JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 107, NO. D24, 4808, doi:10.1029/2001JD001245, 2002
Impact of doubled CO2 on global-scale leaf area index and evapotranspiration: Conflicting stomatal conductance and LAI responses Laurent Kergoat,1 Se´bastien Lafont,2 Herve´ Douville,3 Be´atrice Berthelot,2 Ge´rard Dedieu,2 Serge Planton,3 and Jean-Franc¸ois Royer3 Received 22 August 2001; revised 26 April 2002; accepted 1 May 2002; published 31 December 2002.
[1] Current increase in atmospheric CO2 is expected to modify both climate and plant
function, thereby impacting plant structure and gas exchange. We investigate the effects of doubled CO2 on leaf area index (LAI) and evapotranspiration (ETR) using a global vegetation model for present-day and doubled-CO2 conditions. The model assumes that adaptation of plants to the local climate leads to an equilibrium LAI, which depends on resource availability (minimizing water stress, canopy carbon cost and self-shading). The model compares reasonably well with remote sensing estimates of LAI. Four doubled-CO2 simulations are designed to investigate the role of climate, CO2-induced stomatal closure, enhanced photosynthesis, and a combination of these effects. These simulations show that plant physiological responses to doubled CO2 are potentially more important than climate changes for LAI, and equally important for ETR. In addition, even the sign of the simulated changes in LAI and ETR varies with the assumptions on plant responsiveness to CO2. A reduction of stomatal conductance moderates or cancels the water losses caused by a warmer and drier climate, but photosynthesis stimulation counteracts this stomatal effect, especially in the mid-to-high latitudes, because of enhanced LAI. Experimental evidence of LAI and ETR response to CO2 has been reviewed and compared to the different simulations. On the basis of this confrontation we argue that plant response to CO2 doubling may have a relatively small net impact on global ETR and may cause a moderate increase of LAI. Tree stomata may be less responsive to CO2 than was previously assumed, and stimulated plant growth partly cancels the water savings caused by stomatal closure. Understanding the responses of plant canopies to CO2 is therefore critical for land surface hydrology in a CO2 rich INDEX TERMS: 1818 Hydrology: Evapotranspiration; 1851 Hydrology: Plant ecology; 1655 world. Global Change: Water cycles (1836); 3322 Meteorology and Atmospheric Dynamics: Land/atmosphere interactions; KEYWORDS: LAI, CO2, stomatal conductance, global, evapotranspiration Citation: Kergoat, L., S. Lafont, H. Douville, B. Berthelot, G. Dedieu, S. Planton, and J.-F. Royer, Impact of doubled CO2 on globalscale leaf area index and evapotranspiration: Conflicting stomatal conductance and LAI responses, J. Geophys. Res., 107(D24), 4808, doi:10.1029/2001JD001245, 2002.
1. Introduction [2] Early studies have long demonstrated the effects of carbon dioxide on plant physiology and gas exchange [e.g., Heath and Russell, 1954]. Numerous experiments with plants grown in elevated CO2 have shown an increase in leaf photosynthesis rate and often, but not always, a decrease of leaf stomatal conductance; see reviews by Bazzaz [1990], Field et al. [1995], Ko¨rner [1996], Drake et al. [1997], Morison [1998], Medlyn et al. [2001], but see 1 2 3
Laboratoire d’Ecologie Terrestre, CNRS, Toulouse, France. Centre d’Etude Spatiale de la Biosphe`re, Toulouse, France. Centre National de la Recherche Me´te´orologique, Toulouse, France.
Copyright 2002 by the American Geophysical Union. 0148-0227/02/2001JD001245$09.00
ACL
also Saxe et al. [1998], Curtis and Wang [1998]. Given the current rise of atmospheric CO2, it becomes important to understand the consequences that these effects may have on ecosystem function and on the water cycle. Indeed, water vapor fluxes over the continents are largely controlled by plant canopies. Pollard and Thomson [1995], and Henderson-Sellers et al. [1995] (hereinafter referred to as HS95) have shown that a decrease in stomatal conductance significantly impacts 2xCO2 climate simulations and thus has to be considered for the elevated CO2 climate predictions. Sellers et al. [1996a] (hereinafter referred to as S96) further suggested that the biosphere physiological response to CO2 might reinforce the climate warming. The stomatal closure induced by increased CO2 reduces the latent heat flux and favors the sensible heat flux over the forested continents [see also Cox et al., 1999]. The possible impact of CO2-
30 - 1
ACL
30 - 2 KERGOAT ET AL.: CO2 EFFECTS ON GLOBAL LAI AND EVAPOTRANSPIRATION
induced reduction in stomatal conductance on the water cycle and run-off has been highlighted by Idso and Brazel [1984], Ashton [1984], Wigley and Jones [1985], Neilson and Marks [1994] among others. However, large uncertainties remain. Different processes are expected to impact the ecosystem-scale water budget. The first process is a reduction of the leaf-scale transpiration due to a CO2-induced decrease in the leaf conductance for water vapor diffusion. Such a decrease was generally assumed in previous studies [e.g., HS95, S96], but recent works shed a different light on stomatal response, as will be detailed later. The second process concerns the response of canopy-scale latent fluxes. This response depends on the changes in ecosystem structure, and especially on the changes in canopy leaf area index (LAI, projected leaves area per ground area, m2/m2). Both enhanced plant growth and shifts in species composition within the ecosystem can provoke changes in LAI. These processes are particularly relevant to the decade-to-century timescale of the current CO2 increase. Such changes in ecosystem structure are particularly important for the water budget: if the plants benefit from better environmental conditions and greater resource availability, for example, water, CO2, LAI may increase significantly and this can enhance canopy conductance and canopy-scale ETR, in spite of a leaf-level stomatal closure. This balance between leaf-scale and canopy-scale effects is the so-called ‘‘compensatory response’’ to elevated CO2, highlighted by Field et al. [1995]. This effect is a potentially important, still largely uncertain, driver of surface climate and hydrology in a CO2 rich world. [3] At the global scale, Betts et al. [1997] (hereinafter referred to as B97) first demonstrated that ecosystem structural changes in a coupled climate-vegetation simulation could partly cancel the physiological effects on plants. They found that an increased LAI partially compensates the effect of stomatal closure on ETR. These results significantly differ from previous studies [HS95, S96]. Another simulation was carried out by Levis et al. [1999, 2000] (hereinafter referred to as L00). They also found a large impact of increased LAI on the energy and water budgets. Both B97 and L00 studies have convincingly highlighted the importance of LAI in vegetation-climate feedbacks, through changes in ETR, energy budget and precipitation recycling. However, they obtained drastically different simulations of global LAI changes. The lack of consensus between these studies [see also Pan et al., 1998] indicates that the mechanisms which cause changes in LAI and canopy evapotranspiration in global-scale 2xCO2 simulations, need to be discussed in detail and compared with experimental results. [4] In this study, we use a model, which simulates LAI [Kergoat, 1998], for present-day and doubled-CO2 climate conditions, along with different assumptions on plant canopy response to CO2. A dedicated simulation investigates the recent observations that tree stomata may be rather insensitive to CO2 doubling [Saxe et al., 1998; Curtis and Wang, 1998], as opposed to what was assumed in previous studies [HS95, S96, B97, L00]. Based on a series of simulations, we address the following questions: (1) How does the simulated LAI compare to satellite observations for present-day climate? (2) What is the relative importance of physiological versus climatic effect of CO2 on global-scale
LAI and evapotranspiration? (3) Are there compensations between the physiological and structural changes in the response of evapotranspiration to elevated CO2? The recent CO2-enrichment literature, including Free-Air CO2 Enrichment (FACE), natural CO2 sources, whole-ecosystem enclosures, is then reviewed to address the following issues: (4) How do the mechanisms leading to changes in the 2xCO2 simulations of LAI and evapotranspiration compare to canopy-scale experimental data? (5) Is it possible to define the likely global-scale pattern of plant canopies response to increasing CO2?
2. Model and Data Description 2.1. Vegetation Model [5] The vegetation model aims at estimating LAI and vegetation carbon and water budgets, from prescribed climate and biome distribution. The basic underlying assumption is that, on the long term, plant canopies develop until they reach a quasi-equilibrium, which depends on local resource availability [Woodward, 1987; Woodward et al., 1995; Neilson, 1995; Kergoat, 1998]. This balanced functioning is characterized by 3 criteria: The average water stress must be sustainable (this criterion is referred to as the ‘‘water stress criterion’’), the canopy lowest leaves must have a positive annual carbon balance (the ‘‘self-shading criterion’’), and the annual photosynthesis must cover the canopy construction and maintenance costs (the ‘‘carboncost criterion’’). In practice, for each gridcell, these three criteria are computed for increasing LAI values, and the model retains the highest LAI, which fulfills all of them. The first two criteria are described by Kergoat [1998] and are only briefly outlined below. Details on the carbon-cost criteria are given in Appendix A. 2.1.1. Water Stress Criterion [6] As evapotranspiration usually increases with LAI, the soil water deficit also increases with increasing LAI. By assumption, for a given climate, the LAI is limited to prevent critical soil water depletion [Grier and Running, 1977; Eagleson, 1982; Woodward, 1987]. The model takes into account the physiological and ecological adaptations of different plant functional types. It accounts for plant drought-avoidance and drought-resistance strategies, such as drought-deciduousness and deep rooting [Kergoat, 1998]. The water balance submodel is a two source PM equation. Soil evaporation, rainfall interception, transpiration, snowpack and freezing processes are simulated. 2.1.2. Self-Shading Criterion [7] Increasing canopy LAI attenuates the light, which is available at the bottom. As a result, the photosynthesis of the lowest leaf layer decreases, and may be exceeded by leaf respiration. This criterion is thought to limit LAI for dense canopies, for example, moist tropical forests [Alexandre, 1981]. The self-shading criterion uses a leaf photosynthesis and respiration submodel for the lowest leaf layer of the canopy [Kergoat, 1998]. 2.1.3. Carbon-Cost Criterion [8] This criterion corresponds to an ‘energy’ limitation of vegetation growth [Grier and Running, 1977; Gholz, 1982; Neilson, 1995]. It is an addition to the model used by Kergoat [1998], stating that the annual canopy carbon gain must cover the canopy maintenance and construction cost.
