eBird and Birding on the move (1 Viewer)

[raymie]

I did not even know such a thing existed? is there a way to search it?



Frankly, “the official list of species” is not scientific, because science, essentially, accepts diversity and change of opinions.


In case of mammals, hypothetically, it is enough to mark records as unreviewed unless somebody reviews them.

Actually, one of my objections to large public sighting databases is the lack of verification or no traceable verification, which essentially makes them pseudo-science. For the user, this is not a visible problem, until somebody sees a paper based on ebird which lists a population of 10s or 100s of Ivory-billed Woodpeckers

I perhaps should sit and write a post “how to make ebird and others truthful”.

The Ivory-bill was a legitimate record in the database, just a historical one. It was a problem with the model the authors of the paper used to estimate the numbers, not with eBird.

 

[THE_FERN]

Talking about taxonomy and then isolating a group of macrolepidoptera and calling them butterflies grates with me a touch....

Within lepidoptera generally, I think the state of taxonomic flux is probably very fluid & that is before you get to other invertebrate groups.

All the best

Paul

Taxonomy for any group changes with the number of studies and studiers (since all taxonomy is subjective: you impose categories on something which is in fact a continuum). There are far fewer entomologists and botanists than ornithologists so fewer studies. Therefore the taxonomy is more stable. There are many insect groups where the taxonomy hasn't changed since they were first described (often 19th cen.).

Yes, taxonomy of butterflies (or all leps if you like---"butterflies" have been studied far and away the most) has changed and is changing/will change. But there are not 5+ quite different estimates of the total number of world species---as there are for birds because there are far fewer lepidopterists.

 

[THE_FERN]

I mean, most taxonomic groups don't have standardized lists to draw from,

Actually increasingly not true. Of course global lists do exist for pretty much everything---it's just they're in paper databases. But inaturalist, encyclopaedia of life, gbif, observation.org, iGoTera, Species Files etc etc show it's possible and in some cases show how to do it.

...updated once a decade or even less

Taxonomy changes as a function of taxonomists... ...The positive is you can concentrate on other aspects of the organisms instead because you have taxonomic "stability"

...wrong

To misquote: all taxonomy is wrong but nearly all taxonomy is useful (why we use it). The same goes for "pseudo-science" arguments around crowd-sourced data. Truth is some data beats no data every time even if the "some data" is somewhat wrong.

(Also, everyone makes mistakes. Who is the infallible expert who is going to perfectly correctly identify everything they encounter---in any group of organisms? Such a person does not exist)

 

[THE_FERN]

there new exotic protocol? That's way more straight forward.

Don't get me started on the stupidity of this...

 

[Paul Chapman]

Taxonomy for any group changes with the number of studies and studiers (since all taxonomy is subjective: you impose categories on something which is in fact a continuum). There are far fewer entomologists and botanists than ornithologists so fewer studies. Therefore the taxonomy is more stable. There are many insect groups where the taxonomy hasn't changed since they were first described (often 19th cen.).

Yes, taxonomy of butterflies (or all leps if you like---"butterflies" have been studied far and away the most) has changed and is changing/will change. But there are not 5+ quite different estimates of the total number of world species---as there are for birds because there are far fewer lepidopterists.

This is simply not my experience. Far far more species & far far more taxonomic change.

i am not sure what personal experience that you have of entomology. Mine is relatively limited but I know a few entomologists including a compiler of one of the databases to which you refer. I also know simply from the number of my own sightings where the name has changed repeatedly within the last decade. That is before you get to the extent of unnamed taxa.

So I am at a loss as a result to understand what authoritative basis there is for your comments.

If I had access to it as I type, I would refer you to the Aggasiz et al list of British Lepidoptera and some of the repeated changes.

(There are 50 times the number of moths in Britain as butterflies.)

All the best

Paul

 

[THE_FERN]

i am not sure what personal experience that you have of entomology.

Well I have a doctorate which was largely entomological, and I'm a visiting researcher at the NHM London. So I'd like to think I know a bit.

