43

Climate-induced dieback: an escalating global phenomenon?

C.D. Allen

An introduction to emerging global orests, which today cover 30 per- Since most of the world’s are patterns of -induced forest cent of the world’s land surface found in areas where temperature, light mortality. F(FAO, 2006), are being rapidly or nutrients limit growth and produc- and directly transformed in many areas tivity, recent global warming, changes in by the impacts of expanding human atmospheric composition (i.e. increased populations and economies. Less evi- concentrations of nitrogen compounds

dent are the pervasive effects of ongoing and CO2 from massive societal emis- climatic changes on the condition and sions) and local increases in sunlight and status of forests around the world. Recent precipitation have benefited the growth examples of and heat-related of many forests in recent decades, when forest stress and dieback (defined here and where water has not been limiting as tree mortality noticeably above usual (Boisvenue and Running, 2006). mortality levels) are being documented On the other hand, about one-third from all forested continents, making of the Earth’s land is currently too dry it possible to begin to see global pat- to support tree growth, and significant terns. This article introduces these pat- areas of forest and woodland grow in terns and considers the possibility that marginal climate zones where net pri- many forests and woodlands today are mary vegetation productivity is strongly at increasing risk of climate-induced water limited (Boisvenue and Running, dieback. A more comprehensive article 2006). Forests in such semi-arid regions (Allen et al., 2009) addresses this topic may display substantial growth declines in considerably greater detail. or increases in mortality in response to While climate events can damage or warming temperatures (e.g. forests in many ways ranging from ice Peñuelas, Lloret and Montoya, 2001), storms to tornadoes and hurricanes, the as do tree at the drier edges of emphasis here is on climatic water stress, their range of distribution (e.g. Jump, driven by drought and warm tempera- Hunt and Peñuelas, 2006). tures. Growth and mortality in wetter forests throughout the globe, however, from CLIMATE AS A DRIVER OF FOREST tropical moist forests to boreal systems, GROWTH AND MORTALITY are also highly sensitive to drought The Earth’s climate is recognized to be (Clark, 2004; Nepstad et al., 2007; undergoing significant human-caused Soja et al., 2007). Temperate forests changes, with global mean temperatures growing on productive sites may exhibit now outside the historic range of at least major growth declines, high levels of the past 1 300 years (IPCC, 2007). Mark- mortality and delayed multi-year effects edly greater shifts in climatic patterns from extreme drought and heat stress, are projected for the coming decades in as observed throughout Europe from the many regions, including much warmer 2003 drought and (Ciais et Craig D. Allen is with the United States temperatures and altered precipitation al., 2005; Breda et al., 2006). Warmer Geological Survey, Fort Collins Science Center, patterns that drive the availability of temperatures alone can increase forest Jemez Mountains Field Station, Los Alamos, water to plants. water stress independent of precipitation New Mexico, United States.

Unasylva 231/232, Vol. 60, 2009 44

amount (Barber, Juday and Finney, 2000; triggering many extensive forest insect the risk of cavitation through stomatal Angert et al., 2005). As such, it is not and disease outbreaks (Desprez-Loustau closure, which reduces water loss and apparent that any forests globally are et al., 2006; Raffa et al., 2008). subsequent tension within the xylem. Sto- safe from the impacts of drought. Climate-induced water stress may matal closure comes at a cost, however, as

