Документ взят из кэша поисковой машины. Адрес оригинального документа : http://www.sevin.ru/seminars/argos/TRACK%20and%20LOC.pdf Дата изменения: Fri Mar 13 16:48:00 2009 Дата индексирования: Tue Oct 2 19:06:43 2012 Кодировка:

Track & Loc Service Fish Tag Processing
CLS, Christian Ortega
On behalf of Philippe Gaspar, head of Marine Ecosystem Modeling & Monitoring by Satellite (MEMMS) at CLS


Overview

· About Fish tags · Light based geo-location, principles, constraints · Enhancing light based data · Track & Loc service, how it works · Products delivered · Experience & References


About Fish Tags

How pop-up and archival tags work:
All tags measure pressure + temperature+ light level Pop-ups detach from the animal and transmit PART of the recorded measurements through ARGOS Recovery of archival tags is possible only if the fish is recaptured...but then ALL measurements are recovered

Pop-up tag

Archival tag


How Fish Tags measure locations

Principles of light-based geolocation
·Estimates of sunrise and sunset times allows to geolocate the tag Precision in longitude almost constant over the year

Precision in latitude drops at the equinoxes


Light based geolocation

In theory...

High sensitivity of geolocation accuracy to sunset/sunrise estimates

In a perfect world, a 3min error for sunrise/sunset is the best we can expect...


Theory vs. practice for light-based geolocation

In practice...

Errors reach 1 degree in longitude and 3-5 degrees in latitude Many sources of error: water clarity, fish vertical behavior, gaps in ARGOS transmissions, equinoxes, coding issues... Need to filter the data to reduce geolocation error


CLS Tag Processing Service

Objective : improve light-level based positioning by using additional constraints on the position Technique based on algorithms developed by F. Royer during his thesis at CLS/Ifremer and his Postdoc at University of New Hampshire.


Fish tag processing at CLS: How it works 26° 25°
Position Error derived from Light data Using Sea Surface Temperature (SST)

24°
Maximum distance traveled in one day

D Using Bathymetry

D-1 100 m 200 m 300 m

Using Animal Motion


A realistic tracking of Bluefin tuna in the Atlantic


Theoretical Background
More formally, the geolocation problem is formulated in a state space theory context and solved using a Kalman filtering /smoothing approach (Royer et al., 2004; Royer & Lutcavage, 2008) based on : · a simple movement model · observations including light-level, SST & bathymetry
Light level

time

Trade off between noise smoothing and data fitting

SST

time


CLS Tag Processing Service
Correcting for light attenuation at depth... Depth-corrected light

Raw light

Depth


Theory vs. practice for light-based geolocation
Correcting for large errors at the equinoxes


Main steps of a tracking study
1. Get tags + Argos SUA 2. Tag fish!

3. Wait ... to retrieve tag data
4. Retrieve data from Argos 5. Process data
Send data to Microwave Telemetry Through Wildlife Computers' software

6. Improve location accuracy 7. Publish!
User 's responsibility CLS's Track&Loc service


Decoding Argos Raw data
Satellite Transmitted Data ­ Not Usable Directly Decoding of Individual Datasets using Manufacturer 's Software

Decoding Geolocation Filtering Delivery

Example:

Wildlife Computers MK10


Light-based geolocation
Light-level data : first Location Estimation Using Manufacturer 's Software, Record by Record Can be Subjective

Decoding Geolocation Filtering Delivery


Light-based geolocation

Decoding Geolocation Filtering Delivery « Raw », Unconstrained Light-based Geolocations


Reduction of Location Error
Before Filtering
55
55

After Filtering

Decoding Geolocation Filtering Delivery

50

50

45

45

40

40

35

35

30

30

25

25

20 -10

-5

0

5

10

15

20

20 -10

-5

0

5

10

15

20


Location Uncertainty

Decoding Geolocation Filtering Delivery 50% CI 95% CI

Most Probable track


Sea Surface Temperature

Decoding Geolocation Filtering Delivery


Bathymetry
First Argos Message

Decoding Geolocation Filtering Delivery

Probable Detachment (tag at surface)


Delivery : Raw Data
Final Dataset: Raw + Preprocessed Data Binned (Temperature, Depth)

Decoding Geolocation Filtering Delivery


Delivery : Processed Data

Decoding Geolocation Filtering Delivery



Excel File Interpolated Data Positions + Uncertainty Sea Surface Temperature Depth, Speed (min/max)


Interest of Track & Loc for biologists

· Better results, without the effort

· Avoiding the technicalities of tag processing
· Time-saving · Concentrating on interpreting the result &

biology study


References

· European Tuna Tagging program (ETTP) (>100 pop-ups) · Indian Ocean Tuna Commision (40 pop-ups) · Secretariat of the Pacific Community (Oceanic Fisheries Programme) : 200 internal tags over 5 years

+ Ogasawara Fisheries Center, Irish Sea Fisheries Board, WWF, Univ. Of Cadiz.....


THANKS FOR YOUR ATTENTION