Spatially resolved evaluation of Earth system models with satellite column-averaged CO₂

• Main Authors: B. K. Gier, M. Buchwitz, M. Reuter, P. M. Cox, P. Friedlingstein, V. Eyring
• Format: Article
• Language: English
• Published: Copernicus Publications 2020-12-01
• Series: Biogeosciences
• Online Access: https://bg.copernicus.org/articles/17/6115/2020/bg-17-6115-2020.pdf

<p>Earth system models (ESMs) participating in the Coupled Model Intercomparison Project Phase 5 (CMIP5) showed large uncertainties in simulating atmospheric CO<span class="inline-formula">₂</span> concentrations. We utilize the Earth System Model Evaluation Tool (ESMValTool) to evaluate emission-driven CMIP5 and CMIP6 simulations with satellite data of column-average CO<span class="inline-formula">₂</span> mole fractions (XCO<span class="inline-formula">₂</span>). XCO<span class="inline-formula">₂</span> time series show a large spread among the model ensembles both in CMIP5 and CMIP6. Compared to the satellite observations, the models have a bias of <span class="inline-formula">+</span>25 to <span class="inline-formula">−</span>20&thinsp;ppmv in CMIP5 and <span class="inline-formula">+</span>20 to <span class="inline-formula">−</span>15&thinsp;ppmv in CMIP6, with the multi-model mean biases at <span class="inline-formula">+</span>10 and <span class="inline-formula">+</span>2&thinsp;ppmv, respectively. The derived mean atmospheric XCO<span class="inline-formula">₂</span> growth rate (GR) of 2.0&thinsp;ppmv&thinsp;yr<span class="inline-formula">⁻¹</span> is overestimated by 0.4&thinsp;ppmv&thinsp;yr<span class="inline-formula">⁻¹</span> in CMIP5 and 0.3&thinsp;ppmv&thinsp;yr<span class="inline-formula">⁻¹</span> in CMIP6 for the multi-model mean, with a good reproduction of the interannual variability. All models capture the expected increase of the seasonal cycle amplitude (SCA) with increasing latitude, but most models underestimate the SCA. Any SCA derived from data with missing values can only be considered an “effective” SCA, as the missing values could occur at the peaks or troughs. The satellite data are a combined data product covering the period 2003–2014 based on the Scanning Imaging Absorption Spectrometer for Atmospheric Chartography (SCIAMACHY)/Envisat (2003–2012) and Thermal And Near infrared Sensor for carbon Observation Fourier transform spectrometer/Greenhouse Gases Observing Satellite (TANSO-FTS/GOSAT) (2009–2014) instruments. While the combined satellite product shows a strong negative trend of decreasing effective SCA with increasing XCO<span class="inline-formula">₂</span> in the northern midlatitudes, both CMIP ensembles instead show a non-significant positive trend in the multi-model mean. The negative trend is reproduced by the models when sampling them as the observations, attributing it to sampling characteristics. Applying a mask of the mean data coverage of each satellite to the models, the effective SCA is higher for the SCIAMACHY/Envisat mask than when using the TANSO-FTS/GOSAT mask. This induces an artificial negative trend when using observational sampling over the full period, as SCIAMACHY/Envisat covers the early period until 2012, with TANSO-FTS/GOSAT measurements starting in 2009. Overall, the CMIP6 ensemble shows better agreement with the satellite data than the CMIP5 ensemble in all considered quantities (XCO<span class="inline-formula">₂</span>, GR, SCA and trend in SCA). This study shows that the availability of column-integral CO<span class="inline-formula">₂</span> from satellite provides a promising new way to evaluate the performance of Earth system models on a global scale, complementing existing studies that are based on in situ measurements from single ground-based stations.</p>