A STATISTICIAN COLLABORATING WITH ECOLOGISTS: from EXAMPLES to LESSONS LEARNED

1
A STATISTICIAN COLLABORATING WITH
ECOLOGISTSfrom EXAMPLES to LESSONS LEARNED

• N. Scott Urquhart
• Senior Research Scientist
• Department of Statistics
• Colorado State University
• (June, 2004)

2
FUNDING ACKNOWLEDGEMENT
This talk 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 presenter and
STARMAP, the Program he represents. EPA does not
endorse any products or commercial services
mentioned in this presentation.
3
PATH FOR TODAY

• Autobiographical Sketch
• Some lessons learned
• Listed first, then
• Illustrated from experience
• Questions are welcome at any time

4
BIOGRAPHY of NSU

• Grew up in agriculture in Western Colorado
• Very familiar with the business end of dairy
  cows
• Manual labor of 1950s vintage agriculture
• Motivation for higher education HIGH
• Came to CSU in 1958
• Started in math, but considered civil
  engineering through first year

5
BIOGRAPHY - continued

• Statistical calculation of 1950s was done on
  rotary calculators
• IBM electronic accounting equipment
• Two computers in Colorado then!
• Went to work in the CSU Computing Center
  Spring, 1959
• Predecessor of both Computing Center and
• Stat Lab
• Took the graduate statistical methods course the
  summer after my freshman year
• Been at stat ever since

6
LESSON LEARNED

• Your professional world will change in major
  ways during your career!
• Implication Plan for it by learning how to
  think and learn. You will have to engage in
  learning throughout your career.

7
COMPUTING TOOLS OF 1960
8
STATISTICAL COMPUTATION HAS CHANGED! --- SO WHAT?
STATISTICAL COMPUTATION HAS CHANGED!

• The more powerful of the two computers in
  Colorado at that point was an IBM 704
• About 1985 I realized that the IBM XT on my desk
• 10M (not gig) hard disk
• 8088 8087 processors
• More power than the 704 0f 25 years earlier!
• Implication ?
• Be prepared for CHANGE
• Learn how to learn about all of ecology,
  statistics, and mathematics

9
LESSONS LEARNED

• Your professional world will change in major
  ways during your career!
• Implication Plan for it by learning how to
  think and learn. You will have to engage in
  learning throughout your career.
• Most easy problems have been solved.
• ? Relevant problems will require teamwork.
• Corollary Ecologists, Statisticians and
  Mathematicians need to understand each others
  perspectives
• ?Ecologists take collaborating statisticians
  and mathematicians to the field with you.
  Statisticians and mathematicians GO! Invite
  yourself, if necessary.
• Statistics can make a difference
• Mathematics can, too, but my examples deal with
  statistics

10
ILLUSTRATIONS

• Collaboration with plant ecophysiologists in
  Arctic Alaska
• North Slope
• Collaboration with the Grand Canyon Monitoring
  and Research Center (GCMRC), Flagstaff, AZ
• Monitoring of the terrestrial plant ecology near
  the Colorado River in the Grand Canyon
• EPAs Environmental Monitoring and Assessment
  Program (EMAP)
• My major collaboration for the past 14 years

11
ILLUSTRATIONS

• Collaboration with plant ecophysiologists in
  Arctic Alaska
• North Slope

12
HELPING DESIGNING VEGETATION STUDIES IN ARCTIC
ALASKA
13
MAP OF ALASKA
14
VIEW FROM OUTSIDE - REFRIGERATED ENTRANCE
15
IN THE TUNNEL
16
AN OLD ICE WEDGE
17
A SEGMENT OF THE TRANS-ALASKA PIPELINE
18
RESEARCH SITE AFTER JUNE SNOW
19
TYPICAL VALLEY BOTTOM (NOTE MOOSE IN CENTER)
20
(No Transcript)
21
R4D PROJECT ON ALASKASNORTH SLOPE

