CLIMATE SENSITIVITY ANALYSIS USING NEURAL NETWORKS

Despite the progress in climate research, natural climate variability still involves basic questions on the effect of different forcing fact ors on climate. In the present study, the major climate forcing factor s are analysed by means of a neural network using data of an Antarctic ice core. The results show that a neural network can reconstruct some of the main aspects of the development of the temperature changes ove r the past 160.000 years with only minor deviations and might be a val uable tool for understanding the climate system.