Simulating Canadian Arctic Climate at Convection-Permitting Resolution

Gulilat Tefera Diro,Laxmi Sushama

doi:10.3390/atmos10080430

Abstract

Inadequate representation and parameterization of sub-grid scale features and processes are one of the main sources for uncertainties in regional climate change projections, particularly for the Arctic regions where the climate change signal is amplified. Increasing model resolution to a couple of kilometers will be helpful in resolving some of these challenges, for example to better simulate convection and refined land heterogeneity and thus land–atmosphere interactions. A set of multi-year simulations has been carried out for the Canadian Arctic domain at 12 km and 3 km resolutions using limited-area version of the global environmental multi-scale (GEM) model. The model is integrated for five years driven by the fifth generation of the European Centre for medium-range weather forecast reanalysis (ERA-5) at the lateral boundaries. The aim of this study is to investigate the role of horizontal model resolution on the simulated surface climate variables. Results indicate that although some aspects of the seasonal mean values are deteriorated at times, substantial improvements are noted in the higher resolution simulation. The representation of extreme precipitation events during summer and the simulation of winter temperature are better captured in the convection-permitting simulation. Moreover, the observed temperature–extreme precipitation scaling is realistically reproduced by the higher resolution simulation. These results advocate for the use of convective-permitting resolution models for simulating future climate projections over the Arctic to support climate impact assessment studies such as those related to engineering applications and where high spatial and temporal resolution are beneficial.

Highlights

Several studies have documented substantial changes in the Arctic climate during recent decades, including a marked increase in surface air temperature and a reduction in sea ice extent [1,2,3] and these changes are projected to intensify in the future under enhanced greenhouse gas concentration [4]These projected changes can have considerable impacts on the local communities and on the natural and built environment
The aim of this study is to investigate the role of horizontal model resolution on the simulated surface climate variables
The observed temperature–extreme precipitation scaling is realistically reproduced by the higher resolution simulation

Summary

Introduction

Several studies have documented substantial changes in the Arctic climate during recent decades, including a marked increase in surface air temperature and a reduction in sea ice extent [1,2,3] and these changes are projected to intensify in the future under enhanced greenhouse gas concentration [4]. Global and conventional regional climate models usually underestimate precipitation extremes and the extreme precipitation–temperature (P-T) scaling due to their coarse resolution and uncertainties in representing convection [21,22] Recent studies such as Ban et al [23]. Have indicated that high-resolution models are capable of reproducing the magnitude of regional and local extremes as well as the extreme P-T scaling over mid-latitude regions, such study is yet to be done for the Arctic regions Given this documented added value and increasing computing resources, longer simulations with convection-permitting models have been carried out for climate studies over several domains such as Europe [12,24], Africa [25], Australia [26] and part of North America south of 60◦ N [27,28]. Must be noted that only daily precipitation amounts larger than 0.5 mm/day are considered

Validation of Seasonal Means and Annual Cycle

Daily Rainfall Extremes and P-T Relationship

Discussion and Conclusions