Dual-curve-fitting-based wind parameter extraction from shipborne nautical X-band radar data

In this paper, techniques for retrieving wind direction and speed from shipborne X-band nautical radar images are presented. First, by analyzing the radar backscatter intensity histogram and zero-pixel percentage, each individual image is designed to go through a data quality control process. With this processing, the rain cases and all-black images can be recognized. Then, a harmonic function that is least-squares fitted to the radar backscatter intensity as a function of antenna look direction is applied to determine the wind direction. Furthermore, for wind speed retrieval, an empirical third-order polynomial model is derived using the average radar backscatter intensity and the reference wind speed. To improve the accuracy of wind retrieval, a modified technique, which involves a dual-curve-fitting approach, is implemented. For the data presented in this paper, it was found that the second stage of the curve-fitting performed optimally when the data at angles of 60° to the left and right of the first-guess upwind direction were used. Also, only the data for the dual-curve-fitting were used to calculate the average intensity of the radar images for wind speed estimation. The modified method is applied to the radar data and the results are compared with the reference data measured by a ship-based anemometer. It is shown that the dual-curve-fitting algorithm produces improvements in the mean differences between the radar and the anemometer results for wind direction and speed of about 7° and 0.4 m/s, respectively, under low sea state.