Workshop on Environmental Study Design & Analysis
Roger H. Green
Vancouver April 9-12 2001
................Website Menu
Recent updates
March 26: I've added lots more information about participants; I've added
a bit more about software; I've improved font size, colour & readability in
some places; and I've put up a
tentative workshop schedule.
Roger Green
Outline - Environmental study design and analysis: principles & examples
A. Introduction
1. Objectives
2. Background & interests that I'm assuming
3. My sense of why you need this
4. Where I'm coming from - my experience and interests
5. Why this "principles & examples" organization
6. Relationship to computer programs/packages
B. Design & statistical analysis in environmental studies: Principles
1. Environmental study design: Principles
a. General principles
b. Spatial pattern, statistical assumptions, and transformation of variables
c. Choice of sampling method and sample unit size
d. Hypothesis testing - why & how
e. Estimating necessary number of samples
f. Allocation of sampling effort in space and time
g. Criteria for choice of response variable(s)
h. ANOVA designs - random, paired, nested, factorial
i. Reference sites, reference conditions
j. BACI designs
k. Specialized methods: sequential sampling; mark-recapture and removal estimates; artificial
substrates; field transplant experiments; EMAP; Bayesian methods and other current fads
2. Statistical analysis by univariate models: Principles
a. Review of regression analysis models
b. Linear and nonlinear models in biology and ecology
c. Applications of Analysis of Covariance (ANCOVA) models
d. How to do ANOVA and ANCOVA by regression, using dummy variables
e. Effective display of regression and ANCOVA results
3. Statistical analysis by multivariate models: Principles
a. Overview
b. Search for structure (no a priori sample or variable structure)
c. Samples fall into a priori groups
d. Variables fall into a priori groups (responses, predictors, covariates, - -)
e. Effective display of multivariate analysis results
C. Design & statistical analysis in environmental studies: Examples
Tentative workshop schedule
(This is based on an 8:30am to 5:30pm day, with an hour lunch break a half hour break at mid-
morning and mid-afternoon, and no evening time. None of this is carved in stone.)
Mon afternoon - 1st half: Introductory remarks, getting acquainted, handouts, and the like
Mon afternoon - 2nd half: Principles of Study Design (PoSD): general principles
Tue morning - 1st half: PoSD: spatial pattern, statistical assumptions, transformations
Tue morning - 2nd half: PoSD: hypothesis testing; choosing sampling method and sample unit size; estimating
necessary number of samples
Tue afernoon - 1st half: PoSD: allocation of samping effort in space and time; choosing response variables;
Tue afernoon - 2nd half: PoSD: ANOVA designs; Reference sites and times; BACI designs; miscellaneous methods
Wed morning - 1st half: Univariate Models: review of regression; common nonlinear models in ecology; ANCOVA
Wed morning - 2nd half: Multivariate Models (MM): Introduction and overview
Wed afternoon - 1st half: MM: Search for structure - ordination, clustering and similar methods
Wed afternoon - 2nd half: MM: A priori groups of samples - MANOVA, Canonical Discriminant Analysis, etc.;
A priori groups of variables - Canonical Correlation Analysis, Correspondence Analysis, Mantel's Test, etc.
Thurs morning - 1st half: Case studies, discussion
Thurs morning - 2nd half: Case studies, discussion
N.B. - I need some dialogue here. Should we cover less and spend more time on some things? Should
we use evenings? How many participants can't do evenings? What mix of lecture format, hands on
computers, and discussion of case studies is best?
Useful references - books & papers
Books on study design and statistical methods useful for environmental studies Batschelet, E. 1976. Introduction to mathematics for life sciences. Springer-Verlag, New York. (The kind of book that should be the text for a biologist's calculus course, but never is.) Cochran, W.G. 1963. Sampling techniques, 2nd Ed. Wiley, N.Y. (Still the classic reference on the subject.) Cochran, W.G. 1983. Planning and analysis of observational studies. Wiley, New York. (A nice short treatment of principles of design and statistical analysis - including power analysis - for the kinds of studies ecologists most commonly do, i.e. studies that are not designed experiments.) Crawley, M.J. 1993. GLIM for ecologists. Blackwell, Oxford. (See comments on McCullagh and Nelder 1983.) Crowder, M.J., and D.J. Hand. 1990. Analysis of repeated measures. Chapman and Hall, London. (This is the bible on the subject. There are worked examples, discussion of procedures as implemented in various statistical packages, and discussion of assumptions and consequences of their violation.) Draper, N.R., and H. Smith. 