A review of the possible impacts of climate change on forests in the humid tropics

: Increased anthropogenic emissions of greenhouse gases (GHGs) have led to gradual warming of the earth’s atmosphere. It is predicted that by the year 2100, the average annual global air temperature will increase by 1.1 – 6.4 o C. This has triggered a series of long-term changes in the atmospheric and soil environments. As forests form a key component of humid tropical ecosystems, the objective of this review is to synthesize the currently available experimental results and offer insights into how forests in the humid tropics would respond and adapt to different aspects of climate change. The atmospheric CO 2 concentration (C a ), which is currently at 386 μmol mol -1 , is expected to reach at least 600 μmol mol -1 by 2100. As CO 2 is an input in the photosynthetic process, increasing C a increases photosynthetic rates of all plants. In C 3 herbs and woody plants, which dominate humid tropical forests, this is translated into average biomass increases of 45% (maximum 240%) and 48% (maximum 260%) respectively, leading to increased net primary productivity at the ecosystem level. Increasing air temperatures could either enhance or dampen the growth stimulation due to higher C a , depending on the magnitude of temperature increase relative to the optimum temperature for maximum ecosystem productivity. The ability of humid tropical forests to acclimate and adapt to future temperature changes are considered to be lower than that of temperate forests because of the narrower range of seasonal temperatures experienced by forests in the humid tropics. Increasing C a and temperatures could also inﬂuence forest ecosystems through signiﬁcant changes in the soil environment and in the plant-water relations. In all the above-described responses to increased C a and temperature, there is substantial species-to-species variation. Studies on the capacity of tropical forests to sequester carbon and reduce future increases of C a have indicated a substantial capacity for forests in the humid tropics. However, reductions of this capacity have been observed during years of warmer temperatures and lower precipitation resulting from El Niño Southern Oscillation.


INTRODUCTION
Climate is the totality of the above-ground environment and its variation over time. The term 'climate change' has different connotations and interpretations. For example, the Inter-governmental Panel for Climate Change (IPCC) defines climate change as 'the long-term change in the above-ground environment due to human activity or natural variability'. On the other hand, the United Nations Framework Convention for Climate Change (UNFCC) specifically identifies the human influence and excludes the natural variability in its definition of climate change as 'change of climate attributed directly or indirectly to human activity that alters the composition of global atmosphere'.
There is increasingly conclusive evidence that longterm, human-induced climate change is taking place (Jones & Mann, 2004;IPCC, 2007). Climate change is triggered by increased atmospheric concentrations of greenhouse gases (GHGs), leading to a continuous, gradual warming of the atmosphere and thereby setting in motion a complex chain of changes in the climate. Although some of these changes have made their presence felt in recent times through increased frequency of extreme climatic events such as hurricanes, floods and droughts, many of the changes in climate are initiated by increased GHGs that occur over longer time scales (i.e. decades and centuries). Forests are extremely important ecosystems because of their multiple functions in biodiversity conservation, ensuring long-term environmental stability while providing a variety of economically-important products and services. Forest ecosystems will definitely experience the gradual, long-term changes in climate because of their long life spans. However, there is only limited research information on the potential impacts of climate change on individual plant species in forests and their functioning or on the processes and functioning of forests as ecosystems. A large portion of the limited information that is available is on temperate and boreal forests and their component plant species (Saxe et al., 1998(Saxe et al., , 2001. Therefore, the present review aims to provide a synthesis of the possible impacts of long-term climate change on forests in the humid tropics. However, because of the paucity of empirical data on tropical forests, we have had to rely on the basic principles of plant physiology and ecology and supporting data, which are predominantly from experiments on temperate plant species.

MAJOR ASPECTS OF LONG -TERM CLIMATE CHANGE
Some of the major aspects of long-term climate change that are likely to influence forest ecosystems are briefly described below:

Increasing concentrations of greenhouse gases in the atmosphere
Greenhouse gases are those that have the ability to absorb certain wavelengths of longwave radiation (i.e. wavelengths greater than 0.7 μm). Carbon dioxide, water vapour, methane, nitrous oxides and halogenated carbons are the GHGs in the earth's atmosphere. Most of the shortwave radiation coming from the sun passes through the earth's atmosphere. However most of the longwave radiation emitted from the earth's surface is absorbed by the GHGs. This leads to a gradual warming of the earth's atmosphere and this phenomenon is the well-known greenhouse effect (Houghton, 1997). Actual measurements and estimations from proxy sources have shown that the atmospheric concentrations of most GHGs have been rising steadily during the last three centuries, leading to an up-regulation of the greenhouse effect (called the enhanced greenhouse effect) (Houghton, 1997). These increases are beyond the range of fluctuations that these GHGs have shown over longer time scales of thousand/ million years. The current atmospheric levels of GHGs are higher than those at any time during the last 650,000 years (Siegenthaler et al., 2005) and they are increasing at a rate of 2.3 μmol mol -1 of CO 2 equivalents per year (IPCC, 2001;Stern, 2007).
Carbon dioxide is the GHG, which contributes most to the enhanced greenhouse effect. Atmospheric CO 2 concentration (C a ) was stable around 280 μmol mol -1 for several thousand years up to the beginning of the industrial revolution around 1750. Since then, C a has shown an exponential increase (Figure 1) up to the present level of around 386 μmol mol -1 . During the period between 1975 and 2010, C a in the tropics has increased by 14% from 330 to 377 μmol mol -1 (Clark, 2007;IPCC, 2007). In comparison, C a did not exceed 300 μmol mol -1 during the 420,000 year period before industrialization (Petit et al., 1999). As CO 2 is an essential component in photosynthesis, this rise in C a has profound implications on the functioning of forest ecosystems. Hence, the impacts of increasing C a will form a major part of this review.

Increasing air temperatures
When the increases of all GHGs are taken into account, some recent studies have predicted an increase of global mean temperature by 2 -5 o C by 2030-2060 (Wigley & Raper, 2001;Murphy et al., 2004) (Table 1), which would bring the earth to a global average temperature that has not been reached during the last three million years (Hansen et al., 2006). Moreover, there is a 20% probability of a temperature rise greater than 5 o C by 2050 and a warming of 3-10 o C by 2100 (Meinshausen, 2006;Stern, 2007). Global warming predictions of the Fourth Assessment Report of the IPCC (IPCC, 2007) are given in Table 2. During the 30-year period between 1975 and 2005, temperatures in the tropical regions have increased at a rate of 0.26 ± 0.05 o C per decade (Malhi & Wright, 2004). Although the tropics are expected to experience a slightly lower warming than the mid and high latitudes (Stern, 2007), such increases in air temperatures would influence the rates of a wide range of temperaturedependant processes in plants and soils with profound implications for forests in the humid tropics.