• FUNDED BY DEPARTMENT OF ENERGY
• INCLUDED A STUDY ECOLOGICAL PROCESSES
• OF PLANTS
• WITH THE VIEW OF PREDICTING LIKELY RESPONSE TO
  DISTURBANCES,
• LIKE THE DEVELOPMENT OF AN OIL FIELD
• MAJOR STUDY SITE IMNAVIAT CREEK

22
GENERAL GOALS OF R4D

• DEVELOP ECOPHYSIOLOGY MODELS FOR PLANTS ON THE
  NORTH SLOPE
• TO SUPPORT MORE GENERAL MODELS TO
  ALLOW EVALUATION OF LIKELY IMPACT OF
  ENERGY- RELATED DEVELOPMENT
• NSU ON-SITE EXPERIMENTAL AND SAMPLING DESIGN
• FOR MANY STUDIES

23
R4D PROJECT ON ALASKAS NORTH SLOPE -
continued

• A SMALL QUESTION IN THE LARGER QUESTION
• WHAT MAKES STONE STRIPES
• WATER SEEMS AVAILABLE IN VEGETATION BETWEEN STONE
  STRIPES
• ELEVATION DOESNT SEEM TO EXPLAIN
• STUDY COMPARE PHYSIOLOGICAL PROCESSES OF PLANTS
  IN STONE STRIPE TO THOSE OF PLANTS IN IMMEDIATELY
  ADJACENT PLANTS IN HEAVILY VEGETATED AREA
• ?? HOW TO SELECT STUDY PLOTS ??

24
(No Transcript)
25
HILLTOP AND NEARBY STONE STRIPE
26
STONE STRIPES AND EQUIPMENT FOR MEASURING
PHOTOSYNTHESIS RATE
27
POWER FOR MEASURING PHOTOSYNTHESIS RATE
28
EQUIPMENT FOR MEASURING PHOTOSYNTHESIS RATE
29
R4D PROJECT ON ALASKAS NORTH SLOPE -
continued - 2

• PREVIOUS YEAR STAKES HAD BEEN SET ON 100m CENTERS
  (Orange stakes in an earlier image.)
• NSU HAD ARRANGED FOR THREE 100m TAPE MEASURES TO
  BE AVAILABLE
• AND TOOK ALONG A LONG LIST OF PAIRS OF RANDOM
  DIGITS (00 - 99)
• DEFINE THE POPULATION - AWAY FROM THE SITE
• WHAT DO YOU WANT TO SAY IN THE INTRODUCTORY
  PARAGRAPH OF THE PAPER YOU PLAN TO WRITE ABOUT
  THIS?
• gt CRITERIA FOR WHETHER A SPECIFIC m2 PLOT WAS
  OR WAS NOT IN THE POPULATION

30
FIELD RESULT OF RANDOMLY SELECTED PLOTS

• I WOULD NEVER HAVE PICKED THAT PLOT!!!!

BUT I SHOULD HAVE!!!
31
(No Transcript)
32
(No Transcript)
33
(No Transcript)
34
(No Transcript)
35
(No Transcript)
36
LESSON LEARNED - I

• OPINION FOR MOST ECOLOGISTS, A TYPICAL SITE
  IS VARIABLE FOR THE FEATURE OF INTEREST, BUT AS
  HOMOGENEOUS AS POSSIBLE IN ALL OTHER WAYS
• gt MANY ECOLOGICAL FACTS ARE BASED ON VERY
  BIASED SUBPARTS OF THE SITUATION OF INTEREST

37
ILLUSTRATIONS

• Collaboration with plant ecophysiologists in
  Arctic Alaska
• North Slope
• Collaboration with the Grand Canyon Monitoring
  and Research Center (GCMRC), Flagstaff, AZ
• Monitoring of the terrestrial plant ecology near
  the Colorado River in the Grand Canyon

38
PROGRAM EVALUATION FOR THE GRAND CANYON
MONITORING AND RESEARCH CENTER (GCMRC)

• OBJECTIVE
• REVIEW THE TERRESTRIAL RESEARCH PROGRAM
• EXTERNAL PANEL
• BIOLOGISTS QUANTITATIVE SCIENTISTS
• ACADEMIC FEDERAL EMPLOYEES
• FROM GLEN CANYON DAM THROUGH GRAND CANYON TO
  LAKE MEADE
• EVALUATION PANEL OF 8
• 11-DAY RAFT TRIP THROUGH GRAND CANYON
• ACCOMPANIED BY 9 PEOPLE WHO HAD
  CONDUCTED RESEARCH ALONG THE RIVER

