Modeling of Sea Spike Events with Generalized Extreme Value Distribution

For high resolution radar at low grazing angles, the existence of strong sea spikes leads to long tails in the amplitude distribution of sea clutter. The accurate modeling of sea clutter amplitude distribution in the presence of sea spikes is the main contribution of this paper. Considering the impulsive nature of sea spike events, the generalized extreme value (GEV) distribution, which originated from extreme value theory (EVT), is adopted in the modeling. Final distribution models are proposed by adding a compound Gaussian component in the GEV model to account for the contribution of non-spike samples. Validation results with measured data indicate that the proposed models can achieve satisfactory performance improvement in describing the statistical distribution of spiky sea clutter, especially in the tail region.