmental Context.
IUCN, Gland, Switzerland. viii + 98 pp; 4 color plates
NOTE: This paper examines the role of all sources and sinks of carbon dioxide which affect its dynamical cycling and lifetime in the atmosphere, by developing a theory for the carbon dioxide residence time spectrum to calculate the effect of changes in each on the concentration of the gas in the atmosphere. Using data on the metabolism of forests, a model for the impact of reforestation in both the tropics and colder areas is developed, which shows that both areas need to be reforested, but for different reasons: colder areas because they are more efficient sinks for atmospheric carbon, and tropical areas because the reduce the lifetime, and hence the warming effect, of carbon dioxide. This paper was published in:
Tropical species and ecosystems should be extremely sensitive to global warming because the strong temperature dependence of physiological rates places the maximum tolerable temperature of many tropical species just above their optimum temperatures. An eco-physiological model of photosynthesis, respiration, and net primary production is coupled to measurements of forest productivity along polar to equatorial gradients in order to evaluate potential changes in the efficiency of carbon cycling and storage by forest ecosystems in different climates. The model indicates that forests could become less efficient global carbon absorbers under global warming, increasing rates of plant respiration more rapidly than photosynthesis and producing more short-lived foliage carbon rather than long-lived woody biomass. Tropical ecosystems are found to strongly dominate global biogeochemical carbon fluxes, with important implications for global biosphere-climate feedbacks as well as time scales of atmospheric CO2 change and total greenhouse warming. Many corals live near their upper temperature limits, and are increasingly under thermal stress. Ocean warming could trigger further mass coral bleaching, threatening the viability of coral reef ecosystems, reducing marine carbon burial, and limiting tolerable rates of climate change. The residence time spectrum of atmospheric CO2 is derived with regard to all sources and sinks and used to evaluate response rates to major policy options for stabilizing climate change, including reduced fossil fuel emissions and reforestation. Reforestation provides the fastest atmospheric response. Ecophysiological gradients suggest a dual strategy: reforestation of cold areas for efficient carbon storage and reforestation of hot areas to reduce atmospheric CO2 lifetimes.
Introduction: carbon dioxide and climate change
Global warming is widely expected in coming decades to centuries as carbon dioxide and other greenhouse gases increase in the atmosphere (Houghton et al., 1990). There is no scientific doubt about the reality of the greenhouse effect: increased temperature of gases absorbing radiation is a fundamental law of physics, not a hypothesis waiting to be tested. Uncertainty surrounds predicted rates of global warming and potential ecological impacts. Mean global temperatures have risen by 0.7°C in past century (Hansen & Lebedeff, 1987), and tree-ring records from the Southern Hemisphere suggest the past 25 years have been the warmest in over a thousand years (Cook et al., 1991). Supporting evidence for global warming comes from long time temperature trends in ocean basins and permafrost, retreat of montane glaciers, thinning of Arctic ice, and poleward migration of mobile bird and insect species. Studies of ecosystem impacts of global warming have been largely concerned with potential rates of poleward migration of ecosystems in response to shifting cimate zones, as potential estimates of maximum tolerable rates of warming (Overpeck et al., 1991; Prentice et al., 1992). This paper outlines some physiological impacts of high temperatures on tropical species and ecosystems with potentially negative consequences not previously evaluated in global climate change scenarios.
Standard scenarios are derived from computer simulations using atmospheric general circulation models (GCMs). GCMs describe the transport of heat in the atmosphere, with simple models of some physical feedback mechanisms which could amplify or reduce the climate response. These models predict doubling CO2 to 700 ppm will increase mean global temperature by around 2-4°C and sea level by around 0.3-0.5m (Houghton et al., 1990). Predicted changes, widely used to evaluate costs and benefits of possible societal responses, evaluate only part of the transient response to changes in greenhouse gas forcing, and could underestimate full steady state responses.
Given sufficient time, the full temperature and sea level responses to changes in the earth's radiation budget could be considerably larger than expected. The paleoclimatic record suggests long-term sensitivity of temperature to greenhouse gas concentrations has been several times greater in the past than currently predicted in future GCM scenarios. Antarctic ice cores and fossil coral reefs show temperature and sea level sensitivity to greenhouse gas concentrations about an order of magnitude greater than GCMs predict (Goreau, 1990a), implying either that completely different natural climate regulating processes will operate in the future than over the past 160,000 years, or that current models do not include all major positive biogeochemical feedback mechanisms operating in the climate-biosphere-ocean-ice system. A large, but unknown, positive feedback could operate in the climate system to explain orbital forcing of climate change. A possible source of discrepancy between observations and theory lies in failure of most GCMs to include possible effects of climate change on biological sources and sinks of greenhouse gases (Lashof, 1989; Houghton & Woodwell, 1989; Goreau, 1990a).
