Fifth National Climate Assessment
3. Earth Systems Processes

Shifts in climate at the global scale occur primarily in response to processes that change the balance between incoming solar energy and the outgoing energy radiated from Earth at the top of the atmosphere (TOA). The drivers of change are both natural and human-caused and may be transient or long-lived. Changes in TOA net radiation balance resulting from a perturbation in climate drivers can be quantified in terms of the effective radiative forcing (ERF), measured in units of watts per square meter (W/m²). The main natural drivers of climate change—variations in solar radiation and volcanic aerosols—have negligible contributions to long-term climate forcing.^1,2 Warming from fossil fuel emissions of carbon dioxide (CO₂) is expected to last centuries to millennia, because of the slow rate of the natural processes that remove CO₂ from the atmosphere. Changes in climate may be further amplified or diminished through feedbacks in the climate system. Feedbacks are processes changed by the warming that then modify the TOA radiation balance and the overall level of warming. In the surface–albedo feedback, for example, warming melts ice cover over the land and ocean, exposing darker surfaces beneath that absorb more energy rather than reflecting it, contributing to further warming. See Chapter 2 of the Climate Science Special Report³ for further details on emissions sources, radiative forcing, and ERF.

Anthropogenic Drivers

Well-Mixed Greenhouse Gases

Since the release of the Fourth National Climate Assessment (NCA4) in 2018, global atmospheric abundances of the main well-mixed greenhouse gases (WMGHGs), including CO₂, methane (CH₄), and nitrous oxide (N₂O), have continued to increase (KM 2.1). Atmospheric abundances of halogenated gases have also changed, some increasing and some decreasing due to ozone-depletion policies. Preindustrial to present-day (1750–2019) increases in WMGHG concentrations contributed the bulk of the total human-caused forcing, with increases in CO₂ contributing an ERF of 2.16 ± 0.26 W/m², followed by 0.54 ± 0.11 W/m² from CH₄, 0.41 ± 0.08 W/m² from halogenated compounds, and 0.21 ± 0.03 W/m² from N₂O.² The ERF due to changes in CH₄ abundance is lower than that due to changes in emissions because of offsetting effects of other chemical constituents as discussed below.

The growth in global atmospheric CO₂ levels since 1750 was primarily driven by direct emissions from human activities such as fossil fuel combustion, cement manufacturing, and land-use change. About 41% ± 11% of the 700 ± 75 PgC of CO₂ (1PgC = 1 billion metric tons of carbon [GtC]) emissions between 1750 and 2019 remain in the atmosphere today, with the rest absorbed by oceans and the land biosphere.⁴

Methane is considered both a WMGHG and a short-lived climate forcer due to its chemical reactivity and an approximate 10-year atmospheric lifetime. Methane is produced by both natural processes and human activities. Observational evidence points to microbial sources (agriculture, waste, and natural wetlands) as the predominant cause of the observed increase in global growth in atmospheric CH₄ since 2006,^5,6 with a smaller contribution from fossil fuel production.

In addition to its direct effect on radiative forcing, CH₄ also has an indirect influence through its chemical effects on other climate drivers, including CO₂, ozone, stratospheric water vapor, aerosols, and halogenated gases.⁷ This leads to an increased effect of CH₄ on the amount of energy trapped in the Earth system. Over 20 and 100 years, a given mass of CH₄ emissions is about 80 and 30 times, respectively, more efficient at trapping energy in the climate system than the same emitted mass of CO₂.² This comparison of energy trapped due to a given mass of emitted gas compared to the same emitted mass of CO₂ is known as the global warming potential (GWP), and it is always specified for a given time horizon due to the varying chemical lifetimes of non-CO₂ greenhouse gases. For CH₄, the 100-year GWP is about 30 (relative to a GWP for CO₂ being, by definition, equal to 1 over the same 100-year period).

The increase in atmospheric N₂O levels since 1750 is small compared to that of CO₂ and CH_4, although rates have increased in recent years due to increased nitrogen fertilizer use in agriculture.⁸ Nitrous oxide has an atmospheric lifetime of about 116 years and is nearly 300 times more efficient at trapping energy than CO₂ over a 100-year period.⁹

Many halogenated compounds, which are primarily manufactured gases, also contribute to climate change. These include chlorofluorocarbons (CFCs), perfluorocarbons (PFCs), and sulfur hexafluoride (SF₆), with lifetimes of decades to millennia, and hydrogen-containing halogenated compounds like hydrochlorofluorocarbons (HCFCs) and hydrofluorocarbons (HFCs), with lifetimes of months to decades. The atmospheric abundances of most CFCs have continued to decline in response to regulations under the Montreal Protocol on Substances that Deplete the Ozone Layer and its amendments.¹⁰ Atmospheric levels of HFCs are increasing, while the rates of atmospheric growth of major HCFCs have slowed down in recent years.

Non-methane Short-Lived Climate Forcers

Short-lived climate forcers (SLCFs) are chemically reactive in the troposphere (lower atmosphere), with atmospheric lifetimes typically shorter than two decades, and include ozone, aerosols, and methane. Most SLCFs are also air pollutants or precursors for air pollution (see KM 14.5). An assessment of the ERF effects for two primary non-methane SLCFs are given below.

Ozone: Ozone is a greenhouse gas that occurs naturally throughout the atmosphere and is a harmful air pollutant near the surface (KM 14.1). It is formed in the atmosphere through chemical reactions involving sunlight: in the stratosphere, ozone production occurs via chemical reactions involving the breakdown of oxygen molecules by sunlight, while in the troposphere, it is produced by chemical reactions involving emissions of methane, nitrogen oxides (nitric oxide and nitrogen dioxide), carbon monoxide, and non-methane volatile organic compounds in the presence of sunlight. Increases in human-caused emissions of these precursors since preindustrial times are responsible for increases in tropospheric ozone.^7,11 In the stratosphere, increases in human-caused halogenated ozone-depleting substances (ODSs) have contributed to declining ozone abundances. Over the period 2000–2017, stratospheric ozone concentrations have increased slightly in response to declining ODSs because of the Montreal Protocol and its amendments.¹⁰ The combined changes in tropospheric and stratospheric ozone from the preindustrial era to the present have had an overall warming effect of 0.47 ± 0.23 W/m², with a smaller contribution from stratospheric ozone changes.²

Aerosols: Aerosols are small particles that are emitted directly from human activities and natural processes and are also formed in the atmosphere via reactions involving gaseous precursor emissions. Ice core records indicate that aerosol concentrations increased from the preindustrial era until the 1970s and 1980s, driven by northern midlatitude emissions, and declined thereafter.¹⁰ This decline is attributed to reductions in emissions from Europe and North America due to air quality regulations. Satellite data and ground-based records over the modern period confirm the decline in aerosol concentrations over the northern midlatitude and Southern Hemisphere continents but show increases over South Asia and East Africa.^10,12 Globally, aerosol concentrations have been declining since 2000, driven by reductions in some regions. Aerosols from human activities are also air pollutants (KMs 14.1, 2.1) and influence Earth’s radiation balance directly by scattering or absorbing solar radiation, through interactions between aerosols and clouds, and by changing the surface reflectivity when light-absorbing aerosols are deposited on snow and ice. Changes in aerosols over the period 1750–2014 have had an overall cooling effect of −1.3 ± 0.7 W/m².² Since NCA4, the uncertainty in total aerosol ERF has been reduced, and it is now virtually certain that the total aerosol ERF is negative, as discussed in the Technical Summary of the Working Group I contribution to the Intergovernmental Panel on Climate Change Sixth Assessment Report.¹³

