Warming is not one problem. It is a set of coupled ones. Heat is a direct physical hazard. Sea level rise is a slow, near-irreversible commitment. Ocean acidification is a chemistry problem that would occur even if the planet were not warming. Treating them as a single emergency makes them harder to think about, not easier.
This page organises climate change issues by the system affected: the mechanism, what has been observed, and where the honest uncertainty sits. The figures are current to 2025: global mean temperature reached approximately 1.43°C above the pre-industrial baseline, among the three warmest years on record despite the absence of an El Niño, and atmospheric CO2 stood at 425.6 ppm, a 53% increase over the pre-industrial level of about 278 ppm. Energy production accounts for roughly two-thirds of global greenhouse gas emissions, the common origin of almost everything below.
Heat
Heat is the most direct impact and, in many countries, the deadliest weather hazard — and where the headline temperature is least useful.
Humans shed heat by evaporating sweat. When the air is close to saturated, evaporation slows and the body cannot cool itself, whatever the shade or rest. The wet-bulb temperature captures this: the lowest temperature a surface can reach by evaporation alone. A survivability limit of 35°C wet-bulb is widely cited, but laboratory work at Pennsylvania State University puts the practical limit for healthy young adults nearer 30 to 31°C, and lower for older people and those with cardiovascular conditions.
A 38°C day in a humid coastal city can therefore be more dangerous than a 45°C day in a desert, and heat mortality concentrates in the elderly, outdoor workers, and people whose nights never cool enough for recovery. Cities amplify it: dark surfaces, dense construction and little vegetation produce urban heat islands several degrees warmer than surrounding countryside, most markedly at night.
Water
A warmer atmosphere holds more water vapour. The relationship is the Clausius-Clapeyron equation, and it is close to fixed physics rather than a model output: saturation vapour pressure rises by roughly 7% per degree Celsius of warming.
More moisture means heavier rainfall when conditions do produce rain, so extreme precipitation intensity rises broadly in line with, and in some convective storms faster than, that figure. Meanwhile a thirstier atmosphere pulls moisture from soils and vegetation faster during dry periods, deepening droughts. The water cycle does not shift wetter or drier; it becomes more variable at both ends.
Snowpack is seasonal storage, releasing winter precipitation gradually through spring and summer. When more of it falls as rain and melt begins earlier, the same annual total arrives at the wrong time of year, and systems built around a reliable late-season release face shortfalls even in normal years.
The oceans
The oceans have absorbed almost 90% of the excess energy trapped in the Earth system over the past half century. That absorption both moderates surface warming and makes it slow to reverse. 2025 was the warmest year on record for global ocean heat content.
Marine heatwaves — extended periods of anomalously high sea surface temperature — are more frequent, longer and more intense. Their most visible consequence is coral bleaching: corals expel their symbiotic algae under sustained thermal stress, losing colour and their main energy source, and recover only if the stress ends. The event running from 2023 into 2025 exposed roughly 84% of the world’s reef area to bleaching-level heat stress, on NOAA Coral Reef Watch and International Coral Reef Initiative figures, against 68% during the 2014 to 2017 event.
Ocean acidification is a separate problem with a separate mechanism; conflating the two is a common error. CO2 dissolving in seawater forms carbonic acid, releasing hydrogen ions and reducing the carbonate ions that shell-forming organisms need — a function of CO2 concentration, not temperature, and one that would continue even without warming. Surface ocean pH has fallen about 0.1 units since pre-industrial times, roughly a 30% increase in hydrogen ion concentration on the logarithmic scale.
Ice and sea level
Only some ice raises sea level: sea ice floats and displaces its own mass, so its loss does not directly change sea level, though it does reduce reflectivity and accelerate regional warming. Land ice — glaciers and the Greenland and Antarctic ice sheets — adds water that was not previously in the ocean.
Global mean sea level reached a record high for the fourteenth consecutive year, about 111.2 mm above the 1993 average, rising at roughly 3.5 mm per year and accelerating. Since 2005, thermal expansion has added 1.6 ± 0.3 mm per year and ice sheets and glaciers 2.0 ± 0.4 mm per year: land ice is now the dominant term.
In the Arctic, 2025 brought the lowest maximum sea ice extent in the 47-year satellite record; the September minimum was the eleventh lowest. Age is the more telling figure: ice older than four years covered 95,000 km² in September 2025, down from about 1.5 million km² in the 1980s. The Arctic is warming at roughly three times the global rate, and 2025 was its second-warmest year in the 126-year record.
Committed rise is locked in by changes that have already occurred and continues for centuries regardless of future emissions; projected rise adds what future emissions produce. The IPCC’s likely ranges to 2100 span roughly 0.3 to 1.0 metres by scenario, and outcomes approaching 2 metres cannot be ruled out under low-confidence ice sheet processes. That width is not vagueness: ice sheet instability and ice cliff failure are unresolved, and they dominate the upper end.
