As many of you would know, there are many Australian species on iNat that have their locations automatically obscured; at time of writing, the total number of species is ~2,800.
As a brief aside, the 'true' number is actually a bit lower than that, possibly several hundred lower. This seems oddly vague/inaccurate, but it's due to a mechanism on iNat to prevent bad actors unobscuring individual observations. Consider the following situation:
Someone uploads an observation of something with a sensitive location, it gets rightfully obscured. Three bad actors come along and each intentionally misidentify the observation as a common, unobscured species, eg a feral pigeon. The overall ID now shifts to said common species, and so the location gets unobscured, and they can access the sensitive location.
To combat this behaviour, iNat is setup so that if an ID of a sensitive species is added to an observation at any time, even if it is wrong + overruled or later withdrawn, the observation will remain obscured as a safeguard. So at least some of the species being included in that 2,800 is not because the actual species itself is obscured for Australia, but rather because there is at least one individual observation with at least one wrong ID (but now corrected) of an obscured species from somewhere else in the world. A classic example is the sulphur-crested cockatoo. Currently, 41 observations of this species are being automatically obscured in Australia. This is not because this species has any obscuration applied whatsoever, but because, on each of these observations, someone mistakenly added an ID of yellow-crested cockatoo (a threatened species from Southeast Asia obscured on iNat). All of these observations have since been correctly identified, but that initial wrong ID still applies obscuration.
But even accounting for that, take home message is that a lot of species are currently being obscured in iNat within Australia. Whilst many of these are being obscured based on state or federal sensitive species lists, there are also many that are being obscured only because of IUCN-based statuses. In many cases, these statuses do not correspond at all to Australian state or federal statuses as our conservation bodies/instruments and the IUCN apply different methodologies. On top of that, a lot of species being obscured due to IUCN statuses shouldn't actually have hidden locations, eg there are some Western Australian plant species that have an IUCN status of Near Threatened or Vulnerable despite being quite common and widespread, and having no federal or state status.
There has to be a reason for obscuring locations. If a species is threatened by collection, poaching, etc., then absolutely obscuration is a useful tool and will be implemented. But for other species, the threats are completely unrelated to these factors. For some species, the threats are clearing of habitat due to urban development, stochastic events such as fires due to small population size, increasing temperatures or sea levels due to climate change, etc. So obscuring the locations would not actually ameliorate any of these threats, as they are entirely unrelated to whether locations are public or not. Indeed in cases like these, it would almost certainly actually be detrimental to the species to obscure its locations on iNat. Consider a situation where eg a block of bush is being cleared for a housing development. Someone finds a species there and uploads an observation to iNat, but the location gets obscured, and by the time someone gets around to getting access to the true location, the vegetation has been cleared because no-one knew it was there due to the obscuration.
So obscuration should always be justified. As the main part of my job at the ALA, I'm working with iNat data on both the iNat and ALA ends. Under the National Framework for Restricted Access Species Data, we're trying to act consistently with the framework and make iNaturalist consistent with the state and territory sensitivity lists. So over the next few weeks, I will be opening the locations for all species on iNat where Australian observations are being automatically obscured due to an IUCN status only. Some of you may have noticed I have already done this for a small handful of species, mostly Tasmania stuff (see https://www.inaturalist.org/journal/thebeachcomber/86826-deobscuring-tasmanian-species).
As a few extra points of clarification:
1 . Perhaps the most important point here is to clarify how obscuration works on iNat just for the benefit of those who aren't completely familiar with the process.
a) When an observation is obscured on iNat, the true coordinates entered by the observer are randomly scrambled into a ~500 sq. km box around them. Any other user viewing that record will see the randomly generated coordinates with a box around them; they know the true coordinates are somewhere inside that box, but not the actual location. In addition to the coordinates, the date and time are also removed from the observation, so that only the month and year are displayed. Further, any identifications added to the record will have their timestamp altered to also only show the month and year.
b) When these records get exported to the ALA each week, the obscuration accompanies them, so the records are also obscured in the ALA, ie the coordinates shown in the ALA are the randomly generated ones, and the coordinate uncertainty is listed as somewhere around 29,000-30,000 m.
c) There are many different channels for getting access to the true coordinates if you're a researcher, land manager etc. I won't list them all here, but they include having users 'trust' you on iNat (can be turned on in profile settings) or directly requesting the data from the ALA (the true coordinates do go into the ALA, it's just that they are not made publicly available). So locations of obscured records are by no means lost, just more difficult to access. Also, when updates to coordinates/obscuration are made on iNat, these changes are automatically reflected in the ALA after a week or so.
d) There are two different types of obscuration on iNat.
i . Geoprivacy refers to users manually obscuring their own records. This is done at an individual observation level.
ii . Taxon geoprivacy refers to the automatic obscuration of records by the system. This is done at a taxon level.