Increased frequency of droughts and other extreme climatic events
Warmer air temperatures will accelerate the hydrologic cycle with significant shifts in the amount and temporal distribution of rainfall. Although there is uncertainty about whether annual total rainfall in the tropical zone would change significantly (Stern, 2007), there is increasing evidence that the same annual rainfall will come in a fewer, more intense rainfall events (Wetherald & Manabe, 2002;Huntington, 2006). On the other hand, annual rainfall is predicted to decrease significantly in Mediterranean climates (Stern, 2007). Globally, the percentage of land area under moderate drought at a given time is expected to increase from the current level of 25% to 50% while the corresponding percentage for extreme drought is projected to increase from 3% to 30% (Burke et al., 2006). Hence, forests in the humid tropics could experience drought of greater intensity and longer duration. During the last three decades, Journal of the National Science Foundation of Sri Lanka 39 (4) December 2011 Figure 1: Long-term variation of the atmospheric carbon dioxide concentration along with predicted increases under different emission scenarios: A1 -Rapid economic growth, with convergence between different regions of the world. A1F -Economic growth with high dependence on fossil energy sources; A1T -Economic growth with non-fossil energy sources; A1B -Economic growth with a balanced mixture of energy sources; A2 -Regionally-oriented and unevenly-distributed (i.e. divergent) economic growth, which is slower than in A1; B1 -Convergent world which is rapidly moving global solutions to economic, social and environmental sustainability with the introduction of clean and resource-efficient technologies; B2 -Intermediate rates of economic development focusing on environmental protection and social equity, but with emphasis on local solutions and less rapid and more diverse technological change than in B1 and A1; IS92a -'Business-as-usual' scenario. there have been decreases in rainfall over some tropical regions, with tropical Asia recording a marginal decrease while tropical North Africa showing a 3-4% decrease (Houghton et al., 2001;Malhi & Wright, 2004) and Amazonia showing no significant variation (Malhi & Wright, 2004). Furthermore, there has been an increase in the intensity of El Niño events, causing episodes of significant drought and high temperatures. This has also increased the frequency of regional scale forest fires releasing significant quantities of aerosols into the atmosphere and causing up to 65% reduction of incident photosynthetically active radiation in South East Asia (Kobayashi et al., 2005) and Amazonia (Nepstad et al., 2004).

December 2011
Journal of the National Science Foundation of Sri Lanka 39 (4)

Experimental approaches
The present review will focus on the impacts of increasing C a and air temperatures (T a ) and varying water availability on forests in the humid tropics. The type and magnitude of climate change impacts on a given forest ecosystem is partly determined by the environmental factors that control its functioning and productivity. Under the current climatic conditions, for forests in the humid tropics, radiation receipt and temperature are the main factors of control as water and nutrient availability are unlikely to be limiting (Churkina & Running, 1998;Nemani et al., 2003;Running et al., 2004).
It is important, at the outset, to appreciate the practical experimental difficulties involved in collecting valid research data on the impacts of climate change on a complex ecosystem such as a forest. The slow, gradual nature of climate change necessitates observations over a long time period to detect its impacts on a forest ecosystem, which itself is likely to have a rather long response time. Hence, alternative approaches, which would yield valid predictions within a shorter period of time need to be sought. Such an approach is to quantify the response of seedlings of key plant species to certain   specified increases of C a and T a , which are expected to occur during the next 50-100 years. Accordingly, for C a , elevated levels have ranged from 570 μmol mol -1 (a 200 μmol mol -1 increase, which is expected to occur in another 40-50 years, IPCC, 2007;Stern, 2007) to 700 μmol mol -1 (a doubling of the current ambient C a , which is expected by 2100) while the temperature elevations have ranged from 2 o C to 4 o C (which are expected in the next 40-70 years, IPCC, 2007;Stern, 2007). Plants growing under elevated C a have been compared with those growing at the current ambient levels of C a , with the implicit assumption that the response of plants within this range of C a is linear. It is already known that response of plant processes to increasing tissue or air temperature is 'bell-shaped' (Saxe et al., 2001) with increasing rates of processes up to an optimum temperature followed by a decrease. Hence, a combination of these trends is used in predicting the impacts of increasing C a and T a .
In the majority of experiments, CO 2 enrichment to achieve the required elevation of C a has been carried out on seedlings growing in open top chambers (Whitehead et al., 1995;Norby et al., 1997;De Costa et al., 2003), which are ventilated enclosures with an open top and transparent side walls to simulate 'near natural' conditions. A limited number of experiments on forest plants have been conducted in closed top chambers (CTCs) with re-circulation of air (Beerling & Woodward, 1996;Forstreuter, 1996;Tingey et al., 1996). Free Air CO 2 Enrichment (FACE) is a more expensive method of enriching a larger, un-enclosed (open) area with CO 2 to simulate completely natural conditions . A limited number of FACE studies have been carried out on temperate forest stands, but none on tropical forests. Responses to increased T a has been carried out either in open top chambers heated with heating coils, temperature gradient chambers or in completely enclosed growth chambers.

Quantification of the ecosystem response
Even the few FACE systems are established on monocultures of forest stands. Open top chambers and enclosed growth chambers can only house potted individual seedlings of trees. Extrapolation of plant growth responses obtained under such experimental conditions to predict the response of a complex and multispecies ecosystem such as a forest in the humid tropics is difficult and questionable. A more valid alternative would be to quantify the response of basic plant and soil processes to the different aspects of climate change and use this information to synthesize the ecosystem response (Landsberg, 2003;Mäkelä, 2003). Hence, this approach is followed in the present review. However, even with this approach, the results on many key processes are not unequivocal (Boisvenue & Running, 2006) and therefore, extrapolations, generalizations and predications can be made only with extreme caution (Luo et al., 1999;Beier, 2004). Furthermore, climate change involves the variation of several environmental factors (i.e. C a , T a and water availability) simultaneously and complex interactions in their effects. Therefore, the net result of overall climate change on a given forest ecosystem could be site-and environment-specific (Norby & Luo, 2004).

Inter-and intra-species variation in the magnitude of responses
A review of experimental results on the response of trees to different aspects of climate change show clear differences among species and also among provenances/ populations/families within a species (Eamus & Jarvis, 1989;Poorter, 1993;Berryman et al., 1994;Battaglia et al., 1996;Beerling, 1996;Basso & Bazzaz, 1998;Faria et al., 1998;Atkin et al., 1999;Jablonski et al., 2002;Arulmageswaran et al., 2003;De Costa et al., 2003;Poorter & Navas, 2003;Körner et al., 2005). These variations have to be taken into consideration when reconstructing ecosystem response from studies on individual processes of single plants.

