| 摘要: |
| 深层叶绿素最大值(deep chlorophyll maximum,DCM)现象是海洋与湖泊中普遍存在的生态现象。对其进行数值模拟时,参数不确定性是导致模拟结果出现误差的重要原因。基于一个经典海洋生态模式(nutrients-phytoplankton model,NP),本文通过最优参数敏感性分析(optimization parameter sensitivity analysis,OPSA)方法探讨了模式参数不确定性对DCM模拟的影响。研究表明,背景场浑浊度、垂向湍流扩散系数、浮游植物营养盐含量和硝酸盐再循环系数为模式中的敏感参数,它们的扰动将导致DCM模拟发生显著改变。进一步,设计观测系统模拟试验评估了消除敏感参数误差DCM模拟的改进程度。结果显示,去除4个敏感参数误差DCM模拟平均改进了56.83%,约是去除不敏感参数误差平均改进程度(4.51%)的13倍。而且,去除敏感参数误差模拟改进的稳定性更好,变异系数仅为9.44%,去除不敏感参数误差模拟改进的变异系数达到了14.76%,稳定性较差。据此,可优先发展与敏感参数直接相关的动力过程参数化方案,或在有限的观测资源下优先对敏感参数展开目标观测,进而为提高DCM模拟与预测提供科学指导。 |
| 关键词: DCM 参数敏感性 最优扰动 OPSA方法 目标观测 |
| DOI：10.11759/hykx20220517003 |
| 分类号:731.26 |
| 基金项目:国家自然科学基金项目(92158202,41576015) |
|
| Parameter sensitivity analysis of a biological ocean model based on OPSA method |
|
GAO Yong-Li1, WANG Ji-chao1, SUN Guo-dong2, ZHANG Kun3, JIANG Xiang-yang4, WANG Ning4
|
|
1.China University of Petroleum(East China), Qingdao 266580, China;2.Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing 100029, China;3.Institute of Oceanology, Chinese Academy of Sciences, Qingdao 266071, China;4.Shandong Marine Resources and Environment Research Institute, Yantai 264006, China
|
| Abstract: |
| The deep chlorophyll maximum (DCM) is a common ecological phenomenon in oceans and lakes. Numerical simulation has emerged as an important tool for studying this phenomenon, and parameter uncertainty is a primary source of uncertainty in the simulations. By optimization parameter sensitivity analysis, the sensitivities of 10 parameters in the nutrient–phytoplankton model related to DCM simulation were investigated. The results revealed the sensitive parameters in the DCM simulation to be background turbidity, vertical turbulent diffusion, nutrient content of phytoplankton, and recycling coefficient of nitrate; perturbations in these parameters lead to considerable changes in the DCM. In addition, the observing system simulation experiment was designed to evaluate the improvement in DCM simulation while eliminating sensitive parameter errors. The results revealed that the average improvement in DCM simulation resulting from the removal of the sensitive parameter errors is 56.83%, which is approximately 13 times that obtained from the removal of the insensitive parameter errors (4.51%). Moreover, the coefficient of variation was calculated to examine the stability of simulation improvement. The values obtained were 9.44% for the removal of sensitive parameter perturbations and 14.76% for the removal of insensitive parameter perturbations, indicating decreased stability. This study suggests that prioritizing the parameterization scheme and target observation related to sensitive parameters may provide valuable insights for the advancement of DCM simulations and predictions. |
| Key words: DCM parameter sensitivity optimal perturbation OPSA target observation |
