基于航次观测和再分析资料的南海海表二氧化碳分压反演及变化机制分析*

海表二氧化碳分压(pCO₂)是指海洋表层水和大气之间的二氧化碳(CO₂)交换处于动态平衡时CO₂的含量, 是描述海-气CO₂交换的一个主要因子。本文利用2008—2014年覆盖南海大部分海域的海表pCO₂观测资料, 结合现场海表温度和海表盐度以及卫星观测的叶绿素a数据, 构建了基于多元线性回归方法的分区域反演模型。模型在水深浅于30m的区域均方根误差为5.3μatm, 其余海区均方根误差为10.8μatm, 与前人基于个别航次的有限区域反演结果的均方根误差相当。利用该模型公式和HYbrid Coordinate Ocean Model(HYCOM)再分析海表温、盐数据及MODIS-Aqua卫星观测的叶绿素a数据进行反演, 得到了时空分辨率为5'×5'的2004—2016年的逐月南海海表pCO₂数据。该数据能较好地反映南海海表pCO₂在海表温度影响下, 春夏高、秋冬低的季节变化特征, 与前人基于航次观测的研究结果相似, 表明反演模型具有较高的可信度。进一步分析发现, 南海及邻近海域平均海表pCO₂具有显著的准十年振荡特征: 2012年附近出现了极小值, 之前表现为降低的趋势, 之后略有升高的趋势。受海表pCO₂的影响, 南海海盆平均海-气CO₂通量在2012年之前出现了显著降低的趋势, 表明南海释放到大气中的CO₂减少, 并在2007年之后的冬季出现了负值(从碳源变为碳汇), 2012年之后变化较为平缓。热带太平洋年代际振荡引起的南海区域海表盐度变化是造成海表pCO₂及海-气CO₂通量准十年变化的主要原因。分区分析的结果表明, 南海北部海表pCO₂变化最为显著, 在南海海表pCO₂的季节和准十年变化中都起到非常重要的作用。

Multi-linear regression of partial pressure of sea-surface carbon dioxide in the South China Sea and its mechanism

The partial pressure of carbon dioxide (pCO₂) refers to the content of CO₂ gas at the sea surface when the CO₂ exchange between the sea surface and atmosphere is in dynamic equilibrium, which is an important factor to calculate air-sea CO₂ flux. Based on the observations of pCO₂ covering most of the South China Sea (SCS) from 2008 to 2014, combined with in-situ observations of sea-surface temperature (SST), sea-surface salinity (SSS) and Modis-Aqua satellite observations of chlorophyll a (Chla), we construct a regional inversion of sea-surface pCO₂ based on a multi-linear regression method. The root mean square error is estimated to be 5.3 μatm in the area with depth shallower than 30 m, and 10.8 μatm in the remaining sea area, which are consistent with previous results based on cruise observations. Using the equation acquired from our method and combining with the HYbrid Coordinate Ocean Model (HYCOM) reanalysis SST and SSS data and MODIS-Aqua remote sensing Chla data, we obtained monthly sea-surface pCO₂ of the SCS from 2004 to 2016 with spatial resolution of 5'×5'. The data can well reflect the seasonal variation of sea-surface pCO₂ in the SCS under the influence of SST, which is high in spring and summer, and low in autumn and winter. These findings are similar to previous results based on cruise observations, indicating our method has rather high reliability. Further analysis shows the average sea-surface pCO₂ of the SCS and adjacent sea areas has a significant quasi-decadal oscillation: a minimum value appeared around 2011, which first showed a decreasing trend and then an increasing trend. Due to the influence of sea-surface pCO₂, the average air-sea CO₂ flux in the SCS and adjacent areas decreased significantly before 2012, and changed into negative values during winter, then changed slowly since 2012. The variation of SSS in the SCS caused by the Pacific Decadal Oscillation is the main reason for the quasi-decadal oscillation of sea-surface pCO₂ and air-sea CO₂ fluxes. Our results indicate the variation of sea-surface pCO₂ in the northern SCS is the most significant, which plays an important role in the seasonal and quasi-decadal oscillation of pCO₂ in the whole area.