INFLUENCE OF ELEVATED CARBON DIOXIDE AND TEMPERATURE ON BASIC PLANT AND SOIL PROCESSES
An overview of the inter-connected network of key plant and soil processes that can be influenced by climate change is given in Figure 2.

Response of photosynthesis, respiration and biomass production to increased C a
Photosynthesis is one of the key processes affected by increasing C a because CO 2 is an essential input for it. A wide body of evidence has shown that the net photosynthetic rate per unit leaf area (P n ) increases with increasing C a (Eamus & Jarvis, 1989;Curtis, 1996;Drake et al., 1997;Saxe et al., 1997;Wullschleger et al., 1997;Luo et al., 1999;Nowak et al., 2004). This is primarily because of two reasons. Firstly, a greater CO 2 concentration in the outside air would allow a greater airleaf CO 2 concentration gradient and thereby increase the rate of CO 2 uptake in to the leaves. If the leaves do not have a limitation of photosynthetic structures (i.e. chlorophyll, components of the electron transport pathway, thylakoid membranes etc.) increased uptake of CO 2 would lead to a greater production of primary carbohydrates (i.e. glucose, fructose, sucrose, starch, etc.) through gross photosynthesis. This is called the 'CO 2 fertilization effect'. Secondly, the greater CO 2 concentration at the

December 2011
Journal of the National Science Foundation of Sri Lanka 39 (4) chloroplast partly suppresses photorespiration (Drake et al., 1997) and thereby channels a greater percentage of the absorbed CO 2 through photosynthesis, thus increasing P n . Compilations of responses of a wide range of plant species have shown a significant species-to-species variation in the magnitude of increase of P n in response to a doubling of C a . Among the tree species, Saxe et al. (1998) showed that long-term photosynthetic stimulation by elevated C a was slightly higher for conifers (62%) than for deciduous tree species (53%). Specifically for a selected set of tropical tree species growing in Sri Lanka, De Costa et al. (2003) observed increases of P n ranging from 18% to 224% in response to an increase in C a from 370 to 570 μmol mol -1 (Table 3).
This stimulation of photosynthesis by increasing C a has been shown to be greater for C 3 than for C 4 plants (Drake et al., 1997;Long et al., 2004). Kranz anatomy and the consequent increase of internal CO 2 concentration within the C 4 leaves makes CO 2 a lesser limitation for C 4 photosynthesis (Sage, 2001). In addition to the decrease of photorespiration (R p ), there is considerable evidence  that mitochondrial respiration (R d ) also decreases with increasing C a (Griffin et al., 1996;Bunch, 1997;Drake et al., 1997;Drake et al., 1999). A review of existing literature incorporating all types of plants has shown that R d decreases by an average of 17-20% for a doubling of C a (Eamus & Jarvis, 1989;Drake et al., 1999). This has been attributed to a CO 2 -induced reduction of the activity of two key enzymes of the respiratory pathway (i.e. cytochrome oxidase and succinate dehydrogenase) (Gonzàlez-Meler et al., 1997) and reduced respiratory cost for tissue synthesis under elevated C a (Wullschleger et al., 1994;Amthor, 1997). Reductions in R d in different forest tree species have been observed in several studies (Stewart & Hoddinott, 1993;Karnosky et al., 1999;McDowell et al., 1999). However, CO 2 -induced reduction of R d is not as conclusive as the decrease of R p because there is also evidence against a significant reduction of R d (Saxe et al., 1998;Ceulemans et al., 1999;Amthor, 2000). For example, Overdieck (1993), Pontailler et al. (1994) and Saxe et al. (1998) have observed increases in R d due to elevated C a in Pinus sylvestris (13-19%) and Fagus sylvatica (53-84%). An increase in R d under elevated C a is possible because of increased carbohydrate availability due to increased P n (Amthor, 1995). Moreover, even if specific R d (i.e. dark respiration rate per unit plant biomass) may be reduced with increasing C a , the greater plant biomass produced under elevated C a could increase total plant R d because of increased maintenance respiration. In contrast, Martin et al. (1994) concluded that there is no direct effect of elevated CO 2 on woody tissue respiration because of the higher CO 2 concentrations (i.e. 1%) in wood. However, Carey et al. (1996) observed an increase in R d per unit volume of stem wood due to elevated C a . Hence, a clear scientific consensus has not emerged yet on the direction or magnitude of response of R d to CO 2 enrichment (Amthor, 1997;Saxe et al., 1998;Ceulemans et al., 1999).
It is also possible that the response of R d to elevated C a depends on the type of tissue. Ryan et al. (1996) showed that in stems and coarse roots, R d was positively correlated with biomass (resulting in an increase of R d under elevated C a ) whereas in foliage and fine roots, R d was directly correlated with N concentration (causing a reduction of R d with increasing C a ). It is notable that in trees R d of leaves constitutes more than 50% of total R d . There is fairly clear evidence that root respiration is greater at elevated C a Voce et al., 1995;Rouhier et al., 1996). Mousseau (1996) quoted that in model ecosystems, elevated C a decreased leaf R d by 28%, increased stem R d by 40% and reduced root R d by 13%, when all R d values are expressed on a per unit dry weight basis. The total stand respiration was increased by 52% primarily because of a 75% increase in total biomass.
Another reason for the conflicting results obtained could be variation in the respiratory response at different stages of tree growth. Wullschleger et al. (1995) observed an increase in R d in response to increased C a only during periods of leaf growth. Out of the 111% increase in R d observed under elevated C a during the active growing period, a higher proportion was attributed to increased growth respiration. The direction and magnitude of change in total plant respiration in a future climate will partly determine whether forests would act as a sink or source of atmospheric CO 2 . Therefore, more research is urgently needed to resolve the uncertainties involved in this response.
The net result of the increased gross photosynthesis and decreased R p under elevated C a is an increase in biomass production in plants. A possible reduction of R d would further increase this stimulation of plant growth at increased C a . Inter-species variation in the magnitudes Species P n (μmol m -2 s -1 ) after 75 days P n (μmol m -2 s -1 ) after 105 days of CE of CE C a at 370 μmol C a at 570 μmol C a at 370 μmol C a at 570 μmol mol -1 mol -1 mol -1 mol -1  of stimulation of P n and reductions of R p and R d means that a considerable species-to-species variation could be expected in the stimulation of biomass production by elevated C a . When plants are growing with adequate water and nutrients, for C 3 woody plants, the growth stimulation with a doubling of C a has ranged from 0 to 260% with an average of 48% while for C 3 herbs, it has ranged from 0 to 240% with an average of 45%. These values were obtained in a meta-analysis of experimental data for 160 woody plants and 144 herbs (Poorter & Navas, 2003). Curtis and Wang (1998) and Idso (1999) have reported biomass increases up to 30-50% with woody species subjected to long-term CO 2 enrichment. Specifically, the inter-species variation in the growth stimulation due to increased C a for seedlings of some tropical forest tree species is shown in Figure 3. Schlesinger and Lichter (2001) observed that carbon accumulation in the forest floor was also greater under elevated C a as compared to ambient. Moreover, Naumburg and Ellsworth (2000) showed that shaded leaves of saplings under elevated C a maintained greater rates of P n for longer periods than those of saplings under ambient C a , indicating that the contribution from the lower vegetation strata of a forest to carbon accumulation could increase in a future climate.
In an analysis of experimental data from different sources, Novak et al. (2004) concluded that the observed increases in tree biomass production in response to elevated C a , which are largely based on observations on seedlings growing in open top chambers, are adequately supported by observations on mature trees growing under FACE. However, they emphasize that both species composition and resource availability play important roles in determining the magnitude of response. The CO 2 -induced growth stimulation of trees occurs because of both increased photosynthetic rate per unit leaf area and increased leaf area per plant (Norby et al., 1992;Ceulemans et al., 1995;Norby et al., 1995).