Ecosystems
Species are moving. Terrestrial ranges are shifting poleward and upslope, marine species poleward and deeper, in thousands of documented populations.
The subtler problem is phenological mismatch. Plants and insects respond largely to accumulated temperature, so they advance as springs warm. Many birds cue on day length, which does not change at all. A migratory bird can therefore arrive on schedule to find the caterpillar peak it feeds chicks on already past, the synchronisation broken by a signal only one of them reads.
For most ecosystems the rate of change matters more than the amount. Species have survived comparable absolute changes in the deep past, over millennia; compressing that into a century removes the time range shifts and adaptation require.
Food and agriculture
Staple crops have narrow thermal windows, and heat during flowering is particularly damaging: short exposures above critical thresholds can cut yields substantially even when seasonal averages look acceptable.
Elevated CO2 does have a real fertilising effect, but its limits are well documented. Free-air enrichment experiments show gains smaller in the field than in enclosed chambers, concentrated in C3 crops such as wheat and rice rather than C4 crops such as maize, and materialising only where water and nitrogen are not the constraint. There is also nutrient dilution: crops grown under elevated CO2 show lower concentrations of protein, zinc and iron, so more tonnage does not translate cleanly into more nutrition.
The systemic concern is concentration. A small number of regions supply most internationally traded grain, and simultaneous poor harvests in several of them, driven by correlated atmospheric patterns rather than independently, would move prices far more than any single regional failure. The physical plausibility of that is established; the probability attached to it is not.
Health
Direct heat mortality is the clearest pathway: the Lancet Countdown’s 2025 report estimates the rate of heat-related mortality has risen 23% since the 1990s, with such deaths averaging around 546,000 per year.
Vector-borne disease is a question of range. Mosquitoes carrying dengue, chikungunya and malaria are constrained by temperature at their cold margins, and those margins are moving into higher latitudes and elevations. Whether transmission establishes depends on housing, water management and public health capacity.
Air quality interacts with warming in both directions: wildfire smoke carries fine particulate matter across continental distances, and ground-level ozone forms faster on hot, stagnant days. Mental health effects, including distress following disasters and displacement, are real and increasingly measured, though harder to quantify.
People and displacement
This section needs more care than any other, because it is where the weakest numbers circulate.
Climate acts as a risk multiplier, interacting with governance, economics, land tenure and existing conflict rather than causing outcomes on its own. Drought does not by itself cause displacement; drought in a place with no crop insurance and few alternative livelihoods can contribute to it. Attributing any specific movement of people to climate alone is rarely defensible.
Most climate-related movement observed so far has been internal, short-distance and often temporary. Figures projecting hundreds of millions of international climate migrants by mid-century are widely repeated but poorly sourced, tracing to early estimates that assumed everyone in an exposed area would move. The better-supported statement is narrower: climate pressure raises displacement risk where vulnerability already exists.
Tipping elements
Some parts of the system may have thresholds beyond which change continues on its own, independent of further forcing. This deserves accurate treatment rather than dramatic treatment.
The main candidates are the Greenland and West Antarctic ice sheets, the Atlantic Meridional Overturning Circulation (AMOC), permafrost carbon release, Amazon dieback and warm-water coral reefs. They differ enormously in evidence and timescale.
The ice sheets have the clearest physical basis. Marine ice sheet instability is well understood, and once initiated the sea level contribution would unfold over centuries to millennia, though the threshold temperature remains uncertain across a wide range of estimates.
AMOC weakens in models under warming and shows some observational indications of change, but the direct record is short. The IPCC’s AR6 assessment is that it will very likely decline this century, with medium confidence that the decline will not involve an abrupt collapse before 2100. Studies proposing collapse within decades rest on statistical extrapolation from proxy indicators and are not the consensus position.
Permafrost holds more carbon than the atmosphere currently does. Thaw is observed and its emissions are real, but the evidence points to gradual release rather than a sudden pulse. Amazon dieback is the least constrained and depends as much on deforestation as on climate. Coral reefs are the exception: their threshold is not speculative, and it has largely been crossed for many reef systems already.
The honest summary is that thresholds probably exist, their temperatures are poorly constrained, most consequences unfold over centuries rather than years, and the risk warrants caution rather than alarm. Presenting low-confidence outcomes as imminent damages the credibility of the well-evidenced ones.
Where to go next
These systems are coupled, but they are not the same problem and do not respond to the same interventions. For the observational basis behind the figures here, start with the evidence; for how the public argument has been shaped, the climate communication pages cover that ground; and for what can actually be done, the action page is the destination.