tion. Temporal resolution ranges from subannual to multimillennial. Dating uncertainties are in a similar range. Many of the limitations on the climatic variables that can be reconstructed, their time resolution, and dating uncertainty are set by nature, but others are related to available resources. With more resources and effort, it will be possible to advance our understanding of abrupt climate change (NRC, 1999c).
Intensive, multiparameter, often multi-investigator projects can be especially valuable. As one example, in the ice-core projects from central Greenland, duplication of the measurements by independent international teams provides exceptional confidence in most data and reveals which datasets do not warrant confidence. Sampling at very high time resolution to produce datasets complementary to those of other investigators gives an exceptionally clear picture of past climate. Such projects require more funding and effort than are typical of paleoclimatic research, but they provide an essential reference standard, or “type section,” of abrupt climate change to which other records can be compared. A difficulty is that this reference standard is from one place in high northern latitudes and is inappropriate for study of much of the climate system. Not all paleoclimatic records can be studied in the same detail as those from Greenland, but generation of at least a few similar highly resolved (preferably annually or subannually) reference standards, including a North Atlantic marine record comparable with Greenland records, would be of great value. The ultimate goal is to develop a global network of records with at least decadal resolution. Terrestrial and marine records of climate change and ecological response from the regions of the western Pacific warm pool (the warmest part of the global climate system) and the Southern Ocean and Antarctic continent (the southern cold pole of the climate system) are among the most critical targets for future paleoclimate research, including generation of reference standards.
Temperature is probably the easiest climatic variable to measure, and the public often focuses on it. But water availability over land is probably more important to economic and ecological systems. Focus on measures of precipitation, evaporation, and the quantitative difference between them is particularly important. For investigations of circulation involving the deep ocean, reconstructions of water-mass density in polar and subpolar regions are central. More broadly, improved efforts to develop and calibrate more proxies for additional paleoclimatic indicators are required to understand past changes fully.
Traditional “time-slice” reconstructions (e.g., the world 6,000 years ago) have provided insight into climate processes but are not the best ve-