The potential of using ensemble-mean climate scenarios for crop yield projection.
Di Ma, Qi Jing, Yue-Ping Xu, Alex J Cannon, Taifeng Dong, Mikhail A Semenov, Budong Qian (2020) The potential of using ensemble-mean climate scenarios for crop yield projection. Abstract, The 54th CMOS Congress, May 24-28, 2020, Ottawa.
Uncertainty is large in climate projections based on ensembles of global climate models (GCMs). Thus it has been a common practice to use climate scenarios derived from multiple GCMs in climate change impact studies to account for uncertainties in impact projections. However, it is often time-consuming to run process-based models, such as watershed hydrological models or crop simulation models, in climate change impact studies when a large number of climate scenarios are used. In this study, we used averages (ensemble means) of the change factors derived from 20 GCMs included in the Coupled Model Intercomparison Project Phase 5 (CMIP5) to perturb the parameters in the LARS-WG weather generator and generated a single “ensemble-mean” climate scenario (En-WG) for selected locations across Canada. We used the En-WG scenarios to drive crop growth models in DSSAT v4.7 to simulate crop yields for canola and spring wheat under RCP4.5 and RCP8.5. In comparison, we also ran simulations using LARS-WG generated climate scenarios (WG) based on change factors derived from each of the 20 GCMs, as well as climate scenarios (GCM) downscaled with a multivariate quantile mapping bias correction method. Our results showed that the means and interannual variability of the simulated crop yields using the En-WG climate scenarios were often close to the ensemble means of the simulated crop yields using 20 WG climate scenarios. Our results also demonstrated that the mean crop yields simulated using the WG climate scenarios were often comparable to those simulated using the downscaled GCM data; however, the interannual variability of the crop yields was always lower for the WG scenarios than the latter due to reduced dispersion in stochastic weather generation. In summary, we found that using a single “ensemble-mean” climate scenario in crop yield projections could often produce mean crop yields comparable to the ensemble means or medians of mean crop yields simulated using climate scenarios from each of the multiple GCMs.