Global change--the cumulative impact of anthropogenic changes to water, energy, and biogeochemical cycles--is one of the defining issues for the geophysical and life sciences in the 1990s and beyond. The increasing emphasis on understanding the potential future climatic consequences of anthropogenic interference in the global carbon and trace gas cycles has driven an increased effort to understand the behavior of the climate system on decadal-to-centennial and longer timescales. The issues related to the long-term behavior of the climate system are both of fundamental scientific interest and of importance to society, as a reading of the Climate Convention [ United Nations, 1992] will immediately reveal. Variability on long timescales is challenging and exciting because much of the climate system's low frequency behavior and many long-term changes arise from the coupling of the atmosphere to the ocean, biogeochemical cycles, and the cryosphere, as well as from internal dynamics [ Pielke and Zeng, 1994]. Variability on decadal and longer timescales must be studied in an interdisciplinary fashion and challenges existing models and data sets, which have typically been developed or collected to address more limited problems.
The subject of decadal and longer timescale changes in the earth climate system is broad. For example, knowledge of earth system variability is essential background for modeling future climate because: (1) studies of observed records document substantial system variability, showing that climate is not stable on any timescale, (2) observational and theoretical investigations allow identification of crucial feedbacks and interactions that cause and modulate long-term changes, and (3) model development, verification, and validation require knowledge of observed system variability. Studies of low frequency behavior and long-term change have applications beyond projections of future climate, and substantial work was done prior to the advent of ``global change'' as an issue. Observation of long-term changes such as the coordinated temperature-trace gas changes in ice core records contributed to the recognition of forced changes in the climate system as a significant issue.
This chapter is focused on research conducted during 1991-1994, and will address investigations of decadal and longer timescale changes and rates of change observed in the earth climate system through studies of the paleorecord, the instrumental record, and simulations with simple and three-dimensional models. We will not comprehensively review studies of potential future climate, as discussed for example in the Intergovernmental Panel on Climate Change reports [ IPCC, 1990; 1992]. Several aspects of earth system variability that have been the subject of recent research and which are especially relevant to understanding possible future changes are: (1) the magnitude and causes of observed low frequency natural variability, (2) estimates of rates of change from observations and modeling, and (3) thresholds and nonlinearities. In each of these sections we will attempt to review observations and modeling of the atmosphere, oceans, biosphere, and cryosphere, and, where possible, effects of coupling.