39
MAP OF THE GRAND CANYON AREA
40
THE PANEL SUPPORT CREW
41
THE BEGINNING
42
THE REAL START _at_ LEES FERRY
43
CAMPED LIKE ANY OTHER TRIP
44
WE SAW MAJESTIC SCENERY
45
WENT THROUGH RAPIDS!!!!!!!!
46
THE END WAS IN SIGHT
47
_at_ THE END
48
DESIGNING AND IMPLEMENTING A VEGETATION STUDY FOR
GCMRC

• OBJECTIVE - LOCATE AND LAY OUT VEGETATION
  TRANSECTS
• NSU SELECTED 100 POTENTIAL SITES
• RANDOMLY, SUBJECT TO SPATIAL RESTRICTIONS
• EACH POTENTIAL SITE WAS
• DETERMINED TO BE VERTICAL FACED, EFFECTIVELY
  UNVEGETATED, NO TRANSECT
• OR TRANSECT WAS LAID OUT AND DOCUMENTED
• RESULT
• 20 SITES TO BE REVISITED ANNUALLY
• 40 SITES TO BE VISITED ONLY ONCE IN THREE YEARS

49
NO VEGETATION HERE!!!(MILE 135.2)
50
ROCKY, BUT VEGETATED SITE _at_ 12.3
51
A VERY ROCKY SITE _at_ 95.0
52
A SANDY SITE _at_ 171.5
53
THE CREW AFTER TWO WEEKS ON THE RIVER
54
LESSON LEARNED II

• NSU has continued a low level of collaboration
  with GCMRC
• After a presentation to the Technical Advisory
  Committee, he was asked, Can this design be used
  to estimate the amount of vegetation in the
  region of interest?
• Answer was YES.
• Response from a long-time researcher in GCMRC
• While developing the environmental impact
  statement for a particular activity, we needed to
  estimate this quantity, but realized our data
  (gathered using traditional ecologists methods
  of selecting sites) didnt support it. Im glad
  we can now do that.

55
ILLUSTRATIONS

• Collaboration with plant ecophysiologists in
  Arctic Alaska
• North Slope
• Collaboration with the Grand Canyon Monitoring
  and Research Center (GCMRC), Flagstaff, AZ
• Monitoring of the terrestrial plant ecology near
  the Colorado River in the Grand Canyon
• EPAs Environmental Monitoring and Assessment
  Program (EMAP)
• My major collaboration for the past 14 years

56
DESIGNING STUDIES FOR EPAs ENVIRONMENTAL
MONITORING AND ASSESSMENT PROGRAM (EMAP)

• DISTINCTIVE EMAP PERSPECTIVE
• DEFINE THE POPULATION OF INTEREST
• CONDUCT A PROBABILITY SURVEY OF IT
• CAREFULLY DEFINE THE SAMPLING FRAME
• VARIABLE PROBABILITY SELECTION OF SITES,
  BUT WITH SPATIAL BALANCE
• CAREFULLY DEFINE RESPONSES TO BE EVALUATED
• TRAIN FIELD CREWS WELL
• MANAGE DATA WITH CARE AND AN AUDIT TRAIL
• LEARN FROM PAST MISTAKES, THROUGHOUT
• NSU HAS BEEN VERY INVOLVED WITH PROCESS FOR
  EVALUATING BUGS IN STREAMS

57
Policy Questions Are Often Very Simple
What do you mean you dont know how many acid
lakes there are? William Ruckelshaus - EPA
Administrator - early 1980s
58
EMAP Lesson 1 - Not all Experts are expert
in all things
Thats a stupid question and one that you cant
possibly answer. Chair of NRC Indicator Panel
59
NAPAPNational Surface Water Surveys
60
NAPAP-NSWS Survey ResultsAnswering Ruckelshaus
Question
61
NAPAP-NSWS Survey ResultsAnswering Ruckelshaus
Question
62
EcoSynthesisThe Very Early Stages of EMAP
There are plenty of data out there to know what
is going on in the environment and what our major
problems are. Do the hard work and synthesize
it! Conventional Wisdom
63
EMAP Lesson 2 Conventional wisdom is just
that, conventional, and seldom leads to important
breakthroughs.

