Chinese, Korean, Japanese, Russian, and U.S. scientists, coordinated through the international Pacific Arctic Group6 and, within the United States, through the IARPC DBO Interagency task team. As conceived, the DBO is a holistic approach to track and understand the effects of changing oceanographic and sea ice conditions on the marine ecosystem. Until recently, biophysical sampling has occurred at several shelf biological hotspots from research vessels-of-opportunity that transit the region. The biological sampling, which samples water column and benthic organisms, seabirds, and marine mammals to evaluate species composition, biomass, and the size and condition of key organisms, also includes standard physical oceanographic and nutrient measurements. The shipboard sampling is largely limited to the open-water season but is supplemented by satellite measurements and data from oceanographic moorings (two of the DBO sites have biophysical mooring arrays, and two sites have only physical mooring arrays). However, at present many of the moorings are temporary components of limited-duration process studies, under national or international auspices, being undertaken in the region. Although the DBO program provides an emerging opportunity for assessing biophysical changes over western Arctic shelves, a more concerted effort to coordinate and systematize the sampling over seasonal and interannual scales will be necessary. As a result of western Arctic DBO activities, the Norwegian government is proposing a similar DBO project in the marine waters surrounding Svalbard.
The sampling strategy (duration, sampling rate, spatial extent, locations) of a particular monitoring effort will vary, depending upon the process or variable of interest. There will be a need to measure key system attributes at multi-decadal time scales at relevant rates and obvious locations. Other monitoring efforts need to be adaptive, taking into consideration results that emerge from retrospective (including paleoclimatic) studies, models, and other observations. These may suggest a hypothesis-based observation approach, perhaps of shorter duration (3 to 5 years) with a specific focus. If the results are found to address a critical need, then the sampling may transition into a longer-term effort. An adaptive monitoring effort also allows for the findings of an intensive process study to adjust monitoring activities. Statistical approaches or data assimilation models can aid in devising optimal sampling strategies. However, it is almost certain that resources will be inadequate to execute an optimal sampling strategy for many relevant variables. Here again, data assimilation models might clarify the trade-offs in designing options for sub-optimal (from a statistical perspective) sampling designs. Periodic evaluation can be used to determine whether the monitoring efforts need to be modified, augmented, or suspended.