KERGOAT ET AL.: CO2 EFFECTS ON GLOBAL LAI AND EVAPOTRANSPIRATION
ACL
30 - 3
Table 1. Simulation Design Simulations
Climate
Leaf Level Assumption
CLIM RAD STOM FERT MAX STOM2 MAX2
Present-day, Leemans and Cramer ARPEGE:2xCO2 ARPEGE:2xCO2 ARPEGE:2xCO2 ARPEGE:2xCO2 ARPEGE:2xCO2+STOM iterated ARPEGE:2xCO2+MAX iterated
-
Canopy photosynthesis and net primary productivity saturate with increasing LAI, whereas the carbon requirement is approximately proportional to LAI. As a result, there is a physical limit to LAI development, which is reached when carbon supply equals carbon demand (Appendix A). This equilibrium strongly depends on the environmental conditions. For example, a short and cold growing season, typical of high latitudes, will drastically reduce annual photosynthesis and net productivity. The carbon-cost criterion will then produce a low LAI. Such an energy limitation also coincides with a nutrient constraint on plant growth. Indeed, high latitude ecosystems are often nutrient limited, and low soil temperature reduces nitrogen mineralization. In the model, warming is expected to stimulate photosynthesis in a cold climate, which is what is expected from both a direct effect of temperature on photosynthesis and a nutrientmediated effect on growth. Overall, the carbon-cost criterion captures the primary constraint of low temperature and solar radiation on LAI and productivity [Neilson, 1995; Haxeltine and Prentice, 1996]. [9] In order to estimate the equilibrium LAI from these three criteria, the model is run for five years, at the daily time step, using climate data interpolated from monthly climate inputs. It requires the distribution of 7 vegetation types: tundra, evergreen needle-leaved forest, evergreen broad-leaved forest, dry-deciduous forest, cold-deciduous forest, grasslands and crops, and Mediterranean vegetation or shrubland. Parameters, ranging from physiological characteristics, like maximum stomatal conductance, to ecological adaptations like drought resistance, are taken from the literature [Kergoat, 1998]. For deciduous types, LAI is set to 0 during a dormancy period, which is prescribed from satellite data [Moulin et al., 1997]. Details of the gas exchange equations can be found by Cayrol et al. [2000] and in the Appendix A. 2.2. Satellite Derived LAI and fPAR Data Sets [10] Satellites provide unique data sets to check global models of terrestrial ecosystems. The reflectance factors in the visible and near-infrared domain and their combination (NDVI) have been related to the fraction of PAR (photosynthetically active radiation) absorbed by plant canopies (fPAR), and to other surface parameters, like LAI and vegetation cover. The AVHRR global archive has been processed to produce different estimates of fPAR and LAI [Sellers et al., 1996b; Myneni et al., 1997a; Tian et al., 2000]. Here, modeled LAI and fPAR are compared to their counterparts derived from the AVHRR-FASIR 1987 data set [Sellers et al., 1996b, ISLSCP CD-ROM] and from the MODIS-LASUR data set. This data set is based on the MODIS algorithm [Tian et al., 2000] prototyped with the LASUR surface reflectances [Berthelot et al.,
Stomatal closure Enhanced photosynthesis Stomatal closure and enhanced photosynthesis Stomatal closure Stomatal closure and enhanced photosynthesis
1997, LASUR CD-ROM]. It benefits from atmospheric corrections [Ouaidrari et al., 1997] and state-of-the art retrieval algorithm [Tian et al., 2000]. fPAR estimates are more accurate than satellite derived LAI [Myneni et al., 1997a], because of a reduced sensitivity to signal noises.
3. Simulation Design [11] Given the variability of plant response to CO2, we use a number of different assumptions on leaf-level response to CO2. Besides a control simulation, four simulations investigate the model sensitivity, whereas two additional simulations address the impact of atmospheric feedback on global LAI (Table 1). 3.1. Simulation CLIM: Control Simulation [12] The CLIM simulation estimates the LAI in equilibrium with present-day climate (350 ppmv) using the standard input of the IGBP model intercomparison [Cramer et al., 1999]. It includes the 0.5 0.5 monthly climatology of Leemans and Cramer [1991] (updated, Cramer, personal communication, 1995), and the Zobler soil map. For consistency reasons, the biome distribution map is the one used by the ARPEGE climate model [Wilson and Henderson-Sellers, 1985]. 3.2. Simulation RAD: Climate Only Simulation [13] The second simulation, RAD, uses a 2xCO2 climate (700 ppmv) that accounts solely for CO2 radiative effect. Monthly anomalies obtained from the ARPEGE climate simulation have been super-imposed to the climatology: additional biases for air temperature, proportional biases for precipitation, vapor pressure and incoming short-wave monthly fields. The doubled-CO2 climate simulations are performed with the ARPEGE-Climat model of Me´te´oFrance [De´que´ et al., 1994; Douville et al., 2000] forced by sea surface temperature and sea-ice anomalies obtained from the ‘‘GHG’’ transient coupled experiment of the Hadley Centre [Mitchell and Johns, 1997]. With a 2C global warming and a 6% precipitation increase, the ARPEGE model shows a moderate response to CO2 doubling, which is within the range of similar studies, but closer to the less responsive models [Houghton et al., 2001]. 3.3. Simulation STOM: Stomatal Response [14] The third simulation, STOM, uses the same climate as RAD, but accounts for a decrease of stomatal conductance in response to elevated CO2. We choose to investigate leaf-level responses of stomata and photosynthesis to elevated CO2 independently. In fact, it is still unclear whether a single set of prognostic equations, like coupled photosynthesis-stomata models, is able to represent the variety of
ACL
30 - 4 KERGOAT ET AL.: CO2 EFFECTS ON GLOBAL LAI AND EVAPOTRANSPIRATION
contrasted experimental results. Calibration of key-parameter is often necessary (e.g., nitrogen content, Bunce [2000], Luo et al. [1999], Medlyn et al. [2001]). Here, a maximum CO2 impact on plant stomata [Field et al., 1995; Drake et al., 1997; Morison, 1998] is obtained by increasing the leaflevel minimum stomatal resistance (Appendix A). Medlyn et al. [2001] have shown that such a framework successfully reproduces the CO2 effect on stomatal conductance for a set of enrichment experiments. For consistency reasons, we apply the same resistance bias as Douville et al. [2000]. This bias is taken from a 2xCO2 simulation with the Hadley Center GCM [Cox et al., 1999] (hereinafter referred to as C99). It averages +51%. Simulations using a uniform +50% response, (not shown) lead to similar results. Compared to the results of elevated CO2 field experiments, this is a strong leaf-level physiological response. It is closer to the change prescribed by Sellers et al. [1996a] than to the +100% increase prescribed by Pollard and Thomson [1995] and Henderson-Sellers et al. [1995]. 3.4. Simulation FERT: CO2 Fertilization [15] In contrast, the FERT simulation assumes unchanged stomatal conductance, but includes a CO2 fertilization effect on photosynthesis. This simulation is motivated by some recent studies, which suggest that tree stomata do not close or close slightly in elevated CO2 [Curtis and Wang, 1998; Saxe et al., 1998; Medlyn et al., 2001]. This possibility has been overlooked in previous global modeling studies. The CO2 fertilization effect is applied to the leaf-level photosynthesis. Both the maximum photosynthesis rate and the quantum yield are increased by 25% (Appendix A). Such a leaf-level photosynthesis response is commonly observed, at least for short-term CO2 enrichment [Drake et al., 1997]. Many experiments produce even larger stimulation of the assimilation rate (40% to 60%), but the biomass increases usually less and acclimation often reduces CO2 fertilization. In the model, carbon allocation and annual cost of leaves are not modified by elevated CO2. As a result, changes in carbon available for leaf biomass parallel changes in photosynthesis. Therefore, a 25% increase is reasonable. Since phenological changes are also likely to occur, the growing season is shifted according to the climate: Temperature and rainfall thresholds associated with phenological changes are inferred by matching present-day phenology prescribed from satellite data and climate (CLIM). In simulation FERT, leaf emergence and leaf shedding are triggered by the same thresholds, and follow the temporal shifts in temperature seasonal course for cold-deciduous biomes, and the temporal shift in rainfall for dry-deciduous biomes. 3.5. Simulation MAX: Stomatal Response and Fertilization Effect [16] The MAX simulation combines CO2-induced stomatal closure, enhancement of photosynthesis rate and phenology adjustment. The FERT and MAX simulations aim at capturing maximum changes in plant carbon economy in elevated CO2. 3.6. Feedback Simulations [17] In addition, the importance of feedbacks between vegetation and the atmosphere is assessed with the STOM2 and MAX2 simulations. These simulations correspond to an
asynchronous coupling between vegetation and the atmosphere, for the STOM and MAX scenarios. The stomatal resistance and LAI anomalies from simulations STOM and MAX are imposed to the ARPEGE climate model, through changes in the land surface scheme. This in turn provides new 2xCO2 climates. These climatic scenarios are fully described by Douville et al. [2000]. They are used here as a forcing for the LAI model, which leads to the STOM2 and MAX2 simulations. The comparison with STOM and MAX provides insights on the feedback in the vegetation/climate loop.