Land-Use Effects on Surface Albedo

Anthropogenic water storage, agricultural practices, and forest cover changes modify the land surface and alter the surface energy balance. Increased water storage on land reduces surface reflectivity and has a localized cooling effect due to evaporation. Irrigation of crops has a similar localized cooling effect through both increased evaporation and plant transpiration.¹⁴ Forests can induce warming because they absorb surface energy or induce cooling because of transpiration.¹⁵ Deforestation can cause cooling by brightening the surface and increasing ground evaporation, but it may also cause warming by reducing ground shading and plant transpiration.¹⁶ Overall, global changes in land use have contributed a net negative ERF (cooling effect) of –0.20 ± 0.10 W/m².²

Natural Drivers

Solar Irradiance

Climate forcing from changes in solar irradiance are small relative to changes in anthropogenic greenhouse gases over the industrial era. Changes in solar irradiance over the period 1750–2019 have contributed an ERF of 0.01± 0.07 W m^–2.²

Volcanic Aerosols

Sulfate aerosols formed in the stratosphere from injections of sulfur dioxide (SO₂) from explosive, episodic volcanic eruptions and more frequent smaller eruptions reduce incoming solar radiation and lead to a cooling effect. Volcanic aerosols can also trigger changes in upper-tropospheric clouds, leading to warming,^17,18 although the magnitude of this effect is highly uncertain.¹⁹ Quantitative assessment of the net ERF over the industrial era for volcanic aerosols shows negligible effects on the long-term temperature trends.

Global Surface Temperature Response to Climate Drivers

A notable recent advance is the quantification of the contributions to global surface air temperature change over the preindustrial to present-day period due to forcings from changes in different climate drivers.² Panels a and b in Figure 3.1 show that observed warming in 2010–2019 (compared to 1850–1900) is dominated by contributions from WMGHGs and ozone. This warming has been partially counteracted by the cooling effect from aerosols and land-use change. Because of declining aerosols, the global aerosol cooling effect has weakened since about 1980, and the rate of global warming has increased, mainly due to increasing greenhouse gases (Figure 3.1c). The aerosol cooling effect has a strong regional dependence,⁷ contributing to differences in regional climate change (KM 3.4). Natural drivers over the historical period have contributed a small, highly variable cooling effect. Uncertainties in the ERF values, especially of aerosols, contribute to uncertainty in the attribution of observed warming.

URL

Alternative text

The warming observed over the industrial era was driven by emissions from human activities, with greenhouse gas warming partly masked by aerosol cooling.

Figure 3.1. The figure shows (a) observed change in global average surface temperature in 2010–2019 relative to 1850–1900; (b) temperature change over the same period (also relative to 1850–1900) attributed to total human influence, including changes in well-mixed greenhouse gas concentrations (including carbon dioxide [CO₂], methane [CH₄], nitrous oxide [N₂O], and halogenated gases); combined changes in aerosols, ozone (O₃), and land use (land-use reflectance); and solar and volcanic drivers and natural climate variability; and (c) time evolution of observed temperature (2010–2019, relative to 1850–1900; black line) attributed to different climate drivers from 1850 to 2019, as well as total human influence (“Total anthropogenic”; purple line) and combined natural and human influence (“Total”; lavender line). Whiskers in (a) and (b) show the very likely range, while shaded uncertainty bands in (c) show very likely (5%–95%) ranges. Note that in (b), the warming effect of ozone is largely offset by the cooling effect of aerosols, resulting in a net cooling when the effects of aerosols, ozone, and land-use change are combined. (a, b) Adapted with permission from Figure SPM.2 in IPCC 2021;²⁰ (c) adapted with permission from Figure 7.8 in Forster et al. 2021² and Figure 2.11c in Gulev et al. 2021.¹⁰

Close

The total climate response for a given forcing is an important indicator of the expected climate change impacts. This response is commonly characterized by the equilibrium climate sensitivity (ECS), defined as the change in global average surface temperature after the climate system reaches a steady state following a doubling of the atmospheric CO₂ concentration. A larger value of ECS indicates larger increases in global warming for a given increase in greenhouse gases. ECS depends both on the ERF from a doubling of CO₂ and the sum of climate feedbacks that can either amplify (a positive feedback) or dampen (a negative feedback) the temperature change. The primary feedbacks arise from increased emissions of longwave radiation that cools the warmer planet (Planck feedback); increases in atmospheric water vapor (a greenhouse gas); changes in the vertical profile of atmospheric temperature; reductions in the surface area covered by reflective snow and ice; and changes in cloudiness. Cloud feedbacks are the largest source of ECS uncertainty (Figure 3.2).^2,21 The response of the carbon cycle to climate warming contributes additional uncertainty (KM 3.4).

URL

Alternative text

Multiple feedback processes in the climate system amplify or partially diminish the response to radiative forcing; quantifying their values is necessary to determine the climate response to human activities.

Figure 3.2. The graph shows assessed values of feedbacks acting in the climate system. The total feedback is the feedback that contributes to the assessed value of climate sensitivity. Bars denote the average feedback values, and uncertainties denoted by whiskers represent very likely ranges. Negative feedbacks are denoted by blue bars and positive feedbacks by orange bars. The feedback values are estimated by climate models based on the equilibrium change in Earth’s energy balance in response to a doubling of carbon dioxide (CO₂) with and without the feedback processes. W/m² is watts per square meter. Adapted with permission from Figure TS.17 in Arias et al. 2021.¹³

The key development since NCA4 is that the ECS uncertainty range has narrowed.^2,21 Recent assessments place ECS between 4.5°F and 7.2°F (2.5°C and 4.0°C), a 50% uncertainty reduction from previous assessments (Figure 3.3). It is virtually certain that ECS is greater than 2.7°F (1.5°C), guaranteeing significant climate change impacts from ongoing increases in CO₂ and other greenhouse gases.

URL

Alternative text

Uncertainties in climate sensitivity and feedbacks have been reduced by synthesizing multiple lines of evidence.

Figure 3.3. The graph (left) shows estimates of equilibrium climate sensitivity (ECS; y-axis) and the considered lines of evidence as a function of the years (x-axis) in which various assessment reports were released. These assessments include the “Charney Report”—a study by an ad hoc US research group that was published in 1979—and the first through sixth assessment reports produced by the Intergovernmental Panel on Climate Change (IPCC), indicated here by the acronyms FAR, SAR, TAR, AR4, AR5, and AR6, respectively. NCA adopts the climate sensitivity values assessed by IPCC reports. Thick vertical bars indicate the likely range, and the horizontal tick marks on these bars, where present, indicate the best estimate (tick marks for the AR5 indicate a less than 5% probability of ECS being below 1.8°F and less than 10% probability of being greater than 10.8°F). The reduced estimated range of ECS in AR6 reflects advances in understanding of how climate feedbacks operate across timescales and an improved ability to combine physical constraints with observational data from recent decades, the industrial era, and paleoclimate records. The images of a satellite, the surface temperature time series, and a woolly mammoth (right) are emblematic of these three lines of evidence used to assess climate sensitivity. Adapted with permission from Figure TS.16a in Arias et al. 2021.¹³

This uncertainty reduction results from advances in combining observations and targeted modeling results from nearly independent lines of evidence, each of which generally agrees on the values of ECS. The lines of evidence include 1) observed present-day variations on monthly to interannual timescales from which cloud and other climate feedbacks are inferred; 2) observed temperature changes between the preindustrial period and present and associated relationships to ERF; and 3) estimated temperature and radiative forcing changes for multiple warm and cold periods in paleoclimate records.²¹

Advances since NCA4 include the following:

• Increased use of present-day observations to determine cloud and other climate feedbacks. Confidence in inferring feedbacks for centennial-scale climate changes from present-day variations on monthly to interannual timescales is bolstered by “emergent constraints” found in Coupled Model Intercomparison Project (CMIP) model ensembles (see App. 3).

