4 Topics from Chapter 4: Future Global Climate: Scenario-Based Projections and Near-Term Information
What’s in a model? Modeling basics and understanding the system behind the 6th IPCC Report
Raven Benko
Dr. Shuyi Chen, in a seminar on the intersection between climate change and public health stated that, “the largest uncertainty in climate change predictions is not in scientists’ ability to accurately model climate change, but in what humanity is going to do with the information.” The most recent Intergovernmental Panel on Climate Change (IPCC) report (IPCC-AR6) provides several potential future scenarios of what Earth’s climate future may look like in the face of varying levels of anthropogenic climate change using low to high emission schemes. The probability of any one emission scheme occurring is unknown, as it is difficult and even impossible to predict how governments and nations are going to react and change emissions behaviors. Therefore, climate scientists use models to predict what could happen to Earth’s climate given different global emission levels to provide some sort of context about the impact of human’s current and future emissions activity.
But what exactly is a climate model? What is a model in general? How do scientists predict the future when surrounded by so much uncertainty? In plain language, the following report will give a basic overview of how models – and specifically the climate models used in the 2021 IPCC report – work and what they can tell us about our potential future climate.
I. What is a climate model?
Models are used throughout science to explain natural phenomena predict natural scenarios. In biology, we use model organisms to study how diseases progress in animals related to humans, so that we can learn how diseases progress and medications may work in humans. Models can also be a simple physical representation of a thing – like a miniature train or building. Climate models are largely mathematical equations that mimic physical processes overlayed over a computerized version of Earth – equip with topography, clouds, water, rain, and heat transfers.
As shown in Figure 1, this first component of a climate model is to build a computerized version of Earth, with all its oceans, mountains, land masses, clouds, and ice sheets. This Earth miniature is then partitioned into many 3D cubes – kind of how a photo is made up of pixels. These cubes stretch around Earth’s surface as well as the ocean and up into the atmosphere. Once the earth is partitioned into 3D pixels, a supercomputer applies an array of mathematical equations that represent physical climate processes. These equations dictate the movement of wind, clouds, water masses, as well as the transfer of energy, moisture, and chemicals. Each equation is calculated within one cube and is averaged over space and time to give us an overall prediction of climatic processes. The outputs of climate models are often graphics as in Figure 1 that show average annual precipitation or surface temperatures. These “global climate indicators” included in the model outputs for the IPCC-AR6 report include global surface air temperature, global land precipitation, Arctic sea-ice area, global mean sea level, the Atlantic Meridional Overturning Circulation, global mean ocean surface pH, carbon uptake by land and ocean, the global monsoon, the Northern and Southern Annular Modes, and the El Nino-Southern Oscillation.
II. What types of models are used in the IPCC-AR6 report?
The IPCC-AR6 report focuses primarily on two types of models: comprehensive climate models (atmosphere-ocean general circulation models, AOGCMs) and Earth System Models (ESMs). AOGCMs incorporate the physics of ocean and atmosphere circulation and how it interacts with heating from the sun and warming due to trapped heat from greenhouse gases. The ESMs differ in that they incorporate some biogeochemical cycles as well – carbon cycling within the biotic (living) environment.
Models take an immense amount of computing power; each one of those 3D pixels have dozens of calculations occurring simultaneously and interacting with other pixels. Therefore, only a few research stations around the world “run” (i.e. make a supercomputer compute all the required mathematical calculations within each 3D grid around a computerized version of Earth) climate models. The IPCC-AR6 report incorporates information from experiments run with multiple models that help ensure that differences in models are taken into when interpreting results. These multi-models are called CIMPs. There is high confidence in these multi-models for accurately predicting changes in climate.
III. Model Uncertainty
But how do we know if a model works? Well, every model will have a certain amount of uncertainty, simply because there are so many factors that influence the climate, it is impossible to incorporate them all. Some factors, like biological processes, are extremely difficult to incorporate and so they are often left out of climate models. Other processes, like positive and negative feedbacks are also extremely difficult to capture with mathematical models as they continue to grow and change. For example, melting icecaps represent a positive feedback for climate warming in that as more ice melts, there are less reflective surfaces on earth to reflect solar radiation back to space, therefore more radiation is absorbed at Earth’s surface causing even more warming, greater icecap melting – and the cycle repeats.
One main source of uncertainty in climate models is in how humanity will continue to act and emit. It is recognized in the IPCC-AR6 that this virtually impossible to predict how humans are going to mitigate – or fail to mitigate – climate change inducing emissions. Therefore, climate scientists in the IPCC-AR6 run models at different levels of potential greenhouse gas emission levels that correspond to different levels of warming. These “emission scenarios” allow policymakers to understand what the climate could look like if certain actions are or are not taken. Another human induced uncertainty includes geoengineering practices that could occur in the future like carbon capture or aerosol manipulation which would work to reduce the impact or amount of solar radiation received on Earth.
Another source of uncertainty in climate models is natural internal variability which is most problematic with near-term climate predictions. Problems with internal variability in models can be mitigated by running the same model multiple times under identical forcing conditions but with small changes to initial conditions to better understand the impact of normal climate variability.
With all this uncertainty, how do climate scientists know if the model works (predicts accurately)? For one, models can be used to predict historic climactic states and see if they line up with what we know from ice core records or written climate records. Additionally, climate models can be compared to real-time date to see if they are accurately representing the various cycles on the Earth.
A particularly interesting example of scientists comparing a model’s predictions to real-time data occurred in a case of ocean circulation and marine debris in 1992. In the face of an ocean storm, a shipping tankard dropped several shipping containers into the ocean between China and America, releasing thousands of bath toys including rubber duckies. These toys were obviously designed to float in water and offered an amazing opportunity for scientists to compare their models of ocean circulation to the real thing. As these toys floated along the ocean currents and showed up on various shores across the world, scientists could compare their models of surface water parcel movement to see if their models accurately predicted movement of currents! Model corroboration in action!
Regardless of potential uncertainties in our ability to accurately model changes in the climate under varying emissions and greenhouse gas concentration scenarios, there is high confidence that the latest models used in the IPCC-AR6 accurately capture most aspects of the observed climate change and does even better than the models used in the IPCC-AR5. Some studies have analyzed how previous climate models faired in the years after they were created (comparing the model output to what occurred in real life) and found reasonable agreement with the changes in climate over the short term. However, as policymakers and people. We are often concerned with what could happen in the next century as opposed to decade, and it is simply impossible to compare our model output to real data at that timescale. Therefore, models give us the best estimation of how our climate may change under various scenarios and help us understand how changes in emissions and mitigation behavior may impact climate change.