4. Results 4.1. Control Simulation [18] The fPAR and LAI maps simulated for present-day climate are shown in Figure 1a and 1c. The carbon-cost criterion limits LAI in high latitude and mountainous biomes such as tundra and most boreal forests. The canopy self-shading effect plays the major role in limiting the LAI of tropical forests and some temperate forests and grasslands. As expected, the water stress effect is dominant in arid areas, including both hot and cold continental climate zones. In some cases, like for most of the temperate forests areas, the three criteria often give very similar LAI values, suggesting that the 3 limitations are almost balanced. 4.1.1. Evaluation With AVHRR Satellite Data [19] Figure 1 shows the fPAR obtained by the model and derived from the AVHRR for year 1987 (FASIR data set), along with the LAI from the model and MODIS-LASUR data set for July. Pixel-based scatterplots for fPAR and LAI are displayed on Figure 2. [20] The agreement between modeled fPAR and remotely sensed fPAR is good (Figures 1a 1b, and Figure 2). This indicates that the model mechanisms capture the global patterns of vegetation development and light-harvest. For LAI, the agreement is not as good but still reasonable, the main discrepancies being over the Amazon Basin and the high latitude forests (Figures 1c, 1d and 2b, 2c). The LAI simulation successfully reproduces aridity gradients, e.g., East-West in North America, North-South in central Asia, and also temperature gradients (high latitudes, mountains). The model provides reasonable LAI values for the moist tropical and temperate forests. The disagreement between modeled and satellite-derived LAI over the Amazon basin is likely to be caused by persistent cloud cover that lower the NDVI signal. The discrepancy for the high latitude forest may indicate a model bias. However, there is growing evidence that the FASIR satellite-derived LAI product may be overestimated for the high-latitude forests. Recent remotely sensed LAI maps for the Canada (Chen, unpublished data) show significantly lower values, which are closer to the modeled LAI in Figure 1c. The MODISLASUR data set also gives lower LAI values for both Canada and Eurasia, and the agreement is better for Figure 2c than Figure 2b. The same discrepancy between modeled and remotely sensed LAI was found by Dickinson et al. [1998] with a different LAI model. Besides, field measurements found LAI values of 1.5 and 2.5 for some Siberian mature pines and larches stands respectively [Kelliher et al., 1997, 1998]. We acknowledge, however, that our LAI model might be biased toward
KERGOAT ET AL.: CO2 EFFECTS ON GLOBAL LAI AND EVAPOTRANSPIRATION
ACL
30 - 5
Figure 1. Comparison of the LAI model results with remote sensing estimates. Maximum fPAR (a) simulated for present-day climate and (b) retrieved from the FASIR/AVHRR data set. (c) July maximum LAI ( projected, m2/m2) simulated for present-day climate and (d) retrieved from the LASUR/AVHRR data set processed with the MODIS prototype algorithm. See color version of this figure at back of this issue.
low values for high latitude forests. Since we are investigating the response of vegetation to a change in resource availability, it is important to note that the agreement for fPAR is good in the high latitudes. Indeed, fPAR represents the light resource. Model and observations agree that there is a potential for LAI development and increased light capture in the high latitudes. Considering the good agreement for fPAR and the caveat in remotely sensed LAI, we conclude that the model predicts a reasonably accurate global LAI. The level of agreement between the model and observations suggests that the model might have the same sensitivity to climate change as real vegetation. 4.2. Elevated CO2 4.2.1. Response of LAI [21] In experiment RAD, where only the CO2 radiative effect is considered (Figure 3a), the patterns of LAI anomalies show both decreases and increases. As expected, there is an increase of LAI at high latitudes, on the order of 0.3 m2/m2 (Table 2). This increase is related to the air warming, which favors photosynthesis and therefore enhances the annual productivity. As a result, LAI values that were limited by the carbon-cost criterion in CLIM reach larger
values in RAD. A second point to note is the decrease of LAI over Southern Eurasia and the western continental US, which is caused by an alteration of the water balance. Indeed, the atmospheric demand for water vapor is larger in RAD, leading to high water losses and increased water stress. As a consequence, the equilibrium hypothesis adjusts LAI to lower values for these areas so that the water stress remains sustainable. The tropical zones display both positive and negative LAI anomalies, directly related to local changes in the precipitation field. [22] When the stomatal closure is considered, experiment STOM (Figure 3b), most of the drought-induced LAI decrease is compensated for, and LAI is turned back to its CLIM values or above. This result means that stomatal closure induces a reduction of water losses, which is large enough to significantly improve the yearlong water availability, in contrast to the RAD scenario. This affects plant canopies that are limited by the water criterion, and, to a lesser extent, plant canopies that are limited by the carboncost criterion where water availability is not optimal. [23] The FERT simulation (Figure 3c) is closer to RAD, with both increase and decrease of LAI, with the exception that high-latitude LAI increase is much stronger, commonly reaching +1 m2/m2. Such stimulation is explained by an
ACL
30 - 6 KERGOAT ET AL.: CO2 EFFECTS ON GLOBAL LAI AND EVAPOTRANSPIRATION
Figure 2. (a) Scatterplot of simulated fPAR versus satellite derived fPAR from the FASIR data set. The data points have been pooled into bins for clarity. (b) Same but for July satellite derived FASIR-LAI. (c) Same but for LAI derived from LASUR data set and MODIS prototype algorithm.
enhanced carbon balance due to warmer temperature and increased photosynthesis, which alleviates the carbon-cost criterion. The shift in leaf phenology, corresponding to an earlier spring warming and a later fall cooling, also contributes to enhance the carbon budget. This shift averages 7 days both in spring and fall for the temperate and boreal deciduous ecosystems, and is nearly nonexistent for the tropical biomes. In addition, enhanced photosynthesis has an impact on canopies limited by the self-shading criterion (e.g., moist Tropics), because it decreases the light compensation point of the lowest leaves. A new equilibrium is reached for a slightly higher LAI value. As a result, even dense canopies display positive anomalies (e.g., Amazon basin). Conversely, the areas limited by the water criterion are almost identical in the RAD and FERT experiments. [24] When both photosynthesis enhancement and stomatal closure are considered, experiment MAX (Figure 3d), there is an overall increase of LAI. The LAI anomaly map approximates the sum of the STOM and FERT maps, but the processes are not strictly additive (Table 2). As in simulation FERT, there is a positive feedback between LAI and photosynthesis, in areas where the carbon-cost criterion is the limiting factor. Indeed, when photosynthesis is enhanced, the LAI also increases, which in turn allows additional photosynthesis. Table 2, for instance, shows that a 25% increase of the photosynthesis rate gives a LAI increase of 41%, from the STOM to the MAX simulations of tundra. This positive feedback, pointed out by Lloyd and Farquhar [1996], is important where the fPAR is low. 4.2.2. Response of Evapotranspiration and Soil Water [25] In response to the atmosphere warming (simulation RAD), there is a general increase of the latent heat flux wherever soil moisture is available (Figure 4a). This increase reaches 10 to 15% ( 9 mm/mo) of the CLIM evapotranspiration during the tropical rain season and the boreal summer. Some spatial variability in the Tropics is related to variations in the precipitation. As a result, the soil water in the root zone decreases significantly, mostly in the Tropics (Figure 5a). However, when a CO2-induced stomatal closure is taken into account, (simulation STOM Figure 4b), the latent heat flux anomaly is reversed, in spite of the higher atmospheric demand. This is especially true for fully developed canopies, with high LAI values and a small contribution of soil evaporation. Sellers et al. [1996a] and Betts et al. [1997] have underlined this feature. Compared to RAD, the soil water anomaly is reversed in STOM, and the difference with present-day climate becomes positive (Figure 5b). Such a pattern, however, is completely changed if CO2 is assumed to enhance photosynthesis without affecting stomatal conductance (simulation FERT). In this case, ETR is considerably enhanced compared to present-day simulation (Figure 4c). For the northern mid-to-high latitudes, ETR is significantly larger than in the climate-only RAD simulation, and the soil moisture is depleted in the summer (Figure 5c). This is a straightforward impact of increased LAI and longer growing season. In FERT, the water cycle in the Tropics is similar to the climate-only simulation RAD. When both photosynthesis and stomatal changes are considered (MAX), the Tropics behave like in the STOM simulation; Stomatal closure has a larger impact than LAI increase and this maintains ETR to present-day level, despite increased
KERGOAT ET AL.: CO2 EFFECTS ON GLOBAL LAI AND EVAPOTRANSPIRATION
30 - 7
ACL
Figure 3. Effect of a doubling of CO2 concentration on Leaf Area Index (m2/m2). (a) LAI anomaly between the RAD (climate only) and CLIM (present-day) simulations. (b) Anomaly between STOM (maximum stomatal response) and CLIM. (c) Anomaly between FERT (maximum photosynthesis response) and CLIM. (d) Anomaly between MAX (stomatal and photosynthesis response) and CLIM. See color version of this figure at back of this issue. atmospheric demand. In the Northern latitudes, the LAI effect prevails, especially during the spring (Figure 4d). The overall soil moisture is close to present-day values (Figure 5d). From Figures 4 and 5, it appears that even the sign of the effect of CO2 on ETR and soil moisture is uncertain and depends on plant response to CO2. 4.3. Atmospheric Feedback [26] The RAD, STOM, FERT and MAX simulations are performed with the same climate forcing. Since these simulations lead to large differences in LAI and canopy gas exchange, feedbacks to the atmosphere may be expected.
