Bird surveying in the field using GPS

The Ipswich RSPB Local Group is looking at surveying Ipswich’s house sparrows, ten years on after I took part in the first Ipswich Sparrow Survey. Most bird surveys cover a wide, fairly homogeneous area and avian subjects that are mobile over several hundred metres, or specific colonies. Surveying sparrows in the heterogeneous urban environment, with subjects that stay within 70m of the nest site when feeding young are a different kind of challenge.

We didn’t have the benefit of this BTO House Sparrow Survey protocol [ref]Protocol for censusing urban sparrows, DeLaet, Peach and Summers-Smith, British Birds, 104, May 2011, p255 ff[/ref]. As such the results show sample bias from the observer locations, survey effort and the difficulty to getting negative reports. However, it did answer the question ‘where are there sparrows in Ipswich?’, although not the question ‘where are all the sparrows in Ipswich’. It also answered the question ‘where are there no sparrows’ although the sparrowless areas were not exhaustively surveyed. Proving a negative is hard, though the BTO paper shows the way.

In the intervening time I solved the need to create waypoints on a handheld GPS and announce them to tie the audio to the location. Modern audio recorders contain clocks and timestamp the recording, this can be synchronised with the GPX tracks created by the GPS handheld. Smartphone applications like viewranger can also create GPX tracks. However, although a smartphone does many things it doesn’t do any of them very well.

The same trip tracked using Viewranger on a Samsung Android phone versus a Garmin Vista HCx. each machine was in different outside pockets of a coat at the same height above ground. The lighter red tarce is the Garmin which follows the paths better
The same trip tracked using Viewranger on a Samsung Android phone versus a Garmin Vista HCx. each machine was in different outside pockets of a coat at the same height above ground. The lighter red trace is the Garmin which follows the paths better

I set the Samsung to give the most precise location. The issue doesn’t seem to be as simple as the Viewranger Samsung app updating slower than the Garmin, there are enough points, but they aren’t in the right place, particularly when the border trees by the road and railway line

the save trace showing the trees
the save trace showing the trees

The aim of all this was to write a web app to record sparrows in the field. However, it turns out that the Location API is very hit and miss. Sometimes I got 20m accuracy. Sometimes I get 3km accuracy, and I can’t easily see how to fix that, having already selected high geolocation accuracy. If I can’t do that for myself, I certainly can’t support other people using it on two different operating systems and on some mobiles that may not actually have a GPS receiver.

 

Leave a Reply

Your email address will not be published. Required fields are marked *