| |
|
|
How can climate projections be used?
Climate projections can be used in various ways:
In detail
To indicate the expected direction of change:
- Climate
projections can suggest an expected direction of change, even if the
exact magnitude and timings of changes are not known or are not
important.
- For single
projections, this could be the case where the projected change is large
(considerably greater than natural variability) or where other model
results are in agreement (e.g. those presented in IPCC AR4),
Single projections are not necessarily well suited to indicating that
there is no trend, as they contain no quantification of the potential
range of plausible changes. Understanding the natural variability
(natural changes from year to year) is important when considering
future climate projections. The observed climate datasets can be used
to get an impression of natural variability for some climate variables.
- For probabilistic climate
projections, this could be the case where the probabilistic range is
significantly (or entirely) either greater or less than zero.
- For communications purposes, an expected trend might be sufficient. For example, it might be possible to say simply that average winter temperature will continue to increase.
- Where
the probabilistic range is relatively wide, sensitivity testing might
be an appropriate strategy, to test the implications of different
climate outcomes that are consistent with the complete range of
projections spanned by the curve. For example, a projected increase in
summer temperature that ranges from 0.5°C to 6.0°C could be
investigated with sensitivity testing of different amounts of change
between these extremes.
To indicate that there is no trend:
- Climate
projections might highlight that no clear trend is apparent, or that
projections are uncertain due to incomplete understanding of the
underlying processes.
- For
single projections, this could be the case where the projected change
is (close to) zero or where the projected change is less than the
natural variability of the variable in
question. Single projections are not necessarily well suited to
indicating that there is no trend, as they contain no quantification of
natural variability and uncertainty. Understanding the natural
variability (natural changes from year to year) is important when
considering future climate projections. The observed climate datasets
can be used to get an impression of natural variability for some
climate variables.
- For probabilistic climate projections,
there could be no trend where the probabilistic range spans zero, such
that increases and decreases are both equally plausible outcomes,
particularly where the range is relatively wide or approximately
symmetrical either side of zero change.
To rule out some changes:
- Probabilistic climate projections might suggest that certain projected changes are not likely to occur.
- For
probabilistic climate projections, this could be the case for changes
that are completely outside of the probabilistic range.
- For
communications purposes, ruling out certain changes might be a helpful
strategy. For example, it might be possible to say that colder winters can effectively be ruled out.
To justify the selection of a single estimate:
- In
some applications, a single projection of future climate change might
be required which would provide direction of change and an order of
magnitude of change.
- Probabilistic
projections provide far more information than a single estimate, but in
those cases where a single estimate is required they can be used to
justify this choice and place it in the context of alternative
outcomes.
To provide a plausible range:
- Climate projections might suggest a plausible range of future changes.
- Such an approach is not possible using a deterministic projection, and requires as a minimum several single projections (for example those from the IPCC AR4).
- Probabilistic projections allow a plausible range of projections to be defined. This range has been described as providing an envelope
of change which allows an initial scoping or screening of the potential
climate change impacts on a particular system or decision.
- It
is important to understand just what the range provided by a UKCP09
probabilistic projection includes (the main modelling uncertainties and
natural climatic variability, based on the UKCP09 method) and what it
does not include (projections resulting from alternative methods and
emissions scenarios). Even if the full range of UKCP09 emissions
scenarios are considered, they will inevitably only capture some of the
modelling uncertainty and emissions uncertainty that exists. This means
that future climate could be outside of the probabilistic ranges
reported in UKCP09.
- Full
recognition of this might mean that complementary strategies are
required for some applications or decisions, such as the development of
plausible worst case scenarios or sensitivity testing outside of the
UKCP09 probabilistic range.
To provide a probability distribution:
- UKCP09
provides projections as a probability distribution. The shape of this
probabilistic distribution can be used to provide more detailed
information about the relative likelihood of different amounts of
projected change. Probabilistic climate projections have been developed
to be integrated into impacts assessments using approaches that can
handle probabilistic information, such as formal risk assessment
techniques or Monte Carlo analysis.
- It
is important to understand just what the range provided by a UKCP09
probabilistic projection includes (the main modelling uncertainties and
natural climatic variability, based on the UKCP09 method) and what it
does not include (projections resulting from alternative methods and
emissions scenarios). Even if the full range of UKCP09 emissions
scenarios are considered, they will inevitably only capture some of the
modelling uncertainty and emissions uncertainty that exists. This means
that future climate could be outside of the probabilistic ranges
reported in UKCP09.
- Full
recognition of this might mean that complementary strategies are
required for some applications or decisions, such as the development of
plausible worst case scenarios or sensitivity testing outside of the UKCP09 probabilistic range.
|
|
You may also be
interested in…
|
|
