| 亚北极锋表层参数提取的算法与数据敏感性实验 |
| Algorithm and Data Sensitivity Experiments for Surface Parameter Extraction of the Subarctic Front |
| |
| DOI：doi：10.3969/j.issn.1003-2029.2026.02.006 |
| 中文关键词: 亚北极锋 温度锋 锋面检测 参数特征 敏感性实验 |
| 英文关键词:subarctic front temperature front frontal detection parameter characteristics sensitivity experiments |
| 基金项目: |
|
| 摘要点击次数: 253 |
| 全文下载次数: 93 |
| 中文摘要: |
| 亚北极锋是亲潮冷水与黑潮延伸体暖水交汇的关键区域，锋面结构直接影响跨锋面物
质与能量交换，其参数的精确提取有助于量化跨锋面输运、刻画混合及能量耗散规律。本文聚焦于西北太平洋的亚北极锋，基于2014—2023 年三种不同分辨率的日均海表温度数据，结合绝对梯度法和Sobel、Canny 两种边缘检测算法，系统评估了锋面强度、位置、连续性、长度、时空摆动、频率、主导纬度等特征参数对数据和算法的依赖性和敏感度。结果显示：锋面核心区域的空间分布及其强度的季节变化特征在不同数据和算法组合下具有高度一致性，而锋面强度、连续性、锋线长度对数据和算法较为敏感。高分辨率数据得到的锋面强度变化范围更大，锋轴线连续性更强，更能揭示锋面的精细化结构，但锋面主导纬度对数据的敏感性较低；绝对梯度法适用于刻画大尺度锋面的气候态特征，而Sobel 算法、Canny 算法在提高边缘连续性和抗噪精度的同时，也会引入人为噪声，导致锋面的空间分布较散乱。综合而言，Sobel 算法、Canny 算法更适用于分析中小尺度锋面结构及其相关的能量与混合过程，而绝对梯度法则更适合提取大尺度水团边界与气候态锋面位置，本文明晰了数据分辨率与检测算法对亚北极锋提取的综合影响，为甄选海洋锋面提取算法提供了实验基础。 |
| 英文摘要: |
| The Subarctic Front, a critical convergence zone between the cold Oyashio Current and warm Kuroshio Extension, governs crossfrontal material and energy exchange. Accurate extraction of its parameters is essential for quantifying cross -frontal transport and characterizing mixing processes and energy dissipation mechanisms. Focusing on the Subarctic Temperature Front in the Northwest Pacific, this study evaluates how multiple frontal characteristics depend on data source and algorithm selection. These characteristics include intensity, location, continuity, length, spatiotemporal migration, frequency, and dominant latitude. It integrates three types of daily sea surface temperature (SST) data (2014-2023) with differing spatial resolutions and combines the absolute gradient method with the Sobel and Canny edge detection algorithms. The results indicate that the extraction of frontal parameters is jointly influenced by the data source and algorithm employed. The spatial distribution of the frontal core region and the seasonal variability of frontal intensity show high consistency across different data -algorithm combinations. In contrast, parameters such as frontal intensity, frontal length, and frontal continuity exhibit significant sensitivity to both data choice and algorithm design. Higher-resolution data yield a wider range of frontal intensity variations and improved frontal axis continuity, thereby revealing more refined structural features of the front. However, the dominant latitude of the front shows low sensitivity to data resolution. While the traditional absolute gradient method effectively captures the climatological characteristics of large-scale fronts, edge detection algorithms (Sobel, Canny) enhance edge continuity and noise resistance but may introduce artificial noise, leading to fragmented spatial patterns of the front. In summary, Sobel and Canny algorithms are better suited for analyzing meso to small scale frontal structures and their associated energy and mixing processes, whereas the absolute gradient method is more appropriate for extracting large scale water mass boundaries and climatological frontal positions. These findings clarify the combined effects of data resolution and detection algorithms on the extraction of the Subarctic Front, establishing an experimental basis for the selection of optimal oceanic front detection methodologies. |
| 查看全文 查看/发表评论 下载PDF阅读器 |
|
|
|