Implications of the stimulation of biomass production under increased C a
Comparison of the inter-species variation in growth stimulation due to increased C a has shown that the stimulation is 30% greater for fast-growing plant species (i.e. mainly pioneers in the succession process) as compared to slow-growing species (i.e. mainly latesuccession species) (Poorter, 1993;Lovelock et al., 1998;Atkin et al., 1999;Winter & Lovelock, 1999;Poorter & Navas, 2003). This would have important implications for the species composition of forests in the humid tropics. Hence, a discernible shift in species composition in favour of fast-growing pioneers could be expected. As most of the alien invasive species are fast-growing, there could be a greater probability of invasive species degrading the humid tropical forests. This could also result in greater rates of extinction for threatened and endangered species. With the proportion of fast-growing species in the ecosystem increasing, the species turnover in the forest could be faster.
Review of species responses to elevated C a has also shown that the photosynthetic and growth response is clearly greater for C 3 plants than for C 4 plants. In the meta-analysis of Poorter and Navas (2003), the estimated biomass increase in response to a doubling of C a was only 12% (an average from 40 C 4 species) in C 4 plants as compared to 45% (an average from 160 C 4 species) in C 3 plants. Therefore, over longer time scales, the proportion of C 4 species in a forest ecosystem could decrease along with an increase in the proportion of C 3 species.
A widely-observed response of plants growing under elevated C a is that a higher proportion of increased biomass is partitioned to roots , especially increasing the fine root biomass (Norby et al., 2000;Pregitzer et al., 2000). In addition, there is greater root branching in trees under elevated C a . These responses increase the capacity of the tree root systems to explore a greater volume of soil and absorb more soil resources such as nutrients and water. Although forests in the humid tropics do not experience substantial shortages of water or nutrients under the current climate, C a -induced increases in root absorption capacity will enable them to better withstand any shortages of these soil-based resources (especially, the soil water) in a future climate.

Photosynthetic and growth response to elevated C a under non-optimum growing conditions
The question of whether the above reported stimulation of P n and biomass production by increased C a under optimum conditions (i.e. when plants are grown with adequate water and nutrients under light and temperature regimes to which they are best-adapted), would occur to the same extent under non-optimum growing conditions have been investigated. A review of results by Poorter and Pérez-Soba (2001) shows that there is a reduction of the photosynthetic and growth stimulation under low nutrient availability and low temperatures. The average magnitude of reduction was 20-22% as compared to the growth stimulation under optimum conditions. Interestingly, Poorter and Navas (2003) show that under low nutrient availability, the greater CO 2 -induced growth stimulation of C 3 plants as compared to C 4 plants is nullified. The only exception here is the C 3 legumes, which maintain an advantage over C 4 species because of their N-fixing ability. In an analysis of data from a FACE experiment, with C a elevated up to 550 μmol mol -1 , CO 2induced stimulation of net primary productivity (NPP) of Pinus taeda were 27% and 19% under high and low N availability (Finzi et al., 2002;Norby et al., 2005). In contrast, growth stimulation due to increased C a occurred to the same magnitude under conditions of lower water availability, shade and salinity. However, Oren et al. (2001) concluded that C a -induced increases in NPP would be minimum or zero under prolonged water deficits on poor soils. Generally, forests in the humid tropics do not experience nutrient shortages because of their closed and efficient recycling of nutrients. Low temperatures would be experienced only in those forests in the humid tropics, which are located at higher elevations. Therefore, it is highly likely that humid tropical forests would show greater rates of biomass production and NPP in a future climate of increased C a (Melillo et al., 1993;Field et al., 1998). This is also likely to increase their capacity for carbon sequestration.

