Adding cultivated species to algorithm?

[Charlie Hohn]

My understanding is captive/cultivated observations never get added to the algorithm even if they are the equivalent of research grade.

I have two reasons I think they should be included:

Firstly, i think it is resulting in more wrong IDs. for instance, Canary island pine is probably the most common pine planted in the LA area. It's also very distinctive. However these keep getting identified as Ponderosa pine which looks vaguely similar... i think it's because most obs of Canary island pine are marked as captive (only native to the Canary islands where there aren't enough iNat observations yet)... so it either isn't in the algorithm or is too poorly represented to get picked up. This results in lots of wrong IDs of Ponderosa pine in LA, when in reality it doesn't seem to survive in LA even when planted (it occurs in high mountain areas nearby). For instance https://www.inaturalist.org/observations/10190945

Secondly.. i know we don't encourage adding cultivated things to iNat, but it's fun and easy to add those to Seek, and that is a harmless activity that is likely to be a nice gateway to kids because they are so easy to find and observe. Adding captive/cultivated plants would make Seek work better and be more fun.

I know it's a bit tricky since they aren't tracked as research grade. Maybe they should be, but that's a whole other discussion. but either way I think they need to go into the algorithm!

[Tony Iwane]

A well written proposition, Charlie - what do others think?

Tony Iwane

[bouteloua]

Yes please! Computer Vision for captive/cultivated species, in combination with the way the system votes on commonly captive/cultivated spp., would greatly improve my Identify experience, as I'm mainly interested in wild organisms.

cassi

[Scott Loarie]

Just to clarify if a species is included in the model (our threshold is at least 20 Research Grade observations by at least 20 different people) then we do use captive/cultivated obs of that species to train the model.  eg. the houseplant Golden Pothos is in the model because its invasive in Hawaii etc. so iNat has lots of RG obs of it, but the model also uses the captive/cultivated data to train.

But a species only represented by captive/cultivated obs wouldn't yet be in the model.

--
You received this message because you are subscribed to the Google Groups "iNaturalist" group.
To unsubscribe from this group and stop receiving emails from it, send an email to inaturalist+unsubscribe@googlegroups.com.
To post to this group, send email to inatu...@googlegroups.com.
Visit this group at https://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

--

--------------------------------------------------
Scott R. Loarie, Ph.D.
Co-director, iNaturalist.org
California Academy of Sciences
55 Music Concourse Dr
San Francisco, CA 94118
--------------------------------------------------

[Charlie Hohn]

aha, i see, because the canary islands sadly don't have a ton of iNat yet it probably doesn't have 20 Rg observations of this species. I feel like if possible the algorithm should create a proxy 'research grade' for cultivated plants that have 20+ observations with community ID. But maybe that is very confusing and hard to program, i'm not sure. Overall newbies and people using Seek are the most likely to rely on the algorithm and also the most likely to enter cultivated plants.

On Monday, March 26, 2018 at 2:26:21 AM UTC-4, Scott Loarie wrote:

Just to clarify if a species is included in the model (our threshold is at least 20 Research Grade observations by at least 20 different people) then we do use captive/cultivated obs of that species to train the model.  eg. the houseplant Golden Pothos is in the model because its invasive in Hawaii etc. so iNat has lots of RG obs of it, but the model also uses the captive/cultivated data to train.

But a species only represented by captive/cultivated obs wouldn't yet be in the model.

On Sun, Mar 25, 2018 at 8:40 PM, bouteloua <cassi...@gmail.com> wrote:

Yes please! Computer Vision for captive/cultivated species, in combination with the way the system votes on commonly captive/cultivated spp., would greatly improve my Identify experience, as I'm mainly interested in wild organisms.

cassi

On Sunday, March 25, 2018 at 10:20:04 PM UTC-5, Tony Iwane wrote:

A well written proposition, Charlie - what do others think?

