I’ve been trying out the nature identification app Seek on various plants and animals I see around my neighborhood. Some IDs surprise me (a suspected invasive weed really is a native wildflower); some are so general and obvious they are useless (Dicot); and others are broadly informative but nothing I didn’t already know (Mustard Family, true moss). Seek is sharp on European wildflowers we call weeds in the US and I suspect this is due to abundant observations in the iNaturalist database for the image recognition program.
Warning — there’s a known bug in the app that can result in a blank white page instead of your data. It happened to me on Friday as I tried to look back at my saved observations. iNaturalist developers know about this and are working to get an update fix out asap. Don’t delete the app or you’ll lose all your data.
When the app is working properly, this is what the app’s home screen shows before it is ready to scan using your phone camera. (The wild life in the top of screen will be synched to your area.)
When the camera has latched onto an organism and attempted an ID, the target image appears in the middle of the screen with a name above it. I’ve found that by moving the phone (camera), the name can shift. For example it might go from a general ID like “Dicot” to the specific ID “nipplewort.” I’d never seen that name before so I clicked on the link provided on the screen read info (often from Wikipedia) like the scientific name Lapsana communis, Family Asteraceae as well as species and habitat descriptions. The leaves were the only plant part Seek saw to make the ID and since the plant wasn’t flowering, I’m not 100% sure it was correct but the leaf looks spot on.
The Daily Bucket is a nature refuge. We amicably discuss animals, weather, climate, soil, plants, waters and note life’s patterns spinning around us.
We invite you to note what you are seeing around you in your own part of the world, and to share your observations in the comments below.
|
This is the plant I’d suspected was a weedy Apiaceae but it is Pacific bleeding heart (Dicentra formosa).
This next photo is to illustrate what Seek sees and still can ID. I snapped this blurred photo using the app just as the name appeared (bedstraw). And the app didn’t pay attention to the lesser burdock leaves in the background with sharper focus, although when I pointed the camera just on those leaves it accurately ID’d the burdock (European weed).
Here’s an insight into Seek’s ID ability — the app has expectations. Same plant, different fruits gave different answers.
A few more yard plants and then we’ll go to the beach. This low brick wall is along the lawn side of the sidewalk in front of my house.
I decided to concentrate on the mosses because someone commented in a past Bucket that Seek was good for ID-ing mosses. I had mixed results.
Seek did better on this one — maybe because the sporangia are easily seen?
Lichen was picked up easily but only identified as Shield Lichen, nothing more specific. I think it is genus Parmella (that is shield lichen) but since my lichen guide book burned up, I don’t have a further ID and iNaturalist is not more specific (so far).
On the trail to the beach, I spotted these native plants (spore producing). I only knew the genus for certain because different species are common in the Sierra where I lived.
The beach is a small patch of Elliott Bay rocky shoreline on the southeast edge of Magnolia.
Beach life was mostly seaweeds.
This seaweed was trickier for Seek. Depending on how I moved the camera (very subtly) I either got a general ID of “algae” or the more specific ID of “vesicled rockweed.”
I was hoping for more marine inverts but the beach wild life predominantly was sessile barnacles (per Seek — no more specific name) and two mollusks (one is the devoid of life “bivalve” in the top photo). I tried Seek on some gulls in the water’s edge (no response — too far away) and on the bald eagle in a tree (also too far away).
Finally I found one animal that was close enough, big enough, and slow moving enough for Seek to ID, but only to genus. I think Seek said “whelk,” but I can’t go back into the app and look at the name (due to the white screen app bug) to be sure no species name was attached. The sessile barnacles are also in this photo along with some vesicled rockweed. I don’t know what the greenish grey stuff is — some lichen?
I’ve yet to use Seek in a less developed habitat with more native herbs and shrubs. Trees seem most difficult for Seek, but the big-leaf maples are flowering and opening first leaves and were readily ID’d.
Using it to ID unfamiliar organisms (like the purple whelk) is entertaining and easy. Since I’ve only been in Seattle for 3 months of winter, there are many plants popping up I can’t easily identify and the app helps narrow down possibilities. When it is useless it isn’t any more useless than I am and I can still take a photo for ID later. If I upload that photo to iNaturalist (can’t do it within the Seek app though — need to do it separately) and get an ID put on it, I’m helping Seek learn to identify that organism. Plus, once the IDs are confirmed, the observations add to biodiversity data and I’m doing science.
Today, Saturday, is CitSciDay2019. Grab bincos, camera, phone app, field guides and get out there. Be a scientist.
Get in CitSci shape for the City Nature Challenge 2019 beginning on April 26th. Cities around the world
The challenge occurs in two phases. From 26 through 29 April collect and submit observations through iNaturalist or your city’s chosen platform. From 30 April through 5 May, observations will be identified (you can do this even if you’ve not submitted any). Results will be announced 6 May. Visit the website for more details.
Last year San Francisco swept all three race categories: Most Observations (41,737), Most Species (3,211), and Most Participants (1,532).
SPOTLIGHT ON GREEN NEWS & VIEWS
EVERY SATURDAY AT 3 P.M. PACIFIC TIME
DON’T FORGET TO VIEW METEOR BLADE’S COLLECTION OF LINKS AND EXCERPTS FROM ENVIRONMENTALLY ORIENTED POSTS PUBLISHED ON DAILY KOS DURING THE PREVIOUS WEEK
|