Page 14
The information contained in these documents is confidential, privileged and only for the information of the intended recipient and may
not be used, published or redistributed without the prior written consent of Wildlife Computers.
Sampling
Interval
(seconds)
Argos Messages Per Day
Message
Period
(hours)
Depth or Temp
(One channel)
Depth and Temp
(Two channels)
75
24
48
1
150
12
24
2
300
6
12
4
450
4
8
6
600
3
6
8
Figure 11—The number of Argos messages created per day and the corresponding message sample period for the five time-
series sample intervals.
The sample interval determines the number of messages generated per day. Selecting a short (75-second)
time-series sampling interval will rapidly generate many messages. This may be appropriate for a short
deployment. However, if too many messages are generated, they may not all be received. The result will be
random gaps of time during the deployment for which there is no time-series data. Selecting a longer
sampling interval will improve the odds that there will be no gaps; however, the temporal resolution of
each datum will be reduced. Different study objectives will warrant different trade-offs between coverage
and temporal resolution.
Each time-series message contains:
•
The time-series data sampled at the specified interval.
•
The minimum and maximum values encountered during the period of time covered by the time-
series message period as measured at the fast archive sampling rate.
The absolute Min/Max values and point sample values may not match as the absolute values are
determined from all archived data collected during the message period. This can give insight to
the amount of aliasing that has occurred when generating the time-series message.
Duty-cycling can be used to reduce the number of time-series messages generated. Duty-cycle settings do
not affect the tags archive sample rates.
Example Time-Series Sampling Settings
For the MiniPAT tag, on average approximately 2,000 Argos messages are received following release. With
this number in mind, one can work backwards to determine how quickly time-series can be sampled given
a preferred deployment length. Following are some typical set-ups:
