Scholarlink

<p>We propose a new statistical model and inference algorithm for estimating generic Thermal Comfort Index (TCI), incorporating two key practical considerations: (1) climate model outputs are inherently noisy due to model imperfections, and (2) climate variables exhibit complex, nonlinear dependencies and non-Gaussian distributions. To capture these dependencies, we employ statistical Copula modeling. We show that our approach can be interpreted as a hierarchical statistical model, which poses significant challenges for inference due to its mathematical intractability. To address this, we develop a novel importance sampling-based inference procedure to estimate the statistical properties of the TCI. Through extensive empirical evaluations, we demonstrate that our method consistently outperforms existing approaches across a variety of settings, achieving lower mean squared error (MSE) in TCI estimation. Finally, we validate the applicability of our algorithm using simulations from the WRF climate model and the Heat Index as a representative TCI. Our approach offers a valuable tool for policymakers to support health-aware and climate-resilient urban design.</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

Estimation of Thermal Comfort Index Under Climate Model Uncertainty

  • 期刊论文

  • 发表日期：

    2025年8月28日

查看全文
  • Yidan Gao,
  • Ido Nevat,
  • François Septier

摘要

We propose a new statistical model and inference algorithm for estimating generic Thermal Comfort Index (TCI), incorporating two key practical considerations: (1) climate model outputs are inherently noisy due to model imperfections, and (2) climate variables exhibit complex, nonlinear dependencies and non-Gaussian distributions. To capture these dependencies, we employ statistical Copula modeling. We show that our approach can be interpreted as a hierarchical statistical model, which poses significant challenges for inference due to its mathematical intractability. To address this, we develop a novel importance sampling-based inference procedure to estimate the statistical properties of the TCI. Through extensive empirical evaluations, we demonstrate that our method consistently outperforms existing approaches across a variety of settings, achieving lower mean squared error (MSE) in TCI estimation. Finally, we validate the applicability of our algorithm using simulations from the WRF climate model and the Heat Index as a representative TCI. Our approach offers a valuable tool for policymakers to support health-aware and climate-resilient urban design.