Sugar maple (Acer saccharum Marshall) growth in the species’ southern range has been declining since the 1980s, putting at risk a variety of ecosystem services that the species provides. Heatwaves, drought, frosts, acidic deposition, and insect defoliation, all reducing photosynthetic activity, have been suggested to be behind the phenomenon. Because the geographic scope of previous studies on maple growth is limited to the southern temperate biome, it is not currently understood whether the same negative trends and factors affecting growth rates apply to the species in more northern regions of its distribution range. Here we used annual ring-width data of 1675 trees from a network of 21 sites in Quebec and Ontario between 45˚N and 48˚N to reconstruct maple growth and to analyze its trends and climatic drivers since 1950 C

Despite growing interest in predicting plant phenological shifts,advanced spring phenology by global climate change remains debated. Evidence documenting either small or large advancement of spring phenology to rising temperature over the spatio-temporal scales implies a potential existence of a thermal threshold in the responses of forests to global warming. We collected a unique data set of xylem cell-wall-thickening onset dates in 20 coniferous species covering a broad mean annual temperature (MAT) gradient (?3.05 to 22.9°C) across the Northern Hemisphere (latitudes 23°?66°?N). Along the MAT gradient,we identified a threshold temperature (using segmented regression) of 4.9?±?1.1°C,above which the response of xylem phenology to rising temperatures significantly decline. This threshold separates the Northern Hemisphere conifers into cold and warm thermal niches,with MAT and spring forcing being the primary drivers for the onset dates (estimated by linear and Bayesian mixed-effect models),respectively. The identified thermal threshold should be integrated into the Earth-System-Models for a better understanding of spring phenology in response to global warming and an improved prediction of global climate-carbon feedbacks.

Fire is a major disturbance agent in the boreal forest, affecting the structure, dynamics and biogeochemical cycles in this biome. In the Asian section of boreal forest, the records of long-term fire history are few that limits our understanding of factors forcing regional fire dynamics. We presented an annually-resolved 352-year (1666–2017) fire chronology based on fire scars of Scots pine (Pinus sylvestris L.) and Siberian larch (Larix sibirica Ledeb) from the Transbaikal area in the southeastern Siberia. Fire activity showed an increasing trend from 1720 to 1929 (R2 = 0.80, P < 0.0001), and a significant decreasing trend from 1920 to 2010 (R2 = 0.62, P < 0.001). We assessed the potential relationships between drought (as represented by the Palmer Drought Severity Index, PDSI, and the Monthly Drought Code, MDC), ocean-atmosphere circulation and forest fire by Superposed epoch analyses, cross-wavelet analysis and Granger causality analysis. Increased fire activity was associated with stronger drought from previous winter to current summer of fire event years and positive Arctic Oscillation (AO) before and during major fire season (February and April to May), as revealed by superposed epoch analysis. Granger causality pointed to the significant role of drought in driving forest fires. Our findings provide insights into the climate drivers of forest fire activity and its prediction in the Transbaikal region.

The boreal forests of Central Asia play a vital role in biodiversity protection and regional economic development. It is important to study potential changes in the growth dynamics of boreal species in a context of global change. In this study, we developed a network of 34 tree-ring chronologies for two tree species, Siberian larch (Larix sibirica Ledeb.) and Siberian pine (Pinus sibirica Du Tour). The network extended across a large latitudinal gradient (45°N to 55°N). Principal component analysis (PCA) was used to detect spatial patterns in tree radial growth during a common period 1943–2004. Results indicated an obvious clustering pattern with chronologies being divided into a northeastern (NR) and a southwestern (SR) region. Bootstrapped correlation analyses of regional climate versus aggregated chronologies showed that tree radial growth in both regions was positively associated with summer temperature (June and July). Tree radial growth in the northeastern region was however positively associated with early spring precipitation and spring Palmer Drought Severity Index (PDSI) whereas, in the southwestern region, it was characterized by negative correlations with early summer precipitation and summer PDSI. The warm pool El Niño-Southern Oscillation (WP ENSO) and North Atlantic Oscillation (NAO) regulated tree radial growth through their influence on regional precipitation and temperature. Results suggest that tree radial growth in the region may decline with future projected climate change. This study provides a more comprehensive understanding to tree growth-climate associations across Central Asia.