Tree mortality commonly involves directly cause tree mortality through it prevents CO2 diffusion into the foliage, multiple, interacting factors, ranging short-term acute effects such as irrever- thereby reducing photosynthesis. Chronic from drought to insect pests and dis- sible disruption of water columns within water stress over long periods will weaken eases, often making the determination of tree stems and leaves (cavitation). Tree and ultimately kill , either directly a single cause unrealistic. Abiotic stress species vary widely in their resistance through carbon starvation or indirectly factors, however, commonly underlie and vulnerability to cavitation, a key through the attacks of pests such as bark forest health problems, with climate determinant of drought resistance. When beetles which overwhelm the diminished stresses thought to be a primary factor subject to water stress, trees minimize defences of such chronically starved trees (McDowell et al., 2008). Climatic condi- Example of drought-related mortality worldwide tions also directly affect the population dynamics of forest insects and fungal Region/country Forest type (e.g. Hicke et al., 2006). Thus, Africa some massive outbreaks of tree-killing Algeria Cedrus atlantica forest insects may be attributed to climate Namibia Aloe dichotoma drivers (Raffa et al., 2008). Regardless of Senegal Acacia, Cordyla, Nauclea and Sterculia species the exact mechanism, dieback is often a South Africa Dichrostachys, Pterocarpus and Strychnos species in the northeast non-linear process; it can emerge abruptly Uganda Uvariopsis and Celtis species in tropical moist forest at a regional scale when climatic condi- Asia and the Pacific tions exceed a tree species’ physiological Australia Eucalyptus and Corymbia species in the northeast thresholds of tolerance or trigger out- China Pinus tabulaeformia in east and central regions, Pinus yunnanensis breaks of insect pests (Allen, 2007). in the southwest Many reports link increased forest mor- India Acacia, Terminalia and Emblica species in the northwest tality to various combinations of notable Malaysia Dipterocarpaceae in tropical moist forests in Borneo dry and/or hot conditions, such as drought Republic of Korea Abies koreana in the tropics from severe El Niño events in Russian Federation Picea and Pinus species in temperate and boreal forests of Siberia 1988 and 1997–1998, the persistent warm- Europe ing and widespread drought over much of France Abies, Fagus, Picea, Pinus and Quercus species western since the 1990s, and the extreme heat wave and drought of Greece Abies alba in the north summer 2003 in western Europe. Norway Picea abies in the southeast Russian Federation Picea obovata in the northwest GLOBAL PATTERNS OF RECENT Spain Fagus, Pinus and Quercus species FOREST DIEBACK Switzerland Pinus sylvestris Forest mortality associated with drought Latin America and the Caribbean has been documented recently from all Argentina Austrocedrus and Nothofagus species in Patagonia wooded continents (Figure, p. 46) and from Brazil Atlantic tropical semi-deciduous forest in the southeast diverse forest types and climatic zones. Costa Rica Tropical moist forest Forest dieback is commonly reported near Panama Tropical moist forest the geographic or elevational margins of Near East a forest type or tree species (Jump, Hunt and Peñuelas, 2006), presumably near its Turkey Pinus and Quercus species in the central region historic thresholds of climatic suitability, Saudia Arabia Juniperus procera where the most sensitive response to cli- North America OCVGƀWEVWCVKQPUYQWNFDGGZRGEVGF Canada Acer, Picea, Pinus and species Some examples of forest mortality United States Abies, , Juniperus, Picea, Pinus, Populus, Pseudotsuga driven by climatic water and heat stress and Quercus species since 1970, based on a global review of Source: Allen et al., 2009 (where complete references can be found).