During the last interglacial period, 130,000 years ago, temperatures were around 1 to 2°C warmer than today (Jouzel et al., 1987), sea levels were about 5-8 m higher, and crocodiles and hippopotamuses flourished in southern England. As CO2 concentration was then 27% lower than today, conditions then should underestimate changes that should ultimately result from current CO2 levels, and be much less than those resulting from greenhouse gas doubling (Goreau, 1990a). Full climatic responses should take centuries to millenia since they depend on rates of deep sea warming and oceanic turnover times. Societal responses could continue much longer than CO2 doubling times, greatly increasing ultimate costs of adapting to climate change beyond those of standard short-term transient response scenarios.
A major focus of concern about global warming has been boreal and temperate ecosystems, where models predict greatest warming. GCM scenarios generally predict warming rates around three times higher in boreal zones than at the equator. Consistency among predictions using different GCMs is not a guarantee of accuracy as most models consider only changes in CO2. Models with CO2 alone predict most warming in polar regions, but models which include nitrous oxide, methane, and halocarbons warm the tropical atmosphere more (Wang et al., 1991) since these gases are more effective molecular heat absorbers, have lower height distributions (Lelieveld & Crutzen, 1992), and spectroscopic absorption lines are more unsaturated.
Tropical ecosystems have been assumed to be well adapted to higher temperatures and less sensitive to increases than boreal ones. If tropical terrestrial and marine ecosystems are extremely temperature sensitive, more stringent limits on climate change may be needed than are now recognized. This paper reviews aspects of tropical ecophysiology which point strongly in that direction.
Tropical terrestrial ecosystems: respiration, temperature, and net productivity
Many evaluations of ecological impacts of climate change assume increased CO2 acts as a plant fertiliser, increasing rates of CO2 removal from the atmosphere (Strain & Cure, 1985). This requires productivity to be carbon limited, which can occur in brightly lit, well watered, highly fertilised greenhouses, but which is exceedingly rare in nature. Most ecosystems are unable to achieve optimal productivity due to inadequate availability of water, nutrients, trace elements, light, or temperature over part or all of the growing season. Recent experimental studies indicate the CO2 "fertilisation effect" may be a myth (Bazzaz & Fajer, 1992). Several lines of argument suggest instead that global warming could stimulate positive biological feedbacks:
1) Tropical soils generally have low organic carbon compared to colder zones because of greatly accelerated breakdown of soil organic matter by bacteria and fungi at high temperatures (Tinker & Ineson, 1990). Global warming could cause large amounts of peat in arctic bogs and boreal soils to be oxidised to CO2 (Houghton & Woodwell, 1989; Jenkinson et al., 1991).
2) Many models assume that species will be immediately replaced by neighbours from warmer zones, but the existing flora will respond immediately to climatic change while migration could take centuries to millenia even in the absence of deforestation and human barriers. Plant communities could be affected by increased evapotranspiration and reduced soil moisture before more drought-adapted species can replace them (Neilson et al., 1992), resulting in CO2 buildup in the atmosphere (Neilson & King, in press).
3) High temperatures increase respiration more rapidly than photosynthesis (Larcher, 1980). Many tropical plants and algae live near their optimal temperatures, above which net production decreases due to increased respiration (Stoskopf, 1981), making them increasingly inefficient at retaining carbon.
4) It is often assumed that plant water use efficiency should increase when CO2 concentration rises (Strain & Cure, 1985), but transpiration rates are greatly elevated with rising temperature, until plants reach the wilting point. Plant transpiration. release of water to the atmosphere by leaves opening stomata to take up CO2, puts many molecules of water into the atmosphere for each CO2 molecule admitted. Water is the major global greenhouse gas, so higher transpiration acts to amplify global warming (Henderson-Sellers, 1991). Models which predict transpiration of water vapour from net productivity, rather than gross productivity, could underestimate this major natural climatic feedback mechanism.
Many tropical plants are very close to their optimum temperatures. High respiration rates under hot conditions are observed to greatly reduce agricultural productivity in the tropics (Stoskopf, 1981). The equatorial Brazilian state of Amazonas needs to import its major staple food crop, cassava, from subtropical Sao Paulo, as much higher yields are obtained in the colder south although the plant is an Amazon native. Thus is partly due to better soils and fertilizers, but is also due to very high respiration losses during warm humid Amazonian nights, when plants consume most of the carbon fixed in the day, while plants in Sao Paulo have greatly reduced respiration losses during the cool nights. Global tropicalisation could reduce efficiency of biosphere carbon storage if relatively more photosynthetically fixed carbon is respired.
