Jeffrey Hall — July 03, 09:54PM
How is Distance Too Long flag applied?
I recently had a checklist flagged with the notation "distance too long." My trip was somewhat akin to breeding bird survey procedures. I traveled 10.9 miles on a back country road in Doddridge County, WV (a significantly under-birded area) over the course of 1 hour and 20 minutes. I identified the road I traveled and the one 'side trip' I took. The entire route traversed wooded areas with some farm fields, so the habitat types were not too variable. There a number of checklists from this same county, during the actual breeding bird survey, that covered a distance of 10 miles, with durations ranging from 1 hour 45 minutes to 2 hours--a slightly shorter distance over slightly greater time, admittedly, but is the difference that significant? In addition, I have seen recent checklists from Sax-Zim Bog in Minnesota with distances such as 13.54 miles (covered in 1 hour) and 29.24 miles (!) in 2 hours, 10 minutes. If 10.9 miles is too long a distance, why are greater distances accepted? The ebird Help page describes 5 miles as "ideal," but cites 50 miles as the disqualifying distance, and suggests breaking down long distances into 15 mile segments, which I essentially did. I'm quite content to stop one checklist at the 5 mile mark and start another if that is the preferred method, but I'd just like to be clearer about what happened in this case.
meg — July 05, 09:45AM
Unfortunately the distance too long flag seems at discretion of the local or regional reviewer (but cannot find that actually written anywhere). We find the category more and more frequently as if the flag is a rather recent selection not previously applied. We see no reason to flag these entries out of public view although some researchers may wish to remove them from certain analyses of the data. Yet another example of poor science applied to eBird where the data management seems to be occurring before the data is even submitted.
To resolve the situation, we suggest that you shorten routes by stopping at any county lines that you may cross. then start another entry. that will ease the consternation in the minds of some reviewers who always want the rest of us to bird in the same manner that they conduct their own birding.
Unfortunately eBird never thought through the different methods of birding at the beginning of the process and so makes no distinction between birding by foot, boat, car. Only stationary and travelling. We have one travelling site only 0.5 miles long that traverses at least 9 different habitat types. Some reviewers want separate locations for each habitat; we indicate that the habitat is "central Maine rural" which does not satisfy those other birders. Another friend isolates each township to itself and spends hours at townlines confirming time and distance as well as the accuracy of his lists. We drive and tally birds between convenient points independent of cultural boundaries that are not easily identified on most maps
You are not likely to satisfy all the reviewers who will either automatically or manually check your lists, so just satisfy yourself and enjoy your birding. Always better to have the data submitted than not to have any data at all. Despite all the flags and review, the best advice hidden deep in the eBird "manual" is the likelihood that some researcher will eventually find value in data that others too intent on perfection presently find unsatisfactory.
Jeffrey Hall — July 05, 10:17AM
Thank you, Meg, it is helpful to know that others have found this process somewhat bewildering.
Given the hodge-podge of approaches taken by different birders (and of different levels of identification ability) and different reviewers, and problems like the inability to differentiate birding by car and on foot, I'm a little dubious about some of the more grandiose claims about the utility of the ebird data. It's great to have such an extensive set of data, they are extremely useful and informative for many types of research, and in most large-scale efforts, individual problems or errors tend to get subsumed within the mass of information, but it's a good idea to keep its limitations in mind.