V1, 7/18/05 10 such we agree with the recommendation. This is a challenging goal and will require partnerships with well thought out collaborations. Such partnerships are presently being pursued along at least two different fronts. · The present SEACOOS fixed measurement array coverage is relatively robust along the west Florida shelf, moderately robust along the Georgia-Carolinas and thin along the east Florida, with a significant gap along the northeast Florida shelf. We support the SEACOOS investigators plan to prepare for Coastal Ocean Data Assimilation Experiments (CODAE), which will help scientifically guide future deployment (and/or redeployment) of observational resources. · Establish a formalized mechanism whereby the performance of the model output (skill) is used to inform the optimal configuration of the observing system design to narrow regions of overlap. CODAE is an important activity, in which we must engage, especially for exploring and determining open boundary condition issues.  However, even without a CODAE, we can design many of the observing system attributes.  That requires conceptual thought based on sound physical (and other multidisciplinary) reasoning, and will benefit from a systems engineering approach.   Metrics of model performance are an obvious aspect of the overall systems evaluation and something we are presently engaged in. Data Management · SEACOOS has not systematically identified target users of its information management system. As a result specific, “customer-driven” requirements for the IM system have not been articulated, nor has the resulting IM system been evaluated against such requirements.  The scientists within the SEACOOS project fit the loosely defined profile of “super users”, so the IM system should meet their needs among others. Yet these internal users appear to make only limited use of the SEACOOS IM system in their day to day work. SEACOOS should make an effort to identify clear target user groups (“customers”); contact representatives of those groups; determine specific requirements; and design the next generation of the SEACOOS information management system that will meet those requirements. Thus a requirements-driven approach based upon identified users must be formulated. We agree that the Information Management component of SEACOOS is at a stage in its development that identification of specific user groups is critical for the next generation of information products.  To date, our primary emphasis has been on “aggregation” and the identification of necessary standards and protocols to accomplish that aggregation.  We have identified two primary ways in which we can place increasing emphasis on identifying how best to support SEACOOS users.  One mechanism is the production of visualizations of aggregated data that provide new, useful data presentations to users.  A second