28
Statistical guidelines for sampling marine avian populations Statistical guidelines for sampling marine avian populations Elise F. Zipkin Brian Kinlan Allison Sussman Mark Wimer Allan F. O’Connell Elise F. Zipkin Brian Kinlan Allison Sussman Mark Wimer Allan F. O’Connell USGS Patuxent Wildlife Research Center NOAA National Ocean Service 4 th International Wildlife Management Conference – July 2012

Statistical guidelines for sampling marine avian populations... · Statistical Guidelines for sampling marine avian populations Zipkin_IWMC_2012_Session 30.ppt [Compatibility Mode]

  • Upload
    others

  • View
    14

  • Download
    0

Embed Size (px)

Citation preview

Page 1: Statistical guidelines for sampling marine avian populations... · Statistical Guidelines for sampling marine avian populations Zipkin_IWMC_2012_Session 30.ppt [Compatibility Mode]

Statistical guidelines for sampling marine avian populations

Statistical guidelines for sampling marine avian populations

Elise F. ZipkinBrian Kinlan

Allison SussmanMark Wimer

Allan F. O’Connell

Elise F. ZipkinBrian Kinlan

Allison SussmanMark Wimer

Allan F. O’Connell

USGS Patuxent Wildlife Research Center NOAA National Ocean Service

4th International Wildlife Management Conference – July 2012

Page 2: Statistical guidelines for sampling marine avian populations... · Statistical Guidelines for sampling marine avian populations Zipkin_IWMC_2012_Session 30.ppt [Compatibility Mode]

Seabirds in the Atlantic

Page 3: Statistical guidelines for sampling marine avian populations... · Statistical Guidelines for sampling marine avian populations Zipkin_IWMC_2012_Session 30.ppt [Compatibility Mode]

Where are the birds?

Not a lot known about the distribution and abundances in the Atlantic

• Difficult to survey• Rough conditions• Patchily distributed• Highly mobile

Page 4: Statistical guidelines for sampling marine avian populations... · Statistical Guidelines for sampling marine avian populations Zipkin_IWMC_2012_Session 30.ppt [Compatibility Mode]

Off shore wind power garnering lots of interest

• Many states have implemented a 20% renewable energy by 2020 mandate

• Public perception of oil spills is poor

Where are the birds? Wind development

Page 5: Statistical guidelines for sampling marine avian populations... · Statistical Guidelines for sampling marine avian populations Zipkin_IWMC_2012_Session 30.ppt [Compatibility Mode]

U.S. Bureau of Ocean and Energy Management (BOEM) 

• 5km x 5km lease blocks

• Along the Outer Continental Shelf of the Atlantic Ocean 

All Lease Blocks

Patuxent Wildlife Research Center

Page 6: Statistical guidelines for sampling marine avian populations... · Statistical Guidelines for sampling marine avian populations Zipkin_IWMC_2012_Session 30.ppt [Compatibility Mode]

Objectives

Develop a framework for assessing:

1) which lease blocks are “hot spots” and “cold spots”

2) the required surveying effort to guide BOEM and industry in determining wind turbine placement

Page 7: Statistical guidelines for sampling marine avian populations... · Statistical Guidelines for sampling marine avian populations Zipkin_IWMC_2012_Session 30.ppt [Compatibility Mode]

What is a hot/cold spot?

Hot spot  = A lease block with an average species specific abundance that is three times the mean of the region 

Cold spot = A lease block with an average species specific abundance that is one third the mean of the region 

Page 8: Statistical guidelines for sampling marine avian populations... · Statistical Guidelines for sampling marine avian populations Zipkin_IWMC_2012_Session 30.ppt [Compatibility Mode]

The Atlantic Seabird Compendium• >250,000 seabird observations from U.S. Atlantic waters

• Collected from 1978 through 2011

• Data collected using a mix of methods including non‐scientific approaches

Page 9: Statistical guidelines for sampling marine avian populations... · Statistical Guidelines for sampling marine avian populations Zipkin_IWMC_2012_Session 30.ppt [Compatibility Mode]