1981. Applied regression analysis, 2nd Ed. Wiley, New York. (The bible on regression analysis and modelling, examining residuals, nonlinear models, etc. Worked examples.) Edgington, E.S. 1995. Randomization tests, 3rd Ed. Marcel Dekker, New York. (See comments on Manly 1991.) Elliott, J.M. 1977. Some methods for the statistical analysis of samples of benthic invertebrates. FBA Sci. Publ. No.25, Windermere, Cumbria, U.K. (A "best buy". Good on spatial distributions, sampling designs, transformations.) Green, R.H. 1979. Sampling design and statistical methods for environmental biologists. Wiley, N.Y. (A "handbook" with examples, univariate and multivariate approaches, my prejudices, and a large topic-coded bibliography - now rather out of date. A new book is in prep, the draft title & outline of which is the same as for this workshop.) Harris, R.J. 1985. A primer of multivariate statistics, 2nd Ed. Academic Press, N.Y. (A good multivariate stats text and reference, but “primer” is misleading. Don't start with it - start with Pielou 1984, Pimentel 1978 or Manly 1994.) Hunt, R. 1978. Plant growth analysis. Edward Arnold, London. (Thin soft cover treatment of quantitative models of growth and form - generally applicable - not just to plants.) Jongman, R.H.G., C.J.F. ter Braak and O.F.R. van Tongeren. 1987. Data analysis in community and landscape ecology. Pudoc Wageningen, The Hague. (The latest crazes in statistical analysis of biological community data. The offspring of a mating between a Dutch school and the Cornell school. Trendy methods such as Canonical Correspondence Analysis plus coverage of traditional methods such as Principal Components Analysis and analysis of spatial pattern. Cajo ter Braak wrote a statistical package called CANOCO that does most of it. If you are interested in analysis of community data, do not begin with this one. Read Pielou 1984 or Manly 1994 first.) Keough, M.J., and B.D. Mapstone. 1995. Protocols for designing marine ecological monitoring programs associated with BEK mills. Report No. 11. CSIRO, Canberra. (It’s a book-size report. Excellent coverage of general principles and practice of environmental study design. A non-commercially published word-of-mouth gem, like Elliott 1977. If you write to CSIRO in Canberra they’ll probably mail you a copy for free.) Kirk, R.E. 1982. Experimental design: procedures for the behavioral sciences. Brooks/Cole, Monterey, California. (Another good reference on experimental and observational study design. His fans, who tend to be fanatics, always ask "But what does Captain Kirk say?" (apologies to Trekkies). One of his cleverer efforts is to show how to do "pseudo-F tests" by constructing composite error terms. SAS will sometimes do this for complex designs without being asked. See Winer 1971 or Underwood 1997 for a more classical treatment of the subject.) Legendre, L., and P. Legendre. 1983. Numerical ecology. Elsevier, Amsterdam. (Very thorough text/reference. Covers matrix algebra diversity indices, multivariate analyses, time series, matrix population models, etc.) Manly, B.F.J. 1991. Randomization and Monte Carlo methods in biology. Chapman and Hall, London. (Randomization tests have largely replaced “nonparametric” tests. This is a good reference to them, and Manly’s RT package implements them. Also see Edgington 1995.) Manly, B.F.J. 1994. Multivariate statistical methods: a primer. Chapman and Hall, London. (Another good one is Pielou 1984.) Mead, R. 1988. Statistical principles for practical applications. Cambridge Univ. Press. (Some say it’s the best general reference on principles in study design for applied studies.) McCullagh, P., and J.A. Nelder. 1983. Generalized linear models. Chapman and Hall, London. (Most texts and statistical packages limit "linear models" to ones with normally distributed error, such as ANOVA and ordinary least squares regression. Be sure you're on top of those sorts of linear models first, then look at how this book brings ”other-than-normal error distribution” models such as probit analysis, log- linear analysis of contingency tables, logistic models, and proportional hazards models for survival data, underneath the "linear models" umbrella. The statistical package GLIM implements these methods. GLIM repels many people with its different philosophy and strange syntax, but it is worth knowing how to use. See Crawley 1993.) Pielou, E.C. 1984. The interpretation of ecological data. Wiley, N.Y. (The subtitle is "A primer on classification and ordination". It's a good introduction to descriptive multivariate statistics applied to ecology. There are clear worked examples. Another good one is Manly 1994.) Pimentel, R.A. 1978. Morphometrics: The multivariate analysis of biological data. Kendall/Hunt, Dubuque, Iowa. (A good introduction to multivariate statistics. Measurements on painted turtles feature a lot in his examples, hence the title. Lots of errata but good value nonetheless.) Popper, K. 1980. The logic of scientific discovery, 10th Ed. Hutchinson, London. (It is the statement of what the philosophy of science is, what the scientific method is. For expression of this philosophy in ecological study design and statistical analysis see Underwood 1997. Popper is worth reading and understanding because this sense of “what science is” is currently under attack by Bayesian “risk analysis” types, e.g. Suter 1993, Suter 1996 Human and Ecological Risk Assessment 2: 331-347, and Stewart-Oaten in Schmitt and Osenberg 1996 in Detecting Ecological Impacts: Concepts and Applications in Coastal Habitats, Schmitt and Osenberg, eds., p. 17-27, Academic, New York. See Dennis 1996 Ecol. Appl. 6: 1095-1103, for a good response i.e. a critique of Bayesians in ecology.) Ripley, B.D. 1981. Spatial statistics. Wiley, New York. (As the preface says, "This is a guide to the analysis of spatial data". Relevant topics are spatial autocorrelation and how to deal with it, and testing hypotheses about spatial patterns of organisms. See also chapter 7 of Jongman et al. 1987.) Schmitt, R.J., and C.W. Osenberg. 