ROLES FOR STATISTICS IN 21ST CENTURY MONITORING AND ASSESSMENT SYSTEMS

1
ROLES FOR STATISTICS IN 21ST CENTURY MONITORING
AND ASSESSMENT SYSTEMS 
N. Scott Urquhart Director of STARMAP Department
of Statistics Colorado State University Fort
Collins, CO 80523-1877 - USA
2
OVERVIEW OF THIS THEME

• PERSPECTIVES
• PLENARY SESSIONS
• Today
• CASE STUDIES
• Tomorrow Wednesday
• LINKING THE TWO PERSPECTIVES
• Thursday Morning
• TUTORIAL How To Design and Implement Natural
  Resource Surveys
• Thursday Afternoon

3
SURVEY PERSPECTIVES

• REMOTELY-SENSED RESPONSES HAVE A MAJOR ROLE
• Ground-Based Responses Have an Auxiliary Role
• Often as ground truthing
• IN CONTRAST TO
• GROUND-EVALUATED RESPONSES HAVE A MAJOR ROLE
• Remotely-Sensed Responses Have an Auxiliary Role
• Often serving as covariates

4
REMOTELY-SENSED RESPONSES

• WHAT ARE THEY?
• Usually They are Sensed from Images Obtained from
  an Aerial Platform
• Aerial photography
• Imaging from a space vehicle
• Spectral reflectance fairly well established
• Radars emerging
• Often complete coverage
• Devices attached to a place, animal or robot
• Stream flow
• Sense things like location and temperature
• Deer to bees

5
REMOTELY-SENSED RESPONSES(Continued)

• WHAT ARE THEY?
• Usually They are Sensed from Images Obtained from
  an Aerial Platform
• Aerial photography
• Imaging from a space vehicle
• Classify parts of the image
• Automatic computer based
• Manually
• Evaluate the size of various classes, like
• Land use classes
• Vegetation type

6
GROUND-EVALUATED RESPONSES

• CONTRASTING PERSPECTIVE
• Data is Obtained by Personnel Visiting the
  Field Site of Interest
• At the field site, personnel may
• Collect material for subsequent lab evaluation
• Directly evaluate responses
• Or both common in aquatic studies
• Frequent realities
• Many responses will be evaluated
• Design can not be optimized for all responses

7
GROUND-EVALUATED RESPONSES(Continued)

• Site Selection Process
• Area of interest may be partitioned into disjoint
  areas
• A sample of areas will be visited
• Points may be selected in some manner
• Field crews go to site
• Resource of interest may, or may not, be there

8
OVERVIEW OF THIS THEME

• PERSPECTIVES
• PLENARY SESSIONS - OVERVIEW
• Today
• CASE STUDIES EXAMPLES
• Tomorrow Wednesday
• LINKING THE TWO PERSPECTIVES
• Small area or local estimation
• Thursday Morning
• TUTORIAL How To Design and Implement Natural
  Resource Surveys
• Thursday Afternoon

9
PLENARY SESSIONS OVERVIEW

• Remotely-Sensed Responses (This session 130
  300)
• On Remotely-Sensed Responses
• Raymond (Ray) Czaplewski
• Ground-Evaluated Responses (Next session 345
  515)
• Statistical Perspective on the Design and
  Analysis of Natural Resource Monitoring Programs
• Anthony (Tony) R. Olsen
• Overview of FIA
• Ronald (Ron) McRoberts
• Schedule change
• Hans Schreuder to Wednesday _at_ 130
• Steven Fancy to Thursday _at_ 1145

10
CASE STUDIES EXAMPLES

• Programs Utilizing Remotely-Sensed Responses
• Session 031101 Chair Trent McDonald
• National Resources Inventory
• Wayne Fuller, others
• National Wetlands Inventory
• Tom Dahl
• Date/Time Tomorrow Wednesday, 9/22/04
• 830 930
• 930 1000 time for discussion

11
CASE STUDIES EXAMPLES(Continued)

• Programs Utilizing Ground-Evaluated Responses
• Session 031101 - Continued
• The United States National Agricultural Survey
• Carol House
• Integrated State-Federal Partnership for Aquatic
  Resource Monitoring in the United States for
  Groundwater Using Existing Wells
• Anthony (Tony) R. Olsen
• Forest Inventory and Analysis Program of the
  United States Department of Agriculture
• Michael (Mike) Williams others
• Date/Time Tomorrow Wednesday, 9/22/04
• 1045 1215

12
CASE STUDIES EXAMPLES(Continued)

• Realities of Conducting Natural Resource Surveys
  Chair Mike Williams
• Session 041102 Continued
• The Past, Present, and Future of Sampling Natural
  Resources An Economic and Statistical
  Perspective
• Hans Schreuder
• Interagency Cooperation in Natural Resource
  Surveys
• J. Jeffery Goebel
• Wildlife Monitoring Success Requires More than a
  Good Sampling Design
• Kenneth P. Burnham
• Date/Time Tomorrow Wednesday, 9/22/04
• 130 300

