The objective of this project is to develop a prototype statistical framework supporting Wide Area Assessment (WAA) and Remedial Investigation (RI) decisions relating to the risk of unexploded ordnance and other military munitions concentrated in underwater environments. Decision making involving underwater munitions is inherently complex due to the high degree of uncertainty in the environmental conditions that force munitions responses (burial, decay, migration) and associated risks to the public. The framework will provide a consistent approach to accurately delineating contaminated areas at underwater munitions sites through the estimation of most probable concentrations. It uses site information, environmental conditions, bottom clutter, and detection sensor characteristics. This risk surface can then be used to evaluate costs associated with various remediation approaches. An overall technical objective is to integrate trained statistical belief networks with detailed geophysical knowledge of sites, of sensors, and of the underwater environment to probabilistically estimate the most likely outcomes and tradeoffs while managing uncertainty associated with military munitions response.

The approach is to adapt existing probabilistic technological solutions to develop a prototype decision support system to manage environmental uncertainty at underwater munitions response sites. Through interaction with appropriate federal organizations, the team will document the problems facing munitions response managers so as to select specific statistical modules most useful towards the estimation of munition concentration considering site environmental characteristics, sensor capabilities, bottom clutter, and vehicle deployment strategies. This prototype framework will statistically categorize risk, even in areas with poorly sampled environmental attributes. Realistic evaluation of environmental attributes impacting underwater munitions assessment will include expected munitions type/condition, spatial distribution of initial contamination, time since initial contamination, state of burial, water depths in areas of interest, seafloor sediment types, hydrodynamic conditions, turbidity, and salinity. Limited evaluation of the assessment framework skill will be demonstrated using actual test site data.

This work will expand capabilities to accurately and cost-effectively characterize underwater munitions response sites through the integration of performance prediction modules guiding remediation decisions given uncertain and diverse site conditions. A primary benefit will be an initial assessment of underwater munition risk for use in the determination of Remediation Action (RA) versus Long Term Monitoring (LTM). This will allow recommendation of follow-on actions for the site, thus reducing costs associated with physically locating and clearing munitions. The prototype risk assessment framework will also allow future extension to include fusion of observations from multiple sensors, thereby leading to optimized survey design decisions (i.e., timing the placement of the most appropriate sensor types at the most favorable locations within the site). Scientifically, these tools will help the larger underwater UXO community demonstrate integrated sensing technologies so as to better understand the potential benefits and limitations of various advanced characterization approaches. (Anticipated Project Completion - 2015)