PNNL

Abstract

Long-term scenarios developed by integrated assessment models are used in climate research to provide an indication of plausible long-term developments in the global energy system and land-use patterns and the associated emissions. The phenomena that determine these longterm developments (several decades or even centuries) are very different than those that operate on a shorter time-scales (a few years). Nevertheless, in the literature, we still often find direct comparisons between short-term observations and long-term developments that do not take into account the differing dynamics over these time scales. In this letter, we discuss some of differences between the factors that operate on in the short and long term and use long-term historical emissions trends to show that short-term observations are very poor indicators of long-term future emissions developments. Based on this, we conclude that the performance of long-term scenarios should be evaluated against the appropriate, corresponding long-term variables and trends. The research community may facilitate this by developing appropriate data sets and protocols that can be used to test the performance of long-term scenarios and the models that produce them.