There are significant uncertainties in predicting future climates as a consequence of (a) natural climate variability; (b) the potential for uncertain or unrecognized climatic forcing factors (e.g., explosive volcanism, new or unknown anthropogenic influences, etc.); and (c) inadequate understanding of the climate system (e.g., a recent discovery that clouds absorb significant short-wave radiation which is not included in models, or potential errors in fluxes at the top of the atmosphere associated with changes in relative humidity). These uncertainties may (or may not) lead to re-assessment of model capability. We must expect that new observations or results from process studies may yield information that causes us to re-evaluate and improve the capability of climate models. Our estimates of the credibility of climate system models can, of necessity, be consistent only with known facts and based only on the "best" current knowledge.
The list given below is not a "forecast" but rather a scientific judgment based on the assumption that the concentrations of anthropogenic greenhouse gases will increase. Many of the model experiments on which these points are based have an even more specific assumption, a scenario of a 1% per year increase in CO2 to mimic the radiative forcing of the projected increase in the concentrations of all greenhouse gases.
(1) Large stratospheric cooling will result from the increase in CO2 concentration and ozone depletion; the start of such cooling has been predicted by models and observed in the upper stratosphere.
Basis-Increased emission of infrared radiation is an automatic consequence of the increased infrared optical depth in the stratosphere due to the increased concentration of CO2. This will lead to cooling in the upper stratosphere, and satellite observations and other records show this is occurring (figure 5). In the lower stratosphere, temperature changes due to increased CO2 concentrations are complicated by the changes in ozone concentration due to volcanic aerosols, chlorofluorocarbons, nitrogen oxides from aircraft, and changes in tropospheric chemistry (further affected by other surface emissions), which can variously lead to cooling and warming in the lower stratosphere. Volcanic aerosols have a direct radiative effect on the lower stratosphere, producing a warming.
Basis-This estimate of global warming is based on projections of emissions and on a combination of results from simple model studies, general circulation models, the observed record, and estimates of the response of the climate to various forcings in the geological past, based on reconstructions carried out in paleoclimate studies. The estimated warming would be reduced if sulfur emissions are not controlled, and the actual temperature change may be outside this range if natural climate variations (e.g., significant changes in the frequency of explosive volcanic eruptions) are large. (3) Global mean precipitation will increase. The distribution of this change is less certain (see below).
Basis-A warming of the surface temperature over the globe will lead to an increase in global-mean precipitation because of the relationship between evaporation rate and surface temperature. The underlying physics on this is well-established, and all models confirm this relationship. 
(4) Northern hemisphere sea ice will be reduced (the magnitude of the change will depend on the amount of the warming, and the reduced extent will initially be most evident in the transition seasons). Projected changes and their timing in the Southern Hemisphere sea-ice extent are less certain.
Basis-Studies of past climates provide evidence for polar amplification of warming and reduced sea-ice extent. Modeling studies also suggest that this will occur, although some have suggested local sea-ice expansions.
(5) Arctic land areas will experience wintertime warming.
Basis-Paleoclimate and model studies provide evidence for polar amplification of warming and reduction in land-surface snow cover. The magnitude of the surface warming will be dependent on potential changes in the poleward heat flux, which are uncertain, and the magnitude of the global warming. Current models may significantly underestimate the magnitude of polar amplification while exaggerating tropical changes if poleward heat flux by the ocean-atmosphere system is not properly simulated by models.
(6) Global sea level will rise at an increasing rate, although with some probability that the rate of rise may not be significantly greater than at present. The most reasonable estimates for the rate of sea-level rise are for a rise of 5-40 cm by 2050, as compared to a rise of 5-12 cm if rates of rise over the past century continue (figure 7).
Basis-The most tractable part of making the estimate for the next several decades is projecting the component due to seawater expansion, whose rate of change is closely dependent on the atmospheric warming. Reasonable estimates of the retreat of mountain glaciers are also available. The mass balances of the polar ice sheets are highly uncertain, and are likely to be important only on longer time scales. It is important to recognize that these estimates ignore long-term issues relating to both the slow response of the major ice caps (Greenland and Antarctica), potentially different responses on Antarctica and Greenland, and the continuing (centuries-long response) rise in sea level as the deep ocean only slowly experiences the warming at the surface. These estimates also do not consider the sea-level rise that would result from a potential catastrophic collapse of the west Antarctic ice sheet, which has been proposed but remains a subject of considerable debate.
(7) Solar variability over the next 50 years will not induce a prolonged forcing that is significant in comparison with the effects of the increasing concentrations of CO2 and other greenhouse gases.
Basis-The magnitude of the forcing from known levels of solar variability (variations of roughly 0.2 W/m2 have been occurring as a result of the solar cycle) is small compared to projected changes in greenhouse gas forcing (forcing of roughly 2 to 3.5 W/m2 is projected to occur as a result of likely emission scenarios, according to the IPCC). Changes on centennial to millennial time scales may be as large as 0.5 W/m2, but are still small compared to projected greenhouse forcing.
