There is a growing recognition in the scientific community and more broadly that:

• The Earth functions as a system, with properties and behaviour that are characteristic of the system as a whole. These include critical thresholds, �switch� or �control� points, strong non-linearities, teleconnections, chaotic elements, and unresolvable uncertainties. Understanding the components of the Earth system is critically important, but is insufficient on its own to understand the functioning of the Earth system as a whole.
  
• Humans are now a significant force in the Earth system, altering key process rates and absorbing the impacts of global environmental changes. The environmental significance of human activities is now so profound that the current geological era can be called the �Anthropocene� (Crutzen and Stoermer, 2000).

A scientific understanding of the Earth system is required to help human societies develop in ways that sustain the global life support system. The clear challenge of understanding climate variability and change and the associated consequences and feedbacks is a specific and important example of the need for a scientific understanding of the Earth as a system. It is also clear that the scientific study of the whole Earth system, taking account of its full functional and geographical complexity over time, requires an unprecedented effort of international collaboration. It is well beyond the scope of individual countries and regions.

The world�s scientific community, working in part through the three global environmental change programmes (the International Geosphere-Biosphere Programme (IGBP), the International Human Dimensions Programme on Global Environmental Change (IHDP), and the World Climate Research Programme (WCRP)), has built a solid base for understanding the Earth system. The IGBP, IHDP and WCRP have also developed effective and efficient strategies for implementing global environmental change research at the international level. The challenge to IGBP, IHDP and WCRP is to build an international programme of Earth system science, driven by a common mission and common questions, employing visionary and creative scientific approaches, and based on an ever closer collaboration across disciplines, research themes, programmes, nations and regions.

We need to build on our existing understanding of the Earth system and its interactive human and non-human processes through time in order to:

• improve evaluation and understanding of current and future global change; and
  
• place on an increasingly firm scientific basis the challenge of sustaining the global environment for future human societies.

The climate system is particularly challenging since it is known that components in the system are inherently chaotic, and there are central components which affect the system in a non-linear manner and potentially could switch the sign of critical feedbacks. The non-linear processes include the basic dynamical response of the climate system and the interactions between the different components. These complex, non-linear dynamics are an inherent aspect of the climate system. Amongst the important non-linear processes are the role of clouds, the thermohaline circulation, and sea ice. There are other broad non-linear components, the biogeochemical system and, in particular, the carbon system, the hydrological cycle, and the chemistry of the atmosphere.

Given the complexity of the climate system and the inherent multi-decadal time-scale, there is a central and unavoidable need for long-term consistent data to support climate and environmental change investigations. Data from the present and recent past, credible global climate-relevant data for the last few centuries, along with lower frequency data for the last several millennia, are all needed. Research observational data sets that span significant temporal and spatial scales are needed so that models can be refined, validated, or perhaps, most importantly, rejected. The elimination of models because they are in conflict with climate-relevant data is particularly important. Running unrealistic models will consume scarce computing resources, and the results may add unrealistic information to the needed distribution functions. Such data must be adequate in temporal and spatial coverage, in parameters measured, and in precision, to permit meaningful validation. We are still unfortunately short of data for the quantitative assessment of extremes on the global scale in the observed climate.

In sum, there is a need for:

• more comprehensive data, contemporary, historical, and palaeological, relevant to the climate system;
  
• expanded process studies that more clearly elucidate the structure of fundamental components of the Earth system and the potential for changes in these central components;
  
• greater effort in testing and developing increasingly comprehensive and sophisticated Earth system models;
  
• increased emphasis upon producing ensemble calculations of Earth system models that yield descriptions of the likelihood of a broad range of different possibilities, and finally;
  
• new efforts in understanding the fundamental behaviour of large-scale non-linear systems.

These are significant challenges, but they are not insurmountable. The challenges to understanding the Earth system including the human component are daunting, and the pressing needs are significant. However, the opportunity for progress exists, and, in fact, this opportunity simply must be realised. The issues are too important, and they will not vanish. The challenges simply must be met.

IPCC Homepage