The nearshore wind stress, $u_{\ast }^2$, is examined using machine-learning models for air-ocean data collected via new flux buoys deployed across four experiments. Consistent with prior nearshore studies, existing open-ocean models predict nearshore $u_{\ast }^2$ with a large error of 0.0152 m²/s². Gaussian Process Regression (GPR) for nearshore $u_{\ast \text{BM}}^2$ is developed, reducing errors to 0.0108 m²/s². Nearshore air-sea parameterizations are examined with wind speed (61%) and the wind-wave frequency of encounters (16%) being the most important. A simpler nearshore, GPR-derived, wind-dependent-only model ($u_{\ast \text{NSU}}^2$) is developed, with errors of 0.0135 m²/s². GPRs, evaluated using identical variables, were applied to nearshore observations, and these observations modeled with open-ocean formulations for an initial comparison of parameterizations between these two regimes. The parameterizations are similar, though with subtle nonlinear differences. The new nearshore data set and machine-learning models enhance the accuracy of predictions and understanding of differences from the open-ocean.

中文翻译：

---

近岸风应力的观测见解和高斯过程回归的参数化


近岸风应力， $u_{\ast }^2$ ，使用机器学习模型对通过四个实验中部署的新通量浮标收集的空气-海洋数据进行了检查。与之前的近岸研究一致，现有的公海模型预测近岸 $u_{\ast }^2$ 误差较大，达0.0152 m ² /s ² 。近岸高斯过程回归 (GPR) $u_{\ast \text{BM}}^2$ 开发出来，将误差降低至0.0108 m ² /s ² 。近岸海空参数化以风速（61%）和遭遇风浪频率（16%）最为重要。一个更简单的近岸、探地雷达衍生的、仅依赖风的模型（ $u_{\ast \text{NSU}}^2$ ）得出，误差为0.0135 m ² /s ² 。使用相同变量评估的探地雷达被应用于近岸观测，这些观测用公海公式建模，以对这两种制度之间的参数化进行初步比较。参数化相似，但存在细微的非线性差异。新的近岸数据集和机器学习模型提高了预测的准确性和对公海差异的理解。