Long-term acclimation of photosynthesis to elevated C a
Analysis of the biochemical composition of plant tissue produced under elevated C a has shown a clear increase in the proportion of carbohydrates, and more importantly, a clear decrease in the concentration of nitrogen and hence of proteins (Drake et al., 1997;Cotrufo et al., 1998). In leaves, a reduction in the protein content would mean a reduction of essential photosynthetic enzymes such as Rubisco (ribulose 1,6-bis phosphate caroboxylase oxygenase), thus down-regulating the photosynthetic capacity (Gunderson & Wullschleger, 1994;Curtis, 1996). Therefore, if P n of plants grown under elevated C a throughout is measured at ambient C a , its P n would be lower than that of a plant grown at ambient C a throughout. This is called photosynthetic acclimation. In addition to a reduction in the concentration of Rubisco, a reduction of its activity has also been observed under elevated C a (Tissue et al., 1996). Photosynthetic acclimation is thought to occur because of the inability of a plant to utilize the extra carbohydrates produced by photosynthetic stimulation under elevated C a (Long & Drake, 1992). As trees have a large capacity for storage of extra carbon in their woody tissue and a large number of growing points, which can act as sinks for additional photosynthates (Wullschleger et al., 1997;Janssens et al., 1999), it is argued that photosynthetic downregulation is rare in trees . This is supported by the findings of Idso et al. (1991) who did not observe any down-regulation in either P n or growth in field-grown orange trees subjected to long-term CO 2 enrichment. The meta-analyses by Curtis (1996) and Curtis and Wang (1998), which involved a wide range of tree species, found only 10% down-regulation. In a review of 16 FACE studies involving different vegetation types, Novak et al. (2004) found that while elevated C a reduced leaf N in herbaceous species, there was little change in leaf N in most of the woody species.
In spite of any down-regulation of photosynthetic capacity, the gross and net photosynthetic rates would still be increased under elevated C a , as compared to ambient C a , because of the greater magnitude of the CO 2 fertilization effect and suppression of photorespiration. However, the altered biochemical composition of plant tissues produced under elevated C a would definitely alter the rates of litter decomposition and nutrient release and thereby influence the nutrient recycling process (Swift et al., 1979;Cotrufo et al., 1994;Berntson & Bazzaz, 1996). Several studies have shown increased C/N (Köner, 2000) and lignin/N ratios in tissues grown under elevated C a resulting in lower rates of litter decomposition (Coûteaux et al., 1991;Cotrufo et al., 1994;Boerner & Rebbeck, 1995;Scherzel et al., 1998;De Angelis et al., 2000). On the other hand, Hirschel et al. (1997) found that elevated C a had no significant effect on litter decomposition rates of tropical rainforest species. Feeding behaviour and the population of leafeating insects could also be affected by the changed leaf biochemical composition (Lindroth et al., 1993). Carbon

December 2011
Journal of the National Science Foundation of Sri Lanka 39 (4) inputs to the soil through litter fall and rhizodeposition will increase because of increased biomass production under elevated C a .

Some pertinent questions on the predicted increase of forest growth under increased C a
While predicting an increase of P n , biomass production and NPP under elevated C a , some pertinent questions need to be raised as these predictions are based largely on responses observed in seedlings growing in open top or closed chambers. Firstly, the down-regulation of photosynthetic capacity raises several questions of the sustainability of photosynthetic and growth stimulation of elevated C a in a perennial ecosystem such as a forest: (a) Will there be a point (i.e. a ceiling level of C a ) in a future climate at which the photosynthetic stimulation due to increasing C a would be limited by the amount of photosynthetic machinery available in leaves?, (b) As a consequence, will the growth stimulation level-off at a certain level of C a ?, (c) As stimulation of growth has to be supported by increased capture of essential growth resources (i.e. light, water and nutrients), will resource capture be a limitation to further stimulation of growth at a certain point of C a in the future?
Answers to questions (a) and (b) have to be sought on the physiological basis of photosynthetic acclimation. As the internal CO 2 concentration at the site of photosynthesis (i.e. chloroplast), C i , is sub-optimal in C 3 plants (because of the relatively lower CO 2 concentration in the atmosphere and the absence of a CO 2 concentrating mechanism such as that found in C 4 plants), a greater amount of photosynthetic enzymes (primarily Rubisco) has to be present in the chloroplast to achieve appreciable rates of photosynthesis (Drake et al., 1997;Long et al., 2004). With the elevated C a increasing the air-leaf CO 2 concentration gradient and thereby increasing C i , the essential photosynthetic enzymes need not be present at higher concentrations to achieve higher rates of photosynthetic carboxylation. Biosynthesis of proteins require greater respiratory energy (and a greater amount of photosynthates for respiration) than biosynthesis of carbohydrates (Penning de Vries et al., 1974). Hence, with the reduced requirement of photosynthetic enzymes (i.e. proteins) under elevated C a , allocation of photosynthates in leaves shifts from proteins to carbohydrates, thus causing photosynthetic acclimation. This follows that if the essential photosynthetic enzymes become a limiting factor to achieve the required rates of photosynthesis to utilize the increased CO 2 uptake at a future point, the plants should be able to re-adjust and allocate more photosynthates again to protein synthesis. Hence, it is likely that photosynthetic acclimation may not impose a ceiling on growth stimulation in response to increasing C a . However, as stated earlier, it is possible that resource limitation [question (c)], especially nutrients, could curtail CO 2 -induced growth stimulation, especially in non-leguminous species.
A second pertinent question is whether the growth stimulation observed in the seedlings, which are at the juvenile stage, would be reproduced in fully-grown trees that have come to the reproductive stage. The few FACE studies have shown that the growth stimulation due to increased C a does occur in fully-grown trees as well. On the other hand, fully-grown trees have a much greater respiration cost (especially maintenance respiration) in order to maintain their large biomass. Therefore, whether mitochondrial respiration is reduced or not by elevated C a would be a critical factor in determining the magnitude of growth stimulation in fully-grown trees.

RESPONSE OF PHOTOSYNTHESIS, RESPIRATION AND BIOMASS PRODUCTION TO INCREASED AIR TEMPERATURES (T a )
Long-term, gradual increase of C a is accompanied by an increase of T a because of the enhanced greenhouse effect. Hence, plants and their internal processes respond simultaneously to increases of both C a and T a . Generally, all plant physiological and metabolic processes respond in a similar manner to increasing T a (Figure 4) (Johnson & Thornley, 1985). When the tissue temperatures (which are closely correlated with T a ) increase from low levels, rates of processes increase (often linearly) until an optimum temperature (T o ) is reached. This is because the reaction molecules would have a greater kinetic energy at higher temperatures. When T a increases beyond T o , process rates decrease primarily because of the gradual denaturation of enzymes (Berry & Björkman, 1980) and other effects of heat stress such as damage to thylakoid membranes (Méthy et al., 1997). Therefore, the direction and magnitude of photosynthetic and growth response to a simultaneous increase of T a and C a would be determined by the respective T o and the magnitude of increase of T a in relation to T o . If the increase of T a occurs within the sub-optimal range of temperatures for photosynthesis and biomass production, the rates of these processes would be further stimulated by increased T a . This is most likely to occur in the forests of the humid tropics, which are located at higher altitudes (situation a in Figure 4). In a review of 27 studies of temperature effects on net photosynthesis of different tree species, Saxe et al. (2001) concluded that the temperature increases that are expected to occur during this century would increase the net photosynthesis