• Increased use of very-high-resolution cloud-resolving models to determine expected changes in response to the warmer conditions

• Greater understanding of how changes in the spatial distribution of sea surface temperature over the historical period affected climate feedbacks

• Greater availability of reconstructed temperatures and ERF values for paleoclimate and greater confidence in how to account for the state-dependence of climate feedbacks and departures from equilibrium

The reduction of uncertainty in ECS was accompanied by reduced uncertainty in the cloud feedback. There is greater confidence that the sum of feedbacks over all cloud types is positive (i.e., a warming effect) and primarily results from increases in the altitude of high-level clouds and decreases in the amount of marine stratocumulus and continental low-level clouds. A previously identified negative feedback arising from the transition of cloud phase from ice to liquid as the planet warms²² is now thought to be substantially smaller based on new observational evidence.^23,24

Importantly, the new assessed range of ECS relies on observational analyses and selected modeling evidence but does not consider the ECS values from climate models directly. Models with ECS values outside of the assessed range are thought to have unlikely values of ECS. In particular, the CMIP6 models with ECS values greater than 9°F (5°C) may simulate an unrealistically early timing of reaching a given global warming level (GWL). Despite their unrealistic timings, these models may still be used to estimate the climate impacts occurring at a given GWL.²⁵

Close

Advances in Earth System Observations

Surface and satellite observational products continue to provide deep insights into recent changes in the Earth system. New analyses based on long-term surface observations provide improved constraints on regional plant productivity and the moderating influences of water and nutrients (e.g., AmeriFlux sites and National Ecological Observatory Network [NEON] domains), on surface and subsurface runoff (e.g., USGS observing networks), on surface energy balance (e.g., Atmospheric Radiation Measurement [ARM] and other Baseline Surface Radiation Network [BSRN] sites), on uncertainty in near-surface temperatures (e.g., Goddard Institute for Space Studies Surface Temperature [GISTEMP]), and on atmospheric CO₂ and related gases (e.g., NOAA Earth System Research Laboratories Global Monitoring Laboratory [NOAA ESRL GML]). In situ measurements of ocean temperature, salinity, and key biogeochemical concentrations are provided by buoys, ship tracks, floats and drifters (NOAA Global Drifter Program [GDP]), and Saildrones. Extended records from the Clouds and the Earth’s Radiant Energy System (CERES) and Atmospheric Infrared Sounder (AIRS) satellite missions now provide increasing confidence in warming of the Earth’s atmosphere and surface. Continued analysis of the now 30-year record of sea level heights from a series of satellite altimeters provides observations of a recent acceleration in sea level rise. However, continuing declines in the number of active in situ precipitation monitoring stations threaten to produce gaps in our precipitation observation record.²⁶

Recently deployed observing systems promise a deeper understanding of Earth’s physical systems and reduced uncertainty in climate projections. Recent NASA missions include the ECOsystem Spaceborne Thermal Radiometer Experiment on Space Station (ECOSTRESS) thermal imager, focused on vegetation temperature and response to climate stressors; the Global Ecosystem Dynamics Investigation (GEDI) ecosystem lidar, focused on the forest canopy; the Earth Surface Mineral Dust Source Investigation (EMIT) spectrometer; and the Surface Water and Ocean Topography (SWOT) mission in conjunction with the Centre National d’Études Spatiales, the Canadian Space Agency, and the United Kingdom Space Agency, focused on measuring surface water on land and ocean sea surface height. Other recently launched or forthcoming missions include NASA’s Time-Resolved Observations of Precipitation structure and storm Intensity with a Constellation of Smallsats (TROPICS) and NOAA’s Joint Polar Satellite System-2 (JPSS-2), focused on tropical cyclones and other extreme weather events. Additional frontiers in satellite-based climate monitoring include both hydrological (e.g., snow thickness) and atmospheric (e.g., clouds and atmospheric composition) properties. Besides satellite observations, NASA Jet Propulsion Laboratory’s Airborne Snow Observatory (ASO) and its commercial successor have performed numerous snow surveys, enabling extremely detailed maps of mountain snowpack. Also, recent efforts by USGS on the Next Generation Water Observing System (NGWOS) are enhancing the quality of real-time data on water quantity and quality from fixed and mobile instruments.

Combined with newer algorithms and data assimilation techniques, a growing number of observational data products enable tighter constraints for modern reanalysis datasets, such as Modern-Era Retrospective analysis for Research and Applications, Version 2 (MERRA2) and the European Centre for Medium-Range Weather Forecasts Reanalysis, Version 5 (ERA5). Improvements in the number and quality of observational data products have also enabled new process-oriented metrics and diagnostics (e.g., Leung et al. 2022; Maloney et al. 2019; Simpson et al. 2020^27,28,29), which in turn enhance the validation of Earth system models (ESMs).

New Scenarios and Climate Projections

A new generation of ESMs has produced an updated set of projections under many new policy-relevant scenarios. A number of ESMs have been run at significantly higher resolution (about 20–50 km) than the previous standard (about 100–200 km) for focused experiments to explore particular aspects of climatic changes, like tropical cyclones, that cannot be simulated by coarser-resolution models.

The Scenario Model Intercomparison Project (ScenarioMIP)³⁰ organized the main set of 21st-century projection experiments under the latest phase of the Coupled Model Intercomparison Project, CMIP6. These CMIP6 projections used newly developed scenarios based on Shared Socioeconomic Pathways (SSPs; Figure 3.4).^31,32,33 Five SSPs represent alternative plausible trajectories of 21st-century GDP and population growth and the pace and pervasiveness of socioeconomic and technological progress within individual nations and across the world (see Table 3 in the Guide to the Report). SSP-specific drivers are used in integrated assessment models (IAMs) to simulate the corresponding evolution of the energy system and the resulting greenhouse gas trajectories. These are in turn used by ESMs to drive their 21st-century projections. Some of these trajectories are also modified by imposing mitigation policies that meet specific warming targets (e.g., 1.5°C [2.7°F] or 2°C [3.6°F] above preindustrial levels) by the end of the century.³⁴ Figure 3.4 describes this process step-by-step.

Note that the mitigation assumptions in the various scenarios are not directly relatable to the mitigation discussion in Chapter 32, where the analysis is specifically centered on the current mitigation targets of the US in view of the Paris goals. The stringent goals of net-zero CO₂ emissions by 2050 that Chapter 32 discusses for the US are, however, consistent with the lowest scenario adopted in ScenarioMIP.

URL

Alternative text

Projections of future climate involve a multistep process using scenarios about future socioeconomic developments, policy goals, and emissions to drive Earth system models.

Figure 3.4. The graphic shows the chain of development leading to Earth system model experiments under CMIP6 ScenarioMIP, the most up-to-date 21st-century climate change projections. Five societal development pathways (the SSPs) were produced. Then, assumptions about climate mitigation policies that could be consistently applied to those socioeconomic futures were developed (the SPAs). Integrated assessment models took these baseline or mitigated pathways and produced alternative plausible trajectories of greenhouse gas emissions and land-use change. Earth system models used those emissions and land-use changes as inputs to produce the new CMIP6 scenarios of climate outcomes. Figure credit: Pacific Northwest National Laboratory, NOAA NCEI, and CISESS NC.

