An Interactive Multi-Model for Consensus on Climate Change

Kocarev, Ljupco¹ 

[1] University of California, San Diego

关键词: weather/climate models;    weather predictions;    climate projections;    ensemble modeling;    imperfect models;   

DOI  :  10.2172/1136784
RP-ID  :  DOE-UCSD-0004940-Final
PID  :  OSTI ID: 1136784

美国|英语

来源: SciTech Connect

【 摘 要 】

This project purports to develop a new scheme for forming consensus among alternative climate models, that give widely divergent projections as to the details of climate change, that is more intelligent than simply averaging the model outputs, or averaging with ex post facto weighting factors. The method under development effectively allows models to assimilate data from one another in run time with weights that are chosen in an adaptive training phase using 20th century data, so that the models synchronize with one another as well as with reality. An alternate approach that is being explored in parallel is the automated combination of equations from different models in an expert-system-like framework.