NCEAS, the National Center for Ecological Analysis & Synthesis in Santa Barbara, California, is a major centre for ecological workshops. The workshop I attended, organised by Sandy Liebhold and Jessica Gurevitch in 2000 and 2001, was entitled

'Integrating the Statistical Modeling of Spatial Data in Ecology'.

The workshop participants became good friends. Here is a photo of us at NCEAS.


. From left: Mike Hohn, Steve Citron-Pousty, Phil Dixon, Marie-Josee Fortin, Joe Perry, Jennifer Dungan, Mark Dale, Maria Miriti, Mike Rosenberg, Jessica Gurevitch, Sandy Liebhold, Don Myers, Pierre Legendre, Ottar Bjornstad, Tim Keitt, Dawn Kieling.





The group produced several papers, published in a special issue (Volume 9, number 2) of the journal Ecography, in November 2002. Two of these papers (see below) contain a considerable number of SADIE examples:



'Illustrations and Guidelines for Selecting Statistical Methods for Quantifying Spatial Pattern in Ecological Data' is a comparison of different methods to measure spatial pattern, built around a set of four case studies, and designed to try to help ecologists make the right choice of spatial methods to suit their data and questions. The paper also features a taxonomy of data. The 23 figures attempt to give a empirical basis to the paper, and to facilitate a visual feel for what the analysis of spatial pattern entails. The full reference is:

Perry, J.N., Liebhold, A., Rosenberg, M.S., Dungan, J., Miriti, M., Jakomulska, A. & Citron-Pousty, S. (2002). Illustration and Guidelines for Selecting Statistical Methods for Quantifying Spatial Patterns in Ecological Data. Ecography, 25, 578-600.

What were the conclusions? - you'll have to download the paper to find out!



'A balanced view of scaling in spatial statistical analysis', led by Jennifer Dungan, is a discussion, with examples, of the issue of spatial scale - what it is, how to deal with it, and why ecologists get so exceited about it. Lots of people talk about scale, but the subject area is bedevilled by lack of agreed definitions. We look at the effects of scale on measures of spatial pattern, by defining it in terms of changes in 'support', 'extent' and 'lag'. The full reference is:

Dungan, J.L., Perry, J.N., Dale, M.R.T., Citron-Pousty, S., Fortin, M.-J., Jakomulska, A., Legendre, P., Miriti, M. & Rosenberg, M. (2002). A balanced view of scaling in spatial statistical analysis. Ecography, 25, 626-640.

What are 'support', 'extent' and 'lag'?
If you want to find out, download Jennifer et al.'s paper!






The other papers published by the group

Below, download the other papers published by the group in the special issue of Ecography.

Sandy & Jessica, the co-organizers of the workshop,
wrote an introduction to the issue.

Mark Dale, Phil, Marie-Josee, Pierre, Don & Mike R., described the
conceptual and mathematical relationships between methods for spatial analysis.
There is a section in Mark et al.'s paper showing how SADIE fits in (not very well!)
with the traditional cross-product derivation common to many spatial methods,
and what other methods SADIE is closest to in form and function.

Pierre Legendre led Mark, Marie-Josee, Jessica, Mike H. & Don,
in a group that generated thousands of simulations to study
the effect of spatial structure on design and analysis for field surveys.

Tim Keitt, Ottar, Phil & Steve considered different classes of
statistical approaches that allow for spatial autocorrelation in models
of interactions between an organism and its environment.

Alternatively, you can download all of the working group's papers direct
from the group's page on the NCEAS website





Some of the people at the workshop

Here are links to some of the people that shared in the work I did at NCEAS.


Mike Rosenberg is at Arizona State University. He has a brilliant, free, spatial software package entitled Passage. First formally released on 6 December 2001, Passage is very wide ranging and can handle many of the analyses that appeared in our papers.
Go to Mike's website to download a copy. Highly recommended!


Mark Dale is at the University of Alberta, Edmonton, Alberta, Canada.
Mark's book entitled 'Spatial Pattern Analysis in Plant Ecology', published by Cambridge University Press, (ISBN 0-521-45227-9, cloth US$69.95, 326 pp.) has received excellent reviews. One reviewer, for amazon.com said: "Would that I had owned this book when I was a graduate student sixteen years ago! Mark Dale has dedicated his entire career to the mechanics of quantitatively describing spatial pattern, and this book is an excellent summary and synthesis of twenty years of research". Not content with this, Mark is currently writing another book, with Marie-Josee Fortin, on spatial processes in ecology, which promises to become another standard text.






Marie-Josee Fortin is based in Canada, at the
University of Toronto. She is one of the Editors of the Canadian journal Ecoscience. A special issue (2002, Volume 9, number 2) was published by Ecoscience, devoted to spatial statistics, and edited by Marie-Josee.
I was very pleased to be invited to submit a paper for that issue, jointly with Phil Dixon:
(Perry, J.N. & Dixon, P., 2002, A new method for measuring spatial association in ecological count data. Ecoscience, 9, 133-141).
This paper describes the methodology behind the latest SADIE method to measure and test for spatial association, both globally and locally. Follow this link to
download a copy of the manuscript









Phil Dixon works in Ames, at Iowa State University. Phil is one of the leading statistical consultants in North America, and is an editor for the journal Ecology. Phil's favorite research projects are to develop and evaluate statistical methods to answer interesting biological questions. A lot of this work is collaborative. Major themes include the use of likelihood inference in non-standard situations and computer-intensive methods. Phil has been interested in spatial association and the measurement of seggregation for some time. Phil had a sole-authored paper in the special issue of Ecoscience mentioned above.