Bud and leaf development are important phenological events and help in defining the growing period of trees. Canopy greenness derived from PhenoCam has been used to investigate leaf phenology. Questions remain on how much the continuous records of canopy greenness represent bud developmental phases, and how growing period boundaries are related to canopy greenness and bud phenology. In this study, we compared bud phenology of black spruce [Picea mariana (Mill.) B.S.P] during 2015, 2017 and 2018 with the canopy greenness, represented by Green Chromatic Coordinate (GCC), derived from PhenoCam images of a boreal stand in Quebec, Canada. Logit models were applied to estimate the probability of observing sequential phenological phases of bud burst and bud set along with GCC. GCC showed a bell-shaped pattern, with a slow increase in spring, a peak in summer and a gradual decrease in autumn. The start and end of budburst, and bud set, occurred when GCC reached 72% and 92% (spring), and 94% (autumn) of its maximum amplitude, respectively. These GCC values are reliable thresholds indicating the growing period boundaries. Our study builds a bridge between phenological observations and automatic near-surface remote sensing, providing a statistically sound protocol for calibrating PhenoCam with field observations.

Wood formation consumes around 15% of the anthropogenic CO2 emissions per year and plays a critical role in long-term sequestration of carbon on Earth. However, the exogenous factors driving wood formation onset and the underlying cellular mechanisms are still poorly understood and quantified, and this hampers an effective assessment of terrestrial forest productivity and carbon budget under global warming. Here, we used an extensive collection of unique datasets of weekly xylem tissue formation (wood formation) from 21 coniferous species across the Northern Hemisphere (latitudes 23 to 67°N) to present a quantitative demonstration that the onset of wood formation in Northern Hemisphere conifers is primarily driven by photoperiod and mean annual temperature (MAT), and only secondarily by spring forcing, winter chilling, and moisture availability. Photoperiod interacts with MAT and plays the dominant role in regulating the onset of secondary meristem growth, contrary to its as-yet-unquantified role in affecting the springtime phenology of primary meristems. The unique relationships between exogenous factors and wood formation could help to predict how forest ecosystems respond and adapt to climate warming and could provide a better understanding of the feedback occurring between vegetation and climate that is mediated by phenology. Our study quantifies the role of major environmental drivers for incorporation into state-of-the-art Earth system models (ESMs), thereby providing an improved assessment of long-term and high-resolution observations of biogeochemical cycles across terrestrial biomes.

Contrary to the generally advanced spring leaf unfolding under global warming, the effects of the climate warming on autumn leaf senescence are highly variable with advanced, delayed, and unchanged patterns being all reported. Using one million records of leaf phenology from four dominant temperate species in Europe, we investigated the temperature sensitivities of spring leaf unfolding and autumn leaf senescence (ST, advanced or delayed days per degree Celsius). The ST of spring phenology in all of the four examined species showed an increase and decrease during 1951–1980 and 1981–2013, respectively. The decrease in the ST during 1981–2013 appears to be caused by reduced accumulation of chilling units. As with spring phenology, the ST of leaf senescence of early successional and exotic species started to decline since 1980. In contrast, for late successional species, the ST of autumn senescence showed an increase for the entire study period from 1951 to 2013. Moreover, the impacts of rising temperature associated with global warming on spring leaf unfolding were stronger than those on autumn leaf senescence. The timing of leaf senescence was positively correlated with the timing of leaf unfolding during 1951–1980. However, as climate warming continued, the differences in the responses between spring and autumn phenology gradually increased, so that the correlation was no more significant during 1981–2013. Our results further suggest that since 2000, due to the decreased temperature sensitivity of leaf unfolding the length of the growing season has not increased any more. These finding needs to be addressed in vegetation models used for assessing the effects of climate change.

