Scholarlink

<p>Effect of climate model horizontal resolution on snow simulation is examined in the Chinese Academy of Sciences Earth System Model version 2.0 (CAS-ESM2). Monthly simulated snow water equivalent (SWE), snow cover fraction (FSNO), and snow depth (SDP) at four resolutions (0.25°, 0.50°, 1.00°, and 1.40°) over 1980–2014 in the Northern Hemisphere (NH) are compared with ERA5L, MERRA2, and MODIS products. Further, the sensitivity of snow simulations to temperature biases are explored using CRU product. The results indicate that model simulations at 0.25° exhibit the smallest mean absolute error (MAE) compared to observation, with FSNO simulations outperforming those of the other two snow variables in particular. Specifically, CAS-ESM2 tends to overestimate snow in mountainous regions, such as western North America and the Qinghai-Xizang Plateau, especially at 1.40° resolution, but underestimates across most portion of Eurasia continent. Although spatial differences exist between observed and simulated snow cover trends, the overall NH snow cover trend exhibits a consistent decreasing tendency. Compared to the 1.40° simulation, MAE in SWE, FSNO and SDP at 0.25° are reduced by 11.03%, 9.19%, and 12.07%, respectively. Further exploration indicates a coupling between surface air temperature biases and snow underestimation, with higher-resolution model simulations reducing both temperature bias and snow-related discrepancies. The findings of this study corroborate the hypothesis that high-resolution simulations can enhance the representation of snow accumulation processes, particularly in plateau regions.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Impacts of horizontal resolution on snow simulation and its relationship with temperature in CAS-ESM2

  • 期刊论文

  • 发表日期：

    2026年6月15日

查看全文
  • Jun Zhou,
  • Aihui Wang,
  • Xianghui Kong,
  • He Zhang

摘要

Effect of climate model horizontal resolution on snow simulation is examined in the Chinese Academy of Sciences Earth System Model version 2.0 (CAS-ESM2). Monthly simulated snow water equivalent (SWE), snow cover fraction (FSNO), and snow depth (SDP) at four resolutions (0.25°, 0.50°, 1.00°, and 1.40°) over 1980–2014 in the Northern Hemisphere (NH) are compared with ERA5L, MERRA2, and MODIS products. Further, the sensitivity of snow simulations to temperature biases are explored using CRU product. The results indicate that model simulations at 0.25° exhibit the smallest mean absolute error (MAE) compared to observation, with FSNO simulations outperforming those of the other two snow variables in particular. Specifically, CAS-ESM2 tends to overestimate snow in mountainous regions, such as western North America and the Qinghai-Xizang Plateau, especially at 1.40° resolution, but underestimates across most portion of Eurasia continent. Although spatial differences exist between observed and simulated snow cover trends, the overall NH snow cover trend exhibits a consistent decreasing tendency. Compared to the 1.40° simulation, MAE in SWE, FSNO and SDP at 0.25° are reduced by 11.03%, 9.19%, and 12.07%, respectively. Further exploration indicates a coupling between surface air temperature biases and snow underestimation, with higher-resolution model simulations reducing both temperature bias and snow-related discrepancies. The findings of this study corroborate the hypothesis that high-resolution simulations can enhance the representation of snow accumulation processes, particularly in plateau regions.