Tony Iwane

On Sunday, March 25, 2018 at 6:08:43 AM UTC-7, Charlie Hohn wrote:

My understanding is captive/cultivated observations never get added to the algorithm even if they are the equivalent of research grade.

I have two reasons I think they should be included:

Firstly, i think it is resulting in more wrong IDs. for instance, Canary island pine is probably the most common pine planted in the LA area. It's also very distinctive. However these keep getting identified as Ponderosa pine which looks vaguely similar... i think it's because most obs of Canary island pine are marked as captive (only native to the Canary islands where there aren't enough iNat observations yet)... so it either isn't in the algorithm or is too poorly represented to get picked up. This results in lots of wrong IDs of Ponderosa pine in LA, when in reality it doesn't seem to survive in LA even when planted (it occurs in high mountain areas nearby). For instance https://www.inaturalist.org/observations/10190945

Secondly.. i know we don't encourage adding cultivated things to iNat, but it's fun and easy to add those to Seek, and that is a harmless activity that is likely to be a nice gateway to kids because they are so easy to find and observe. Adding captive/cultivated plants would make Seek work better and be more fun.

I know it's a bit tricky since they aren't tracked as research grade. Maybe they should be, but that's a whole other discussion. but either way I think they need to go into the algorithm!

--
You received this message because you are subscribed to the Google Groups "iNaturalist" group.

To unsubscribe from this group and stop receiving emails from it, send an email to inaturalist...@googlegroups.com.

To post to this group, send email to inatu...@googlegroups.com.
Visit this group at https://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

[Colin Meurk]

Can I please change the direction of this a little and go back to the broader question of legitimacy of cultivated/domesticated/planted organisms as real biological information and therefore deserving of research grade status.

As an ecologist (and urban ecologist) I would declare that all organism observations are valid pieces of information about where something can live (whether it's a house plant, a spider in a house, or an epiphytic orchid in a tropical rain forest). The key information required to make sense of the observation/location (in all cases) is the nature of the environment (and history). Yes, a tropical plant can only grow in Alaska if it is under artificially heated cover; but it IS THERE and we can figure out why and a record will provide useful knowledge to some of the community. Remember this is citizen science - there are both citizens and scientists to be satisfied. We need to educate users to record what the circumstances of a record are; but it is usually fairly obvious and often can be added after. The problem with making an arbitrary decision, that 'wild' is good and 'planted' is bad, is that it brushes over all kinds of grey areas like introduced species (whether they are deliberate or hitch-hikers) that become naturalised and maybe hybridise with local 'natural' species. But it will also encourage people who don't like seeing their doubly identified records (entered in good faith and failing to be registered as research grade) to withhold important information (in other words decide not to tick the 'cultivated' box). It devalues the work that observers and identifiers put into records to not see that 'research grade' status come up. I repeat, it is a totally arbitrary decision with no ecological foundation to prescribe that garden plants, street trees, eco-restoration plantings, etc are not valid biological information.  It should be made also clear that GBIF has no problem whatsoever with such records provided they are appropriately labelled. It has nothing to do with whether the data is useful for research or not. So, the status name itself (Research Grade) is misleading right there. The other important consideration is that often it is cultivated or domesticated spp (the pet dog or cat) and the garden rose that are the entrée for people to open their eyes and explore the natural world around them.  It is very demotivating for newbies to see their initially joyful observations deemed inferior and fail to receive the tick of approval. I hear many people, when starting out, say they don't want to record anything because they think their 'silly common records' are not worthy, and we are only for elite, snobby people to record some rare or weird creature that only an expert would know about.  This is counterproductive.  I apologise for raising these matters again, but they seem to have been ignored unless I missed it, or because they have been in the wrong topic thread, but I know that Carrie Seltzer has made a similar point recently, and I can say that all the NZ administrators and associates are unanimous in wanting this policy changed. Please can this happen now, or at least hear a rational reason for not doing it other than one just doesn't happen to like exotic/planted/caged things.  Look, I'm a field ecologist and I love the great outdoors and my life is dedicated to saving primary (wild) habitats, and we know more than anyone about the negativity of biosecurity threats and people spreading harmful or undesirable organisms, but this doesn't negate all the other arguments above and the important ecological information regarding domesticated things - knowing where they are, maybe anticipating the next ecological disaster, and applying this knowledge in a proactive way. Again this is valid applied RESEARCH. We definitely should be promoting the wilding of nature [it is estimated there won't be any wild macro-fauna left on the planet by middle of this century if we don't do something about it now]. The corollary of all this is the desirability of providing a range of pin types for mapped records so distribution maps make a distinction between 'wild/natural', 'planted/domesticated/caged/cultivated', 'naturalised exotic spp', and 'imported native species naturalised outside their natural range'.  This last category will require some knowledge and may be added by other observers, or as discussed elsewhere in this forum, matching it to known natural distributions with a suitable buffer, and automatically categorising it as naturalised if it falls outside the buffer.  I hope this makes sense and gathers some productive discussion - and isn't just some late night (NZ time) raving of a kiwi. We are a bird too you know! :-).

