I am exploring the open data set on breeding birds, from the North American Breeding Bird Survey (BBS). The data is orgainzed in multiple files, including raw count data and information on survey protocols. The protocols say that each BBS route is run once each year, but the file RunProtocolID.txt includes "run protocol ID" (RPID) codes that seem to indicate that a given route can be censused multiple times (replicates) in a given year. For example, these lines from the file:
RPID RunProtocol_English RunProtocolDesc 101 Standard BBS 3-min, 1 observer, single run per year 102 Standard BBS Replicate 1 3-min, 1 observer (same or different person), second run in year 103 Standard BBS Replicate 2 3-min, 1 observer (same or different person), third run in year 104 Standard BBS Replicate 3 3-min, 1 observer (same or different person), fourth run in year
How am I to interpret the term "Replicate"? Does it really mean that data for routes with RPID = 104 have been collected 4 times in a given year? How does this affect the species counts provided for that route in that year? Do they represent averages over all runs?