Jordan (1983) has summarized latitudinal measurements of net forest productivity. His data splits net production into woody biomass and leaf and litter production, grouped according to solar radiation input. These data (Figure 1) show that boreal forests are better woody biomass producers than equatorial forests, but hot forests produce much more leaf and litter biomass. Boreal forests grow slowly, but turn most carbon production into wood which is stored for a long time in trees and soil. Tropical forests grow very rapidly, but primarily produce leaves, fruits, flowers, and twigs which are rapidly eaten by insects, decomposed by fungi and bacteria, with carbon returned to the atmosphere. Biosphere carbon retention ranges from as low as days to months (for rapidly growing tropical leaves, fruits, and flowers) to centuries or millennia (for old growth boreal coniferous forests or peat bogs) so total carbon storage is considerably greater in cold than warm zones (Adams et al., 1990).
Considering only net primary production in global models seriously underestimates carbon flow between atmosphere and biosphere. Net productivity is readily measured by directly harvesting and weighing biomass to determine changes over time. Determining the amount of respiration is much more difficult, requiring careful measurements of CO2 uptake and release to the atmosphere, which have mainly been made on plants in temperate greenhouses. Actual rates of respiration in nature are very poorly known, especially in the tropics. Physiological data suggests that respiration losses amount to around 25% in temperate plants but rise to around 80% in equatorial zones (Larcher, 1980). If we assume that respiration losses increase roughly linearly from boreal to equatorial ecosystems, the net productivity values of Jordan can be used to estimate forest plant respiration and gross productivity. Calculated values are shown in figure 2. Carbon cycling between atmosphere and biosphere greatly increases towards the equator because of increased respiration. Ecosystem carbon retention efficiency, which is defined as:
E = N/P = 1- (R/P) (1)
(where N is net production, P is gross photosynthesis, and R is respiration) declines as the respired fraction increases.
Figure 2 indicates that tropical ecosystems strongly dominate global carbon recycling. Their high carbon cycling rates lower the lifetime of added CO2 molecules in the atmosphere, which reduces total atmospheric warming. Under global warming the optimal CO2 stabilization strategy, along with fossil fuel emission reduction, would increase global gross productivity to reduce the CO2 lifetime, and increase global net productivity to increase carbon storage in biomass and soils. Human activities currently do the opposite: deforestation and soil degradation have already reduced global biomass and productivity significantly.
Figure 1. Average net forest productivity of long-lived and short lived biomass based on measurements summarized in Jordan (1983). The solar energy scale spans boreal to equatorial ecosystems. Figure 2. Gross photosynthesis and plant respiration calculated from total net production from figure 1, assuming that respiration to photosynthesis ratios increase approximately linearly with solar radiation inputs, as shown based on physiological data from Larcher (1980). If R/P increases non-linearly, as might be expected, then gross photosynthesis and respiration could increase even more rapidly towards the tropics.
Tropical marine ecosystems: temperature and coral reef bleaching
Recent observations suggest that many coral reefs are undergoing significant temperature stress, and may be near their upper temperature limits (Glynn, 1991; Hayes & Goreau, 1991). In the 1980s mass regional coral reef bleaching began to be reported in the Pacific, Atlantic, and Indian Oceans (Williams & Bunkley-Williams, 1990). Bleaching results from stress-induced expulsion of symbiotic unicellular algal populations which provide corals with most of their food, oxygen, waste removal, and ability to deposit skeleton (Goreau et al., 1979). Coral color is almost entirely due to the algae, so bleached coral tissues turn transparent, exposing underlying white limestone skeleton. Bleached corals are starving and practically cease growth and reproduction for up to ten months each time (Porter et al., 1989; Goreau & Macfarlane, 1990; Szmant & Gassman, 1990). Corals over much of the Caribbean were bleached in four of the five years between 1987 and 1991, with increasing mortality.
A large number of stresses, including extremely high or low temperatures, high or low salinities, high or low light levels, and excessive sedimentation can induce bleaching under laboratory conditions. Of all factors capable of causing bleaching, only excessively high sea water temperatures is consistently reported to correlate with mass bleaching episodes (Williams & Bunkley Williams, 1 990; Goreau et al., 1993; Goreau & Hayes, 1994). The only other environmental factor known to have caused bleaching is high freshwater and sedimentation stress. This affects reefs near river mouths during extreme hurricane flood conditions (Goreau, 1964) and cannot account for bleaching under the warm, dry, cloudless weather conditions which precede most regional bleaching events. Fresh water bleaching is localized in shallow reefs near river mouths which recover in weeks, but regional bleaching events last months, take place in areas completely free from river inputs like small desert islands, and affect corals to their maximum depth.
