The Dilemma of Including 'Hot' Models in Climate Impact Studies: A Hydrological Study

Abstract

Abstract. Efficient adaptation strategies to climate change require estimating future impacts and the uncertainty surrounding this estimation. Over- or under-estimating future uncertainty may lead to maladaptation. Hydrological impact studies typically use a top-down approach in which multiple climate models are used to assess the uncertainty related to climate model structure and climate sensitivity. Despite ongoing debate, impact modelers have typically embraced the concept of "model democracy" in which each climate model is considered equally fit. The newer CMIP6 simulations, with several models showing a climate sensitivity larger than that of CMIP5 and larger than the likely range based on past climate information and understanding of planetary physics, have reignited the model democracy debate. Some have suggested that hot models be removed from impact studies to avoid skewing impact results toward unlikely futures. This large-sample study looks at the impact of removing hot models on the projections of future streamflow over 3,107 North American catchments. More precisely, the variability of future projections of mean, high, and low flows is evaluated using an ensemble of 19 CMIP6 GCMs, 5 of which are deemed "hot" based on their global equilibrium climate sensitivity (ECS). The results show that the reduced ensemble of 14 climate models provides streamflow projections with reduced future variability for Canada, Alaska, the Southwest US, and along the Pacific coast. Elsewhere, the reduced ensemble has either no impact or results in increased variability of future streamflow, indicating that outlier climate models do not necessarily provide outlier projections of future impacts. These results emphasize the delicate nature of climate model selection, especially based on global fitness metrics that may not be appropriate for local and regional assessments.

Improving Earth System Model Selection Methodologies for Projecting Hydroclimatic Change: Case Study in the Pacific Northwest

“…Exclusion or inclusion of high ECS models is particularly complicated for water security applications. Asenjan et al (2023) found that including high ECS models for hydrologic change studies significantly changed the projections in only some of the regions they examined. Here, six of our top twenty models are from only two distinct modeling systems (CESM/E3SM and CNRM-CM6/ESM2) (Fig.…”

“…Other novel approaches using the concepts of emergent constraints over the globe or a region, or global trend constraints have been used to assess ESMs and develop relationships to scale (or constrain) responses for particular variables or regions (Hausfather et al, 2020;Simpson et al, 2021;Lyu et al, 2021;Tokarska et al, 2020;Ribes et al, 2022). However, focusing on global evaluation only for regional water security impact studies may be problematic as ESM performance and subsequent hydrologic response varies across regions (Melsen et al, 2018;Asenjan et al, 2023), thus the emphasis on regional metrics in regional studies.…”

Improving Earth System Model Selection Methodologies for Projecting Hydroclimatic Change: Case Study in the Pacific Northwest

“…Asenjan et al. (2023) found that including high ECS models for hydrologic change studies significantly changed the projections in only some of the regions they examined. Here, six of our top 20 models are from only two distinct modeling systems (CESM/E3SM and CNRM‐CM6/ESM2) (Figure 5) that have an ECS greater than 4.5°C, (CESM2, CESM2‐WACCM, E3SM‐1‐1‐ECA, CNRM‐ESM2‐1, CNRM‐CM6‐1).…”

Assessing the Hydrological Impact Sensitivity to Climate Model Weighting Strategies

“…Following the standard procedures for hydrological climate change impact analysis, a top-down hydroclimatic modeling chain was used (as outlined in Arsenault et al, 2020a;Rahimpour Asenjan et al, 2023). Precipitation and temperature data were extracted from 22 CMIP6 climate models under the SSP5-8.5 scenario for both the reference and future periods.…”

“…These considerations are crucial for improving the accuracy and robustness of future change estimates and the uncertainties associated with them, as noted by 490 Singh & AchutaRao (2020). The exclusion of GCMs might also be guided by factors other than performance, such as excluding models with a climate sensitivity considered too high (Hausfather et al, 2022;Rahimpour Asenjan et al, 2023), or based on more specific criteria, like omitting GCMs that do not physically represent the North American Great Lakes for a study focused on that region. 495…”