The Atlas and eBird

By Birding-Aus, on June 17th, 2015

I also watch the discussion about eBird with interest and from my limited observation it can be great fun as a personal birding tool – but here’s some thoughts that may get me “flamed” – I hope not – just a contribution to make sure we use the right application for the right purpose and understand their limitations (in the Australian context). Please tell me if I am misguided. My main concern is around data integrity. I see the following on the eBird Australia site and that gives me comfort for the time being. i.e. that there is no data transfer yet to the Birdlife Australia Atlas – thank heavens!: “All data entered through Eremaea eBird will be passed onto the Atlas. We have not yet set up a transfer link between the Atlas and Eremaea eBird, but we are in discussions with them — so watch this space!” Do they intend eBird to eventually provide clean data to the Atlas? or will the Atlas staff have to apply the moderation because I see no evidence of adequate eBird moderation of data entry. It is a concern to me because as I see it, eBird is a great personal tool and very feature rich. Fabulous for personal recreational birding. However that is not necessarily compatible with the data quality required for a scientific data set used for EIS work (for a Birdlife fee) for example. I have read the eBird statement on filtering and reviewing data and that is fine and dandy but on my reading their filtering for species occurrence is at State level in Australia? Whoa!!! So if someone records Mallee Fowl for Bondi, the eBird filter wont find that one?? Maybe I am misinterpreting? I currently see regular duplicated data in eBird as each person naturally wishes to have a personal data set – so for example, three people out together generate three lists of the same or similar species for the same location. That of course may not matter if you are only interested in a binary record of species presence at a site. However some records I see entered, for areas I know well, would never get near the Atlas database without requests for verification and URFs. e.g. the Warriewood Wetlands in Sydney which is one of my Atlas “favourite sites” and is an eBird “hotspot” has a count of 10 Western Gerygone for a date in 2013. Now I know you never say never with bird behaviour and distribution, but I reckon that might have attracted the Atlas moderators attention and a question or three?? But then yes, they do occur in NSW and have been recorded in (generally western) Sydney! Graeme Stevens disclaimer:Long term Atlasser with lots of skin in that game! – and very pleased to be corrected on any of the above. I also recognise that the current version of the Atlas has real limitations as a personal tool and record – but that was not it’s original purpose. A new portal release is under development I understand.

Birding-Aus mailing list
Birding-Aus@birding-aus.org
To change settings or unsubscribe visit:
http://birding-aus.org/mailman/listinfo/birding-aus_birding-aus.org

chris.sanderson

June 19, 2015 at 4:13 pm

Michael I remember back in the early days of eBird they used polygons for sites, I presume this was moved away from for a good reason. Accuracy estimates could come from gps recorders (I know my iphone gives me an accuracy estimate when I make a waypoint for a species, generally between 5-10m). I think as long as data is tagged with sources and collection methods in the metadata, then the more the merrier. Why exclude data because of possible errors? Yes, your consultant report is a good example of possible issues, but what are the risks of turning away data? How much valuable information do we then lose that stays in someone’s notebook? eBird, as stated already a few times, has a pretty robust moderation system, as does the Atlas. They work differently, and they both miss things (the moderators are only human after all). But they are the best we have and personally I think are doing a pretty amazing job when you consider the state of databases for mammals, frogs, reptiles, butterflies, dragonflies, etc. on a national scale. Cheers, Chris On Fri, Jun 19, 2015 at 2:50 PM, Michael Norris < menorris@ozemail.com.au> wrote:

menorris

June 19, 2015 at 2:50 pm

Hi all Just checked BirdLife and there is an Atlas FAQ that says there are plans and discussions underway to exchange data between eBird/Eremaea and the Bird Atlas. That strengthens my concern that there should be ways, especially on eBird, to report apparent errors. Is there any? I know of quite a few errors from my patch, some of which appeared in a consultant’s report. I very much doubt that the errors affected any conclusions but it offends my sensibilities – call me old-fashioned? – to see inaccuracies on the public record. Some of the difficulties/errors are because eBird, like the BirdLife Atlas and the ALA, uses specific points for records although people will often (for good reasons) refer to a wider area. Bunyip State Park in Victoria is an example where this 80? sq km hotspot includes half-a-dozen more specific locations. This becomes a particular problem when data is transferred to the Atlas of Living Australia (that uses an accuracy estimate – how?) or BirdLife where the focus is on 2 hectare and area searches. Some of the older records use grid squares. I suspect defining sites as polygons or circles would be better – but we are so lucky to have what we’ve got already. Michael Norris

peter

June 18, 2015 at 9:13 pm

There was a thread about birdata a few weeks ago. Apparently, a programming error means results are being returned from a wider area than expected in postcode searches. I.e the base data is actually much better than the search results indicate. However, the species distribution map does show a Black Currawong near Goulburn. If the quality of the birdata data is good, but the user interface is lacking, perhaps the answer is to abandon the birdata interface for eBird, and get the eBird programmers to add a flag that can be set by a Birdlife Australia representative to indicate that a survey has been checked and approved by them. Peter Shute Sent from my iPad

