Transcript urquhart.2004.pps - Colorado State University`s Department of

STARMAP
YEAR 3
N. Scott Urquhart
STARMAP Director
Department of Statistics
Colorado State University
Fort Collins, CO 80523-1877
STARMAP/DAMARS 9/10/04
#1
STARMAP FUNDING
Space-Time Aquatic Resources Modeling and Analysis Program
The work reported here today was developed under the STAR
Research Assistance Agreement CR-829095 awarded by the
U.S. Environmental Protection Agency (EPA) to Colorado
State University. This presentation has not been formally
reviewed by EPA. The views expressed here are solely those
of presenters and STARMAP, the Program they represent.
EPA does not endorse any products or commercial services
mentioned in these presentation.
This research is funded by
U.S.EPA – Science To Achieve
Results (STAR) Program
Cooperative
# CR - 829095
Agreement
STARMAP/DAMARS 9/10/04
#2
TODAY’S
COMMENTS ABOUT STARMAP



Background Comments
Research
Outputs



Extension/Outreach/Cooperation




Publications/presentations
Training – future generations of environmental statisticians
Meetings
Learning Materials
Recruiting
Plans for the year ahead


Research
Meetings
STARMAP/DAMARS 9/10/04
#3
PROGRESS SINCE LAST YEAR:
STARMAP RESEARCH - A FOCUS
 EPA released
the report,
Response of Surface Water Chemistry to the Clean
Air Act Amendments of 1990 (EPA/620/R-02/004)



Was submitted to Congress in January, 2003
Statistical techniques available for it were
severely limited in several dimensions
Good illustration of tools we need to develop


Used as presentation illustration
STARMAP investigators are using


The context for model assumptions
Data on which this report was based
STARMAP/DAMARS 9/10/04
#4
PROGRESS SINCE LAST YEAR:
STARMAP RESEARCH – Project 1
 Combining/analyzing

Bayesian methods





Extension of contingency table methods
» Devin Johnson
» Finished PhD, now at University of Alaska; continuing
Selection of spatial models – Andrew Merton
Megan Dailey is starting work on this project
Book on Computational Statistics
Spatial statistics




environmental data
Designing to estimate semivariograms – Kerry Ritter & Molly Leecaster
How much spatial correlation in aquatic responses? - nsu
Alix Gitelman (OSU) & graph methods
 Originally under Project 2; work is closer to Project 1
 Steve Jensen – short talk
Jennifer Hoeting (PI) will comment further
STARMAP/DAMARS 9/10/04
#5
PROGRESS SINCE LAST YEAR:
STARMAP RESEARCH – Project 2
 Local


Jointly funded by STARMAP and DAMARS
Methods to improve survey summaries using
auxiliary information






Estimation/Survey Methods
From design-based to model-based
Major collaboration with Jean Opsomer, Iowa State U
 Ji-Yeon Kim and Gerda Claeskens
Giovanna Ranalli = post doc, not here right now
Two from cooperating fellows from Taiwan
 Nan-Jung Hsu and Hsin-Cheng Huang
Progress on procedures for smoothing data from river networks
Jay Breidt will expand on this during his talk

STARMAP/DAMARS 9/10/04
Mark Delorey and Bill Coar
#6
PROGRESS SINCE LAST YEAR:
STARMAP RESEARCH – Project 3

Indicator Development





GIS work & support is funded here
SUBSTANTIAL PROGRESS in GRTS sampling
 Will be reported at Monitoring Science and Technology
conference week after next (Denver)
Good cooperation with Region 8 and Western Ecology Division
Erin Poston, who will give a talk tomorrow, is funded here.
Dave Theobald (PI) will comment further
 Promising GIS tools are under development and testing
STARMAP/DAMARS 9/10/04
#7
PROGRESS SINCE LAST YEAR:
STARMAP RESEARCH – Project 4

Extension/Outreach


Urquhart (PI)
I’ll return to this as next to the last major topic
 Only opportunity to talk about outreach here
STARMAP/DAMARS 9/10/04
#8
TRAINING - RESULTS

Governing RAF states


“… cadre of graduates would be developed who would
have the expertise and experience in survey design and
analysis needed to fill a gap in the expertise required to
successfully monitor the condition of the Nation’s
aquatic resources.”
Students & early career professionals

Students




One PhD + 4 Masters completed
5 PhD, 4 stat + 1 landscape ecology
2 MS;
Early-career professionals - 8
STARMAP/DAMARS 9/10/04
#9
PROGRESS SINCE LAST YEAR:
TRAINING: CONTINUED

Field experience




Visits to other EPA facilities



Several + ongoing
All toured past EMAP sites near OSU last year
All had a tour of WED/EMAP last year
Theobald & Poston interacted with personnel at
Region 8 recently
All CSU students attended a meeting a Region 8 in
May of this year
Gitelman will comment about her
interactions with Ken Reckhow later today.
STARMAP/DAMARS 9/10/04
# 10
OUTPUTS

Relevant statistical research in the form of



The Generalized Random Tessellation Stratified
sampling process


Publications
Worked examples in the “Case studies” part of the
learning materials
Implemented correctly in ARCView
Prediction of probability of perennialness of
stream traces

Add to sampling procedure by end of Program
STARMAP/DAMARS 9/10/04
# 11
OUTPUTS in YEAR 3
 PUBLICATIONS

