Technology and birding: Merlin Bird ID

As a birder, I often struggle with my eyesight.  I wear very strong prescription glasses, and this makes birding in rain or humidity a proper pain in the neck.  Droplets on my glasses and steaming up when using binoculars are frustrating.  I find a change of focal length disorientating and difficult to adapt to; moving between binoculars and naked eye, or telescope to camera, or in particular from phone to optics leaves my vision momentarily fuzzy and blurred.  So I often depend on my ears when I'm birding.  I'm not brilliant at instant ID when it comes to bird song, but I usually use birds calling as a way of homing in on them and confirming identification visually.  I don't know how I'd manage to go birding without that sense, given how slow my eyes are to focus.  

Positively accessible

The Merlin Bird ID app from Cornell Labs has real application here, giving instant access to at least an approximation of this sense that a hearing impaired person struggles with in birding.  It doesn't give directionality, and so won't allow a person to home in on a singing or calling bird, but it does alert you to what's close by, what you could be looking for.  This inclusivity and accessibility is, on its own, worth all the negatives that we could (and will) explore when it comes to Merlin, and this has to be the headline: more people can go birding better and more often because of this app.  I have friends who have the kind of hearing range contraction that sometimes comes with age or with certain occupations (and I have friends who have never had the top end of frequencies) and Merlin helps them to identify when Goldcrest, Treecreeper, or Grasshopper Warbler are present.  Many people can hear those calls in isolation, when played back to them on a computer in a quiet indoor environment, but it's a whole different world when in the field, in strong wind, or a noisy place with traffic, industry, or even just a raucous dawn chorus.  There, many people struggle to pick out the smaller, more subtle sounds and Merlin does a good job of hearing what we cannot hear.

An audio fieldguide

I tend to use Merlin to confirm that I'm on the right track with my initial thoughts when I hear an unfamiliar call.  If I can hear a call but I can't place it, and having exhausted my limited store of expertise I still can't get hold of it, I'll flip Merlin on to see what it says.  It's like an audio field guide for me, a way of confirming what I think about a call.  I usually run it in short bursts, where it picks up most of what I can hear, and then stop it recording, and use the playback function to (quietly) listen again.  Very useful if a bird stops calling.  It's not infallible, and it's only as good as the microphone pickup on your brand of smart phone, but it has been very helpful to me when birding overseas in France and Spain, where I'm surrounded by almost familiar calls, but mixed in with the exotic.  Hearing Serins twittering over Melodious Warbler and Golden Oriole fluting past Subalpine species is overload for this UK birder and I just need a little help to separate them out, learning the calls from the way Merlin reacts to them, so that I don't need the app next time.  

A flatter learning curve

Beginner birders have a significant difficulty when it comes to learning bird calls.  A lot of newer birders struggle with them because we assume when we take up birdwatching that our eyes will be our primary sense.  It took me a long, long time to get to grips with the calls of even the small number of species I need to know in my own 10km circle, and every year I have some memory slip with summer migrant birds (especially Lesser Whitethroat - I never quite remember that song!).  For those of my generation who spent hours and hours in the field learning bird calls and experiencing the vast range of sounds birds make, building up a hinterland of knowledge, Merlin represents a shortcut, a leg up for new birders to be able to access a form of birding that is often opaque for the first few years of birding life.  It's fair to say that Merlin has a levelling effect, reducing the steepness of the learning curve and therefore increasing inclusiveness.  Less experienced people can bird better, quicker, if they use Merlin.  This is important in a world where birding mentors are less and less about who you go birding with and more about who you follow on social media.  I had some excellent birders to teach me, but the bird clubs are mostly gone.  Where will a new birder turn to for help?  The internet.  And Merlin is right there.  Nobody needs to lose face and feel foolish asking for help, especially given some of the passive aggressive, sarcastic and self-appointed alpha-type idiots who respond to any public request for help.  So quicker success means more people stick to it.  More people coming into birding find it less opaque and that it requires less arcane and obscure knowledge to succeed more quickly.

