Climate describes the usual conditions produced by physical processes in the atmosphere. These conditions come from long-term observations, not from a single unusual season or storm.

Climate and weather aren't the same. Weather is the state of the atmosphere at a particular place over a short period, while climate covers the characteristic pattern and variability expected over much longer periods.

Temperature and precipitation

An area's climate is mainly described through seasonal variations in temperature and precipitation. Timing matters just as much as annual averages. Two places may receive the same yearly rainfall but have very different climates: one might get rain throughout the year, while the other has a single short wet season.

These seasonal patterns are influenced by latitude, altitude, distance from the sea, prevailing winds, ocean currents and topography. A mountain, for example, may force moist air upwards. This produces precipitation on the windward side and drier conditions in the rain shadow.

A climate graph is a combined graph that shows a location's mean monthly temperature and precipitation. First, use the temperature line to identify the annual range and seasonal peaks. Then check the precipitation bars for the total amount and any wet or dry seasons. Finally, look at how the patterns interact—for example, whether the warmest months are also the driest.

The historical increase

Anthropogenic emissions are releases produced by human activities. Human emissions of $CO_2$ have driven a significant rise in its atmospheric concentration. This increase began during the Industrial Revolution in late eighteenth-century Europe, as coal-powered machinery started to replace human, animal and renewable energy on a large scale.

Industrialization then spread, fossil-fuel use expanded and the human population grew. Emissions rose especially quickly during the twentieth century and after 1950, driven by greater electricity production, transport, manufacturing and cement production. Deforestation released additional emissions through burning and decomposition. It also removed vegetation that could absorb $CO_2$.

Before widespread industrialization, the atmospheric concentration was about 280 parts per million. During the 2020s, it exceeded 420 parts per million—roughly a 50% increase. Concentration and annual emissions aren't the same: emissions are an input flow, whereas atmospheric $CO_2$ is a storage. It accumulates when inputs exceed removal by oceans and terrestrial ecosystems.

The long-term atmospheric record shows an upward trend, along with a small annual oscillation caused mainly by seasonal plant growth and decay. This oscillation is particularly noticeable across the land-rich Northern Hemisphere. For anthropogenic climate change, the long-term rise is what matters; the seasonal saw-tooth pattern doesn't cancel it.

The historical evidence strongly supports anthropogenic attribution. Natural carbon-cycle flows continue, but the timing and scale of the modern increase closely match industrial fossil-fuel use and land-use change.

Records of past climates

Natural archives provide evidence from before instrumental observations began. An ice core is a cylindrical sample removed from an ice sheet or glacier whose layers preserve evidence of earlier atmospheric and climatic conditions. Scientists can analyse trapped air bubbles to find past concentrations of $CO_2$ and other gases. Isotopic composition also helps them reconstruct temperature.

A tree ring is an annual layer of woody growth whose width and density reflect environmental conditions during its formation. In places where temperature or moisture limits growth, wide rings may show relatively favourable years, while narrow rings suggest unfavourable ones. Ring growth also depends on tree age, disease and local competition, so the relationship must be calibrated against modern observations.

Deposited sediments are accumulated layers of mineral or organic material that preserve evidence of past environments. Pollen, shells, organic remains and sediment grain size can reveal the vegetation, temperature, water conditions or aridity present as each layer formed. Once the layers are dated, the evidence becomes a time series.

Correlation over glacial cycles

Records from approximately 800,000 years show repeated glacial–interglacial cycles. Atmospheric $CO_2$ concentration and reconstructed global temperature generally rise and fall together. Warm periods coincide with relatively high $CO_2$, whereas cold glacial periods coincide with lower $CO_2$. This is a positive correlation: higher values of one variable tend to occur with higher values of the other.

To investigate this type of graph, first identify repeated peaks and troughs. Compare their timing and ranges, then compare the recent value with the earlier range. In this record, modern $CO_2$ lies above the concentrations recorded during previous cycles.

Correlation by itself doesn’t establish a one-way cause. Orbital forcing could have initiated past changes, with carbon-cycle feedback then amplifying them. However, laboratory physics, the atmospheric energy balance and modern emissions data provide an independent causal mechanism through which additional $CO_2$ produces warming.

From emissions to climate change

The greenhouse effect is an atmospheric process in which greenhouse gases absorb and re-emit some outgoing long-wave radiation, keeping the lower atmosphere and surface warmer than they would otherwise be. This process is natural and essential.

The enhanced greenhouse effect is the additional warming caused when human activities increase atmospheric greenhouse-gas concentrations, so a larger proportion of outgoing long-wave radiation is retained within the lower atmosphere–surface system. Human activities release large quantities of carbon dioxide, methane and nitrous oxide, as well as smaller quantities of other greenhouse gases.

