Paleoclimatic reconstructions for the Mid-Holocene, Eemian, Mid-Pliocene, and the Last Glacial Maximum are used to test the paleoanalog hypothesis and to develop a regional climate change scenario based on linear scaling by one parameter - the mean Northern Hemisphere temperature change with respect to present, ΔTHN. The empirical verification of the paleoanalog hypothesis is extended to a cold epoch for zonal means and to regional distributions of temperature for the warm epochs. The best agreement among the scaled paleoanomolies from different epochs is obtained if the seasonal temperatures are scaled with ΔTHN of the corresponding season. Preferential areas are identified where the paleoanalog hypothesis works relatively well; these areas coincide with the areas of the most pronounced warming. It is shown that the geographical distributions of the winter temperature anomalies over land in the paleodata are similar to those in the 1980-1990 period. From the three warm periods, a paleodata-based scenario is deduced for the spatial distribution of temperature in a future climate, on the scale of continents. The conditions under which scenarios based on paleodata can be applied are discussed.