Journal of the National Science Foundation of Sri Lanka 39 (4)
December 2011 of 66% of the 27 studies reviewed. This review largely focused on species growing in temperate environments, which are equivalent to high altitudes of humid tropics, at least in terms of the temperature regime.
On the other hand, if the future increases of T a occur around T o , increased T a would not cause a significant change in process rates with only the growth stimulation due to C a occurring (situation b in Figure 4). However, if the T a is pushed beyond T o during future warming of the atmosphere, increased T a would have a negative effect on the rates of photosynthesis and biomass production. Therefore, in a future climate where C a is increased along with an increase of T a above T o , the direction and magnitude of change in the rates of photosynthesis and biomass production would be determined by the relative magnitudes of the positive effect of increasing C a and the negative effect of increasing T a (situation c in Figure 4). For most processes of tropical plant species, T o is within the range of 25 -32 o C (Squire, 1990). At present, most lowland forests within the humid, tropical zone experience an average air temperature of about 26 -28 o C throughout the year. This is just below or near the T o for most physiological processes responsible for plant growth. Therefore, it is likely that increases in T a in the near future (i.e. the next 10-20 years) may not cause significant changes in the biomass production rates of forests located at lowlands in the humid tropics. During the same period increasing C a is likely to cause an increase in growth rates (situation b in Figure 4). However, it is also highly probable that at some point of time in the future (i.e. probably in the next 50-60 years), the T a would be pushed significantly beyond T o , bringing the lowland humid tropical forests into situation c in Figure 4. Several global ecosystem process models (White et al., 2000;Cramer et al., 2001;Fung et al., 2005) have predicted that the productivity of tropical forests would decrease with increasing T a . There are also predictions of substantial die-off of tropical forests from mid-century onwards (Jones et al., 2003;Cowling et al., 2004). Leverenz et al. (1999) have shown that even in temperate tree species, photosynthesis could be decreased when the increase in T a is greater than 2-3 o C above the summer ambient.
On the other hand, because the increase in T a is gradual, plant processes and ecosystems can adapt to increasing T a to a certain extent. For example, Battaglia et al. (1996) showed that T o for maximum light-saturated photosynthesis (P max ) of Eucalyptus can be increased by up to 10 o C due to acclimation to high temperatures. If this happens the majority of plant species in forests of the humid tropics could remain in situations a or b for a longer period. Increasing T a in the range between 20 o and 30 o C has been reported to have positive impacts on tree photosynthesis by increasing the contents of photosynthetic pigments (Ormrod, 1999) and increasing the apparent quantum yield of photosynthesis (i.e. net photosynthetic rate per unit of radiation absorbed) (Wang et al., 1996;Lewis et al., 1999;Saxe et al., 2001).
It is also important to note that the response of photosynthesis to temperature involves interactions with other environmental factors such as vapour pressure deficit and irradiance and plant factors such as leaf temperature and stomatal conductance. For example, Bassow and Bazzaz (1998) showed that variation of photosynthesis in four temperate deciduous tree species was only weakly correlated with variation in leaf temperature and that when averaged over a season, temperature variations explained only a small percentage of the variation in photosynthesis. Beyond certain thresholds, both higher and lower temperatures could lead to partial stomatal closure (Jarvis, 1976). When combined with high irradiance, lower stomatal conductance could lead to formation of reactive oxygen species and phototoxic damage to the photosynthetic tissue (i.e. photoinhibition) (DeLucia et al., 1991;Saxe et al., 2001). Increases in T a could also cause stomatal closure due to increased water deficits in leaves (Jarvis, 1976) and thereby reduce photosynthesis and accelerate photoinhibition. However, some tree species already have the capability of evolving photo-protection mechanisms to avoid photoinhibition (Faria et al., 1998) and avoid potential adverse effects of higher T a .

Inter-species variation in T o and the magnitude of response to increasing T a and its implications
Experimental work has shown considerable speciesto-species variation in T o and temperature-sensitivity (Table 4) (Saxe et al., 2001;Battaglia et al., 1996;Mebrahtu et al., 1991;Sun & Sweet, 1996;Arulmageswaran et al., 2003). Very often, plant species that are adapted to grow at higher temperatures (i.e. low altitudes in the tropics) have higher T o values than those that are adapted to grow under cooler temperatures (i.e. higher altitudes in the tropics). This inter-species variation in T o could modify or shift the responses shown in Table  4. A higher T o for species growing in a lowland forest would make them less vulnerable to future increases in T a as it would take a longer duration for T a to be pushed significantly above T o . On the other hand, a lower T o for species growing in high altitude forests would make them more vulnerable to future increases of T a as T a could be increased above T o within a shorter period.
Inter-species variation in T o within a given forest ecosystem would mean that with time, the ecosystem composition would shift in a such way that the proportion of species adapted to warmer temperatures (i.e. having higher T o ) would increase at the expense of those adapted to relatively cooler temperatures (i.e. having lower T o ), which may become threatened and endangered. Accordingly, the ecosystem boundaries could also shift, with especially the high altitude forests shrinking in area.
However, the capacity of an individual species or a whole ecosystem to adapt to environmental change should not be discounted. As the increase of T a due to enhanced greenhouse effect occurs at a very gradual rate, at least some individual species and ecosystems may have the capacity to adapt to the gradually-increasing temperatures. However, in this regard, it has been shown that tropical forest species have a lower capacity for adaptation to higher temperatures than temperate forest species (Cunningham & Read, 2003). This is not surprising as temperate species experience a much greater fluctuation in temperature, both within a 24-hour day and during a year, even at present. Therefore, temperate species have developed internal mechanisms and adaptations to withstand these temperature fluctuations. In contrast, such mechanisms and adaptations are largely absent in tropical species, which experience only a narrow range of temperatures both diurnally and annually.

INFLUENCE OF INCREASING C a AND T a ON PLANT WATER RELATIONS
Both increasing C a and T a have important influences on water relations of forests in the humid tropics. However, it should be noted that under the current climatic conditions water is not a serious limiting factor for forest growth, reproduction and ecosystem function in humid tropics where rainfall exceeds evapotranspiration almost throughout the year. However, with the predicated increase in the frequency of high rainfall episodes and the associated probability of longer rainless periods in a future climate, forests in the humid tropics could also experience water deficits for at least short periods of the year. Hence, the effects of C a and T a on processes determining the water balance of forest plants could increase in importance in a future climate.  Arulmageswaran et al., 2003). Table 4: Response of absolute and relative growth rates to a 4 o C increase in air temperature in seedlings of selected tropical tree species growing in controlled environmental growth chambers. Plants were grown with adequate water and nutrients. Growth rates were computed for a 100 -day (=100 day period)period from the time of imposing temperature treatments.