Douville et al. [2000] used the leaf conductance and LAI of the STOM and MAX simulations to perform 2xCO2 climate experiments with the ARPEGE model. They found that the climate response to CO2-induced changes in the vegetation was globally a second-order effect compared to the radiative effect, but regional impacts proved to be important. These ‘‘new’’ 2xCO2 climatic fields are used here to simulate equilibrium LAI an ETR. The resulting LAI simulations, referred to as STOM2 and MAX2, therefore represent the second iteration of an asynchronous coupling between the atmosphere and the vegetation systems. As shown in Table 3, biome-averaged LAI values in STOM and STOM2 are
Table 2. Biome-Averaged Leaf Area Index (L, m2/m2), Annual Evapotranspiration (E, mm/yr), and Available Water in the Root Zone (W, mm) for the CLIM, RAD, STOM, FERT, and MAX Simulations Needle-Leaved Evergreen Forest
Tundra CLIM RAD STOM FERT MAX
Cold Deciduous Forest
Dry Deciduous Forest
Broad-Leaved Evergreen Forest
Mediterranean Vegetation/ Shrubland
Grass/Crops
L
E
W
L
E
W
L
E
W
L
E
W
L
E
W
L
E
W
L
E
W
1.4 1.6 1.7 2.3 2.4
97 121 106 141 126
35 34 35 34 34
2.9 3.2 3.6 3.9 4.4
318 381 319 411 351
64 63 68 62 66
2.8 2.9 3.0 3.8 4.0
389 429 372 469 408
66 66 69 64 68
3.5 3.1 4.2 3.3 4.7
887 876 846 878 856
127 122 136 121 132
4.9 4.9 5.2 5.3 5.6
1159 1258 1100 1262 1119
250 242 255 241 253
0.8 0.8 1.2 0.8 1.2
237 246 253 246 253
94 94 93 94 93
2.5 2.5 2.7 3.2 3.7
883 933 891 956 923
48 48 49 47 48
ACL
30 - 8 KERGOAT ET AL.: CO2 EFFECTS ON GLOBAL LAI AND EVAPOTRANSPIRATION
Figure 4. Latitude versus time maps of monthly evapotranspiration anomaly (mm/mo). (a) RADCLIM. (b) STOM-CLIM. (c) FERT-CLIM. (d) MAX-CLIM. nearly the same. The reduction of latent heat flux in STOM produces a small additional warming [Douville et al., 2000] which does not affect the equilibrium LAI so much, and the differences between STOM and STOM2 are small in comparison to the differences between the RAD, STOM, FERT, and MAX scenarios themselves. This is also true for the MAX and MAX2 simulations. Overall, at the biome level, the differences between the first and second steps of the coupling are smaller than 0.1 m2/m2, with only the grass/crop biome showing a 0.2 m2/m2 increase.
[27] Ecosystem averages (Table 3) tend to mask some regional patterns. Figure 6 displays the strongest regional feedbacks found over 10 10 areas - the Amazon basin, North-Western Eurasia - along with an example of a weak response over North-Eastern Eurasia. [28] Over the Amazon basin, the STOM2 and MAX2 simulations show an enhancement of ETR and nearly no change in LAI (Figure 6a). This is due to the slightly enhanced air temperature and atmospheric water demand simulated by ARPEGE forced by the STOM and MAX
KERGOAT ET AL.: CO2 EFFECTS ON GLOBAL LAI AND EVAPOTRANSPIRATION
ACL
30 - 9
Figure 5. Latitude versus time maps of monthly available water in the root zone anomaly (mm). (a) RAD-CLIM. (b) STOM-CLIM. (c) FERT-CLIM. (d) MAX-CLIM. surface changes [Douville et al., 2000]. Over North-Western Eurasia, the atmospheric feedback-loop stimulates LAI and ETR, as expected from a slightly warmer climate (Figure 6b). However, over North Eastern Eurasia, these effects are much smaller and even reversed (Figure 6c). Such a heterogeneous response contributes to make biome-averaged LAI very similar for the STOM/STOM2 and MAX/MAX2 simulations. In most regions, ETR is slightly increased in STOM2 and MAX2 compared to STOM and MAX respectively, and the feedbacks are therefore mostly negative. This suggests that results from STOM and MAX are robust and not
underestimated. Atmospheric feedbacks, which may depend to some extent on the atmospheric model, proved to be second order effects, as far as plant structure is concerned. This study, as well as Levis et al. [2000], points out that the basic assumptions on plant response to CO2 are preeminent in biosphere/climate coupling [Foley et al., 1998].
5. Discussion [29] Given such a large uncertainty in LAI and ETR response to CO2, we have reviewed the literature of CO2
ACL
30 - 10 KERGOAT ET AL.: CO2 EFFECTS ON GLOBAL LAI AND EVAPOTRANSPIRATION
Table 3. Biome-Averaged Leaf Area Index (L, m2/m2) for the STOM and MAX Simulations and Their Iterations STOM2 and MAX2
CLIM STOM STOM2 MAX MAX2
Tundra
Needle-Leaved Evergreen Forest
Cold Deciduous Forest
Dry Deciduous Forest
Broad-Leaved Evergreen Forest
Mediterranean Vegetation / Shrubland
L
L
L
L
L
L
L
1.4 1.7 1.7 2.4 2.4
2.9 3.6 3.6 4.4 4.5
2.8 3.0 3.1 4.0 4.0
3.5 4.2 4.3 4.7 4.7
4.9 5.2 5.2 5.6 5.6
0.8 1.2 1.1 1.2 1.2
2.5 2.7 2.7 3.7 3.5
enrichment experiences, looking for effects of physiology and structure on plant canopies and ETR. This section presents a discussion of these experimental results, with a focus on the mechanisms that lead to changes in LAI and ETR in our global simulations. We also attempt to define a ‘‘likely’’ response for the different ecosystems. 5.1. Self Shading Criterion [30] According to our simulations, CO2 fertilization results in a lower light compensation point for the lowest leaf layer and therefore produces an increase in LAI, where this criterion is limiting. In line with this mechanism, several experiments report an increased growth in forest-floor plants in response to elevated CO2, for both temperate and tropical forests [Osborne et al., 1997; Ha¨ttenschwiler and Ko¨rner, 1996; Wu¨rth et al., 1998; Naumburg and Ellsworth, 2000]. Close examinations have demonstrated that plant carbon balance was improved in typical understory light regimes [Naumburg and Ellsworth, 2000]. Fewer studies have considered a whole dense canopy growing in near natural conditions. So far, recent results from FACE facilities point toward an insignificant increase or an uncertain trend of LAI for dense forest canopies. Wullschleger and Norby [2001] did not measure any significant effect of CO2 on LAI for mature sweetgum plantations (LAI = 6). The Duke forest FACE experiment led to uncertain changes in LAI (Lichter et al. [2000] do not discard a small increase), with understory saplings showing stimulated photosynthesis [DeLucia and Thomas, 2000]. Unfortunately, such tree plantations, where the dominant strata is monospecific, often even-aged, and species diversity is usually low, may overemphasize the individual response of the dominant trees. Multispecific and multistrata canopies have a greater plasticity [Reich et al., 2001] and might show a stronger LAI response. Overall, although experiments with mature and multispecies forest ecosystems are still scarce, a moderate increase of LAI is likely to affect dense canopies growing in favorable environments. This is compatible with the moderate increase in LAI (10%) simulated for tropical dense forests and some temperate closed canopies in the FERT and MAX simulations (Figures 3c, 3d, Table 2). [31] As a consequence of such a weak LAI response, the impact of LAI on ETR is expected to be small for dense canopies limited by the self-shading criteria. Canopy conductance is not sensitive to changes in LAI for high LAI values [Schulze et al., 1994]. Therefore, ETR will mainly reflect the changes in leaf-level stomatal conductance. In contradiction with what was commonly assumed in several global-scale studies, there is growing evidence that tree stomatal conductance may not respond strongly to elevated
Grass/Crops
CO2. More precisely, recent experimental results suggest than needle-leaved trees are less responsive than broadleaved deciduous trees, which are less responsive than grasses. Even more important, mature trees seem to be less responsive than saplings [Saxe et al., 1998; Medlyn et al., 2001]. Conductance decrease on the order of 0 –10% is commonly found in conifers, whereas 5 – 25% and 20– 40% are often reported for broad-leaved deciduous trees [Medlyn et al., 2001; Saxe et al., 1998] and herbaceous plants respectively. Therefore, there is growing suspicion that dense forests may exhibit a moderate or hardly significant decrease in canopy conductance in response to CO2. Ellsworth [1999] found that elevated CO2 had no significant impact on canopy evapotranspiration of the Duke forest loblolly pines FACE. Stomatal conductance and soil moisture were unchanged by CO2 treatment. For the sweetgum FACE, Wullschleger and Norby [2001] estimated a 12% reduction in stand transpiration and did not notice any LAI compensation effect. These two experiments point toward a relatively small impact of CO2 on dense temperate forests ETR. So far, there is no evidence that tropical trees stomatal response would differ from those of temperate broad-leaved deciduous and evergreen trees [Lovelock et al., 1999; Ko¨rner, 1998]. We therefore argue that dense forests response to CO2 may consist of a moderate canopy conductance decrease, ranging from 0 to 15%, resulting from moderate stomatal responsiveness and minor increase in LAI. 5.2. Carbon-Cost Criterion [32] In our simulations, boreal and arctic ecosystems are especially constrained by the carbon cost criterion, which also colimits some temperate ecosystems. Experiments on these systems have investigated the effects of both temperature and CO2 increase. There is ample evidence that productivity is enhanced by experimental warming at ambient CO2 (e.g., Chapin et al. [1995], ITEX experiment Henry and Molau [1997]). Such results are in line with the conclusions of many modeling studies, including this one (climate only simulation RAD, Figure 3a). As far as CO2 is concerned, Oechel et al. [1994] have shown that warming was able to sustain CO2 fertilization in tundra, which otherwise tends to decrease with time. They stated that warming stimulates nitrogen mineralization, allowing enhanced plant growth. Hoosbeeek et al. [2001] found no significant effect of CO2 on plant biomass for nutrient-poor bogs. Cook et al. [1998] found that grasses growing near CO2 vents in Iceland show a significant, long-term downregulation of photosynthesis and decreased LAI. According to Saxe et al. [1998] review, downregulation in trees is rarely large enough to offset the photosynthetic gains. The
