DesignbasedModelassisted Survey Methodology for Aquatic Resources

1
Design-based/Model-assisted Survey Methodology
for Aquatic Resources

• Don L. Stevens, Jr.
• Presented at
• SECOND ANNUAL CONFERENCE
• STATISTICAL SURVEY DESIGN AND ANALYSIS
• FOR AQUATIC RESOURCES
• Department of Statistics
• Oregon State University
• August 11-12, 2003

2
This presentation was developed under STAR
Research Assistance Agreement No. CR82-9096-01
awarded by the U.S. Environmental Protection
Agency to Oregon State University. It has not
been formally reviewed by EPA. The views
expressed in this document are solely those of
the author and EPA does not endorse any products
or commercial services mentioned in this
presentation.
3
Discussion Outline - 2002

• Program description
• What we are trying to do
• Program context
• Where we are how we got here
• Integration Outreach
• Who we are working with
• Accomplishments

4
Discussion Outline - 2003

• Program description
• What we are trying to do
• Program context
• Where we are how we got here
• Accomplishments
• Programmatic
• Outreach
• Research

5
Most Significant Programmatic Accomplishments

• We devised an acronym
• DAMARS
• Program on
• Designs And Models for Aquatic Resource Surveys

6
Most Significant Programmatic Accomplishments

• We devised a logo

7
DAMARSatYEAR 1.75

• Don L. Stevens, Jr.

8
Discussion Outline

• State of the Program
• Outreach(DAMARS STARMAP)
• State of the Research
• Synopsis Integration
• Preview of Presentations

9
State of the Program

• Personnel
• OSU Faculty
• Steve Carroll leaving DAMARS
• Alix Gitelman
• Alan Herlihy
• Virginia Lesser on sabbatical in UK
• Paul Murtaugh leaving DAMARS
• Don Stevens

10
State of the Program

• Personnel
• CSU Faculty
• Scott Urquhart
• Jay Breidt
• ISU Faculty
• Jean Opsomer
• UW Faculty
• Loveday Conquest

11
State of the Program

• Personnel
• OSU Research Associates
• Breda Munoz-Hernandez
• Ruben Smith
• Jeannie Sifneos
• Dan Dalthorpe joining DAMARS this fall

12
State of the Program

• Personnel
• OSU Graduate Students
• Cynthia Cooper Leigh Ann Harrod
• Kathy Georgitis Susan Hornsby
• CSU Graduate Students
• Siobhan Everson-Stewart Mark J. Delorey
• Alicia A. Johnson
• UW Graduate Students
• Rebecca Buchanan

13
State of the Program

• Personnel
• OSU Graduate Students
• Cynthia Cooper Leigh Ann Harrod
• Kathy Georgitis Susan Hornsby
• CSU Graduate Students
• Siobhan Everson-Stewart Mark J. Delorey
• Alicia A. Johnson
• UW Graduate Students
• Rebecca Buchanan

14
Program Objectives

• To support advances in (statistics) and
  hierarchical survey design and analysis and
  spatial and temporal modeling
• Describe current state of aquatic ecosystems
• Detect quantify change and trend
• Predict at non-sampled locations
• Identify relationships between stressors
  condition (cause/effect)

15
Program Objectives

• To develop and extend the expertise on design and
  analysis to States and Tribes
• distance learning
• seminars
• workshops
• demonstrations
• Both DAMARS STARMAP have same objectives
• STARMAP -- learning materials
• DAMARS -- demonstration projects

16
State of the Program

• Success depends on having States, Tribes, ( EPA)
  use techniques tools
• Foster client use of both design analysis
  tools
• Cooperative agreement
• Work with EPA
• Work on EPA-sponsored projects
• Use data generated by EPA projects

17
State of the Program

• Research is driven by issues that arise in
  aquatic monitoring
• Indicator development/Monitoring Design/Analysis
  (huge area)
• Research topics
• Identified in RFA our experience
• Arise from collaboration with EPA, State, other
  STAR researchers

18
State of the Program

• Collaborative statistical research with another
  statistician
• Common perspective
• Complementary backgrounds
• Example Stevens Olsen
• History of joint papers presentations
• Continuingpresentation on GRTS extensions

19
State of the Program

• Collaboration that brings statistical perspective
  to multi-disciplinary research team
• Source for new challenges that drives the
  development of new methodology theory
• Requires high level of commitment
• Substantial time requirement
• Example GLEI, an EaGLE STAR Program

20
Collaboration with GLEIGreat Lakes Environmental
Indicators

• Stevens, Munoz, Smith spent a week visiting
  GLEI in February
• Identified areas where our skills could
  complement GLEIs
• Species-area curves wetland loss
• Sifneos, White, Anderson poster
• Probability distribution of condition indicator
• Stevens JSM presentation
• Statistical basis of ecological indicators

21
State of the Program

• Demonstration Projects
• Archetypes used as models
• Real-life aquatic monitoring by real State
  agencies
• Push the envelope of State-level monitoring
• Design to State-articulated needs
• Make them succeed!