Major sources of $CO_2$ include fossil-fuel combustion, cement production and deforestation. Methane, $CH_4$, comes from livestock digestion, flooded rice cultivation, landfill, fossil-fuel extraction and wetlands. Fertilized agricultural soils and manure management are strongly associated with nitrous oxide, $N_2O$.

Global warming is the long-term increase in Earth's mean surface temperature caused predominantly in the modern period by the enhanced greenhouse effect. Climate change is the broader range of persistent changes in climatic conditions, including altered precipitation, circulation, extremes and sea level as well as warming.

The causal chain works like this: human activities increase greenhouse-gas emissions, causing atmospheric concentrations to rise. More outgoing long-wave radiation is then absorbed and re-emitted. The global energy balance gains energy, global temperature rises, and connected atmospheric, oceanic and ecological processes change.

This gives a careful answer to the attribution question. Natural influences still affect climate from year to year and across geological time. However, the rapid, sustained modern trend is primarily anthropogenic: observed warming matches human greenhouse forcing and cannot be reproduced adequately by natural drivers alone.

Scale and type of impact

Climate change affects ecosystems at every scale, from individual sites to the global Earth system. A biome shift is a movement in the geographical distribution of a biome as climatic conditions suitable for it move to a different latitude or altitude. Because species move at different rates, whole communities may be reorganized instead of shifting together intact.

Coral bleaching and desertification are local impacts. During severe marine heatwaves, corals may expel their symbiotic algae. They then lose their colour and much of their energy supply. If bleaching is repeated or prolonged, corals may die and reef habitat can decline. Exceptional marine heat in the Florida Keys during 2023 caused widespread bleaching and led to the temporary removal of nursery corals. This shows a local ecosystem responding to a wider climatic trend.

Desertification is land degradation in drylands that reduces biological productivity through climatic variation and human activities. Greater evaporation and less reliable rainfall can worsen damage from vegetation clearance, overgrazing or unsuitable cultivation.

On a global scale, warming and freshwater released by melting land ice can change ocean density and circulation. Sea level rises as warming seawater expands and melting land-based ice adds water to the oceans. Coastal wetlands, mangroves and nesting beaches may disappear if they cannot migrate inland.

Productivity, biodiversity and resilience

Some cool or high-latitude regions may initially become more productive naturally. Warmer conditions can lengthen growing seasons, while higher $CO_2$ can increase photosynthesis. The benefit depends on other conditions, though. Nutrient shortages, heat damage, drought, fire, pests and extreme weather may eventually outweigh it.

Resilience is the capacity of a system to resist disturbance, recover and retain its essential structure and functions. Biodiversity often strengthens ecosystem resilience because species respond differently to disturbance. If one species declines, another may continue to perform a similar function. Low genetic or species diversity and fragmented habitat leave fewer options for recovery, as do rapid change and repeated disturbance. Climate change can therefore reduce biodiversity, which then weakens resilience—a reinforcing effect.

Investigating evidence

Climate graphs from different locations can reveal shifts in seasonal temperature and precipitation, rather than just changes in annual means. Long-term records of atmospheric $CO_2$ and temperature support anthropogenic warming. In ocean records, increasing dissolved $CO_2$ is associated with declining pH and ocean acidification.

For a database investigation, select one ecosystem and one measurable response, such as annual heat stress and coral cover. Keep the location and time period consistent, then plot the variables and look for any lag or anomalous years. Check other possible influences, including storms, disease or local pollution. A relationship provides useful evidence, but these additional variables limit a simple causal claim.

Climate change affects societies from the household and community level to national and international scales. However, exposure varies sharply, as does the ability to respond. Societal resilience is the capacity of a human community to anticipate, withstand, adapt to and recover from disturbance while maintaining essential functions.

Main areas of impact

• Health: heatwaves cause more heat exhaustion, cardiovascular stress and deaths. Shifts in temperature and rainfall can change where disease vectors are found. Smoke, flooding, food insecurity and displacement add further health risks.
• Water supply: drought and increased evaporation deplete surface-water and soil-moisture stores. As glaciers shrink, runoff may rise temporarily before dry-season supplies decline. Floodwater can also contaminate drinking water.
• Agriculture: heat stress, drought, floods and changing pests can lower crop yields or livestock productivity. Cooler regions may benefit from longer growing seasons, though only where fertile soil, water and suitable infrastructure are available.
• Infrastructure: coastal settlements face threats from sea-level rise and storm surges. Heat can damage roads, railways and power systems. Intense rainfall may overwhelm drainage and wastewater systems.

