19.7. Limitations of Methods and Directions for Future Research
This section discusses the strengths and limitations of the analytic approaches
used to address the reasons for concern, mainly with regard to whether they
can, with the confidence levels given, indicate the severity of impact or risk
as a function of increase in global mean temperature. This discussion identifies
key uncertainties inherent in each method and offers directions for future research
that could improve our confidence in the results produced with each approach.
The organization of this section parallels that of the previous sections of
this chapter. The strengths, limitations, uncertainties, and directions for
each approach are discussed in the same order in which they were discussed in
the preceding sections. However, integrated assessment frameworks are considered
separately from aggregate approaches. Last is a discussion of integration across
methods and reasons for concern.
Advantages: Because observations are based on observed effects rather than
models, they can be used to indicate whether climate change is causing impacts
and whether impacts lead to positive, negative, or indeterminate outcomes. They
also can be used to validate hypotheses and models that formalize hypotheses
on cause and effect.
Disadvantages: The problem with relying on observations to determine the severity
of impacts or risk from climate change is that there has been only 0.7°C
of mean global warming over the past century (although some regions have experienced
much more warming). Because many impact thresholds may not be crossed until
greater magnitudes or rates of warming are reached, it is not clear how to interpret
an observed effect of warming or a group of such observations. Such observed
impacts to date often will be of only minor consequence, even though they may
tend to confirm our understanding of impact processes. Moreover, lack of observed
impacts may be simply because climate change has not yet reached critical thresholds
for such effects. Finally, attribution of causality is very difficult with observed
effects or groups of effects. One must be able to demonstrate that a regional
change in climate is a significant cause of an observed effect and that the
regional change in climate is linked to global climate change.
Uncertainties: Uncertainties include the magnitude of climate change that has
occurred, the extent to which impacts can be attributed to climate change that
has occurred, and whether the relationship between climate change and possible
impacts is linear or nonlinear and continuous or discontinuous.
Research Needs: For climate change impact detection to advance, there is a
need for continued, improved, and augmented data collection and further development
of analytical techniques. Geographical diversity is needed to balance the current
bias of study locations in North America and Europe; more observation studies
are needed in developing countries, with emphasis on those where physical, biological,
and socioeconomic systems have higher vulnerability to climate change (see Chapter
Because climate and impact systems are linked over a range of temporal scales,
longer time series of data allow better understanding of the relative magnitudes
of short- and long-term responses (Duarte et al., 1992; McGowan et al., 1999).
Large-amplitude temporal changes usually involve large spatial dimensions, so
broad-scale spatial/temporal studies are necessary as well. Satellite measurements
of the Earth's surface provide a very useful monitoring capability for
ocean, ecosystem, and land-cover changes. For example, satellite measurements
of the Earth's surface offer the potential for aggregation of observed
impacts with regard to broad-scale ecological responses such as vegetative responses
to increasing lengths of growing seasons (e.g., Myneni et al., 1997), complemented
by meteorological and vegetation data (e.g., Schwartz, 1998).
For ecosystem impacts, continuing observations are needed at sites where studies
already have been conducted, at long-term ecological research sites (e.g., Chapin
et al., 1995), and in protected areas. Programs that provide continued long-term
monitoring of marine and terrestrial environments also are important (Duarte
et al., 1992; Southward et al., 1995). Large-scale spatial/temporal ecosystem
studies are necessary because effects from local changes cannot be extrapolated
to large areas without evidence (McGowan et al., 1999; Parmesan et al., 1999).
Definition of indicator species or systems is a useful element of detection
studies (e.g., Beebee, 1995; Nehring, 1998; Cannell et al., 1999). Coupled with
monitoring programs, such data may then provide a consistent set of evidence
with which to study past, present, and future impacts of climate changes.
A further critical research need is to strengthen analytical tools for understanding
and evaluating observed climate change impacts. Robust meta-analyses of studies
that present good quality, multivariate data from a diversity of settings around
the world will help to define further the global coherence among impacts now
observed. Care also must be taken to ensure that the sample of studies is representative
across time and space, is not biased in its reporting, and uses appropriate
statistical tests. Also needed is development of methods to analyze differential
effects of climate across a range or sector. Individual and grouped studies
need to address possible correlations with competing explanations in a methodologically
Also needed are refinements in the fingerprint approach (e.g., Epstein et al.,
1998), including more precise definition of expected changes and quantitative
measurement techniques, similar to that used in detection of climate changes
(see TAR WGI Chapter 12). For climate,
fingerprint elements include warming in the mid-troposphere in the southern
hemisphere, a disproportionate rise in nighttime and winter temperatures, and
statistical increases in extreme weather events in many locations. These aspects
of climate change and climate variability have implications for ecological,
hydrological, and human systems that may be used to define a clear and robust
multidimensional "expected impact signal" to be tested in a range
of observations. A more refined and robust fingerprint approach may aid in the
study of difficult-to-detect, partially causal climate effects on socioeconomic
systems such as agriculture and health.
