Observing Fish

1
Observing Fish Populations

• Goals
• Identify and discuss potential sources of
  fisheries data
• Explain fisheries survey methodology
• stratification
• systematic vs random
• Discuss the use of tagging data in fisheries
• Discuss the implication of fish distributions
  to CPUE abundance indices

2
A Simple Problem In Fisheries Management
Communication

• Keep Informal communication open
• Know where numbers come from

3
Sources of Data
4
Observed Data

• Encounter rates
• CPUE
• estimate N and or F
• count per transect
• Sampled Catch
• Age estimates
• from hard parts
• validation important
• length
• weight
• reproductive condition
• genetics, morphometrics
• stomach contents
• Location
• infer movement

5
Data From Commercial Fisheries

• Observers
• assume fishing unchanged
• usefully for discard studies
• difficult to randomly subsample

• Port Sampling/Creel Survey
• collected from processors
• collected by agent
• tends to aggregate spatially and temporally

• Log Books
• assume accurate reporting
• lacks detailed age/spp. data
• can get detailed spatial data
• potentially high coverage

6
Patterns in Commercial CPUE
CPUE
c.
b.
a.
Abundance

• a. Proportional
• ideal situation
• b. Hyperstable (CPUE poor index)
• constant density in schools
• handling time constraints
• c. Hyperdepletion (CPUE safe index)
• refuge (area, size, etc.)

7
Suggested Use of Commercial Catch and Effort Data

• Map the CPUE values
• sketch CPUE distribution
• Know your fishery
• how is it conducted?
• how could this influence CPUE?
• Use additional information if possible
• fish movements between areas
• tagging studies
• survey data
• operational studies of fishing gear
• DO NOT USE CPUE TO CALCULATE
• A SINGLE, SIMPLE INDEX OF
  ABUNDANCE!!!

8
Research Surveys

• Use dedicated equipment or charters
• Expensive (operating expenses)

9
Survey Characteristics
10
Survey Design

• Stocks range is often heterogeneous
• different bottom types
• different currents
• different gear efficiencies

• Stratification
• A stratum is defined to be more homogeneous
  than the whole area sampled
• Results in more precise estimates
• if variation within stratum is less than random
  sample over whole survey area

11
Stratum Boundaries

• set apriori
• should not be changed within survey season
• stratum boundary becomes non-random
• may bias variation estimate (appear too
  uniform)
• can redefine strata between survey seasons
• BUT caution for comparisons between seasons

• INDEX weighted mean of strata
• proportional to stratum area
• inversely proportional to stratum uncertainty

12
Systematic vs Random Sampling

• Systematic PROS
• efficient to implement
• better for mapping
• minimize risk of missing aggregations
• facilitates comparisons between survey seasons
• more precise Iff Var among obs. is greater than
  random obs.

• Systematic CONS
• easily biased by gradients
• precision not easily defined
• resample
• IN CONTRAST standard error easily estimated for
  random samples

13
Random designs can beimplemented on large scales.

• First Time Survey Design in Stages
• first pass define strata
• fast, rough (sonar?)
• second stage estimate abundance
• more precise gear/protocol
• weighted estimate
• third stage / more stages
• subsample or resample as required for research
  questions

14
Research Tagging

• Used to estimate
• abundance
• survival
• movement
• fishing mortality

• Usually very expensive
• 5,000 per yellowfin tag return for South
  Pacific Tuna Comm.
• more expensive with trackable tags
• Low emphasis on collection
• Hard to determine Parameters
• tag related mortality
• tag shedding

15
Some Tagging Issues

• Double Tagging to estimate
• shedding rate
• tagging mortality
• Post-tagging holding
• estimate tagging mortality
• why might this not work well??
• Tag under reporting
• agent sampling of catches
• statistical techniques
• reward programs
• Not recommended for abundance
• Recommended for movement