基于回归模型的南海表层漂流浮标轨迹模拟研究  

Simulation of surface drifting buoy trajectory in the South China Sea based on regression model

作　　者：綦梦楠 张娟[1,2] QI Meng-nan;ZHANG Juan(South China Sea Marine Forecasting Center of State Oceanic Administration,Guangzhou 510310 China;Key Laboratory of Marine Environment Exploration Technology and Application,Ministry of Natural Resources,Guangzhou 510310 China)

机构地区：[1]国家海洋局南海预报中心,广东广州510310 [2]自然资源部海洋环境探测技术与应用重点实验室,广东广州510310

出　　处：《海洋预报》2021年第4期87-98,共12页Marine Forecasts

基　　金：广东省重点领域研发计划(2020B1111020005);国家重点研发计划项目(2017YFC1404701);国家海洋局南海分局海洋科学技术局长基金(180216)。

摘　　要：分别利用地转流、风海流和ROMS模式表层海流中的一项或者多项作为自变量,建立了关于表层漂流浮标漂移速度的回归模型,模拟了2017年12月—2018年2月和2019年12月—2020年2月的南海海域表层漂流浮标轨迹。根据回归模型结果对浮标进行72 h的漂移轨迹模拟,对比结果显示:基于地转流、风海流和ROMS模式表层海流3项作为自变量的回归模型(MEGR)效果最好,其浮标模拟轨迹的72 h平均距离误差为38 km,平均角度误差为35°,平均综合技术得分为0.34。将风海流和地转流作为自变量加入到以ROMS模式表层海流结果为自变量的回归模型中,浮标模拟轨迹的72 h平均距离误差减少10 km,平均角度误差减少5°,平均技术得分提高0.09,特别是在涡旋附近区域的模拟效果得到显著提升。该方法对利用ROMS模式结果进行漂移浮标轨迹预测具有较好的校正效果。此外,风海流叠加地转流数据和ROMS模式数据在南海漂流浮标轨迹的预测方面具有较好的互补性;在南海流速相对稳定的区域,利用MEGR模型得到的拟合流速可以较为准确地模拟漂移浮标轨迹,在南海涡旋活跃的区域,该模型效果有待进一步提升。Using one or more of geostrophic current,wind-driven current and sea surface current simulated by ROMS as independent variables,a regression model of drifting buoy velocity is established to simulate the trajectory of surface drifting buoy in the South China Sea from December 2017 to February 2018 and from December 2019 to February 2020.The results show that the regression model(MEGR)based on geostrophic current,wind current and ROMS surface current is the best in simulating the buoy trajectory with a 72 h-average error of 38 km and 35°for distance and angle,respectively,and an average technical score of 0.34.When the wind-driven current and geostrophic current are added as independent variables to the regression model using ROMS surface current as independent variables,the 72 h-average distance and angle errors of the simulated buoy trajectory are reduced by 10 km 5°,respectively,and the average technical score is increased by 0.09.In particular,the performance of regression model around eddy areas is significantly improved.This method has a good correction effect on the prediction of drifting buoy trajectory based on ROMS model results.In addition,the wind-driven current superimposed with geostrophic current is complementary to ROMS model data in predicting the trajectory of drifting buoys in the South China Sea.The buoy trajectory can be accurately simulated using the fitting velocity obtained by the MEGR mode in in areas with relatively stable current speed in the South China Sea,while the results of the model need to be further improved in areas with active eddies.

关 键 词：表层漂流浮标 回归模型 地转流 ROMS模式 

分 类 号：P731.21[天文地球—海洋科学]