Ancillary Data Task Force Report

10 November 1997

Introduction

On 11 August 1997, a Task Force was formed to address issues regarding the storage of ancillary data in the North American Bird Banding Laboratory (BBL). John Sauer was appointed leader of the Task Force, and Peter Blancher, Mary Dodd, R. Michael Erwin, Mary Gustafson, Frances Hicks, Judy Hildenbrand, James Kelley, Kathy Klimkiewicz, and Jeff Spendelow were appointed as members.  The Task Force met on 27 August, 17 September, 7 October, 28 October, and 5 November 1997, and this report is coauthored by Task Force members.

After an initial identification of the relevant ancillary data, we examined the value of each data type, and considered additional information presented by Erica Dunn on the possible value of measurement data (Appendix 1). Minutes of each Task Force meeting are attached as Appendices 2-6 to this report.

This report has 5 sections: (1) Task Force charge; (2) comments regarding the Scope of the Task Force, (3) general conclusions about storage of ancillary data; (4) specific comments about several types of candidate ancillary data considered by the Task Force; and (5) summary statement.

1. Task Force Charge

   Tasks: The Panel Report recommends, but the Implementation Team does not necessarily endorse, the collection, processing, storing and dissemination of ancillary data such as weights and measurements collected by banders in addition to the basic, universal data currently reported to BBL. The Task Force should evaluate potential value of additional data relative to costs and recommend which, if any, should be collected, and how they should be managed. Theoretical considerations such as the potential use of ancillary data as co-variables in contemporary analytical models, and practical considerations such as usefulness in verifying basic, universal data such as species, age, and sex, should be made.

   Operations: All interests and ideas are to be considered, but in its deliberations the Task Force's over-riding consideration should be what is best for the North American Bird Banding Program. We recommend that the Task Force meet frequently and that the Task Force chairperson provide biweekly progress reports. When the Task Force completes its tasks, it will disband.

   Expected Products: An evaluation report with specific recommendations for collecting (or not collecting) and managing ancillary data.

   Completion Goal Date: November 15, 1997

2. Scope of Task Force Report

   Recommendations-We reached consensus on recommendations for types of ancillary data to be incorporated into the BBL database. We also discussed several other ancillary data types that are frequently collected by banders and have been suggested as possible additions to the BBL database. These additional data types were not recommended for addition to the BBL database at this time, in light of both the present state of our knowledge of the uses and users of the data and the current constraints on BBL budgets and staffing.

   Specification of Protocols-Most types of ancillary data are limited in their scientific use by inconsistent protocols for collection. In our consideration of these data, we did not attempt to specify the "best" protocols, and we did not specify the exact categories and data formats of the information that we recommended be collected by the BBL. These matters are best left to BBL staff to reconcile. Instead, we provided general guidance on the relevant features of each data type, and attempted to identify the data in which the nature of data collection techniques lead to the possibility of inaccurate or biased estimation of the parameters of interest.

   Identifying Costs and Benefits-In conducting our discussions, we compared the present benefits of storage of several types of ancillary data to the costs. When common costs or benefits could be identified, we summarized them in introductory statements rather than repeat them in several contexts. Our fundamental criterion for balancing costs and benefits was whether the information had current value for management or research; in the absence of well-defined uses, resources devoted to editing and storing information represent potentially wasted effort.

3. General Conclusions About Storage of Ancillary Data in the BBL
   
   1. Almost all ancillary data types we considered will only be relevant to a subset of the BBL records. Existing BBL data do not have the information, and it is anticipated that most types of ancillary data will only be collected for, or be applicable to, a subset of new and existing records. Therefore, addition of ancillary data should occur in the context of a relational database, in which the additional information can be stored as independent data tables that can be related to other tables in the database. In this report, we refer to all banding information stored by the BBL as the "database"; and to the storage location of a specific data type as a "table."
      
   2. Ancillary data can be categorized as either covariates for analysis of banding data (such as body mass as a covariate of survival) or as aids in editing and administration (such as how sexed or how aged, or auxiliary marking data). Ancillary data recommended for collection by the Task Force were associated with editing and administration.
      We recognize the potential value of storage of additional scientific data. However, collection of much of this information is limited by inconsistent collection methods and by poorly-defined uses of the data. In the absence of clear management and scientific programs for use of these data, and given the costs associated with developing procedures for collecting, storing, and retrieving the information, it did not seem appropriate for the BBL to develop the programs and store these data at this time.
      