At all sites where bleaching was reported in the past decade, experienced long term observers noted such events were novel. It is unlikely that all divers and spear fishermen could have failed to recognize past mass bleaching events as they are absolutely distinctive: the reef suddenly turns white and appears to be dead. The longest continuous observational studies of reefs by trained scientific observers go back 40 years in Jamaica and in French Polynesia. Local bleaching and reef deterioration around pollution and freshwater sources were seen at both sites, but mass bleaching was completely absent in the 1950s, 1960s, and 1970s. The distribution of bleaching in time and space is entirely unlike any known anthropogenic stress such as erosion, sewage, construction, over fishing. (Goreau, 1992b). Bleaching is observed to be most extensive at sites remote from stress, with up to 80 or 90% of all corals affected (Goreau & Macfarlane, 1990). In reefs subject to regular stress from freshwater, sediments, and sewage, bleached corals are relatively less abundant, apparently because their corals are more resistant to stress (Goreau, 1992b). Bleaching is most severe in previously un-stressed areas which had separated the locally stressed reefs near population centers and rivers.
Sea surface temperature records from high resolution satellite data in the Caribbean, calibrated against in-situ data, show identical spatio-temporal patterns of coral bleaching intensity and positive sea surface temperature anomalies between 1980 and 1991 (Goreau et al., 1993). At all eight sites whose mean monthly sea surface temperature values were examined, there was a threshold temperature above which bleaching was always observed and below which it never was reported. The apparent threshold temperature is proportional to mean annual temperature, implying local adaptation, yet bleaching occurred in both warmest and coolest reef areas. Positive sea surface temperature anomalies are seen in satellite records to precede all major reported bleaching events in the Pacific and Indian Oceans (Goreau & Hayes, 1994). Caribbean bleaching has occurred during each of the warmest temperature anomalies of the past decade, perhaps coincidentally the hottest decade in the past century. No multi-decade water temperature records are available from coral reef sites, so it is uncertain whether temperature anomalies to date result from local or global warming.
Experiments on corals during cooler times of year show they bleach readily at ambient temperatures experienced during bleaching episodes, and field and laboratory evidence suggests temperatures which induce bleaching have not changed with time. The novel and increasing appearance of large areas of bleached corals in all three oceans, following extended periods of high sea surface temperature, implies that many coral reef communities worldwide are currently under temperature stress. Whether global warming or local weather has triggered currently existing high temperature stress is immaterial for future scenarios of greenhouse warming. So many coral populations are now poised at or near thermal shock during warmest times of the year that they are, ipso facto, endangered by any future warming which may occur regardless of its cause.
Many coral researchers now regard reefs one of the ecosystems at most risk from climate change (Goreau,1990b; Williams & Bunkley-Williams, 1990; Hayes & Goreau, 1991; Glynn, 1991). Nearly half of global carbon burial in sediments takes place in coral reef limestone, but direct effects on atmospheric CO2 are small (Fujita et al., 1991). Coral reefs are vital for tropical marine biodiversity, biological productivity, fisheries, tourism, and shoreline protection. Their sensitivity to temperature places strong constraints on acceptable rates of global warming if the most climate-sensitive, species-rich, and productive ecosystems are to be protected (Goreau & Hayes, 1994). Like the ozone hole, coral bleaching is an unanticipated phenomenon indicating the extent of climate change may be greater than recognized, deleterious impacts may already present, and corrective action may be more urgent than thought.
Tropical fluxes and atmospheric CO2 residence times
Total warming by CO2 is the product of the rate of molecular energy absorption and emission, gas concentration, and lifetime in the atmosphere. The first factor decreases slightly as concentration rises, the second is increasing rapidly, and confusion exists over the residence time of carbon in the atmosphere. Depending on how it is computed, estimates of the residence time range from around three to 120 years (Rodhe, 1990). Low values are obtained by dividing the amount of atmospheric CO2 by rates of gross removal by photosynthesis, but much of this carbon is quickly reoxidised and returned to the atmosphere by respiration or decomposition. The effective lifetime of added carbon in the atmosphere-ocean-biosphere system before removal and burial in sediments is orders of magnitude longer. Because of recycling, no single number can characterise the "lifetime" of atmospheric CO2. Instead there is a spectrum of lifetimes, dependent on all processes which add or subtract it.