Emissions of aerosols and trace gases from wildfires and their direct shortwave radiative forcing (DSRF) at the top of atmosphere were studied using satellite observations from Moderate?Resolution Imaging Spectroradiometer, Atmospheric Infrared Sounder, Clouds and Earth Radiant Energy System on Aqua, and Ozone Monitoring Instrument on Aura. The dominant fuel types of the selected fire cases in the northeast of China (NEC), Siberia (Russia), and California (USA) are cropland, mixed forest, and needle?leaf forest, respectively. For the cropland fire case in NEC, the fire radiative power?based emission coefficients (Ce) of aerosol is 20.51 ± 2.55 g/MJ, half that of the forest fire cases in Siberia (40.01 ± 9.21 g/MJ) and California (45.23 ± 8.81 g/MJ), and the carbon monoxide (CO) Ce (23.94 ± 11.83 g/MJ) was about one third and half that of the forest fire cases in Siberia and California, respectively. However, the NOx (NO2 + NO) Ce (2.76 ± 0.25g MJ?1) of the cropland fire in NEC was nearly 3 times that of those forest fire cases. Ratios of NOx to aerosol, HCHO, and CO in the cropland case in NEC show much higher values than those in the forest fire cases. Despite the differences of the Ce and the composition ratios, the DSRF efficiency of smoke aerosol at the top of atmosphere showed similar values among those fire cases. Our results highlight the large variability of emission rate and relative chemical composition but similar DSRF efficiencies among wildfires, which would provide valuable information for understanding the impact of fire on air quality and climate.

Under current global warming, high-elevation regions are expected to experience faster warming than low-elevation regions. However, due to the lack of studies based on long-term large-scale data, the relationship between tree spring phenology and the elevation-dependent warming is unclear. Using 652k records of leaf unfolding of five temperate tree species monitored during 1951–2013 in situ in Europe, we discovered a nonlinear trend in the altitudinal sensitivity (SA, shifted days per 100 m in altitude) in spring phenology. A delayed leaf unfolding (2.7
0.6 days per decade) was observed at high elevations possibly due to decreased spring forcing between 1951 and 1980. The delayed leaf unfolding at high-elevation regions was companied by a simultaneous advancing of leaf unfolding at low elevations. These divergent trends contributed to a significant increase in the SA (0.36
0.07 days 100/m per decade) during 1951–1980. Since 1980, the SA started to decline with a rate of 0.32
0.07 days 100/m per decade, possibly due to reduced chilling at low elevations and improved efficiency of spring forcing in advancing the leaf unfolding at high elevations, the latter being caused by increased chilling. Our results suggest that due to both different temperature changes at the different altitudes, and the different tree responses to these changes, the tree phenology has shifted at different rates leading to a more uniform phenology at different altitudes during recent decades.

Ecosystem productivity estimated with a model calibrated with eddy-covariance data was related to tree-ring growth of two different boreal conifers along a latitudinal gradient. The relationship between ecosystem productivity and growth changed with species and site. Greater photosynthesis in spring and summer increased annual anomalies of radial growth in both species, and the response of growth to productivity was earlier in warmer southern stands particularly for pine. Radial growth of jack pine increased in the long-term with higher productivity, whereas this relationship was more reduced in black spruce. This could express species-specific differences in carbon allocation strategies but likely it is a consequence of the limiting marginal soils where spruce is found in the south. Only tree-rings of jack pine at some sites showed certain potential as direct proxies for ecosystem productivity at the low and high-frequency responses. Introduction Climate warming and the increase in atmos-pheric-CO 2 concentrations cause changes in forest growth and ecosystem productivity. There are reports of contrasting growth responses to warming over recent decades in different types of forests. Although some boreal species show negative growth trends in response to recent climate change (Hoofgaard et al. 1999, D'Arrigo et al. 2004), net ecosystem productivity in boreal and temperate conditions is generally expected to increase with increasing temperatures (Myneni et al. 1997, Boisvenue and Running 2006). Forest growth measurements and models assume that there is a close connection between the stem