KERGOAT ET AL.: CO2 EFFECTS ON GLOBAL LAI AND EVAPOTRANSPIRATION
CLIMEX experiment has exposed a small catchment of undisturbed boreal vegetation to CO2 enrichment and climate warming [Beerling, 1999]. No effect of CO2 on water runoff was noted, a fact attributed to a small LAI increase fully compensating the small reduction in stomatal con-
ACL
30 - 11
ductance. Taken together, these results point toward a clear positive impact of temperature and a weak CO2 fertilization effect, reduced in nutrient-poor habitats. Synergy between warming and CO2 causes larger stimulation of plant productivity and LAI. Whether plant carbon budgets can improve and result in denser canopies seems to be closely related to the future of nutrient availability. Unfortunately, the potential role of nutrients in a CO2 rich world is still a matter of hot debate; see discussion by Gifford [1994], McGuire et al. [1995], Lloyd and Farquhar [1996], and Schimel [1998] among others. Nitrogen (N) use efficiency is sometimes considered to be an absolute constraint on growth, whereas others state that the nitrogen cycle could ‘‘follow’’ the carbon cycle on the long term. Biotic N fixation, atmospheric N deposition and temperature-stimulated decomposition rates [Rustad et al., 2001] favor N availability, but the low N content of plants grown in elevated CO2 may provide a negative feedback to N availability on the long term. Interestingly, several experiments have shown an increase of the fraction of N-fixing plants [e.g., Warwick et al., 1998, Zanetti et al., 1997]. Moreover, plant carbon balance, as well as the whole nutrient cycle, will benefit from a shift in phenology leading to earlier springs and delayed falls. Such a modification of plant phenology has already been recorded in the temperate and boreal zones, and attributed to the rise in temperature observed during the last decades [Menzel and Fabian, 1999; Myneni et al., 1997b]. For these reasons, we suggest that the likely response of LAI to warming and CO2 doubling is a moderate but significant increase at high latitudes and high elevations. Such a LAI response is consistent with, and possibly slightly weaker than our FERT and MAX simulations, which lead to an LAI increase on the order of 1 m2/m2 (Figures 3c, 3d, Table 2). Furthermore, given the current uncertainty on tree stomatal response to CO2 (especially for old conifers), it is well possible that the CO2 net effect at boreal latitudes will be an increase rather than a decrease in ETR, as illustrated by simulation FERT (Figure 4c). 5.3. Water Stress Criterion [33] Interactions between CO2, water and plant growth have been documented by a fair number of field experiments, dealing with different types of grasslands, rain-fed crops, and, marginally, with shrubland and dry woodland ecosystems. For herbaceous canopies, elevated CO2 is generally found to decrease evapotranspiration rate, to enhance soil moisture, and to increase plant growth and LAI. Elevated CO2 tends to promote a significant but incomplete LAI compensation effect, with either a larger maximum LAI or an increase in late-season LAI caused by the delay of soil
Figure 6. (opposite) Atmospheric feedbacks on regional LAI and ETR for 10 by 10 degrees areas located in (a) Amazonia, (b) North-West Eurasia, and (c) North-East Eurasia. Stomatal closure in STOM and MAX reduces ETR, increases atmospheric demand for water vapor, which in turn causes ETR to increase in simulation STOM2 and MAX2, for Amazonia and North-West Eurasia (maximum regional responses) but not for North-East Eurasia. LAI is slightly or not affected by atmospheric feedbacks.
ACL
30 - 12 KERGOAT ET AL.: CO2 EFFECTS ON GLOBAL LAI AND EVAPOTRANSPIRATION
Figure 7. Effect of doubled CO2 on LAI (m2/m2) and ETR (mm/day). Differences between global LAI and ETR from STOM, FERT, MAX simulations and from RAD (climate only) simulation are displayed, along with the results of some previous studies. HS95 is for Henderson-Sellers et al. [1995], S96a and S96b are for Sellers et al. [1996a], C99 is for Cox et al. [1999], L00a and L00b (B97a and B97b, respectively) are for simulations with fixed and dynamical canopies by Levis et al. [2000] and Betts et al. [1997], respectively. The shaded zone figures a tentative response of plant canopies to doubled CO2 accounting for a possibly low response of tree stomata (see text). moisture depletion (C3 annuals Field et al. [1997], C4/C3 tallgrass prairie Owensby et al. [1999], C4 shortgrass prairie Morgan et al. [2001], calcareous grassland Niklaus et al. [1998], Mediterranean community Nijs et al. [2000]). Site fertility may dampen these effects [Field et al., 1997], and there is a consensus that water relations drive plant growth more than CO2 fertilization per se. Experiments with crops usually follow this pattern, with a significant species-specific variability. Stomatal closure is usually strong (30– 50% in doubled CO2), and the compensation by LAI growth varies from nonexistent to more than complete, depending on the crop. An increase in ETR, resulting from a strong LAI enhancement has been reported for sunflower canopies [Hui et al., 2001], an almost exact compensation effect has been noted for cotton in a FACE [Kimball et al., 1994], whereas partial compensation has often been found, for instance for wheat [Kimball et al., 1999]. Much less is known for shrublands and dry forests. Moderate stomatal conductance decrease (10% to 20%) has been reported (Nevada FACE facility Pataki et al. [2000] and Nowak et al. [2001], natural CO2 springs Bettarini et al. [1998]). [34] Overall, for water limited herbaceous systems, experimental results support the assumption of the STOM and MAX simulations, which suggest that plant water savings can alleviate the increased atmospheric demand for water vapor (climatic effect of CO2) and even slightly increase growth and LAI. For shrubland and dry woodlands,
LAI response could fall between FERT (slight decrease, Figure 3c) and STOM (slight increase, Figure 3b). Most often, annual ETR does not change much, because of LAI partial compensation and delayed onset of drought. Changes in ETR seasonality are expected, though. According to our model, water limited LAI is rather insensitive to photosynthesis fertilization, as the MAX and STOM simulations are similar for these areas. This is consistent with the results of Ha¨ttenschwiler et al. [1997] (for Quercus ilex) and Morgan et al. [2001] (C3 grassland). Water limited ecosystems show a strong decoupling between leaf-level response (stomatal conductance and photosynthesis rate) and canopy scale response, caused by the dynamical adjustment of LAI to soil moisture. 5.4. Global-Scale Response [35] When the different plant types and climate zones are considered, this review points toward a moderate increase of global LAI and a relatively small decrease of global annual ETR caused by vegetation response to a doubling of CO2 concentration. The main causes are a moderate response of tree stomata to CO2 and a compensation by stimulated growth (grasslands, crops and boreal zones). Shifts in phenology and soil moisture seasonality also contribute to this result. A tentative global response of LAI and ETR is illustrated by the shaded area in Figure 7. As large areas may be closer to the FERT or CLIM simulations than to the
KERGOAT ET AL.: CO2 EFFECTS ON GLOBAL LAI AND EVAPOTRANSPIRATION
STOM and MAX simulations (boreal zone, tropical forests), the STOM simulation has been replaced by the middle-point of STOM and CLIM, and the MAX simulation has been replaced by the middle-point of MAX and FERT. Other studies, with either fixed [HS95, S96, C99, B97a, L00a] or dynamical LAI [B97b, L00b], are also displayed for comparison. Figure 7 clearly illustrates the impact of LAI adjustment on ETR, since studies allowing for LAI changes show a weak negative or even positive effect of plant response to CO2 on ETR. [36] The compensating role of LAI found in this study is in line with the conclusion of Betts et al. [1997] and Levis et al. [2000], but it is much closer to the former than to the latter. In fact, our likely response of LAI and ETR agrees fairly well with Betts et al. [1997] estimates but differs noticeably from the strong LAI increase of Levis et al. [2000], especially in the tropics (+4.5 LAI averaged over 10S-10N). Whether this is a consequence of differences in modeling, or in the climate-coupling procedure is unclear [Levis et al., 2000]. Differences in LAI response to CO2 are larger and more difficult to analyze than differences in stomatal response. One can note that Betts et al. [1997] vegetation model relies on ‘‘ecosystem-physiology’’ rules to determine LAI [Woodward, 1987; Woodward et al., 1995], as does our model. Such an approach has the ability to decouple the leaf and canopy scale responses [Ko¨rner, 1996], which results in moderate response to elevated CO2.