   3. The database used by the BBL must maintain the flexibility to add tables containing additional types of ancillary data. It is likely that research will identify additional uses of ancillary data that are important for managers and other BBL clients. Although it is beyond the present resources of the BBL to conduct this research, we recommend that the BBL take an active role in encouraging the development of projects that can define best scientific uses of ancillary data. If research partners can successfully develop uses, protocols, and storage methods for ancillary data, the BBL can then add additional data tables as needed.
      
   4. Costs to the BBL of adding ancillary data tend to fall into two general categories. First, the expense of developing protocols, entry programs, editing, storage, and retrieval for most types of ancillary data may be considerable, depending on how much of the development is done by partners and how much of the editing can be done by banders prior to data submission. Second, for some types of ancillary data, there are potential costs associated with misinformation from biased or imprecise estimates due to poorly-defined protocols or use of indices. Including inappropriate or incorrect information could undermine the credibility of all BBL data.
      
4. Recommendations about Candidate Ancillary Data

   We categorize data into (1) types that we recommend for immediate addition to the BBL database; and (2) types that we recommend not be added to the database at this time.

   1. Data Recommended for Immediate Inclusion-We recommend that how aged and how sexed codes, information on auxiliary marking, and a flag for data verification be included as tables in the BBL database. These ancillary data are primarily of benefit for administrative and editing functions of the BBL, allowing staff and users to increase efficiency of editing and data verification and decrease staff time associated with answering public queries about auxiliary marking information. Common costs associated with the information are primarily (1) development of appropriate formats and standard codes; (2) development of programs for entering the data into the tables; (3) development of supplemental tables for each data type; and (4) costs of editing, storage, and retrieval of information in the tables.
      
      1. How Aged/Sexed Codes. Collecting and storing the actual information used in assessing age and sex is difficult because a large number of difficult-to-quantify primary characters (molt, measurements, other observations) are often used. We decided that the BBL does not presently have sufficient staff to identify, standardize, and store these characters. However, summary information regarding how aging and sexing was accomplished can be more readily categorized and stored. Such information will only be applicable to a subset of species, as codes are commonly recorded during songbird banding but are not generally used by waterfowl banders.

      Benefits

      1. Provides additional information to allow BBL staff and users to assess validity of age and sex identifications. Provides convenient summary of a great deal of information, such as reproductive status, measurements, plumage, and skull ossification. However, actual measurements/observations are not quantified for future users, hence, the information is of limited scientific use.
         
      2. Allow future users (and BBL staff) to reassess identifications in light of advances in sexing/aging techniques.
         
      2. Auxiliary Markers.

         Occasionally, additional auxiliary markers are placed on birds during banding. These markers are generally more visible than bands and are often reported to the BBL, but no easy way exists for BBL staff to answer these public requests for information. We recommend that individually coded, unique markers become a permanent part of the banding record. The costs of programming these data into the system would quickly be surpassed by the benefits in quick association of marked birds from reports of observations of auxiliary markers. Also, costs will be minimized as most information can be received as computer files from banders as part of the banding data entry system.

         Only species with individually coded markers (that is, unique marks for individuals of the species that can be associated with individual band numbers) should be included initially as they are much more frequently observed and reported in an identifiable manner. Birds banded with individual color band combinations (e.g. Yellow Warbler with L:red/yellow and R:pink/silver) should not be incorporated initially as combinations may be repeated on different individuals of the same species in several areas across the US and Canada. Also, colors fade and change with exposure to UV light, and observed colors may be difficult to match with original colors at banding. At present, few birds with these non-unique color band combinations are reported to the BBL.

         Individually marked birds that are a part of an international protocol should have the following data reported: type of marker (patagial/wing marker, neck collar, coded leg band), color of marker (color of codes can be incorporated, but is much less important), and individual codes on marker. These data should become part of an ancillary data table.

         The following species and markers, all of which have existing data and large numbers of resightings, would be important to include in this program at the initial stages. All are coordinated internationally and use unique marker codes.

         • Goose and swan neck collars and colored leg bands
         • Swan patagial tags
         • Eagle patagials tags and colored leg bands
         • Peregrine Falcon colored leg bands
         • Roseate Tern visual identification (VID) leg bands

         Other individually coded bands (Black Oystercatchers, Harlequin Ducks, etc.) should be included eventually, but are of lower priority as few sightings are reported compared to the other groups. However, where a good data base exists (e.g., Harlequin Ducks) that can be easily incorporated into the BBL, these data should be added to increase the usefulness of the database.