13
CASE STUDIES EXAMPLES(Continued)

• Not Represented
• Alberta Biodiversity Monitoring Program (ABMP)
• http//www.abmp.arc.ab.ca/
• Cooperative venture government, academia
  industry
• Minimally Represented Surveys of animal
  populations
• Very different study requirements from most of
  the cases discussed here
• Often, finding the animals constitutes a major
  undertaking
• Frequently, some sort of modeling plays a major
  role
• Nevertheless, many of the same ideas have to be
  addressed

14
LINKING THE TWO PERSPECTIVES

• Small Area Estimation and Model-Based Inference
• Session 051105 -- Gretchen Moisen organized this
• Small Area Estimation for Natural Resource
  Surveys
• F. Jay Breidt
• Evaluating Standards Using Data Collected From
  Regional Probabilistic Monitoring Programs
• Eric P. Smith others
• Non-linear Small Area Estimation in the National
  Resources Inventory Survey
• Tapabrata (Taps) Maiti
• Date/Time Thursday, 9/23/04
• 830 1000

15
LINKING THE TWO PERSPECTIVES(Continued)

• Small Area Estimation and Model-Based Inference
• Session 051105
• Use of Model-based Stratifications for Sampling
  Rare Ecological Events Lichens as a Case
  Example
• Thomas C. Edwards others
• Developing Risk-based Guidelines for Water
  Quality Monitoring and Evaluation The
  Australian Experience
• David Fox
• Long-term Monitoring of Large, Remote Areas with
  Minimal Funding Hope and Encouragement for
  Natural Area Managers
• Steven Fancy
• Date/Time Thursday, 9/23/04
• 1045 1215

16
TUTORIALHOW TO DESIGN AND IMPLEMENT NATURAL
RESOURCE SURVEYS

• A Tutorial on Designing Natural Resource Surveys
  Concepts to Implementation
• Session 061104
• Instructor Urquhart
• Structured around the Anatomy Of Sampling Studies
  Of Ecological Responses Through Time
• Urquhart Olsen
• Date/Time Thursday, 9/23/04
• 130 300

17
TUTORIALHOW TO DESIGN AND IMPLEMENT NATURAL
RESOURCE SURVEYS(Continued)

• A Tutorial on Designing Natural Resource Surveys
  Concepts to Implementation
• Session 061104
• The Generalized Random Tessellation Stratified
  Sampling Design for Selecting Spatially-Balanced
  Samples
• Don L. Stevens
• GRTS for the Average Joe Implementing GRTS in
  Windows and S-Plus
• Trent L. McDonald
• Robust Spatial Sampling of Natural Resources
  Using a GIS Implementation of the GRTS Algorithm
• David M. Theobald
• Date/Time Thursday, 9/23/04
• 345 515

18
SUMMARY SESSION

• Unified Knowledge-Based Strategies and Solutions
• Ray Czaplewski
• USDA, Forest Service
• Richard W. Guldin
• Science Policy, , USDA Forest Service-Research
  Development
• Keith Pezzoli
• University of California _at_ San Diego
• Greg Reams
• Forest Health Monitoring, USDA Forest Service
• Carl Reed
• Specification Program, Open GIS Consortium, Inc
• Date/Time Friday, 9/24/04
• 830 1200
• If any of these people are here, please see me.

19
FIRST PLENARY SPEAKER

• Raymond (Ray) Czaplewski
• Project Leader, Forest Invent. Monitoring
  Envi..
• USDA-Forest Service-Rocky Mountain Research
  Station, Fort Collins, CO
• Received his PhD in Range Science from Colo State
  Univ
• Earlier in his career he held positions as a
  statistician and landscape ecologist.
• Professional interests include
• Integration of remotely sensed data from
  earth-observing satellites into monitoring
  processes
• Ecological process models and
• Field observation. 
• On Remotely-Sensed Responses

20
SECOND PLENARY SPEAKER

• Anthony (Tony) R. Olsen
• Statistics Lead, Environmental Monitoring
  Assessment Program (EMAP)
• EPAs-Western Ecology Division, Corvallis, OR
• Received his PhD in Statistics from Oregon State
  Univ
• His professional interests include
• Statistical aspects of monitoring and assessment,
  monitoring design
• Survey sampling
• Exploratory data analysis and
• Graphical data analysis and graphical
  communication.
• Statistical Perspective on the Design and
  Analysis of Natural Resource Monitoring Programs

21
THIRD PLENARY SPEAKER

• Ronald (Ron) McRoberts
• Group Leader for Research for the Forest
  Inventory Analysis Program
• North Central Research Station, USDA-Forest
  Service
• He received a PhD in biostatistics from the Univ.
  of Minnesota.
• His research interests include
• Nonlinear modeling,
• Land cover land change, and
• Map-based estimation of forest attributes.
• Overview of FIA