IV. Likelihood of model predictions
Climate models provide us with an upper and lower bound of likelihood for the given predictions simulated. These boundaries may be influenced by missing processes or fundamental lacks of knowledge in aspects of the climate system. While the likelihood of some scenarios may be relatively low, their impact on society and biota on Earth may be relatively high, in which case it is useful to predict these outcomes anyway. Especially given that it is difficult, nay impossible, to corroborate long-term climate model predictions with real data (because we are predicting too far into the future), it is important to be able to understand the worst case scenario in case circumstances arise that bring about these low likelihood but high impact scenarios.
While low-likelihood, high impact scenarios may not (to the best of our predictive knowledge) be likely to occur, many of the models do accurately represent certain aspect of our current climate, therefore they carry useful information. Additionally, these low likelihood scenarios are often those with the most extreme warming, and with extreme warming comes increased variability and uncertainty. For instance, some low-likelihood, high-impact scenarios predict increased extreme precipitation events but decreased mean precipitation locally. These changes in precipitation not only impact rainfall but also impact atmospheric humidity, evapotranspiration, soil moisture, and runoff – all complicated processes.
Figure 4.42 in the IPCC-AR6 report shows the differences between different liklihood scenarios and the changes can be drastic. Local upper estimates in annual mean precipitation increase by 70% in regions over Northern Africa, the equatorial Pacific, and the Arctic. Lower estimates instead show I decrease around 50% in annual mean precipitation around the equatorial Atlantic, central America, and some parts of southern Africa, South America, and Australia.
It may seem that with such differences in a key climatic output from a climate model, we may not know where to start in understanding what might happen in our future and what these models tell us. Simply put, the models do an accurate (and often very accurate) job in estimating a variety of potential climate futures based on potential trends in human emission behaviors. While some processes that are complicated to calculate are not incorporated into climate models, scientists work to corroborate the accuracy of climate models with historical or real time climate data as well as with other models. The process of constructing a model is built on elemental theories in physics for how our climate is dictated, so they work on known physical processes, albeit extremely complex and interactive. While uncertainties remain in our ability to accurately predict and incorporate every aspect of our climate into a climate model, the largest uncertainty may be how humanity will react to the knowledge of our changing climate; whether humanity will work to mitigate, incorporate geoengineering to reverse climate change, or will continue with business as usual and even emit more in the future are societies across the world become more developed.
In the words of statistician George E. P. Box, “all models are wrong, but some are useful”. Climate scientists have been working for decades to continue to refine our climate models, using complex multi-model schemes, running experiments, and constantly calibrating inputs so that they most closely match what we know about the climate today and in the past. It is humanity’s job, the role of governments and institutions, to take this painstaking information and do something about it.
Chapter 4 Sections 4.3 & 4.4
Jacqueline Shaff
Chapter four summarizes different projections of the physical indicators of global climate change. The climate models are produced by the Coupled Model Intercomparison Project (CMIP6), which is essentially the combination of many distinct climate models produced by modeling groups all over the world. The CMIP is a framework to keep all of the individual models done in a systematic way so the results can be comparable. The “coupled” part of the name means that all of the models include the physical processes of the atmosphere and ocean.
4.3 Projected Changes in Global Climate Indices in the 21st Century
This section summarizes the indicators of global climate change over the 21st century across Earth’s different climate systems. The major climate system components include the atmosphere, the cryosphere, the oceans, and the biosphere. These systems and their interactions with one another determine our daily weather and longer-term averages that make up our climate. The projected changes in global climate indices are modeled across several different “Shared Socioeconomic Pathways” (SSPs) that represent scenarios of socioeconomic global changes up to 2100. The indicators described here are used in Sections 4.4 through 4.6 to describe the climate models and their implications for climate policy.
4.3.1 Atmosphere
4.3.1.1 Surface Air Temperature
The atmosphere refers to the layer of gases and aerosols (the mixture of fine liquid or solid particles suspended in the air) that surround the planet. These gases and aerosols, particularly the ones that make up a small fraction of the atmospheric mass like water vapor, CO2, and ozone, determine the amount of shortwave radiation (from the sun) and longwave radiation (from Earth’s surface and atmosphere) via scattering, absorption, or transmission. Incoming solar radiation and outgoing radiation from Earth’s climate systems balance out in a stable climate. Ultimately these interactions are what heat our planet.
The temperature projections in this section specifically focus on the air temperature at Earth’s surface and are represented by global mean surface air temperatures (GSAT) as a measure of global temperature change. Overall, the GSAT is projected to rise throughout the 21st century if greenhouse gas emissions continue to increase. Simulations are displayed using the mean GSAT changes between both 1850-1900 and 1995-2014 and 2021-2040, 2041-2060, and 2081-2100 (Table 4.22). For all temperature differences regardless of time period or SSP, there is an increase in GSAT.
4.3.1.2 Precipitation
In addition to air temperatures, the moisture content of the atmosphere and the hydrologic cycle are vital pieces of our climate system. Water evaporates from the ocean or wet land surfaces and then condenses in the atmosphere to form precipitation. Warmer temperatures lead to more moisture in the air because warmer air can hold more water vapor. Water vapor content increases with a warmer climate, and when there is more moisture in the atmosphere, there can be more moisture convergence leading to more precipitation in areas where there is already rain (“the wet areas get wetter”). Specifically, tropical regions, monsoons, storm tracks, and high latitudes are expected to get rainier with warmer temperatures.
Precipitation changes are modeled as percent (%) changes relative to averages over 1995-2014 for the time periods of 2021-2040, 2041-2060, and 2081-2100 (Table 4.3). Precipitation changes are also specified as being over land or ocean; however, most of the modeling work focuses on precipitation over land since that serves greater relevance to society. All mean percent changes across time periods and SSPs are positive, indicating projected increases in precipitation. Although, some of the ranges do contain negative percent changes, which suggests there could be cases where there would be less precipitation averaged over Earth’s land area. Regardless, there are greater increases in precipitation over land compared to over ocean or even globally across all scenarios.