The new scenarios are labeled “SSPX-Y,” where SSPX (with X ranging from 1 to 5) identifies the SSP used to produce the greenhouse gas trajectories and Y indicates the magnitude of the radiative forcing by 2100 in W/m². The main 21st-century trajectories from the CMIP6 ensemble are SSP1-2.6, SSP2-4.5, SSP3-7.0, and SSP5-8.5, along with SSP1-1.9, the lowest emissions trajectory, which is designed to stay below 1.5°C (2.7°F) of warming above preindustrial levels. These scenarios are intended to provide a representative set of plausible alternative pathways of greenhouse gas and aerosol emissions and land-use changes, according to alternative societal and economic development trends over the 21st century. Three new scenarios correspond to three of the Representative Concentration Pathway (RCP) scenarios used in CMIP5 in their overall forcing levels (SSP1-2.6 with RCP2.6, SSP2-4.5 with RCP4.5, and SSP5-8.5 with RCP8.5), albeit with different details in their composition of gases and land use. Scenario projections can be used to explore climate outcomes under a coherent future trajectory of greenhouse gases and other anthropogenic forcings.

An alternative perspective on the effects of climate change uses GWLs as an organizing principle. One can ask when a certain GWL will be reached for a given scenario and climate model, thus connecting the two perspectives. CMIP6 includes many experiments performed by many ESMs at about 100 km resolution.³⁵ A multimodel approach, the High Resolution Model Intercomparison Project (HighResMIP),³⁶ is featured for the first time in CMIP6 to systematically investigate the impact of horizontal resolution. Models participating in HighResMIP have resolutions between 20 and 50 km and show more realistic simulations of intense storms and resulting precipitation.^37,38 At these resolutions, they can explicitly represent tropical cyclones,^39,40,41 and their simulations support the conclusion of a global decrease in tropical cyclone frequency⁴² together with an increase in intensity with warming. Their more refined topography also enhances their representation of local processes, such as the effects of warming on mountain snowpack—an important water source for the western United States.⁴³

Storyline simulations are another strategy to develop future projections.^44,45,46 In one variation of this approach, ensembles of short initialized weather hindcast simulations under imposed warming conditions can provide actionable information at regional or local scales.^47,48,49 Longer multiyear simulations can also be performed to understand changes in the statistics of weather extremes and storm characteristics (e.g., Chen et al. 2023; Gutmann et al. 2018^50,51) or to explore future change in climatic extremes such as drought (e.g., Ullrich et al. 2018⁵²).

Large Ensemble Simulations

Many realizations of the past and future climate are generated from climate model simulations started from initial conditions that are slightly perturbed from one another. These small random perturbations, the so-called butterfly effect, can subsequently alter the chaotic sequence of weather and climate events and, hence, the future course of the variability simulated by the model. Recently, climate model experiments have been conducted using multiple climate models with small random perturbations at the start of the simulations. The advent of these initial-condition large ensembles (henceforth referred to as large ensembles) in climate modeling provides an explicit framework for quantifying the relative contributions of external drivers and natural variability to regional climate change by sampling different possible sequences of natural variability that share a common forced response.⁵³ The application of large ensembles to assess the possible range of climate trends at local and regional scales is a major advance in recent years (Figure 3.5). New experiments from multiple climate models that capture uncertainty due to natural variability, changes in variability, and model uncertainty have improved our ability to understand and quantify expected climatic changes and the relative contributions of human-caused climate change and natural climate variability at regional scales. A new insight from the large ensemble archive is that different models have different representations of natural climate variability and how such variability responds to anthropogenic climate change. For some variables, such as precipitation, variability consistently increases with global warming, and large ensembles provide important context for quantifying and understanding these changes.

URL

Alternative text

Large ensemble simulations provide a plausible range of trends in winter surface air temperature combining the human-caused climate change and natural variability. The observed trend falls within the range simulated by the large ensemble for the historical period.

Figure 3.5. This figure shows past and projected future trends (in °F per decade) in winter (December–February) average air temperature over the United States and its territories (except the US-Affiliated Pacific Islands [USAPI]). The top row shows the observed trends during 1972–2021 (panel labeled “Observations”) and the plausible model range of trends simulated by the CESM2 100-member Large Ensemble during 1972–2021 (panels labeled “Warmer,” “Average,” and “Cooler”). Here, “warmer” means the ensemble member with the 5th largest US-average trend, “cooler” means the ensemble member with the 5th smallest US-average trend, and “average” means the average trend across all 100 ensemble members. Note that the model range results from the combined influences of simulated natural variability and human-caused climate change. The bottom row of maps shows the plausible model range of trends simulated by the Community Earth System Model 2 (CESM2) 100-member Large Ensemble projected for 2022–2071 under the SSP3-7.0 radiative forcing scenario (panels labeled “Warmer,” “Average,” and “Cooler”; see above for definitions). White areas on the maps are major lakes, including Great Salt Lake, Lake Okeechobee, and Lake Tahoe. Trend values in the USAPI (not shown) are very similar to those for the Hawaiian Islands. The box plot at the lower left shows the distribution of US-average trends simulated by the CESM2 100-member Large Ensemble for the period 1972–2021 (orange) and projected for the period 2022–2071 (red). The thin white line within each box denotes the average value, and the boundaries of the box show the 25th–75th percentile range. The black dot shows the observed US-average trend during 1972–2021, which lies within the plausible model range of trends. Figure credit: National Center for Atmospheric Research.

Emergent Constraints on Future Projections

An approach to reduce uncertainty in climate change projections has matured over the past decade. It is known as “emergent constraints.”⁵⁴ The term refers to strong statistical relationships between highly uncertain future climate parameters and observable trends or variations in the current climate, along with a physical explanation of this relationship. Observations of the current climate are used to quantify the difference between simulated and observed values in a model simulation (referred to as model bias) and subsequently constrain the future climate parameters. Research on emergent constraints has targeted a wide variety of geophysical processes.⁵⁵ For example, projections of western US runoff in CMIP5 models can be constrained by the observed sensitivity of runoff to precipitation in the historical climate.⁵⁶ A linear relationship between the projected summer warming and the model warm bias over the central US in CMIP5 models can be used to correct the future temperature projections,⁵⁷ and the observed interhemispheric asymmetry of the intertropical convergence zone has been used to correct future projections of California winter precipitation.⁵⁸

Extreme Event Attribution

The science of evaluating the effects of human-caused climate change on extreme weather and climate events has advanced significantly. Human influence has changed the frequency and intensity of some types of extreme events, and it is now possible to quantify the influence of anthropogenic climate change on certain types of specific extreme events.

Extreme event attribution quantifies the current human influence on observed severe weather events, primarily through changes in magnitude and frequency.⁵⁹ Recent methodological advances have widened the classes of weather events analyzed and extended these analyses beyond single events to include hazards throughout an entire season (e.g., Herring et al. 2022, 2019, 2018; Reed et al. 2022; Wehner et al. 2019^{60,61,62,63,64}). Confidence in attribution statements is increased when multiple methods, observations, and models lead to similar conclusions⁶⁵ and similar underlying trends are detected.⁶⁶

Recent methodological advances include graphical methods to identify cause and effect pathways,⁶⁷ using physical insight to inform statistical models,^68,69 factual and counterfactual simulations,⁷⁰ and large ensembles.⁷¹ The CMIP5/6 databases are often used in event attribution studies, but the relative coarseness of the model grids (100 km or more) limits applicability to large-scale events such as certain heatwaves⁷² and winter storms.⁷³ Finer-scale events such as intense storms are often amenable to storyline hindcast simulations (see New Scenarios and Climate Projections above), using higher-resolution regional models to compare the “storm that was” with observed climate change to the “storm that might have been” without the human changes to the climate system.^74,75 However, by construction, such storyline analyses inform attribution statements about changes in the magnitude of an event not the frequency.

Attribution methods have advanced such that rapid statements can be made just weeks after an event^76,77 and even forecasted,⁷⁸ and these analyses could be made operational.⁷⁹ Despite recent progress, the human influence on some extreme weather events may not be attributable because of model limitations.^80,81

Extreme event attribution has shown that some extreme events are happening with greater frequency, magnitude, and duration due to anthropogenic climate change. For example, climate change very likely made a 2016 extreme precipitation event in Louisiana heavier⁷⁷ and increased both precipitation^82,83,84 and flooding⁸⁵ during Hurricane Harvey. While these studies reveal how climate change affects extreme events, they also reveal that historical observations of climate are often insufficient for characterizing future risks.