Immediate phenotypic variation and the lagged effect of evolutionary adaptation to climate change appear to be two key processes in tree responses to climate warming. This study examines these components in two types of growth models for predicting the 2010–2099 diameter growth change of four major boreal species Betula papyrifera, Pinus banksiana, Picea mariana, and Populus tremuloides along a broad latitudinal gradient in eastern Canada under future climate projections. Climate-growth response models for 34 stands over nine latitudes were calibrated and cross-validated. An adaptive response model (A-model), in which the climate-growth relationship varies over time, and a fixed response model (F-model), in which the relationship is constant over time, were constructed to predict future growth. For the former, we examined how future growth of stands in northern latitudes could be forecasted using growth-climate equations derived from stands currently growing in southern latitudes assuming that current climate in southern locations provide an analogue for future conditions in the north. For the latter, we tested if future growth of stands would be maximally predicted using the growth-climate equation obtained from the given local stand assuming a lagged response to climate due to genetic constraints. Both models predicted a large growth increase in northern stands due to more benign temperatures, whereas there was a minimal growth change in southern stands due to potentially warm-temperature induced drought-stress. The A-model demonstrates a changing environment whereas the F-model highlights a constant growth response to future warming. As time elapses we can predict a gradual transition between a response to climate associated with the current conditions (F-model) to a more adapted response to future climate (A-model). Our modeling approach provides a template to predict tree growth response to climate warming at mid-high latitudes of the Northern Hemisphere.

• Premise of the study: In a warming climate, boreal trees may have adjusted their growth strategy (e.g., onset and coordination of growth among different organs such as stem, shoot, and foliage, within and among species) to cope with the extended growing seasons. A detailed investigation into growth of different organs during a growing season may help us assess the potential effects of climate change on tree growth in the boreal forest.

• Methods: The intra-annual growth of stem xylem, shoot tips, and foliage area of Pinus banksiana, Populus tremuloides, and Betula papyrifera was monitored in a boreal forest in Quebec, Canada during the growing season of 2007. Xylem formation was measured at weekly intervals, and shoot elongation and foliage expansion were measured three times per week from May to September. Growth indices for stem, shoot, and foliage were calculated and used to identify any climate–growth dependence.

• Key results: The time periods required for stem growth, branch extension, and foliage expansion differed among species. Of the three species, P. banksiana had the earliest budburst (20 May) yet the latest completion date of the foliage growth (2 August); P. tremuloides had the latest budburst (27 May) yet the earliest completion date of the foliage growth (10 July). Air temperature positively affected shoot extension growth of all three species. Precipitation positively influenced stem growth of the two broadleaf species, whereas growing season temperature positively impacted stem growth of P. banksiana.

• Conclusion: The results show that both the timing of growth processes and environmental dependences differ among co-occurring species, thereby leading to different adaptive capability of these boreal tree species to climate change.

• Premise of the study : Climate warming might have resulted in altered initiation and termination dates of stem xylem growth in
  boreal stands. A systematic understanding of intra-annual xylem formation is thus needed for precise simulation of future
  growth in the context of sustainable forest management.
• Methods : A recently developed novel microsampling approach was employed over two growing seasons (2005 – 2006) to investigate
  the intra-annual stem xylem formation of Picea mariana at three sites along a latitudinal gradient (approximately 47.5 –
  50 ° N) in western Quebec, Canada. The critical timings of xylem cell formation were determined and compared among sites/
  years. The relationships between xylem cell formation and meteorological variables were examined.
• Key results : From south to north, the onset of xylem cell production was detected on 20 May (SD ± 3) at Angliers, 24 May (SD ± 3)
  at Chicobi and 24 May (SD ± 4) at Muskuchii in 2005, and on 12 May (SD ± 4) at Angliers, 14 May (SD ± 3) at Chicobi and 20
  May (SD ± 3) at Muskuchii in 2006, respectively. Xylem cell production at each respective site terminated on 11 August (SD ± 4),
  7 August (SD ± 3), and 7 August (SD ± 4) in 2005, and on 8 August (SD ± 4), 4 August (SD ± 4), and 4 August (SD ± 4) in 2006,
  respectively.
• Conclusion : Our study implies that despite the expected occurrence of earlier phenological development due to early spring
  climate warming, boreal trees like P. mariana might not be producing wider rings if cold temperatures occur later in the growing
  season in June to August. These results may challenge the view that boreal trees could be benefi ting from spring warming
  to enhance growth.