4.3.2 Cryosphere, Ocean, and Biosphere
4.3.2.1 Arctic Sea Ice
The cryosphere refers to the frozen places on Earth’s surface where water is in its solid form. Specifically, the cryosphere is composed of sea ice (frozen ocean water floating), continental ice sheets (glaciers that cover large amounts of land in Greenland and Antarctica), permafrost (frozen soil), mountain glaciers (permanent ice on top of mountains), and seasonal snow cover. The cryosphere is an important component of Earth’s climate system because it has a high albedo (reflectivity of solar radiation), low thermal conductivity (ability to conduct or transfer heat), and drives oceanic and atmospheric circulation. Ice sheets also store water, and changes in their volume can influence sea level.
In the Arctic, sea ice grows over winter when there is no sunlight and it is extremely cold and melts in the summer. The minimum ice coverage in the Arctic is around mid-September, which is why the models often compare the sea ice levels in September across periods of time. Arctic sea ice is typically only 0.5 to 3 m thick but can be as thick as 10-30 m in pressure ridges when sea ice bumps into other sea ice. Since sea ice is already on the surface of the ocean, it does not lead to sea level rise when it melts.
Arctic sea ice averages are measured in 106 km2 for the time periods of 2021-2040, 2041-2060, and 2081-2100 for the month of September and the month of March (Table 4.4). Under the SSP2-4.5, SSP3-7.0, and SSP5-8.5 scenarios, the Arctic will become essentially ice-free in September by 2100. Arctic sea ice levels are also modeled to decrease to a higher degree in September compared to March for all of the scenarios.
4.3.2.2 Global Mean Sea Level
The ocean is an important regulator of Earth’s climate and is responsible for storing and transporting a large amount of energy. Global ocean circulation is driven by wind and density (thermohaline circulation). Pacific Ocean heat transport gives you a sense of how much heat is transported by the wind-driven circulation. Atlantic Ocean heat transport is northward everywhere due to the Atlantic Meridional Overturning Circulation (AMOC) described in the next subsection. The interaction between the ocean and atmosphere systems also leads to exchanges of gases, such as carbon dioxide, making the ocean an important carbon sink.
Global mean sea level rise is measured in meters (m). The previous assessment report (AR5) projected that the rate of global mean sea level rise throughout the 21st century would exceed the rates observed between 1971-2010 mainly due to increases in the ocean temperatures and the loss of glaciers and ice sheets. This report came to similar conclusions that under any SSP, we are projected to see increases in global mean sea level rise in the 21st century.
4.3.2.3 Atlantic Meridional Overturning Circulation
The AMOC is a system of ocean currents that can be thought of as a conveyer belt driven by water’s density. Saltier and colder water is denser than fresher and warmer water. So, when warm water moves northwards, it will cool and some freshwater evaporates, making it saltier. This colder and saltier water sinks near Greenland because it is now denser and will begin to move southwards. As it moves southward, it will begin to move back up towards the surface as strong winds over the Southern ocean create upwellings of deep water to the surface and warm. Overall, the AMOC transports relatively warm waters northward and cold waters southward.
Relative to 1995-2014, the CMIP6 models suggest that the AMOC strength will decrease throughout the 21st century for each of the SSP scenarios (Figure 4.6). When the atmosphere warms, the surface ocean temperatures will not lose as much of their heat. Additionally, increases in precipitation and ice melting will make the ocean fresher in parts. Warmer and less salty North Atlantic surface waters will reduce the sinking as a result of the density gradients, weakening the AMOC. A decrease in AMOC strength under global warming will not change global heat transport significantly but could affect the delivery of heat to the high northern latitudes.
4.3.2.4 Ocean and Land Carbon Uptake
The last part of the climate system discussed is the biosphere. The biosphere refers to all regions of the previously described systems that are occupied by living organisms. The marine and terrestrial biospheres interact with the atmosphere when it comes to carbon uptake. Living organisms are responsible for taking in and releasing greenhouse gases. For example, marine and terrestrial plants take in and store carbon dioxide through the process of photosynthesis. And you, reader, are one of the many organisms on this planet that release carbon dioxide through the process of respiration. So, both the ocean and land are responsible for carbon uptake from the atmosphere. The cumulative uptake of carbon by both the ocean and land is modeled to increase throughout the 21st century across all emission scenarios. Ocean carbon flux is an important component of mitigating the rise of atmospheric CO2 and is responsible for many changes in the ocean biosphere, like ocean acidification that will be discussed in the next subsection.
4.3.2.5 Surface Ocean pH
When the ocean absorbs carbon dioxide from the atmosphere, the ocean pH decreases. This is because more carbon dioxide increases the number of hydrogen ions in the ocean. The models predict that for all scenarios except SSP1-1.9 and SSP1-2.6, the ocean surface pH will decrease throughout the 21st century (Figure 4.8). With more carbon dioxide emissions on land, there will be an increase in carbon storage by the ocean and more ocean acidification in the future.
4.3.3 Modes of Variability
Climate variations can result from external sources, like those described above, or from internal interactions from components that are a part of the climate system themselves. There are natural climate variations from internal components varying, and this section specifically focuses on annular modes and El Niño Southern Oscillation (ENSO).
4.3.3.1 Northern and Southern Annular Modes
Annular modes refer to patterns of climate variability on the hemisphere scale that results from atmospheric dynamics. Annular modes are the important source of climate variability in middle and high latitudes in the Northern and Southern Hemispheres and can be responsible for anomalies in surface temperatures and precipitation.
The Northern Annular Mode (NAM) in the Northern Hemisphere has an index in this chapter defined as the difference in mean sea level pressure (SLP) (atmospheric pressure at mean sea level) between 35°N and 65°N. Only when modeled with the highest emission scenarios (SSP3-7.0 and SSP5-8.5) does the NAM show a change by the end of the 21st century. Those models project that the boreal wintertime (wintertime in the Northern Hemisphere) surface NAM will be more positive by 2100. Additionally, there is little evidence that anthropogenic forcings are driving the recent variations across decades of the NAM.
The Southern Annular Mode (SAM) in the Southern Hemisphere has an index in this chapter defined as the difference in mean SLP between 40°S and 65°S. Under the highest emission scenarios (SSP3-7.0 and SSP5-8.5), the SAM in the austral summer (summer in the Southern Hemisphere) becomes more positive by 2100. For the lowest emission scenarios (SSP1-1.9 and SSP1-2.6), the SAM is projected to be less positive due to recovery of the ozone hole.