2 . Crucial point: any records that you have manually obscured have not been deobscured. This only applies to species that were getting automatically obscured by the system, and only those being affected by an IUCN status.
Comentarios
Thank you for explaining all the above. I have been asking others and no one seems clear about much of it. Excellent article.
Thanks for this post! I have a quick question: Are the true coordinates of manually-obscured records available to approved users via the ALA? Or can they only access records obscured by taxon geoprivacy?
my original understanding was yes, no matter the circumstance. But based on a recent discussion I was part of, it seems like there are some specific cases where no, true coordinates of manually-obscured records are not always available to approved users via the ALA. I need to confirm this though
I've now finished this project. Just in case there are some species that some people strongly believe should have remained obscured, or for special cases like eg https://www.inaturalist.org/flags/643838, please comment here with the species in question and we can discuss/reassess. Because this project has affected Australian species in general rather than specific taxa or states, I'm just tagging the top IDers and observers that are from Australia (I can only tag so many people without spending hours), so please tag anyone else who you think should be notified
Here are the species that I assessed that you can check: https://docs.google.com/spreadsheets/d/1-v5-KcigUXfkTuKbTdyWEAihxrNIGiFi/edit?usp=sharing&ouid=109399214372663799877&rtpof=true&sd=true
Please note the emphasis on assessed; many of these species are still obscured, either across Australia or in the relevant states.
@george_seagull @reiner @bushbandit @benkurek__ @leoncrang @louisb @cesdamess @ben_travaglini @sofiazed1 @alx4mtmel @bwjone432155 @twan3253 @baronsamedi @matthew_connors @michaelcincotta @alan_dandie @suzieandjim @gumnut @torhek @deborod @meta4 @pewin @scottwgavins @jadonald @kallies @wildroo @tony_d @christopherburwell @iancastle @pennywort_man @joswan12 @ellurasanctuary @elizabethhatfield @w_martin @gregtasney @ratite @eremophila @rfoster @chuditch @vireyajacquard @dustaway @nicklambert @ethmostigmus @jackiemiles @cobaltducks @adrian2370 @gposs @kevinbonham @davidsando @imcmaster @cinclosoma @mononymous @davemmdave @wellsii @saltmarshsteve @daviaker @kenharris @natashataylor @martinbennett @dianneclarke @domf @martin487 @triciastewart @mattcampbellaus @rattyexplores @larney @douch @possumpete @tjeales @kim-tarpey @arthur_chapman @dj_maple @suzieandjim @hatwise @suzieandjim @a_kurek @jmartincrossley @paul2george @tonydiver @dianadavey @elusiveorchids @eschlogl @pardalotebellion @peregrine80 @hdavid @russellcumming @cher63 @em_lamond @ethan241 @harryrosenthal @ladydawn @thistlemouse
Hi Thomas,
Thanks for the article, it’s good that you guys take the pro-active approach to obscuring protected species.
Just a note on the example of it being detrimental, when we do desktop-surveys before any on site surveys (for proposed wind farms in my case), usually we have a 50km buffer/radius around the project to formulate an initial species list. So there’s still a reasonable chance that obscured species will be picked up and known ahead of time.
The problem is that developers and government usually don’t give a crap even if you do find these things :(
Very interesting, thank Thomas 🙂
These three species had locations obscured at species level in QLD.
Ninox strenua – Photographers disturbing the birds at roost sites during the day. Particularly, metropolitan roosts.
Petauroides Volans – Spotlighting localised populations including regular paid tour guide visits.
Petaurus australis - Spotlighting localised populations including regular paid tour guide visits.
@gregtasney Ninox strenua is obscured across the board on iNat, no change was made there
Petaurus australis is also already obscured for QLD
Petauroides volans was unobscured for QLD, but I have now obscured it; QLD site says it is sensitive, so the previous change (not me) that set that to open was mistaken
Very infomative, thanx
Thanks for the info, I'm fine with it, don't have anything to add.
Hi Thomas, appreciate all of your hard work and dedication to this matter. Always good to know the birds and animals come first with obscuring locations. Thank you for all you do.
As I was reading it I was thinking "I know a few who obscure their locations for other reasons" but good to see than those manually obscured will be left as such.
