The impact of model variation in CO 2 and temperature impulse response functions on emission metrics

Abstract

Abstract. Emission metrics are necessary to determine the relative climate effect of emissions of different species, such as between CO2 and CH4. Most emission metrics are based on Impulse Response Functions (IRFs) derived from singe models. There is currently very little understanding on how IRFs vary across models, and how the model spread propagates into the metric values. In this study, we first derive three CO2 IRF distributions from Carbon-Cycle models in the inter-comparison projects C4MIP and LTMIP, and three temperature IRF distributions from AOGCMs in the inter-comparison projects CMIP3 and CMIP5. Each distribution is based on the behaviour of several models, and takes into account their spread. The derived IRF distributions differ considerably, which is partially related to differences among the underlying models, but also to the specific scenarios (experimental setup) used in the inter-comparison exercises. For example, the very high emission pulse in LTMIP leads to considerably higher CO2 IRFs, while the abrupt forcing scenario in CMIP5 leads to a relatively high temperature IRF the first four to five years. The spreads within the different IRF distributions are however rather similar. In a second part of the study, we investigate how differences among the IRFs then impact GWP, GTP and iGTP emission metric values for time horizons up to 100 yr. The spread in the CO2 IRFs causes rather similar impacts in all three metrics. The LTMIP IRF gives 20–35% lower metric values, while the C4MIP IRFs give up to 40% higher values for short time horizons shifting to lower values for longer time horizons. Within each derived CO2IRF distribution, underlying model differences give similar spreads on the metrics in the range of −15% to 25% (10–90% spread). The GTP and iGTP metrics are also impacted by spread in the temperature IRFs, and this impact differs strongly between both metrics. For GTP, the impact of the spread is rather strong for species with a short life time. For BC, depending on the time horizon, 50% lower to 85% higher values can be found using the CMIP5 IRF, and slightly lower variations are found when using the CMIP3 IRFs (10% lower to 40% higher). For CH4 the impact from spread in the temperature IRF is still considerable, but it becomes small for longer-lived species. On the other hand, the impact from spread in the temperature IRF on iGTP is very small for all species for time horizons up to 100 yr as it is an integrated metric. Finally, as part of the spread in IRFs is caused by the specific setup of the inter-comparison exercises, there is a need for dedicated inter-comparison exercises to derive CO2 and temperature IRFs.