4.3.3.2 El Niño-Southern Oscillation
ENSO results from the ocean and atmosphere interactions in the tropical Pacific and explains most of the variance in patterns of climate variability in the tropics. During ENSO, there is warmer water and more precipitation over the central Pacific. There will very likely be intensified amplitudes of ENSO rainfall variability throughout the 21st century with increasing emissions, however, there was not a model consensus on changes in the amplitude of SST variability.
4.4 Near-term Global Climate Changes
This section summarizes near-term climate change projections between 2021-2040 compared to 1995-2014 using the global climate indicators from Section 4.3.
4.4.1 Atmosphere
4.4.1.1 Average Global Surface Air Temperature
When averaged across the near-term (2021-2040), GSAT is very likely to be higher than the average between 1995-2014 by between 0.4°C–1.0°C (Table 4.5).
4.4.1.2 Spatial Patterns of Surface Warming
The model projections also specify where the surface temperatures will increase the most spatially. For the SSP1-2.6 and SSP3-7.0 projections (Figure 4.12), the largest warming occurs in high altitudes and overall more warming over land than over the ocean. Specifically, for both projections, the seasonal mean surface temperatures on land in the Northern Hemisphere are projected to exceed 1°C increase relative to 1995–2014. In the near-term, the northern North Atlantic, parts of India, parts of North America, Europe and Asia in winter, and the Southern Hemisphere subtropical eastern Pacific under CMIP6 models do not show robust warming.
4.4.1.3 Precipitation
Precipitation changes can be separated into two components: changes in intensity (“wet areas get wetter”) and changes in position (where the precipitation is). In regards to the intensity, precipitation is projected to increase at high latitudes, over ocean regions, and in wet regions in the tropics, while decreasing in already dry regions, such as the subtropics (Figure 4.13).
4.4.2 Cryosphere, Ocean, and Biosphere
4.4.2.1 Arctic Sea Ice
In September in the near-term, Arctic SIA is projected to decrease (Figure 4.15). Most (79%) of the simulations across just ten years for all SSPs project that there will be decreasing Arctic sea ice in September; and that rises across 30 years to almost all (98%) of the models.
4.4.2.2 Ocean and Land Carbon Flux
Under the higher emission scenarios (SSP2-4.5, SSP3-7.0, and SSP5-8.5), ocean carbon exchange will increase (Figure 4.16). These trends are much more clear than for land carbon exchanges, where it is unlikely that there will be a detectable increase in land carbon flux in the near-term.
4.4.3 Modes of Variability
4.4.3.1 Northern and Southern Annular Modes
In the near-term, there are slightly more positive NAM for higher emission scenarios (SSP2-4.5, SSP3-7.0, and SSP5-8.5) during boreal fall, winter, and spring (Figure 4.17a). However, the magnitude in those changes caused by increases in GHG emissions is about the same, if not a little less, than if the changes were caused by just internal variability in the NAM. There is also a projection of a more positive SAM, this time in the austral winter (Figure 4.17b). Again, in the near-term there is also some internal variability expected.
4.4.3.2 El Niño-Southern Oscillation
In the near-term, there are no robust changes in ENSO sea surface temperature variability or rainfall variability associated with ENSO. However, the amplitude of rainfall variability may be likely to increase because of an increase in moisture. As previously discussed with other precipitation changes, warmer temperatures mean more moisture can exist in the air, allowing for more rainfall in that area.
4.4.4 Response to Short-Lived Climate Forcers and Volcanic Eruptions
Short-lived climate forces are substances such as methane, ozone, and aerosols that have a shorter lifetime in the atmosphere compared to carbon dioxide. Short-lived climate forcers are important to understand because mitigation efforts will influence future climate projections that we have been looking at. And because of their shorter lifetimes, these SLCF can heavily influence near-term projections.
Mitigation efforts of SLCFs can influence other aspects of the global climate system. Reduction in aerosol emissions in the near-term will lead to warmer temperatures. Aerosols can absorb or reflect sunlight and incoming solar radiation, and right now reflection is currently larger. So when aerosols decrease, temperatures will actually rise at first as more incoming solar radiation reaches Earth’s surface.
Another phenomenon that can influence near-term projections would be volcanic eruptions. Volcanic eruptions can have large cooling effects by the release of dust or aerosols that will block incoming solar radiation from reaching Earth’s surface. This cooling effect can last for anywhere from months to years depending on the eruption. If there was another large eruption similar to Mount Pinatubo in the Philippines in June 1991, there would likely be cooling in the Northern Hemisphere with peak amplitude between 0.09°C and 0.38°C, lasting for three to five years.
Mid- and Long-Term Climate Change
Gabriela Carr
Cross-Chapter Box 4.1: The Impact of Volcanic Eruptions on Climate
(Figure 1)
Models in this report do not include volcanic eruptions because they are simply too unpredictable. However, that does not mean that they will not occur in the 21st century, and they can have a big climate impact. Just like anthropogenic emissions, eruptions can be measured in terms of effective radiative forcing (ERF), i.e., the change they cause in the Earth’s energy balance (measured in watts/meter2, or W/m2.) Prior to 1900 and broad-scale industrialization, volcanic eruptions were the largest source of ERF in the world. Today, CO2 emissions from volcanic eruptions are less than 1% the magnitude of emissions from human activity.
How often do major eruptions occur?
Eruptions with an ERF stronger than -1 W/m2 have happened twice a century, on average, since about 500 BC. For reference, the eruption of Mount Pinatubo in 1991 fits into this category.
How does the climate respond to major eruptions?
That depends. Is the volcano in the tropics? If so, volcanic aerosols, i.e., air-borne particles, can spread in the stratosphere, i.e., the atmospheric layer above most clouds, of both the northern and southern hemisphere. This broadens the eruption’s impact. Is the volcano outside the tropics? Then aerosols will spread in the stratosphere of the hemisphere in which the volcano is located, narrowing impact. The height of the eruption, the amount of sulfur (which reflects radiation, and is therefore cooling) released, and the time of year all determine the total mass released, as well as both the time the aerosols remain in the air and how those aerosols impact radiation.
There are still some generalizable impacts, however. A single eruption can cause a decrease in global surface air temperature (GSAT) for multiple years, while a cluster of eruptions can stretch that impact to decades or even a century. Average land precipitation decreases, including monsoon rains. Dry regions get wetter and wet regions drier for up to a few years after the eruption. In the Arctic, sea ice increases with cooler temperatures for up to decades. The cooling feedback between sea ice and the ocean furthers cooling, even after volcanic aerosols are gone, while the ocean also stores and feeds back cool temperatures into the atmosphere.
