11.1.3 Some Unifying Themes

The basic pattern of the projected warming as described in Chapter 10 is little changed from previous assessments. Examining the spread across the MMD models, temperature projections in many regions are strongly correlated with the global mean projections, with the most sensitive models in global mean temperature often the most sensitive locally. Differing treatments of regional processes and the dynamical interactions between a given region and the rest of the climate system are responsible for some spread. However, a substantial part of the spread in regional temperature projections is due to differences in the sum of the feedbacks that control transient climate sensitivity (see also Chapter 10).

The response of the hydrological cycle is controlled in part by fundamental consequences of warmer temperatures and the increase in water vapour in the atmosphere (Chapter 3). Water is transported horizontally by the atmosphere from regions of moisture divergence (particularly in the subtropics) to regions of convergence. Even if the circulation does not change, these transports will increase due to the increase in water vapour. The consequences of this increased moisture transport can be seen in the global response of precipitation, described in Chapter 10, where, on average, precipitation increases in the inter-tropical convergence zones, decreases in the subtropics, and increases in subpolar and polar regions. Over North America and Europe, the pattern of subpolar moistening and subtropical drying dominates the 21st-century projections. This pattern is also described in Section 9.5.4, which assesses the extent to which this pattern is visible over land during the 20th century in precipitation observations and model simulations. Regions of large uncertainty often lie near the boundaries between these robust moistening and drying regions, with boundaries placed differently by each model.

High-resolution model results indicate that in regions with strong orographic forcing, some of these large-scale findings can be considerably altered locally. In some cases, this may result in changes in the opposite direction to the more general large-scale behaviour. In addition, large-area and grid-box average projections for precipitation are often very different from local changes within the area (Good and Lowe, 2006). These issues demonstrate the inadequacy of inferring the behaviour at fine scales from that of large-area averages.

Another important theme in the 21st-century projections is the poleward expansion of the subtropical highs, and the poleward displacement of the mid-latitude westerlies and associated storm tracks. This circulation response is often referred to as an enhanced positive phase of the Northern or Southern Annular Mode, or when focusing on the North Atlantic, the positive phase of the North Atlantic Oscillation (NAO). In regions without strong orographic forcing, superposition of the tendency towards subtropical drying and poleward expansion of the subtropical highs creates especially robust drying responses at the poleward boundaries of the five subtropical oceanic high centres in the South Indian, South Atlantic, South Pacific, North Atlantic and, less robustly, the North Pacific (where a tendency towards El-Niño like conditions in the Pacific in the models tends to counteract this expansion). Most of the regional projections of strong drying tendencies over land in the 21st century are immediately downstream of these centres (south-western Australia, the Western Cape Provinces of South Africa, the southern Andes, the Mediterranean and Mexico). The robustness of this large-scale circulation signal is discussed in Chapter 10, while Chapters 3, 8 and 9 describe the observed poleward shifts in the late 20th century and the ability of models to simulate these shifts.

The retreats of snow and ice cover are important for local climates. The difficulty of quantifying these effects in regions of substantial topographic relief is a significant limitation of global models (see Section 11.4.3.2, Box 11.3) and is improved with dynamical and statistical downscaling. The drying effect of an earlier spring snowmelt and, more generally, the earlier reduction in soil moisture (Manabe and Wetherald, 1987) is a continuing theme in discussion of summer continental climates.

The strong interactions between sea surface temperature gradients and tropical rainfall variability provides an important unifying theme for tropical climates. Models can differ in their projections of small changes in tropical ocean temperature gradients and in the simulation of the potentially large shifts in rainfall that are related to these oceanic changes. Chou and Neelin (2004) provide a guide to some of the complexity involved in diagnosing and evaluating hydrological responses in the tropics. With a few exceptions, the spread in projections of hydrological changes is still too large to make strong statements about the future of tropical climates at regional scales (see also Section 10.3). Many AOGCMs project large tropical precipitation changes, so uncertainty as to the regional pattern of these changes should not be taken as evidence that these changes are likely to be small.

Assessments of the regional and sub-regional climate change projections have primarily been based on the AOGCM projections summarised in Table 11.1 and an analysis of the biases in the AOGCM simulations, regional downscaling studies available for some regions with either physical or statistical models or both, and reference to plausible physical mechanisms.

To assist the reader in placing the various regional assessments in a global context, Box 11.1 displays many of the detailed assessments documented in the following regional sections. Likewise, an overview of projected changes in various types of extreme weather statistics is summarised in Table 11.2,

which contains information from the assessments within this chapter and from Chapter 10. Thus, the details of the assessment that lead to each individual statement can all be found in either Chapter 10, or the respective regional sections, and links for each statement are identifiable from Table 11.2.