Close

Natural Variability

At all spatial scales, the climate response is forced by anthropogenic drivers, which are external to the climate system. This response occurs against a background of natural climate variability (i.e., internal to the climate system). Such variability is generated by natural processes, for example, by atmosphere–ocean interaction (e.g., El Niño and La Niña events), atmosphere–land interaction, or chaotic variability within the atmosphere itself. Depending on the sequencing and magnitude of natural variability and the magnitude of the climate change response, natural variability can mask the climate change signal, amplify it, or be overwhelmed by it. Natural variability therefore adds uncertainty to climate change projections. Because the sequencing of natural variability is largely unpredictable, this component of uncertainty in climate change projections is irreducible. This is in contrast with the uncertainty associated with the climate change response, which arises from lack of information about future greenhouse gas and aerosol emissions, as well as incomplete understanding of climate processes and associated limitations in terms of how well models reproduced observed changes. These are knowledge gaps that can, in principle, be filled (KM 3.2).

The magnitude of natural variability generally increases as spatial scales decrease (Figure 3.6). At the scale of a typical Earth system model (ESM) grid cell, natural variability is the dominant source of uncertainty in precipitation and temperature projections over much of the US. On continental scales, model and scenario uncertainties become more important.⁸⁶ Therefore, at the regional scales and the multidecadal time horizons relevant for adaptation planning, uncertainty due to natural variability can be a larger contributor to overall uncertainty in climate change than model or scenario uncertainties (Figure 3.6; e.g., Dong et al. 2021⁸⁷). Furthermore, it is important to note that anthropogenic forcing also changes the variability of some Earth system processes, and these forced changes contribute to changes in the frequency, duration, and intensity of extreme events, such as heatwaves and heavy precipitation.

URL

Alternative text

The sources of uncertainty in climate projections vary depending on timescale and geographical scale and for different aspects of the climate system.

Figure 3.6. These charts show the relative importance of different sources of uncertainty (natural variability, model uncertainty, and scenario uncertainty) for projections of moving decadal average temperature (top row) and precipitation (bottom row) for the globe (first column), Northern Hemisphere land (second column), and the Southern Great Plains (third column) from 1999 to 2099 (decadal averages are plotted mid-decade; for example, the x-axis starts in 2004) relative to the reference period of 1995–2014, based on CMIP6 models. Model uncertainty is calculated as the variance across models’ forced response estimates. Scenario uncertainty is calculated as the variance across multimodel averages for different scenarios. The shadings are constructed as a symmetric 90% range around the multimodel, multiscenario average projection. Figure credit: Cornell University, National Center for Atmospheric Research, and Pacific Northwest National Laboratory.

Atmospheric Circulation Changes

Past and future climate changes in the United States are strongly modulated by atmospheric circulation features such as the semipermanent North Pacific and North Atlantic subtropical high–pressure systems, the Aleutian Low, the meandering jet stream stretching from the North Pacific to the North Atlantic, extratropical storm tracks and fronts, and the North American Monsoon. These regional- to continental-scale circulation features are in turn modulated by even-larger-scale overturning circulations such as the Hadley cell (see Perlwitz et al. 2017⁸⁸). Regional circulation features over North America are also modulated by recurring remote patterns of variability such as the El Niño–Southern Oscillation (ENSO) and North Atlantic Oscillation and by variability due to the chaotic dynamics within the atmosphere (the so-called butterfly effect).

Anthropogenic forcing alters the vertical and horizontal energy and moisture distributions in the atmosphere. Atmospheric circulation is directly impacted by the resulting changes in the temperature gradient from the equator toward the poles (both in the lower and upper troposphere), the decrease in the rate that temperature falls with height above the surface, and increased latent heating.

A synthesis of observational and modeling studies estimates that the tropics have widened by about 0.5° of latitude per decade since 1979. Although climate models driven by external forcing simulate, on average, a larger expansion rate than observed, the observed rate is within the bounds of the model simulations.⁸⁹ The discrepancy between the observed and simulated rates of tropical widening is partly due to patterns of natural variability such as ENSO and the Pacific Decadal Oscillation, as well as natural atmospheric variability.^90,91

Consistent with the tropical expansion, climate models project poleward shifts of the jet streams and storm tracks and poleward expansion of the subtropical dry zones with warming,⁸⁸ with associated impacts on the US. However, the Northern Hemisphere jet shifts are regionally and seasonally dependent (e.g., Oudar et al. 2020; Zhou et al. 2022^92,93). Simulations show a poleward jet shift in the central North Pacific with warming, along with an eastward jet extension that steers more Pacific storms toward California.^94,95 However, large uncertainty remains because the jet extension is also influenced by natural variability.⁸⁷ With global warming shifting the North American westerly jet poleward throughout the warm season, climate models project late-spring wetting and late-summer drying in the Midwest.⁹⁶ Poleward expansion of the North Atlantic subtropical high may intensify and extend the Great Plains low-level jet poleward, affecting warm-season precipitation in the Midwest.⁹⁷

Besides changes connected to the tropical expansion, climate models project a weakening of the North American Monsoon with warming due to increased atmospheric stability.⁹⁸ Similar to other monsoon systems, the onset of North American Monsoon rainfall is projected to be delayed by warming.^99,100,101 In the North Pacific, warming is projected to reduce the frequency of atmospheric blocking during winter.¹⁰²

Water Cycle Changes

Many processes relevant to the water cycle have already begun to change and are projected to continue changing as the planet warms. These include the atmospheric circulation changes noted above as well as changes in atmospheric moisture, patterns of natural variability, the magnitude of variability of the water cycle, and the modulating role of vegetation on evaporation. These changes are driving changes in the intensity of precipitation in extreme events; snowfall, snowpack, and snow melt; and the seasonality of average precipitation and evaporation.

Seasonal and annual average precipitation and evaporation patterns have been changing with global warming. Precipitation from extreme events is projected to increase with warming (e.g., Neelin et al. 2017¹⁰³). Heavy precipitation events constitute a large fraction of total precipitation and also a large fraction of the change in precipitation in both observations and projections.^{104,105,106,107} However, it is possible to have increases in extreme precipitation but decreases in annual average precipitation at the same location due to large declines in non-extreme precipitation (Figure 3.7)^98,108,109 and overall changes in the variability of precipitation. Precipitation variability generally is projected to increase at all timescales and in most locations in response to anthropogenic forcing.¹¹⁰ This implies increases in both heavy precipitation extremes and drought,¹¹¹ a phenomenon called whiplash.¹¹²

Water vapor factors into the intensity of precipitation extremes, such that increased water vapor due to warmer air temperatures alone would drive relatively uniform projected increases in precipitation extremes over much of the planet. Additionally, changes in circulation modulate the spatial pattern of extreme precipitation intensity change, increasing it in some regions and reducing it in others.¹¹³ For extratropical cyclones in summer, precipitation intensity increases, and more energetic convective storms are projected despite weakening circulation.¹¹⁴ While understanding of these changes in extreme precipitation has improved, recent work has also highlighted the oftentimes large magnitude of natural variability in extreme precipitation (KM 3.5).

URL

Alternative text

As the climate warms, extreme precipitation events become more intense and make up a larger fraction of total precipitation, while moderate events become less common.