9 professional publications – in print or accepted




Including a book
12 in various stages of submission and review
26 manuscripts in various stages of development
Presentations



37 completed
5 more by end of year 3
To a variety of audiences, domestic and international
STARMAP/DAMARS 9/10/04
# 12
OUTPUTS in YEAR 3
continued
 STUDENT

PhDs


PROGRESS
Four nearing completion
 Three Stat
 One geoscience
Masters

Two completed this year
 One @ CSU, one @ OSU
 Two more in progress
STARMAP/DAMARS 9/10/04
# 13
MEETINGS

Graybill Conference – success

Proceedings will be published as an issue of
Environmental and Ecological Statistics

Friendly outlet for young environmental statisticians
 Topic
contributed paper session at 2004 Joint
Statistics Meetings – with the Chicago Center


Success
Monitoring Science and Technology Symposium


Two and a half days of sessions on statistical aspects of
monitoring
Half day tutorial on GRTS sampling
STARMAP/DAMARS 9/10/04
# 14
MEETINGS - WHY?

TRAINING

Provide opportunities for students to gain relevant
professional experience
 DEVELOPMENT
OF YOUNG
ENVIRONMENTAL STATISTICIANS


Early career environmental statisticians
Gain stature in our discipline
 OUTREACH/EXTENSION


to
Statistical community – Graybill Conference
Users of statistics – Monitoring Science & Technology
STARMAP/DAMARS 9/10/04
# 15
RECRUITING

Post-doctoral fellows



Modest success – will continue efforts
Giovanna Ranalli was here for four months, and will
return in October
Director will continue to be active in recruiting


Students for the CSU statistics graduate program
In high school advanced placement statistics




We need undergraduate majors in relevant areas to recruit for
future graduate students
Urquhart visited 8 AP classes in four high schools last year
 More visits are being arranged for next year
Developed some dynamic graphics for power and r2
Example
STARMAP/DAMARS 9/10/04
# 16
LEARNING MATERIALS


Jointly funded by STARMAP & DAMARS
Browser based (web access NOT needed)


To be delivered via CD ROM
Substantial opportunities for individualization


Perspective
 Senior administrator to
 Implementer to
 Researcher
Landscape setting
 Evaluation & feedback - OSU test –
 Evaluators – state, tribal & EPA regional
 Article in Tribal News
STARMAP/DAMARS 9/10/04
was positive
# 17
CONTENT OF LEARNING MATERIALS
PROGRESS

Why Monitor?

Draft completed and tested
 Where


Draft completed and tested
What to monitor = Indicators


to Monitor? (= GRTS-type sampling)
Nothing started on this yet
How to Monitor? (= Field Operations)




EPA Field Operations Manuals
Training videotaped in May for inclusion in this part
We have PowerPoint presentations from training
Student is working on this (MS)
STARMAP/DAMARS 9/10/04
# 18
CONTENT OF LEARNING MATERIALS
PROGRESS
(continued)


…
How to summarize.


Case studies


Student has started on this
Several current efforts will be turned into these
Distribution of completed materials

Preliminary discussions with the Council of State
Governments have been very encouraging.
STARMAP/DAMARS 9/10/04
# 19
DYNAMIC ILLUSTRATION OF POINT
LOCATION


Strahler order 1, 2 & 3
represented by trace
thickness
Relative probability by
order




First: 0.25
Second: 0.50
Third 1.00
Points appear in sampling
order



Slowly at first
Then faster
Loops back to first point
STARMAP/DAMARS 9/10/04
# 20
PLANS FOR THE YEAR AHEAD

Project 1 - continue present work



Project 2 – continue developments



Emphasize small area (local) estimation
Potentially valuable for 305(b) => 303(d) reporting
Project 3 - complete GRTS sampling in GIS



Expand designs for spatial modeling
Explore data resulting from San Diego study
Apply tools to develop indicators
Regional species richness
Outreach – continue development of cooperation


Case studies
Continue implementation & testing

At any suitable venue
STARMAP/DAMARS 9/10/04
# 21
ACCUMULATING AQUATIC
SPATIAL DATA

How much spatial correlation really is present in
aquatic responses, after accounting for habitat
features?

Urquhart has been seeking data sets to look at this:





Streams in Virginia
Ohio River – about 400 points
Detroit River – 100 or so over short distance
Estuaries – one in northeast
Near coastal – San Diego Project
 Discussed under Kerry Ritter’s presentation
STARMAP/DAMARS 9/10/04
# 22
FUTURE MEETINGS

Computational Environmetrics, October 21-23, 2004, in
Chicago, Illinois


Statistics and Environment Section of the American Statistical
Association (ASA) and
The Center for Integrating the Statistical and the Environmental
Science (CISES) at the Department of Statistics, University of
Chicago.



Three short courses – Jennifer Hoeting will be involved in two
Jay Breidt & Devin Johnson are invited speakers
DAMARS/STARMAP in 2005



At Oregon State University
Tentative dates: September 9 & 10, 2005
Advice – Content & Audience
STARMAP/DAMARS 9/10/04
# 23
FUTURE MEETING POSSIBILITIES
Special session for next year’s North American
Benthological Society Meeting
 Symposium on the interface between GIS and
statistics


Especially relevant as ArcGIS now has spatial statistics
components



As distinct from its geostatistics modules
Code is open and alterable
Dave Theobald can say more about this possibility
STARMAP/DAMARS 9/10/04
# 24