Ocean Ecosystems. Satellite data indicate that there has been a decline in ocean NPP since 1999. A dominant factor has been significant NPP reductions in tropical regions due to natural variability. The global declines are partially offset by increased productivity in Arctic waters. These changes suggest that regional ocean ecosystems can be very sensitive to climate change and variability. As oceans continue to warm, further increases in productivity in mid to high latitudes will likely exceed any reductions in the tropics. This implies that the current decline in total ocean NPP will not persist. Experts caution that these changes need to be carefully monitored694-696.
There are also reports of large changes in regional ocean ecosystems that appear to be climate related. In Bering Sea ecosystems, for example, a recent regional change in climate has caused a noticeable shift from species that feed off sea bottoms in shallow waters, such as ducks, walruses and gray whales, to fish that feed within the water column. This change may be linked to a shift in phases of the Arctic Oscillation. In the southwest Pacific Ocean, there is evidence of enhanced growth of several fish species in upper waters that have warmed, but a decline in deeper waters that have cooled. Similarly, in the northwest Atlantic, the combination of overfishing and a regional cooling of shelf waters due to changes in ocean circulation and an increase in the influx of freshwater from the Arctic have contributed to a collapse in the abundance of some important local species, particularly Atlantic cod. There are also indications that, over the past 50 years, tropical regions of oxygen depleted waters in the central and eastern Atlantic and the equatorial Pacific have expanded and intensified, to the detriment of marine organisms within these regions. These examples are further reminders of the vulnerability of regional ocean ecosystems to future changes caused by a warming global climate697-701.
Surveys along the Pacific coast of North America show increased upwelling of abnormally acidic waters, indicative of broader ocean acidification due to rising carbon dioxide concentrations and a general decline in the pH of coastal waters. There has been a corresponding decrease in productivity of local calcareous species relative to non-calcareous ones702-703.
Analysis of global production of six key food crops indicates that there has been a small but significant decline in yields of wheat, maize and barley between 1981 and 2002. This implies that the negative impacts of the combined environmental stress of high temperatures, drought and other factors on food production have exceeded the benefits of enhanced CO2 fertilization and technological advances704.
British hospital admission reports show that the respiratorial virus seasons in that country end about 2-3 weeks earlier for every degree rise in seasonal temperature. There does not seem to be a corresponding change for the onset of the flu season705.
In northern Canada, changing river, lake and ocean ice conditions are altering traditional winter transportation and hunting patterns. This has resulted in a change in the consumption of natural foods by Inuit communities706.
5.3 Detection/Attribution Over the past few years, numerous studies have examined the various trends in climate noted in the preceding sections and sought to attribute these to specific causes. The following is a brief summary of the current status of such attribution studies.
5.3.1 Temperature Various statistical and modeling studies continue to show that, while internal natural climate variability has an important influence on climate trends on decadal time scales, particularly at the regional scale, global trends in temperature over the past century cannot be adequately explained by internal climate variability. For example, one study indicates that the recent cluster of high global average surface temperatures has less than a 0.001 probability of occurring by chance. Furthermore, while there is abundant evidence that most of the warming in the early part of the 20th century can be attributed primarily to solar and volcanic forcings, changes during the past three decades are almost entirely attributable to anthropogenic forcings associated with changes in atmospheric concentrations of greenhouse gases and aerosols and related feedbacks. In fact, in the absence of such human influences, net warming of the global climate over the past four centuries would likely have been only about 0.1°C. Multi-model simulations also confirm that the observed changes in temperatures in both the Antarctic and Arctic regions are at least partly attributable to human influences. Likewise, the large increase Atlantic and Pacific Ocean SSTs, the upper ocean warming in all ocean basins, as well as the amplified warming in the mid-depth Southern Ocean relative to upper ocean warming are all consistent with model simulations of response to anthropogenic forcing. Some experts caution, however, that statistical analyses of the relationships between observed temperature change patterns and both empirical data for natural and anthropgenic forcings differ significantly from model simulations. Hence the latter may still be missing some important regional feedbacks270,410,420,707-714.
Regional temperature trends are more difficult to attribute to external forcings. This is partly due to the regional influences of precipitation and changes in atmospheric circulation on temperature. However, improved techniques in identifying and removing the effects of natural variability (or noise) within the climate system together with the increasing strength of the anthropogenic forcing signal now make attribution of regional temperature trends possible. Analyses show that the pattern of changes in climate around the world over the past half century can be explained by the combined effects of greenhouse gas and aerosol forcing, but not by natural forcings or variability. Human influences can now be consistently detected at regional scale, and the added contribution of natural climate oscillations on regional climate can also be better explained. Studies have shown, for example, that changes in NAO and Aleutian Low behavior can explain much of the enhanced warming over land masses above 40°N, relative to the global mean trend715-718.