Is there likely to be a major eruption in the 21st Century?
There will likely be at least one major eruption before 2100. A cluster of large eruptions, an unlikely possibility, could dramatically alter this century’s climate trajectory.
4.5: Mid-Term (2041-2060) to Long-Term (2081-2100) Global Climate Change
I. 4.5.1: Atmosphere
a. 4.5.1.1: Global Surface Air Temperature (GSAT)
(Figure 4.19)
On both mid-term and long-term scales, air temperature near Earth’s surface, and with it land and ocean temperatures, will warm on average across the globe, regardless of which Shared Socioeconomic Pathway (SSP) you choose. SSPs group climate trajectories by potential policy choices, depending on the global push for climate mitigation and/or adaption, and range from SSP1 to SSP5. They are usually coupled with different effective radiative forcing (ERF) levels in 2100 based on varying projected emissions; ERF ranges from 1.9 to 8.5 in this chapter. Unless otherwise specified, trends described throughout this paper should be assumed to intensify with higher emission SSP/ERF combinations, in the following order from low to high emission scenarios: SSP1-1.9, SSP1-2.6, SSP2-4.5, SSP3-7.0, SSP5-8.5.
i. Land-Ocean Warming Contrast
On average, warming over land will be higher than warming over the ocean; in fact, this has already been happening since 1900. The difference has been particularly dramatic for dry subtropical continents, while the difference between land and ocean has been smaller in the wetter tropics and mid-latitudes. The land-ocean warming contrast happens because air over the ocean is wetter than air over land. As air over the land warms this will increase the contrast, because the air over land will become increasingly dry and therefore warm more easily.
ii. Polar Amplification
The Arctic will experience more warming than the global average; depending on which SSP you choose, the projections range from 3° Celsius to 12° Celsius of warming by the end of the century. This polar amplification occurs for many reasons. For example, as the region warms, snow melts, decreasing the albedo, or reflective capabilities, of the Arctic. Unable to reflect as much radiation back out from the surface, the region warms further. As temperature warms, the lapse rate, or rate that temperature decreases with height, will increase in the Arctic. This makes it more difficult for radiation from the surface of the Arctic to make it out to space, furthering warming at the surface. The ocean will increasingly bring warm temperatures to the Arctic as well. These are only a few of the reasons the Arctic will warm; there is great uncertainty in how much warming will occur, depending on how much weight a model places on each of these warming mechanisms. The models agree, however, that the Antarctic will experience less warming than the Arctic. One reason for this difference is that upwelling in the Southern Ocean uptakes heat from the surface air, bringing that heat north and away from the Antarctic.
iii. Seasonal Warming Patterns
(Figure 4.20)
Poleward of 55° North or South, winters will warm more than summers, reducing the difference between seasons. In most regions between 30° and 55° North or South, i.e., the subtropics and mid-latitudes, summers will warm more than winters, increasing the difference between seasons.
iv. Changes in GSAT Variability
(Figure 4.21)
As changes in GSAT variability are not well understood, both because of existing variability and variability under ERF scenarios, I will not go over them in detail here. However, there is some confidence that GSAT variability will decrease on a year-to-year timescale in the mid-latitudes and poles during the winter months.
b. 4.5.1.2: Annual Mean Atmospheric Temperature
(Figure 4.22)
In the coming century, the largest temperature change in the troposphere, the atmospheric layer closest to the Earth’s surface and the one which contains most clouds, will be in the high latitudes of the Northern Hemisphere. In the tropics, warming will likely be larger in the upper troposphere than at the surface. Throughout all latitudes, the troposphere will warm; however, depending on which ERF scenario you choose, the stratosphere above will either warm or cool. A major determinant of stratospheric temperature will be ozone hole recovery; as the hole closes, the lower stratosphere of the Antarctic will warm. The Arctic will very likely have stronger warming in the troposphere than the global average. If a higher ERF is chosen, the average stratospheric cooling across the globe will increase by the end of the century.
c. 4.5.1.3: Near-Surface Relative Humidity (RH)
(Figure 4.23)
There will likely be a decrease in near-surface relative humidity (the amount of water vapor in the air divided by the maximum amount of water vapor the air can hold before condensation) over many land areas, particularly in the tropical and sub-tropical latitudes. There will be a stronger decrease in the summer than winter months and in higher than lower emissions scenarios. Both changes in transport of ocean moisture to land and changes in evapotranspiration contribute. As temperatures and CO2 levels rise, soils dry and the stomata, the respiratory openings on plant leaves, become less open to retain moisture. This prevents moisture from entering the atmosphere either from soil or leaves themselves. Decreases in near-surface relative humidity will be smaller over the ocean than over land.
d. 4.5.1.4: Precipitation
(Figure 2.4)
On land, higher latitudes will experience more rain for two reasons: the warming troposphere will have an increased specific humidity (the ratio of water vapor to total air), and, under a high emissions scenario, more water vapor will travel pole-ward from the tropics. In the mid-latitudes and subtropics, dry and semi-dry land areas will get drier (although there are discrepancies depending on the model used), while wetter mid-latitudes will get wetter under a high-emissions scenario. Per degree Celsius of GSAT warming, average precipitation globally will increase by 1-3%. This increase is primarily driven by increases in water vapor, and secondarily by warming. The impact of sea surface temperature feedbacks to precipitation is uncertain in the long-term. Overall, variability in precipitation will increase over land areas due to warming. As for the ocean, precipitation will likely decrease over subtropical regions and increase over many monsoon regions.
e. 4.5.1.5: Global Monsoon Precipitation and Circulation
It was already predicted in the previous report that in the long-term, monsoons will have increased area, precipitation, and intensity, but a weaker circulation. The increase in precipitation is due to an increase in atmospheric moisture. The increase in monsoon precipitation will be greater in the Northern Hemisphere than the Southern Hemisphere for three reasons: Hadley circulation (the circulation of water vapor away from 30° North or South to the wet tropics), the temperature difference between the hemispheres, and moisture in the atmosphere will all increase. There will be a greater difference between dry and wet seasons in monsoon areas, as well as a greater variability between monsoon seasons.