PREFACE

The present thesis is comprised of seven parts, including the General Introduction, the Chapter I, Chapter II, Chapter III, Chapter IV, the General Conclusion, and an Appendix I presenting a publication associated with my Ph.D study. All the papers and contexts involved in the thesis together are my original contributions to my Ph.D in Environmental Science that I pursued and accomplished at Université du Québec en Abitibi-Témiscamingue, Canada. The Chapters I-IV are correspondingly based on the following four publications:

1. Huang J.G., Tardif J., Denneler B., Bergeron Y., and Berninger F (2008). Tree-ring evidence extends the historic northern range limit of severe defoliation by insects in the aspen stands of western Quebec, Canada. Canadian Journal of Forest Research, 38:2535-2544. (Impact Factor 2008: 1.434)
2. Huang J.G., Tardif J., Bergeron Y., Denneler B., Berninger F., and Girardin M. (2010) Radial growth response of four dominant boreal tree species to climate along a latitudinal gradient in the eastern Canadian boreal forest. Global Change Biology, 16:711-731. doi:10.1111/j.1365-2486.2009.01990.x (IF 2008: 5.876)
3. Huang J.G., Bergeron Y., Berninger F., Zhai L.H., Tardif J., and Denneler B. Impact of future climate on radial growth of the four dominant boreal tree species along a latitudinal gradient in the eastern Canadian boreal forest. (in preparation to Global Change Biology) (IF 2008: 5.876)
4. Huang J.G., Bergeron Y., Zhai L.H., and Denneler B. Variations in intra-annual radial growth (formation of the xylem) of black spruce along a latitudinal gradient in western Quebec. (in preparation to Trees-Structure and Function) (IF 2008: 1.629)

The Appendix I publication was the result of my Ph.D Synthesis Exam, which is often done during the second year of the Ph.D study through writing a 50-pages report to answer a hot question associated more or less with the candidate‟s study field within three months from the Ph.D committee. It aims to assess if a Ph.D candidate in Environmental Sciences is eligible for doing this Ph.D at University of Quebec. The Appendix I paper is listed as following:¸

5. Huang J.G., Bergeron Y., Denneler B., Berninger F., and Tardif J. (2007) Response of forest trees to increased atmospheric CO2. Critical Reviews in Plant Sciences, 26(5): 265-283. DOI:10.1080/07352680701626978 (IF 2008: 6.206)

The leading author developed all the experimental designs with Dr. Yves Bergeron and the late Dr. Bernhard Denneler, and conducted all the fieldwork with Dr. Denneler. The leading author conducted all the laboratory work of Chapters I and II. The leading author conducted all data analyses as well as wrote the five manuscripts and the thesis. Dr. Bergeron constantly financed, supervised, and discussed all PhD project, and commented the early versions of the manuscripts of all five papers and the thesis. Dr. Jacques Tardif guided the data analysis work in the Chapter I, and discussed and commented the early versions of the manuscripts of the four papers and the thesis. Dr. Frank Berninger discussed and commented the early versions of the manuscripts of the four papers and the thesis. Dr. Martin P. Girardin (Canadian Forest Service) calculated Canadian drought code data used in the Chapter II and commented the early version of it. Ms. LiHong Zhai (UQAM) calculated Canadian drought code data and processed large sets of climate change scenarios data in the Chapter III, and conducted all the laboratory work and provided key help on data analysis for Chapter IV, as well as discussed and commented the early versions of these two chapters. The comments and suggestions from all the coauthors greatly improved the quality of the papers, and are much appreciated.