Yes there are lots of entomological taxonomy studies—more than ever before—and each will likely lead to changes for the reasons I've noted. However, if you want to find groups which +/- no-one (or only one or two) people in the world work on, then start with insects or other invertebrates. You might be the sole taxonomist working on them. Then the official taxonomy can be yours and yours alone.

I should note that I'm talking about taxonomy overall. Higher level taxonomy (families and above) is much more liable to change in any group. That's because a) it's relatively simple to get a bunch of representative taxa which can stand in for the groups lower down, b) your study will have big taxonomic implications (probably) and c) molecular techniques are now such that we might believe you. So until fairly recently there were quite a few alternative arrangements for the families of the plants of the world.

Anyway, the broad point is that the more taxonomists, the more proposed changes and the greater instability.

Meanwhile, happy for you point me to the various dissenting lists of the butterflies of the world. I'm aware of Gerardo Lamas's work of course, but I'm less clear there are any real alternatives (i.e. contra IOC, Clements, HBW etc etc for birds). [I do know there are couple of quite different opinions about European species though: generally molecular study-related splits afaik]

Edit: butterflies are a reasonable group to think about here as they've roughly the same no of species as birds, and their diversity seems to be driven by the same forces—in South America there's a broad equivalence between nos of species on a country-by-country basis, for example

 

[Paul Chapman]

Well I have a doctorate which was largely entomological, and I'm a visiting researcher at the NHM London. So I'd like to think I know a bit.

Yes there are lots of entomological taxonomy studies—more than ever before—and each will likely lead to changes for the reasons I've noted. However, if you want to find groups which +/- no-one (or only one or two) people in the world work on, then start with insects or other invertebrates. You might be the sole taxonomist working on them. Then the official taxonomy can be yours and yours alone.

I should note that I'm talking about taxonomy overall. Higher level taxonomy (families and above) is much more liable to change in any group. That's because a) it's relatively simple to get a bunch of representative taxa which can stand in for the groups lower down, b) your study will have big taxonomic implications (probably) and c) molecular techniques are now such that we might believe you. So until fairly recently there were quite a few alternative arrangements for the families of the plants of the world.

Anyway, the broad point is that the more taxonomists, the more proposed changes and the greater instability.

Meanwhile, happy for you point me to the various dissenting lists of the butterflies of the world. I'm aware of Gerardo Lamas's work of course, but I'm less clear there are any real alternatives (i.e. contra IOC, Clements, HBW etc etc for birds). [I do know there are couple of quite different opinions about European species though: generally molecular study-related splits afaik]

Edit: butterflies are a reasonable group to think about here as they've roughly the same no of species as birds, and their diversity seems to be driven by the same forces—in South America there's a broad equivalence between nos of species on a country-by-country basis, for example

I am genuinely puzzled. This is just massively irrelevant to eBird trying to run lists of other taxa which are constantly changing at species level?!?

All the best

Paul

 

[THE_FERN]

I am genuinely puzzled. This is just massively irrelevant to eBird trying to run lists of other taxa which are constantly changing at species level?!?

All the best

Paul

compared with birds they're simply not. And plenty of other systems let you record those other things as you go [i.e. +/- like ebird but including other taxa]

Edit: also to be clear, as long as the taxon and date are recorded, then it doesn't matter if the taxonomy changes later. That's only important if you want to summarise the information against the new taxonomy. It's not important from the point of view of collecting the data "on the move" [thread title]

 

[Paul Chapman]

compared with birds they're simply not.

Not my experience at all as alluded to here...

 

[opisska]

Ebird's data model is wrong which doesn't help. Ideally it'd record the location every time you pressed the button (with the provision to record placeholders like "bird 1", "bird 2" etc for things you id later). With a "background mode" the app could be always on and recording incidentals would be quicker, easier and locationally more accurate.

Instead, ebird ties things to "transects" or "hotspots", a collection of one or more bird sightings. This makes the location of any particular bird within the session uncertain, and for long sessions muddles the location so much it can be meaningless (why people tend to add additional coordinates).

