During the August-September 2003 AOSN-II experiment, the Harvard Ocean Prediction System (HOPS), a system of integrated software for multidisciplinary oceanographic research and forecasting, was used to provide real-time forecasts of physical variables and model errors for dynamical investigations and as a guide for adaptive sampling.

Over the period 4 August to 3 September, HOPS provided 23 sets of real-time nowcasts and forecasts of temperature, salinity and velocity. Additionally, ten sets of real-time model error forecasts, utilizing a total of 4323 ensemble members, were issued using the Error Subspace Statistical Estimation (ESSE) methodology. Adaptive sampling recommendations were suggested on a routine basis. The forecasts and data analyses were distributed via a local web site (http://people.deas.harvard.edu/~leslie/AOSNII/) links to other AOSN-II related web sites. Data from ships (Pt. Sur, John Martin, Pt. Lobos), gliders (WHOI and Scripps) and aircraft (SST) were assimilated generally within 24 hours of appearance on the data server. Results from the operational forecasts will be presented as well as forecast skill estimates.