To address the central question of how climate change influences tree growth within the context of global warming, we used dendroclimatological analysis to understand the reactions of four major boreal tree species –Populus tremuloides, Betula papyrifera, Picea mariana, and Pinus banksiana– to climatic variations along a broad latitudinal gradient from 46 to 54°N in the eastern Canadian boreal forest. Tree-ring chronologies from 34 forested stands distributed at a 1° interval were built, transformed into principal components (PCs), and analyzed through bootstrapped correlation analysis over the period 1950–2003 to identify climate factors limiting the radial growth and the detailed radial growth–climate association along the gradient. All species taken together, previous summer temperature (negative influences), and current January and March–April temperatures (positive influences) showed the most consistent relationships with radial growth across the gradient. Combined with the identified species/site-specific climate factors, our study suggested that moisture conditions during the year before radial growth played a dominant role in positively regulating P. tremuloides growth, whereas January temperature and growing season moisture conditions positively impacted growth of B. papyrifera. Both P. mariana and P. banksiana were positively affected by the current-year winter and spring or whole growing season temperatures over the entire range of our corridor. Owing to the impacts of different climate factors on growth, these boreal species showed inconsistent responsiveness to recent warming at the transition zone, where B. papyrifera, P. mariana, and P. banksiana would be the most responsive species, whereas P. tremuloides might be the least. Under continued warming, B. papyrifera stands located north of 49°N, P. tremuloides at northern latitudes, and P. mariana and P. banksiana stands located north of 47°N might benefit from warming winter and spring temperatures to enhance their radial growth in the coming decades, whereas other southern stands might be decreasing in radial growth.

Abstract: A dendrochronological reconstruction of insect outbreaks was conducted along a latitudinal gradient from 46°N to 54°N in the boreal forest of western Quebec, Canada. Tree-ring chronologies of the host species, trembling aspen (Populus tremuloides Michx.), were constructed to identify periods of severe defoliation and comparisons were made with tree-ring chronologies of nonhost species. In addition, the frequency of white and narrow rings was used to further confirm the occurrence of insect outbreaks at these latitudes. Some major outbreaks occurred in relatively close synchrony at the regional scale, but the initiation year, intensity, and extent of the outbreaks varied spatially. For example, the 1950s outbreaks were observed from 1951 to 1952 at 46°N, from 1953 to 1954 at 47°N, and from 1954 to 1956 at 48°N. Other major outbreaks like the 1964 and 1980 outbreaks were fairly well synchronized at northern latitudes. The observed outbreaks in trembling aspen stands at 54°N also provided clear evidence that severe insect defoliation occurs much further north than the currently reported range limit, that is, between 49°N and 51°N, of the most important trembling aspen defoliator, the forest tent caterpillar (Malacosoma disstria Hubner). Our study demonstrated that careful identification of white rings in host species can provide valid information allowing the expansion of the forestry insect inventory database both at temporal and spatial scales.

Résumé : Une reconstitution dendrochronologique des épidémies d’insecte a été réalisée le long d’un gradient allant de 46°N à 54°N dans la forêt boréale de l’ouest du Québec, au Canada. Les chronologies de l’espèce hôte, le peuplier faux-tremble, ont été construites de manière à identifier les périodes de défoliation sévère et des comparaisons ont été effectuées avec les chronologies d’espèces non hôtes. De plus, la fréquence des cernes pâles et étroits a été utilisée pour valider l’occurrence des épidémies d’insecte à ces latitudes. Quelques épidémies majeures sont survenues avec une synchronicité relativement étroite à l’échelle régionale mais l’année du début, l’intensité et l’étendue des épidémies variaient dans l’espace. Par exemple, les épidémies des années 1950 ont été observées de 1951 à 1952 à 46°N, de 1953 à 1954 à 47°N et de 1954 à 1956 à 48°N. D’autres épidémies importantes comme celles de 1964 et 1980 étaient assez bien synchronisées aux latitudes nordiques. Les épidémies observées dans les peuplements de peuplier faux-tremble à la latitude 54°N fournissent des preuves manifestes que des défoliations sévères causées par les insectes surviennent beaucoup plus au nord que la limite de l’aire de répartition couramment rapportée, soit entre 49°N et 51°N, dans le cas du plus important défoliateur du peuplier, la livrée des forêts. Notre étude montre que l’identification minutieuse des cernes pâles chez l’espèce hôte peut fournir une information valide qui permet d’élargir la base de données de l’inventaire des insectes forestiers tant à l’échelle temporelle que spatiale.