The 2022 floods in Pakistan show how these impacts connect. Extreme monsoon rainfall and glacier melt inundated settlements and farmland, damaged transport and health facilities, contaminated water supplies and displaced millions. Climate change wasn’t the only factor: land use, poverty, drainage and settlement patterns also shaped the outcome. Even so, warming increased the likelihood and intensity of extreme rainfall.

Why resilience differs

Wealth can pay for cooling, flood protection, insurance, healthcare and reconstruction. Adaptive capacity also rises with effective government, trusted warnings, education, social networks and diversified livelihoods. By contrast, poverty, marginalization, weak institutions, conflict, poor housing and reliance on climate-sensitive work reduce it.

Scale makes a difference. Even a wealthy country may have highly vulnerable neighbourhoods. Meanwhile, a lower-income community may draw on strong local knowledge and mutual support. Ahmedabad in India, for example, developed a heat action plan that uses forecasts, public warnings, healthcare preparation and access to cooling measures. This shows that governance and planning—not income alone—shape resilience.

Climate impacts can deepen existing socio-economic inequalities. People with the least secure housing, poorest health or least political influence often suffer the greatest harm and take the longest to recover.

Representing cause and effect

In a systems diagram, boxes represent variables or storages, while arrows show causal relationships. A plus sign on an arrow shows that the variables change in the same direction. A minus sign shows that they change in opposite directions. The signs describe the relationship; they don’t indicate whether an outcome is desirable.

A positive feedback loop is a feedback process that amplifies an initial change and moves a system farther from its previous equilibrium. A negative feedback loop is a feedback process that counteracts an initial change and tends to stabilize a system.

Albedo and methane feedbacks

Albedo is the proportion of incoming solar radiation reflected by a surface. As warming melts snow and ice, darker land or water is exposed. Albedo falls, so more solar radiation is absorbed. Further warming follows, making this positive feedback.

Warming may also thaw permafrost and speed up anaerobic decomposition in waterlogged soils. This releases methane, strengthening greenhouse forcing. The additional warming causes more thaw. The timing and size of the feedback remain uncertain, but its direction is reinforcing.

Solar radiation and the global energy balance

The global energy balance is the relationship between incoming solar radiation and energy leaving Earth as reflected short-wave and emitted long-wave radiation. When this balance is disrupted for a sustained period, the amount of energy stored in the climate system changes.

More solar radiation can warm Earth. In response, the warmer surface emits more long-wave radiation to space, which opposes the initial warming. This is a negative radiative feedback. Solar variation may initiate change, with albedo and greenhouse-gas responses modifying it. However, recent solar variation alone cannot explain current warming.

Every intermediate link should appear when drawing a loop. In the ice–albedo loop, drawing a direct link from melting ice to warming leaves out the assessed mechanism: less ice lowers albedo, causing greater absorption of solar energy.

A planetary boundary is a scientifically proposed limit for an Earth-system process beyond which the risk of large-scale, abrupt or irreversible environmental change rises substantially. It isn’t a cliff edge where catastrophe suddenly becomes certain. Instead, it marks the point at which conditions move beyond a comparatively safe operating space.

Published evidence

The planetary-boundaries framework assesses climate change through indicators such as atmospheric $CO_2$ concentration and changes in Earth's energy balance. The proposed boundary is 350 parts per million $CO_2$, yet modern concentrations exceed 420 parts per million. For this reason, the framework’s 2023 update classified the climate boundary as transgressed.

Several independent observations support this finding. Global warming has continued, while glaciers and ice sheets have experienced widespread loss. Ocean heat content is increasing, mean sea level is rising, and heat extremes are becoming more frequent or intense. Climate-system change, then, isn’t identified from a single indicator.

Some uncertainty remains. Boundary values are estimates, not perfectly known thresholds, and impacts vary between regions. Passing a boundary also doesn’t prove that every tipping point has already been crossed. But uncertainty works both ways: some responses could happen earlier or faster than central estimates suggest.

Published evidence strongly supports the assertion that humanity has moved beyond the proposed safe boundary for climate change. Further action is still worthwhile. Every avoided increment of warming lowers additional risk and makes adaptation more feasible.

A perspective is a viewpoint formed from a person's or group's values, knowledge, experiences and circumstances. It affects which evidence people trust, which risks they put first and which responses they see as acceptable.

Factors shaping perspectives

Direct experience of heat, flooding, crop loss or coastal erosion can make climate risk feel immediate. Other influences include wealth, education, occupation, age, culture, political identity, place of residence and access to reliable media. For example, a farmer who depends on seasonal rainfall may view risk differently from an urban employee whose work is less directly affected by weather.