f. 4.5.1.6: Sea Level Pressure, Large-Scale Atmospheric Circulation, Storm Tracks and Blocking
i. Sea Level Pressure
(Figure 4.25)
On average, atmospheric pressure at sea level will increase in the mid-latitudes while decreasing in the high latitudes. In the Northern Hemisphere winter, sea level will decrease in northeastern Asia and North America. In the Northern Hemisphere summer, sea level pressure will decrease in the Mediterranean and Middle East due to regional warming.
ii. Zonal Wind and Westerly Jets
(Figure 4.26)
As the atmosphere’s stratification, the upper troposphere’s temperature gradient, sea surface temperature, the land-sea warming contrast, and loss of sea ice all increase, jet streams will likely shift towards the poles. Changes to the stratospheric polar vortex will also contribute to this shift. Clouds contribute the most uncertainty to these predictions. The highest confidence for these projections is in the Northern Hemisphere, where autumn and summer low-level zonal-mean westerly jet streams will likely shift towards the poles.
iii. Storm Tracks
(Figure 4.27)
As Hadley cells expand and the origins of extratropical cyclones (ETCs) shift towards the poles, the number of ETCs will slightly decrease. However, ETCs with heavy precipitation will increase with warming, as the warmer atmosphere can hold more moisture. Individual ETCs will spread poleward due to a stronger jet stream in the upper troposphere as well as stronger precipitation. In Northern Hemisphere winter, strong ETCs are somewhat likely to decrease except in the northern North Pacific. In the southern hemisphere, there will be an increase in strong ETCs, with high wind speeds in high emissions scenarios. Southern Hemisphere ETCs will shift towards Antarctica due to a poleward shift of the jet stream as well as slower warming in the Southern Ocean.
iv. Atmospheric Blocking
(Figure 4.28)
High pressure weather systems, both in the mid- and high-latitudes, can block the flow of westerly winds; this blocking causes extreme weather, intensifying seasonal heating or cooling. While there is a lot of uncertainty surrounding the impact of emissions on such blocking, under SSP3-7.0 and SSP5-8.5 such blocking events will likely decrease in Greenland and the North Pacific during Northern Hemisphere winter.
II. 4.5.2: Ocean
a. 4.5.2.2: Ocean Acidification
(Figure 4.29)
Surface ocean acidification will continue in the mid- and long-term and will travel to deeper parts of the ocean. Acidification will be greater in the high latitudes, and particularly in the Arctic. However, uncertainty in projections increases with depth.
III. 4.5.3: Modes of Variability
a. 4.5.3.2: El Nino-Southern Oscillation (ENSO)
The previous report already determined that ENSO will continue to be the main source of climate variability within a given year. However, there is no single prediction of how ENSO sea surface temperature will change in the coming century, or where ENSO events will shift in location. It is very likely that rainfall related to ENSO events will increase in variability, and likely that North Pacific and North American impacts of ENSO events will shift east.
4.6 Implications of Climate Policy
Yian Lin
4.6.1 Patterns of Climate Change for Specific Levels of Global
Section 4.6.1 presents the estimated spatial patterns of change in global temperature, precipitation, and atmospheric circulation when global warming of 1.5°C, 2°C, 3°C, and 4°C relative to the period 1850–1900 happens. I will focus on the estimated change in global temperature and precipitation in this report.
Global warming of a specific degree means an increase, relative to 1850-1900, of that degree in global mean surface temperature (GMST) averaged over a 30-year period. For example, when GMST averaged over a 30-year period increases by 1.5°C, we say that the global warming of 1.5°C is reached.
For those who wonder what “spatial patterns of change” means and or why we may want to assess it, note that the effects of global warming are not uniform across the globe. For example, some regions may warm more than others, and the change in precipitation may also vary across different places. It may help to take a quick look at Figure 4.31(a) of AR6 which shows estimated change in global temperature under the 1.5°C of global warming. There is no need to pay too much attention to details and the caption yet. What we can see in Figure 4.31(a) is a global map with colors on it, and different regions on the map may have different colors. This shows that different regions may experience quite different temperature changes. Now it becomes clearer why we are interested in “spatial patterns of change,” and we are ready to dig a little deeper what the estimated spatial patterns of change in global temperature and precipitation are when we reach 1.5°C, 2°C, 3°C, and 4°C of global warming.
For global temperature (section 4.6.1.1), the simulation results show that global warming of 1.5°C, 2°C, 3°C, and 4°C all implies an increase in mean temperature in almost all regions of the world when compared to 1850-1900 (Figure 4.31 (a) – (d)). The higher the levels of global warming, in general the larger the increase in mean temperature is (Figure 4.31 (e) – (g)). Also, Figure 4.31 shows that, no matter for which of the four levels of global warming, there are at least three general spatial patterns of the increases in the annual average near-surface temperature. Firstly, warming over land is generally larger than warming over ocean areas. Secondly, warming in high latitudes, especially in the Arctic, is larger compared to low latitudes. Thirdly, the Northern Hemisphere warms more than the Southern Hemisphere.
Now I will present some possible explanations for the three general spatial patterns. The first pattern that land warms more than ocean can be explained majorly by two reasons. The first is that water has a larger heat capacity than land, which means that the same amount of heat can raise land’s temperature more than water’s. The second reason is that oceans have unlimited amounts of water to evaporate, and evaporation leads to cooling of the oceans. In contrast, the land surface may dry out quickly, so cooling by evaporation is constrained. Another way to understand this is through energy balance at the Earth’s surface–the amount of energy absorbed by the Earth’s surface must be equal to the amount of energy going out from the surface. Given that greenhouse gases trap a large amount of the heat emitted by the Earth’s surface through longwave radiation and radiate back heat to the Earth’s surface, energy fluxes into the Earth’s surface increase. To reach energy balance, the earth must increase the energy fluxes out from it. Evapotranspiration is one way to go. Because land surfaces cannot use as much evapotranspiration as oceans do to lose heat, they must increase their temperature to increase energy lost to the atmosphere through longwave radiation or dry heat exchange. Once we know the explanations for the first general pattern, the third general pattern is easy to understand. The first pattern together with the fact that there is more land area in the Northern Hemisphere than in the Southern Hemisphere implies the third pattern that the Northern Hemisphere warms more than the Southern Hemisphere.
