Evidence from marine ecosystems documents changes in species abundance and diversity and spatial distributions associated with air and ocean temperature rises (Chapters 5 and 6). Several studies document changes from the Antarctic region: Increases in chinstrap (Pygoscelis antarctica) penguins, stability or slow declines in Adelie (Pygoscelis adeliae) penguins, and declines in rockhopper penguins in recent decades are attributed in part to differential responses to warming climate conditions that are altering bird habitats (Fraser et al., 1992; Cunningham and Moors, 1994; Smith et al., 1999). Loeb et al. (1997) report effects on the Antarctic food web resulting from decreased frequency of winters with extensive sea-ice development; krill abundance is positively correlated with sea-ice extent, and salp abundance is negatively correlated. Smith (1994) reports a significant and relatively rapid increase in the numbers of individuals and populations of the only two native Antarctic vascular plant species at two widely separated localities in the maritime Antarctic.
Increases in abundance of southern macroinvertebrate species and declines in
northern species in a rocky intertidal community on the California coast are
consistent with recent climate warming (Sagarin et al., 1999). Warming annual
temperature has been suggested as a possible cause of increases in abundance
of plankton in the German Bight, but numerous factors, including regional eutrophication,
also have been noted (Greve et al., 1996). Lehman (2000) found that the distribution
of phytoplankton biomass in northern San Francisco Bay Estuary was influenced
by environmental conditions resulting from an interdecadal climate regime shift
between 1975 and 1993; precipitation regimes were primarily implicated, with
water temperatures also playing an important role. Ross et al. (1994) document
the loss of low-elevation pine forests in the Florida Keys because of rising
Evidence of observed impacts of regional climate changes from socioeconomic
systems is much sparser than from physical and biological systems, and methodologically
it is much more difficult to separate climate effects from other factors such
as technological change and economic development, given the complexities of
these systems. Vulnerability to climate change and climate variability is a
function of exposure and adaptive capacity (see Chapter 18).
Exposure varies from region to region, sector to sector, and community to community,
and adaptive capacity may be even more variable. The adaptive capacity of socioeconomic
systems also contributes to the difficulty of documenting effects of regional
climate changes; observable effects may be adaptations to a climate change rather
than direct impacts. Evidence of observed adaptation of many of these systems
to multiple stresses, including climate variability, suggests that complexities
inherent in socioeconomic systems could be a source of resilience, with potential
for beneficial adaptations in some cases. Studies that have explored some of
these complex relationships are briefly reviewed in the following subsections,
but they are not included in the summary tabulation or figure.
It has been proposed that observed impacts of changes in regional climate warming that are relevant to agriculture are related to increasing yield trends in Australia, lengthening growing seasons at high latitudes, improved wine quality in California, and expansion and advanced phenologies of agricultural pests. However, links between changes in regional climate variables and such changes are hard to prove because agriculture is a multifactored biophysical and socioeconomic system (see Chapter 5).
Nicholls (1997) analyzed Australian wheat yields from 1952 to 1992 and concluded that climate trends appear to be responsible for 30-50% of observed increases, with increases in minimum temperatures (decreases in frosts) the dominant influence (Nicholls, 1997); this conclusion has been questioned, however, by Godden et al. (1998) and Gifford et al. (1998). Possible confounding socioeconomic factors in identifying the effects of climate change on crop yields are responses of farmers to growing conditions (e.g., farmers may increase fertilizer application in good years, thereby exaggerating the impact of climate variables on yield), technological progress, changes in market structure, and changes in agricultural subsidies. Crop responses to increasing atmospheric CO2 concentrations also may affect yield trends.
Carter (1998) found that the growing season in the Nordic region (Iceland, Denmark, Norway, Sweden, and Finland) lengthened between 1890 and 1995 at all sites except Iceland, with likely but undocumented impacts on crop phenologies and timing of farm operations.
Nemani et al. (2001) relate warming at night and during spring in California over the period 1951-1997 (especially since 1976) to improved vintage quantity and quality.
Recent expansion and advances in insect phenologies may be associated with regional increases in mean or minimum temperatures (e.g., advances in flight phenology of aphid species in Britain) (Fleming and Tatchell, 1995; Zhou et al., 1995). Such increases in insect pests may be contributing to agricultural losses at least partially related to recent climate trends, but these effects have not been examined analytically.
