不同混合层深度算法的全球适用性对比分析

混合层深度是研究海洋上层动力过程及海气相互作用的一个至关重要的物理量，准确估算混合层深度对上层海洋动力学和热力学的深入研究具有重要的科学意义。本文基于Argo实时观测剖面数据，分海域、分季节对比分析了目前常用的几种混合层深度算法的异同与优缺点。结果表明，理论上最大角度法的精确度最高，曲率法其次，然后是阈值法和最优线性拟合法。最大角度法和曲率法的结果比较接近，实测数据表明曲率法的时空适用性更广。阈值法、最优线性拟合法分别受梯度阈值和密度（或温度）梯度变化的制约，其计算的混合层深度相对较浅。各种算法的差异性随着季节跃层的增强而逐渐减小，且北半球的差异小于南半球。

Comparative analysis of global applicability for different mixed layer depth algorithms

The depth of mixed layer is a critical physical parameter to study the dynamic process and air-sea interaction of the upper ocean. Accurate estimation of the depth of mixed layer is of great scientific significance for the further research of dynamic and thermodynamic processes. Based on Argo real-time observation profile data, this paper compares and analyzes the similarities, differences, advantages, and disadvantages of several commonly used algorithms of the mixed layer depth in different areas and seasons. The results show that the maximum angle method has the highest accuracy in theory. That is followed by the curvature method, the threshold method, and the optimal linear fitting method. The results of the maximum angle method and the curvature method are relatively similar. The curvature method has wider space-time applicability by comparing with the observations. The methods of threshold and optimal linear fitting are constrained by gradient threshold and density (or temperature) gradient change respectively. Both the calculated results of them are relatively shallow. The differences of various algorithms gradually decrease when the seasonal pycnocline enhance with the differences in the northern hemisphere are smaller than that in the southern hemisphere.