1996. Detecting ecological impacts caused by human activities. Academic Press, New York. (It’s a mixed bag, with chapters by various people. But the people are mostly very good - e.g. Stewart-Oaten, Underwood, Keough, Jones & Kaly, Mapstone, Kingsford - and the topics are mostly very current and important. I totally disagree with some of it, but that’s OK - some of the chapter authors disagree with other chapter authors within the book.) Schneider, D.C. 1994. Quantitative ecology: spatial and temporal scaling. Academic Press, San Diego. (Nice coverage of how temporal and spatial scales of observations, and of what is being observed, influences results and interpretations of studies.) Seber, G.A.F. 1984. Multivariate observations. Wiley, New York. (A "bible" for multivariate statistics, including the math theory and algorithms that underlie it.) Snedecor, G.W., and W.G. Cochran. 1989. Statistical methods, 8th ed. Iowa State Univ. Press, Ames, Iowa. (Perhaps the best biologically oriented statistics text. See Zar 1996 for a more introductory level text.) Suter, G. W. 1993. Ecological risk assessment. Lewis, Boca Raton, Florida. (See also Suter 1996 Human and Ecological Risk Assessment 2: 331-347. He advocates a Bayesian-style “risk analysis”approach to environmental management decision-making, as opposed to the traditional Popperian scientific method approach. Also see Stewart-Oaten in Schmitt and Osenberg 1996 in Detecting Ecological Impacts: Concepts and Applications in Coastal Habitats, Schmitt and Osenberg, eds., p. 17-27, Academic, New York. I’m not a fan, but it’s currently fashionable scientifically and politically. For an antidote see Popper 1980, Underwood 1997, and Dennis 1996 Ecol. Appl. 6: 1095-1103. The last is a short and easy read.) Underwood, A.J. 1997. Experiments in ecology: their logical design and interpretation using analysis of variance. Cambridge Univ. Press. (Finally we have the book which brings together all the ecological study design and statistics principles scattered throughout Underwood’s papers. Available in soft cover. You should seriously consider buying it.) Winer, B.J. 1971. Statistical principles in experimental design, 2nd Ed. McGraw-Hill, New York. (One of the classics on experimental design, from a social sciences perspective but that doesn’t matter very much.) Zar, J.H. 1996. Biostatistical analysis, 3rd Ed. Prentice-Hall, Englewood Cliffs, N.J. (Introductory biostatistics, and a well-done job. Besides a good introduction to all the obvious things, there is also power analysis, circular distribution statistics (e.g., times of day, stages of the tide, directions of animal movement and orientation, etc.), nested ANOVA, and other gems.)
Papers on study design and statistical methods useful for environmental studies (some have comments, some don’t) Chapman, P.M. 1996. Presentation and interpretation of Sediment Quality Triad data. Ecotoxicology 5:327-339. (Chapman's updating of the SQT in response to other papers on it.) Clarke, K. R. and R. H. Green. 1988. Statistical design and analysis for a 'biological effects' study. Mar. Ecol. Prog. Ser. 46: 213-226. Dennis, B. 1996. Discussion: Should ecologists become Bayesians? Ecol. Appl. 6: 1095-1103. Eberhardt, L. L. 1976. Quantitative ecology and impact assessment. J. Envir. Man. 4: 27-70. (A classic that should be more famous than it is. Many current issues were considered in this paper two and a half decades ago. I didn’t know this paper when I wrote my 1979 book.) Fairweather, P. G. 1991. Statistical power and design requirements for environmental monitoring. Austr. J. Mar. Freshw. Res. 42: 555-567. Green, R. H. 1984. Statistical and nonstatistical considerations for environmental monitoring studies. Environ. Monit. Assessm. 4: 293-301. Green, R. H. 1989. Power analysis and practical strategies for environmental monitoring. Environm. Res. 50: 195-205. Green, R. H. 1993. Application of repeated measures designs in environmental impact and monitoring studies. Austr. J. Ecol. 18: 81-98. Green, R. H. 1994. Aspects of power analysis in environmental monitoring. In “Statistics in Ecology and Environmental Monitoring”, D. J. Fletcher and B. F. J. Manly, eds., p. 173-182, Otago Conference Series. University of Otago Press, Otago, New Zealand. Green, R.H., J.M. Boyd, and J.S. Macdonald. 1993. Relating sets of variables in environmental studies: the Sediment Quality Triad as a paradigm. Environmetrics 4:439-457. (The data analyzed here are from Vancouver Harbour.) Green, R.H., and P. A. Montagna. 1996. Implications for monitoring: study designs and interpretation of results. Can. J. Fish. Aquat. Sci. 53:2629-2636. (In the GOOMEX 8-paper set. Three applications to GOOMEX data: sediment quality triad, increased error variance as result of impact, future study design recommendations.) Green, R.H. and S. R. Smith. 1997. Sample program design and environmental impact assessment on coral reefs. Proc. 8th Int. Coral Reef Symp. 2: 1459-1464. (Repeated measures design for coral reefs, with examples.) Green, R. H., and R. C. Young. 