Method

One of the techniques for determining geographic location (geolocation or GLS - Global Location Sensing) is from a daylight levels record. Just as ancient mariners monitored the celestial bodies to navigate, we can determine location from monitoring sunlight. Longitude can be determined because the Earth spins and latitude can be determined because the Earth's axis is tilted. For a very good animation illustrating why we can determine location using the Sun, see the animation at mathsisfun or this movie:




Threshold analysis theory

Light levels are recorded with reference to an internal clock/calendar. From the data the time of local sunrise and of local sunset are estimated on a daily basis by assuming that a certain light level determines that the Sun is at a certain elevation. The time midway between the times of sunrise and sunset is local noon (and between sunset and sunrise is local midnight). This determines the longitude. The day length on a particular date determines the latitude, eg. in June the farther north the location, the longer is the day length. In December the situation is the opposite.

The largest uncertainty in applying the method comes from the assumption of Sun elevation from light level. This is because of the unknown amount of shading present causing an unknown amount of attenuation of solar irradiance (e.g. cloud, weather, terrain, foliage etc). In contrast, the conversion of the timestamped Sun elevation level to location can be calculated quite accurately from well known astronomical algorithms (e.g 'Astronomical Algorithms' by Jean Meeus, 2nd ed 1998), except for latitude at equinox. Two position fixes per day can be calculated based around local noon and local midnight.

Longitude is usually more accurately determined than latitude. Latitude has greater error close to equinox (when threshold geolocation cannot be used alone to calculate latitude) and close to the equator where seasonal variations resulting in daylength variations are minimal.

The critical times for our devices are dusk and dawn. When shading issues are to be considered, it is at these times that you must consider the habit and habitat of the bird. Shading at any other time during the day is irrelevant for threshold analysis.


The device

Mk4 geolocator on albatross legThe logger (geolocator) is a battery powered instrument with an in-built microprocessor and a memory for data storage. It is encapsulated in a clear, water resistant package, with two external terminals for command and data transfers. The recorder can be reused for the life of the battery (non replaceable). It may be put into a recording mode for collecting and delivering data, or into a sleep mode to save battery power for longer shelf life (note that sleeping only reduces battery consumption by about 20%). When in recording mode it may be attached onto the bird.

The logging engine takes a light level reading at certain intervals (for our loggers usually every minute) and then saves to memory the maximum light measurement every recording interval (for our loggers this can be 10, 5 or 2 minutes). Currently, light measurement is 6-bit and the range is limited to dim light levels.

A number of our loggers also record wet/dry information. This is achieved by measuring conductivity; hence the 'wet' state will only be recorded if the water is salty. Most of our devices sample for wet/dry every 3 seconds and make a record of the total number of samples wet every 10 minutes. Some of our larger loggers equipped with wet/dry recording, also record temperature to allow satellite sensed sea surface temperature (SST) to aid geolocation. For these devices, a temperature record is taken 20 minutes after a change from the dry to continuous wet state.


Application

serial interface box for downloading geolocation dataWhenever the deployment of a tag on a flying bird is considered, the size and weight of the instrument are two of the most important factors. The objective is to reduce any effect of the tag on the bird by keeping size, drag and weight of the instrument to a minimum. For greatest success, attachment and retrieval is usually performed at sites where the birds are easiest to recapture; usually at the breeding location.

Leg mounting the devices onto a Darvic (now Salbex) ring has been most popular with seabirds and large waders. For smaller species where the leg is relatively weak and terrestrial species, back mounting is recommended. Back mounting allows a greater attachment:body weight ratio and also positions the light sensor above the bird so that shading by the bird does not contribute to error. The use on terrestrial species is still at an early stage; most uncertainty comes from not knowing how much foliage shading will effect the data. Studies over the next few years should also give us feedback on the best mounting arrangements and materials. Early studies (see publications) have already shown that dominantly aerial birds such as the Purple Martin will produce much cleaner data than, for example, Wood Thrush. The good news is that most birds, when migrating, will spend considerable time in the air and so away from sources of shading.

To operate these recorders, a general purpose computer (eg. PC compatible, Macintosh, etc.) and an interface box are required.The interface box is purpose-built instrument. It has a 2-wire port with two crocodile clips for connecting it to a logger and a  USB connector for the PC.

We supply software with our interface boxes to decompress logger data, allow the user to analyse the data, and to calculate location fixes. Substantially more expensive, alternative software is available from Jensen Software called Multitrace.





 
 
  Site Map