As for the rest, there's sone odd ones on the list that you linked to. Maybe they were just obscured at State level?
Thanks Thomas for the post - valuable. Obscuring should also best be used in conjunction with locality/region, etc. For example, your mention of yellow-crested cockatoo above - that species should not be obscurred in Australia. Location (via a region, for example) is important because a species may be endangered/threatened in one area (and thus the location should be obscurred), but be a weed/pest in another area or region (especially in another country) where the actual locations may be important for control.
The other issue I'd like to raise is randomization versus generalization. For details and reasons see my paper where I argue for generalization over regionalization - see Chapman AD (2020). Current Best Practices for Generalizing Sensitive Species Occurrence Data. Copenhagen: GBIF. https://doi.org/10.15468/doc-5jp4-5g10. For non-Australians, that paper is also available in Spanish and French.
Thank you for this explanation and for all your work involved.
Very interesting article, from which i have learned a great deal. Knowing that "trusted researchers" can always access the true locations is a real comfort. Thanks for the detailed article.
@mattcampbellaus which ones were you thinking about Matt
The two that really stood out Thomas were Pogona barbata and Phaps chalcoptera. They were just names that popped out at me as I scrolled down the page.
@ricardosimao
@tere_c
Pogona barbata is currently considered a sensitive species by the Victorian Government (https://www.environment.vic.gov.au/__data/assets/pdf_file/0024/48831/VBA-Restricted-Taxa.pdf) so has its observations obscured on iNat for Vic (that one was a surprise to me)
Phaps chalcoptera is only on the list because it has an IUCN status, but it's least concern and not obscured (the list I linked is everything that had a global status of some kind on iNat, both obscured and open)
Thanks for the information
Making you aware of changes to taxon-specific geo-location obscuring.
@terra_australis @xanthosia @deborahec @cumberlandplain @porcoespinho15 @rwl
No worries Thomas, thanks for the clarification. I had heard that barbata was considered to be in decline, especially in Vic, but no idea that it had any sort of listing as of yet.
some extra relevant information for everyone relating to taxa that get auto-obscured on iNat as they have sensitive locations (= 'taxon geoprivacy') vs. individual observations being manually obscured by users (= 'geoprivacy')
If you are an iNat user who has uploaded observations from Australia, but you haven't affiliated your account with the iNaturalist Australia node, then the ALA will receive the true locations only for your observations obscured via taxon geoprivacy. The ALA can then share these with researchers. The ALA will not even receive the true locations of your manually obscured records at all.
If you have affiliated your account with iNaturalist Australia, the ALA will receive the true locations of your observations that are obscured both automatically and manually
BUT, even though the ALA does store the true locations of manually obscured records on their backend, they do NOT share these with researchers etc. They only share the true locations of taxon geoprivacy-affected species.
Interesting stuff thanks Thomas - I have always thought iNat approach to obscuring was a bit haphazard. Generally speaking I am fairly anti-obscuring for most species, but fully recognise there are instances where it is valuable - I appreciate you pointing out that for some species / instances it may actually be detrimental.
Bird-wise I would suggest that at a minimum iNat should obscure the same species as eBird makes sensitive in Australia, namely Orange-bellied Parrot, Night Parrot, Plains-wanderer and Double-eyed Fig-Parrot (Coxen's). The latter I can't imagine receiving any iNat observations any time soon so probably not much of an issue. I can also see an argument being made for species that are especially vulnerable to disturbance at roost/nest sites, i.e. rare raptors and owls, although personally I don't believe that this action would have a huge impact. Most other species (e.g. shorebirds) I don't think there is any reason to obscure.
Butterfly-wise, I can't think of particularly many taxa that I think definitely should be obscured. A few taxa (Swamp Tiger the best example off the top of my head) currently are obscured and I think need not be. Maybe some of our most endangered and range-restricted taxa should be obscured, but equally for the most part these species' locations are well known already and obscuring iNat records probably has very little impact. Species like Bulloak Jewel etc.
Perhaps @brandon_hewitt @braden1 @pewin have opinions on the butterflies?
thanks Louis
Swamp tiger is now unobscured for Australia (that was one of the ones I opened). The only butterfly currently being obscured in iNat is Ogyris subterrestris (but only for South Australia)
Orange-bellied parrot: obscured NSW, SA
Night parrot: obscured in NT, WA, QLD
Plains-wanderer: obscured NSW, VIC
Double-eyed Fig-Parrot: obscured NSW
Añade un comentario