A few researchers continue to argue that many of the assumptions involved in such attribution studies are questionable, and that the conclusions are thus to be accepted with caution. They note, for example, that human development may have also affected global temperatures through albedo effects of land clearing, that there are still biases in the global temperatures records used in these studies, and that attribution studies assume the models used in the studies are perfect in their projections. However, others note that these concerns have been addressed in climate record reconstructions and in the error margins of study results, and that the conclusions with respect to the strong human fingerprint in recent global temperature trends remains robust719-721.
5.3.2 Hydrology Model simulations also indicate that many of the observed changes in the global hydrological cycle are consistent with those expected under enhanced greenhouse gas and aerosol forcing. For example, while recent trends in global surface specific humidity and in total increases in atmospheric moisture content over oceans can be partly explained by natural variability linked to ENSO and other oscillations, a significant component of these trends is also attributable to anthropogenic forcing. Recent patterns of droughts like that in the Sahel also appear to be linked to changes in regional SST gradients associated with anthropogenic radiative forcing (particularly that caused by atmospheric aerosols), although local feedbacks appear to amplify the pattern. Likewise, increased global precipitation and river water discharge into oceans, as well as the global pattern of changes in precipitation are all consistent with model simulations responding to recent anthropogenic forcing (including the direct effects of enhanced CO2 concentrations on plant physiology), but cannot be correctly simulated without such forcing. Even some regional hydrological changes (such as those for snowfall over Asia, for precipitation and runoff in the Arctic and for changes in the hydrological cycle in the western United States) are now at least partly attributable to human influences221,431,484,493,494,497,722-724.
Figure 14. Time series (left panel) of observed annual zonal mean precipitation anomalies in 10° latitude bands (thin black trace) together with ensemble mean annual zonal mean precipitation anomalies in the 50 available ALL simulations (thin blue trace). Straight dashed black and red lines indicate the trends. Green (or yellow) shading identifies latitude bands with increasing (or decreasing) trends in both observations and models; grey shading indicates disagreement between observed and simulated trends. The map (right panel) indicates the different 10° latitude bands and whether trends agree in sign. Areas with insufficient data are shown in white. Only land precipitation data are used. (Zhang et al., 2007, ref. #723).
Past changes in global atmosphere and ocean circulation patterns have been dominated by the combined effects of a variety of well known periodic fluctuations, including the North Atlantic Oscillation (NAO), the Atlantic Multi-decadal Oscillation (AMO), the Pacific North America (PNA) pattern, the El Niño Southern Oscillation (ENSO), the Arctic Oscillation (AO), and the Antarctic Oscillation (AAO). These in turn are influenced by, and influence, changes in ocean circulation patterns, including overturning rates. Model studies indicate that some of these have low frequency components that can be replicated throughout the past century without external forcing. However, there is increasing evidence that recent changes in regional pressure patterns and winds, and related shifts in temperature, cannot be fully explained by these natural fluctuations. In the Arctic, for example, dramatic reductions in ice cover, rather than AO changes, appears to be a dominant contributor to recent meridional changes in pressure patterns in the region. The sea ice decrease, in turn, has been at least partly linked to human induced global warming. In Antarctica, human induced stratospheric ozone depletion may now be amplifying the natural AAO pattern611,725-727.
Recent increases in mid North Atlantic salinity are also consistent with that expected due to anthropogenic forcing. However, the forcing signal for similar changes in its sub-polar regions still appears to be lost in the noise of climate variability728.
Several studies now suggest that some recent extreme heat waves are sufficiently unusual to be attributable to human induced global warming. Multiple model simulations show, for example, that the record setting temperatures over the United States in 2006 cannot be attributed to ENSO influences, but that more than half of the anomaly can be explained by the climate effects of greenhouse gas and aerosol forcing. While past studies have suggested that the record setting European summer heat wave of 2003 also had a human-induced climate change link, one recent study indicates that the record temperatures only occurred near the surface and did not extend very far into the troposphere. Hence, local feedbacks such as that related to low soil moisture may have been an important contributor to its extreme nature729-731.
Recent shifts in other climate variables, including increases in regional daily maximum and minimum temperatures, declines in frost days, and increases in degree days, are also increasingly difficult to attribute entirely to natural variability and are therefore also likely linked to anthropogenic forcings732.
Model simulation ensembles suggests that 50% or more of the rapid decline in late summer sea ice cover in the Arctic since 1979 can be attributed to external forcing526a,b.