This is seen by considering the simple mass balance equation for the amount of carbon in the atmosphere:
dC/dt = jFi (2)
where t is time, C is the amount of CO2 in the atmosphere, and Fi is the flux of CO2 per unit time added to the atmosphere by process i. A negative flux indicates sinks which remove CO2 from the atmosphere.
Each process has a characteristic time scale, or a partial residence time, Tj, where Ti = C/Fi (3)If positive, these represent times it would take each source, acting alone, to double CO2 in the atmosphere, and if negative they represent the time for each sink, acting alone at a constant rate, to consume all atmospheric CO2.
It is useful to consider the total residence time,
Tt = C/ j Fi (4)
Substituting (3) into (4), and simplifying, yields:
1/Tt = j (1/Ti) (5)
So partial residence times are not linearly additive, following rules for addition of parallel resistances. Tt values can be misleading, as it is not the mean lifetime of CO2 molecules in the atmosphere, and since infinite values result when sources precisely balance sinks. Atmospheric carbon is recycled through other carbon pools at much faster rates than indicated by Tt. We can decompose equation 5 into two separate parts, summing sources and sinks separately, to obtain residence times with respect to all sources, and to all sinks, T+ and T- respectively. These are equal and opposite only when there is no net change in C. Their difference is an indicator of net rate of change of CO2, corresponding to Tt. Table 1 lists the various partial and total residence times of atmospheric CO2.
The spectrum of atmospheric CO2 lifetimes for all major sources and sinks in table 1 is shown in figure 3a, with values of the fluxes plotted against their residence times. These points appear to cluster near the axes of the graph, but actually fall on the two segments of a unit rectangular hyperbola, better shown in figure 3b, which magnifies the inner part of the previous figure. One advantage of viewing atmospheric CO2 dynamics in terms of residence time spectra is that uncertainties or changes in the absolute value of these fluxes shift them along the hyperbola. Dynamic time responses of atmospheric CO2 to changes in fluxes also shift along the hyperbolas. Resulting changes in partial and net lifetimes can be calculated from equations 2-5 or read off the graph, providing information about how long different processes take to affect atmospheric CO2. Improved data on fluxes could move many sources or sinks to some degree, but it is likely that figure 3 captures the essence of dynamical CO2 response to major sources and sinks.
The major processes affecting atmospheric carbon dioxide.
These are listed with their fluxes to or from the atmosphere (expressed as per cent of atmospheric CO2 cycled per year), and their partial residence times (in years). T+ is the residence time with respect to all sources, T- is the residence time with respect to all sinks, and Tt is the residence time with respect to the difference between sources and sinks. Global carbon fluxes compiled from Solomon et al., (1985), Goreau (1992a), Gerlach (1991), Schlesinger (1990) & Jenkinson et al. (1991), converted to percentages of atmospheric carbon dioxide changed per year. Only the fossil fuel emission flux is known with great accuracy, other fluxes may be uncertain by 10 to 50%.
Process (symbol, figure 3) Flux (%/Yr) Time (yrs)
Ocean-atmosphere (OA) 12.990 7.7
Plant respiration (R) 7.810 12.8
Decomposition (S) 7.810 12.8
Fossil-fuels (F) 0.714 140.0
Deforestation (D) 0.338 295.8
Volcanism (V) 0.012 8500.0
Terrestrial photosynthesis (P) -15.625 -6.4
Atmosphere-ocean (AO) -12.990 -7.7
Sedimentary burial (B) -0.390 -256.4
Soil formation (H) -0.052 -1925.0
Weathering (W) -0.010 -10000.0
3) Net balance
Sources (T+) 29.674 3.37
Sinks (T-) -29.067 -3.44
Difference (T,) 0.607 165.00
Atmospheric CO2 residence time spectra for major sources and sinks.
Figure 3a shows scales from plus to minus 10,000 years, while figure 3b shows scales from plus to minus 300 years. Fluxes listed in table 1 are shown indicated by a letter symbol with a numeric sign indicating whether this source or sink is currently increasing, decreasing, or uncertain. Figure 3a identifies only sources or sinks with partial residence times greater than 1,000 years, and figure 3b identifies those with shorter response times.