22
Outreach- Demonstration

• Surface Waters Ambient Monitoring Program (SWAMP)
• San Francisco Estuary Regional Monitoring Program
  for Trace Substances (RMP)
• West Coast Tidal Wetland Monitoring and
  Assessment Venture (CRAM)
• Sampling Coho salmon in Oregon coastal streams
  (ODFW)

23
SWAMP

• Run by California State Water Resource Control
  Board (Cal EPA)
• Low level of activity this year
• Budget
• Management change
• Workshop Understanding Surface Water Monitoring
  Requirements
• SWAMP SPARC
• Scientific Planning and Review Committee

24
RMP

• Designed monitoring plan for San Francisco Bay
• Pulse of the Estuary, 2000 Update
• Re-design team SFEI, USEPA Region 9, DAMARS,
  USGS, others
• Nice example of using prior information to guide
  design
• Implemented in 2001
• Rotating panel GRTS
• Intensity varies by Bay segment
• Separate designs for water column sediment

25
RMP

• Plans for an analysis workshop later this
  fall/winter
• Draft report on re-design
• Renewed activity
• Should result in several spin-off articles
• Sarah Lowe, SFEI, will attend Wednesdays workshop

26
CRAMCalifornia Rapid Assessment Method

• Funded by EPA
• Joint effort
• SFEI CCC WED
• SCCWRP DAMARS Region 9

27
CRAMCalifornia Rapid Assessment Method

• Modeled on OhioRAM
• Extended to cover CA
• Salt marshes
• Tidal influenced
• DAMARS (Stevens) represented on the Core
  Development Team
• Metric/indicator development
• Planning for verification/validation study
• Pilot assessment

28
ODFW

• Focus is on Coho Salmon
• High visibility issue in NW (OR, WA, CA, BLM,
  BPA, USFS, USFWS, NOAA, USEPA)
• Rotating panel GRTS is basic sampling design for
  The Oregon Plan for Salmon Watersheds
• ODFW scientists very willing to share their
  scientific insight, their time, and their data
• ODFW, OWEB, ODEQ have become our advocates

29
ODFW

• Ideal test bed for design analysis
• Historical data, both probability convenience
• Dynamic frame
• Missing data, ignorable non-ignorable
• Ancillary data
• Rotating panel through time
• Rich source of analysis issues

30
ODFW

• Response (population level) influenced by many
  factors
• Ocean conditions
• Watershed conditions
• Weather
• Bottom line Are management efforts having an
  impact?

31
ODFWIntegrated Monitoring Design

• There are two primary objectives of the sampling
• describe current status
• describe population trend
• Coho adult juveniles, habitat, water quality
• Relevant Information
• There is spatial pattern in the population
• Coho have a 3-year life cycle
• Natural regional stratification into Monitoring
  Areas (MAs) which partition ESUs

32
Sampling Design

• A rotating panel design was used to balance
  between extensive population coverage (for
  status) and repeat visits to same site (for
  trend)
• Panel definition
• One panel of sites visited every year
• Three panels of sites visited every 3 years
• Nine panels of sites visited every 9 years
• Twenty-seven panels of sites visited every 27
  years
• Four panels visited each year the annual panel,
  a 3-year panel, a 9-year panel, and a 27-year
  panel

33
Sampling Design

• Sites on streams were selected using a GRTS
  design
• Sites were apportioned to panels so that
• Each panel is in-itself spatially well balanced
• Each annual sample is spatially well balanced
• Equi-probable sample within MA

34
Research Overview

• Survey Design Methodology for Aquatic Resources
• Parametric Model Assisted Survey Methods for
  Environmental Surveys
• Nonparametric Model Assisted Survey Estimation
  for Aquatic Resources

35
Survey Design Methodology

• Presentations for this meeting
• A Cost Analysis of Ranked Set Sampling to
  Estimate a Population Mean, RA Buchanan, LL
  Conquest, and J-Y Courbois
• Some Extensions of the Generalized Random
  Tessellation Stratified (GRTS) Sampling
  Methodology, DL Stevens AR Olsen
• Convergence of Classes of Model-based and
  Design-based Spatial Samples, C Cooper and DL
  Stevens
• A Hierarchical Bayesian Spatial Model for Count
  Data Modeling Counts of Coho in the State of
  Oregon, RA Smith, B Munoz-Hernandez and DL
  Stevens Jr.

36
Survey Design Methodology

• Manuscripts
• 1 in press Environmetrics
• 1 reviewed, revised, returned to JASA
• 2 in preparation
• Presentations
• 3 at JSM
• 1 at ENAR
• 3 at workshops

37
Parametric Model Assisted Survey Methods

• Presentations this meeting
• Using Bayesian Spatial Models to Facilitate Water
  Quality Monitoring, Steven S. Carroll and Anthony
  R. Olsen
• Evaluation of Adjustment Procedures for Handling
  Item Nonresponse in Environmental Data, B
  Munoz-Hernandez, RA Smith and V Lesser
• Weighting Adjustments for Estimates of Coho
  Salmon Abundance when Survey Sites Are Missing at
  Random, LA Harrod, V Lesser, and B
  Munoz-Hernandez
• Exploring Design and Model-Based Approaches for
  Handling the Non-ignorable Missing Data Mechanism
  for Inference in Environmental Surveys, B
  Munoz-Hernandez, RA Smith, and V Lesser

38
Parametric Model Assisted Survey Methods

• Manuscripts
• 2 in preparation
• Presentations
• 2 at JSM

39
Non-Parametric Model Assisted Survey Methods

• Presentations this meeting
• Nonparametric Small Area Estimation for the
  Northeastern Lakes Survey, Jean D. Opsomer, F.
  Jay Breidt, and Siobhan Everson-Stewart
• Distribution Function Estimation in Small Areas
  for Aquatic Resources, Mark J. Delorey and F. Jay
  Breidt

40
Non-Parametric Model Assisted Survey Methods

• Manuscripts
• 3 submitted, 1 being revised (Biometrics, Survey
  Methodology, JASA)
• 4 in preparation
• Presentations
• 3 at JSM
• 3 at ENAR
• Lots at other meetings