Description of need:
My request is to make the default radius of “Nearby observations” smaller. 50km away doesn’t seem very “nearby”. Yes, I know you can change it from the default once you get to the Explore page.
I think the current radius is a good default, especially since many areas of the world are still under-represented and thus the “seen nearby” function would not be as useful if it was narrower.
In most of California, the habitat can change almost completely in just 5 km. 10 km can take you from the mountain tops at 10000 feet elevation to the desert floor at 0 feet elevation. Those two areas have no species in common at all, so it is quite unhelpful to suggest nearby species from 10 km away in those cases, if they come from those elevation extremes. Examples include from Mt. San Jacinto to Palm Springs, and Laguna Mountain to the Vallecito Valley floor.
I understand that in some parts of the world, the habitat doesn’t change much in 50 km. In those cases, a larger radius might be appropriate if there are few nearby observations.
The best solution might be to have a radius that depends on the local variation in elevation.
Lacking that, I’d suggest using 5 km as the default, but if there are few nearby species within that radius, perhaps doubling the radius until a given number is retrieved.
Maybe a dynamic circle that’s something like “smallest radius with 100+ observations” or something would make sense. Essentially, the 100 closest observations, but still mappable with a circle.
Where I live it’s more often change in soil composition than elevation. Also a change based on how densely populated (and thus disturbed) the area is effects what species of plants you get (I don’t know how much other taxa are effected).
Most species have distributions far greater than that, so within that boundary is perfectly reasonable.
If anything it’s actually pretty conservative unless you’re dealing with an endemic species with an extremely small range, and even then you can’t rule out outliers.
Maybe I’m using it differently than other people? If I want to know all possible species in the area, we already have that functionality elsewhere. For nearby, I want to look to see if something I’m looking at (my photo or other people’s) has already been observed in the near vicinity. Maybe someone has recorded the exact same one as me or maybe another one nearby in a very similar habitat. It can be helpful when IDing. Or I might want to see what else has been observed in an area so I know things I can look for when I’m there in the future. Or maybe I want to ID other observations in the very nearby area after visiting myself because I know what I saw there (I don’t photograph and observe every single organism).
I suspect that it’s more common to be using it as an aid to help filter potential species for the area though, in which case the 50km radius is very small.
For your specific application I do that on the browser by selecting an appropriately sized AOI (usually vastly larger in my instance as the nearest similar observation can often be hundreds of kilometers to more than a thousand kilometers away), and filter the search within that AOI.
The best option would be an easy toggle with several radius selections in the ID stage, but failing that I’d opt to leave it as it currently is and save the more detailed analysis for browser specific exploration in a separate ‘Explore’ tab.
Would it be at all possible to allow us to specify ranges in our Account Settings: Content and Display?
One size certainly does not fit all. Then the default can stay 50,000m (but why is it in km, not in metres, like elsewhere on the system), but users can tailor it to their needs.
Perhaps this feature “Observations Nearby” should use information from the geomodel and not necessarily be circular. The useful size varies from place to place, and a circle does not take into account habitat changes e.g. due to elevation.
Yes, but it maybe should take mountains into account, maybe not be circular, and maybe have variable sized regions depending on the location (terrain, sparsity of observations).
On a similar note - could we make the maximum size of the location circle smaller? I’ve seen observations that cover multiple states - one time, half of North America even. I’d argue that if you can’t get within five or ten miles, it shouldn’t be counted as a valid observation, because it’s impossible for a second person to go and confirm the observation.
I disagree that observations with large accuracy values shouldn’t be allowed. The data are the data - anyone can filter out observations with large accuracy values if they wish (and many do).
I’m not sure what you mean by
but if this is intended to mean someone going out to actually find the same individual organism, I think that this occurs for a vanishingly tiny proportion of iNat observations. It’s possible for static individuals (plants) or very detectable ones (birds), but for most it will not be possible. In fact, for many organisms, even an accuracy as low as 1 km (or maybe even less!) would preclude most attempts to relocate an individual.
Many biodiversity checklists are done at state/county levels (or their equivalents) which can correspond to accuracies of >100 miles in some cases. Those data can be quite valuable. Observations with large accuracy values can still be useful for phenology, training the CV, and many other things, even if their coordinates are not very precise.
I never use this link (actually I kind of forgot about its existence) so I’m curious what people use it for? Do you mostly use it in the ways @lappelbaum says here?
I generally use the Compare tool becuase I’m mostly wanting to ID something if I’m looking at it, or I use the link to check observations made that day by the observer to see if they’ve made observations for instance where multiple species are interacting.
I wouldn’t be against making it a bit smaller (like 25-30km) but either way one can quickly and easily resize the circle to an area of one’s liking. I don’t think a dynamic approach based on elevation or number of observations would be feasible. We’re just trying to get people to a reasonable starting place and then they can take it from there.