Points close to the vertical axis cause rapid atmospheric CO2 responses and points far away require very large changes and long time lags to significantly affect atmospheric CO2. Figure 3b suggests that only large changes in major biological fluxes or in exchange of CO2 between atmosphere and ocean can appreciably affect CO2 over decadal time scales since all other process act over centuries to millenia. Global imbalance of sources and sinks causes increase of CO2 corresponding to a Tt of over 150 years. Tt is about 50 times larger than the residence times due to all sources or all sinks, so CO2 added to the atmosphere is recycled through biosphere and oceans about 50 times before accumulating in the atmosphere because of current carbon cycle imbalances, and more before sedimentary burial.
Most sources and sinks are strongly affected by human activity. These are indicated in figure 3 by plus or minus signs next to each point. At present most sources are increasing and most sinks are decreasing, so the net balance is positive and increasing, and Tt is getting shorter. Sources and sinks most directly controllable by human activity are fossil fuel combustion, deforestation, and photosynthesis. Stable CO2 concentration can be obtained by decreasing fossil fuels and net deforestation, but they are too small with regard to natural sources to have much impact on T+, whereas T- is most strongly affected by changes in photosynthesis. Replacing net deforestation by net reforestation would have the fastest impacts in moderating current CO2 increase. Although small increases in global photosynthesis have much more rapid effects than large decreases in fossil fuel combustion or deforestation both supply and demand side measures are required for other reasons, such as abatement of other combustion-generated pollution, and since the cheapest CO2 to remove is that which is never emitted. If photosynthesis were increased sufficiently worldwide, Tt could move down and rightward along the positive branch towards the asymptote, becoming infinite when sources and sinks are in exact balance. and then jump to the left hand end of the negative branch if atmospheric CO2 decreases.
Another mechanism which could stabilize CO2 is a large increase in atmosphere-ocean fluxes due to increased solubility of CO2 in the ocean, increased burial of carbon in sediments, or increased turnover of the deep sea. CO2 is more soluble in cold water, so this mechanism absorbed most CO2 during Ice Ages, when atmospheric carbon dioxide is known to have been considerably lower than today (Barnola et al., 1987). Increased marine productivity has only small effects because marine decomposition is far more rapid than on land, so accumulation is very small. Increased sediment burial rates could only change CO2 over many centuries or millennia. Changes in deep ocean circulation could be significant but is largely beyond direct human control. Ocean warming should increase stable thermal stratification and decrease CO2 uptake and transport into the deep ocean. The major source of deep water, the Greenland Sea, has reduced in recent years because surface waters have become too fresh to sink (Schlosser et al., 1991), apparently due ice melting. This could lead to more rapid surface warming as heat transport to the deep sea is slowed.
Implications for global carbon cycle management
Tropical ecophysiology has important implications for efforts to manage global carbon cycles to stabilize atmospheric CO2 and limit global warming. Large scale reforestation is capable of removing atmospheric carbon at a cost which varies from place to place, being least in the tropics because of higher net production and cheaper labor. Costs of carbon removal are estimated from as low as $2-3 per tonne for the tropics to higher values in temperate regions (Myers & Goreau, 1991; Howlett & Sargent, 1991). The results of this paper suggest a dual strategy would be required: tropical reforestation is required because growth is so fast and because of its strong impact on keeping down CO2 lifetimes, but temperate and boreal reforestation is needed because these forests are more efficient at storing their carbon for long periods. Reforestation strategies to mitigate global warming must be global in their extent but latitudinally divergent, aiming at increasing woody and soil biomass in cold forests, and increasing leaf, flower, and fruit production in the tropics. Environmentally sound development strategies should seek to sustainably harvest excess wood production in colder zones, and food, fibre, fuel, and biochemical producing trees in the tropics. Unfortunately, most efforts to harvest tropical forests focus on lumber, the form of carbon they produce least efficiently, rather than short-lived carbon products unexcelled in range, variety, and productivity. A new research paradigm is needed to replace ecosystem "development" strategies which have failed badly in many temperate zones and are disastrous on easily degraded tropical soils (Goreau, 1992c).
One way to evaluate the potential of reforestation as a tool to contain climate change and produce sustainable development is to modify methodology used in studies of agricultural sustainability (FAO, 1984). Soil maps and climatic data were used to evaluate potential productivity of a wide range of crops according to soil fertility and length of growing season resulting from temperature and rainfall constraints. Three cases were considered: rain-fed subsistence agriculture, rain-fed agriculture with high levels of technological inputs such as tractors and fertilizers, and high inputs plus irrigation. Potential productivities compared to food needs of existing populations in each area revealed many tropical regions could not produce enough food for local needs with low technology rain-fed agriculture, and some were sufficiently overpopulated that they could not do so even if resources permitted the most productive technology to be used. Productivity estimates were confined to harvestable portions of major crops.