Unasylva 231/232, Vol. 60, 2009 45

more than 120 documented examples One consequence of substantial forest significant increases in problems asso- (Allen et al., 2009), are presented in the dieback is redistribution of within- ciated with forest health and dieback. Table. While forest dieback is commonly ecosystem carbon pools and rapid losses Given the dieback problems already noted in semi-arid regions where trees are of carbon back to the atmosphere. For reported under relatively modest recent near the physiological limits of dryness instance, climate-driven effects of forest increases in global mean temperature for growth (e.g. Fensham, dieback, insect and disease mortality (about 0.5ºC since 1970) and drying Fairfax and Ward, 2009), it is clear and fire impacts have recently turned climate in some areas (e.g. Seager et al., that climate-induced drought and heat Canada’s temperate and boreal forests 2007), far greater chronic forest stress stress have the potential to cause forest from a net into a net carbon and mortality risk could be expected dieback across a broad range of forest source (Kurz et al., 2008). Similarly, it is because much greater increases in mean and woodland types around the world. possible that “widespread forest collapse temperature (about 2º to 4ºC globally, Examples are particularly well docu- via drought” could transform the world’s and more in some areas) and significant mented from southerly parts of Europe tropical moist forests from a net carbon long-term regional drying in some places (Peñuelas, Lloret and Montoya, 2001; sink into a large net source during this are projected to occur by 2100 (IPCC, Breda et al., 2006) and in temperate and century (Lewis, 2005). 2007). Beyond changes in mean climate boreal forests of western North America, Given the potential risks of climate- conditions, other climate changes such where background mortality rates have induced forest dieback, increased man- as extreme droughts, elevated maximum increased rapidly in recent decades (van agement attention to adaptation options temperatures and longer-duration heat Mantgem et al., 2009) and widespread for enhancing forest resistance and resi- waves, which are projected to increase death of many tree species in multiple lience to projected climate stress can be in frequency and severity (IPCC, 2007), forest types has affected well over 10 expected, for example stand might be expected to exacerbate forest million hectares since 1997 (Breshears densities to reduce competition, selec- dieback. et al., 2005; Raffa et al., 2008). tion for different genotypes (e.g. drought A number of information gaps and resistance) or translocation of species to scientific uncertainties currently limit CONSEQUENCES OF BROAD-SCALE match expected climate changes. the conclusions that can be drawn about FOREST MORTALITY trends in forest mortality and the pre- Assessing the potential for, and conse- FOREST DIEBACK – AN EMERGING dictions that can be made about future quences of, extensive climate-induced GLOBAL TREND? climate-induced forest dieback. First, forest dieback is fundamentally impor- and ecologists have long known despite many national and even regional tant because trees grow relatively slowly that climate stress has major effects on forest monitoring efforts, there is an but can die quickly. A 100-year-old tree forest health. Awareness of, and inter- absence of adequate global data on forest may be killed by severe drought within a est in, climate-induced forest dieback health status (FAO, 2006). Reliable long- few months to a few years. As a result, is not new (Auclair, 1993; Ciesla and term, global-scale forest health moni- drought-triggered forest mortality can Donaubauer, 1994). It is known that toring, combining remote-sensing and result in rapid ecosystem changes over natural climate variation historically ground-based measurements, is needed huge areas, far more quickly than the triggered episodes of widespread forest to determine the status and trends of gradual transitions that occur from tree mortality (Swetnam and Betancourt, forest stress and mortality on the planet regeneration and growth. Land-use 1998). So, one might ask, is anything new accurately, as well as to understand eco- impacts such as anthropogenic burns or different occurring now? Certainly system responses after dieback events. and forest fragmentation, interacting the Earth is currently experiencing sub- Second, adequate quantitative knowl- with climate-induced forest stress, are stantial, rapid, directional global climate edge of the physiological thresholds of likely to amplify forest dieback in some change driven by major and pervasive individual tree mortality from chronic or regions, for example the Amazon Basin human alterations of the Earth’s atmos- acute water stress is available for only a (Nepstad et al., 2008). If current forest phere, land surface and waters (IPCC, few tree species (McDowell et al., 2008), ecosystems are forced to adjust abruptly 2007). Concurrent with these changes, and associated temperature sensitivities to new climate conditions through mas- climate-related forest mortality is appar- are largely unknown. Further, there is sive forest dieback, many pervasive and ently increasing in many parts of the little detailed understanding about the persistent ecological and social effects world. While the available evidence is place-specific sequences and ranges of will result from the loss of forest pro- not yet conclusive, it is possible that mean and extreme climatic conditions ducts and ecosystem services – including the increasing reports of dieback rep- that can trigger species-specific tree sequestration of atmospheric carbon. resent just the beginning of globally mortality in forests on real landscapes

Unasylva 231/232, Vol. 60, 2009 46

Localities with increased forest mortality related to climatic stress from drought and high temperatures

Drought-induced mortality of Pinus sylvestris, Andalucia, Spain (April 2006)

Severe mortality of overstorey aspen (Populus tremuloides) following the 2001–2002 drought in the parkland zone of Saskatchewan, Canada (August 2004) R.NAVARRO M. MICHAELIAN C.D. ALLEN C.D. ALLEN Mortality after warm drought in the early 2000s, Jemez Mountains, New Mexico, United States: left, Pinus ponderosa mortality (July 2006); right, mass mortality of Pinus edulis and scattered Juniperus monosperma survivors (May 2004)

A dust storm blows through a stand of Acacia albida in the Senegalese Sahel where dieback was documented in the last half of the twentieth century (1993) T. KITZBERGER Mortality of Nothofagus dombeyi in mixed N. dombeyi–Austrocedrus chilensis stand, induced by a warm drought in 1998–1999, northern Patagonia, Argentina (September 2004) P. GONZALEZ

Note: Only localities from the Table are shown; many additional localities are mapped in Allen et al., 2009.

Unasylva 231/232, Vol. 60, 2009 47

Climate-induced mortality of Pinus sylvestris, Valais, Switzerland (1999) Pinus yunnanensis stand, Yunnan Province, China, showing mortality induced by a drought that resulted in outbreaks of Tomicus yunnanensis and Tomicus minor shoot beetles from 2003 to 2005 (July 2005) A. RIGLING Z. ZHANG

Drought-induced death of Acacia aneura, eastern Australia (2007) R. FENSHAM FAO/FO-6298/G. ALLARD A. BRIKI H. CHENCHOUNI & M. BENSACI Cedrus atlantica mortality triggered by drought, Belezma National Park, Algeria, with surviving understorey including Quercus ilex (2007)

Dieback and decline of Juniperus procera, Saudi Arabia (March 2006)

Unasylva 231/232, Vol. 60, 2009 48

and potentially lead to extensive forest C.C., Buermann, W., Pinzon, J., Tucker, Ciesla, W.M. & Donaubauer, M.E. 1994. dieback. C.J. & Fung. I. 2005. Drier summers Decline and dieback of trees and forests:

Third, scientists lack adequate knowl- cancel out the CO2 uptake enhancement a global overview. FAO Paper No. edge of the feedback and non-linear induced by warmer springs. Proceedings 120. Rome, FAO. interactions between climate-induced of the National Academy of Sciences of Clark, D.A. 2004. Sources or sinks? The forest stress and other climate-related the United States of America, 102(31): responses of tropical forests to current disturbance processes, such as insect 10823–10827. and future climate and atmospheric outbreaks and fire, that can cause wide- Auclair, A.N.D. 1993. Extreme climatic composition. Philosophical Transcripts spread forest mortality (Allen, 2007). ƀWEVWCVKQPU CU C ECWUG QH HQTGUV FKGDCEM of the Royal Society of London, Series B, These scientific uncertainties about KP VJG 2CEKſE 4KO Water, Air and Soil 359: 477–491. fundamental tree mortality processes , 66(3–4): 207–229. Desprez-Loustau, M.-L., Marcais, B., represent a key limitation to more accu- Barber, V.A., Juday, G.P. & Finney, Nageleisen, L.-M., Piou, D. & Vannini, rate quantitative modelling of future B.P. 2000. Reduced growth of Alaskan A. 2006. Interactive effects of drought and climate-induced forest dieback (e.g. white in the twentieth century pathogens in forest trees. Annals of Forest Huntingford et al., 2008). Accordingly from temperature-induced drought stress. Science, 63: 597–612. they also limit the ability to predict the Nature, 405: 668–673. FAO. 2006. Global Forest Resources implications of dieback for the poten- Boisvenue, C. & Running, S.W. 2006. Assessment 2005 – Progress towards tial of global forests to sequester excess Impacts of on natural forest sustainable . FAO atmospheric carbon or, alternatively, productivity – evidence since the middle of Forestry Paper No. 147. Rome. to become carbon sources and thereby the 20th century. Global Change Biology, Fensham, R.J., Fairfax, R.J. & Ward, D.P. contribute to amplified climate change 12(5): 862–882. 2009. Drought-induced tree death in savanna. (Lucht et al., 2006). Breda, N., Huc, R., Granier, A. & Dreyer, Global Change Biology. (In press) Overall, additional monitoring of glo- E. 2006. Temperate forest trees and Hicke, J.A., Logan, J.A., Powell, J. & bal forest health and new research are stands under severe drought: a review of Ojima, D.S. 2006. Changing temperatures needed to improve scientific certainty ecophysiological responses, adaptation KPƀWGPEGUWKVCDKNKV[HQTOQFGNGFOQWPVCKP regarding risks of future climate-induced processes and long-term consequences. pine beetle (Dendroctonus ponderosae) forest dieback for more accurate input to Annals of Forest Science, 63: 625–644. outbreaks in the western United States. policy decisions and forest management Breshears, D.D., Cobb, N.S., Rich, P.M., Journal of Geophysical Research, 111: worldwide. X Price, K.P., Allen, C.D., Balice, R.G., G02019. Romme, W.H., Kastens, J.H., Floyd, Huntingford, C., Fisher, R.A., Mercado, L., M.L., Belnap, J., Anderson, J.J., Myers, Booth, B.B.B., Sitch, S., Harris, P.P., Cox, O.B. & Meyer, C.W. 2005. Regional P.M., Jones, C.D., Betts, R.A., Malhi, vegetation die-off in response to global- Y., Harris, G., Collins, M. & Moorcroft, change-type drought. Proceedings of the P. 2008. Towards quantifying uncertainty Bibliography National Academy of Sciences of the United in predictions of Amazon “die-back”. States of America, 102(42): 15144–15148. Philosophical Transcripts of the Royal Allen, C.D. 2007. Cross-scale interactions Ciais, P., Reichstein, M., Viovy, N., Society of London, Series B, 363(1498): COQPI HQTGUV FKGDCEM ſTG CPF GTQUKQP Granier, A., Ogee, J., Allard, V., 1857–1864. in northern New Mexico landscapes. Aubinet, M., Buchmann, N., Bernhofer, Intergovernmental Panel on Climate Change Ecosystems, 10: 797–808. C., Carrara, A., Chevallier, F., De (IPCC). 2007. Climate change 2007 –the Allen, C.D., Macalady, A., Chenchouni, H., Noblet, N., Friend, A.D., Friedlingstein, physical science basis. Contribution of Bachelet, D., McDowell, N., Vennetier, P., Grunwald, T., Heinesch, B., Keronen, Working Group I to the Fourth Assessment M., Gonzales, P., Hogg, T,. Rigling, A., P., Knohl, A., Krinner, G., Loustau, D., Report of the IPCC. Cambridge, UK, Breshears, D.D., Fensham, R., Zhang, Manca, G., Matteucci, G., Miglietta, F., Cambridge University Press. Z., Kitzberger, T., Lim, J.-H., Castro, Ourcival, J.M., Papale, D., Pilegaard, Jump, A., Hunt, J.M. & Peñuelas, J. 2006. J., Running, S.W., Allard, G., Semerci, K., Rambal, S., Seufert, G., Soussana, Rapid climate change-related growth A. & Cobb, N. 2009. Drought-induced J.F., Sanz, M.J., Schulze, E.D., Vesala, decline at the southern range edge of forest mortality: a global overview reveals T. & Valentini, R. 2005. Europe-wide Fagus sylvatica. Global Change Biology, emerging climate change risks. (Submitted reduction in primary productivity caused 12: 2163–2174. for publication) by the heat and drought in 2003. Nature, Kurz, W.A., Stinson, G., Rampley, G.J., #PIGTV#$KTCWF5$QPſNU%*GPPKPI 437(7058): 529–533. Dymond, C.C. & Neilson, E.T. 2008.