[Murray Dawson]

I'm certainly in support of Colin's views. In an increasingly urbanized world, peoples first engagement with nature is often plants that are cultivated, and that's all part of their iNat journey. 

[Charlie Hohn]

I agree that research grade should be attainable for cultivated plants, but want to add the caveat that they should still remain absent from the normal species range maps or else get different symbology. They are valuable for all the reasons mentioned but are not valuable for spatial ecology

[Charlie Hohn]

...and Colin said that in his post too... sorry for the repetition. but yeah, I agree. There could be for instance a great opportunity for promoting documenting things planted for pollinators. Especially if 'private place' functionality is added for yards. Document the native plants or other bee plants you hvae in your yard, then document the bees. Why not? Document anything planted in your area that might become invasive. Document phenology of planted species. Etc

[Whitney Mattila]

I agree. I dud a thread similar to this, with the added suggestion of there being a 'research grade' for garden and landscaping plants, just with a different name to make it clear that they're different. Maybe, if that happens, they can be a selectable option for maps for specific species/families/etc as well, if the casual option is adding too much noise by itself.

[Alysa Joaquin]

I also agree with Colin, especially considering random people can wander into my observation list, decide something is cultivated/captive, and quietly make it casual grade (and therefore absent from searches for the species). I've had escapee plants marked as captive because they weren't native, even after I explained that there was no habitation anywhere near the area and it couldn't possibly be cultivated. It's really frustrating to do a ton of work with rearing immature insects I find in my yard to document life cycle changes (and discover an identification error experts have been making!), only to have them marked "captive" by strangers. The life cycle graph that was recently added to species pages had a huge potential to reveal information about species, and my observations from rearing some species made a significant impact on some of those graphs. All that information is totally lost when people check the "captive" box and I don't receive any kind of notification.

Similarly--some species are only rarely seen (or when they are seen they are dead). Captive observations of these species can do a lot of good in seeing anatomical features and behaviors, but when they're marked captive they don't show up in searches. I have a few observations of an ambassador bat Austin Bat Refuge brings to outreach events (her name is June Bug!), and there really aren't that many clear, close-up photos of this species. But, since she's captive... nobody sees her.

I'm not sure how I feel about people posting their pets, doing nothing but "human" observations, observations of zoo animals, observations from botanical gardens in otherwise under-represented areas (the top species lists for those places may end up all being non-natives from the garden). I think this is a really complex problem and I'm not sure how I would address it. The biggest issue for me is legitimate and informative observations being excluded from research grade because it either is or appears to be captive--but then how do you separate these observations out from observations that don't contribute to our knowledge of the natural world (a photo of a houseplant or a domestic cat)?