Analyses of a broad range of studies into trends within the natural physical and biological system show exceptional agreement of the global and continental scale patterns of change within these systems with those for temperature. These include changes in the cryosphere, in surface hydrology and in the phenology of species. Thus, they are almost certainly not due to natural variability. About 90% of the changes are in the direction projected by climate change impact studies based on model simulations of climate response to greenhouse gas and aerosol forcing672,733-734.
6.0 Impacts 6.1 Hydrological resources and events Changing precipitation, together with increased evaporation due to rising temperature, can have large influences on water resources and soil moisture. Since soil water holding capacity can vary significantly, runoff response to changing precipitation is non-linear. One study, for example, projects that a 10% decline in rainfall would only cause a 17% drop in runoff in areas of Africa that receive more than 1000mm of rain each year, but a 50% decrease in areas where annual precipitation is less than 500mm. Over global land areas, on average, the projected increase in evaporative losses due to rising temperatures is expected to exceed that for rainfall. Hence, average global soil moisture is projected to decrease, with the spatial extent of areas with severe soil moisture deficits and the frequency of short term droughts likely to double over the next century. That for longer term droughts lasting more than 6 months may triple. Hardest hit regions will likely include the Mediterranean, west Africa, central North America and central Asia. Some models suggest that the changes in soil moisture in mid-latitudes will be relatively small. However, because of the large natural variability of soil moisture over time and space, additional changes due to human-induced climate change may not be detectable until 2050332,735-736.
Across North America, megadroughts are likely to become more common in the interior drylands and the extent of areas currently with desert-like conditions may expand. Models are less consistent with respect to whether or not these regions of severe drought might expand into southern Canada. Changing seasonality and more intense precipitation events can also alter river flow cycles and flood risks. In the Châteauguay River basin of Quebec, for example, peak flows in spring and summer are projected to occur earlier and the spring flood may become of lower magnitude if reduced snow melt exceeds any increase in rain. In the region of southern BC and northern Washington State, projected runoff appears to have an advanced peak-to-minimum seasonal cycle under warmer climates but little net change in total annual flow or on regional groundwater and aquifer recharge. One study, using a simple trend extrapolation for the Great Lakes water resources, projects an increase in lake levels and flows in coming decades. However, this is not in agreement with other studies using more complex climate and lake hydrology models that project decreasing Great Lakes water levels and flows736,737-743.
In polar regions, local factors such as ice cover change and earlier snow melt can significantly modify the effect of global hydrological and circulation changes on regional precipitation and runoff. In the Arctic, spring freshets are expected to be lower, and lake and river ice cover durations to shorten. Decaying permafrost in the region will also affect local drainage in complex ways, and increased vegetation will likely alter water chemistry. In Antarctica, precipitation increases over the continent will likely exceed evaporation changes, resulting in a small net accumulation of snow on the ice sheet surface that is likely to slightly offset sea level rise due to iceberg calving and other losses at the margins of the ice sheet744-746.
Characteristics of precipitation will also change. Satellite data indicate that deep convective activity in the tropics increases by about 45% per degree of warming. Over moist continental areas such as the southern and eastern United States, increased latent heat in the boundary layer will also cause updraft to increase significantly. This increase may not affect the number of thunderstorms, but will increase the potential for their severity. In general, such enhanced convection also increases the intensity of rainfall. Models generally agree that rainfall intensity increases on average by about 2% per degree of warming, but that intense events are enhanced by about 30% per degree of warming. Greatest increases in both precipitation intensity and frequency appear likely to occur in wet regions of the world and in the high Arctic, while dry regions will likely see less activity. One study, using outputs from a single model, also projects that intense daily winter precipitation will increase in frequency for much of North America, particularly in the south-west and central United States, but decrease over the Canadian Prairies. Another single model study suggests that an increase in convective type precipitation will likely cause extreme short duration rainfall events in Southern Quebec to double in frequency by mid century, and longer 12-24 extreme events to become 50% more frequent. Comparisons of satellite observations of precipitation intensity with changes in temperature imply that these model projections of precipitation intensity may be too low633,747-752.
Changes in hydrological processes will also affect water quality in lakes and streams, often in surprising ways. One experimental field study found, for example, that biomass productivity within cold, boreal lakes increased much more under the combined effects of climate change and acid deposition than if each factor was considered individually. The positive impacts, however, were largely due to the stress tolerance of one specific species, and hence may have limited applicability elsewhere. Studies also indicate that, if future climate change also increases the risk of very dry years, the recovery of Ontario lakes from acidification would be much slower than under a less variable climate. That is because low water levels during dry years allow sulphur stored in sediments to be re-oxidized and thus re-introduced into the water. Increased decomposition of organic matter can also increase NO3 content in acidified lakes753-755.