Some changes in marine and coastal ecosystems have links to commercial fisheries,
but it is difficult to separate regional climate effects from human use of fish
stocks (see Chapter 6). Recent warming trends and coincident
overfishing and eutrophication have been noted in the English Channel and North
Sea, with potential future consequences for fish of high mass-market value (e.g.,
haddock, cod, plaice, lemon-sole codSouthward et al., 1995; O'Brien et
al., 2000). Diminished krill supplies in the Antarctic associated with decreases
in annual sea-ice cover and warmer air temperature documented by Loeb et al.
(1997) between 1976 and 1996 may have long-term negative effects on upper tropic
levels, affecting commercial harvests. These observations, in part, have prompted
the Commission for the Conservation of Antarctic Marine Living Resources (CCAMLR)
to request updated krill data currently used in krill management. CCAMLR manages
and sets limits on the international harvest of Antarctic krill (Loeb et al.,
Associations between regional climate trends and impacts related to energy, industry, and human settlements are sparse. One documented example is rapid coastal retreat along ice-rich coasts of the Beaufort Sea in northwestern Canada (Dallimore et al., 1996). Where communities are located in ice-rich terrain along the shore, warmer temperatures combined with increased shoreline erosion can have a very severe impact (see Chapter 6).
Determining the relationship between regional climate trends and impacts relating
to financial and insurance services is difficult because of concurrent changes
in population growth, economic development, and urbanization. Trends have been
analyzed regarding increased damages by flooding and droughts in some locations.
Global direct losses resulting from large weather-related disasters have increased
in recent decades (see Chapter 8). Socioeconomic factors
such as increased coverage against losses account for part of these trends;
in some regions, increases in floods, hailstorms, droughts, subsidence, and
wind-related events also may be partly responsible (see Chapter
8). Attribution is still unclear, however, and there are regional differences
in the balance of these two causes. Hurricane and flood damages in the United
States have been studied by Changnon et al. (1997), Changnon (1998), and Pielke
and Downton (2001). Pielke and Downton (2001) found that increases in recent
decades in total flooding damage in the United States are related to climate
factors and societal factors: increased precipitation and increasing population
and wealth. Hurricane damages, on the other hand, are unaffected by observed
climate change (Changnon et al., 1997; Changnon, 1998).
There is little evidence that recent trends in regional climates have affected health outcomes in human populations (see Chapter 9). This could reflect a lack of such effects to date or difficulty in detecting them against a noisy background containing other more potent influences on health (Kovats et al., 1999). The causation of most human health disorders is multifactorial, and the socioeconomic, demographic, and environmental context varies constantly. With respect to infectious diseases, for example, no single epidemiological study has clearly related recent climate trends to a particular disease.
Various studies of the correlation between interannual fluctuations in climatic conditions and the occurrence of malaria, dengue, cholera, and several other infectious diseases have been reported. Pascual et al. (2000) report a relationship between cholera and El Niño events. Such studies confirm the climate sensitivity of many infectious diseases, but they do not provide quantitative information about the impact of decadal-level climate change. Fingerprint studies examine the patterns of collocated change in infectious diseases and their vectors (if applicable) in simpler physical and ecological systems. This is an exercise in pattern recognition across qualitatively different systems.
One example is the set of competing explanations for recent increases in malaria in the highlands (see Chapter 9). A fingerprint study has hypothesized possible connections of plant and insect data, glacier observations, and temperature records to global climate change in high-altitude locations, with implications for patterns of mosquito-borne diseases (Epstein et al., 1998). Loevinsohn (1994) notes a connection between climate warming and increased rainfall with increased malaria incidence in Rwanda, whereas Mouchet et al. (1998) emphasizes the importance of nonwarming factors (e.g., land-use change in response to population growth, climate variability related to ENSO) in explaining variations in malaria in Africa.
Changes in disease vectors (e.g., mosquitoes, ticks) are likely to be detected before changes in human disease outcomes. Furthermore, a change in vector does not necessarily entail an increase in health impacts because of simultaneous processes related to the disease itself and the human population at risk. For example, the presence or absence of sanitation systems, vaccination programs, adequate nutritional conditions, animal husbandry, irrigation, and land-use management influences whether the presence of a disease in wild vectors leads to disease outbreaks in human populations. The effects of changes in frequency of extreme events may entail changes in health impacts, but these have not been documented to date.
Other reports in this collection