[Charlie Hohn]

there are a lot of grey areas and gradients here but with plants there is a big divide. On the one hand there are plants which are planted buy survive for many years in the same place - trees, shrubs hardy to the climate, bulbs that slowly spread, restoration plantings. On the other hand there are houseplants, annuals from wal mart you stick in the ground for a few weeks, and vegetables you plant that don't overwinter and are carefully tended. The former has a lot more data valuable to conservation and science than the latter. With animals, there seem to be a bit less grey areas - any pet tells you essentially nothing other than maybe helping the imaging algorithm. My preferred nerd answer that no one else wants would be to have divisions for all these grey areas and have people fill those out, but i realize that won't work. 

[Colin Meurk]

Thx Charlie, Murray, Whitney, and Alysa for the lively and informative discussion on this. Just to pick up on a couple of points. Note in relation to 'dead' things; there are several fields available for alive, dead, dead or alive, etc.  Cleaning up such replication is another topic, being partly addressed now by annotations :-). With regards pets - well, I have no problem, again if they are marked 'domesticated' and show up as such by a distinctive pin on the map separating them from wild dogs, cats or rabbits. And as I said previously it is often the entry point for new iNat recruits. And the test should be is it of scientific merit (hence research grade). Scientific interest will vary greatly among scientists - from Alysa's life cycle studies (therefore capturing all stages of life, not arbitrarily deciding that only adults merit scientific study) to biogeographic distributions (and being able to differentiate/detect naturally occurring but perhaps expanding ranges - for example with biosecurity issues like the myrtle rust from Sth America now spreading across the Pacific and having recently reached NZ probably from Australia - we have a myrtle rust reporter app!); AND social science (what are the distributions of pets, dog sizes, hair colours, in relation to say climate, urbanisation, disease or rodent presence, and socio-economic indicators). I periodically record (domestic) dogs and cats because as with common wild organisms, mapped gaps will otherwise be artificial in themselves. People often don't like to or think anyone is interested to record common species like Poa annua or house sparrows (declining in parts of Europe) - so we should encourage people to record every/anything. And if one generated a map of dogs or cats it would give a very distorted picture if only my cat showed up on the map of NZ :-). Or sheep - there were once 30 x more sheep in NZ than people - where did they all go if there was a blank map? Well there are a few wild/escaped sheep - one celebrated case was Shrek (https://teara.govt.nz/en/photograph/16641/shrek-the-sheep ) :-).  Am I boring everyone now?  from the sample of comments on this so far it would seem the desire line is to reluctantly (in some cases) accept domesticated/cultivated organisms (and maybe a couple of other grey categories) as warranting research grade PROVIDED they are clearly distinguished as such in maps and data feeds. cheers c

[Scott Loarie]

So you guys want a way to search for captive obs that would otherwise be research grade? ie have a community ID of species rank or lower and don't have any issues with location etc? 

I think we can do all of that except the community ID piece currently: https://www.inaturalist.org/observations?captive=true&photos&place_id=any&subview=grid&taxon_id=152870&verifiable=any&rank=species

One issue we were talking about in the office is that people don't tend to ID captive stuff (even before we buried them by defaulting to showing non-captive obs in search) I think because:

1) naturalists tend not to be interested in captive stuff (I know I'm not)

2) location is really key to ID'ing things and its not super useful for captive stuff

3) often captive species don't work well with taxonomies (weird cultivars and breeds and hybrids)

If folks are interested in ID'ing captive stuff here's a handy URL:

https://www.inaturalist.org/observations/identify?quality_grade=needs_id%2Ccasual&verifiable=any&captive=true&photos=true

--

You received this message because you are subscribed to the Google Groups "iNaturalist" group.

To unsubscribe from this group and stop receiving emails from it, send an email to inaturalist+unsubscribe@googlegroups.com.

To post to this group, send email to inatu...@googlegroups.com.
Visit this group at https://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

[Colin Meurk]

Thx Scott – i/we want to break down the arbitrary divide of what has scientific merit. I guess I want to search for research grade full stop (that have a photograph and therefore are verifiable – non-photographed records by acknowledged experts is another debate), and I want to see my (and others) records have a ‘research grade’ banner if they have been verified.  Just because some people don’t like identifying  captive things doesn’t mean there isn’t a whole subset of users who do as I explained. And it is a real thing regardless of how it got there. Again I think it is an entry point for new ‘identifiers’.  A novice can identify a cat, dog or sheep (and maybe even add the breed) whereas they may feel intimidated or frustrated by the level of taxonomic experience required for many other things. We should be welcoming J.