6. Concluding Remarks [37] In this study, we analyzed LAI and evapotranspiration response to a doubling of CO2 concentration using a vegetation model. The respective roles of leaf-level stomatal conductance decrease and photosynthesis enhancement were investigated with the help of several simulations. We showed that evapotranspiration is intimately tied to plant response to elevated CO2, and that uncertainty on plant response could either lead to a significant decrease or a small increase of evapotranspiration, compared to a situation where only the effect of CO2 on climate is accounted for. However, after reviewing the literature on CO2 enrichment experiments, we conclude that plant response to CO2 doubling may have a relatively small net impact on global evapotranspiration, and may provoke a moderate, but locally more important, increase of LAI. This conclusion is mainly based on the fact that (1) tree stomata may be less responsive to CO2 than was previously assumed, (2) stimulated plant growth and phenology adjustment have the potential to partly cancel the water savings caused by stomatal closure in elevated CO2. Although plant types significantly differ in their response to CO2, as do the mechanisms controlling LAI in different climatic conditions, moderate LAI increase and small or even negligible cumulative ETR decrease are remarkably common features of experimental studies. The amplitude of plant growth stimulation in the boreal zone, related in an uncertain manner to the nutrient cycle, and the stomatal response of mature trees in the moist Tropics are the main sources of uncertainties. Clearly, a number of other factors have the potential to impact biosphere-atmosphere interactions in a CO2 rich world, the first of them being human-induced land-use change. Nevertheless, this work demonstrates the
ACL
30 - 13
importance of plant response to CO2 for land surface hydrology.
Appendix A: Description of the Vegetation Model [38] The vegetation model is extensively described in Kergoat [1998], the gas-exchange equations are detailed in Cayrol et al. [2000]. Additions to these papers are briefly listed here. A1. Carbon Cost Criteria [39] The annual leave carbon requirements depend on plant specific leaf area (SLA) and leaf lifetime (t). The carbon cost criterion estimates SLA.t according to Reich et al., [1992] relationships, following Haxeltine and Prentice [1996]. The LAI that optimizes the carbon-cost criterion is computed according to equation: LAI=t:SLA ¼ Al :Rc:GPPðLAIÞ; 2
ð1Þ
1
where t.SLA is equal to 0.02 m g C yr, Rc is the respiration fraction (set to 0.5 for grasses and 0.6 otherwise), Al is the fraction of carbon allocated to the leaves, set to 0.33 for all the vegetation types, GPP is the canopy photosynthesis. [40] Another modification, following Kergoat [1998], is the increase of the rooting depth of tropical trees (4.5 m). A2. Photosynthesis and Conductance Submodels [41] For each grid cell, a leaf level photosynthesis model is integrated to the canopy level according to LAI. The daily rate of leaf photosynthesis, P, is given by P¼
Dl ; rs þ rr
ð2Þ
where is the difference in CO2 concentration between the chloroplast and the atmosphere, rs and rr are the stomatal and nonstomatal resistances to CO2 transfer. Dl is the day length. The stomatal resistance is a ‘‘Jarvis’’ type resistance [Jarvis, 1976], with a minimum resistance rsmin modulated by soil available water, vapor pressure deficit and incoming photosynthetically active radiation: rs ¼ rsmin :f1 ðWÞ:f2 ðVPDÞ:f3 ðIÞ:
ð3Þ
The residual resistance accounts for mesophyll diffusion and carboxylation processes and is controlled by PAR level and temperature: rr ¼ rrmin :g1 ðIÞ:g2 ðTaÞ:
ð4Þ
Parameters for the residual resistance are derived from the maximum rate of photosynthesis, Pmax, the quantum yield, a and the minimum stomatal resistance, rsmin [see Cayrol et al., 2000] taken from Ko¨rner [1994]. Integration to the canopy gives Z Pcan ¼ Dl:
LAI
P:dL
¼ D=½kðrs0 þ rr0 Þ :Log½rs0 Is þ rr0 Ir þ k I0 ðrs0 þ rr0 Þ
= rs0 Is þ rr0 Ir þ ðrs0 þ rr0 ÞkI0 :ek:LAI ;
ð5Þ
30 - 14 KERGOAT ET AL.: CO2 EFFECTS ON GLOBAL LAI AND EVAPOTRANSPIRATION
ACL
where Pcan is the daily canopy photosynthesis, I0 the surface incoming PAR, k the PAR extinction coefficient, LAI the canopy leaf area index. rs’ and rr’ are equal to rs/f3 and rr/g1 respectively. Is and Ir are parameters of g1 and f3. Similarly, the canopy conductance for water vapor diffusion, Gcan is given by Z
LAI
Gcan ¼
1:6 dL=rs ¼ 1:6=ðk:rs0 Þ:Log ðIs þ k:I0 Þ= Is þ k:I0 :ek:LAI :
ð6Þ
A3. Doubled CO2 Parameterizations [42] The CO2 impact on stomatal resistance is rsmin 0 ¼ g:rsmin :
ð7Þ
The bias g is prescribed on a pixel basis (see text). [43] The CO2 impact on photosynthesis is rsmin 00 ¼ rsmin 0 ¼ g:rsmin ðMAXÞ or rsmin 00 ¼ rsmin ðFERTÞ
ð8Þ
P max00 ¼ 1:25 Pmax
ð9Þ
a00 ¼ 1:25 a
ð10Þ
00 ¼ 2::
ð11Þ
[44] Acknowledgments. The authors acknowledge Lili Ioanidu for her help in the preparation of the manuscript and the anonymous reviewers for helpful comments and suggestions. Thanks also to Richard Betts (for providing the stomatal bias), Sam Levis, and Bernard Saugier for stimulating discussions.
References Alexandre, D. Y., L’indice foliaire des foreˆts tropicales, Acta Oecol., 2, 299 – 312, 1981. Ashton, A. R., The effect of doubling atmospheric CO2 on streamflow: A simulation, J. Hydrol., 67, 273 – 280, 1984. Bazzaz, F. A., The response of natural ecosystems to the rising global CO2 levels, Ann. Rev. Ecol. Syst., 21, 167 – 196, 1990. Beerling, D. J., Long-term responses of boreal vegetation to global change: An experimental and modelling investigation, Global Change Biol., 5, 55 – 74, 1999. Berthelot, B., S. Adam, G. Dedieu, P. Maisongrande, L. Kergoat, and F. Cabot, A global dataset of surface reflectances and vegetation indices derived from AVHRR/GVI time series for 19891990: The Land Surface Reflectances (LASUR) data, in Physical Measurements and Signatures in Remote Sensing, edited by G. Guyot and T. Phulpin, pp. 685 – 690, Balkema, Rotterdam, 1997. Bettarini, I., F. P. Vaccari, and F. Miglietta, Elevated CO2 concentrations and stomatal density: Observations from 17 plant species growing in a CO2 spring in central Italy, Global Change Biol., 4, 17 – 22, 1998. Betts, R. A., P. M. Cox, S. L. Lee, and F. I. Woodward, Contrasting physiological and structural vegetation feedbacks in climate change simulations, Nature, 387, 796 – 799, 1997. Bunce, J. A., Response of stomatal conductance to light, humidity and temperature in winter wheat and barley grown at three concentrations of carbon dioxide in the field, Global Change Biol., 6, 371 – 382, 2000. Cayrol, P., L. Kergoat, S. Moulin, G. Dedieu, and A. Chehbouni, Calibrating a coupled SVAT/vegetation growth model with remotely sensed reflectance and surface temperature: A case study for the HAPEX-Sahel grassland sites, J. Appl. Meteorol., 39, 2452 – 2472, 2000. Chapin, F. S., III, G. R. Shaver, A. E. Giblin, K. J. Nadelhoffer, and J. A. Laundre, Response of arctic tundra to experimental and observed changes in climate, Ecology, 76, 694 – 711, 1995.
