Study On Comprehensive Model Of Sea State Bias Estimation In Radar Altimeter

The radar altimeter mainly used to measure the sea surface height(SSH), significant wave height(SWH) and wind speed(U) and other sea state information. Measurement process will produce a variety of error. Sea state bias(SSB) is the largest source of error of satellite altimetry. The empirical model used in practical application includes the parametric model, nonparametric model and direct estimation model.In this paper, through the distance weighted method to collinear about eighty-six million group data from 1th to 144th cycle in JASON-2 and establish collinear data set. Based on the data set establish the sea state bias direct estimation. Direct estimation of the method results for sea state bias is the true value table. Based on the true value table establish the 32 kinds sea state bias parametric estimation model, prioritizing the best form from all the models. Because of the coefficient error caused by the sea surface height difference least squares has been overcome. We can get the improved parametric model has good extensionality and also excellent accuracy.Based on the crossover point data set from 67th to 106th 40cycles of JASON-2 and kernel smoothing statistical technique establish the sea state bias nonparametric estimation model. Obtain the more accurate SSB related SWH and U and test the validity of the nonparametric model. Comparative analysis the results and the SSB of JASON-2 GDR, the two results are coincide it indicates that we recurrent the model used in the JASON-2 GDR. That lays a foundation for the future work to improve the measurement accuracy of HY-2.After analysis we know that SWH and U have great difference of different latitude areas in the northern and southern hemisphere. As SSB has great relationship with SWH and U. It will have great error only use single global model. We calculate the goodness of fit and analysis the advantage of parametric and nonparametric model. Determine that use the nonparametric model above the 30° in northern hemisphere. And other area use the parametric model can get higher estimation accuracy of SSB.