1993. Sampling to detect rare species. Ecol. Appl. 3: 351-356. (This resulted from a request to help with a contract from the U.S. Office of Endangered Species, re. how to sample when trying to find such a species in a habitat. The answer turned out to be simple and elegant. The database is unionid molluscs in Tennessee and Virginia rivers. I have had lots of reprint requests for this one.) Jones, G.P. and Kaly, U.L. 1996. Criteria for selecting marine organisms in biomonitoring studies. In: “Detecting Ecological Impacts: Concepts and Applications in Coastal Habitats”, R. J. Schmitt and C. W. Osenberg, eds., p. 29-48. Academic, New York. Kennicutt, M.C., II, R.H. Green, P. Montagna and P.F. Rosigno. 1996. Gulf of Mexico Offshore Operations Monitoring Experiment (GOOMEX), Phase I: sublethal responses to contaminant exposure - introduction and review. Can. J. Fish. Aquat. Sci. 53:2540-2553. (See comments on Peterson et al 1996, below. This one is the introductory paper of the 8-paper set. It presents the overall study design and planned statistical analysis, which was my responsibility in the project.) Mapstone, B. D. 1995. Scalable decision rules for environmental impact studies: effect size, Type I, and Type II errors. Ecol. Appl. 5: 401-410. McDonald, L. L. and W. P. Erickson. 1994. Testing for bioequivalence in field studies: Has a disturbed site been adequately reclaimed? In “Statistics in Ecology and Environmental Monitoring”, D. J. Fletcher and B. F. J. Manly, eds., p. 183-197, Otago Conference Series. University of Otago Press, Otago, New Zealand. Olsgard, F., P. J. Somerfield, and M. R. Carr. 1997. Relationships between taxonomic resolution and data transformations in analyses of a macrobenthic community along a established pollution gradient. Mar. Ecol. Prog. Ser.: 1-9. (Patterns of macrobenthic data in vicinity of North Sea oilfield show high degree of consistency up to taxonomic level of order.) Peterman, R. M. 1990. Statistical power analysis can improve fisheries research and management. Can. J. Fish. Aquat. Sci. 47: 2-15. Peterson, C. H. 1993. Improvement of environmental impact analysis by application of principles derived from manipulative ecology: Lessons from coastal marine case studies. Austr. J. Ecol. 18: 21-52. Peterson, C.H., M.C. Kennicutt, II, R.H. Green, P. Montagna, D.E. Harper, Jr., E.N. Powell, and P.F. Rosigno. 1996. Ecological consequences of environmental perturbations associated with offshore hydrocarbon production: a perspective on long-term exposures in the Gulf of Mexico. Can. J. Fish. Aquat. Sci. 53: 2637-2654. (This is the “summary of results & implications” paper, the 8th of an 8-paper set reporting the results of a 2-3 year study by a half-dozen principal investigators. This set of papers has had a major impact, and its conclusions are somewhat controversial -- for example that there is no apparent biological impact beyond a few hundred meters from drilling platforms.) Peterson, C.H., L.L. Macdonald, R.H. Green, and W.P. Erickson. 2001. Sampling design begets conclusions: the statistical basis for detection of injury to and recovery of shoreline communities after the Exxon Valdez oil spill. Marine Ecology Progress Series 210: 255-283. (This is a major review paper which will be controversial. Its main theme is that the Exxon-funded study design whatever the intentions were and whatever their design’s other virtues, was bound to have low power to detect impacts by the oilspill on the biological community. And it did, in comparison with the government-funded & supervised studies. Mind you, I was Chair of the Statistical Working Group for the latter and Peterson was Chief Scientist, so we can’t claim to be disinterested.) Stewart-Oaten, A. 1996. Goals in environmental monitoring. In: “Detecting Ecological Impacts: Concepts and Applications in Coastal Habitats”, R. J. Schmitt and C. W. Osenberg, eds., p. 17-27. Academic, New York. Suter, G. W. 1996. Abuse of hypothesis testing statistics in ecological risk assessment. Human and Ecological Risk Assessment 2: 331-347. Underwood, A. J. 1981. Techniques of analysis of variance in experimental marine biology and ecology. Ann. Rev. Oceanogr. Mar. Biol 19: 513-605. Underwood, A. J. 1991. Beyond BACI: experimental designs for detecting human environmental impacts on temporal variations in natural populations. Austr. J. Mar. Freshw. Res. 42: 569-587. Underwood, A. J. 1992. Beyond BACI: the detection of environmental impacts on populations in the real, but variable, world. J. Exp. Mar. Biol. Ecol. 161: 145-178. Underwood, A. J. 1993. The mechanics of spatially replicated sampling programmes to detect environmental impacts in a variable world. Austr. J. Ecol. 18: 99-116.. Underwood, A. J. 1994. Things environmental scientists (and statisticians) need to know to receive (and give) better statistical advice. 33-61. Underwood, A. J. 1994. On beyond BACI: sampling designs that might reliably detect environmental disturbances. Ecol. Appl. 4: 3-15. Underwood, A. J. 1997. Experiments in ecology: their logical design and interpretation using analysis of variance. Cambridge, UK: Cambridge University Press. Walters, C.J. and R.H. Green. 1997. Valuation of experimental management options for ecological systems. J. Wildl. Man. 61(4):987-1006. (Roughly bashed together during a sabbatical period I spent at UBC, then finished by back-and-forth fax and email. Walters and I come from opposite philosophical directions re. estimation vs. testing, and Bayesian vs. Fisherian. This paper is an attempt to reconcile these philosophically different approaches in ecological applications.)