The same approach could readily be extended to ecosystem productivity to evaluate potentials of each region to absorb carbon and convert it into wood or useful biomass via reforestation in comparison to rates of CO2 production from fossil fuel combustion. This requires wordwide maps of fossil fuel use, soil fertility, land use, soil slopes, temperature, and rainfall, with ecophysiological regressions of net productivity and respiration on critical variables. Information now available in global data bases can be processed pixel by pixel for models of potential changes in ecosystem distribution (Leemans, 1989; Cramer & Leemans, 1993; Prentice et al., 1992). Combining methods could provide a basis for evaluating costs and benefits of biosphere options to mitigate CO2 buildup under a wide variety of scenarios, and perhaps forestall potentially unwelcome impacts of climate change on the ecophysiology of the most productive and species-rich tropical ecosystems and the global biogeochemical cycles and climatic feedbacks they dominate.
Adams, J., H. Faure, L. Faure-Denard, J. McGlade & F. Woodward. 1990. Increases in terrestrial carbon storage from the Last Glacial Maximum to the present. Nature 348: 711-714.
Barnola, J., D. Raynaud, Y. Korotkevich & C. Lorius. 1987. Vostok ice core provides 160,000 year record of atmospheric CO2. Nature 329: 408-414.
Bazzaz, F. & E. Fajer. 1992. Plant life in a CO2-rich world. Scientific American 266: 68-74.
Cook, E., T. Bird, M. Peterson, M. Barbetti, B. Buckley, R. D'Arrigo, R. Francey & P. Tans. 1991. Climatic change in Tasmania inferred from a 1089 year tree ring chronology of Huon Pine. Science 253: 1266-1268.
Cramer, W. & R. Leemans. 1993. Assessing impacts of climate change on vegetation using global vegetation models. Pp 190-217 in: Solomon, A. & Shugart, H. (Eds) Vegetation dynamics modelling and global change. Chapman-Hall, New York.
Food and Agriculture Organisation. 1984. Land, food, people. FAO, Rome, Italy.
Fujita, R, S. Gaffin, M. Oppenheimer, J. Marston & T.J. Goreau.1991. Emissions from coral reefs associated with mass coral bleaching. Proc. Intl. Soc. Reef Studies, p 26. Berkeley, California.
Gerlach, T. 1991. Etna's greenhouse pump. Nature 351: 352-353.
Glynn, P. 1991. Coral reef bleaching in the 1980s and possible connections with global warming. Trends in Evolution and Ecology 6: 175-179.
Goreau, T.F. 1964. Mass expulsion of zooxanthellae from Jamaican reef communities after Hurricane Flora. Science 145: 383-386.
Goreau, T.F., N. Goreau, & T.J. Goreau. 1979. Corals and coral reefs. Scientific American 241: 124-136.
Goreau, T.J. 1990a. Balancing atmospheric carbon dioxide. Ambio 19: 230-236.
Goreau, T.J. 1990b. Coral bleaching in Jamaica. Nature 343: 417.
Goreau, T.J. 1992a. Control of atmospheric carbon dioxide. Global Environmental Change 2:5-11.
Goreau, T.J. 1992b. Bleaching and reef community change in Jamaica: 1951-1991. American Zoologist 32:683-695.
Goreau, T.J. 1992c. Technological options to minimize the loss of biological diversity. Environmentally sound technology for sustainable development: Advanced Technology Assessment System Bulletin 7:67-73; United Nations Centre for Science and Technology for Development.
Goreau, T.J., R. Hayes, J. Clark, D. Basta & C. Robertson. 1993. Elevated sea surface temperatures correlate with Caribbean coral reef bleaching. Pp 225-255 in: Geyer, R. (Ed) A global warming forum: scientific, economic, and legal overview. CRC Press, Boca Raton, Florida.
Goreau, T.J. & R. Hayes. 1994. Coral bleaching and ocean "hot spots". Ambio 23:176-180.
Goreau, T.J. & A. Maclarlane. 1990. Reduced growth rate of Montastrea annularis following the 1987-1988 coral bleaching event. Coral Reefs 8: 211-216.
Hansen, J. & S. Lebedeff. 1987. Global trends of measured surface temperature. Science 92: 13345- 13372.
Hayes, R. & T.J. Goreau. 1991. Tropical coral reef ecosystem as a harbinger of global warming. World Resources Review 3: 306-322.
Henderson-Sellers, A. 1991. Developing an interactive biosphere for global climate models. Vegetatio 91: 149- 166.
