Climate models simulate how the atmosphere, oceans, and land surface respond to increasing concentrations of GHGs and other climate drivers that vary over time (see Chapter 6). These models are based on numerical representations of fundamental Earth system processes, such as the exchange of energy, moisture, and materials between the atmosphere and the underlying ocean or land surface. Climate models have been critically important for understanding past and current climate change and remain an essential tool for projecting future changes. They have also been steadily increasing in detail, sophistication, and complexity, most notably by improving spatial resolution and incorporating representations of atmospheric chemistry, biogeochemi-cal cycling, and other Earth system processes. These improvements represent an important integrative tool because they allow for the evaluation of feedbacks between the climate system and other aspects of the Earth system.
As discussed in Chapter 6, there are a number of practical limitations, gaps in understanding, and institutional constraints that limit the ability of climate models to inform climate-related decision making, including the following
• The ability to explicitly simulate all relevant climate processes (for example, individual clouds) on appropriate space and time scales;
• Constraints on computing resources;
• Uncertainties and complexities associated with data assimilation and parameterization;
• Lack of a well-developed framework for regional downscaling;
• Representing regional modes of variability;
• Projecting changes in storm patterns and extreme weather events;
• Inclusion of additional Earth system processes, such as ice sheet dynamics and fully interactive ecosystem dynamics;
• Ability to simulate certain nonlinear processes, including thresholds, tipping points, and abrupt changes; and
• Representing all of the processes that determine the vulnerability, resilience, and adaptability of both natural and human systems.
As discussed in Chapter 6, climate modelers in the United States and around the world have begun to devise strategies, such as decadal-scale climate predictions, for improving the utility of climate model experiments. These experimental strategies may indeed yield more decision-relevant information, but, given the importance of local- and regional-scale information for planning responses to climate change, continued and expanded investments in regional climate modeling remain a particularly pressing priority. Expanded computing resources and human capital are also needed.
Progress in both regional and global climate modeling cannot occur in isolation. Expanded observations are needed to initialize models and validate results, to develop improved representations of physical processes, and to support downscaling techniques. For example, local- and regional-scale observations are needed to verify regional models or downscaled estimates of precipitation, and expanded ocean observations are needed to support decadal predictions. Certain human actions and activities, including agricultural practices, fire suppression, deforestation, water management, and urban development, can also interact strongly with climate change. Without models that account for such interactions and feedbacks among all important aspects of the Earth system and related human systems, it is difficult to fully evaluate the costs, benefits, trade-offs and co-benefits associated with different courses of action that might be taken to respond to climate change (the next subsection describes modeling approaches that address some of these considerations). An advanced generation of climate models with explicit and improved representations of terrestrial and marine ecosystems, the cryosphere, and other important systems and processes, and with improved representations and linkages to models of human systems and actions, will be as important as improving model resolution for increasing the value and utility of climate and Earth system models for decision making.
Was this article helpful?