19.7.2. Studies of Unique and Threatened Systems
Assessments of unique and threatened systems tend to be based on studies of
particular exposure units such as coral reefs, small islands, and individual
Advantages: These studies contain richness of detail and involve many researchers,
often from developing and transition countries. In contrast to aggregate studies,
studies of exposure units can be used, at least in principle, to analyze distributional
effects by focusing on impacts on particular systems, species, regions, or demographic
Disadvantages: One of the main disadvantages is that exposure-unit studies
often are not carried out in a consistent manner. Exposure-unit studies often
examine related sectors in isolation and do not examine linkages or integration
among sectors and regions; for example, studies of the effects of climate change
on ecosystems or individual species often are conducted without examining the
potential effects of societal development on such systems. Local processes and
forces (e.g., urbanization, local air pollution) often can be more important
than global ones at the local scale, complicating the task of measuring the
influence of global climate change at the local scale.
Another key disadvantage is incompleteness of coverage. For example, in spite
of many and extensive country studies, there still are many gaps in coverage
in terms of countries, regions within countries, and unique and important potential
impacts that have not been assessed. The choice of exposure units may not necessarily
cover the most vulnerable systems. Topics such as impacts on biodiversity or
unique ecosystems often are not covered. There also has been little attention
to impacts on poor and disadvantaged members of society. Even where particular
critical exposure units have been covered, there may be just a single study.
Drawing conclusions with high confidence on the basis of one study may be inappropriate.
Uncertainties: Uncertainties include the likely magnitude of climate change
at the spatial resolution required by the study of the particular unique and
threatened system, masking of global change effects by nonclimate factors, the
degree of linearity/ nonlinearity in the relationship between stimulus and response,
and the degree to which results from individual studies can be extrapolated
Research Needs: It would be desirable to have more studies of individual systems,
according to some set of priorities concerning the likely immediacy of the impacts.
Additional work on standardizing methods and reporting of results also would
be extremely useful. It also would be useful to devote more effort to integration
of results from existing studies. Again, it would be especially useful to increase
monitoring of changes in organisms, species, and systems that have limited range
now or are near their limits and to try to separate out or consider other causal
mechanisms such as local air pollution, loss of habitat, and competition from
invading pests and weeds.
19.7.3. Distributional Impacts
Advantages: Distributional impact studies draw attention to likely heterogeneity
in impacts among different regions and social and economic groups. They also
help to identify and assess the situation of the "most vulnerable"
people and systems. Thus, such studies bring equity considerations to center
Disadvantages: Distributional impact studies require regional climate change
projections and impact projections at the regional to local scale, where GCMs
may not be very accurate. They also require projections of demographics and
socioeconomic structure over a long time horizon.
Uncertainties: Research into the distribution of impacts of climate change
is recent (see Section 19.4). There are some findings
on which there is virtual unanimity. Some findings are broad conclusionssuch
as that more resource-constrained regions are likely to suffer more negative
impacts, as are people whose geographic location exposes them to the greatest
hazards from climate change. (Such people often live in regions with marginal
climate for food growing or in highly exposed coastal zones.) Others are more
specific but to date have been more conclusive with regard to the direction
of different impacts among regions, rather than the magnitudes. For example,
we know that impacts in developing low-latitude countries are more negativein
part because those countries tend to be operating at or above optimum temperatures
alreadyand, in some cases, in regions where rainfall will decrease, leading
to water stress. There also is limited capacity for adaptation in these areas.
In some mid-latitude developed countries, agriculture would benefit initially
from warmer conditions and longer growing seasons. Beyond such sweeping statements,
uncertainties are vast. Resource constraints and (climatic) marginality are
multidimensional and complex phenomena. Currently, it is not known which components
of resource constraints or climatic marginality are more important or which
components may compensate for others or may have synergistic effects. There
are suggestions in other literature, but these have not been systematically
applied to the impacts of climate change, conceptually or empirically.
In sum, there is virtual consensus about the broad patterns. There is much
less knowledge about the details, although that situation is slowly improving.
Research Needs: Development of appropriate indicators of differences in regional
impacts and ways of comparing them across regions and socioeconomic groups would
be extremely useful. Improved methods for characterizing baseline demographics
and socioeconomic conditions in the absence of climate change or climate change-motivated
policies also would be useful. There is a need to quantify regional differences
and to develop estimates of the cost of inequity in monetary or other terms
(e.g., effect on poverty rates and trade, social and political instability,
and conflict). More accurate projections of regional climate change would increase
confidence in predictions of regional climate change impacts.