Achieving the full potential of a scale item in scat observations

This post supplements the following one that I posted in Sep 2018:

https://www.inaturalist.org/projects/scats-dung-s-afr/journal/27547-submitting-inat-observations-of-animal-scats-droppings-some-suggestions

Sometimes it is possible to ID a scat with reasonable confidence just knowing its approximate size and, of course, such things as habitat, the scat contents, its exact location (buried, elevated, in midden), its shape, likely occurrence in that area and so on. But this is often not possible. This is when knowing the exact size of the scat can be useful.

The problem

A challenge with animal scats is that the average sizes for different species can be very similar. Consider this example. There are 22 non-antelope species with an average scat width between 10 and 15 mm. How about this: There are 20 antelope species with an average pellet width less than 10 mm.

I know these numbers because I have a database of thousands of scat measurements covering many southern African species.

It’s not difficult to imagine that if average widths are very similar then the ranges (e.g., smallest width to largest width) for the species overlap. Hugely. That is the real problem.

How I use the scale item

I can measure on your photo on screen (i) the scat width and (ii) the size of the scale item. (Say they are 45 mm and 60 mm respectively.) Using the actual size of the scale item (say a South African 50 cent coin so its diameter is 22 mm) I calculate the exact width of the actual scat as follows:

Scat width = 22 X (45/60) = 16.5 mm.

I will sometimes calculate the average of multiple measurements of the scat or pellets in the photo. I’m most likely to do this for the antelope because the average width of the pellets is more representative than the width of a single pellet.

Producing the shortlist

You may well wonder how I can possibly establish anything useful from ranges that overlap so much - a very reasonable question.

Applying a statistical approach to analyse the data in the database gives me my best chance. I use (i) the average scat width and (ii) the width range (represented by its standard deviation) for each species in the database. Specifically, I calculate the statistical likelihood of each species being your species.

I’m most interested in the shortlist of species with the highest likelihoods. Your species is most likely in that list.

However, “The problem” described above is acutely manifest in these lists. If the width of the scat in your photo is not accurate – because you or I had to guess it - I get one list. An accurate width will give a different list. This is when the full potential of a scale item with accurately known size is realised.

Scale item recommendations

A ruler as scale item makes things easy, for me at least. It also works for any size scat or pellets, a 10 cm ruler length being quite sufficient.

A coin is good, placed so that the value is clear in the photo. The exact dimensions of most coins on the planet are available on the web, if I don’t happen to have your chosen one to hand.

You can measure the scat yourself though I don’t advise this, for various reasons. It’s simpler for you to display some scale item.

If the exact length of the scale item is not obvious (like it is for a coin), then please measure it exactly and state the size.

If you don’t want an ugly distracting unnatural scale item in your photo, then just submit two photos, the first without and the second with the scale item.

Don’t use fingers unless you (i) specify their width, (ii) where on your finger the with applies and (iii) you take note of the suggestions below on how to take the scale item photo. I understand perfectly why fingers are often used as scale items – you are always sure to have some with you no matter where you are. But as scale items, they are problematic.

[ I said in the post mentioned above that “a finger is not ideal although it’s better than nothing”. This is true. However, a finger can also be downright misleading. If I use the above method to determine a scat width, what value do I use for the width of your finger? I can only base it on the width of my own finger, i.e., assume that your finger is the same size as mine (or maybe guessing that your finger is a little wider or narrower than mine). It has become evident that I can easily be out by tens of percent! Not that surprising, in retrospect. ]

Please use scale items of an equivalent scale to the size of the scats or pellets. Even if you state the exact maximum width of your shoe, it won’t help much if I need to calculate the width of tiny antelope pellets. Use your shoe (with stated width) for the big stuff. Conversely, using a tiny coin for a megaherbivore dropping doesn’t help much either. Just use your common sense remembering that I’m not looking for a rough idea of how big the scat is – I need it exactly.

Any other scale item can be used but please state what its exact size is. It can help me and therefore you.

Taking the scale item photo

The following is worth noting: Cameras can distort perspective. For scale item photos, wide angle photos are the worst. Zooming in is best.

I also suggest the following, in the interests of accuracy:

Place the scale item as close as possible to the scat or pile of pellets.

Take the photo straight down onto the surface of the ruler or coin (or finger, whatever).

The graduation lines on the ruler should be parallel (not look like they will converge somewhere in the background).

A coin should be perfectly circular, not oval.

The same applies to any scale item for which you have stated the size: it shouldn’t be distorted in any way.

The return on investment

I understand entirely that the quest for better accuracy using a scale item with known or stated size requires effort from you, the observer. The only return on that investment I can offer is an increased confidence in the ID of the animal that produced the scat.

Also, don’t forget that the supplementary information (habitat, contents, etc.) can also clinch the deal.

Kevin

Publicado por kevinatbrakputs kevinatbrakputs, 09 de junio de 2020

Comentarios

Thumb

thank you

Publicado por gillian21 hace 3 meses (Marca)
Thumb

yes, I see what you mean. I will keep this in mind in future and always stick a 50 c coin into my pocket in future. Greetings, Angela

Publicado por angelanoske hace 3 meses (Marca)
Thumb

Are fingerprints too fine? It is about 0.7mm per ridge? I presume that they will vary less than finger width or digit length, and so if you count 15 that will equal ~10mm.
Alternatively, could we perhaps have a post of personalized items. For instance the stitching size in my boot (I seldom include the entire width.
My problem is I regularly lose things like hats, GPS's, jerseys and such: coins and rulers will probably have a half life of one use ....

Publicado por tonyrebelo hace 3 meses (Marca)
Thumb

An interesting thought, Tony. A concern that some might have is that photos of fingerprints may be a security risk. There are opinions for and against this (Google “fingerprint photos security risk”). Although I have produced such iNat photos, I did start smudging my fingerprints before uploading. I am not advocating this – people must make up their own minds. And one might get the whole spectrum of responses from “don’t be ridiculous” to hysteria. I’m generally risk averse. I’m also not into hacking so I can’t personally judge the risk. :-)

But to return to the main issue, yes, I’m well aware of the practical difficulties of scale items. Probably the most flexible approach is to use whatever you have and then measure that item when you get home.

I haven’t got any polite suggestions for people who might lose such items, unfortunately. :-)

Publicado por kevinatbrakputs hace 3 meses (Marca)

Agregar un comentario

Acceder o Crear una cuenta para agregar comentarios.