MARB 403 Lab 6

« previous | Tuesday, January 8, 2012 | next »

Telemetry

Authors:

• Andrew J. Read

Obtaining data remotely by transmitting or storing for retrieval.

• radio tags
• From office

Archival tags (records data and then stores it for retrieval) Transmitting (sends data to be collected elsewhere)

• early devices: Small gadgets (e.g. kitchen timers)
• mordern are much more sophicsticated: water temperature, depth, light level, etc.
• physiological data: heart rate, body temp, feeding (stomach temp)

DTAG: digital acoustic tags

• Developed by Mark Johnson and Peter Tyack, Woods Hole Oceanographic Institution
• Records mammal acoustics and surrounding sounds
• used in North Atlantic Right whales response to ship alert systems

Pinnipeds are easy to recover, but cetaceans are harder to find. Suction cups or fired into blubber

Limitations:

• Radio systems only work when above water
• Short battery life
• Short radio transmission distance
• large size can be uncomfortable

Satellite linked transmitters

Marine mammals spent 95% of time submerge

Allows us to see what goes on unseen

• mating, feeding, predator avoidance
• brain fart lol (dolphin meme)
• Deep-diving marine mammals
• Crittercams document behavior

Future development

• Make smaller, more sensitive instruments
• Collect even more information

Problems

Increased data set size (difficult to analyze)

How do animals feel about it?

Behavioral sampling ...

Recording cetacean behavior is a challenge:

• travel fast
• submerged
• long travels
• disappear during dives
• no trails

No standardized way to record telemetry

Analysis of 74 journal articles

which subjects? how long?

Follow Protocols

1. survey
   • what is cetacean doing right now?
2. group follow
   • most common
   • fission fusion society?
   • observe behavior over 30 minutes
3. individual follow
   • Follow 1 animal in a group
   • Who does individual approach/avoid? (animal perspective)
   • How quickly can that animal be picked out of group
4. tracking
   • tags, hydrophones
   • continual data recording
   • expensive, so not widely used
   • attachment can affect behavior
5. anecdote
   • description of a single event
   • usually bizarre

Sampling Methods

3 basic "categories" for behavior:

1. State: are long behaviors of measurable duration (e.g. foraging, traveling)
   
2. Event: brief behavior, measured in frequency (e.g. fish in mouth)
   
3. Bout: series of short events (e.g. fish-flipping over and over)

Sampling Methods:

Ad lib

most common

"typical field notes"

better for sampling when signalers and recipients can be determined

Continuous

systemic record of frequencies or durations for a specified set of behaviors

Altmann recommends using this method on only for 1-2 individuals at a time (Don't miss behavior activities)

Successful when individuals: are rapidly identifiable, live in small groups, dive for short periods

richest source of info on social behavior and relationships

Focal Gathering

continuous assessment of group activity

"focal subgroup sampling" (continuous sampling with more than one individual)

Accuracy determined by group size, cohesiveness, and animal activites

One-Zero Sampling

Scoring whether behavior occurs during an interval

High rates of sampling error

Does not represent frequency or duration

Point Sampling

Scoring activity as 'snapshot' at a given moment

"point" often occurs underwater

"point" is often missed, when does it happen? NOW! no THEN!

Scan Sampling

Point or instantaneous sample of individuals' behavior or location before moving to next

Interval sampling; midpoint determines behavior

Used 3% of time in studies

Difficulty: in time it takes for researcher to decide behavior, animal's behavior may have changed

Scan fixed intervals to asses groups

Assess nearest neigbors

Identify activity based on surfacing individuals

Predominant Activity Sampling (PAS)

Scoring individual behavior as predominant activity over some interval (e.g. orcas spend 40% of time spy-hopping)

Useful to sample animals going in and out of view for brief periods

Incident Sampling

Scoring all behavioral events of a specific type in a group

Must be infrequent for adequate records

Distinctions of individuals may or may not be made

Focus on dramatic/recognizable events with few animals

Sequence Sampling

Focus on sequences of behavior or particular interaction

Record even from start to finish regardless of other events around

Excellent to determine conditional probabilities of behavior sequences

Recommendations

Shortcomings/problems

• Large portion used ad lib. sampling, which is a biased way of sampling things.
• Lack protocol and subject info (important to indicate # of animals and observation length)

Scan, incident, sequencing, and PAS are recommended