6.2 Agriculture Multiple scenario projections for the response of global crop yields to future climate change suggest that, without adaptation measures, total wheat production may increase slightly by 2030. Barley production is likely to decline slightly and that for maize could decrease by between 4 and 28%. These projections are particularly sensitive to future changes in temperature. Greatest vulnerability for related food shortages will be in those regions, such as south and south-east Asia and southern Africa, where local populations rely on crops that are projected to decline in yield. Unless there are significant offsets through adaptive measures, there could be a 50% increase by 2030 in undernourished people in Africa alone. In India, warmer temperatures could cause average farm productivity to decrease by 4 to 26% by 2100, while that in Brazil could decline by 1 to 39%. Recent trends in crop yields are also a reminder that the ability of CO2 fertilization and advances in agricultural technologies to mitigate such impacts may have been overestimated in related impact studies. Furthermore, such studies must also consider the concurrent effects of other environmental and economic stressors. For example, some plants such as soybeans appear to have lower defences against pests under high atmospheric CO2. Warmer winters and higher degree-days may also increase the expansion and outbreaks of pests. In sub-Saharan African countries, increased local purchasing power may reduce hunger risks in some of the impacted regions. However, in other regions, decreases in purchasing power may exacerbate the problem756-764.
Warmer temperatures can also cause stress to crops growing in southern regions of North America. For example, most of the premium wine production regions in the United States could fail to meet requirements for quality wine. However, temperature increases can provide significant benefits for agriculture in cooler regions of the continent. In Southern Quebec, for example, productivity in the northern and higher elevation regions of agricultural lands will improve significantly, although there may be declines for potato and wheat production in the more southerly regions. However, hot days and water shortages may be concerns in some areas. In Ontario, cool season crops such as cole can decline in yield by as much as 10% for every 10 growing season days with temperatures exceeding 30°C. In the Okanagan region of British Columbia, warmer temperatures and longer growing seasons are projected to increase water demand for crop production by about 50% by 2085, with greatest increases during times of lowest water supplies765-769.
6.3 Natural Ecosystems 6.3.1 Global Ecosystems On average, growing seasons around the world are projected to lengthen by more than one month by 2100. Although warmer temperatures will benefit ecosystems and individual species in some regions, it will affect others negatively. The long-term response of ecosystems may also be differ significantly from that projected for the short term. Furthermore, individual species will also be affected by the impact of changing climates on their habitats, food supply and predators, and on the spread of diseases and pests. Migrant bird species, for example, would need to cope with changes in both summer and winter habitats, which may respond to changing climates at differing rates. Deforestation activities in tropical regions will exacerbate the risks to habitats. The boundaries of many bird species could shrink dramatically, and be displaced by as much as 1000 km. Dynamic vegetation models suggest that ecosystem productivity will likely increase in northern Europe and the Arctic but decline in other parts of Eurasia, Canada, China, Central America and the Amazon. Semi-arid and savanna regions are likely to expand. There is a significant risk that current global land carbon sinks could become a source by 2100. In North American forests, the average distribution range of plant species by 2100 is projected to decrease by 12 to 58% and be displaced northward by 700 km. Furthermore, wild fires may be significantly enhanced in many regions due to warmer temperatures and reduced rainfall, with far reaching implications for forest species. In general, species richness is expected to decline, with those species with limited range and special habitats being hardest hit672,770-776.
Adaptability of flora and fauna to projected changes in CO2 concentrations and climate vary significantly. Short lived species, such as insects, algae and annual plants, can be highly adaptable to shifts in climate, but are more negatively affected by increased climate variability than long-lived species, and vice versa. Some species adapt through effective dispersal processes, which facilitate migration. However, various modeling studies suggest that many endemic plants and vertebrate species in key biodiversity hot spots of the world could be at risk of extinction. Those species that are currently sparse and have narrow temperature or habitat ranges, or have limited genetic variability are most at risk. Some researchers caution that there is little evidence for high extinction rates during the last deglaciation. They argue that adaptive capacity at the species level may be greater than many models suggest777-784.
The combined effects of changing climates and CO2 and nitrogen fertilization will also change plant phenology and the physical environment. Some of these effects can be offsetting. For example, while warmer temperatures tend to advance flowering dates of grass species, higher CO2 concentrations tend to delay flowering. Elevated CO2 concentrations also appear to delay the end of the growing seasons of species such as aspens. Lake recovery from acidification in parts of eastern North America and Europe may also be affected. Furthermore, direct CO2 effects alone could increase continental runoff by 6% over next century through reduced evapotranspiration. This would ameliorate projected water shortages in some regions of the world769,785-787.