Journal of the National Science Foundation of Sri Lanka 39 (4) December 2011
Both C a and T a primarily influence plant water relations by influencing the rate of transpiration. As transpiration occurs through the stomata, stomatal conductance (i.e. a measurement of how easily water is able to move through the stomata in the vapour phase) has a direct relationship with transpiration rate. Stomatal conductance (g s ) is determined by the stomatal density (i.e. the number of stomata per unit leaf area) and the degree of stomatal opening. Increasing C a has been shown to decrease g s in a range of tree species (Mousseau & Saugier, 1992;Berryman et al., 1994;Field et al., 1995). However, the magnitude of reduction of g s has been found to be lower in trees (i.e. around 20%, Field et al., 1995) as compared to herbaceous plants in which the reduction is around 40% (Morison, 1987). Moreover, the magnitude of CO 2induced change in g s has shown considerable variation between species (Field et al., 1995;De Costa et al., 2003;Norby & Luo, 2004) with the g s of some species showing no response (Curtis, 1996;Eamus, 1996) or even an increase (Barton et al., 1993;Heath & Kerstiens, 1997). A significant contribution to the above decreases of g s has been the reduction of stomatal densities with increasing C a . Analysis of fossilized plant samples that grew several hundred to thousand years ago at a lower C a has shown higher stomatal densities as compared to those grown in recent times at higher C a (Woodward, 1987;Van de Water et al., 1994;Beerling, 1996;Heatherington & Woodward, 2003). However, further studies have shown that stomatal densities are much less sensitive to increases of C a from the current ambient levels upwards than they have been to historical increases in C a (Woodward & Bazzaz, 1988). However, g s could still decrease with future increases of C a through reduced stomatal apertures as a direct response of guard cells to increased intercellular CO 2 concentration (Morison, 1987;Mansfield et al., 1990;Mott, 1990). Interestingly, different ecotypes of a given species or genus growing across an altitude gradient have shown a gradient in stomatal density and g s with both showing an increasing trend with increasing altitude (Woodward et al., 2002). This has been shown to be a response to the natural decrease of C a with increasing altitude.
Actual measurements have shown that reduced g s at increased C a results in a parallel decrease in transpiration rate per unit leaf area (E l Costa et al., 2003). However, a reduction in E l may not necessarily decrease total transpiration, either on per tree or per unit land area basis, because of the associated increase of leaf area per tree under elevated C a (Ellsworth et al., 1995;Lindroth, 1996;Ceulemans et al., 1999). In situations where the proportional reduction in g s is lower than the proportional increase of leaf area, total transpiration could even increase under elevated C a .
As elevated C a causes a simultaneous increase of the net photosynthetic rate per unit leaf area (P n ), transpiration efficiency, i.e. the ratio between P n and E l and sometimes called water use efficiency, is increased with increasing C a (Morison, 1993;Drake et al., 1997). Significant increases in water use efficiency with increasing C a have been shown for forest trees growing under natural conditions (Fernandez et al., 1998;Feng, 1999). This means that at elevated C a , plants are able to produce more biomass per unit of water used as transpiration, thus increasing their ability to tolerate water-limited periods.
While increasing C a reduces transpiration rates, simultaneous increases of T a tend to increase E l . This is primarily because of the increased leaf-to-air vapour pressure gradient at increased T a . As increases of T a beyond T o could also decrease P n , the predicted increases of transpiration efficiency with increasing C a would be dampened down by the simultaneous increase of T a . Here again, this dampening down of transpiration efficiency is most likely to be felt by the forests in the lowlands of humid tropics, which already experience T a levels that are closer to T o . Because of the inherent ability of C 4 plant species to function at higher temperatures, CO 2 -induced increases of transpiration efficiency of C 4 plants (which is higher than that of C 3 plants at the current levels of C a and T a ) are less likely to be reduced by increasing T a .

INFLUENCE OF INCREASING C a AND T a ON REPRODUCTIVE BIOLOGY
Effects of climate change on reproductive biology of plant species in a forest will have important implications for possible long-term shifts in the species composition of the ecosystem. A meta-analysis by Jablonski et al. (2002) using the wide range of plant species has shown that increasing C a within the range of 500 -800 μl l -1 (ppm by volume) has increased flowering by an average of 19%, fruiting by 18% and seed production by 16%. The individual seed weight also has shown an average increase of 4%. Although these increases in the reproductive effort were proportionately lower than the 31% increase observed for the total biomass of the same set of species used in the meta-analysis, long-term increases in C a are likely to have a positive influence on the reproductive biology of plants, irrespective of the species. Notably, with the exception of leguminous species, the seed nitrogen concentration decreased with increasing C a , by an average of 14%.
While the influence of increasing C a on reproductive biology appears to be positive, increasing T a could have both positive and negative effects. All plants have to complete a specific, genetically-determined thermal

December 2011
Journal of the National Science Foundation of Sri Lanka 39 (4) duration before flowers or other forms of reproductive structures are initiated (Roberts et al., 1997;Saxe et al., 2001). With increasing T a , this specific thermal duration requirement would be fulfilled within a shorter period of time. In plant species, which flower seasonally, this could prompt earlier flowering than at present. Hence, flowering and fruiting periods for different species within a humid tropical forest ecosystem could undergo temporal shifts in a future, warmer environment. This could set off a chain of events by influencing the behavioural patterns of the fauna, which depend on the fruits of different plant species within the ecosystem.
In addition to its influence on the thermal duration requirement, increasing T a could influence the processes of flower development, pollination and seed/fruit formation. Generally, these stages are highly sensitive to supraoptimal temperatures. Therefore, increasing T a beyond a genetically-determined optimum could adversely affect the development of reproductive structures. However, there can be substantial variation in the degree of heat tolerance between different species. Accordingly, species with a relatively greater heat tolerance, especially at the reproductive phase, could increase proportionately and change the composition of future forest ecosystems in the humid tropics.

INFLUENCE OF INCREASING C a AND T a ON SOIL PROCESSES
Soil is the major source of nutrients to plants in a forest ecosystem and nutrient recycling plays a major role in supporting the rich diversity of flora and fauna in forests of the humid tropics. Several processes related to nutrient recycling are affected by increasing C a and T a . The increased biomass production under elevated C a will increase the addition of organic material to the soil through litter fall . Proportionately greater stimulation of root biomass by elevated C a would increase the organic matter inputs to the soil through root exudates and through root decomposition and turnover (Norby, 1994). However, decomposition rates of litter produced under higher C a would be slower because of its greater carbon and lower nitrogen concentrations (i.e. higher C:N ratio) (O'Neill & Norby, 1996). In lowland forests of the humid tropics, where the temperatures are higher, slowing down of the litter decomposition may be an advantage because it would lead to slower release of nutrients and lower leaching losses caused by higher rainfall. However, the increase of T a that accompanies the increase of C a would ultimately result in soil temperatures (T s ) also increasing. This, in turn, will increase the rates of litter decomposition in forests of the humid tropics located at all altitudes. Increased T s would also result in greater mineralization rates for nutrients (i.e. N and P) that are fixed in soil minerals. The resulting increase in nutrient availability in the soil could support the increased nutrient demand required for CO 2 -induced stimulation of photosynthesis and biomass production. These changes in the soil environment could also modify its microbial population and their activity (O'Neill, 1994;Sadowsky & Schortemeyer, 1997). Because of the longer time scales involved, there could be many positive and negative feedbacks between these interacting processes.