Papers on biomonitoring especially with bivalve molluscs (some have comments, some don’t) Bailey, R. C., and R. H. Green. 1988. Within-basin variation in the shell morphology and growth rate of a freshwater mussel. Can. J. Zool. 66: 1704-1708. Green, R. H. 1971. A multivariate statistical approach to the Hutchinsonian niche: bivalve molluscs of central Canada. Ecology 52: 543-556. Green, R. H. 1972. Distribution and morphological variation of Lampsilis radiata in some central Canadian lakes: a multivariate statistical approach. J. Fish. Res. Board Can. 29: 565-1570. Green, R. H., R. C. Bailey, S. G. Hinch, J. L. Metcalfe, and V. H. Young. 1989. Use of freshwater mussels to monitor the nearshore environments of lakes. J. Great Lakes Res. 15: 635-644. Green, R. H., S. M. Singh and J. M. McCuaig. 1983. An arctic intertidal population of Macoma balthica: genotypic and phenotypic components of population structure. Can. J. Fish. Aquat. Sci. 40: 1360-1371. Green, R. H., S. M. Singh, and R. C. Bailey. 1985. Bivalve molluscs as response systems for modelling spatial and temporal environmental patterns. Sci. Tot. Envir. 46: 147-169. Hinch, S. G., and R. C. Bailey. 1988. Within- and among-lake variation in shell morphology of the freshwater clam Elliptio complanata. Hydrobiologia 157: 27-33. Hinch, S. G. and L. A. Stephenson. 1987. Size and age specific patterns of trace metal concentrations in freshwater clams from an acid sensitive and a circumneutral lake. Can. J. Zool. 65: 2436-2442. Hinch, S. G., and R. H. Green. 1988. Shell etching on clams from softwater Ontario lakes: a physical or chemical process? Can. J. Fish. Aquat. Sci. 45: 2110-2113. Hinch, S. G., and R. H. Green. 1989. The effects of source and destination on growth and metal uptake in freshwater clams transplanted among south-central Ontario lakes. Can. J. Zool. 67: 855-863. Hinch, S. G., L. J. Kelly, and R. H. Green. 1989. Morphological variation of Elliptio complanata in softwater lakes exposed to acidic deposition. Can. J. Zool. 67: 1895-1899. McCuaig, J. M., and R. H. Green. 1983. Unionid growth curves derived from annual rings: a baseline model for Long Point Bay, Lake Erie. Can. J. Fish. Aquat. Sci. 40: 436-442. Metcalfe-Smith, J. L., and R. H. Green. 1992. Ageing studies on three species of freshwater mussels from a metal-polluted watershed in Nova Scotia, Canada. Can. J. Zool. 70: 1284-1291. Metcalfe-Smith, J.L., R.H. Green and L.C. Grapentine. 1996. Influence of biological factors on concentrations of metals in the tissues of freshwater mussels (Elliptio complanata and Lampsilis radiata radiata) from the St. Lawrence River. Can. J. Fish. Aquat. Sci. 53:205-219. (Brings together the literature, integrating the marine work with our freshwater work, and showing how powerful statistical methods can elucidate different components of biological variation caused by metal vector patterns.) Smith, A. L., R. H. Green & A. Lutz. 1975. Uptake of mercury by freshwater clams (Family Unionidae). J. Fish. Res. Board Can.32: 1297-1304.
Statistical analysis software
We will use Minitab Release 13 for Windows as our main working statistical analysis software. Please note that you can download a demo from their website. Minitab has made available to me a "timed" copy of Minitab Release 13 which I will bring to the workshop for installing in workshop computers (and participants' computers too if they wish). They will expire on 1 October 2001 and you'll have to buy one then if you want to keep using it. We can see how many want it at that point and I'll see what sort of "group price" I can get. We will try some other statistical software for particular kinds of analysis, for example Glim (for linear models where the response doesn't need to be normally distributed or additively related to the predictors, e.g. probit, logistic and loglinear models), Simca (for Canonical Analysis), and RT (randomization tests of hypotheses). Maybe SAS which is a real workhorse for big projects with several Principal Investigators, a "Statistical Analysis System" to be sure but also a good database manager.
