Figure  1.  Seasonal cycle of MLD from B-SOSE (a–d) and WOD18 (e–h). The values represent the climatological mean over the period during 2013 to 2018. A region within the deep MLD band is randomly selected to test the credibility of B-SOSE (black box in Fig. 1). The seasons are defined as spring including March, April, and May (MAM), summer including June, July, and August (JJA), autumn including September, October, and November (SON) and winter including December, January, and February (DJF).

Figure  2.  Seasonal cycle of Chl-a concentration from B-SOSE (a–d) and VIIRS (e–h) sensor. The values represent the climatological mean over the period from 2013 to 2021. The black box in the subplot represents the same region as the black box in Fig. 1.

Figure  3.  Time series of monthly MLD anomaly from B-SOSE and WOD18. The values are calculated by the black box depicted in Fig. 1. The asterisk (*) represents that the results are statistically significant at the 95% confidence level.

Figure  4.  Time series of monthly Chl-a anomaly derived from B-SOSE and VIIRS sensor. The values are calculated by the black box depicted in Fig. 2. The asterisk (*) represents that the results are statistically significant at the 95% confidence level.

Figure  5.  Maximum values of MLD (a) and Chl-a (d) and the occurrence time of maximum MLD (b, c) and Chl-a (e, f) from B-SOSE. b and e show the months when the maximum value occurs, c and f show the seasons when the maximum value occurs. In c and f, the colored contours represent the edges of sea ice in four seasons: the black, cyan, blue, and pink contours represent austral spring, summer, autumn, and winter, respectively.

Figure  6.  Strongest positive (a) and negative (d) correlation coefficients between the MLD and Chl-a and the time that Chl-a lags the MLD when the strongest correlation coefficient occurs (b, c, e, f). Figures 6a and d are statistically significant, and the statistically insignificant points have been removed. The colored contours represent the edges of sea ice in four seasons. The black, cyan, blue, and pink contours represent austral spring, summer, autumn, and winter, respectively.

Figure  7.  Time series of MLD (a), Chl-a (b), and iron (c) in the selected regions (P1, P2, N1, and N2). The error bars represent the standard deviation.

Figure  8.  Time series and connections of MLD, Chl, and iron in annual spring (a), summer (b), autumn (c), and winter (d) in Region P1. The magnitudes are calculated by domain-averaged over Region P1.

Figure  9.  Time series and connections of MLD, Chl-a concentration, and iron in annual spring (a), summer (b), autumn (c), and winter (d) in Region P2. The magnitudes are calculated by domain-averaged over Region P2.

Figure  10.  Time series and connections of MLD, Chl-a concentration, and iron in annual spring (a), summer (b), autumn (c), and winter (d) in Region N1. The magnitudes are calculated by domain-averaged over Region N1.

Figure  11.  Time series and connections of MLD, Chl-a concentration, and iron in annual spring (a), summer (b), autumn (c), and winter (d) in Region N2. The magnitudes are calculated by domain-averaged over Region N2.

Figure  12.  Correlation coefficients between MLD anomaly and Chl-a anomaly in the austral spring (a), summer (b), autumn (c), and winter (d) MLD and spring Chl-a. The black, cyan, and pink contours represent the edges of sea ice in the austral spring, summer, and winter, respectively. White regions represent correlations with p > 0.05.

Figure  13.  Correlation coefficients between anomaly of iron and Chl-a (a), MLD (b) in the austral summer. White regions represent correlations with p > 0.05.

Figure  14.  A schematic summarizing the response of Chl-a anomaly to the physical variables in the austral summer. The regions in blue, purple, and red indicate the areas where the influence of austral summer MLD, SST, and sea ice is significant, respectively.

Figure  15.  Time series and connections of the three terms in Eq. (5) [$ {\partial {h}} $/${\partial {t}} $, the rate of MLD deepening; $ {{w}}_{{b}} $, the vertical velocity at the ML base; $ \overrightarrow{{V}}\cdot \nabla {h} $, the horizontal advection (adv) of water in the ML] and iron anomalies in annual summer. The magnitudes are calculated by domain-averaged over Fig. 14a, where the MLD works in the austral summer.

Figure  16.  Correlation coefficients of winter iron anomaly and spring Chl-a (a) and winter MLD (b). White regions represent correlations with p > 0.05.

Figure  17.  A schematic summarizing the response of austral spring Chl-a anomaly to the physical variables in the austral winter. The regions in blue, purple, and red indicate the areas where austral winter MLD, SST, and sea ice influence are significant, respectively.

Figure  18.  Time series and connections of three terms in Eq. (5) ($ {\partial {h}} $/${\partial {t}} $, the rate of MLD deepening; $ {{w}}_{{b}} $, the vertical velocity at the ML base; $ {{\boldsymbol{V}}}\cdot \nabla {h} $, the horizontal advection (adv) of water in the ML) and iron in annual winter. The magnitudes are calculated by domain-averaged over Fig. 17a, where the MLD works in winter.

Figure  19.  A schematic summarizing the response of phytoplankton biomass to various primary physical variables in the SO on seasonal (a, b) and interannual (c, d) time scales. a. Positively correlated regions, regions in pink (P1) and red (P2) represent regions of deep MLD with high Chl-a (positively correlated). b. Negatively correlated regions, regions in light blue (N1) and blue (N2) represent regions of deep MLD with low Chl-a (negatively correlated). Regions P1 and N1 exhibit synchronous responses of Chl-a to the MLD, while Regions P2 and N2 show a one-season lagged responses of Chl-a to the MLD. c. Synchronous influence regions, regions showing positive anomalies in summer Chl-a are associated with negative anomalies in summer SST (purple), positive anomalies in summer MLD (blue), and positive anomalies in summer SIC (red). d. Delayed influence regions, regions showing positive anomalies in spring Chl-a are associated with negative anomalies in winter SST (purple), positive anomalies in winter MLD (blue), and negative anomalies in winter SIC (red).