[Traduit par la Rédaction]

The CO2 fertilization hypothesis stipulates that rising atmospheric CO2 has a positive effect on tree growth due to increasing availability of carbon. The objective of this paper is to compare the recent literature related to both field CO2-enriched experiments with trees and empirical dendrochronological studies detecting CO2 fertilization effects in tree-rings. This will allow evaluation of tree growth responses to atmospheric CO2 enrichment by combining evidence from both ecophysiology and tree-ring research. Based on considerable experimental evidence of direct CO2 fertilization effect (increased photosynthesis, water use efficiency, and above- and belowground biomass), and predications from the interactions of enriched CO2 with temperature, nitrogen and drought, we propose that warm, moderately droughtstressed ecosystems with an ample nitrogen supply might be the most CO2 responsive ecosystems. Empirical tree-ring studies took the following three viewpoints on detecting CO2 fertilization effect in tree-rings: 1) finding evidence of CO2 fertilization effect in tree-rings, 2) attributing growth enhancement to favorable climate rather than atmospheric CO2 enrichment, and 3) considering that tree growth enhancement might be caused by synergistic effects of several factors such as favorable climate change, CO2 fertilization, and anthropogenic atmospheric deposition (e.g., nitrogen). At temperature-limiting sites such as high elevations, nonfindings of CO2 fertilization evidence could be ascribed to the following possibilities: 1) cold temperatures, a short season of cambial division, and nitrogen deficiency that preclude a direct CO2 response, 2) old trees past half of their maximum life expectancy and consequently only a small increase in biomass increment due to CO2 fertilization effect might be diminished, 3) the elimination of age/size-related trends by statistical detrending of tree-ring series that might remove some long-term CO2-related trends in tree-rings, and 4) carbon partitioning and growth within a plant that is species-specific. Our review supports the atmospheric CO2 fertilization effect hypothesis, at least in trees growing in semi-arid or arid conditions because the drought-stressed trees could benefit from increased water use efficiency to enhance growth.

The CO2 fertilization hypothesis stipulates that the rising atmospheric CO2 concentration has a positive effect on tree growth due to the increasing availability of carbon. The purpose of this paper is to review the recent literature on the field CO2-enriched experiments and empirical dendrochronological studies testing this hypothesis. A large body of studies based on field CO2-enrichment experiments provided evidence for increased photosynthesis, water use efficiency, and above- and below-ground biomass. Empirical dendrochronological studies showed three viewpoints on detecting CO2 fertilization effect in tree rings: 1) findings of evidence of CO2 fertilization effect in tree rings, 2) growth enhancement was attributed to rather favourable climate than atmospheric CO2 enrichment, 3) tree growth enhancement might be caused by synergistic effects of some or all of the factors such as favourable climate change, CO2 fertilization, and anthropogenic atmospheric deposition (e.g. nitrogen). Based on evidence of CO2 fertilization effect from considerable short-term field CO2 experiments and some empirical tree-ring studies, I am convinced that atmospheric CO2 fertilization effect does occur in some natural forests, at least in trees growing at the semi-arid or dry condition because the drought-stressed trees could benefit from increased water use efficiency to enhance growth. Under the rising greenhouse gases in the atmosphere and continually future climate warming, and to accurately assess impacts of CO2 fertilization on forest ecosystem and global carbon equilibrium, it is necessary to not only develop more long-term CO2 field experiments on mature trees to help resolve the uncertainties of mature trees’ responses to atmospheric CO2 enrichment in natural forests, but also strengthen empirical dendrochronological studies and other approaches such as carbon isotope analysis and building computer model. I advocate that empirical dendrochronological studies testing CO2 fertilization effect should 1) focus on trees growing in semiarid or arid region, 2) take caution to detrend age-related growth trends, 3) try to use BAI or age classes etc. when detrending fails, 4) understand variations in water use efficiency with the aid of carbon isotope analysis. When taking into account all those aspects, I believe that it could help better understanding of CO2 fertilization effect in natural forests, and certainly contribute to assessment of the impacts of atmospheric CO2 enrichment on environment, economy, and society in the future. © 2006 UQAM tous droits réservés.