STATISTICAL STORM SURGE ESTIMATES TO AID SURGE FORECASTING ALONG THE NORTHEASTERN USA

The winter weather near the northeast Atlantic coast is highly influenced by frequently occurring extratropical storm systems, referred to as Nor’easters. The storm surge associated with these systems is an important factor contributing to inundation of coastal areas, and the potential for property damage and loss of life due to storm surge and flooding necessitates accurate predictions of high water levels associated with storm conditions. In addition, the expected global sea level rise over the next century will extend the zone of impact from storms and storm surge farther inland, therefore research on the characterization of these storms and their impact in a future climate is also necessary.

In a series of previous studies (Salmun et al., 2009; 2011) we suggested the use of statistical models to aid in surge forecasting. We developed and evaluated a regression equation relating the storm-maximum value of storm surge at The Battery, N. Y., to the average significant wave height during the storm measured at a nearby buoy. We tested the predictive capability of the regression relation with a series of retrospective forecasts of storms and significant wave heights and showed that the statistical prediction of storm-maximum surge had a smaller mean error than that of NOAA's uncorrected extratropical storm-surge forecast and equivalent to the bias corrected operational forecast.

In this presentation, we discuss two aspects of the extension of our results. 1) We investigated the applicability of the statistical model to a large geographical region and showed that the regression relation can be "trained" with data from water gauge measurements and near-shore buoys along the east coast from North Carolina to Maine, and that it provides reasonable estimates of storm maximum storm surge. 2) We used the statistical model as a bias correction for operationally produced predictions using NOAA’s dynamical models, and showed that the use of the statistical surge prediction as a bias correction for the entire time series of the storm surge related to a storm event is statistically equivalent to the existing bias correction technique. We argue that this “correction” is applicable for much longer lead times and can also be used to aid in the characterization of storm surge and inundations in future climates scenarios.