The Atlantic Seabird Compendium• >250,000 seabird observations from U.S. Atlantic waters

• Collected from 1978 through 2011

• Data collected using a mix of methods including non‐scientific approaches

We used:• 32 scientific data sets – 28 ship‐based,  4 aerial

• Transects were standardized to 4.63km

• 44,176 survey transects representing 463 species 

Page 10: Statistical guidelines for sampling marine avian populations... · Statistical Guidelines for sampling marine avian populations Zipkin_IWMC_2012_Session 30.ppt [Compatibility Mode]

Two part approach

1) Determine the best statistical distribution to model the count data for each species in each season

2) Use the best fitting distribution to produce power analyses

Page 11: Statistical guidelines for sampling marine avian populations... · Statistical Guidelines for sampling marine avian populations Zipkin_IWMC_2012_Session 30.ppt [Compatibility Mode]

The rest of the talk

1) Describe the broad two part approach 

2) Integrate an example using Northern Gannets

Page 12: Statistical guidelines for sampling marine avian populations... · Statistical Guidelines for sampling marine avian populations Zipkin_IWMC_2012_Session 30.ppt [Compatibility Mode]

Two part approach

1) Determine the best statistical distribution to model the count data for each species in each season

2) Use the best fitting distribution to produce power analyses

Page 13: Statistical guidelines for sampling marine avian populations... · Statistical Guidelines for sampling marine avian populations Zipkin_IWMC_2012_Session 30.ppt [Compatibility Mode]

Part 1: Model the data

Test eight statistical distributions:

Poisson                                    Negative binomial         Geometric                          Logarithmic                 Discretized lognormal        Zeta decay                        Yule                                                         Zeta (power law)

Northern Gannet spring count data

Page 14: Statistical guidelines for sampling marine avian populations... · Statistical Guidelines for sampling marine avian populations Zipkin_IWMC_2012_Session 30.ppt [Compatibility Mode]

Examples of the distributions

1 2 5 10 20

1e-0

51e

-03

1e-0

1

Positive Poisson (simulated)

1 2 5 10 20 50 100 200

1e-0

51e

-03

1e-0

1

Positive neg binomial (simulated)

1 2 5 10 20 50 100

1e-0

51e

-03

1e-0

1

Positive geometric (simulated)

1 2 5 10 20 50 100 200

1e-0

51e

-03

1e-0

1

Logarithmic (simulated)

1 5 10 50 100 500 1000

1e-0

51e

-03

1e-0

1

Discretized lognormal (simulated)

1 100 10000

1e-0

51e

-03

1e-0

1

Zeta (simulated)

1 100 10000

1e-0

51e

-03

1e-0

1

Yule (simulated)

Page 15: Statistical guidelines for sampling marine avian populations... · Statistical Guidelines for sampling marine avian populations Zipkin_IWMC_2012_Session 30.ppt [Compatibility Mode]

Part 1: Results

Spring Summer Fall  Winter Total

Number species with >500 observations 12 10 15 11 48

Page 16: Statistical guidelines for sampling marine avian populations... · Statistical Guidelines for sampling marine avian populations Zipkin_IWMC_2012_Session 30.ppt [Compatibility Mode]

Part 1: Results

Spring Summer Fall  Winter Total

Number species with >500 observations 12 10 15 11 48

Discretized lognormal 

Yule

Negative binomialLogarithmicZeta decay

Page 17: Statistical guidelines for sampling marine avian populations... · Statistical Guidelines for sampling marine avian populations Zipkin_IWMC_2012_Session 30.ppt [Compatibility Mode]

Part 1: Results

Spring Summer Fall  Winter Total

Number species with >500 observations 12 10 15 11 48

Discretized lognormal  7 (4*) 4 (3*) 8 (3*) 8 (2*) 27 (12*)

Yule 1* 3* 1* 1 1 (5*)

Negative binomialLogarithmicZeta decay

3* 0 (3*)

Page 18: Statistical guidelines for sampling marine avian populations... · Statistical Guidelines for sampling marine avian populations Zipkin_IWMC_2012_Session 30.ppt [Compatibility Mode]

