6.2 Climate change—causes and impacts

Weather is a short-term atmospheric state; climate is a long-term pattern described by averages, ranges or frequencies.

Classify the observation by time window and statistic. Temperature and precipitation are core seasonal variables, but climate also includes variability and extremes.

Today’s storm is weather; a 30-year rise in summer rainfall variability is climate evidence.

No. It is weather noise; test the long-term distribution.

Climate is not a single average temperature and weather is not irrelevant—it is the data climate summarizes.

Emissions are a flow added over time; atmospheric CO₂ concentration is the stock left after emissions and natural removals are combined.

Anthropogenic CO2 emissions began raising atmospheric concentration from the Industrial Revolution in late 18th-century Europe; the global emission rate accelerated especially after 1950 as industrialization and population increased. Concentration still rises whenever total emissions exceed ocean and ecosystem removal.

If annual emissions fall but remain greater than removals, atmospheric CO2 continues to increase more slowly; concentration falls only when removals exceed emissions for long enough.

Use both views: the historical timeline identifies the human source, while the stock–flow balance explains why slower emissions growth is not the same as falling concentration.

Do not confuse a lower inflow with a smaller stock. A bathtub still fills when the tap slows unless the drain removes water faster than the tap adds it.

A proxy is an indirect record of past climate; interpret its measured variable before claiming a temperature trend.

Ice cores preserve past gas concentrations and isotopic evidence; tree rings record growth conditions; deposited sediments preserve layered biological or chemical signals. Together, records spanning the past 800,000 years show glacial-cycle co-variation between atmospheric CO2 and global temperature.

A narrow tree ring may reflect drought, cold or another stress, so calibrate each proxy and compare independent records before attributing a temperature change.

Check dating, calibration, local representativeness and missing intervals. Positive CO2-temperature correlation supports a relationship but does not by itself identify the full causal mechanism.

A positive correlation is evidence of co-variation, not proof that one proxy caused the other.

The chain is natural greenhouse effect → human enhancement → warming → changes in climate hazards and impacts.

Short-wave sunlight enters; outgoing infrared is absorbed by added GHGs; energy balance shifts until the system warms. Impacts then depend on circulation, water, ice, ecosystems and society.

More CO₂ slows infrared loss, raises temperature, then increases heat extremes and water stress in some regions.

As an observed hazard/impact, not as the cause of greenhouse-gas forcing.

Do not jump from emissions to one local impact without the intermediate physical and social mechanisms.

Climate change alters ecosystems from local sites to global circulation, affecting resilience and sometimes shifting biome boundaries.

Some cooler regions may gain a longer growing season or higher productivity, while heat- or water-limited systems lose productivity; evaluate location, season and biodiversity rather than a global average alone.

Trace climate driver → ecosystem exposure → biological mechanism → impact → resilience or biome response at a stated scale.

A local benefit does not cancel unequal global harm, and one ecosystem example cannot define every region's response.

A hazard becomes an impact through exposure and vulnerability; recovery depends on infrastructure, health, income, governance and social support.

Trace hazard → exposed people or assets → mechanism → sector outcome → capacity to recover. Relevant sectors include health, water supply, agriculture and infrastructure, and the same hazard can produce different outcomes across socioeconomic conditions.

A heatwave may raise illness, water demand and crop stress while damaging transport or power systems; early warning, robust services, income and governance can reduce losses and speed recovery.

A physical event becomes a social impact only when people, livelihoods or infrastructure are exposed; resilience depends on resources, institutions and social support.

Recovery capacity is not a moral trait; it reflects resources, institutions, exposure and historical inequality.

A feedback is a change in output that alters the original input or process; positive amplifies, negative dampens.

Use input → process → output → feedback. Less ice lowers albedo and increases absorbed sunlight; warming can thaw permafrost and release methane, adding forcing. Label sign, not just sequence.

Ice loss → darker ocean → more absorbed energy → further ice loss is positive feedback.

No. It means the initial change is amplified; the outcome can be harmful or helpful depending on context.

A correlation or chain is not a feedback until the output returns to alter the system.

The climate-change planetary boundary is a proposed safe-operating-space limit for Earth-system pressure; published evidence suggests that this boundary has already been passed.

Evaluate which climate indicators and control variables the source uses, the reference period, spatial scale and uncertainty. Transgression means rising systemic risk, not a guaranteed date of collapse.

