CLIM2002 Chap.3 Hazards, Projections and Confidence
Hazards, Projections and Confidence
Climate hazards may be acute events, chronic trends or combinations in which a changing baseline alters an extreme. Confidence in a projection comes from converging evidence about physical processes, observed trends, model performance and agreement across projections.
This chapter shows how to read climate graphics without jumping from a displayed variable to a claim about damage that still depends on exposure and vulnerability. Acute hazards occur over relatively short periods, while chronic hazards develop through persistent shifts in average or background conditions. The distinction affects which records, warnings, design standards and responses are relevant.
It is also possible for the two to interact: an extreme event can occur on top of a changing baseline, so an analyst may need to describe both the long-term shift and the event-scale departure. Projection reading begins with the display rather than its visual impression. Identify the variable, unit, geographic scale, reference period, future period, scenario and whether the colours show absolute conditions or change.
Check whether different panels use the same legend. Only then interpret direction, magnitude and pattern. A map or ensemble summary is evidence about a hazard, not a complete estimate of losses. Confidence should be built from several lines of evidence. Physical understanding asks whether the process producing the projected change is well understood.
Observations indicate whether a related signal can be detected and whether the record is adequate. Model evaluation asks whether relevant behaviour is represented at the necessary scale. Agreement considers the consistency of results without assuming that similar models are independent. These lines can support different conclusions about direction, magnitude and timing, which should be reported separately.
Decision making remains possible when evidence is incomplete. Robust measures can perform acceptably across several futures, staged investments can preserve flexibility, and monitoring can identify when a threshold is approaching. The correct response to uncertainty is therefore to show how it affects the choice, not to hide it behind an overconfident number or use it as a reason for indefinite delay.
What this chapter covers
- 01
Acute hazards occur as extreme events
- 02
Chronic hazards develop over longer periods
- 03
Changing baselines alter threshold frequency
- 04
Process understanding supports confidence
- 05
Observations and model skill provide separate evidence
- 06
Projection agreement concerns direction and magnitude
Worked example · free
Read a future heat map
- +1Identify variable, baseline, future period, unit and scenario.
- +1Assess process understanding, observations, model skill and projection agreement.
- +1Translate the heat signal into consequences only after adding exposure and vulnerability.
Key terms
- Acute hazard
- A damaging event that occurs over a relatively short period.
- Chronic hazard
- A slowly emerging change in typical climate conditions.
- Projection confidence
- A judgement informed by process, observations, model performance and agreement.
Hazards, Projections and Confidence FAQ
Is sea-level rise acute or chronic?
The rise in mean sea level is chronic, while storm surge is an acute event acting on the higher baseline. Coastal assessments often need both because the chronic shift changes the frequency and reach of extremes.
Does narrow model spread guarantee high confidence?
No. Agreement is one line of evidence. Sparse observations or weak representation of the relevant process can limit confidence even when model outputs cluster closely, so the assessment should report those limitations.
How should uncertainty change action?
Match the response to consequence and flexibility. Robust measures can work across several futures, staged investment can preserve options, and monitoring signposts can trigger stronger action as evidence or conditions change.
Why separate confidence in direction from confidence in magnitude?
Evidence can strongly support whether a hazard increases or decreases while leaving a wide range for how much and when. Reporting those dimensions separately preserves useful information and helps a decision maker choose robust measures, margins and review points.
Assessment move
Use a four-column confidence audit for process understanding, observations, model performance and projection agreement. For every graphic, annotate the variable, baseline, future period, scenario, unit and limitation. Begin practice with a blank reading protocol before looking at the colours. Write what is measured, where, when and relative to what.
Then record whether the display gives an ensemble mean, a range, individual models or scenario differences. Summarise direction, magnitude and timing in separate sentences so strong evidence for one dimension does not inflate confidence in the others. Build a hazard card for each major example with fields for time scale, affected process, supporting observations, model limitation and decision relevance.
Compare cards to learn why confidence can differ among hazards and spatial scales. For an applied question, add the missing impact pathway: exposed system, vulnerability, threshold and consequence. Propose a response that remains useful across the displayed range, and state a measurable signpost that would justify changing course.
When two sources appear inconsistent, check their variables, periods, resolutions and legends before concluding that the science conflicts. In your final review, underline every confidence word and ask which evidence line supports it. Replace unsupported precision with a bounded statement that tells the reader what is known, what varies and how that variation enters the decision.
