An approach to studying the impact of climate change that avoids dealing with model uncertainties has been urged by Glantz [1992]. Case studies of particular events in which decadal shifts in regional weather patterns have resulted in physical impacts and societal responses can indicate some general features of human perception of climate change and the resulting response. A somewhat different use of actual climate events is represented by a study [ Rosenberg, 1993] in which records of a decade of actual unusual weather are used as input to a simulation of impacts in the central United States.
Observational data have also been used to create a typical or average sequence of temperatures or precipitation; the scenarios so created can then be modified in a manner to simulate a changed climate and then applied to information about the area crops in order to estimate the climate change impacts [ Carbone and Schwartz, 1993].