BUT my reading of it is that we basically want to generate map data that differentiates the categories – distribution maps are one of the most significant scientific outputs from iNat in my opinion. And it’s pleasing to note that increasingly GBIF map overlays are based on ‘our’ data – we are having a big impact on biogeographic knowledge J.

1.       I wouldn’t say we’ve found any particular resistance to identifying captive things.  I speed date thru masses of records and often back-arrow through someone’s records; I don’t skip the ones that are pets or garden plants if I can identify them.  that is the only determinant – can I identify it?  To clarify – I’m not a great advocate or get great kicks out of seeing animals in zoos – but it is my only location for a cassowary!

2.       Re captive animals in say zoos – again I don’t see where confusion arises.  Maybe it is (or soon may be) the only live example in existence (e.g. the white rhino).  It says something to conservation science that this the (sad) case. I agree location can be important for identification (part of the gestalt) and it is annoying when occasionally people hide rather than obscure locations. Certainly, for example, I don’t bother identifying Hebe (NZ woody veronica) from gardens beyond genus because there are over 100 spp and probably a thousand cultivars and/or hybrids. But that’s cool. It does raise the question as to whether an agreed genus should be research grade tho.  Why not? How does it confuse anything?

3.       Cultivars etc can often only be identified to genus level.  that’s cool. And personally I’m not at all interested in cultivar names, but again a subset of users are, and I think these can be perfectly well handled in ‘tags’ or custom fields. Same with dog breeds.

 

Thx for alerting/reminding me to the obvious filter of ‘captive’ species. I’m embarrassed to say I had forgotten that was a filter item!  The site is so rich one is always learning or being reminded of the many ways it can be used. It doesn’t seem like there is a key word search for topics in the help window? However, weirdly when I tried that on NatureWatchNZ I only got 4 records – that can’t be right; must have done something wrong:

http://naturewatch.org.nz/observations?captive=true&place_id=6803 .

My contribution to iNat thru NatureWatch NZ is to see the site be more relevant to more people J.  You guys have done a great thing – thankyou!

C

To unsubscribe from this group and stop receiving emails from it, send an email to inaturalist...@googlegroups.com.

To post to this group, send email to inatu...@googlegroups.com.
Visit this group at https://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

 

--

--------------------------------------------------
Scott R. Loarie, Ph.D.
Co-director, iNaturalist.org
California Academy of Sciences
55 Music Concourse Dr
San Francisco, CA 94118
--------------------------------------------------

--
You received this message because you are subscribed to a topic in the Google Groups "iNaturalist" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/inaturalist/lbllbJ3YVhU/unsubscribe.
To unsubscribe from this group and all its topics, send an email to inaturalist...@googlegroups.com.

To post to this group, send email to inatu...@googlegroups.com.
Visit this group at https://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

---

Please consider the environment before printing this email
Warning: This electronic message together with any attachments is confidential. If you receive it in error: (i) you must not read, use, disclose, copy or retain it; (ii) please contact the sender immediately by reply email and then delete the emails.
The views expressed in this email may not be those of Landcare Research New Zealand Limited. http://www.landcareresearch.co.nz

[Scott Loarie]

Try this http://naturewatch.org.nz/observations?captive=true&photos&place_id=6803&verifiable=any

To unsubscribe from this group and stop receiving emails from it, send an email to inaturalist+unsubscribe@googlegroups.com.