Figure 3.7. The graphic depicts example changes in the precipitation distribution at many locations as projected by climate models. (top) In general, warming shifts the distribution toward a greater probability of large events and a reduced probability of light to moderate events, with only a modest change in the average total precipitation over a season or year. (bottom) Greater levels of water vapor in the atmosphere in a warmer world drive more precipitation during storms, when moist air masses converge. The increased water vapor convergence, or moist inflow, in very stormy areas also transports more water vapor out of surrounding areas, reducing any light to moderate precipitation there. The changes in total precipitation are often modest, because they reflect this tug-of-war between opposing changes in heavy and light to moderate precipitation. Changes in circulation can also affect the changes in all parts of the precipitation distribution. Certainty in projected changes differs among the underlying processes: the increase in water vapor and associated increase in extremes is very certain, while the changes in circulation are less certain. Further discussion related to top panel can be found in Fischer and Knutti (2016).¹¹⁵ Figure credit: Cornell University and University of California, Los Angeles.

Changes in precipitation seasonality can occur with or without changes in average annual total precipitation. Increases in the amplitude of the precipitation seasonal cycle may be expected, as precipitation increases proportionally to its present-day seasonal cycle. However, other changes in the seasonal cycle of precipitation have also been found. Examples of these changes include a sharpening of the seasonal cycle in California,^112,116,117 a substantial increase in late-spring precipitation and a substantial decrease in late-summer precipitation in the US Midwest,⁹⁶ and a delayed onset of rainfall in many monsoon regions.^100,101

Even if precipitation increases in some regions in the future, one concern is that evaporation will increase with warming, leaving the land surface drier (also see Ch. 4). Model projections and observational analyses for the US suggest that the net effect will be for warming to increase evaporation and surface drying in much of the country.

With warming temperatures, more precipitation falls as rain instead of snow. Snowpack in many mountainous regions of the US has already decreased on average over the last few decades; for example, the western US snowpack-driven reservoirs are already exhibiting changes.¹¹⁸ These declines are expected to be exacerbated by greater increases in winter temperatures compared to summer temperatures in some regions due to snow albedo feedbacks. Another factor that decreases snowpack is the larger temperature increases at higher altitudes, which increase the frequency of multiyear snow droughts.¹¹⁹ Decreases in warm-season streamflow have also been observed and projected because of the increase in evaporation with warmer temperatures.^120,121,122 Decreasing snowpack and increases in flooding are both concerns for water management. Increasing flooding is driven by the increasing intensity of extreme precipitation discussed above, by the shift from snowfall to rainfall, and potentially also by increases in rain-on-snow events, which can lead to rapid snowmelt (for one example, see Box 29.2).

Changes in the Carbon and Biogeochemical Cycles

Understanding of the biogeochemical responses to human-caused greenhouse gas emissions has increased due to expanded observations and improvements in models. However, uncertainty in the future evolution of the global carbon cycle remains high (KM 2.3).

Terrestrial ecosystems and the oceans take up a little more than half of the CO₂ emitted from human activities, partially offsetting the climate effects of carbon emissions. From 2011 to 2020, annual emissions averaged 10.6 ± 1.2 PgC, and the emitted carbon was ultimately distributed between the atmosphere (47%), vegetation and soils in terrestrial ecosystems (29%), and the oceans (24%; Figure 3.8).¹²³ Over the past six decades, the average fraction of anthropogenic CO₂ emissions that has accumulated in the atmosphere has remained nearly constant at 41%, even as fossil fuel emissions have rapidly increased.

URL

Alternative text

While the land and ocean take up some of the carbon dioxide from human activities, the rest continues to accumulate in the atmosphere every year.

Figure 3.8. The primary sources of anthropogenic carbon are fossil fuel and land-use change emissions (upward thick arrows). While the land biosphere and ocean continue to take up the same proportion of anthropogenic carbon every decade (downward thick arrows), atmospheric carbon dioxide (CO₂) levels continue to rise as fossil fuel and land-use change emissions continue over time. Represented here is the budget of the global carbon cycle each year from estimates averaged globally for the decade 2011–2020, with flux estimates from Canadell et al. 2021.⁴ Values are in Petagrams of carbon (PgC). Adapted from Friedlingstein et al. 2022¹²³ [CC BY 4.0].

An important question is whether the same fraction of fossil fuel emissions will continue to be offset by ocean and land biosphere uptake or whether the offset will slow or reverse over time. Answering this question depends on understanding feedbacks between the carbon cycle and climate change. The additional carbon sink on land is a consequence of multiple ecosystem processes, including increased photosynthesis responses to rising atmospheric CO₂,¹²⁴ nitrogen deposition, fire suppression, and forest regrowth after disturbances such as land clearing for agriculture.¹²⁵ Large-scale observational constraints on increased vegetation productivity due to increased atmospheric CO₂ are uncertain.¹²⁶ Ocean waters take up an increasing amount of carbon due to chemical dissolution of CO₂ in seawater, biological fixation through photosynthesis, shell formation in some organisms, and subsequent transport of carbon into deeper waters.

As atmospheric CO₂ increases, plant productivity and soil carbon also increase and are partly responsible for increases in land carbon storage, which functions as a negative feedback on climate change. However, several processes could reduce terrestrial carbon uptake, including increasing temperature and drought frequency or intensity, limitation of other necessary nutrients, and land-use changes such as deforestation. In the Arctic, potentially large biogenic releases of CO₂ and CH₄ to the atmosphere due to permafrost thaw—the timing of which may be accelerated by increasing wildfire at high latitudes¹²⁷—could be a potentially large positive climate feedback that may alter the effectiveness of climate mitigation strategies.¹²⁸ ESMs predict a total loss of near-surface permafrost by 2100 for intermediate (RCP4.5) and very high (RCP8.5) emissions scenarios, and permafrost carbon emissions are projected to outpace increased carbon uptake from higher vegetation productivity in a warmer Arctic.¹²⁹

Uptake of CO₂, an acid gas, by oceans leads to acidification of seawater through a series of chemical reactions that lower the pH and availability of carbonate ions in addition to other chemical changes (see Chapter 2 for ocean acidification trend) that have been shown to be harmful to marine life.¹³⁰ Over time, uptake of carbon could slow as ocean waters become saturated with dissolved CO₂.¹³¹ Recent observations show that ocean carbon processes are starting to change in response to the growing ocean carbon sink, and these changes are expected to contribute to future weakening of the ocean carbon sink under medium to high emissions scenarios (similar to SSP2-4.5 and SSP5-8.5 of Figure 3.7 in Jiang et al. 2019¹³²).¹³³

A comparison between the CMIP5 and CMIP6 generation of ESMs with respect to the representation of the CO₂ fertilization effect and carbon cycle–climate feedbacks suggests that uncertainties in these processes have remained large and virtually unchanged in recent years.¹³⁴ However, the latest generation of land models have improved representation of biogeochemical cycles,⁴ and ocean models have exhibited improved representation of coupled physics and biogeochemical cycles.¹³⁵

In the United States, the land carbon sink is dominated by forests, which have expanded in the last century due to fire exclusion in the West and to secondary forest regrowth following agricultural abandonment in the East. The growth of the North American forest biomass carbon sink is expected to become more saturated and less effective over time (Focus on Western Wildfires).^125,136 The climate change–wildfire relationship is also expected to play an increasingly dominant role in the strength of the US land carbon sink, requiring increased management of forests to meet carbon storage and other needs of society.¹³⁷ In addition, increasing temperature and moisture in some regions could lead to increased microbial emissions of CH₄ from wetlands and Arctic tundra.