sonja.ross7

June 18, 2015 at 9:05 pm

Hi Michael & Martin, I believe BirdLife Australia is working on a new portal which will be more user friendly and utilise the information we incorporate in our e-Bird lists. Budget restrictions don’t allow making major changes fast or easy always. I imagine they won’t spend time trying to fix inaccuracies while trying to get the new system up and running. The new app is an update/improvement on the BirdLog one I think. Sonja

carlsclifford

June 18, 2015 at 8:57 pm

The reason that Cornell Lab uses eBird, is that it is a joint project between CLab and the Audubon Society. I think it runs on servers at CLab. Carl Clifford

mickramsey

June 18, 2015 at 7:46 pm

Agree with Martin here, some comments have been all about the inaccuracies if eBird data. A 2 minute search of Birdata data today revealed White-throated Honeyeaters, White-browned Treecreepers and Splendid Fairy-wrens in the eastern suburbs of Melbourne, completely out of range. When or will ever this inaccuracies be fixed by Birdata. They have been there at least 10 years to my knowledge. At least eBird has a dedicated and passionate group of reviewers that are consistently trying to ensure accurate data. Sure some things slip through but in large it’s pretty accurate. Birdata and the atlas to remain clunky and old in technology terms. EBird has just released a new app that is easy to use. Birdata and Birds Australia are not using a valuable resource here by not using Ebird data. Cornell use eBIRD it in the USA, why can’t Birds Australia? Michael Sent from my iPhone

martinflab

June 18, 2015 at 4:26 pm

There are also errors which creep in to Atlas data. I remember in particular noticing a record of Black Currawong in the Goulburn NSW area some time back. As I type the Birdata list for postcode 2620 (some rural parts of the ACT and Queanbeyan NSW) contains records for Red-winged Parrot, White-browed Treecreeper, Red-browed Pardalote, and Yellow-throated Scrubwren . To my knowledge all of these are just about impossible in the COG Area of Interest (roughly Cooma to Goulburn to Yass with Canberra more of less central). There are also a much larger of species that I don’t recollect ever being reported to COG from postcode 2620 – although have been recorded once or twice in other parts of the ACT – and are thus just about possible in the area. My point is not to dump on the Atlas but to point out that in any huge data set there are going to be a few outliers and its up to users to approach them with caution! Thinking about this and others matters to do with eBird and other data systems in Australia birding led me to compile a rather lengthy blogpost on the topic. Martin Martin Butterfield http://franmart.blogspot.com.au/ On 18 June 2015 at 10:31, Peter Shute < pshute@nuw.org.au> wrote:

June 18, 2015 at 10:31 am

Because the observer names are recorded, there’s probably nothing stopping researchers from excluding records from people who have entered what they consider to be incorrect ids, or only using records from people they consider to be reliable. Peter Shute

mickhhb

June 18, 2015 at 8:08 am

I think you’ve hit the nail on the head Graeme. It is more of a recreational birders’ tool and not one with any scientific rigour for data analysis. I mean, it just can’t be really and I don’t know if that has ever been its intention. eBird is also very attractive to ‘travelling’ birders and I am constantly seeing erroneous records appearing in lists. I assume these are most often made by birders in unfamiliar territory because it is often things like Corvids that get misidentified. Just this week we’ve seen someone reporting Masked Woodswallow from Sydney’s Eastlakes Golf Course (single bird in winter). I’ve seen this many times on other eBird lists where the bird was obviously a Black-faced Cuckoo-shrike. The same person has entered Olive Whistler and Banded Stilt from Western Sydney in the past week. Clearly they are misidentifications. In the Hunter I constantly see Little Ravens reported from places I know only Aussie Ravens occur. It must be very difficult for the very few moderators to keep on top of it all. I think the duplication of data issue has been discussed here earlier. Yes, the BirdLife portal is currently being developed and is advancing. Funding assistance has been provided from the NSW Twitchathon funds from last year. As far as I know it is still being decided how the eBird data will be used. Mick From: Graeme Stevens < gestev45@hotmail.com> To: “birding-aus@birding-aus.org” < birding-aus@birding-aus.org> Sent: Wednesday, 17 June 2015, 21:07 Subject: [Birding-Aus] The Atlas and eBird I also watch the discussion about eBird with interest and from my limited observation it can be great fun as a personal birding tool – but here’s some thoughts that may get me “flamed” – I hope not – just a contribution to make sure we use the right application for the right purpose and understand their limitations (in the Australian context). Please tell me if I am misguided. My main concern is around data integrity. I see the following on the eBird Australia site and that gives me comfort for the time being. i.e. that there is no data transfer yet to the Birdlife Australia Atlas – thank heavens!: “All data entered through Eremaea eBird will be passed onto the Atlas. We have not yet set up a transfer link between the Atlas and Eremaea eBird, but we are in discussions with them — so watch this space!” Do they intend eBird to eventually provide clean data to the Atlas? or will the Atlas staff have to apply the moderation because I see no evidence of adequate eBird moderation of data entry. It is a concern to me because as I see it, eBird is a great personal tool and very feature rich. Fabulous for personal recreational birding. However that is not necessarily compatible with the data quality required for a scientific data set used for EIS work (for a Birdlife fee) for example. I have read the eBird statement on filtering and reviewing data and that is fine and dandy but on my reading their filtering for species occurrence is at State level in Australia? Whoa!!! So if someone records Mallee Fowl for Bondi, the eBird filter wont find that one?? Maybe I am misinterpreting? I currently see regular duplicated data in eBird as each person naturally wishes to have a personal data set – so for example, three people out together generate three lists of the same or similar species for the same location. That of course may not matter if you are only interested in a binary record of species presence at a site. However some records I see entered, for areas I know well, would never get near the Atlas database without requests for verification and URFs. e.g. the Warriewood Wetlands in Sydney which is one of my Atlas “favourite sites” and is an eBird “hotspot” has a count of 10 Western Gerygone for a date in 2013. Now I know you never say never with bird behaviour and distribution, but I reckon that might have attracted the Atlas moderators attention and a question or three?? But then yes, they do occur in NSW and have been recorded in (generally western) Sydney! Graeme Stevens disclaimer:Long term Atlasser with lots of skin in that game! – and very pleased to be corrected on any of the above. I also recognise that the current version of the Atlas has real limitations as a personal tool and record – but that was not it’s original purpose. A new portal release is under development I understand.

