While Landsat data have been used in assessing changes in land use and cover, including wetlands, and aircraft-based inventories have been used in monitoring submersed aquatic vegetation for some time, the use of remote sensing in monitoring the Bay proper is a more recent development. Dr. Lawrence Harding of my Center has conducted a program of remote sensing of phytoplankton biomass using an aircraft borne Ocean Data Acquisition System (ODAS) for the last six years (Harding et al.1994). ODAS consists of three radiometers in the blue-green region of the visible spectrum that measure water-leaving radiance at 460, 490 and 520 nm, a region of the spectrum responsive to changes in concentrations of chlorophyll. Since 1990 a regular series of 25-30 flights per year has been conducted. The near-synoptic maps of surface chlorophyll that have resulted have provide new insight into the spatial distribution of plant biomass, the importance of relatively small scale patchiness, and the relationship of plankton production to events which were not possible through the less-synoptic, sparse grid sampling from vessels that is routinely done.
In 1995, ODAS was accompanied in the over flights by a Sea-viewing Wide Field-of-view Sensor (SeaWlFS), which contains sensors at seven wavebands, including the 6 visible bands of the SeaWlFS to be included on a satellite to be launched later this year. These additional bands allow improved recoveries of chlorophyll in highly turbid conditions or extremely high concentrations associated with blooms. Based on this intercomparison, we expect to transition to applications of space borne SeaWlFS data soon after it is deployed.
c. Continuous profiling
Scientists in my Center are using towed profiling instruments to conduct research on the importance of physical discontinuities and shear structures to food chain dynamics in the Chesapeake Bay. While these instruments provide powerful new insight on these processes, towed profiling sensors also show great promise for environmental monitoring of such large and heterogeneous coastal ecosystems as the Chesapeake. We are using a Scanfish, one of ten or more commercially available underway profiling platforms, which was selected because it can be electrically actuated, rather than pressure actuated, an essential for operation in such shallow waters. The vehicle undulates up and down in the water column in a saw-toothed pattern (wave length 50-200 m depending on water depth) as it is towed behind a vessel. It is equipped with salinity, temperature, depth, dissolved oxygen, and chlorophyll and optical backscatter sensors and an optical plankton counter. Data are recorded aboard ship together with GPS data and, on occasion, with split-beam acoustic profiles of fish density. Conventional station based monitoring generally has had a horizontal spacing of the order of 10 kin, while continuous profiling produces an effective sampling intensity of the order of 100 m along its course. It also provides greater synopticity in that the entire Bay can be surveyed underway along a 600 km course in 4-5 days. Continuous profiling also allows the resolution of fronts, internal waves and ecosystem processes on the scale of I m in the vertical.
4. EXPANDING INFORMATION NETWORKS
The deluge of data, including the large volumes produced from temporally or spatially continuous data collection; the wide distribution of computation power; and the growth of the information super highway are causing a revolution of how we manage and use monitoring data. For a long time, the Chesapeake Bay Program labored on the development of a central data center at which the monitoring data analysts worked. Data generators would "turn over' data, much effort was spent on "gathering" data and ensuring that it was in the same format, and statisticians would analyze the data largely in isolation from scientific experts. This centralized approach did not work very well and after a decade the otherwise exemplary Chesapeake Bay monitoring program found itself in the embarrassing situation of not being able to provide data readily to external or internal users.
Inertia made this centralized model difficult to change, but now the Program is moving in the very different direction of distributed data and information networks. Data are increasingly being maintained and updated on an array of
