Research On Wide-field Wave Height Field Fusion Method Based On Scatterometer And Altimeter

The significant wave height of the ocean wave is of great significance to marine engineering and marine environmental safety.China's recently developed HY-2A satellite carries instruments such as scatterometers and altimeters for obtaining marine elements.The scatterometer can obtain wide-width sea surface wind field data,while the altimeter can only obtain the significant wave height data of the sub-satellite point.In this paper,the advantages of scatterometer and altimeter are combined,and the significant wave height in the scatterometer is measured by intelligent algorithm,which improves the space-time resolution of the significant wave height of the altimeter.The traditional fusion method uses the empirical relationship between wind and waves to convert the wind field of the scatterometer into a wave field,and corrects the wind wave field under the scatterometer by the significant wave height data obtained by the altimeter.Although the traditional fusion method can quickly obtain the significant wave height of the wide amplitude,the premise assumes that the measurement of the altimeter is accurate.In the specific implementation process,the main person determines the important details such as the search radius,and there are problems such as subjectivity and large error.Therefore,using the support vector machine and long-short-term memory two intelligent algorithms to calculate the significant wave height field under the scatterometer.Two intelligent algorithms extract the significant wave height under the altimeter and the wind field data under the same latitude and longitude scatterometer in the corresponding time period.The following three experiments are involved: wind prediction significant wave height,wind wave experience relationship directly obtain significant wave height,wind wave relationship obtained significant wave height to predict significant wave height.The wind and wave experience proposed by the predecessors has many relationships.The empirical relationship between the wind and wave involved in this paper uses the empirical relationship between wind and significant wave height proposed by Pierson and Moskowitz(1964).Then,the wind wave experience relationship directly obtained the significant wave height experiment obtained by the formula.The other two groups have the same experimental idea,that is,the experimental data is divided into training set and test set and normalized.The support vector machine uses the cross-validation method to find the optimal parameters through the selected kernel function(long-short-term memory defines neural network variables such as input layer,output layer,and offset,and determines the data dimension of input and output),predict the test set by the model established by the training set,and denormalize the predicted data.Finally,the experimental results of the two intelligent algorithms are compared with the traditional fusion methods to draw conclusions.Theoretical and experimental results show that the long-short-term memory intelligent algorithm can effectively predict the significant wave height under the wide-width scatterometer.