Southern California Bight 2008 Regional Monitoring Survey

II. APPROACH TO INFORMATION MANAGEMENT

The IMS will be based on a centralized data storage model. A centralized system was selected because Bight’08 is an integrated project and the typical data user will be interested in obtaining the whole data set (or large parts thereof), rather than smaller units of data (individual parameters, subset of the geographic range) residing in individual laboratories. The centralized system was selected over the alternative of a distributed system linked through a series of FTP sites because sophisticated tools would have to be developed for users to access those sites, plus the difficulty of maintaining a linked-distributed system over an extended number of years.

The centralized database will be developed using standardized data transfer protocols (SDTP) for data exchange. The SDTP details the information to be submitted with each sample collection or processing element, the units and allowable values for each parameter, and the order in which that information will be submitted. They are necessary to ensure that data submitted by the participants are comparable and easily merged without significant effort or assumptions by the organization responsible for maintaining the centralized data system. Use of SDTP allows each participating organization to retain data they generate in their local data management system while providing a mechanism for data exchange among project participants and a means for populating a centralized database.

The SDTP will be organized through a relational structure. A relational structure involves use of multiple data tables linked through one or more common fields. A relational structure allows data created at different times (e.g. lab data vs. field data) to be entered at the time of data production, minimizing the possibility of data loss. A relational structure also minimizes redundant data entry, by allowing data that are recorded only once (e.g. station location) to be entered into separate tables rather than to be repeated in every data record.

a. Standardized Data Transfer Protocols

The first level in all of the components is the Station table, which includes a single data record for each site sampled. The Station table includes descriptors such as latitude, longitude, and general location. The Station table will also contain inclusion probabilities/area weights for each sample type at each sample site to ensure that samples are properly weighted in data analysis; since a stratified random sampling design was used to select sample sites for many Bight’08 stations, data are not equally weighted in their contribution to an overall project mean. One difference between the Bight’08 survey and previous surveys is that approximately half of the sample stations will be revisits to previous survey sites (¼ Bight’98 and ¼ Bight’03) in order to produce a trends assessment.

The second level is the Station Occupation table, which requires a record for each visit to a sampling site. Date, time, and environmental descriptors such as weather and sea state are included in this table. The Station Occupation table is linked to the Station table by a StationID field, which resides in both tables.

The third level is the Sample Event table, which contains a record for each sampling activity conducted during a visit to a site. All sampling efforts will be recorded to the event level regardless of whether the sample attempt was successful or not. This information is recorded not only to show level of effort, but also to indicate areas that potentially cannot be sampled, for future surveys. There are currently three types of Sampling Event Tables (more will be added later) corresponding to the three types of sampling activities conducted in the Coastal Ecology and Rocky Habitat components of Bight’08 (Grab, Trawl and Dive). The Sampling Event table is used to record information about each of these events, such as trawl duration, observations about sediment type in a grab, etc. Both StationID, SampleDate link the Sampling Event ­tables to Station Occupation table.

The fourth level is the Results table, which contains a separate record for every measurement result. These Results tables corresponding to different types of measurements conducted in the field (e.g. fish abundance) or in the laboratory (e.g. chemical concentration). StationID and Date link the Results tables to Sampling Event tables. For some tables, they are also linked by Sample Time and Trawl Number.

Within the Results tables, all data are captured at the level of individual replicate, rather than in a summarized form. Fields for summary quality assurance information are also included. Detailed laboratory QA data will be retained at the laboratory where the data were generated.

Figure 1 provides a listing of all of the tables in the SDTP for Coastal Ecology and Rocky Habitat, and illustrates their relational structure. The detailed descriptions of each field in each data table in the SDTP are found in each section related to specific portions of the project. Appendix 1 provides lookup lists of acceptable entries for each field, where a constrained list is required. Appendix 2 provides the structure in which metadata for the project will be stored.



Figure 1. Coastal Ecology and Rocky Habitat Relation Data Structure.