Changes in the Ocean

The ocean has a large capacity to store and release heat and has been able to absorb 91% of the excess heat attributed to greenhouse gas emissions.² The ocean both exchanges heat with the atmosphere and moves heat from the tropics toward polar regions, where warm surface ocean waters transform into cooler, high-density waters that sink (Figure 3.9),¹³⁸ taking high carbon concentrations with them.¹³⁹ The Atlantic Meridional Overturning Circulation (AMOC) is an important component of the global ocean circulation, transporting heat and carbon and affecting hydroclimate, hurricane activity, and coastal sea level. It is hypothesized from some proxy evidence that the AMOC has declined since the Industrial Revolution.¹⁴⁰ It remains uncertain whether observations support the anthropogenically forced weakening of the AMOC during the past four decades predicted by climate models.¹⁴¹

URL

Alternative text

Climate change has multiple effects on the ocean, atmosphere, and cryosphere and their complex interactions.

Figure 3.9. The figure shows important physical processes that play a role in the ocean and cryosphere, along with their linkages. Associated climate change–related effects, including sea level rise, increasing ocean heat content, ocean acidification, marine heatwaves, and ice mass loss, are also shown. The arrows indicate an exchange taking place between ice, ocean, and atmosphere. Adapted with permission from Figure TS.2 in IPCC 2019.¹⁴²

On short timescales (annual to decadal), ocean circulation dominates the pattern of changes in ocean heat content; although on longer timescales, the spatial pattern is primarily associated with the addition of excess heat into the ocean.¹⁴³ Since 2005, Argo profiling floats have provided observations of temperature and salinity changes in the global ocean to a depth of 2,000 m. These measurements show an increase in global ocean heat content, although there is geographic variability. On shorter timescales, marine heatwaves are periods of extreme high ocean temperature relative to the long-term average seasonal cycle.¹⁴⁴ Persistent marine heatwaves have been observed in the northeastern Pacific since around 2014,^145,146,147 with associated negative impacts on ocean ecosystems (Ch. 27; Figure 10.2).

Changes in the Cryosphere 

The cryosphere is the frozen part of the Earth system and includes ice sheets, glaciers, sea ice, permafrost, and snow (Figure 3.9). Observations of the cryosphere have expanded in recent years, including from satellites like Gravity Recovery and Climate Experiment Follow-On (GRACE-FO) and Ice, Cloud, and Land Elevation Satellite-2 (ICESat-2).¹⁴⁸ These and other complementary observations of the cryosphere show declines in ice sheet and glacier mass, snow cover, and Arctic sea ice.¹⁴⁹ Through efforts like the Ice Sheet Mass-Balance Inter-comparison Exercise (IMBIE),¹⁵⁰ there is clear evidence of loss of ice sheet mass from 1993 to 2020, and the rate of loss has increased for both the Greenland and Antarctic ice sheets. There is increased consensus regarding the behavior of many ice sheet processes. However, as the climate warms, the future response of the ice sheets and some associated processes are still uncertain.¹⁴⁹ Two specific processes, marine ice cliff instability^151,152 and marine ice sheet instability,^{153,154,155,156,157} could lead to rapid ice sheet loss over several decades, but the physical processes that would result in these instabilities and how they would progress once triggered remain uncertain. The decline in sea ice affects several critical functions: sea ice serves to regulate climate by reflecting solar radiation; inhibits ocean–atmosphere exchange of heat, momentum, and gases; and supports global deep-ocean circulation, polar species, and livelihoods of people in the Arctic.

Art × Climate

Pia De Girolamo
Invasion
(2020, acrylic on canvas)

Artist’s statement: My painting depicts a polar bear in the Arctic viewing a distant cruise ship. I was on such a ship, visiting the Arctic Circle in 2019. The polar bear's look is wondering and poignant; the cruise ship seems innocent, but it represents another human incursion into this place of beauty. Climate change has already affected the area; the ice that allows the polar bear to travel and find food is shrinking. Bear and boat are on opposite sides in this painting. We need to be on the same side as the bear.

View the full Art × Climate gallery.

Artworks and artists’ statements are not official Assessment products.

Sea Level Rise

Over long time periods, the main drivers of changes in global mean sea level (GMSL) are thermal expansion due to heating of the ocean and the addition of water associated with melting ice from ice sheets and glaciers. Human-caused changes in the movement of water between ocean and land, including from groundwater depletion and water impoundment associated with dam building, have a minor impact on GMSL, although they can increase in importance for specific time periods.¹⁵⁸ The increase in GMSL during the 20th century estimated from tide-gauge records has been explained by the individual processes contributing to it.¹⁵⁸ The rate of the 20th-century increase in sea level was faster than in any other century in at least the last 3,000 years.^159,160 In the past two decades, the changes in GMSL measured by satellite altimetry have matched the expected sea level rise based on the combination of in situ measurements of the Argo profiling floats and the observations of water-mass change from the GRACE and GRACE-FO satellites. Data from the now three-decades-long satellite altimetry record shows that the rate of GMSL rise has been increasing, driven by accelerating contributions from the underlying processes.^158,161,162

Sea level rise is not uniform across the globe. Relative sea level rise at any specific location responds to processes that are important at regional and local scales.^{149,163,164,165} On short timescales and in short records, natural variations on interannual to decadal timescales can affect estimates of rates and accelerations. Over long time periods, there are three main causes of regional variations in relative sea level: 1) sterodynamic changes (the combination of thermal expansion and ocean dynamics that is driving global sea level rise); 2) gravitational, rotational, and deformational changes that result from water moving from the cryosphere to the ocean; and 3) vertical land movement (subsidence or uplift) due to glacial isostatic adjustment, tectonics, sediment compaction, groundwater and fossil fuel withdrawals, and other non-climatic factors.^166,167

An improved understanding of the drivers of regional mean sea level rise and how processes combine to cause sea level change at the coast has led to better assessments of the frequency, duration, and timing of high-water levels and coastal flooding events (Figure 3.10).¹⁶⁷ Regional sea level change has been the main driver of changes in extreme water levels.¹⁶⁷ Due to ongoing regional relative sea level rise and narrowing of the gap between the typical high tide and flooding threshold, the frequency of high tide flooding has increased. Natural ocean variability, including that associated with tides and large-scale climate signals that did not necessarily result in flooding in the past, is expected to lead to a rapid increase in the amount of high tide flooding in the coming decades when combined with future sea level rise (see Ch. 9).^167,168,169

URL

Alternative text

Sea level rise is increasing the probability of coastal flooding and associated impacts.

Figure 3.10. The schematic shows the physical factors affecting coastal flood exposure. Nearshore processes like storm surge, wave setup (the increase in water level due to the presence of breaking waves), and swash (wave runup and wave rundown) combine with other large-scale processes to drive impacts. Due to the clear and strong relative sea level rise signal—that is, the combination of global increases in sea level and the fact that land is sinking in many US coastal areas—the probability of flooding and impacts is increasing along most US coastlines. Source: Sweet et al. 2022.¹⁶⁷

Regional-Scale Changes

Many of the changes in Earth system processes discussed above can vary across a range of spatial scales. For example, at the continental scale (less than 10,000 km), the climate system’s warming response to anthropogenic forcing is affected by differences in how quickly land and ocean areas warm, with land generally warming more quickly than the ocean. This is because land areas have a lower heat capacity and thus respond more quickly than the ocean to anthropogenic forcing, and because of the larger cooling effect of evaporation from the ocean surface. Therefore, North America generally warms more than adjacent oceans (Figure 3.11). The high latitudes of the Northern Hemisphere warm the most of any region, although the reasons for this are complex and may involve processes poorly captured by global climate models.¹⁷⁰ Precipitation responses, arising from the changes in the atmospheric circulation and water cycle processes discussed above, tend to be organized in latitude bands such as the tropics, subtropics, midlatitudes, and high latitudes.

URL

Alternative text

While temperatures have been rising almost everywhere, warming has not occurred uniformly over the planet.

