Page 509

contemporary analogues of future climate changes and studies that are based on scenarios generated by numerical models can be useful. Studies are needed of such subjects as water and ecosystems, and they should be integrated to create a larger, coherent picture. Social, demographic, economic, and ecological data for both the United States and other countries need to be improved to aid impact assessments and adaptations.

Although estimates of impact and suggestions of adaptation abound, many suffer from four shortcomings. First, they may assess, say, the fall in yield of 1990 wheat caused by a 1°C warming in 2030, ignoring the proven adaptability of farmers, who will not behave in 2030 just as they do in 1990 if their environment has changed. Second, the studies may ignore technological changes, such as improved wheat strains. Third, the assessments are usually made without regard for the background of other changes that will affect impacts, for example, how markets for food products are changing and how production is shifting around the world from one region to another because of changes in comparative advantage. Fourth, suggestions of adaptations may fail to anticipate such side effects as salinity from irrigation.

Similar issues arise in impact studies of unmanaged ecosystems, which often study the response of a single species, neglecting the impact of other blows, such as chemical pollution, and amidst the competition and contributions of other species that grow with it. How will one species succeed another as the system adjusts at a place? How will a system of plants and animals migrate if climatic zones shift over half a century?

Many studies of adaptation must be conducted outdoors and in the current climate. The need for specialized research and development concerned with changing climate is moderated by the fact that the climates that may be experienced in 2030 are, for the most part, climates that are today being experienced somewhere, probably nearby. If the climate of Nebraska is going to become like that in Oklahoma today, experiments in Oklahoma fields now help later adaptation in Nebraska. Nevertheless, keeping in mind knowledge from simplified and controlled experiments and searching for global principles rather than catalogs of empiricisms, scientists must learn how disparate, entire systems of species live and reconstitute themselves outside as the environment, especially the concentration of CO2, changes. Analysis must include so-called pests whose depredations depend on the quality of the host and environment and alter the outcome outdoors. For crop varieties a sound strategy continues to be maintaining diverse strains and adapting them to the weather of the current decade, because the climate of the next decade will not be vastly different, even if climate is changing over a century, and because the useful economic life of a cultivar is only about a decade. It will be useful to demonstrate in the reality of outdoors how to shorten long renewal times so that man-made things can be, and natural things will be, promptly adapted to climate.