Unasylva 231/232, Vol. 60, 2009 49

Risk of natural disturbances makes future Filho, B. & Merry, F. 2008. Interactions J. & Naik, N. 2007. Model projections of an contribution of Canada’s forests to the among Amazon land use, forests and imminent transition to a more arid climate global highly uncertain. climate: prospects for a near-term forest in southwestern North America. Science, Proceedings of the National Academy of tipping point. Philosophical Transcripts 316: 1181–1184. Sciences of the United States of America, of the Royal Society of London, Series B, Soja, A.J., Tchebakova, N.M., French, 105: 1551–1555. 363(1498): 1737–1746. N.H.F., Flannigan, M.D., Shugart, H.H., Lewis, S.L. 2005. Tropical forests and the Nepstad, D.C., Tohver, I.M., Ray, D., Stocks, B.J., Sukhinin, A.I., Varfenova, changing earth system. Philosophical Moutinho, P. & Cardinot, G. 2007. E.I., Chapin F.S. & Stackhouse, P.W. Jr. Transcripts of the Royal Society of London, Mortality of large trees and lianas following 2007. Climate-induced boreal forest change: Series B, 361: 195–210. experimental drought in an Amazon forest. Predictions versus current observations. Lucht, W., Schaphoff, S., Erbrecht, Ecology, 88(9): 2259–2269. Global and Planetary Change, 56(3-4): T., Heyder, U. & Cramer, W. 2006. Peñuelas, J., Lloret, F. & Montoya, R. 2001. 274–296. Terrestrial vegetation redistribution and Severe drought effects on Mediterranean Swetnam, T.W. & Betancourt, J.L. 1998. carbon balance under climate change. YQQF[ƀQTCKP5RCKPForest Science, 47: Mesoscale disturbance and ecological Carbon Balance and Management, 1: 6. 214–218. response to decadal climatic variability McDowell, N., Pockman, W.T., Allen, Raffa, K.F., Aukema, B.H., Bentz, B.J., in the American Southwest. Journal of C.D., Breshears, D.D., Cobb, N., Kolb, Carroll, A.L., Hicke, J.A., Turner, M.G. & Climate, 11: 3128–3147. T., Sperry, J., West, A., Williams, D. & Romme, W.H. 2008. Cross-scale drivers of van Mantgem, P.J., Stephenson, N.L., Yepez, E.A. 2008. Mechanisms of plant natural disturbances prone to anthropogenic Byrne, J.C., Daniels, L.D., Franklin, J.F., survival and mortality during drought: CORNKſECVKQP VJG F[PCOKEU QH DCTM DGGVNG Fulé, P.Z., Harmon, M.E., Larson, A.J., why do some plants survive while others eruptions. Bioscience, 58(6): 501–517. Smith, J.M., Taylor, A.H. & Veblen, T.T. succumb to drought? Tansley Review. New Seager, R., Ting, M., Held, I., Kushnir, Y., 2009. Widespread increase of tree mortality Phytologist, 178: 719–739. Lu, J., Vecchi, G., Huang, H.-P., Harnik, rates in the western United States. Science, Nepstad, D.C., Stickler, C.M., Soares- N., Leetmaa, A., Lau, N.-C., Li, C., Velez, 323: 521–524. X

Unasylva 231/232, Vol. 60, 2009