So Merlin helps me with my hearing, and it helps me to learn and ID birds.  It opens up bird identification to more people than would otherwise have access to it.  Sounds like a positive all the way down.  But as usual, it isn't.  

Merlin as a crutch

Give people an app that does something for them and people don't learn, they lean.  The much-feared AI brain drain is a real phenomenon.  Plenty of people run Merlin in the background all the time they're birding so that it catches what they miss.  At best this is a form of insurance so that they don't miss anything, but at worst it's simply letting an app do their birding for them.  There is no element of learning in running Merlin for an hour and checking it every 5 minutes to see what was there.  It feels lazy and disengaged from the environment to not use Merlin to help hear or see the birds around us, instead using Merlin like some great drag net to sweep up all the sounds we didn't have the attention span to listen for.  People who use Merlin like this lean on the app, it becomes a replacement sense in the wrong way, creating dependency on the technology to be able to identify birds by call.  It may be a leveller for new birders, but it becomes a lazy trap very quickly if not used the right way.

Confidently inaccurate

Merlin is a very confident voice that is wrong more often than we would like it to be.  There are some birds that it finds it hard to separate by call.  Merlin gets Goldcrest/Firecrest, Chaffinch/Redstart and Reed Bunting/Yellowhammer wrong around 40% of the time in my experience, depending on location.  Recent unlikely notifications I've had include Shore Lark in Bolton, Caspian Tern in Salford on the Moss, Bluethroat on the West Pennine Moors, and Firecrest in my tiny Manchester garden.  My camera rig has a high pitched squeaking noise that Merlin sometimes identifies as Sanderling, and something about my scope backpack makes it think there's always an Oystercatcher in the area.  This error-proneness in Merlin is something experienced birders can parse and correct mentally.  They know it's incredibly unlikely to see Sanderling at night in a Manchester forest, and so disregard the record.  Newer birders might accept it at face value, treating Merlin as a trusted opinion and that they have proof of hearing a certain bird.  When added to other technology like eBird helping them report sightings instantly, we get a lot of error in the reports of unusual birds.  This domino-rally chain reaction is having an impact on birds reported on rare bird services.  I've noticed recently that the number of retracted bird sightings on the rare bird apps has rocketed - I've stopped travelling for birds reported until there's been photo conformation, because a significant amount of what is reported is pulled from eBird, which is often informed by uncritical use of Merlin.  This also has a corollary impact on science, which depends on citizen sightings in order to gather information.  How accurate is the data when you factor in the popularity of an app identifying birds with a lack of human skill in the analysis?

Merlin has blindspots

Due to its selective programming which is necessary to narrow down potential species for ID, Merlin sometimes appears "deaf" to an obvious bird.  Wood Warblers were all around me, singing openly and loudly, and yet Merlin wouldn't pick them up.  This is a quirk of the location feature - the app didn't think it likely that Wood Warbler would be in an ornamental park near Bolton despite the fact that we've known about at least half a dozen breeding pairs there for the best part of a decade now.  Merlin selectively hears sometimes, and that can be strange and unsettling.  If I was hearing impaired, I wouldn't have known to wait and watch for the Wood Warblers to come out - a bird that I almost always hear before I see - because Merlin, that I depend upon for my sense of birdsong, had decided it was making an error detecting Wood Warbler and omitted it from my list.  A new birder might miss out on seeing Wood Warbler, trusting the app more than their ears.  

What's the verdict?  Merlin, a brilliant project, an incredible undertaking and an undeniably useful tool brings down barriers to bird ID, confirms what we are hearing and is invaluable to those with a hearing impairment.  But it also encourages laziness, dependency on the app instead of expanding human experiential knowledge, and it can be unreliable in a variety of ways, though there is refinement and development in progress to address these issues.  I think the positives outweigh the negatives here, the cons are things we can learn to manage in order to benefit from the pros. It helps to be around other birders to learn how to learn the bird calls and wean new birders off Merlin-dependency, but that's a necessary part of every birder developing: if not Merlin then we still have other bad habits to un-learn. 

Like all tools in birding, it really depends on how it's being used, but if this piece of gear helps birding to be more inclusive then I support it.