For the second general pattern that the high latitudes warms more than the low latitudes, I will focus on the so-called Arctic amplification–the phenomenon that the high latitudes of the Northern Hemisphere are estimated to warm the most when compared to other parts of the earth given any of the four levels of global warming. One reason for Arctic amplification is the ice-albedo feedback that operates on the state variable of temperature. Simply put, global warming leads to ice melting in the Arctic. The declines in ice will reduce the reflectivity (albedo) of the Arctic surface, which means that more solar radiation will be absorbed by the Arctic surface. As a result, the Arctic will be warmed up even more. This loop of cause and effect is known as the Ice-albedo feedback, and it is a positive feedback because it amplifies the initial increase in temperature. In other words, this feedback makes the increase in temperature more than it would have been without the feedback. Another even more important reason is the lapse rate feedback, described above in section 4.5.1.1ii.
For precipitation (section 4.6.1.2), global mean precipitation increases as global surface average temperature (GSAT) rises. Even more than what we have seen for global temperature, the effects of global warming on precipitation is not uniform across the globe (Figure 4.32 of AR6). Quite a few regions are even projected to experience a decrease in precipitation. Higher levels of global warming have stronger effects, no matter whether the effects are estimated increase or decrease in precipitation. In addition, Figure 4.32 shows that there are some general spatial patterns of the change in precipitation consistent across different levels of global warming. Firstly, the high latitudes in both the Northern and Southern Hemispheres, tropical regions, and large parts of the monsoon region are projected to have more precipitation. Secondly, precipitation is projected to decrease over the subtropical regions.
Before presenting the possible explanations for the two general spatial patterns, I would first provide the background knowledge on basic physical processes that explain how the global mean precipitation is related to GSAT. First of all, it is important to note that warmer air can have much larger water-holding capacity. The maximum amount of water vapor that can exist in air before condensation occurs (saturation vapor pressure) is an exponential function of air temperature. Given that relative humidity — water vapor in the air (vapor pressure) divided by total amount of water vapor the air can hold (saturation vapor pressure) — is expected to stay roughly the same with global warming, the amount of water vapor in the air (water vapor content) rises approximately at a rate of 7%/K as air temperature rises. Now let’s think about what is required for rain. Rain is a result of condensation that happens when moist air cools, and this kind of cooling almost always happens with rising motion. Therefore, both water vapor and rising motion are needed to get rain, and it is impossible to have rain everywhere even when there is more water vapor in the air. However, given the same amount of rising motion, more water vapor in the air does imply more precipitation. This is why we expect precipitation in the wettest regions to increase under global warming. The fact that warm air holds more water vapor is one of the reasons for the first general pattern that high latitudes, tropical regions, and monsoon regions are expected to get rainier. This is referred to as a thermodynamic change. However, it is also worth noting that the confidence in how the precipitation may change in high latitudes is higher when compared to the confidence in the tropics. One reason is that tropical precipitation shifts towards warmth, and it even shifts in response to warming very far away.
Although the global mean precipitation is expected to increase, some regions such as the subtropical regions are projected to experience less precipitation. This is because some other regions will experience much more rainfall while the global mean precip only increases a little. Then you may wonder why the dry regions will be in the subtropics. One of the reasons is because more moisture is projected to be fluxed away from there by the Hadley cell and eddies. Hadley cell takes moisture from the subtropics into the equatorial regions, and eddies carry moisture from the subtropics into higher latitudes.
For those who are interested in how the spatial patterns of change presented in Figure 4.31 and Figure 4.32 are constructed for a given temperature threshold, some information is provided in section 4.6.1 before going into the subsection 4.6.1.1. Model simulations were done for the Tier-1 SSPs of CMIP6 (SSP1-2.6, SSP2-4.5, SSP3-7.0 and SSP5-8.5). For each model simulation considered under each of the SSPs, scientists first computed 20-year moving averages of the global average surface air temperature (GSAT). Then from the resulting time series, the scientists found the first years in which the increase in the simulated GSAT, relative to the 1850-1900, exceeds the 1.5°C, 2°C, 3°C and 4°C thresholds. For each of those first years detected, a 20-year global climatology was subsequently constructed using those first years as the centers. As a result, for each SSP under each of the four levels of global warming, there are some 20-year global climatologies constructed. And the spatial patterns of change for any given level of global warming were computed based on the composite of all the 20-year global climatologies constructed under the four SSPs.
4.6.2 Climate Goals, Overshoot, and Path-Dependence
Section 4.6.2.1 assesses the impacts and challenges of overshoot trajectories when compared to pathways with no or limited overshoot. Generally speaking, warming stops when emissions cease. Overshoot happens when the global mean surface temperature (GMST) temporarily exceeds the pre-specified cap of the level of global warming, such as 1.5°C. Assuming the expected result that warming ceases when emissions stop, to make the GMST decline after an overshoot requires net negative CO2 emissions, in a scenario in which CO2 removed exceeds CO2 emitted. It is important to note that making CO2 emissions zero will not achieve net negative CO2 emissions. A large amount of anthropogenic removal of CO2 is also needed. Although the GMST will drop from the level that it reaches during the overshoot and will recover to lower temperatures, the effects of overshoot on key climate indices such as global thermosteric sea level may not reverse. Also, the longer the duration and the higher the peak temperature of the overshoot may lead to greater impacts. This is why we want to compare the overshoot trajectories with the pathways that have no or limited overshoot.
Section 4.6.2.1 uses the overshoot scenario SSP5-3.4-OS to demonstrate the possible impacts on different climate indices. The number 3.4 in “SSP5-3.4-OS” means the radiative forcing around year 2100 is 3.4 W/m2. Radiative forcing is a change in Earth’s energy balance due to a particular factor such as greenhouse gases. A positive radiative forcing means the factor leads to warming of the earth. In SSP5-3.4-OS, CO2 peaks at 571 ppm in the year 2062 and drops to 496 ppm by 2100 as we can see from Figure 1 in this report. Although the CO2 concentration level drops back down, changes in climate indices have not fully reversed by 2100 as Figure 4.34 of AR6 shows. The average atmospheric CO2 concentrations (ppm) for the two 20-year periods, 2034-2053 and 2081–2100 (shaded grey bar), are the same. However, the mean GSAT, mean global land precipitation, and mean global thermosteric sea-level over 2081-2100 are higher compared to the same mean quantities computed for 2034-2053. And the mean Arctic September sea ice area for 2081–2100 is smaller than the same quantity computed for 2034-2053.
