An Advanced Point-wise Ambiguity Selection Algorithm: Application to SeaWinds

The scatterometer wind estimation process results in several possible wind vectors (ambiguities) at each resolution cell. The current ambiguity selection technique applied to SeaWinds on QuikSCAT data requires outside data as part of the initialization. An advanced ambiguity selection algorithm known as BYU point-wise does not use nudging data; rather, it utilizes a low-order Karhunen Loeve (KL) wind field model to promote self-consistency. In application to a subset of SeaWinds data, BYU point-wise selects 93% of the same ambiguities as the JPL method. On a set of non-storm error regions, BYU point-wise performed subjectively better in 55% of regions and subjectively worse in only 11% of regions. In cyclonic-storm cases, BYU pointwise performed subjectively better in 11% of regions while performing worse in 23% of regions. Thus, BYU point-wise generally produces more consistent results in non-storm regions without the aid of external nudging data.