To post to this group, send email to inatu...@googlegroups.com.
Visit this group at https://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

 

--

--------------------------------------------------
Scott R. Loarie, Ph.D.
Co-director, iNaturalist.org
California Academy of Sciences
55 Music Concourse Dr
San Francisco, CA 94118
--------------------------------------------------

--
You received this message because you are subscribed to a topic in the Google Groups "iNaturalist" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/inaturalist/lbllbJ3YVhU/unsubscribe.

To unsubscribe from this group and all its topics, send an email to inaturalist+unsubscribe@googlegroups.com.

To post to this group, send email to inatu...@googlegroups.com.
Visit this group at https://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

---

Please consider the environment before printing this email
Warning: This electronic message together with any attachments is confidential. If you receive it in error: (i) you must not read, use, disclose, copy or retain it; (ii) please contact the sender immediately by reply email and then delete the emails.
The views expressed in this email may not be those of Landcare Research New Zealand Limited. http://www.landcareresearch.co.nz

--

You received this message because you are subscribed to the Google Groups "iNaturalist" group.

To unsubscribe from this group and stop receiving emails from it, send an email to inaturalist+unsubscribe@googlegroups.com.

To post to this group, send email to inatu...@googlegroups.com.
Visit this group at https://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

[jesse rorabaugh]

"our threshold is at least 20 Research Grade observations by at least 20 different people"

Can you clarify the second part of that? I read it as being 20 observers which does not seem to be how the algorithm is behaving. For example, I recently had the algorithm suggest a Melon Aphid as a possibility. Looking at Research Grade observations I see that there are 5 identifiers and 7 observers. I have 16 of the 23 research grade submissions but it still did not block it from entering the algorithm.

Twenty observations is much easier to get than twenty different observers, also it seems less vital. Once you get to five people you have enough diversity of camera type and photographic style that I wouldn't think it would be an issue. Even with one observer, it seems a lot better to just add it to the algorithm and hope for the best.

[Charlie Hohn]

once it reaches 20 observers do all my photos count? or ust the first fiew or something? if anything you should do my last few... my photos have improved over the years in terms of diagnostic features. (despite what some might think as i still do weird ones sometimes)

[Star Donovan]

On Tuesday, March 27, 2018 at 4:54:22 AM UTC-5, Colin Meurk wrote: 

It is very demotivating for newbies to see their initially joyful observations deemed inferior and fail to receive the tick of approval. I hear many people, when starting out, say they don't want to record anything because they think their 'silly common records' are not worthy, and we are only for elite, snobby people to record some rare or weird creature that only an expert would know about.  This is counterproductive.

As a start, maybe on casual observations with enough IDs to otherwise be RG, iNat could just change the little gray flag that says "casual":

• change the text from "casual" to something like "Cultivated - Research Grade" or "Casual - Research Grade" (oxymoron?) on casual observations with enough community ID 

• change the color from gray to a "happier" color for kids/primary school students that is bright and eye-catching like the research grade's green*
  

I know that doesn't address the main points re: the value of the data and it's inclusion/exclusion, but it could help with Colin's point about motivation (and/or with teachers who require "Research Grade" observations for a grade from their students**).

* Maybe cyan? purple? orange?  The possibilities are almost limitless - except for green (too close to current RG) yellow (sometimes hard to read against white text or background), red or black (people have been conditioned by most graphic design principals to read those as alarms/warnings)

**I know that active curation by teachers is a much larger issue and I do not mean to offer a suggestion that could be seen to enable or encourage laissez-faire attitudes or behaviors, but this might might help address a small part of that (I'd much rather see a flood of "cultivated - research grade" obvs than the ID rings, copyrighted pic theft, and other behaviors that the forum topics have revealed students sometimes use to get to RG; a pic of your own pet/houseplant at least has the integrity of being your own obvs, with accurate time/date)

[Charlie Hohn]

what about "community-supported ID"

[Star Donovan]

On Wednesday, March 28, 2018 at 1:18:58 PM UTC-5, Charlie Hohn wrote:

what about "community-supported ID"

For the banner text? 

