Not all carbon dioxide emission scenarios are equally likely: a subjective expert assessment

Abstract

Climate researchers use carbon dioxide emission scenarios to explore alternative climate futures and potential impacts, as well as implications of mitigation and adaptation policies. Often, these scenarios are published without formal probabilistic interpretations, given the deep uncertainty related to future development. However, users often seek such information, a likely range or relative probabilities. Without further specifications, users sometimes pick a small subset of emission scenarios and/or assume that all scenarios are equally likely. Here, we present probabilistic judgments of experts assessing the distribution of 2100 emissions under a business-as-usual and a policy scenario. We obtain the judgments through a method that relies only on pairwise comparisons of various ranges of emissions. There is wide variability between individual experts, but they clearly do not assign equal probabilities for the total range of future emissions. We contrast these judgments with the emission projection ranges derived from the shared socioeconomic pathways (SSPs) and a recent multi-model comparison producing probabilistic emission scenarios. Differences on long-term emission probabilities between expert estimates and model-based calculations may result from various factors including model restrictions, a coverage of a wider set of factors by experts, but also group think and inability to appreciate long-term processes.

Emissions – the ‘business as usual’ story is misleading

“…Happily -and that's a word we climatologists rarely get to use -the world imagined in RCP8.5 is one that, in our view, becomes increasingly implausible with every passing year 5 . Emission pathways to get to RCP8.5 generally require an unprecedented fivefold increase in coal use by the end of the century, an amount larger than some estimates of recoverable coal reserves 6 .…”

RCP8.5 tracks cumulative CO 2 emissions

“…While this is cause for guarded optimism, given the additional degradation of coupled human–natural systems that 4.5 °C would entail relative to a 3 °C world, it does not make using RCP8.5 “misleading” or useless. Furthermore, expert elicitation-based estimates of 2100 CO 2 emission levels range from 54.4 to 71.4 Gt CO 2 /y (expert median range from three elicitations) ( 13 ). While this central estimate is smaller than the 105.6 Gt CO 2 /y prescribed in RCP8.5, the same elicitation revealed 90th percentile estimates extending to 125 Gt CO 2 /y in each experiment ( 13 ).…”

“…Furthermore, expert elicitation-based estimates of 2100 CO 2 emission levels range from 54.4 to 71.4 Gt CO 2 /y (expert median range from three elicitations) ( 13 ). While this central estimate is smaller than the 105.6 Gt CO 2 /y prescribed in RCP8.5, the same elicitation revealed 90th percentile estimates extending to 125 Gt CO 2 /y in each experiment ( 13 ). Turning to integrated assessment models, the median estimate in 2100 is 94.3 Gt CO 2 /y with a range of 28.5 to 272.7 Gt CO 2 /y (5th to 95th percentile) ( 14 ).…”

The Role of Climate Sensitivity in Upper‐Tail Sea Level Rise Projections

“…In addition, we only consider two forcing scenarios (RCP 2.6 and 8.5), whereas forcing uncertainties may well exceed these bounds, in particular for multidecadal projections (Ho et al, 2019). Moreover, the analysis focuses on the relationship between SLR and Earth's ECS, rather than perhaps more salient aspects of climate change such as the transient climate response (Knutti et al, 2017).…”