Cook, A. C., D. T. Tissue, S. W. Roberts, and W. C. Oechel, Effect of longterm elevated [CO2] from natural CO2 springs on Nardus stricta: Photosynthesis, biochemistry, growth and phenology, Plant Cell Environ., 21, 417 – 425, 1998. Cox, P. M., R. A. Betts, C. B. Bunton, R. L. H. Essery, P. R. Rowntree, and J. Smith, The impact of new land surface physics on the GCM simulation of climate and climate sensitivity, Clim. Dynam., 15, 183 – 203, 1999. Cramer, W., et al., ring global models of terrestrial net primary productivity (NPP): Overview and key results, Global Change Biol., 5(suppl. 1), 1 – 15, 1999. Curtis, P. S., and X. Wang, A meta-analysis of elevated CO2 effects on woody plant mass, form, and physiology, Oecologia, 111, 299 – 313, 1998. DeLucia, E. H., and R. B. Thomas, Photosynthetic response to CO2 enrichment of four hardwood species in a forest understory, Oecologia, 122, 11 – 19, 2000. De´que´, M., C. Dreveton, A. Braun, and D. Cariolle, The ARPEGE/IFS atmosphere model: A contribution to the French community climate modelling, Clim. Dyn., 10, 249 – 266, 1994. Dickinson, R. J., M. Shaikh, R. Bryant, and L. Graumlich, Interactive canopies for a climate model, J. Climate, 11, 2823 – 2836, 1998. Douville, H., S. Planton, J.-F. Royer, D. B. Stephenson, S. Tyteca, L. Kergoat, S. Lafont, and R. A. Betts, Importance of vegetation feedbacks in doubled-CO2 climate experiments, J. Geophys. Res., 105, 14,841 – 14,861, 2000. Drake, B. G., M. A. Gonza`les-Meler, and S. P. Long, More efficient plants: A consequence of rising atmospheric CO2, Annu. Rev. Plant Phys., 48, 609 – 639, 1997. Eagleson, P. S., Ecological optimality in water-limited natural soil-vegetation systems, 1, Theory and hypothesis, Water Resour. Res., 18, 325 – 340, 1982. Ellsworth, D. S., CO2 enrichment in a maturing pine forest: Are CO2 exchange and water status in the canopy affected?, Plant Cell Environ, 22, 461 – 472, 1999. Field, C. B., R. B. Jackson, and H. A. Mooney, Stomatal response to increased CO2: Implication from the leaf to the global scale, Plant Cell Environ., 18, 1214 – 1225, 1995. Field, C. B., C. P. Lund, N. R. Chiarello, and B. E. Mortimer, CO2 effects on the water budget of grassland microcosm communities, Global Change Biol., 3, 197 – 206, 1997. Foley, J. A., S. Levis, I. C. Prentice, D. Pollard, and S. L. Thomson, Coupling dynamics models of climate and vegetation, Global Change Biol., 4, 561 – 580, 1998. Gholz, H. L., Environmental limits on aboveground net primary production, leaf area, and biomass in vegetation zones of the pacific northwest, Ecology, 63, 469 – 481, 1982. Gifford, R. M., The global carbon cycle: A viewpoint on the missing sink, Aust. J. Plant Physiol., 21, 1 – 15, 1994. Grier, C. C., and S. W. Running, Leaf area of mature northwestern coniferous forests: Relation to site water balance, Ecology, 58, 893 – 899, 1977. Ha¨ttenschwiler, S., and C. Ko¨rner, Effect of elevated CO2 and increased nitrogen deposition on photosynthesis and growth of understory plants in spruce model ecosystems, Oecologia, 106, 172 – 180, 1996. Ha¨ttenschwiler, S., F. Miglietta, A. Raschi, and C. Ko¨rner, Morphological adjustments of mature Quercus ilex trees to elevated CO2, Acta Oecol., 8, 361 – 365, 1997. Haxeltine, A., and I. C. Prentice, BIOME3: An equilibrium terrestrial biosphere model based on ecophysiological constraints, resource availability and competition among plant functional types, Global Biogeochem. Cy., 10, 693 – 709, 1996. Heath, O. V. S., and J. Russell, Studies in stomatal behaviour, VI, An investigation of the light responses of wheat stomata with the attempted elimination of control by the mesophyll, Part II, Interactions with external carbon dioxide, and general discussion, J. Exp. Bot., 5, 269 – 292, 1954. Henderson-Sellers, A., K. Mcguffie, and C. Gross, Sensitivity of global climate model simulations to increased stomatal resistance and CO2 increases, J. Clim., 8, 1738 – 1756, 1995. Henry, G. H. R., and U. Molau, Tundra plants and climate change: The International Tundra Experiment (ITEX), Global Change Biol., 3(suppl.1), 1 – 9, 1997. Hoosbeek, M. R., N. van Bremen, F. Berendse, P. Grosvernier, H. Vasander, and B. Wallen, Limited effect of increased atmospheric CO2 concentration on ombrophytic bog vegetation, New Phytol., 150, 459 – 463, 2001. Houghton, J. T., Y. Ding, D. J. Griggs, M. Noguer, P. J. van des Linden, X. Dai, K. Maskell, and C. A. Johnson (Eds.), Climate Change 2001: The Scientific Basis: Contribution of Working Group I to the Third Assessment Report of the Intergovernmental Panel on Climate Change, Cambridge Univ. Press, New York, 2001.
KERGOAT ET AL.: CO2 EFFECTS ON GLOBAL LAI AND EVAPOTRANSPIRATION Hui, D., Y. Luo, W. Cheng, J. S. Coleman, D. W. Johnson, and D. A. Sims, Canopy radiation- and water-use efficiencies as affected by elevated [CO2], Global Change Biol., 7, 75 – 91, 2001. Idso, S. B., and A. J. Brazel, Rising atmospheric carbon dioxide may increase stream flow, Nature, 312, 51 – 53, 1984. Jarvis, P. G., The interpretation of the variations in leaf water potential and stomatal conductance found in canopies in the field, Philos. Trans.R. Soc. London, Ser. B, 273, 593 – 610, 1976. Kelliher, F. M., D. Y. Hollinger, E. D. Schulze, N. N. Vygodskaya, J. N. Byers, J. E. Hunt, T. M. McSeveny, I. Milukova, A. Sogatchev, A. Varlagin, W. Ziegler, A. Arneth, and G. Bauer, Evaporation from a eastern Siberian larch forest, Agric. For. Meteorol., 85, 135 – 147, 1997. Kelliher, F. M., et al., Evaporation from a central Siberian pine forest, J. Hydrol., 205, 279 – 296, 1998. Kergoat, L., A model for hydrological equilibrium of Leaf Area Index at the global scale, Hydrol, J., 212 – 213, 267 – 286, 1998. Kimball, B. A., et al., Effect of free-air CO2 enrichment on energy balance and evapotranspiration of cotton, Agric. For. Meteorol., 70, 259 – 278, 1994. Kimball, B. A., R. L. LaMorte, and P. J. Pinter Jr., et al., Free-air CO2 enrichment and soil nitrogen effects on energy balance and evapotranspiration of wheat, Water Resour. Res., 35, 1179 – 1190, 1999. Ko¨rner, C., Leaf diffusive conductances in the major vegetation types of the globe, in Ecophysiology of Photosynthesis, edited by E. D. Schulze and M. M. Goldwell, pp. 463 – 490, Springer-Verlag, New York, 1994. Ko¨rner, C., The response of complex multispecies systems to elevated CO2, in Global Change and Terrestrial Ecosystems, edited by B. H. Walker and W. L. Steffen, pp. 20 – 42, Cambridge Univ. Press, New York, 1996. Ko¨rner, C., Tropical forests in a CO2-rich world, Clim. Change, 39, 297 – 315, 1998. Leemans, R., and W. Cramer, The IIASA data base for mean monthly values of temperature, precipitation and cloudiness on a global terrestrial grid, Res. Rep. RR-91-18, Int. Inst. for Appl. Syst. Anal., Laxenburg, Austria, 1991. Levis, S., J. A. Foley, and D. Pollard, Potential high-latitude vegetation feedbacks on CO2-climate change, Geophys. Res. Lett., 26, 740 – 750, 1999. Levis, S., J. A. Foley, and D. Pollard, Large scale vegetation feedbacks on a doubled CO2 climate, J. Clim., 13, 1313 – 1325, 2000. Lichter, J., M. Lavine, K. A. Mace, D. D. Richter, and W. H. Schlesinger, Throughfall chemistry in a loblolly pine plantation under elevated atmospheric CO2 concentration, Biogeochemistry, 50, 70 – 93, 2000. Lloyd, J., and G. D. Farquhar, The CO2 dependence of photosynthesis, plant growth responses to elevated CO2 concentrations and their interaction with soil nutrient status, I. General principles and forest ecosystems, Funct. Ecol., 10, 4 – 32, 1996. Lovelock, C. E., A. Virgo, M. Popp, and K. Winter, Effects of CO2 concentrations on photosynthesis, growth and reproduction of branches of the tropical canopy tree species, Luehae seemannii Tr. and Planch, Plant Cell Environ., 22, 49 – 59, 1999. Luo, Y., J. Reynolds, Y. Wang, and D. Wolfe, A search for predictive understanding of plant responses to elevated [CO2], Global Change Biol., 5, 143 – 156, 1999. McGuire, A. D., J. M. Mellilo, and L. A. Joyce, The role of nitrogen in the response of forest net primary production to elevated atmospheric carbon dioxide, Annu. Rev. Ecol. Syst., 26, 473 – 503, 1995. Medlyn, B. E., et al., Stomatal conductance of forest species after long-term exposure to elevated CO2 concentration: A synthesis, New Phytol., 149, 247 – 264, 2001. Menzel, A., and P. Fabian, Growing season extended in Europe, Nature, 397, 659, 1999. Mitchell, J. F. B., and T. C. Johns, On modification of global warming by sulfate aerosols, J. Clim., 10, 245 – 267, 1997. Morgan, J. A., D. R. LeCain, A. R. Mosier, and D. G. Milchunas, Elevated CO2 enhances water relations and productivity and affects gas exchange in C3 and C4 grasses of the Colorado shortgrass steppe, Global Change Biol., 7, 451 – 466, 2001. Morison, J. I. L., Stomatal response to increased CO2 concentration, J. Exp. Bot., 49, 443 – 452, 1998. Moulin, S., L. Kergoat, N. Viovy, and G. Dedieu, Global scale assessment of vegetation phenology using NOAA/AVHRR satellite measurements, J. Clim., 10, 1154 – 1170, 1997. Myneni, R. B., R. R. Nemani, and S. W. Running, Estimation of global leaf area index and absorbed PAR using radiative transfer models, IEEE Trans. Geosci. Remote Sens., 35, 1380 – 1393, 1997a. Myneni, R. B., C. D. Tucker, C. J. Asrar, and G. Nemani, Increased plant growth in the Northern high latitudes from 1981 to 1991, Nature, 286, 686 – 702, 1997b.