As it is, I do record incidentals but it's difficult from a speeding car because of the start up time for the app (dunno why it has to get your location before you can record a sighting: could get location seperately in the background). I know people who operate mainly in incidental mode, and thiys have records which are closer to the ideal with much better spatial resolution

Wait. So when I see my friends walking with the app on and recording each bird as they register those, this still doesn't get posted to eBird as individual locations, but it get bunched up in one point? Does that mean that all the individual points for birds are a) only when the observe went an extra mile and made it a separate observation or b) are there some average representation of the moving checklists?

I really don't understand what is the motivation of eBird of muddling the locations of the observations. The entire "hotspot" business is so counter-productive. Yeah, yeah, "it allows you to see the diversity of a place" - the same would work if you could just draw a small circle on the map and get a list of species in the circle - but the data access of eBird is made as obtuse as possible for some reason.

It's really a shame that this ended up being the one most popular platform.

 

[raymie]

Hotspots are intended for relatively small, easily definable areas, at which point hiding the location doesn't really matter, since everyone who birds in these locations would have a very similar track. The problem arises from shoddy hotspot reviewing that is allowing hotspots to be created for large or undefinable areas.

Please use the hotspots that are the way they were meant to be, so people can easily track specific locations. But know it isn't the right time to use a hotspot and not to use bad hotspots.

 

[Paul Chapman]

Wait. So when I see my friends walking with the app on and recording each bird as they register those, this still doesn't get posted to eBird as individual locations, but it get bunched up in one point? Does that mean that all the individual points for birds are a) only when the observe went an extra mile and made it a separate observation or b) are there some average representation of the moving checklists?

I really don't understand what is the motivation of eBird of muddling the locations of the observations. The entire "hotspot" business is so counter-productive. Yeah, yeah, "it allows you to see the diversity of a place" - the same would work if you could just draw a small circle on the map and get a list of species in the circle - but the data access of eBird is made as obtuse as possible for some reason.

It's really a shame that this ended up being the one most popular platform.

Hotspots are intended for relatively small, easily definable areas, at which point hiding the location doesn't really matter, since everyone who birds in these locations would have a very similar track. The problem arises from shoddy hotspot reviewing that is allowing hotspots to be created for large or undefinable areas.

Please use the hotspots that are the way they were meant to be, so people can easily track specific locations. But know it isn't the right time to use a hotspot and not to use bad hotspots.

Surely, the purpose of eBird is not to allow non eBird users to find a specific species. I have no interest in the exact locations of individual robins or woodpigeons.

It needs to be a useful tool for individual's bird recording which encourages people to use it thereby maximising the data collected and allowing the collateral benefits of the scientific recording that falls out of the data.

Raymie's point on the size of Hotspots is well made but the reality is that on occasion, it simply suits an individual's recording to keep a list for what they see in a day or afternoon or slightly longer drive or walk.

All the best

Paul

 

[opisska]

I still fail to see the benefit of intentionally discarding information. If you want to collect data, why go out if your way to make sure you don't collect them precisely?

As for "the purpose of eBird is not for people to find birds" - this only reinforces my statement that it's a shame that this is the platform that most people choose to record birds. We could have had a platform which would be good for finding birds and there would be no harm to the actual recording done by this, but sadly, eBird has sabotaged this.

 

[THE_FERN]

Wait. So when I see my friends walking with the app on and recording each bird as they register those, this still doesn't get posted to eBird as individual locations, but it get bunched up in one point? Does that mean that all the individual points for birds are a) only when the observe went an extra mile and made it a separate observation or b) are there some average representation of the moving checklists?

Yes.

As far as I know, any GPS track is not used (except, possibly to derive a single representative location to which to assign the checklist, and distance travelled).

It makes ebird very arbitrary with respect to birds' precise actual locations. If distance and time are accurately recorded you can probably derive a broadly accurate surface to represent the probability of occurrence. Maybe you could combine that with actual habitat info to guess where the bird occurred, at least if habitat occurs in Islands (like in cities), and if the (habitat) island isn't too big. But it would be a lot better to actually record where the bird was !