If an assessed climate indicator lies beyond its proposed safe range, that supports boundary transgression; a separate mechanism and local evidence are still needed to claim that a particular ecosystem has crossed a tipping point.

Report the source, indicator definition, threshold method and uncertainty before accepting or challenging the assertion.

Boundary transgression is not synonymous with certain irreversible collapse.

Perspectives shape which risks, values and solutions people prioritize; they do not replace physical evidence about climate processes.

An ecocentric view may prioritize ecosystem integrity; an anthropocentric view may prioritize livelihoods; a technocentric view may favour engineered solutions. Test each response against the same evidence.

Two groups accept the same drought data but prefer restoration versus desalination because they value different risks and controls.

No. It changes interpretation, priorities or decisions, not the observation itself.

Calling a perspective ‘biased’ does not evaluate its evidence; separate values, facts and proposed action.

Choose a climate record by variable, spatial scale and timescale: direct instruments measure current conditions, while proxies extend the record indirectly.

Weather stations and observatories directly record local variables such as temperature and greenhouse-gas concentrations; radar maps precipitation; satellites observe broad atmospheric and land-use patterns. Ice-core isotopes, tree-ring chronology and peat-core pollen are indirect proxies.

Use radar to locate today's storm cells, an observatory for a long direct CO2 series, and an ice core to reconstruct ancient gas or isotope conditions.

Both direct and indirect records inform climate models; match resolution, calibration and uncertainty to the question.

A longer record is not automatically more direct or more precise.

A climate model maps inputs and assumptions through physical processes to outputs that can be tested against observations.

Inputs include forcing and boundary conditions; processes represent atmosphere, ocean, ice and land; outputs include temperature, rainfall or extremes. Hindcasting tests structure but cannot prove every future detail.

A model that reproduces past warming under observed emissions gains support, but unresolved clouds still affect future projections.

It tests consistency with known data, while future forcing and model uncertainty remain.

Model uncertainty does not make a model useless; it tells you how strongly to qualify the output.

A climate scenario is a conditional ‘if–then’ pathway based on stated emissions, technology, policy and socioeconomic assumptions—not a guaranteed forecast.

Compare scenarios by holding the question and model framework constant, then state which emissions, technology, policy and socioeconomic assumptions differ. Read output ranges for sea-level rise, local temperature and precipitation over the same time horizon.

A lower-emissions pathway may project less warming and sea-level rise, but local precipitation can still vary by region and model; the result is conditional on the scenario assumptions.

Report assumptions, time horizon, location, variable, model range and uncertainty before comparing impacts.

Scenario labels are not predictions of what must happen; they describe consequences of choices and conditions.

A tipping point is a threshold where reinforcing feedback can push a system into a substantially different, persistent state.

Name the pressure, threshold evidence, positive feedback and possible new equilibrium. Examples include Antarctic ice-sheet melting, slowing Atlantic thermohaline circulation, and an Amazon Rainforest–Cerrado transition; timing and reversibility remain uncertain.

For Antarctic ice, warming can increase melt and reduce reflective ice area, reinforcing energy absorption; the threshold concerns persistent state change rather than one dramatic day.

A credible claim identifies nonlinear response, reinforcing mechanism, possible persistence and uncertainty—not only a severe impact.

‘Tipping point’ does not mean sudden, exact or certainly irreversible in every case.

Tipping elements can interact: crossing one threshold may alter forcing or resilience and raise risk for another.

Draw element A → changed climate/ecology → pressure on B, then label whether the link amplifies or dampens. Combined prediction is less certain than isolated analysis because interactions and timing vary.

Forest dieback lowers carbon uptake, increasing warming pressure on an ice system; this is a plausible linked pathway, not proof of synchronized collapse.

Each element and cross-link has uncertainty, so errors compound across the chain.

Linked risk is not a deterministic domino sequence; distinguish mechanism, likelihood and timing.

Responsibility measures contribution to causes; vulnerability measures exposure and capacity to cope. They are related policy questions, not the same variable.

Compare historical or per-capita emissions, current emissions and consumption footprints with exposure, income, infrastructure, health and adaptation options. State metric and timescale.

A country may have low historical emissions but high flood exposure and limited finance, while a wealthy high-emitter has more adaptation capacity.

No. Vulnerability depends on exposure and capacity; responsibility needs its own emissions evidence.

Do not collapse justice into a single ranking; separate causal contribution, present need and ability to respond.