gsd37

June 18, 2015 at 8:02 am

I am an eBird reviewer for my region in British Columbia, Canada. I don’t know how your regions are defined in Australia, but here, they are all relatively small geographically. Each region has its own reviewer, and each reviewer can change and adjust the filters as they apply to their own specific region. The filters are designed to filter out unexpected species, out of season occurrences, and unusually large numbers. Reviewers regularly see all the records that have been filtered for one reason or another. At this point they are NOT part of the database; they are flagged as INVALID. They remain that way until the reviewer has had time to evaluate any field notes or photos that might have been submitted with the record, or, contact the observer for those details. The reviewer must then decide, based on the evidence provided, whether or not to deem the record VALID or leave it as INVALID. I believe that if the reviewers do a good job, and the filters are set appropriately, no obviously erroneous data will get through. That’s a big IF, though! It means that first the country has to be divided into small enough regions to be properly managed, and second, there needs to be enough people willing to put their hand up to act as a reviewer. I know that when I took on this region, there were already filters in place – presumably put in on a much broader scale than appropriate. I have had to make adjustments to many of my filters, and I continue to ‘tinker’ with them on an ongoing basis. If there is a weakness to eBird, it is perhaps not in the system itself, but rather in the shortage of people willing to take on a region and then to give it the time required to make it effective.Gary DavidsoneBird ReviewerCentral Kootenay Region From: Graeme Stevens < gestev45@hotmail.com> To: “birding-aus@birding-aus.org” < birding-aus@birding-aus.org> Sent: Wednesday, June 17, 2015 4:07 AM Subject: [Birding-Aus] The Atlas and eBird I also watch the discussion about eBird with interest and from my limited observation it can be great fun as a personal birding tool – but here’s some thoughts that may get me “flamed” – I hope not – just a contribution to make sure we use the right application for the right purpose and understand their limitations (in the Australian context). Please tell me if I am misguided. My main concern is around data integrity. I see the following on the eBird Australia site and that gives me comfort for the time being. i.e. that there is no data transfer yet to the Birdlife Australia Atlas – thank heavens!: “All data entered through Eremaea eBird will be passed onto the Atlas. We have not yet set up a transfer link between the Atlas and Eremaea eBird, but we are in discussions with them — so watch this space!” Do they intend eBird to eventually provide clean data to the Atlas? or will the Atlas staff have to apply the moderation because I see no evidence of adequate eBird moderation of data entry. It is a concern to me because as I see it, eBird is a great personal tool and very feature rich. Fabulous for personal recreational birding. However that is not necessarily compatible with the data quality required for a scientific data set used for EIS work (for a Birdlife fee) for example. I have read the eBird statement on filtering and reviewing data and that is fine and dandy but on my reading their filtering for species occurrence is at State level in Australia? Whoa!!! So if someone records Mallee Fowl for Bondi, the eBird filter wont find that one?? Maybe I am misinterpreting? I currently see regular duplicated data in eBird as each person naturally wishes to have a personal data set – so for example, three people out together generate three lists of the same or similar species for the same location. That of course may not matter if you are only interested in a binary record of species presence at a site. However some records I see entered, for areas I know well, would never get near the Atlas database without requests for verification and URFs. e.g. the Warriewood Wetlands in Sydney which is one of my Atlas “favourite sites” and is an eBird “hotspot” has a count of 10 Western Gerygone for a date in 2013. Now I know you never say never with bird behaviour and distribution, but I reckon that might have attracted the Atlas moderators attention and a question or three?? But then yes, they do occur in NSW and have been recorded in (generally western) Sydney! Graeme Stevens disclaimer:Long term Atlasser with lots of skin in that game! – and very pleased to be corrected on any of the above. I also recognise that the current version of the Atlas has real limitations as a personal tool and record – but that was not it’s original purpose. A new portal release is under development I understand.

Birding-Aus

The Atlas and eBird

10 comments to The Atlas and eBird

Mailing List

Book Search

Meta