         Benefits:

         1. Will allow much more rapid association of reports of sighted birds with records in the banding database, saving time for both banders and BBL staff in identifying marked birds. Will also allow addition of coded marker to the encounter certificates that are issued by the BBL.
            
         2. Will enhance the value of marking projects, as additional resightings will be incorporated into the BBL database and continuity of records of resightings will be ensured.

         Costs:

         Acquiring new and pre-existing files with band numbers and individual codes will require editing and entry of data. Some of these files are extensive, with ca. 500,000 Canada Goose collars forming perhaps the largest source of existing information.

      3. Flag for data verification.
         A series of flags should be developed to indicate quality of data, changes and corrections made to data, and verification of records. This should be done for both banding and encounter data.

         Benefits:

         1. Allows users and BBL staff to determine whether data have been verified, preventing duplicate verification.
            
         2. Indicates data reliability to users.
            
         3. Allows tracking of changes to original records.
            
   2. Data not Recommended for Inclusion at this Time-A variety of measurement, life history, plumage, and other information is collected by some, but not all, banders. Some of this information may be of great importance to future analyses, as survival of birds may covary with body mass or other ancillary data, and variation in measurements among individuals have potential uses in discriminating among species and populations. See Appendix 1 for more details on some of these proposed uses. BBL staff should encourage research into management and scientific uses of ancillary data, and any BBL database must maintain the flexibility to incorporate additional information. However, none of the data discussed below are recommended for immediate inclusion in the BBL database. The lack of management and scientific programs designed to use the information, and incomplete understanding of limitations of the information, mitigate against the BBL assuming the costs of adding these data tables now.

      As recommended by Buckley et al. (1997), the BBL should encourage research into the use of these data as a means of augmenting the value of banding information.

      1. Wing Measurements. Wing length measurements are collected by many banders.

      Benefits

      1. Ecologically useful measure of bird size, and when combined with mass provides an index to bird condition.
         
      2. Potentially useful in distinguishing between similar species, subpopulations within species, and between sexes in some species; Consequently, wing length data may be useful for BBL edits and for distinguishing subpopulations during migration, thereby possibly giving information on source of migrants.
         
      3. Already collected by many banders in many localities across seasons, thus providing information for distinguishing species and populations not already documented
         
      2. Mass (Weight). Body mass is an important component of condition indices, and is collected by many banders.

         Benefits

         1. Potential use as measure of bird condition, especially when combined with an index of body size such as wing length.
            
         2. Potential covariate to index energetic capabilities, that could be used to evaluate consequences of condition on migratory flight range and reproductive potential upon spring arrival.
            
         3. May provide useful information for identifying important migratory stopover sites, especially (but not necessarily) if mass on recapture also collected.
            
         4. May be useful for distinguishing similar species, populations (e.g. during migration) and sex for some species.
            
         5. Useful information for edits, for reason above.
            
      3. Fat Scores. Fat scores are collected by many banders, and provide a relative measure of visible fat reserves. The indexes produced by fat scoring have been strongly criticized due to inconsistencies among observers, leading to a lack of agreement among researchers on the value of the index.

         Benefits:

         1. Provides information on condition, although at a fairly imprecise level. It is clear that instances exist in which these data are useful, such as for assessment of migration status. However, further investigation is needed to understand the value of the information.
            
         2. Assessment of fat reserves may provide useful information for interpreting body mass data.
            
      4. Other aging and sexing information. Several other candidate ancillary data exist that provide information for aging and sexing, and thus also could be used as covariates in analyses. Some of these data are collected by many banders, and are summarized in the how aged and how sexed codes. These other data include, but are not limited to:
         
         • Molt.
         • Degree of skull ossification.
         • Brood patch, presence and score.
         • Cloacal protuberance, presence and score.
         • Additional measurements, such as tail length and culmen length.
           
5. Summary Statement

   The Ancillary Data Task Force recommends that (1) how sexed and how aged codes; (2) unique auxiliary marking data; and (3) a flag for data verification be included as part of the BBL database. Because these data are generally only applicable to a portion of the BBL data, we recommend that the information be included as supplemental data tables that can be linked to the primary database. We also recommend that the BBL maintain the flexibility to add additional ancillary data tables in response to future research into the uses of these data in migratory bird management. Finally, we recommend that the BBL encourage the study of uses of ancillary data in advancing our knowledge of migratory bird populations.