Information about workshop participants (alphabetical)
(E-mail to everyone including participants and organizers.) Kris Andrews Bonnie Antcliffe Julia Beatty, Head, Environmental Assessment Section, BC Environment, Kootenay Region Email: Julia.Beatty@gems4.gov.bc.ca ph: 250-354-6750, fax: 250-354-6367 (Workshop local organiser) Janice Boyd, Environment Canada Email: Janice.Boyd@ec.gc.ca ph: 604-666-5908, fax: 604-666-7294 (Assisting Roger Green with the workshop) Roxanne Brewer Bruce Carmichael Jillian Chown Liz Freyman, biologist, BC Government Email: Liz.Freyman@gems1.gov.bc.ca Expectations: I am looking for tools/techniques where we could use simple statistics to help with our environmental impact assessments, particularly with relatively small datasets. I am interested in learning more about how to use Exel to do simple stats/graphs/etc. I am also very interested in ways of presenting limited data such as using graphical tools, visual interpretation techniques, etc. I am looking for good explanations of more complex statistical tools to help me assess research reports and scientific literature. For example: principal component analysis, canonical correspondence analysis, significant difference techniques, etc. Background: B.Sc. in Soil Science, 2 years experience in data management for GIS, 9 years experience in water quality issues, environmental impact assessment, chemistry and toxicology for aquatic, benthic and terrestrial ecosystems. Our current data sets are very limited. We do investigative sampling and therefore are looking for preliminary indicators of problems. We also do state of environment monitoring but our budgets and staff resources are limited and therefore our data is limited to simple comparisons such as with standards/criteria or upstream/downstream or before/after. Sampling media includes mainly water (marine and freshwater), sediments, benthic invertebrates (taxonomy and tissue chemistry), soils. Statistics background: after university graduation, I have very little experience with stats. My computing experience for stats is outdated: I have used Systat and SPSS in the late 1980's. Bob Grace Colin Gray Basil Hii, Hydrogeologist, Environment Canada Address: 1200 West 73rd Ave Suite 700, Vancouver BC V6P 6H9 Email: basil.hii@ec.gc.ca ph: 604 664 4039, fax: 604 664 9126 Expectations: ? Background: ? Vic Jensen, impact assessment biologist, BC Government Email: Vic.Jensen@gems3.gov.bc.ca Expectations: I would like to better understand the basics; how to determine adequate sample size to find a specified difference of x between two populations of data. I recently ventured into benthic invertebrates and would appreciate a better understanding of multivariate analysis of those sort of data sets. Background: My job as an impact assessment biologist largely deals with ambient water quality near impact zones; ie before after control impact scenarios. Statistics rarely go beyond max min mean stdev. I took stats 24 years ago during my dusty BSc and had a refresher course at Mesatchie Lake a few years ago. Kim Keogh Steve Macdonald, Fisheries & Oceans Canada (at SFU) Email: MacDonaldSt@pac.dfo-mpo.gc.ca ph: 604-666-7910 (Assisting Roger Green with the workshop) Les McDonald Bev McNaughton Brent Moore Narender Nagpal Remi Odense Earl Plain Larry Pommen, Senior Assessment Engineer, Water Management Branch, BC Ministry of Environment Email: Larry.Pommen@gems4.gov.bc.ca ph: 250-387-9516, fax: 250-356-8298 Expectations/needs: A sufficient awareness of environmental study design and analysis to know when to retain statistical expertise to assist with design and analysis and to supervise their work. Background: MSc in environmental engineering with 28 years of experience in the water quality/water pollution field, including monitoring, assessment, reporting, and water quality objectives and guidelines development. Little formal statistical training. Computing skills include word processing and spreadsheets, but no use of statistics software. My work mostly involves measuring water quality constituents in freshwater streams and lakes, with some measurement of constituents in bottom sediments and biological tissues. The primary purposes of the measurements are to detect trends over time, to develop water quality objectives, and to assess compliance with water quality objectives. Robyn Roome, Impact Assessment Biologist, BC Ministry of Environment Email: robyn.roome@gems1.gov.bc.ca ph: 250-354-6356 Expectations: I am interested in impacts to benthic invertebrate communities, fish populations, and measuring impacts to receiving waters. I am also interested in learning more about statistical design in environmental impact assessment. In the future I hope to learn more about applying techniques such as principal components analysis to benthic invertebrate data. Background: I recently started working as an environmental impact assessment biologist at the Ministry of Environment. Prior to that I worked as a fisheries consultant and I am also finishing a M.Sc. at the University of British Columbia. The courses I have taken in statistics include two semesters of biometrics at a second year undergraduate level. I have worked with Systat and SigmaStat (Jandel Scientific) statistical software. I am interested in impacts to benthic invertebrate communities, fish populations, and measuring impacts to receiving waters. I am also interested in learning more about statistical design in environmental impact assessment. In the future I hope to learn more about applying techniques such as principal components analysis to benthic invertebrate data. Andrea Ryan Patrick (Pat) Shaw, Environmental Quality Objectives Specialist, Environment Canada Email: pat.shaw@ec.gc.ca ph: 604-664-4071, fax: 604-664-9126 Expectations: I'm expecting this workshop to be mostly a refresher on statistical concepts and methods, but also an introduction to alternate and novel ways of looking at data. Making best use of residuals from regressions, concepts from ANOVA designs, application of somewhat newer techniques including resampling methods. I'd be quite interested in learning more about nonparametric techniques as well. Background: For the past 10 years, I've worked for Environment Canada conducting a range of environmental surveys and designed experimental