Houghton, J., G. Jenkins & J. Ephraums. 1990. Climate change: the IPCC scientific assessment. Cambridge University Press.
Houghton, R. & G. Woodwell. 1989. Global climate change. Scientific American 260: 36-44.
Howlett, D. & C. Sargent. Eds. 1991. Technical workshop to explore options for global forestry management, Bangkok, Proceedings. International Institute for Environment and Development, London, 349pp.
Jenkinson, D., D. Adams & A. Wild. 1991. Model estimates of CO2 emissions from soil in response to global warming. Nature 351: 304-306.
Jordan, C. 1983. Productivity of tropical rain forest ecosystems and the implications for their use as future wood and energy sources. Pp 117-135 in: Golley, F. (Ed) Tropical rain forest ecosystems: str ucture and function. Elsevier, Amsterdam.
Jouzel, J., C. Lorius, J. Petit, C. Genphon, N. Barkov, V. Kotlyakov & B. Petrov. 1987. Vostok ice core: a continuous isotope temperature record over the last climatic cycle (160,000 years). Nature 329: 403-408.
Larcher, W. 1980. Physiological Plant Ecology Springer Verlag, Berlin.
Lashof, D. 1989. The dynamic greenhouse: feedback processes that may influence future concentrations of atmospheric trace gases and climatic change. Climatic Change 14: 213-242.
Leemans, R. 1989. Possible changes in natural vegetation patterns due to a global warming. Pp 105-122 in: A. Hackl (Ed) Der Treibhauseffekt: das problem. Akademie fur Umwelt und Energie, Laxenburg, Austria.
Lelieveld, J. & P. Crutzen. 1992. Indirect chemical effects of methane on climate warming. Nature 355: 339-342.
Myers, N. & T.J. Goreau. 1991. Tropical forests and the greenhouse effect: a management response. Climatic Change 19: 215-225.
Neilson, R. & King, G. In Press. Continental scale biome responses to climatic change. In. McKenzie, D., Hyatt, E. & MacDonald, J. (Eds) Ecological indicators. Elsevier Science Publishers, Amsterdam.
Neilson, R., G. King & G. Koerper. 1992. Toward a rule-based biome model. Landscape Ecology 7:27-43.
Overpeck, J., P. Bartlein & T. Webb. 1991. Potential magnitude of future vegetation change in Eastern North America: comparisons with the past. Science 254: 692-695.
Porter, J., W. Fitt, H. Spero, C. Rogers & M. White. 1989. Bleaching in reef corals: physiological and stable isotopic responses. Proc Nat Acad. Sci. USA 86: 9342-9346.
Prentice, I., W. Cramer, S. Harrison, R. Leemans, R. Monserud & A. Solomon. 1992. A global biome model based on plant physiology and dominance, soil properties and climate. J. Biogeography 19: 117- 134.
Rodhe, H. 1990. A comparison of the contribution of various gases to the greenhouse effect. Science248: 1217-1219.
Schlesinger, W. 1990. Evidence from chronosequence studies for a low carbon storage potential of soils. Nature 348: 232-234.
Schlosser, P., G. Bonisch, M. Rhein & R. Bayer. 1991. Reduction of deepwater formation in the Greenland Sea during the 1980s: evidence from tracer data. Science 251: 1054-1056.
Solomon, A., J. Trabalka, D. Reichle & L. Voorhees. 1985. The global cycle of carbon. Pp 324 in: Trabalka, J. (Ed) Atmospheric carbon dioxide and the global carbon cycle.
Stoskopf, N. 1981. Understanding crop production. Prentice-Hall, New Jersey, 433pp.
Strain, B. & J. Cure. 1985. Direct effects of increasing carbon dioxide on vegetation. U.S. Department of Energy, Washington D.C.
Szmant, A. & N. Gassman. 1990. The effects of prolonged bleaching on the tissue biomass and reproduction of the reef coral Montastrea annularis. Coral Reefs 8: 217-224.
Tinker, P. & P. Ineson. 1990. Soil organic matter and biology in relation to climate change. Pp 71-87 in: Scharpenseel, H., Schomaker, M. & Ayoub, A. (Eds) Soils on a warmer Earth. Elsevier, Amsterdam.
Wang, W.C., M.Dudek, X.Z.Liang,, & J.Kiehl. 1991. Inadequacy of effective CO2 as a proxy in simulating the greenhouse effect of other radiatively active gases. Nature 350:573-577.
Williams, E. & L. Bunkley Williams. 1990. The world wide coral reef bleaching cycle and related sources of coral mortality. Atoll Research Bulletin 335: 1-71.