Global Biogeochemistry Models and Global Carbon Cycle Research at Lawrence Livermore National Laboratory

Covey, C ; Caldeira, K ; Guilderson, T ; Cameron-Smith, P ; Govindasamy, B ; Swanston, C ; Wickett, M ; Mirin, A ; Bader, D

Lawrence Livermore National Laboratory

关键词: Biogeochemistry;    Diagnosis;    Land Use;    Biology;    Simulation;   

DOI  :  10.2172/15016353
RP-ID  :  UCRL-TR-212604
RP-ID  :  W-7405-ENG-48
RP-ID  :  15016353

美国|英语

来源: UNT Digital Library

【 摘 要 】

The climate modeling community has long envisioned an evolution from physical climate models to ''earth system'' models that include the effects of biology and chemistry, particularly those processes related to the global carbon cycle. The widely reproduced Box 3, Figure 1 from the 2001 IPCC Scientific Assessment schematically describes that evolution. The community generally accepts the premise that understanding and predicting global and regional climate change requires the inclusion of carbon cycle processes in models to fully simulate the feedbacks between the climate system and the carbon cycle. Moreover, models will ultimately be employed to predict atmospheric concentrations of CO{sub 2} and other greenhouse gases as a function of anthropogenic and natural processes, such as industrial emissions, terrestrial carbon fixation, sequestration, land use patterns, etc. Nevertheless, the development of coupled climate-carbon models with demonstrable quantitative skill will require a significant amount of effort and time to understand and validate their behavior at both the process level and as integrated systems. It is important to consider objectively whether the currently proposed strategies to develop and validate earth system models are optimal, or even sufficient, and whether alternative strategies should be pursued. Carbon-climate models are going to be complex, with the carbon cycle strongly interacting with many other components. Off-line process validation will be insufficient. As was found in coupled atmosphere-ocean GCMs, feedbacks between model components can amplify small errors and uncertainties in one process to produce large biases in the simulated climate. The persistent tropical western Pacific Ocean ''double ITCZ'' and upper troposphere ''cold pole'' problems are examples. Finding and fixing similar types of problems in coupled carbon-climate models especially will be difficult, given the lack of observations required for diagnosis and validation of biogeochemical processes.