studies. In a previous academic incarnation, I did a variety of multivariate morphometric studies of crustacea in a seach for an underlying evolutionary trends. Leading to that work was a period of marine consulting, looking at marine environmental effects of various discharges and other stressors. Formal statistical training included a variety of undergraduate and graduate courses in basic statistics, modelling, theoretical ecology and numerical ecology. Stephanie Sylvestre, Environmental Studies Scientist, Environment Canada Email: stephanie.sylvestre@ec.gc.ca ph: 604-664-4099, fax: 604-664-9126, field cell phone: 604-839-2493 Expectations: Am looking for a refresher in statistical areas that I haven’t used in years. My work with Environment Canada has proven to me that no matter how hard I try, statistically good designs are difficult! The site selections are usually politically driven and funding driven. We are expanding into other study areas that require other statistical techniques that I haven't used in years and I am simply wanting a refresher course. I have been also working on stream assessments using fish and crayfish exposure studies. Our focus is on effects of non-point source pollution in agricultural and urban areas. Our exposure studies have been conducted using an upstream downstream approach. Due to the difficulties of finding reference areas we only have one reference site. We are working on two stream systems. One stream system has one upstream, one mid-stream and one downstream site. On the second stream we have only a downstream site. Although design advice will not help this particular study, it would help in additional studies similar to this. It may even help in planning subsequent studies to field validate some of our results. We have a variety of data types from single sample organic contaminant data at each site to regular nutrient and monitoring data (n=30) at each site, as well as individual fish/crayfish health assessment data (n=25) at each site. My primary goal with this refresher course is analyzing this upstream/downstream data in the most effective way. Background: Since my time at Western, I have been working regularly with ordinations and discriminant analysis of invertebrate data to continue with Bob and Trefor’s RCA work. I took a multivariate course instructed by Dr. Eric Smith through the US Fish and Wildlife Service in West Virginia in 1999. It was a good course, a very hands on one with lots of real life examples and a variety of multivariate techniques. Dave Sutherland Taina Tuominen, Head, Aquatic Sciences Section, Environment Canada Email: taina.tuominen@ec.gc.ca ph: 604-664-4054, fax: 604-664-9126 Expectations: I have a need to review study designs and review reports/papers with my staff and, where possible, provide advice on suggested approaches for answering questions/hypotheses in proposed studies. Therefore, I would like to enhance my understanding of environmental study design and statistical analyses available for assessing data. The statistical needs encompass: analysis of trends in long-term data sets, determination of differences in contaminant content or biological effect among study sites and relative to differing stresses, application of multivariate methods and clustering for displaying and interpreting environmental data, suggested approaches for analysing environmental data (often consisting of less than ideal number of replicates and a high degree of variability). Background: I have an MSc in biology. I have basic statistics from university (SFU). I never did take a biostatistics course; however, I have applied parametric and nonparametric statistics to datasets. Recently, I have conducted very limited statistical analyses (Excel); however in the past I have used an interactive package available at SFU and SPSS for analyses, such as ANOVA regression analysis and non-parametric tests. My current work is as section head, and as such, I oversee several projects addressing both surface water (streams and rivers) and groundwater issues. Most of our projects involve either water quality monitoring programs or benthic monitoring programs and assessment of trends detected by these programs or shorter-term studies addressing the effects of human activities on the aquatic environment. Graham Veale Tom Webber Paul Willis Cecilia Wong
Example questions-hypotheses-datasets
(go to ones that have been posted)
If you have an interesting environmental monitoring or impact study problem
(defined by a question and hypotheses) with an associated dataset, that you think
would be of interest to all workshop participants, send it to Roger Green for posting
on this website. Just go back to the menu and click on the "Contact Roger Green"
option. Actually what I think would be best would be to have the dataset on a public
ftp site. If you can have it on one, just let me know the site and filename and that's
what I'll post. If not, send me the dataset as an attachment to an e-mail (an Excel
file would be good) and I'll put it on a public ftp site at UWO. In addition you
should send me (and I'll post): a one-short-paragraph description of the study, the
basic question involved, and the null and at least one alternative hypothesis.
Vic Jensen: I will likely bring some trend data comparing two or more lake sites for one or more parameters. Q - has there been a change outside over time and are the two or more data sets essentially the same. I could bring some taxonomic data if you think it would be useful.
Short professional resume - Roger H. Green
Academic qualifications:
- Bachelor of Science in Biology, College of William & Mary, 1961
- Ph.D. in Zoology, Cornell University, 1965
- Fulbright Postdoctoral Fellowship, University of Queensland, 1965-66
- Resident Ecologist, Marine Biological Lab, Woods Hole, 1966-68
- Asst. and Assoc. Professor, Department of Zoology, University of Manitoba, 1968-76
- Assoc. Professor and Professor, Department of Zoology, University of Western Ontario,
1977-99
- Professor Emeritus, Department of Zoology, University of Western Ontario, 1999-
Publications:
- 55 papers in refereed journals
- 23 papers in refereed conference or workshop proceedings
- 1 book: Sampling Design and Statistical Methods for Environmental Biologists. 1979. Wiley,
New York.