 

[raymie]

Surely, the purpose of eBird is not to allow non eBird users to find a specific species. I have no interest in the exact locations of individual robins or woodpigeons.

It needs to be a useful tool for individual's bird recording which encourages people to use it thereby maximising the data collected and allowing the collateral benefits of the scientific recording that falls out of the data.

Raymie's point on the size of Hotspots is well made but the reality is that on occasion, it simply suits an individual's recording to keep a list for what they see in a day or afternoon or slightly longer drive or walk.

All the best

Paul

Absolutely, one of the main purposes of eBird is allowing people to find birds. If used correctly, eBird does this extremely well. The problem is people don't know how to submit to eBird properly.

"Day lists" are explicitly not allowed in eBird, as are checklists that cover a distance longer then 5 miles.

 

[THE_FERN]

"Day lists" are explicitly not allowed in eBird, as are checklists that cover a distance longer then 5 miles.

A silly and arbitrary limit and, unless things have changed, perhaps not correct. As I've argued before, it doesn't matter what distance a checklist covers if you insist on doing things the ebird-data-model way (i.e. wrongly). Longer distances still yield useful data, it's just less precise.

It's annoying because sometimes you simply can't chunk your lists easily. I had that in Masai Mara where we had a continuous drive which covered perhaps 200 km. (If doing again I might try to record hundreds of 'incidentals' instead).

In this case, I couldn't submit the checklist via the app as the GPS trace was too long (this wasn't documented, and it unhelpfully just said "there was an error..."). Thankfully I know how to get the GPS trace out of the underlying database on my phone or else it would be lost forever. I submitted the list via spreadsheet upload instead, after assigning all my records to the "Masai mara general area" hotspot. This ebird user experience is wrong on so many levels, I almost don't know where to begin...

 

[THE_FERN]

Longer distances still yield useful data, it's just less precise.

E.g. if you were deriving a density surface as part of a '"probability of encountering" calculation, you'd just weight each record by the inverse of the distance travelled during the checklist (so less precise locations contribute less to the final estimate). For a given record, you still have some useful info ("seen somewhere in the Masai Mara"), but again it's just not as precise.

These problems don't apply if you use the correct data model---which is to record where each thing is seen and not condense the data down arbitrarily

(Note that any checklist distance limit doesn't help at all. A circle of circumference 5 mi is different to a straight line of distance 5 mi for example)

 

[THE_FERN]

Absolutely, one of the main purposes of eBird is allowing people to find birds.

If you want to do that, record where the bird was actually seen !

...Not where some random hotspot is...

 

[jurek]

I actually see

The same goes for "pseudo-science" arguments around crowd-sourced data. Truth is some data beats no data every time even if the "some data" is somewhat wrong.

Not, it is a common mistake. Small data usually gives more accurate picture than large but skewed data, unless the bias is understood and corrected.

I perhaps really should write more about it, because Ebird and field biologists are happily unaware of many problems of working with data. In e.g. medicine, work with biased data can cost an arm and a leg. Literally. Not that it is a finished topic.

entomological

butterflies

I agree that non-birds should be recorded, even with the lower quality of taxonomy and identification sources. It will be poorer for decades in the future. Is this a reason not to collect any data?

At the very least, one can reliably record unmistakable species and observations with voucher photos - probably numerically, majority of records. More imaginatively, a form asking to describe how a critter / plant was identified (eg. a photo or ID marks) can be automatically posted, like ornitho.ch does for rare birds.

BTW, poor identification is a problem for birds, too. In ebird or ornitho I regularly see wrongly labelled photos. A handy button to report a suspected mistake could be handy.

 

[opisska]

Well for non-birds, there really is iNaturalist. For birders, it's not practical, because it asks for a photo for each observation, but for hard-to-ID things like most insects, this is the best approach if you want to have some reasonability in the data. So there really isn't much to invent in this direction.