In this study,a random-forest-based model of atmospheric CO2 column concentration over the South China Sea was built with the data of chlorophyll-a concentration,instantaneous photosynthetically active radiation,particulate inorganic carbon,particulate organic carbon,sea surface temperature,wind speed and wind direction,which were from multisource satellite remote sensing data.The accuracy of the model was verified by the data in 2020,with Bias being 0.27 ppm,R2 being 0.59 and RMSE being 1.00 ppm.The results show that the atmospheric CO2 column concentration in the South China Sea pre-sents obvious seasonal characteristics,with the highest value in spring,followed by that in summer,win-ter and autumn in sequence.Moreover,the main impact factors for the seasonal differences of atmospheric CO2 column concentration in the South China Sea vary with time.In January and April,it is affected main-ly by wind direction.In July,wind speed and wind direction are the two major impact factors.In October,sea surface temperature is the major factor.This method established based on the multisource satellite re-mote sensing data can realize the high-frequency and full-coverage monitoring of atmospheric CO2 column concentration in the South China Sea.
关键词
南海/随机森林模型/大气CO2柱浓度/估算模型/季节变化特征
Key words
the South China Sea/random forest model/atmospheric CO2 column concentration/estimation model/seasonal variation characteristic
万方数据