POTENTIAL FOR CARBON SEQUESTRATION BY FORESTS IN THE HUMID TROPICS
The longevity and favourable conditions for plant growth of forests in the humid tropics allows significant potential for carbon sequestration in their biomass and soil in a future high C a environment (Dixon et al., 1994;Grace et al., 1995;Naburrs & Mohren, 1995;Lloyd & Farquhar, 1996;Chambers et al., 1998;Phillips et al., 1998;Schimel et al., 2001) and slow down the pace of climate change Clark, 2004a;Thompson et al., 2004). In fact, Phillips et al. (1998) showed that increased biomass in tropical forests during the last 40 years accounted for 40% of the missing carbon in the carbon balance of the entire world. Therefore, afforestation and prevention of deforestation have been identified as measures for mitigation of climate change (IPCC, 2007). However, there have been uncertainties on the capacity of tropical forests as carbon sinks (Pearce, 1999;Clark, 2002Clark, , 2004bHoughton, 2003Houghton, , 2005. Firstly, carbon sequestration capacity may be reduced when forest ecosystems approach the mature, climax stage (Pregitzer & Euskirchen, 2004) and when the ontogenetic development is accelerated by elevated C a (Hättenschwiler et al., 1997). Some experiments have shown that the CO 2 -induced growth stimulation of forests is transient and diminishes when trees transfer from juvenile to mature stage (McConnaughay et al., 1996;Tissue et al., 1997). Secondly, increasing deforestation and burning of biomass in the tropics could add significant amounts of CO 2 to the atmosphere. Moreover, on re-analyzing Phillips et al. (1998)'s data, Clark (2002) contends that the data do not indicate a significant biomass carbon sink in old-growth forests of the humid Neotropics.
Analyzing the results of four FACE experiments on temperate forest stands, Norby et al. (2005) showed a 23 ± 2% median increase in NPP in response to elevated C a up to 550 μmol mol -1 and concluded that this response is highly conserved across a broad range of productivity. Norby et al. (2005) (Houghton et al., 2001). The observation by Körner et al. (2005) that physiological responses (i.e. photosynthesis, leaf N and non-structural carbohydrate concentrations) to CO 2 enrichment of older and larger trees were similar to younger trees adds support to the view that carbon sequestration capacity is maintained in mature forests as well. In a review involving results from 49 studies on the impacts of climate change (i.e. increased C a and T a and changes in precipitation, incident solar radiation and various other atmospheric and soil factors) during the last 55 years on forest productivity, Boisvenue and Running (2006) showed an increasing trend in NPP in 37 of the studies. They further showed that the combination of changes in various climatic factors since the middle of the 20 th century has had a positive impact on forest productivity when water is not limiting. As water is rarely a limiting factor in forests of the humid tropics, this analysis confirms the predictions based on fundamental physiological processes. Four (Phillips et al., 1998;Malhi & Grace 2000;Baker et al., 2004;Lewis et al., 2004) out of the five tropical forest ecosystems considered in the study of Boisvenue and Running (2006) had shown increases in NPP or net biome production (NBP) in response to climate change that had occurred during the latter half of the 20 th century. In fact, Nemani et al. (2003) estimated that tropical ecosystems accounted for a major portion of the 6% increase in global NPP that occurred during the period from 1982-1999, with the Amazon accounting for 42%. This increased NPP in tropical forest ecosystems is attributed to increased C a and T a (Lewis et al., 2004) and decreased cloud cover (Wielicki et al., 2002;Wild et al., 2005) in these radiation-limited forest ecosystems (Graham et al., 2003).
However, despite the overall increasing trend in NPP in tropical forests during the last 50 years, lower tree growth rates have been recorded in years with higher temperatures (Clark & Clark, 1994;Clark et al., 2003;Clark, 2004b). These warmer years have coincided with the El Niño Southern Oscillation (ENSO), which is a major determinant of temperature and precipitation variation in the tropics (Nemani et al., 2003). Episodes of ENSO have also caused substantial tree mortality, especially larger trees, which store a greater proportion of carbon, due to drought and forest fires (Williamson et al., 2000;Clark, 2004b).
In addition to the NPP, the partitioning of CO 2induced increases of biomass also determines the capacity of a forest ecosystem for carbon sequestration and mitigation of climate change (Dufresne et al., 2002;Luo et al., 2003). Biomass partitioned to above-ground vegetation can be sequestered in stem wood while that partitioned to roots can ultimately be stored as soil C. Chambers et al. (1998) have shown that some of the trees in the Amazon that have lived well over a millennium have been capable of growth and carbon sequestration for over 1,400 years. Idso (1999) has cited data that suggest that even after 100 years of exposure to elevated C a , trees will still be removing more carbon from the air than they would be able to do if atmospheric C a remained at the current level. Similarly, the soil C that gets sequestered in the recalcitrant soil C pool has residence times of the same scale as that of wood C (Leavitt et al., 1994;Parton et al., 1994). Studies at the single species level show a clear preference for greater partitioning of CO 2induced biomass to roots . However, available experimental results and model predictions at the ecosystem level do not show any consistent preference in C sequestration towards either vegetation or soil. In the study by Norby et al. (2005) percentage NPP gain partitioned to stem wood ranged from 11% to 93%. In global productivity models (Dufresne et al., 2002), the fraction of C stored in vegetation has varied from 35% to 85%. On the other hand, Post and Kwon (2000) concluded that in the terrestrial ecosystems of the Northern Hemisphere, the potential for carbon sequestration in vegetation biomass is much greater than in soil. enrichment. Commonwealth Forestry Review 75: 39-47. Eamus D. & Jarvis P.G. (1989). The direct effects of 47.
increase in global atmospheric CO 2 concentration on natural and commercial temperate trees and forests.  (1996). Direct and 60.
Photosynthetic acclimation in trees to rising atmospheric