IPCC Fourth Assessment Report: Climate Change 2007
Climate Change 2007: Working Group I: The Physical Science Basis

9.1 Introduction

The objective of this chapter is to assess scientific understanding about the extent to which the observed climate changes that are reported in Chapters 3 to 6 are expressions of natural internal climate variability and/or externally forced climate change. The scope of this chapter includes ‘detection and attribution’ but is wider than that of previous detection and attribution chapters in the Second Assessment Report (SAR; Santer et al., 1996a) and the Third Assessment Report (TAR; Mitchell et al., 2001). Climate models, physical understanding of the climate system and statistical tools, including formal climate change detection and attribution methods, are used to interpret observed changes where possible. The detection and attribution research discussed in this chapter includes research on regional scales, extremes and variables other than temperature. This new work is placed in the context of a broader understanding of a changing climate. However, the ability to interpret some changes, particularly for non-temperature variables, is limited by uncertainties in the observations, physical understanding of the climate system, climate models and external forcing estimates. Research on the impacts of these observed climate changes is assessed by Working Group II of the IPCC.

9.1.1 What are Climate Change and Climate Variability?

‘Climate change’ refers to a change in the state of the climate that can be identified (e.g., using statistical tests) by changes in the mean and/or the variability of its properties, and that persists for an extended period, typically decades or longer (see Glossary). Climate change may be due to internal processes and/or external forcings. Some external influences, such as changes in solar radiation and volcanism, occur naturally and contribute to the total natural variability of the climate system. Other external changes, such as the change in composition of the atmosphere that began with the industrial revolution, are the result of human activity. A key objective of this chapter is to understand climate changes that result from anthropogenic and natural external forcings, and how they may be distinguished from changes and variability that result from internal climate system processes.

Internal variability is present on all time scales. Atmospheric processes that generate internal variability are known to operate on time scales ranging from virtually instantaneous (e.g., condensation of water vapour in clouds) up to years (e.g., troposphere-stratosphere or inter-hemispheric exchange). Other components of the climate system, such as the ocean and the large ice sheets, tend to operate on longer time scales. These components produce internal variability of their own accord and also integrate variability from the rapidly varying atmosphere (Hasselmann, 1976). In addition, internal variability is produced by coupled interactions between components, such as is the case with the El-Niño Southern Oscillation (ENSO; see Chapters 3 and 8).

Distinguishing between the effects of external influences and internal climate variability requires careful comparison between observed changes and those that are expected to result from external forcing. These expectations are based on physical understanding of the climate system. Physical understanding is based on physical principles. This understanding can take the form of conceptual models or it might be quantified with climate models that are driven with physically based forcing histories. An array of climate models is used to quantify expectations in this way, ranging from simple energy balance models to models of intermediate complexity to comprehensive coupled climate models (Chapter 8) such as those that contributed to the multi-model data set (MMD) archive at the Program for Climate Model Diagnosis and Intercomparison (PCMDI). The latter have been extensively evaluated by their developers and a broad investigator community. The extent to which a model is able to reproduce key features of the climate system and its variations, for example the seasonal cycle, increases its credibility for simulating changes in climate.

The comparison between observed changes and those that are expected is performed in a number of ways. Formal detection and attribution (Section 9.1.2) uses objective statistical tests to assess whether observations contain evidence of the expected responses to external forcing that is distinct from variation generated within the climate system (internal variability). These methods generally do not rely on simple linear trend analysis. Instead, they attempt to identify in observations the responses to one or several forcings by exploiting the time and/or spatial pattern of the expected responses. The response to forcing does not necessarily evolve over time as a linear trend, either because the forcing itself may not evolve in that way, or because the response to forcing is not necessarily linear.

The comparison between model-simulated and observed changes, for example, in detection and attribution methods (Section 9.1.2), also carefully accounts for the effects of changes over time in the availability of climate observations to ensure that a detected change is not an artefact of a changing observing system. This is usually done by evaluating climate model data only where and when observations are available, in order to mimic the observational system and avoid possible biases introduced by changing observational coverage.