Figure 3.11. The maps show observed global (left) and overland US (right) trends (in °F per decade) in annual average near-surface temperature over the period 1972–2021. While temperatures have been increasing almost everywhere, warming has not occurred uniformly over either the planet or the US. Temperatures over land have been increasing faster than over the ocean, and the Arctic has been warming at more than twice the global average rate. Temperatures over the US have increased faster in Alaska, at high elevations, and in regions with significant seasonal snowpack. Figure credit: University of California, Los Angeles; University of California, Davis; NOAA NCEI; and CISESS NC.

At the regional scale (less than 1,000 km), variations in land surface properties are associated with differences in the local climate responses to anthropogenic forcing. Examples from the US (Figure 3.11, US panel) include greater warming at snow margins in western US mountains,¹⁷¹ greater warming in inland areas separated from the coast by California mountain ranges,¹⁷² greater warming in extremely arid landscapes,¹⁷³ shifts of precipitation downwind in mountainous areas,¹⁷⁴ snowfall loss at higher elevations where temperatures increase above the freezing line,^175,176 changes in snowmelt-driven runoff as a function of elevation,¹⁷⁷ changes in snowfall over large lakes,¹⁷⁸ and changes in evaporation and soil moisture due to variations in vegetation. Regional variations in the changes of temperature and precipitation are also driven partly by the regional distributions of aerosols, which interact with clouds and radiation.¹⁷⁹ Downscaling techniques, which convert global model outputs from lower resolution (about 100 km) to high resolution (1–10 km), are often used to simulate these phenomena for adaptation planning (see App. 3).

Close

Extreme Heat and Cold

Changes in temperature extremes in recent decades are driven primarily by trends toward warmer conditions rather than any changes in variability.¹⁸⁰ Consequently, the frequency and intensity of cold extremes have declined over much of the United States while the frequency and intensity of extreme heat have increased.¹⁸¹ Arctic warming may also drive increases in the occurrence and persistence of circulation anomalies that are related to extreme cold and heat,¹⁸² although the evidence that these mechanisms have played a role in recent events is mixed.^183,184 Climate change may also be contributing to “false spring” events,¹⁸⁵ where early warming has caused early budbreak and flowering of plants, exposing them to damaging frost and freeze events.¹⁸⁶

Extreme Precipitation and Flooding

Observed increases in extreme precipitation intensity at the continental scale in North America have been, for the first time, attributed to human influences.¹⁸⁷ With warming, it is expected, and has been documented, that more winter precipitation will fall as rain instead of snow,¹⁸⁸ although the projected intensity of most extreme snowstorms remains uncertain.¹⁸⁹ Projected increases in extreme precipitation events are larger in the winter season since warming in winter is larger, including events related to atmospheric rivers.¹⁹⁰ Precipitation associated with hurricanes increases with warming at least as much as water vapor,⁴² and the heaviest events can increase at an even greater rate.^63,191,192 Where and when increases in extreme precipitation manifest in any given year or even decade is strongly subject to natural variability (see also KM 3.3). Notably, increases in extreme precipitation events do not always directly translate to increases in river flooding, in part because of the many processes at the land surface that affect flood events (KM 4.1; Figure 4.8).

Drought

Drought is broadly defined as a transient period of below-average water availability, typically expressed in terms of fluctuations in precipitation, soil moisture, or streamflow and runoff.¹⁹³ Drought is a complex phenomenon (Figure 3.12) that depends on fluctuations in moisture supply, direct losses of moisture to the atmosphere, and ecosystem and land surface processes.^194,195,196 Western North America experienced several decadal-scale droughts during the 20th century¹⁹⁷ and numerous multidecadal “megadroughts” prior to AD 1600¹⁹⁸ and is currently experiencing an ongoing megadrought largely unprecedented over the last 1,200 years.^199,200 See Chapter 6 for more information regarding the effect of changes in the water cycle over land.

URL

Alternative text

Climate change alters the hydrologic cycle and is expected to increase drought in some regions through various process pathways.

Figure 3.12. Changes in climatic drivers (e.g., precipitation, temperature, wind, etc.) affect different aspects of the hydrologic cycle (e.g., evapotranspiration, snowpack, streamflow, soil moisture). In turn, these hydrologic shifts translate into changes in the severity, frequency, and risk of different drought types. Plus and minus signs denote the direction of change in the driver that would cause increases in drought. For example, where precipitation declines (down arrow), all drought types will increase because this reduces snowpack, streamflow, groundwater and reservoir storage, and soil moisture. Similarly, increasing temperatures (up arrow) are also expected to increase hydrological and biophysical drought by reducing snowpack and increasing evaporative losses from streams, surface reservoirs, and soils. Adapted with permission from Figure 8.6 in Douville et al. 2021.²⁰¹

Observed and projected hydroclimatic changes in response to external forcing are highly season- and region-dependent over the US, especially in the West.²⁰² Natural variability also plays a prominent role in shaping hydroclimate on annual and decadal timescales,²⁰³ as indicated by the decrease in total area coverage of low soil moisture over the US since 1915²⁰⁴ and increasing Missouri River flows,²⁰⁵ a region that is expected to become drier under climate change (e.g., Cook et al. 2015²⁰⁶). Climate change is expected to amplify drought, primarily through warming-induced increases in evaporative demand and surface water losses^207,208 and declines in snowfall and water stored in winter snowpack.⁴³ There is strong evidence that these processes are already amplifying hydrological drought severity in California,^209,210,211 the Pacific Northwest,^212,213 the Colorado River basin,^121,214 and across southwestern North America.²⁰⁰ However, there are uncertainties in how these processes may present in the future. For example, changes in plant water use in response to increasing temperatures and rising atmospheric CO₂ are complex and poorly understood and may either ameliorate¹⁹⁵ or amplify²¹⁵ soil moisture and runoff droughts at the surface.

Wildfire

The direct influence of climate change on modern and future wildfire activity arises from the effect of warming on fuel moisture content and flammability (Focus on Western Wildfires). Flammability is directly related to the vapor pressure deficit (VPD), an integrative measure of atmospheric aridity or dryness. High VPD is strongly associated with the hot and dry weather conditions conducive to the intensification and spread of wildfires, as well as with drier fuels at the surface. In recent decades, warmer temperatures, declines in humidity, and increases in VPD have caused large-scale increases in fire weather²¹⁶ and areas burned by wildfire^217,218,219 across the West. From 1984 to 2015, about half of the increase in burned area across the western United States is attributable to increases in fuel flammability caused by anthropogenic climate change. These climate change–driven increases in wildfire burned area are expected to continue into the coming decades, as fuel availability is not expected to be a limiting factor before 2050.²²⁰

Compound Events

Compound events refer to the combination of multiple weather or climate events—which individually may not be extreme—that together pose risks to human or natural systems (see example in Figure 3.13)^221,222,223 and are described in further detail in Focus on Compound Events. Since climate change may affect the frequency and magnitude of the individual components of compound events in different ways, changes in compound events can be complex. Recent advances in statistical techniques used to describe compound events^{224,225,226,227} offer rigorous ways to quantify changes. Progress has also been made in categorizing compound event types in this rapidly evolving field.^223,228,229

URL

Alternative text

Consecutive events caused significant human health and economic impacts in Southern California from 2012 through 2018.

Figure 3.13. A series of events in Southern California produced a cascade of impacts on human health and the economy. These include (from top left to bottom right): a prolonged drought from 2012 to 2016; above-average winter precipitation in 2017, enhancing growth of fuels; a dry, warm spring and summer, reducing moisture levels and drying existing vegetation; record-setting Santa Ana winds; fires occurring shortly thereafter; and rainfall over the burned area, leading to debris flow event in Montecito, CA. Adapted from AghaKouchak et al. 2020;²³⁰ modified with permission from the Annual Review of Earth and Planetary Sciences, Volume 48 © 2020 by Annual Reviews, http://www.annualreviews.org.

Close