Abstract Deep convection is one of the key components of the Atlantic Meridional Overturning Circulation. The intensity of deep convection (DC) is traditionally estimated as the maximum mixed layer depth (MMLD). In this study, we developed a criterion of the minimum number of casts needed for obtaining the MMLD in the Greenland Sea with a pre-defined accuracy. The criterion depends on convection intensity. For gridded datasets, we introduce a complementary measure for the DC intensity: the area of the region with the mixed layer depth over a predefined value (800 m for the Greenland Sea, notated as S800). For a weak Qor a moderate DC, variations of its intensity is more clear from variations of the MMLD (cluster 1 in the MMLD - S800 parameter space). Then the MMLD can be obtained with the 25% accuracy for at minimum 40 casts during winter. For a well developed DC (cluster 2), variations of the DC intensity are better accessed from variations of S800 and minimum 10 casts are required. In the central Greenland Sea, the number of casts is sufficient for obtaining the interannual variations of the convection intensity only since 1986. If only Argo floats are available, minimum 4 floats should simultaneously operate in the Greenland Sea gyre during winter to reach the abovementioned accuracy. Up to present, the number of floats has been insufficient during most of the winters.

中文翻译：

---

来自有限数量的铸件的深度对流强度的准确性

摘要 深对流是大西洋经向翻转环流的重要组成部分之一。传统上，深度对流 (DC) 的强度估计为最大混合层深度 (MMLD)。在这项研究中，我们制定了以预定精度在格陵兰海获得 MMLD 所需的最少铸件数量的标准。该标准取决于对流强度。对于网格数据集，我们引入了 DC 强度的补充度量：混合层深度超过预定义值（格陵兰海为 800 m，记为 S800）的区域面积。对于弱 Q 或中等 DC，其强度的变化从 MMLD（MMLD - S800 参数空间中的集群 1）的变化中更加明显。然后，在冬季至少进行 40 次施法时，可以以 25% 的准确度获得 MMLD。对于发达的 DC（集群 2），从 S800 的变化中可以更好地访问 DC 强度的变化，并且至少需要 10 次铸型。在格陵兰海中部，仅 1986 年以来的对流强度年际变化就足以获得对流强度的年际变化。 如果只有 Argo 浮标可用，则在冬季格陵兰海环流中至少需要 4 个浮标同时运行才能达到上述精度. 到目前为止，大部分冬季的浮车数量都不足。如果只有 Argo 浮标，冬季至少 4 个浮标在格陵兰海环流中同时运行才能达到上述精度。到目前为止，大部分冬季的浮车数量都不足。如果只有 Argo 浮标，冬季至少 4 个浮标在格陵兰海环流中同时运行才能达到上述精度。到目前为止，大部分冬季的浮车数量都不足。