Towards Modeling Tropical Cyclone Boundary Layer: From Meteorological Perspective to Wind Engineering Applications

Extreme winds in tropical cyclones (TC) are responsible for the considerable loss of civil engineering structures in TC-prone areas. For analyzing and mitigating TC-induced damage, the estimation of TC-induced extreme winds by the Monte Carlo simulation is a primary step in determining the intensity measure in performance-based wind engineering (PBWE). This requires a tropical cyclone boundary layer (TCBL) wind model, which solves for near-surface wind fields from the primitive equations of fluid motion driven by prescribed gradient wind/pressure profiles. This dissertation is thus aimed at utilizing cutting-edge meteorological advances to enhance the TCBL models in the PBWE. Theoretically, the 3D nonlinear TCBL model is the most rigorous, while a family of simplified linear TCBL models exists with various approximations. This dissertation investigates these models from the following perspectives.
(1) The effects of four key factors (algorithm’s order, thermal effects, vertical diffusivity, and surface drag coefficients) on the 3D nonlinear TCBL model are analyzed by the idealized simulation of an ensemble of exhaustive examples based on TC parameters (VMax, VT, and Holland-B). Patterns of various effects are observed by selected typical examples. At the same time, four indicators (differences between maximum surface and overall wind speeds and MAPE (mean average percentage error) of surface and overall wind speeds) are estimated over the ensemble to quantify their effects. The interaction between these effects and with other factors is also analyzed. It is noted that most of these effects center around RMW (radius to maximum winds) and have asymmetric horizontal distribution with both negative and positive portions concurrently. Mostly, the influence of these factors is less than 22%.
(2) A benchmark dataset has been established for validating TCBL models, which consists of 692 sorties (a combination of input and output data from measurements). Each sortie combines the input of Holland radial profiles given by flight-level reconnaissance data (Flight+) and at least one type of output measurement (surface stations, SFMR, H*Wind). The dropsonde data is also included as the output of vertical profiles. Among these, the 82 sorties associated with all three types of outputs are finally utilized to demonstrate the established benchmark dataset by comparing the simulation results of the 3D nonlinear and linear TCBL models. The measurements are also analyzed to update and examine the underlying models.
(3) The hierarchy of linear models has been revisited, enhanced, and further developed to enable expeditious computation of the 3D nonlinear model. The unified governing equations are examined in terms of various types of gradient wind models (asymmetric and axisymmetric). The height-dependent vertical diffusivity model, asymmetric gradient model, and vertical and radial thermal profiles are introduced into the 3D linear model by deriving semi-analytical solutions. The height-resolving column model is expressed and enhanced by wind-dependent drag coefficients and height-dependent vertical diffusivities. Additional enhancements are introduced to assess the effects of vertical velocity and nonlinear horizontal advection. Selected enhanced linear models are numerically examined by extensive examples and compared to the 2D slab model and the benchmark 3D nonlinear model. It is shown that the height-resolved column model may be an acceptable approximation, provided some restrictions are satisfied.
(4) The additional asymmetry resulting from the land-sea roughness contrast for land-falling TCs has been analyzed by the 3D nonlinear model. Even when the coastline is remote from the TC center, this distortion in surface wind speeds is discovered near the TC eyewall. A conceptual model is proposed to represent the contrast-induced asymmetry as the summation of the transitional effect and the global distortion effect, which may interact with the translation-induced asymmetry. Extensive numerical examples show that this asymmetry could influence the overall and surface over-ocean wind speeds by 22% and 14%, respectively. A numerical example also demonstrates its influence on the PBWE.
This systematic research has exhaustively explored the capability and sensitivity of the 3D nonlinear model and a hierarchy of linear models. Accordingly, the enhanced understanding and modeling of TCBL will promise to advance the reliability of the PBWE-enabled wind-resistant design of civil engineering structures.

热带气旋（TC）引发的极端大风，是热带气旋易发区域内土木工程结构遭受重大损失的关键诱因。为分析并减轻热带气旋引发的灾害，通过蒙特卡洛模拟估算热带气旋极端大风，是确定基于性能的风工程（PBWE）中强度指标的核心前置步骤。该分析流程需依托热带气旋边界层（TCBL）风模型，该模型通过由给定梯度风/气压廓线驱动的流体运动原始方程，求解近地面风场分布。据此，本学位论文旨在借助前沿气象学研究进展，完善基于性能的风工程框架下的热带气旋边界层模型。

理论层面，三维非线性热带气旋边界层模型是最为严谨的建模方案，同时现有一系列通过各类近似方法构建的简化线性热带气旋边界层模型。本论文从以下四个维度对各类模型展开系统性研究：
（1）基于热带气旋参数（最大风速VMax、移动风速VT及霍兰德参数B）构建穷尽样本集合的理想模拟，分析四类关键因素——算法阶数、热效应、垂直扩散系数及表面曳力系数——对三维非线性热带气旋边界层模型的影响。通过选取典型样本观测各类因素的影响规律，同时基于该样本集合估算四类指标：最大表面风速与整体风速的差值，以及表面风速、整体风速的平均绝对百分比误差（mean average percentage error，MAPE），以此量化各因素的影响程度，同时分析各因素间的交互作用。研究发现，上述多数因素的影响集中于最大风速半径（RMW，radius to maximum winds）区域，且同时存在正负影响区域的非对称水平分布特征，整体而言，各因素的影响幅度均小于22%。
（2）建立用于热带气旋边界层模型验证的基准数据集，该数据集包含692组探测架次数据，每组数据均由实测输入与输出数据构成。每组架次的输入数据为基于飞行高度层侦察数据（Flight+）得到的霍兰德径向廓线，输出数据则至少包含一类实测结果：地面站观测数据、海面微波辐射计（SFMR）数据或H*Wind分析系统数据。下投式探空仪数据作为垂直廓线的实测输出数据也被纳入数据集。其中，同时涵盖三类输出数据的82组架次最终被用于验证所建立的基准数据集，通过对比三维非线性与线性热带气旋边界层模型的模拟结果实现。同时，对实测数据展开分析以更新并校验相关底层模型。
（3）重新梳理、优化并进一步发展了线性模型层级体系，以实现三维非线性模型的快速计算。针对各类梯度风模型（非对称与轴对称），对统一控制方程展开分析。通过推导半解析解，将高度依赖的垂直扩散模型、非对称梯度模型以及垂直与径向热廓线引入三维线性模型。针对高分辨柱模型，通过引入风速相关的曳力系数与高度依赖的垂直扩散系数实现模型优化，并新增垂直速度与非线性水平平流效应的评估模块。通过大量数值算例对选取的优化线性模型进行数值检验，并与二维平板模型及基准三维非线性模型进行对比。结果表明，在满足一定约束条件下，高分辨柱模型可作为一种可接受的近似方案。
（4）利用三维非线性模型，分析了登陆热带气旋因海陆粗糙度差异引发的额外非对称效应。研究发现，即使海岸线远离热带气旋中心，其表面风速的畸变现象仍会出现在热带气旋眼墙附近。本文提出一个概念模型，将粗糙度差异引发的非对称效应表征为过渡效应与全局畸变效应的叠加，该效应可与平移引发的非对称性产生交互作用。大量数值算例表明，该非对称效应对整体洋面风速与表面洋面风速的影响幅度分别可达22%与14%，此外，通过一个数值算例验证了其对基于性能的风工程的影响。

本系统性研究全面探索了三维非线性模型及线性模型层级体系的性能与敏感性。由此获得的对热带气旋边界层的更深入理解与优化模型，将有望提升基于性能的风工程指导土木工程结构抗风设计的可靠性。