- 1 book in prep: Environmental Study Design and Analysis: Principles and Examples. Wiley,
New York.
- 8 contributions to a book or document
- 4 invited book reviews
Invited contributions to conferences or workshops (since 1997):
- International Conference on Ecology of Estuaries, Deakin University, Victoria, Australia,
February 1997
- Aquatic Toxicity Workshop, Niagara Falls, Ontario, November 1999
- Half-day workshop on design and statistical analysis of ecological data from a barrier
island, University of North Carolina Institute of Marine Sciences, Morehead City, North
Carolina, March 1998
- Workshop on Design of Forest Biodiversity Monitoring Program, Edmonton, Alberta, April 1999
- Ninth Lukacs Symposium on Frontiers of Environmental and Ecological Statistics for the 21st
Century, Bowling Green, Ohio, April 1999 (I received an Outstanding Ecologist award)
- Second ASEAN Experts Meeting on Regional Criteria and Indicators for Sustainable Forest
Management, Brunei, October 1999
- Asia-Pacific Conference on the Biology of the Environment, Singapore, November 1999
- Nature Conservancy planning workshop, Monterey, California, December 1999
- Sable Offshore Energy Program workshop, Dartmouth, Nova Scotia, March 2000
- American Society of Testing & Materials (ASTM) conference, special session on biomonitoring,
April 2000
- Fisheries & Oceans Canada/Ministry of Natural Resources Ontario Workshop on environmental
monitoring of fish habitat, March 2001
Workshops I organized, and other external teaching and academic duties (since 1997):
- External examiner for School of Biological Sciences, National Institute of Education,
Nanyang Technological University, Singapore, 1996-1997
- Workshop on Design and Statistical Approaches to Applied Studies with Emphasis on
Multivariate Methods, Singapore Institute of Biology and Department of Biological Sciences
National University of Singapore, Singapore, March 1999
Graduate and postdoctoral students supervised:
- Masters students: 12 completed and 2 writing up
- Ph.D. students: 8 completed
- Postdoctoral students: 3 completed
Research grants & contracts through University of Western Ontario since 1997:
- Research grants from Natural Sciences & Engineering Research Council of Canada, for studies
on freshwater and marine benthic communities, to 2003
Consulting services provided (since 1997):
- Attorney General's Office State of California - advising and expert witness on design and
statistical analysis of results from environmental monitoring of Diablo Canyon Power Plant
cooling water intake and discharge system (1995-97)
- Natural Resources Canada (CANMET) - advising on sampling design and statistical analysis in
the Aquatic Effects Technology Evaluation Program re. monitoring environmental impacts of
mine effluents (1996-99)
- Jacques Whitford Ltd. (Nfld) - participated in St. Johns Nfld workshop on design of
environmental monitoring for Terra Nova oilfield project, Grand Banks of Nfld (June 1997)
- Martec Ltd. - advised re. design of, assisted with statistical analysis & interpretation
for, and helped prepare reports on environmental effects monitoring re. inshore portion
of underwater gas pipeline from the Sable offshore field to the Nova Scotia shore (1998- )
- Jacques Whitford Ltd. (Nova Scotia) - advised re. design of, assisted with statistical
analysis & interpretation for, and helped prepare reports on offshore field environmental
monitoring design for Sable Offshore Energy Program (1998- )
- Department of Land, Water and Environmental Conservation, State of New South Wales,
Australia - advising on design of long-term monitoring program for large inland rivers
(1998)
- LGL Ltd/U.S. Minerals Management Service - advising on design and analysis for study on
feeding by bowhead whales on the Alaska north slope adjacent to the Beaufort Sea (1998- )
- RL&L Environmental Services, Castlegar, British Columbia - analysis and modeling of
temperature data for an impoundment on the Columbia River in British Columbia (1999-2000)
- Regional Centre for Forest Management, Kuala Lumpur, Malaysia - reviewing documents and
advising on design aspects of managing SE Asian forests to maintain biodiversity, and
periodic reviewing and advising re. environmental concerns & forest management practice
(1999- )
- University of Guelph, Ontario Hydro, and Chippewa First Nation - member of Advisory
Committee for project on "Whitefish Interactions with Nuclear Generating Stations" (2000- )
- Ontario Ministry of Environment, reviewer of research document on mine-waste contamination
in Moira River (2000)
- Assisted with statistical analysis of Ph.D. research data on Malayan monitor lizards at
Sungei Buloh Nature Park, Singapore (2000-01)
- International Council for Science, member of UN scientific panel assessing impact of
uranium mining near Kakadu National Park, Australia - a World Heritage Site (2000)
- Alberta Ministry of Environment, Terrestrial Environmental Effects Monitoring program, re.
statistical issues related to compositing of samples prior to chemical analysis
(2000-2001).
- Analysis of migratory bird data for Sungei Buloh Nature Par, Singapore, and ecological
studies of marine biota (2001- )
January 2001