Part 1: Results

Northern Gannet

Discretized lognormal top distribution for fall and spring

Discretized lognormal and Yule fit equally well in winter and summer

1 5 10 50 500

1e-0

41e

-03

1e-0

21e

-01

1e+0

0

Count (log scale)

Prob

abili

ty (l

og s

cale

)

Discretized lognormalYuleZeta decayZeta

Page 19: Statistical guidelines for sampling marine avian populations... · Statistical Guidelines for sampling marine avian populations Zipkin_IWMC_2012_Session 30.ppt [Compatibility Mode]

Two part approach

1) Determine the best statistical distribution to model the count data for each species in each season

2) Use the best fitting distribution to produce power analyses

Page 20: Statistical guidelines for sampling marine avian populations... · Statistical Guidelines for sampling marine avian populations Zipkin_IWMC_2012_Session 30.ppt [Compatibility Mode]

Part 2: Power analysis

Page 21: Statistical guidelines for sampling marine avian populations... · Statistical Guidelines for sampling marine avian populations Zipkin_IWMC_2012_Session 30.ppt [Compatibility Mode]

Part 2:    Power 

analysis for Northern gannets in the spring

*Focusing only on lease blocks where individuals were observed

Page 22: Statistical guidelines for sampling marine avian populations... · Statistical Guidelines for sampling marine avian populations Zipkin_IWMC_2012_Session 30.ppt [Compatibility Mode]

Part 2: Northern gannet results

Reference mean = 6.9 individuals per lease block conditional on presence

Page 23: Statistical guidelines for sampling marine avian populations... · Statistical Guidelines for sampling marine avian populations Zipkin_IWMC_2012_Session 30.ppt [Compatibility Mode]

5 10 15 20 25

0.0

0.2

0.4

0.6

0.8

1.0

Number of sampling events

Sim

ulat

ed p

ower

Hot spot (3 x mean)Cold spot (0.33 x mean)

Part 2: Northern gannet results

Reference mean = 6.9 individuals per lease block conditional on presence

Page 24: Statistical guidelines for sampling marine avian populations... · Statistical Guidelines for sampling marine avian populations Zipkin_IWMC_2012_Session 30.ppt [Compatibility Mode]

5 10 15 20 25

0.0

0.2

0.4

0.6

0.8

1.0

Number of sampling events

Sim

ulat

ed p

ower

Hot spot (3 x mean)Cold spot (0.33 x mean)

Part 2: Northern gannet results

1 5 10 50 100 500 1000

1e-0

51e

-03

1e-0

1

Counts

Freq

uenc

y

Discretized lognormal

Reference mean = 6.9 individuals per lease block conditional on presence

Page 25: Statistical guidelines for sampling marine avian populations... · Statistical Guidelines for sampling marine avian populations Zipkin_IWMC_2012_Session 30.ppt [Compatibility Mode]

Part 2: Northern gannet results

Page 26: Statistical guidelines for sampling marine avian populations... · Statistical Guidelines for sampling marine avian populations Zipkin_IWMC_2012_Session 30.ppt [Compatibility Mode]

Summary of results

• Seabirds tend to be highly aggregated and require skewed statistical distributions to accurately describe populations

• For many species, we                                                  need a large number of                                             surveys to detect areas                             with atypical abundances

Page 27: Statistical guidelines for sampling marine avian populations... · Statistical Guidelines for sampling marine avian populations Zipkin_IWMC_2012_Session 30.ppt [Compatibility Mode]

Implications for wind power

• Intensive sampling in multiple seasons will be required to determine potential impacts on seabirds

• A possible approach could be to combine data on functionally similar species or species of high conservation value 

Page 28: Statistical guidelines for sampling marine avian populations... · Statistical Guidelines for sampling marine avian populations Zipkin_IWMC_2012_Session 30.ppt [Compatibility Mode]

Acknowledgments

• The many researchers and their crews who collected the data used in our analyses

• Emily Silverman, Diana Rypkema

• The Bureau of Ocean, Energy, Management (BOEM) for funding model development and analysis