Sandy Liebhold works at the Northeastern Research Station - USDA Forestry Sciences Laboratory at Morgantown, West Virginia. I was fortunate to visit this wild state, especially beautiful in winter, and home to the Mountaineers. Sandy is pictured here with daughter Naomi, in a picture taken from his website. Sandy is an entomologist who, more than anyone else, has been responsible for the explosion of interest in the application of geostatistics to entomology and ecology. Sandy was co-organiser of the workshop. We all thank him and his Dad for their hospitality in his Dad's house in the mountains outside Santa Barbara.





Jessica Gurevitch works at SUNY, the famous
State University of New York at Stoneybrook. She has interests in meta-analysis and has co-authored a software package for meta-analysis with the goal of making these techniques more accessible to ecologists (MetaWin, Rosenberg, Adams and Gurevitch, publ. Sinauer Assoc., now in a second version). Have a look at her major review of meta-analysis in ecology and evolution, which has recently been published (Gurevitch et al. 2001, Advances in Ecological Research). Jessica was the other co-organizer of the NCEAS workshop.
Her other major non-teaching effort recently has been co-authoring a textbook, The Ecology of Plants (Gurevitch, Scheiner and Fox), published by Sinauer Associates in Spring 2002. You can find out more about the activities of her lab at Stoneybrook by following this link.


One of Jessica's students is Maria Miriti, now based at Ohio State University. It was Maria's excellent data on the desert shrub Ambrosia dumosa that formed one of the major datasets in our paper comparing methods for spatial pattern analysis with case studies.





Jennifer Dungan works as a Research Scientist in the Ecosystem Science and Technology Branch of the NASA Ames Research Center, California. Jennifer has an M.S. in Environmental Monitoring from the University of Wisconsin and a Ph.D. in Geography from the University of Southampton, UK. She has done research on modeling the global carbon budget for The Ecosystems Center, Marine Biological Lab in Woods Hole, and worked on mapping land use patterns in the tropics using remote sensing and Geographic Information Systems while at the University of Wisconsin. Since coming to NASA Ames in 1987, Jennifer has worked in ecosystem simulation modeling, imaging spectroscopy and geostatistical methods. Her primary research interest is in the elucidation of ecological patterns and processes using remotely sensed data, particularly as a function of scale (which explains her excellent leadership of the subgroup that wrote the NCEAS scale paper!).


Not much web-based information on our colleague Steve Citron-Pousty, but he is a lovely fellow and one of the funniest guys I've ever met. This is a good photo of him and his family.

The late Anna Jakomulska is not in the NCEAS photo above. Not long after the publications appeared in Ecography we were all shocked and saddened to hear of Anna's tragically early death in a motor cycle accident. Anna worked in the area of landscape ecology at the University of Warsaw, Poland. We remember her as a beautiful person and a dedicated young scientist, with a keen interest in geostatistics and its applications. You may learn more of her work and her department from their website.




Here is the famous Pierre Legendre, chef extraordinaire of the Duck with White Peaches recipe, who works at the University of Montreal. Pierre's website is a treasure trove of information and free downloads of much spatial and multivariate software and data.



Also at the workshop were the two ex-NCEAS post-docs:Ottar Bjornstad, a wizard ecologist and statistician now at Penn State, and Tim Keitt, now at SUNY.



Also, let's not forget the two geostatistics gurus: Mike Hohn of the West Virginia Geological Survey, P. O. Box 879, Morgantown, West Virginia 26507-0879, and Don Myers of the University of Arizona.






Philosophy of the workshop

In many areas of ecology there is an increasing emphasis on spatial relationships. Often, ecologists are interested in new ways of analyzing data with the objective of quantifying spatial patterns, and in designing surveys and experiments in light of the recognition that there may be underlying spatial pattern in biotic responses. In doing so, ecologists have adopted a number of widely different techniques and approaches derived from different schools of thought, and from other scientific disciplines. While the adaptation of a diverse array of statistical approaches and methodologies for the analysis of spatial data has yielded considerable insight into various ecological problems, this diversity of approaches has sometimes impeded communication and retarded more rapid progress in this emergent area. Many of these different statistical methods provide similar information about spatial characteristics, but the differences among these methods make it difficult to compare the results of studies that employ contrasting approaches.

The published papers explore possible areas of agreement and synthesis between a diversity of approaches to spatial analysis in ecology. The working group comprised a diverse set of scientists interested in various aspects of the statistical analysis of spatial data in Ecology. The papers appear in a special section of the October, 2002 issue of the journal "Ecography". I thank the Nordic Ecological Society for allowing me to make these papers available here. More information about the journal, Ecography, can be viewed at the Ecography web site.