Research On Red Tide Warning In Pingtan Coastal Zone Of Fujian Province

In recent years,the red tide has frequently occurred offshore waters of China,which generating some adverse effects on society and the economy.Red tide is seasonal and spatially heterogeneous and difficult to predict accurately.Therefore,this article conducts a red tide early warning research in the Pingtan marine area of Fujian province.With the meteorological,water quality,and red tide monitoring data of the Pingtan waters from 2013to 2019,we analyzed the temporal and spatial characteristics and influencing factors of the red tide.On this basis,the input and output parameters were selected,and the research constructed the apparent oxygen increase(AOI)mechanism model and the BP neural network non-mechanism model respectively.And then,the field monitoring data of May2019 were used to verify the model.The following research results were obtained:In the past 11 years,a total of 15 red tides occurred in the Pingtan waters.The red tide usually occurred in April to June.The duration of the red tide concerntrated on 1 to 5 days,and the affected area was mostly below 20 km~2.The dominant algae which causing frequent red tide in Pingtan waters are Skeletonema costatum,Karenia mikimotoi,and Noctiluca algae.The water temperature range varied from 14.9～26.3℃from April to June,and the salinity fluctuates between 14.8～35.2,and average wind speed is 3.0 m/s,which meets the optimal growth conditions of dominant algae.Therefore,red tide in Pingtan occurred frequently from April to June.Longwangtou Bay,Liushui sea zone and Suao Bay,which are located in the east,northeast and northwest of Pingtan,respectively.They are the marine areas occurring red tide frequently.AOI is the amount of contribution to seawater dissolved oxygen produced by phytoplankton photosynthesis.Therefore,the density of algae affects the AOI value.According to the features of Pingtan red tide,the dominant algae data that caused frequent red tide in Pingtan waters were screened,and the relationship between AOI and algae density was used to build an AOI mechanism model.The model fitting was divided into Skeletonema costatum,Karenia mikimotoi,Prorocentrum donghaiensis and various dominant algae.According to the fitting degree and error comparison analysis of every fitting formulas,the AOI fitting formula finally determined to be suitable for the Pingtan sea area is:AOI=0.5992 lg N-2.7518(R~2=0.5443,n=121).With the monitoring data in May2019,the verification results are obtained:the average relative error of AOI and algae density are 26%,37%,respectively.That is to say,the prediction accuracy of AOI is 74%,and the prediction accuracy of algae density is 63%.The AOI red tide warning value is defined as 0.50 mg/L,and the corresponding algal density warning value is 3.00E+05cells/L.Principal component analysis(PCA)was performed on 802 sets of monitoring data of Pingtan marine area from 2013 to 2019,and the environmental factors with high contribution rates were selected:Chl-a,temperature,wind speed,sunlight,salinity,DO,water temperature and p H.Using the BP red tide early warning calculation model,meteorological and water quality factors are used as the input of the model,and Chl-a and algae density are used as the output of the model for calculation.When Chl-a is used as the output index,the optimal input combination is:temperature,sunshine,wind speed,AOI,the model has a high fitting accuracy(R~2=0.651),and the error is small.The model was verified by the monitoring data in May 2019,and its accuracy was as high as 79%.When the algal density is used as the output index,the optimal input combination of the model is:air temperature,wind speed,and Chl-a.The R~2 of this input combination reaches 0.759,the fitting degree is high,and the verification accuracy is as high as 89%.Moreover,the input factors of the two models are in good agreement with the PCA results,confirming the reliability of the analysis results.In addition,the fuzzy probability analysis based on the pre-red tide and red tide data in the Pingtan sea waters determined a Chl-a warning value of 4.0μg/L and algae density warning value of 3.00E+05 cells/L.This warning value was consistent with the AOI warning model.The results of this study can be applied to the study of the time series of the forecast of red tides and provide certain reference for the red tide prevention and control in Pingtan coastal area.