• The data were not there to provide unbiased
  estimates.
• lack of appropriate indicators (e.g. forests) and
  
• lack of an appropriate design (e.g. inland
  aquatics).
• We discover we will have to collect the data
  ourselves!

64
WHAT EFFECT HAS EMAP HAD?(A Statistically
Defensible Approach)

• Drastically altered the sorts of data and reports
  EPAs Office of Water requires of States and
  Tribes.
• Prompted substantial changes in certain parts of
  the Forest Inventory and Analysis (FIA) of the US
  Forest Service
• Prompted substantial changes in outlook and
  cooperation of
• EPAs Office of Water
• National Oceanic and Atmospheric Administration
• United States Geological Service
• Relative to what follows

65
The Clean Water Action Plan (1999)

• OW/NOAA/USGS requirement to assess the condition
  of aquatic resources
• Stalemated 1999 Conference Call
• Invention of the National Coastal Condition Report

66
National Coastal Condition Report I

• Based on 1990-1996 data
• Virginian, Carolinian, Louisianian, selected west
  coast sites

67
National Coastal Assessment

• 1999 EMAP Western Pilot Coastal
• 2000 All coastal states in conterminous US and
  Puerto Rico
• 2001 Includes Alaska and Hawaii (one time)
• 2002 Includes Western Inter-tidal
• 2003 Includes Western Offshore with NOAA
• 2004 Includes SE Offshore with NOAA

68
THE BOTTOM LINE

• For this Statistician, Collaborating with
  Ecologists and Agricultural Scientists, the
  Experience Has Been
• Interesting
• Fun
• Rewarding
• We all need clean water, clean air, and
  wholesome food
• End of this Planned Presentation
• Questions Welcome

69
CADDIS FLIES
70
MESA CREEK (ORC 20)
71
SURBER SAMPLER IN USE
72
COLLECTING IN A STREAM
73
FIELD CLEANING OF A COLLECTION
74
BAGGING PRESERVING A COLLECTION
75
BAGGED COLLECTION - LAB
76
EVALUATING THE EVALUATION ERROR

• HERE, THERE IS NO WAY TO REPEAT THE PROCESS OF
  PULLING CLUMPS APART gt REPEATED MEASURES
  IMPOSSIBLE
• ALTERNATIVE SPLIT COLLECTIONS
• 5050
• SEPARATELY EVALUATE EACH HALF
• EXAMINE SIMILARITY OF SPLITS TO STUDY
  EVALUATION ERROR

77
COLLECTIONEMPTIED INTO A BEAKER PRIOR TO
SPLITTING
78
PLACING A COLLECTION IN THE SPLITTER
79
SPLITTING A COLLECTION
80
COMPLETED SPLIT
81
SPLIT COLLECTION READY FOR ENUMERATION
82
EPAs ARCTIC CONTAMINANTS RESEARCH PROGRAM

• REALITY ARCTIC HAZE HAS INCREASED
  GREATLY OVER THE PAST 30 YEARS.
• APPARENT ORIGIN
• NORTHERN EUROPE
• INCLUDING THE FORMER SOVIET UNION.
• IT CONTAINS LOTS OF BAD STUFF.
• QUESTION
• HAS MUCH OF THIS BAD STUFF BEEN DEPOSITED IN
  ARCTIC ALASKA?
• APPROACH
• USE LICHENS AS A NATURAL ACCUMULATOR
• COLLECT SPECIFIC LICHENS ACROSS A
  POSSIBLE GRADIENT OF SITES
• NSU SELECTED GENERAL SITE LOCATION
  DEVELOPED LOCAL SITE SELECTION CRITERIA