vious section. Most of the models have focused largely on surface oil slicks. Several of these models have graphical user interfaces and integrated Geographical Information Systems (GIS) and are widely used by agencies and industry for:

• Tactical (emergency) spill response. In the event of an accidental spill, a composite model is often used to predict where the spill will go and how it will weather. This kind of information plays a key role in determining protection priorities.

• Strategic (contingency) planning. The models are run in a stochastic manner (multiple-runs) to develop maps of the probability a spill would impact a region given that a spill occurs.

• Post-Spill assessment. The models are used after a spill to fill observation gaps and estimate damage. Another use is to determine the source of unknown oil. In this model, one knows where the spill ended up but needs to determine the most likely origin to identify the responsible party.

There have been two reviews of composite oil slick models published recently: ASCE (1996) and Reed et al. (1999). ASCE (1996) cites more than 50 composite slick models in existence at the time of its study. Clearly a complete review of all these models would be a major publication in itself and is beyond the scope of this report. Nevertheless a closer look at a few representative models does provide valuable insight into the present state of the art. Table 4-6 summarizes and differentiates among four widely used composite models. Some of the differences in Table 4-6 are a result of the fact that the models are intended for different purposes. SIMAP is primarily for strategic planning and perhaps post-spill assessment. OSIS and GNOME/ADIOS2 can be used for all three purposes: tactical planning, strategic planning, and post-spill assessment. OSRA is focused on strategic planning. The intent of Table 4-6 is to provide an overview of the range of models. The rows in the table indicate how each model accounts for the major processes identified in the previous sections. Two of the models are used by U.S. government agencies, while the other two are commercially available and widely used by private industry.

The National Oceanic and Atmospheric Administration has a primary responsibility in providing spill modeling expertise to the U.S. Coast Guard during major spills in U.S. waters. NOAA also works in other parts of the world on an invitation basis. NOAA actually uses two models: GNOME to account for advection, spreading, and first-order evaporation and ADIOS2 to estimate detailed weathering, dispersion, and so forth. The Minerals Management Service (MMS) model known as OSRA (Smith et al., 1982) is used to estimate spill impact probability associated with offshore oil development in the Outer Continental Shelf (OCS). This model is a key consideration in determining the potential environmental impact of future oil and gas developments. OSIS (Leech et al., 1993) is a commercial product frequently used in Europe. SIMAP (French, 1998) is another commercial product that is used in the United States especially for post-spill Natural Resources Damage Assessments