|Working Group I: The Scientific Basis|
|Other reports in this collection|
It is important that the global mean forcing estimates be interpreted in a proper manner. Recall that the utility of the forcing concept is to afford a first-order perspective into the relative climatic impacts (viz., global-mean surface temperature change) of the different forcings. As stated in Section 6.2, for many of the relevant forcings (e.g., well-mixed greenhouse gases, solar, certain aerosol and O3 profile cases), model studies suggest a reasonable similarity of the climate sensitivity factor (Equation 6.1), such that a comparison of these forcings is meaningful for assessing their relative effects on the global mean surface temperature. However, as mentioned earlier, the climate sensitivity factor for some of the spatially inhomogeneous forcings has yet to be fully explored. For some of the forcings (e.g., involving some absorbing aerosol and O3 profile cases; see Hansen et al., 1997a), the climate sensitivity is markedly different than for, say, the well-mixed greenhouse gases, while, for other forcings (e.g., indirect aerosol effect), more comprehensive studies are needed before a generalisation can become possible.
It is also cautioned that it may be inappropriate to perform a sum of the forcings to derive a best estimate “total” radiative forcing. Such an operation has the limitation that there are differing degrees of reliability of the global mean estimates of the various forcings, which do not necessarily lend themselves to a well-justified quantitative manipulation. For some forcings, there is not even a central or best estimate given at present (e.g., indirect aerosol forcing), essentially due to the substantial uncertainties.
The ranges given for the various forcings in Figure 6.6, as already pointed out, do not have a statistical basis and are guided mostly by the estimates from published model studies. Performing mathematical manipulations using these ranges to obtain a “net uncertainty range” for the total forcing, therefore, lacks a rigorous basis. Adding to the complexity is the fact that each forcing has associated with it an assessment of the level of knowledge that is subjective in nature viz., LOSU (Table 6.12). The LOSU index is not a quantitative indicator and, at best, yields a qualitative sense about the reliability of the estimates, with the well-mixed greenhouse gases having the highest reliability, those with “medium” rank having lesser reliability, and with even less reliability for the “low” and “very low” rankings. To some extent, the relatively lower ranking of the non-well-mixed greenhouse gases (e.g., aerosols, O3) is associated with the fact that the forcing estimates for these agents depend on model simulations of species’ concentrations, in contrast to the well-mixed greenhouse gases whose global concentrations are well quantified.
In a general sense, the strategy and usefulness of combining global mean estimates of forcings that have different signs, spatial patterns, vertical structures, uncertainties, and LOSUs, and the resulting significance in the context of the global climate response are yet to be fully explored. For some combinations of forcing agents (e.g., well-mixed greenhouse gases and sulphate aerosol; see Section 6.2), it is apparent from model tests that the global mean responses to the individual forcings can be added to yield the total global mean response. Because linear additivity tests have yet to be performed for the complete set of agents shown in Figure 6.6, it is not possible to state with absolute certainty that the additivity concept will necessarily hold for the entire set of forcings.
Figure 6.6 depicts the uncertainties and LOSUs only for the global mean estimates. No attempt is made here to extend these subjective characterisations to the spatial domains associated with each of the forcings (see Figure 6.7, and Table 6.11 for the Northern to Southern Hemisphere ratios). As in the SAR, we reiterate that, in view of the spatial character of several of the forcing agents, the global mean estimates do not necessarily describe the complete spatial (horizontal and vertical dimensions) and seasonal climate responses to a particular radiative perturbation. Nor do they yield quantitative information about changes in parameters other than the global mean surface temperature response.
One diagnostic constraint on the total global mean forcing since pre-industrial times is likely to be provided by comparisons of model-simulated (driven by the combination of forcings) and observed climate changes, including spatially-based detection-attribution analyses (Chapter 12). However, the a posteriori inference involves a number of crucial assumptions, including the uncertainties associated with the forcings, the representativeness of the climate models’ sensitivity to the forcings, and the model’s representation of the real world’s “natural” variations.
Overall, the net forcing comprises of a large positive value due to well-mixed greenhouse gases, followed by a number of other agents that have smaller positive or negative values. Thus, relative to IPCC (1990) and over this past decade, there are now more forcing agents to be accounted for, each with a sizeable uncertainty that can affect the estimated climate response. In this regard, consideration of the “newer” forcing agents brings on an additional element of uncertainty in climate change analyses, over and above those concerning climate feedbacks and natural variability (IPCC, 1990). Both the spatial character of the forcing and doubts about the magnitudes (and, in some cases, even the sign) add to the complexity of the climate change problem. However, this does not necessarily imply that the uncertainty associated with the forcings is now of much greater importance than the issue of climate sensitivity of models.
Other reports in this collection