Technology and Identification

 In today post I want to illustrate how a variety of techniques is somewhat needed to verify the presence of a species on a site. My case study for this is the Harvest Mouse. Harvest Mice are the UKs smallest rodent weighing only 6g when adult and not a species I expected to find on my land.

The habitat is suitable, thick brambles and stands of tall grasses and reeds are separated from agricultural land by a hedgerow. When I first started to manage my patch, I did wonder if Harvest Mice occupied the site and put out some cut of tennis balls to see if any would nest in them, this was a complete failure a) because even with GPS recording them, I lost them in the dense vegetation and b) because the frequent winter floods swept them away. I saw no signs of their presence although I have looked for nests and whilst I have seen bundles of grass, they were never distinct enough for me to rest an identification on.  

I assumed Harvest Mice if present were restricted to the field which predominantly grew wheat and corn, that was until this year.

I blogged earlier in the year about my purchase of a static batlogger and my success with it. From June I have been putting it out to record the bat species. I am still getting a handle on interpreting that data and in my June post I explained that I had detected Harvest Mouse calls.

Harvest Mice are known to vocalise in ultrasonic pulses between 20 and 70 kHz, with four notable call types that are distinct from other rodents. The data I collect with the Batlogger is analysed through the BTO pipeline. This outputs the results of all recorded sounds with an assigned identification, this identification is qualified by a probability that the sound file matches the sonograms known for that species. It is generally that any probability lower than 0.5 is discarded. This value seems very low to me, and I had decided only to accept probabilities above 0.75 and 0.9. The spread of the data can be seen in the boxplots below.

Total

As you can see the average probability is around 0.6 and the number of identified files is very low. For this reason and the thinking that the detectors’ location did not ideally line up with suitable habitat I decided to not count Harvest Mice as a resident species at present but to keep an eye, or rather, sensor out.

Given their tiny size it is very difficult to see a Harvest Mouse in the wild, specific surveys needed to be conducted. I doubted that my three trail cameras would yield much given their locations and the fact that I have always considered Harvest Mice to be a species that doesn’t come down to the ground much and spends much of its time in the vegetation.

I was proved wrong on the 19^th August. A single image and 10 second clip from my Mostela camera revealed what seemed to be a harvest mouse.

The individual in shot was tiny, its tail thick, its ears small and its nose blunter, but the image was black and white, and I wanted data to back up my gut. I posted the image to the Mammal Society Facebook group to garner opinions and set about researching. My Mostela is regularly visited by Wood Mice and shrews, and this individual had none of the ‘vibe’ of those species. As you can see from the images below the mouse was about 2.265 inches long (Head and body length) this equates to 57mm right at the lower end of the expected range of Harvest Mice and well below that of the Wood Mouse. 

In the image below I show a Wood Mouse so that you can see the size difference.

Research indicates that Harvest Mice do travel along the ground. At this time of the year, they should be breeding and be using breeding nests in dense reeds, grasses, brambles or cereals. The field by the mostelas was sown with wheat which due to the hot weather and the variety grew little higher than 40cm and was harvested on the 5^th July. Maybe the nest was destroyed then and the mice dispersed into my patch.

What I like about this story is the way in which the technology and research work in tandem to identify a hitherto unconfirmed species. This identification used two remote recording devices, a host of Mammal experts on Facebook, and deep dives into two books (Sound Identification of Terrestrial Mammals of Britain and Ireland, and Mammals of the British Isles)

It gives me a new drive to investigate this species to establish if they are a resident under reported species. I will continue to use the Batlogger relocating it closer to the mostelas and conducting a proper winter survey looking for old nests. I ll report back here as and when I find out more.

The application of Apps in Wildlife Watching

 I am of the generation that saw the common use of mobile phones during my late teens. Back then they were simple brick-like affairs with just call and test functions, far removed from the pocket computers we have today. I am by no means a Luddite I am drafting this blog on a Remarkable (TM) tablet but wildlife watching has always been a getaway from technology. I use tech of course, what birder doesn't? , my binoculars, my camera, my Garmin GPS device for plotting trees, and my digital thermometer/anemometer but I had always avoided the application of apps.

This month however I have discovered two apps that I have found easy to work into my survey routine. 

First off is the Mammal Mapper by the Mammal Society. I must admit I have been using this app for a little while to log mammal sightings but I haven't been using it to its full potential. My increased use of the app came from my desire to submit more of my sightings to the Mammal Society national database. In the past, I had used its individual sighting mode which is simple to use and very intuitive. This month I have used it in survey mode and it was amazing.  

I used it on my own patch of course, where else would I test it? The app allows you to continuously track your movements on a map as you walk about. Whenever you spot a mammal you log a sighting by pressing the record button. The record button allows you to add photographs and record notes. You can also record mammals based on their tracks and signs and dead individuals. At the end of your walk, you press submit and this generates a map of where you walked highlighting each sighting. It also states the time taken and the distance traveled. 

The app is so easy to use and creates a great record of all your mammal sightings and the distance and length of your visits. Its greatest benefit is the fact that the information is logged directly with the Mammal Society. Of course, there is an inherent bias in this system, small rodents are likely to be under-reported, and most species are nocturnal and hard to spot but the facility to record mammal signs and record individual sightings makes this a powerful tool for the society to monitor mammal populations.

The second app I have started to use is the Merlin Bird App from Cornell University. I am normally a traditional birder and make field notes of birds whose identities I am unsure of rather than carry a bird book with me. I always do my checks at home late with my good bird books. I used to carry a book as a teenager but I was once told that a good birder never takes one with him, an element of birding snobbery that has stuck with me. So was not the visual ID tools that I was most interested in. 

The power of this app is instead its bird song and call identification. I am not the greatest at identifying birds from their song, I have learned the basics and I know most of the birds on my patch but I am still getting the hang of some, the warbler for instance. I have tried to improve and used  CD's and cassettes in the past but it never really stuck, for me, I think I need to associate the sound with actually seeing the bird a kind of visual tag to attach to the audio. 

Merlin isn't the only app out there for this I have used a few other apps before this one and find most to be lacking in one key area, accurate identification. The Merlin App bowled me over the first time I used it on my iPhone. Cornell is an American university and so I had to download the Britain and Ireland datasets but this was very easy and took up very little space, my other option was to download the birds of the Western Palearctic, but as I don't travel I think the local set will work fine. 

I have only used the bird song function and it is great. You click on the record button and the app will display a sonogram in real-time very much like some bat detectors. Beneath the sonogram, it will generate its estimation of the species and create a list of all being heard and highlights each one as they call.

 I have used it for several hours and not once did it misidentify a bird, an accuracy rate second to none. It helped me locate species on the survey I had missed. At one point a background call, which I would normally have ignored as too obscure, revealed itself to be a Goldcrest, honing in on the sound I was able to see it and confirm its presence.  

This isn't an app that I will use on every visit nor will I record a full survey, although this is eminently possible, all be it risking using up all my memory storage. I will use it to help train myself and on unknown or suspect birds. It will help tighten me up on my neverending battle to accurately identify Marsh Tits and all Willow Tits and will give me a better handle on Willow Warblers in the summer. 

Like all good apps these are tools to help, not take over your visits. It would be able to record a whole visit and not look up and around at all and still have a fairly good species list at the end, all be it lacking in the less vocal residents. 

Technology should augment our enjoyment and not steal the fun from it and these two apps certainly do that. On my Mammal Mapper survey, I was hyper-focused on mammal signs checking every footprint and examing every burrow whilst the Merlin app opened up the world of sound to me and encouraged me to be a better birder.