Multi-model projection of summer season expansion over the Northern Hemisphere land areas

Abstract

Understanding the change in seasonal cycle is important due to its strong influences on human society and environment. The change in length or timing of the warmest season on the year, “summer season”, is one of the intuitive indicators to monitor the ongoing changes in seasonal cycle with the potential impact on extension of favorite conditions to induce heat extremes. We devised an objective algorithm by applying threshold crossing statistics to define the relative timing and length of summer season on each grid. Using the metrics, this thesis investigates future changes in summer season timing and length at global scale using multiple climate simulations over the Northern Hemisphere land areas at specific global warming level and for upcoming twenty-first century. Studies of the uncertainties in the future projection of summer onset and withdrawal have also been limited. In this context, this thesis also explores the uncertainty factors and associated mechanisms for inter-model differences to provide reliable information of the change in summer season in future period. The first chapter quantified the future changes in summer onset and withdrawal at the 1.5°C and 2.0°C warmer conditions relative to pre-industrial levels to assess the risk and impacts of mitigation in the Paris Agreement, using the atmospheric global climate model (AGCM) large-ensemble simulations provided by the Half a degree Additional warming, Prognosis and Projected Impacts (HAPPI) project. The significant expansions of summer season over Northern Hemisphere extratropical lands are projected to be much longer in the 2.0°C than in the 1.5°C warmer conditions with slightly larger contributions by delay in withdrawal. Stronger changes are observed in middle latitudes than high latitudes and largest expansion is found over East Asia and Mediterranean up to three weeks. Associated changes in frequency of hot days like middle of summer is further analyzed focusing on the extended summer edges. The summer-like days occur more frequently in lower latitudes including East Asia, USA and Mediterranean, in accord with largest summer season lengthening. Especially, the benefits of mitigation by half a degree warming indicated that summer season lengthening and associated increases in hot days can be reduced significantly if warming is limited to 1.5°C. Overall, similar results are obtained from the repeated analysis using Coupled Model Intercomparison Project phase 5 coupled GCM simulations, suggesting a weak influence of air-sea coupling on summer season timing changes. In the second part, inter-model uncertainties in summer onset projections are examined over the northern high latitudes in the late 21st century using the latest generation of atmosphere-ocean general circulation models (CMIP6). Focusing on the three subregions with large inter-model spread (Northwest Russia, Bering Sea region and Northeastern Canada), two uncertainty factors are identified, which drives the inter-model spread. First, the influence of models’ different climate sensitivity is quantified based on the linear relation between summer onset projections and transient climate response (TCR) values. TCR is found to be a dominant factor inducing the large inter-model spread with explained variance ranging from 39-50%, depending on regions. Secondly, the additional contribution of regional snow-albedo feedback is estimated by regressing summer onset projections onto snow cover projections but using residual projections after removing TCR’s contributions. The regional snow-albedo feedback exhibited significant inter-model correlations with summer onset explaining 21-30% of the inter-model variances over Bering Sea region and Northeastern Canada. Moreover, the impacts of snow-albedo feedback are found to persist by the late summer or early fall through modulating surface heat budgets and affecting surface temperatures for the two regions. These results suggest that global climate sensitivity as well as regional snow-albedo feedback need to be considered for reliable projection of future summer onset and withdrawal at regional scales. In short, this thesis evaluated changes in summer season length and timing in the Northern Hemispheric lands by applying objective algorithm at Paris Agreement global warming levels and demonstrated the necessity of half a degree less warming by quantifying the possible benefits like frequencies of summer-like days following the expansion of summer season. Also, a comprehensive analysis was conducted to understand the inter-model uncertainty in future summer season focusing on the northern high latitudes. Contributions of models’ global climate feedback and snow-albedo feedback has been quantified, which provides important implications for constraining future changes in summer season timing and associated impact assessments.