Values also shape how people interpret climate change. Someone with an ecocentric perspective may prioritize ecosystem integrity and rapid emission reduction. An anthropocentric perspective may favour policies that protect human welfare while balancing economic costs. By contrast, a technocentric perspective may place greater confidence in innovation, engineering and market mechanisms.

From perspective to societal response

Governments, businesses, scientists, Indigenous communities, campaigners and consumers may agree that climate change is occurring, yet disagree about timing, responsibility, acceptable cost and the balance between mitigation and adaptation. Communities that depend on fossil fuels may fear unemployment caused by a rapid transition. Low-lying communities, however, may see delayed action as an immediate threat.

Behaviour is also affected by personal perspectives. Concern doesn’t automatically lead to action. Cost, convenience, infrastructure and the belief that individual choices are insignificant can create a gap between values and behaviour. On the other hand, personal action can influence social norms and political demand, even when its direct effect on emissions is small.

Different perspectives therefore shape responses to climate change. Conflict and misinformation can delay collective action. Inclusive decision-making, though, can produce fairer, more durable policies by recognizing different knowledge, values and vulnerabilities. A strong judgement separates disagreement over values and policy from the rejection of well-supported physical evidence.

Direct observations

A direct measurement is an observation of the climatic variable being studied using an instrument at the time it occurs. Weather stations and observatories measure temperature, precipitation, humidity, air pressure and wind. Observatories with long-running records also track greenhouse-gas concentrations.

Radar detects precipitation and follows its movement using reflected radio waves. Satellites repeatedly observe large areas, recording cloud, surface temperature, vegetation, ice cover, fires and land use. This coverage is especially useful where ground stations are sparse. However, satellite records are shorter and must be calibrated against surface observations.

Researchers must check long-term records for changes in instruments, observation times and the surroundings of each station. For instance, urban growth around a station may affect its local temperature record. Rather than accepting every reading unchanged, scientists homogenize the data and compare records from neighbouring stations.

Proxy evidence

A proxy measurement is an indirect indicator from which a past climatic variable is inferred after calibration. These proxies extend climate records beyond the instrumental period.

Ice cores preserve atmospheric gases along with stable oxygen-isotope ratios. Ratios involving $^{18}\text{O}$ and $^{16}\text{O}$ change with temperature-dependent evaporation and condensation, so they can be used to infer past temperatures. Trapped air bubbles provide samples of earlier atmospheric greenhouse-gas concentrations.

Dendrochronology is the dating and analysis of annual tree rings to reconstruct environmental change. Where temperature or moisture limits growth, ring width or density may indicate that variable. Pollen preserved in peat cores reveals earlier plant communities. Because plant species have characteristic climatic tolerances, shifts in pollen assemblages indicate climatic and land-use change.

Direct observations tend to offer greater precision and temporal resolution, but they cover a relatively short period. Proxies reach much further into the past, though they involve uncertainties in dating, calibration and interpretation. Climate models rely on both. Direct data describe and test recent behaviour, while proxies show whether models reproduce longer-term climatic variability.

A global climate model is a mathematical representation of Earth's climate system that uses equations to simulate physical, chemical and biological processes and their interactions across a three-dimensional grid. Although models simplify reality, they do far more than plot trend lines.

Inputs, processes and outputs

Models take in data on solar radiation, greenhouse-gas concentrations or emissions, aerosols, volcanic particles and land-use conditions. They then represent atmospheric and ocean circulation, radiation transfer, cloud formation, ice melt, and exchanges of carbon, water and energy. Their outputs may include temperature, precipitation, sea ice, ocean circulation and sea level.

A model cannot directly calculate every process that occurs at a scale smaller than one grid cell, such as the formation of individual clouds. Instead, it uses parameterizations based on observations and theory. The range of outcomes partly reflects different parameter choices and assumptions about future human activity.

Validation through hindcasting

Hindcasting is a model-testing process in which a model is run from an earlier time towards the present and its simulated outputs are compared with observed conditions. Credibility increases when a model reproduces major past trends and spatial patterns, including responses to volcanic eruptions. When it fails to do so, researchers may need to revise its equations, parameters or input data.

Hindcasting tests a model; it doesn’t prove that every future prediction will be correct. For example, a model may reproduce the global mean but still perform poorly when simulating local rainfall. Future human emissions also have no exact historical equivalent.

Uncertainty

Some inputs are uncertain, especially proxy reconstructions and future emissions. Uncertainty also comes from model structure and parameterization. Natural variability can affect when regional changes occur. Scientists therefore compare several models and scenarios. Agreement between them raises confidence, while the spread of results shows a range of possible outcomes rather than a failure to understand climate.

