Large Moth Bias

iNaturalist is a wonderful platform and I use it on a daily basis. I'm honestly driven by my own enjoyment, but I also take pride in knowing I am conducting Citizen Science - the idea that the observations I post can be a valuable tool for biologists. As with any data set, iNaturalist observations must be used with care, and my own choices impact the data set. Biases exist and scientists must take care to identify the biases so that they can minimize their impact.

One of these biases I would like to highlight in this post is what I'll call the "Large Moth Bias." On a daily basis the average adult living in Oklahoma probably passes in close proximity to at least a few moths. How many of those moths do you think the average Oklahoman is going to stop, photograph, and upload to iNaturalist? That's right, the number is so low it might as well be 0%. On the other hand, if the average Oklahoman walks by a large moth, say a Luna or a Polyphemus, how many of those people are going to stop and take notice? Yep, definitely more than 0. This isn't just true of moths, of course. Most anything unusual or more conspicuous is likely to be noticed more than the small, common, or inconspicuous.

Why does this matter with iNaturalist observations, you ask? Well, if you look at the Moths of Oklahoma project landing page you'll see that the most observed species in the state is the Luna Moth (153 observations). Three other moths in the top observed list are also moths that I would put in the "large and conspicuous" category - White-lined Sphinx (115 observations), Polyphemus Moth (98 observations), and Io Moth (94 observations).

If you were to have an accurate count of every single moth in the state of Oklahoma over the course of a year I don't think any of these three species would be in the top 10. I believe the Luna, Polyphemus, and Io have so many iNat observations because when people see them they take notice and feel compelled to share the photo and find out what it is. In general, people are less likely to care what species a moth is when it's small and brown and fits their informal definition of what a moth should look like. While looking at observation numbers I was struck by the low count for Oklahoma's other large moth, the Cecropia. There are only 33 observations of this species in the state. I think that tells us that the Cecropia is quite rare, given that it is actually the largest moth and very showy and has so few observations. That seems significant to me. Another explanation is that the Cecropia only has one brood per year so the amount of time that Cecropias are out flying around is much less than the 2-3 broods of the other large silk moths (Io, Polyphemus, and Luna).

In my opinion, the White-lined Sphinx is more common than the species listed above, but I think it too has been elevated by observation bias because it is often seen during daylight (or dusk) hours and is sort of large.

Another impact on these counts is that iNaturalist is configured to log individual observations of a single organism, whereas some other platforms (like eBird) are built around the idea of logging multiple individuals of the same species, and there are ways to see the quantity and frequency of different species. On eBird you can enter a checklist with each species of bird you saw and how many you saw of each species. For instance, 2 Northern Cardinals, 5 Blue Jays, 100 European Starlings, and 1 Bald Eagle. With iNaturalist you are unlikely to log 100 individual Bluegrass Webworm moths when you see them. In fact, you might not log any of them because you see them so often and have already logged one this week. In that way, even those of us who contribute many observations to iNaturalist are contributing to this bias.

In a future blog post I will highlight how population biases the geographic distribution of species so stay tuned. In the meantime, what are some biases you can think might impact observations in iNaturalist?

Posted on 01 de outubro de 2019, 01:48 PM by zdufran zdufran

Comentários

Nenhum comentário ainda.

Adicionar um Comentário

Iniciar Sessão ou Registar-se to add comments