[Charlie Hohn]

to be the equivalent of 'research grade' for anything cultivated

[Colin Meurk]

Why do we need to go into contortions? It's just research grade as I've argued. no need for some other category.  Most impt thing is that it goes into GBIF! And all the legacy data as well.   cheers c

Sent from my Samsung device

--
You received this message because you are subscribed to a topic in the Google Groups "iNaturalist" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/inaturalist/lbllbJ3YVhU/unsubscribe.

To unsubscribe from this group and all its topics, send an email to inaturalist...@googlegroups.com.

To post to this group, send email to inatu...@googlegroups.com.
Visit this group at https://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

[Star Donovan]

I was just proposing a stop-gap solution in case admins or other users really want to keep cultivated/captive apart from research grade, or as an interim work-around if it cannot be implemented quickly.

Sort of a concession: "Ok, if they must be separate, can we at least do this to encourage users?"

[Charlie Hohn]

Colin my understanding is GBif doesn't want cultivated species. And i agree there. It would make their range maps useless unless they had a way to filter it out. Though as the iNat range maps fill out it matters less I suppose. And it wouldn't be that hard to just not export those.

iNat admins seem pretty strongly set against having cultivated plants as research grade, since this keeps coming up and never really gets any traction or a full explanation. I do think part of it is the difficulty of the grey areas For instance other than having mediocre stock images animals in a zoo or houseplants don't offer much to the scientific community (though it can be fun to track them on here). Persistent planted landscape plants do matter a lot for urban ecology though which is something iNat focuses on. It's a harder to describe distinction since it varies from place to place. My cutoff is whether or not it survives a Vermont winter (either as a plant or as seeds). If you plant it and it dies over the winter without seeding back in, it's not that different from a houseplant you keep outside. But obviously every part of the world has different stressors - water availability, shade, heat, soil nutrients, etc... and a different 'hard season' (or in some cases none) and that gets tricky. For instance... in coastal California which is a mediterranean climate with very dry summers... nearly any plant that is nether native or an invasive from a similar climate needs summer water to survive (and sometimes in urban settings even the native plants need water)...

[tony rebelo]

I never understood this distinction and I dont think I ever will.  

Perhaps the real solution is to change the name "Research Grade" to  "GBIF Grade" and admit openly what this is all about, and stop pretending and confusing.

if we are going to keep the term "Research Grade", then no problem, but why on earth must it be affected by whether a species is cultivated or planted or not?

If things are annotated as planted, caged or gardened, is that not good enough?  How on earth does the annotation of "planted" affect the quality of the identification?  If GBIF wants only wild organisms then let GBIF select Research Grade AND Wild!  Why deprecate the Research Standard because of one organization's needs?  Other organizations do use invasive, naturalized or planted / captive organisms as useful information in a variety of different uses (but even they are usually interested if it is planted versus wild).

[Charlie Hohn]

For what it's worth, invasives spreading on their own, naturalized plants growing from naturally dispersed seed, etc (and the animal equivalents) aren't marked as captive. 

[Lincoln Durey]

I think:
1) definitely keep the current thing as "Research Grade" that name is a super concise way to say what I think iNat has been about for 10 years: that this observation is a piece of "Real Science" that scientists can use and rely upon in their research, and that follows very strict provenance rules. 
2) call the new "Planted RG" Community Grade (CG)
3) report RG2+ observations to GBIF, (as now)
4) never report CG obs to GBIF, we need those maps to stay pure wild plants.

[bobby23]

Community Grade... I like the sound of that. A good intermediate grade between Research and Casual.

Why is Casual Grade "demoralizing"? Why do people assign it less value than Reaearch Grade? No one is telling people to feel this way.

[jesse rorabaugh]

Changing Research Grade to include captive/cultivated organisms seems like a stretch to me. Too many people don't want them in range maps and Research Grade hasn't included them for so long it would confuse people. The data should be made available to anyone who wants to use it, but that is enough.