ACL
30 - 15
Naumburg, E., and D. S. Ellsworth, Photosynthetic sunfleck utilization potential of understory saplings growing under elevated CO2 in FACE, Oecologia, 122, 163 – 174, 2000. Neilson, R. P., A model for predicting continental scale vegetation distribution and water balance, Ecol. Appl., 5, 362 – 385, 1995. Neilson, R. P., and D. Marks, A global perspective of regional vegetation and hydrologic sensitivities from climatic change, J. Veg. Sci., 5, 715 – 730, 1994. Nijs, I., J. Roy, J.-L. Salager, and J. Fabreguettes, Elevated CO2 alters carbon fluxes in early successional Mediterranean ecosystems, Global Change Biol., 6, 981 – 994, 2000. Niklaus, P. A., D. Spinnler, and C. Ko¨rner, Soil moisture dynamics of calcareous grassland under elevated CO2, Oecologia, 117, 201 – 208, 1998. Nowak, R. S., L. A. DeFalco, C. S. Wilcox, D. N. Jordan, J. S. Coleman, J. R. Seemann, and S. D. Smith, Leaf conductance decreased under freeair CO2 enrichment (FACE) for three perennials in the Nevada desert, New Phytol., 150, 449 – 458, 2001. Oechel, W. C., S. Cowles, N. Grulke, S. J. Hastings, B. Lawrence, T. Prudhomme, G. Riechers, B. Strain, D. Tissue, and G. Vourtilis, Transient nature of CO2 fertilization in arctic tundra, Nature, 371, 500 – 503, 1994. Osborne, C. P., B. G. Drake, J. Laroche, and S. P. Long, Does long-term elevation of CO2 concentration increase photosynthesis in forest floor vegetation? Indiana strawberry in a Maryland forest, Plant Physiol., 114, 337 – 344, 1997. Ouaidrari, H., E. Vermote, N. El Saleous, and D. Roy, AVHRR Pathfinder dataset: Evaluation and improvements, in Physical Measurements and Signatures in Remote Sensing, edited by G. Guyot and T. Phulpin, pp. 131 – 137, Balkema, Rotterdam, 1997. Owensby, C. E., J. M. Ham, A. K. Knapp, and L. M. Auen, Biomass production and species composition change in tallgrass prairie ecosystem after long-term exposure to elevated atmospheric CO2, Global Change Biol., 5, 497 – 506, 1999. Pan, Y., et al., Modeled responses of terrestrial ecosystems to elevated CO2: A comparison of simulations by the biogeochemistry models of the Vegetation/Ecosystem Modeling and Analysis Project (VEMAP), Oecologia, 114, 389 – 404, 1998. Pataki, D. E., T. E. Huxman, D. N. Jordan, S. F. Zitzer, J. S. Coleman, S. D. Smith, R. S. Nowak, and J. R. Seemann, Water use of two Mojave desert shrubs under elevated CO2, Global Change Biol., 6, 889 – 897, 2000. Pollard, D., and S. L. Thomson, Use of a land-surface-transfer scheme (LSX) in a global climate model: The response to doubling stomatal resistance, Global Planet. Change, 10, 129 – 161, 1995. Reich, P. B., M. B. Walters, and D. S. Ellsworth, Leaf life-span in relation to leaf, plant, and stand characteristics among diverse ecosystems, Ecol. Monogr., 62, 365 – 392, 1992. Reich, P. B., et al., Plant diversity enhances ecosystem responses to elevated CO2 and nitrogen deposition, Nature, 410, 809 – 812, 2001. Rustad, L. E., J. L. Campbell, G. M. Marion, R. J. Norby, M. J. Mitchell, A. E. Hartley, J. H. C. Cornelissen, and J. Gurevitch, A meta-analysis of the response of soil respiration, net nitrogen mineralization, and aboveground plant growth to experimental warming, Oecologia, 126, 543 – 562, 2001. Saxe, H., D. S. Ellsworth, and J. Heath, Tree and forest functioning in an enriched CO2 atmosphere, New Phytol., 139, 395 – 436, 1998. Schimel, D. S., The carbon equation, Nature, 393, 208 – 209, 1998. Schulze, E. D., F. M. Kelliher, C. Ko¨rner, J. Lloyd, and R. Leuning, Relationship between maximum stomatal conductance, ecosystem surface conductance, carbon assimilation rate and plant nitrogen nutrition: A global ecology scaling exercise, Annu. Rev. Ecol. Syst., 25, 629 – 660, 1994. Sellers, P. J., et al., Comparison of radiative and physiological effects of doubled atmospheric CO2 on climate, Science, 271, 1402 – 1406, 1996a. Sellers, P. J., S. O. Los, C. J. Tucker, C. O. Justice, D. A. Dazlich, G. J. Colatz, and D. A. Randall, A revised land surface parameterization (SiB2) for atmospheric GCMs, Part II, The generation of global fields of terrestrial biophysical parameters from satellite data, J. Clim., 9, 706 – 736, 1996b. Tian, Y., Y. Zhang, Y. Knyazikhin, R. B. Myneni, J. Glassy, G. Dedieu, and S. Running, Prototyping of MODIS LAI and FPAR algorithm with LASUR and LANDSAT data, IEEE Trans. Geosci. Remote Sens., 38, 2387 – 2401, 2000. Warwick, K. R., G. Taylor, and H. Blum, Biomass and composition changes occur in chalk grassland turves exposed to elevated CO2 for two seasons in FACE, Global Change Biol., 4, 375 – 385, 1998. Wigley, T. M. L., and P. D. Jones, Influences of precipitation changes and direct CO2 effects on streamflow, Nature, 314, 149 – 152, 1985.
ACL
30 - 16 KERGOAT ET AL.: CO2 EFFECTS ON GLOBAL LAI AND EVAPOTRANSPIRATION
Wilson, M. F., and A. Henderson-Sellers, A global archive of land cover and soils data for use in general circulation climate models, J. Climatol., 5, 119 – 143, 1985. Woodward, F. I., Climate and Plant Distribution, Cambridge Univ. Press, New York, 1987. Woodward, F. I., T. M. Smith, and W. R. A. Emanuel, A global land productivity and phytogeography model, Global Biogeochem. Cy., 9, 471 – 480, 1995. Wullschleger, S. D., and R. J. Norby, Sap velocity and canopy transpiration in a sweetgum stand exposed to free-air CO2 enrichment (FACE), New Phytol., 150, 489 – 498, 2001. Wu¨rth, M. K. R., K. Winter, and C. Ko¨rner, In situ responses to elevated CO2 in tropical forest understory plants, Funct. Ecol., 12, 889 – 895, 1998. Zanetti, S., U. A. Hartwig, C. Van Kessel, A. Luscher, T. Hebeisen, M. Frehner, B. U. Fischer, G. R. Hendrey, H. Blum, and J. Nosberger, Does
nitrogen nutrition restrict the CO2 response of fertile grassland lacking legumes?, Oecologia, 112, 17 – 25, 1997.
L. Kergoat, Laboratoire d’Ecologie Terrestre, CNRS, 13 Avenue Colonel Roche, BP 4072, 31029 Toulouse Cedex 4, France. (
[email protected]) B. Berthelot, G. Dedieu, and S. Lafont, Centre d’Etude Spatiale de la Biosphe`re, 18 Avenue Edouard Belin, bpi 2801, 31401 Toulouse cedex 4, France. (
[email protected];
[email protected];
[email protected]) H. Douville, S. Planton, and J.-F. Royer, Centre National de la Recherche Me´te´orologique, 42 Avenue Coriolis, 31057 Toulouse cedex 01, France. (
[email protected];
[email protected]; jean-francois.royer@ meteo.fr)
KERGOAT ET AL.: CO2 EFFECTS ON GLOBAL LAI AND EVAPOTRANSPIRATION
Figure 1. Comparison of the LAI model results with remote sensing estimates. Maximum fPAR (a) simulated for present-day climate and (b) retrieved from the FASIR/AVHRR data set. (c) July maximum LAI (projected, m2/m2) simulated for present-day climate and (d) retrieved from the LASUR/AVHRR data set processed with the MODIS prototype algorithm.
ACL
30 - 5
KERGOAT ET AL.: CO2 EFFECTS ON GLOBAL LAI AND EVAPOTRANSPIRATION
Figure 3. Effect of a doubling of CO2 concentration on Leaf Area Index (m2/m2). (a) LAI anomaly between the RAD (climate only) and CLIM (present-day) simulations. (b) Anomaly between STOM (maximum stomatal response) and CLIM. (c) Anomaly between FERT (maximum photosynthesis response) and CLIM. (d) Anomaly between MAX (stomatal and photosynthesis response) and CLIM.
ACL
30 - 7