4.6.3 Climate Response to Mitigation, Carbon Dioxide Removal, and Solar Radiation Modification
Section 4.6.3 is about climate response to mitigation and solar radiation modification (SRM). Here I will provide some background and complementary information to help have a more comprehensive understanding of approaches that have been proposed to mitigate global warming. First, mitigation means “a human intervention to reduce emissions or enhance the sinks of greenhouse gases” (SR 1.5 Glossary). A sink is “a reservoir (natural or human, in soil, ocean, and plants) where a greenhouse gas, an aerosol or a precursor of a greenhouse gas is stored” (SR 1.5 Glossary). Mitigation includes carbon dioxide removal (CDR) options, such as planting new trees (afforestation), that remove CO2 from the atmosphere. Solar radiation modification (SRM) involves blocking the solar radiation that gets to the Earth’s surface. Examples of SRM methods include stratospheric aerosol injection (SAI), marine cloud brightening (MCB), and surface albedo enhancements. CDR and SRM are referred to as geoengineering or climate engineering.
Section 4.6.3.1 assessed the climate response to mitigation by comparing high- and low-emission scenarios. One key takeaway is that variability due to natural internal processes within the climate system (internal variability) may mask the effects of reduced emissions especially during the near term. For example, when the simulations show that the surface air temperature rises less in 2021-2040 in the low emission scenario than in the high emissions scenario, it is hard to know whether the projected difference is due to the difference in levels of emission or the internal variability. On the other hand, when the simulations show that the surface air temperature rises similarly in the two emission scenarios, it is not necessary that the reduced emissions have no effects. It is possible that the effects are marked by the internal variability. The internal variability dominates the changes in climate quantities during the near term. As a result, the benefits of mitigation for climate quantities will emerge only later during the 21st century.
Section 4.6.3.2 assessed the climate response to mitigation through CDR. The good news is that GAST should decrease in proportion to the total amount of CO2 removed from the atmosphere, because the relationship between temperature change and cumulative carbon emissions is nearly linear. However, the bad news is that the possible amount of CO2 removed by CDR is unlikely to reach the target level that is required to limit climate warming to 1.5° C or 2° C. The reality is that current CO2 emissions are very large (about 40 billion tons CO2/year). To just offset the current emissions, we need to filter a volume of air approximately equal to the Grand Canyon every second partly because the absolute concentration of CO2 in air is very small (about 0.04%). Another main takeaway from this section is that even if CDR can be implemented at ideal scale, there will still be substantial lags in the changes in key climate variables (Figure 4.37).
For section 4.6.3.3, it is important to note that SRM is different from mitigation methods in a way that it aims to alter the Earth’s energy budget instead of dealing with the greenhouse gases. SRM was first proposed in the 1970s. A detailed study on SRM options was issued by The National Academy of Sciences in 1992. One reason why people are considering SRM is that the progress in mitigation efforts seems slow and very expensive. However, SRM methods that dim the skies will not be able to perfectly cancel the effects of CO2 because solar radiation and greenhouse gases have different effects on the climate. For example, greenhouse gases warms nights more, but SRM cool days more. Reduction in solar radiation will not mitigate ocean acidification caused by the increase in CO2 levels, so large effects on marine life and the ecosystem will not be prevented by SRM methods. In addition, dimming the skies will have negative side effects. For example, plants will have less sunlight to grow. The tropics could be dried and summer monsoons could be strengthened by some SRM methods. Furthermore, once we begin to use SRM methods, we will need to do it for a very long time. A stop is projected to warm the Earth very rapidly (i.e. 2-4° C warming within 10 years).
4.7: Beyond Long-Term (2100-2300) Global Climate Change
Gabriela Carr
I. 4.7.1: Commitment and Climate Change Beyond 2100
a. 4.7.1.1: Climate Change Following Zero Emissions Commitment (ZEC)
(Figure 4.39, Table 4.8)
CO2 in the atmosphere will decrease for decades after emissions have been stopped, particularly assuming that natural carbon sinks continue to function as normal. ZEC is measured in terms of GSAT following cessation of carbon emissions, and the models differ on how GSAT will respond to the emissions decrease. Some say that temperatures will decline, some that temperatures will remain steady, and others that temperatures will continue to rise. These differences depend on the impact of carbon sinks and the ocean in uptaking emissions that remain in the atmosphere. It is likely that, regardless of whether temperature will increase, decrease, or remain the same, the magnitude of change will be small compared with GSAT’s natural variability.
b. 4.7.1.2: Change in Global Climate Indices Beyond 2100
(Figure 4.40)
i. 4.7.1.2.1: Global Surface Air Temperature (GSAT)
(Table 4.9)
Beyond 2100, GSAT depends strongly on which emissions scenario you choose, and the difference in GSAT between high and low emissions scenarios will get larger with time. In the low emission scenario of SSP1-1.9, GSAT decreases to 0.9° C by 2300. SSP1-2.6, GSAT stabilizes until 2100 and then decreases to 1.5° C by 2300. For the higher emissions scenarios of SSP2-4.5, SSP3-7.0, and SSP5-8.5, GSAT will be greater than 2° Celsius by 2300. A GSAT greater than 2° Celsius last occurred 3 million years ago. Climate models project temperature changes of up to 17o C global warming in the (quite extreme) extended SSP5-8.5 scenario.
ii. 4.7.1.2.2: Global Land Precipitation
While we continue to produce emissions, global precipitation will increase with increasing temperature; if we reduce emissions, precipitation will increase more quickly than temperature. This heightened precipitation with a decrease in emissions occurs because of ocean heat as well as the atmosphere’s quick adaptation to reduced ERF. Under SSP5-8.5 precipitation increases through 2300, while under the lower emissions scenarios of SSP1-2.6 and SSP5-3.4 precipitation stabilizes.
iii. 4.7.1.2.3: Arctic Sea Ice
Under the low emissions scenarios SSP1-2.6, Arctic sea ice expands through 2300. However, under the high emissions scenario SSP5-8.5, the Arctic becomes free of ice in September, or even for almost the entire year.
II. 4.7.2: Potential for Abrupt and Irreversible Climate Change
(Table 4.10)
For a change to be considered abrupt and irreversible, it must be large-scale, take place at most over a few decades, continue to cause impacts to both human and natural systems for multiple decades, and require a long recovery time. The threshold for such a change is known as the tipping point, and the system experiencing it is known as the tipping element. Table 4.10 lists fifteen suspected tipping elements, such as ocean acidification, permafrost carbon, and boreal forest cover. The table updates whether these tipping elements are considered able to experience tipping points according to the AR6 models, their irreversibility, their projected trajectories under continued warming, and any change in their categorization since the previous report.