A climate scenario is a coherent set of assumptions about future drivers used to explore a possible climate outcome. It does not claim that a particular future is certain. Different scenarios may use different assumptions about greenhouse-gas emissions, land use, population, energy systems and climate policy.

When models run under lower-emission scenarios, they generally project less warming and sea-level rise than under high-emission scenarios. These differences grow later in the century as greenhouse gases and heat accumulate over time.

Interpreting scenario data

Most scenario graphs show a central line surrounded by an uncertainty range. Compare the direction, magnitude and timing of change. Look for the point when the ranges separate, but don’t treat the upper and lower bounds as impossible limits.

Sea level responds slowly because oceans retain heat, while ice sheets take time to adjust. As a result, it may continue rising after atmospheric emissions are reduced. Local sea-level change can also differ from the global mean due to ocean circulation, gravitational effects of ice loss, vertical land movement and coastal subsidence.

Temperature patterns aren’t the same everywhere. Land usually warms faster than ocean, and high northern latitudes warm particularly rapidly. Precipitation projections show less uniformity. Warmer air can hold more water vapour, which intensifies heavy rainfall in some places. Elsewhere, altered circulation and greater evaporation increase drying.

Local projections are more uncertain than the global temperature trend because regional precipitation and circulation are difficult to resolve. Comparing scenarios is still useful: it reveals which impacts depend strongly on future human choices and which changes are already difficult to avoid.

A critical threshold is a boundary beyond which a small additional pressure can produce a disproportionately large change in a system's state. A tipping point is a critical threshold at which self-reinforcing processes can commit a system to a substantially different state. After that, a new equilibrium may form around different conditions.

Thresholds occur both in the global climate system and in local ecosystems. Global thresholds are particularly concerning: change may be rapid and difficult to anticipate. It may also be irreversible on human timescales, with catastrophic consequences for exposed societies or ecosystems.

Example: Amazon rainforest–Cerrado transition

The Amazon recycles moisture through evapotranspiration. Trees release water vapour that supports regional rainfall; this rainfall then maintains forest growth. Warming, drought and deforestation reduce tree cover and evapotranspiration. As rainfall declines, the dry season lengthens. Fire and tree mortality increase, causing still more forest loss.

Through this positive feedback, parts of the system may cross a threshold from humid rainforest towards a more open, seasonally dry vegetation state resembling the Cerrado. Such a change would reduce biodiversity and carbon storage. It could also release additional $CO_2$, reinforcing global warming. The exact location of the threshold remains uncertain because it depends on warming, deforestation, fire, rainfall and ecological adaptation.

Other possible thresholds include destabilization of Antarctic ice sheets and slowing of Atlantic thermohaline circulation. Ice-sheet loss can become self-sustaining when retreat exposes thicker ice to warm ocean water or lowers the ice surface into warmer air. Atlantic circulation may weaken when warming and freshwater input reduce the density of North Atlantic surface water, disrupting sinking and heat transport.

Climate models identify risks and ranges, not exact countdown points. Uncertainty about the threshold doesn’t show that the risk is absent. Instead, it gives a reason to avoid pressures that could push the system into a new equilibrium.

A tipping cascade is a sequence of interacting threshold changes in which crossing one tipping point increases the likelihood of crossing one or more others. What makes it a cascade is the interaction, not simply the presence of several separate tipping points.

Individual tipping elements may be biotic, involving living organisms or ecological processes, or abiotic, involving non-living physical components. Some combine both. Amazon forest dieback is strongly biotic, though it also depends on abiotic rainfall and temperature. Ice-sheet loss and changes in ocean circulation are predominantly abiotic.

How tipping elements can interact

As Greenland loses ice, freshwater enters the North Atlantic. This may weaken Atlantic overturning circulation and change heat and rainfall patterns. Shifts in tropical rainfall could place the Amazon under greater drought stress, encouraging fire and forest loss. With less carbon stored in the forest, atmospheric $CO_2$ rises. The resulting warming then accelerates Greenland and Antarctic ice loss.

The proposed links don’t all have the same strength or certainty. Some interactions reinforce each other, while others oppose one another or take effect only after long time lags. With two or more tipping points involved, the scale and pace of future climate change become especially uncertain. If abrupt interactions occur, standard projections based on gradual responses may underestimate the risk.

Uncertainty does not mean that every outcome is equally probable. Models and observations can identify the more credible links, as can palaeoclimate evidence. Even so, exact thresholds and timings remain difficult to determine. The possibility of cascades supports a precautionary approach because preventing the first threshold crossing may avoid several subsequent changes.