Definitely they should be in the algorithm though. It would be much more useful as a way for a non-scientist to identify species if a bunch of zoo/pet store animals, as well as market and garden plants could be included. Species such as corn which are virtually never wild really should be in the algorithm once sufficient observations have been made.

It also makes sense to split casual grade. It is currently a grab bag of different things:

1. Observations with no evidence for them. Typically no photo or a very poor photo.
2. Date/location inaccurate observations.
3. Captive species
4. Fossils

The title causal grade probably makes sense for 1, and maybe 2. For items 3, 4, and maybe 2 it may make more sense to give them a Needs ID tag until confirmed then give them a title such as "Supported ID"

[Upupa epops]

There is at least one other type of casual observation.

If you mark an observation "Community ID cannot be improved" and the taxon is at a higher level than genus, the observation becomes casual. Only if it is genus level or lower does it become research grade.

Also if the user opts of of a community ID, I think the observation becomes casual, even if everyone agrees with their ID.

I'm not sure why either of these are the case.

[Jon Sullivan]

This is an interesting thread.

I agree in general that it would be good for captive/cultivated records to be able to achieve "research grade" status, or some similar word.

I understand why iNat wants to encourage people to get outside and find wild things. I want that too. I also know of plenty of research questions in invasion biology and climate change research that can be answered with the cultivated plant records such as on iNat. They are research grade even if they're not being flagged as such.

iNat has such a strong and diverse community that I don't think we're in danger of being overrun with kitten photos. I'd be quite happy if captive/cultivated organisms could achieve research grade just like wild records. If that leads to some more captive/cultivated records, then that would be useful rather than a distraction. These organisms, at least the cultivated plants, make up most of the living fabric of cities and farmland.

If the first plants a new user adds to iNat are cultivated, as I imagine must often be the case in cities, then I see only benefits in those records being quickly identified and achieving research grade. We can lure those users further into the wild as they get hooked on the iNat community.

I personally don't mind either way whether captive/cultivated records are on or off by default on the Identify page. I can get to them when I want with the filters. However, I do know that some people, after not seeing their captive/cultivated records on the default Identify page, have been purposefully keeping the captive flag off their records to get them identified. This is not ideal. Perhaps it's better for them to be displayed by default and those users that don't want to see them can turn them off.

It is important though that captive/cultivated records continue to be off the default maps and only added with a filter, or flagged separately on maps.

I wonder if it's worth trialing this and seeing if it makes a detectable difference to the proportion of iNat observations added that are captive/cultivated and the proportion of new users that are retained.

[Donald Hobern]

A few comments from the GBIF perspective.  I don't believe that Research Grade was explicitly established only as a tool for GBIF.  

From our standpoint, it would probably be good for a larger proportion of iNaturalist records to make it into the GBIF network.  

This would include observations identified to genus level (which may be useful for less well-recorded or hard-to-identify taxa, and which may help these records to come to the attention of people who can curate them further. 

It might also include cultivated/domesticated organisms at least where these are active components in the environment and therefore de facto a significant component of biodiversity in a region.  This would include crop species, free moving domesticated animals, etc.  Zoo animals and pets would be much less useful.  

I think, as some have mentioned, the key thing would be the quality of the metadata giving context to the observation and as far as possible indicating the human assistance/management aspect.

Donald

--

You received this message because you are subscribed to the Google Groups "iNaturalist" group.

To unsubscribe from this group and stop receiving emails from it, send an email to inaturalist...@googlegroups.com.

[jesse rorabaugh]

If we do start sending cultivated species to GBIF, it may be necessary to find a way to split plausibly useful observations from those which aren't going to be useful to biological scientists.

An extreme example of something not worth sending to GBIF is this one I made yesterday:
https://www.inaturalist.org/observations/10511634

This observation has no scientific value. I just thought it absurd that the identify algorithm can identify all sorts of obscure stuff, but can't identify a peanut. It would be neat if the algorithm could look at a photo of anything in a supermarket, pet store, zoo, or botanical garden and identify it. So I took a bunch of photos of stuff in my pantry.