OK, so assume the GPS receiver takes 60 milliwatts [1] for 60 seconds, four times a day. You'd be using 0.66 milliwatt-hours per day. (Other receivers claim to use as little as 8 milliwatts [2])

A D cell alkaline battery, depending on the manufacturer, can have 20,000 milliwatt-hours or more [3].

Two of those and you could have 4 years of battery life. Assuming your batteries retain their advertised capacity when dragged around in snow and ice by a wolf for a few years!

[1] https://en.wikipedia.org/wiki/SiRFstarIII [2] http://www.csr.com/products/25/sirfstariv-gsd4t [3] https://en.wikipedia.org/wiki/D_battery

Temperature certainly has an effect on batteries¹. But also that the periods of high current demand were spaced so far apart -- the batteries had time to recover. Much like how the battery in your car is a "starting" battery -- high current is needed only occasionally to start the engine, with the alternator making up the demand the rest of the time.

¹ Dad used to work for Union Carbide, which owned the "Eveready" brand at the time. One of the projects he worked on was a battery for the Union Pacific Railroad. It was designed to be buried in the ground in remote locations to run signal systems. Burying it gave it a more constant temperature. If the researchers placed the battery on the inside of the collar next to the wolf's fur, that could really extend it's life.

Ah, spotted an error there. If you were using 0.66 milliwatt-hours per day and had two 20,000 milliwatt-hour batteries, it would of course last for 60,606 days. Which is 160 years not 4 years.

Of course, you'd also need to power the radio module to send the SMS, and presumably wolves spend time in remote regions with poor cell phone coverage Products like the Quectel M10 can use as much as 2 watts while transmitting [1] (or as little as a fifth of that). To make 40,000 milliwatt-hours of batteries last for 5 years would only give you a budget of 21 milliwatt-hours (75 joules) per day. That would let your run a 2 watt transmitter for about 37 seconds a day - so to make 8 transmissions a day you'd have to power on, join the network, transmit, and power down again in less than 5 seconds.

On the other hand, according to [2] it takes between 1.5 and 5.5 joules to send an SMS, allowing us to send several in a budget of 75 joules per day. They don't make much mention of signal strength or the energy cost of connecting to the network, unfortunately.

So I guess it all comes down to how long you have to have your radio module turned on in order to get a message off.

There are also tracking collars with solar cells [3].

[1] http://arduino.cc/en/uploads/Main/Quectel_M10_datasheet.pdf [2] http://mobiledevices.kom.aau.dk/research/energy_measurements... [3] http://www.rvc.ac.uk/SML/People/awilson/Wildlife-tracking-co...

Does SMS use GPRS? The article says it sent its position by SMS.

SMS doesn't use GPRS, it's sent on a special channel

SMS uses the control channel of GSM.

AGPS requires GPRS data.