Basis-Evaporation increases strongly with temperature increases. Current models indicate some general agreement that summer mid-latitude dryness occurs because the evaporation increase is larger than the precipitation increase.  However, uncertainties include the following: (a) increased atmospheric moisture may be transported into continental interiors, resulting in transients of increased precipitation; (b) the vegetation response to increased CO2 is not known outside of highly controlled conditions; and (c) land surface-atmosphere interactions, including the storage of wintertime moisture, are still poorly represented in models.
(9) High-latitude precipitation will increase, with potential feedback effects related to the influence of additional freshwater on the thermohaline circulation and of increased snowfall or rain on the mass balance of polar ice caps.
Basis-With global warming, there will be an increase in the atmospheric mixing ratio of water vapor, producing larger moisture fluxes and more precipitation than at present in high latitudes.
(10) Antarctic and North Atlantic ocean regions will experience warming that is slower than the global average.
Basis-Oceanic regions where surface waters mix downward and where deeper waters upwell to the surface will have a smaller than average surface temperature response to global warming. The suppression or enhancement of the warming will also depend on the nature of changes in the precipitation and freshwater input at the sites of downward mixing.
(11) Transient explosive volcanic eruptions will result in short term relative cooling.
Basis-Historical volcanism records indicate cooling of a few tenths of a degree lasting up to a few years following major eruptions (figure 8). The historical frequency of explosive volcanic events large enough to produce substantial increases in stratospheric aerosols suggests that a few to several such events could occur over the next several decades.
Basis-Many potential changes in variability can be identified, suggesting that some will occur. These possibilities include the following: (a) In all models, standing and transient eddy activity in the mid-latitudes decreases with the reduced meridional temperature gradient associated with global warming, which may lead to reduced wintertime variability for warmer climates; (b) El Niñ o Southern Oscillation (ENSO) frequency may be related to average temperature, as suggested in historical and some modeling studies (models have suggested both more and less persistence during warm climates); (c) variability associated with smaller scale convective activity may increase (e.g., thunderstorms) as a result of greater moisture content in the atmosphere; and (d) greenhouse warming and land-ocean temperature differences may increase the frequency of atmospheric "blocking" events, and may influence low frequency precipitation variability.
(13) Regional scale (100-2000 km) climate changes will be different from the global average changes. However, at present there is only very limited capability to estimate how various regions will respond to global climate change.
Basis-There is a significant mismatch of spatial scales between present climate system models and regional climate variations. This is especially important because of the dependence of regional-scale responses to the details of regional land-surface characteristics, especially orography, hydrological conditions, and land-surface features. The best estimates of regional change are currently based on the large-scale characteristics of model simulations, and differences between global and regional changes are uncertain but are expected to be present.
(14) Tropical storm intensity may change.
Basis-An increase in tropical storm intensity is plausible (e.g., model studies suggest increases in intensity associated with higher sea-surface temperatures), but are uncertain because of potential changes in poleward heat flux, uncertainties in tropical sea- surface temperature response, and the strength of the Hadley circulation in a greenhouse warming condition. Whether the frequency of tropical storms will increase is uncertain, in part because GCMs are currently not run at the appropriate spatial resolution to simulate hurricane formation and other factors that might lead to changes in hurricane generation.
(15) Details of the climate change over the next 25 years are uncertain.
Basis-Uncertainties in the factors controlling the natural variability of the climate, in the model simulations (as described above), and in the perturbations to atmospheric composition make it extremely difficult to predict or even suggest the details of changes in the global or regional climate on the time scale of only a few decades. In any given decade, the changes in temperature and related variables could be substantially less than or more than the model-predicted trend. Warming estimates in terms of degrees per decade and their use to analyze a single decade are, therefore, unwarranted. Focus on a decadal analysis of observations can be equally misleading.
(16) Biosphere-climate feedbacks are expected, but how much these feedbacks will amplify or moderate climate change is uncertain.
Basis-Processes governing changes in the distribution and character of vegetation are not incorporated in climate models. Limited assessments suggest plausible changes in climate may occur as a result of vegetation modifications that result from the greenhouse-gas-induced climate changes, including (but not limited to) (a) replacement of high-latitude tundra by multi-story vegetation, resulting in added warming; (b) regional impacts in response to vegetation changes from forest to grassland, etc.; (c) impacts on the CO2 concentration of changes in carbon storage in vegetation and soils; (d) regional and possibly global impacts of extensive tropical deforestation; and (e) impacts on the N2O, NO, CH4, NMHCs, and O3 cycles. These effects could amplify or moderate expected climatic changes.
Go to Part 3. Opportunities for Reducing Uncertainties