Experimental design in the LTER network: incorporating the human element.
Diane Hope
Key Participants: David Casagrande, Bill Cook, Steve Carpenter, Peter Groffman, Bill Freudenberg, Chris Martin, Nancy Grimm, Scott Collins, Alan Yeakley, Dick Lathrop, Freeda Traub
Summary
The workshop was attended by about 25 participants. Following a pre-arranged schedule of invited short presentations then open discussions, it generated some insightful and lively debate. Several of the workshop participants have since been working on a manuscript generated from some of the ideas discussed, entitled:
“Learning to roll with the punches: Adaptive experimentation in human-dominated systems” by main authors William Cook, David Casagrande, Diane Hope, & Peter Groffman. The manuscript will be submitted to the ESA journal “Frontiers in Ecology and the Environment” this month.
Overview
In the workshop we introduced the topic by reviewing the current use of manipulative experiments in the LTER network (see below) and suggested that manipulative experiments incorporating approaches from both the biophysical and social sciences are critical to understanding pattern and process in human-dominated ecosystems. Several invited contributors each then presented to the group on what they saw as key issues in this area. Steve Carpenter talked about the need for novel approaches and the use of simulation modeling to help determine the correct types of experiments to be performed; Peter Groffman outlined the advantages and disadvantage of the comparative watershed approach as a surrogate for manipulative experiments at larger scales; Bill Freudenberg reviewed some of the unintended consequences of performing experiments involving humans as integral components.
The group then spent some time discussing the compromises, challenges and opportunities required to include humans as an integral component in manipulative experiments. To date, this has rarely been done, largely because of ethical implications and logistical difficulties. So in the last part of the workshop we discussed a case example of the residential landscaping experiment being developed at the Central Arizona-Phoenix Long Term Ecological Research program (CAP LTER), some design details of which were outlined by Chris Martin. This was successful in stimulating much useful discussion on specific issues important to take account of when embarking on this kind of work. Amongst the most important issues to arise pertaining to the planned CAP LTER residential landscaping experiment were:
- sufficient detailed pre-treatment baseline surveys of the biophysical and social environment BEFORE experimental manipulations are started
- an experimental design that is statistically justifiable, but which does NOT give in to over-jealous use of ‘pseudoreplication’ dogma. It was generally agreed that non-replicated block designs may often be necessitated in such experiments, to balance the need to get neighborhood & community effects with the limitations on available resources
- continued monitoring of untreated households in the surrounding ‘matrix’, as well as treated and identified control home owners.
Since the workshop, a number of core workshop participants have been working on a draft manuscript entitled ‘Experimental design in human dominated ecosystems’, which is being prepared for submission to ‘Frontiers in Ecology and the Environment’ or a similar journal. I suggest that an abbreviated version of this manuscript be posted on the LTER Network site, as a kind of ‘Guidelines when initiating experimental designs involving human agents’, once the full version of the manuscript has been accepted for publication.
There are a variety of experimental approaches that investigate at least one direction of human-biophysical feedback. The following is a review of the main types, illustrated by reference to type examples of each major conceptual category.
Controlled, replicated experiments. Plots are used to study impacts of anthropogenic inputs on communities and ecosystem processes at small spatial scales (e.g., the effects of nitrogen deposition, Wedin and Tilman 1996). This approach allows extensive replication, strict controls, and the consideration of multiple variables via factorial treatments, but often does not allow study of processes operating on scales larger than a few meters. Manipulations of larger areas, including control units, have been used to explore the impacts of clear-cuts on water and nutrient flow (Hornbeck et al. 1993), acidification in lakes (e.g. at NTL), fire and livestock grazing in grasslands (Knapp et al. 1998) and agricultural processes (e.g. at KBS) on populations, communities and nutrient cycling. These experiments allow consideration of larger-scale processes than do plot-scale studies, but extensive replication is more difficult and less common.
Large-scale manipulations: the watershed approach. An experiment is a deliberate manipulation of a system to elicit some expectation – it does not necessarily require a control. This broader view of what constitutes experimentation is often required where humans are involved.
Chronosequences and natural experiments. Many unplanned experiments may not be recognized or exploited as such until later in time; the effects of land management on biological processes can often be studied by post-hoc comparisons of landscapes with varying historical management techniques (Mitchell et al. 2002, Currie and Nadelhoffer 2003).
Social science approaches. Popular motivation for the use of particular land managements, and the perceptions and values of different landscape types, are critical for understanding anthropogenic processes in ecosystems. Social scientists have generally used historical or comparative analyses to study such processes, but relevant approaches also include psychological laboratory experiments and, rarely, in situ experiments. Researchers are using historical and comparative approaches for analyzing patterns of land use and sociological variables across space and time (Turner et al. 1996, Hall et al. 2002). Sociological surveys of human perceptions before and after ecological restoration projects can provide quantitative measures of manipulative effects. Studies of human migration also provide valuable insights into human response to ecosystemic change.
References
Currie, W. S. and K. J. Nadelhoffer.
2003. The imprint of land use history: patterns of carbon and nitrogen in
downed woody debris at the
Hall, B., G. Motzkin, D. Foster, M. Syfert, and J. Burk. 2002. Three
hundred years of forest and land-use change in
Hornbeck, J.W., M.B. Adams, E.S.
Corbett, E.S. Verry, J.A. Lynch. 1993. Long-term
impacts of forest treatment on water yield: a summary for northeastern
Knapp, A.K., J.M.
Mitchell, Chris E.; Turner, Monica G.; Pearson, Scott M. 2002. Effects of historical land use and forest patch size on myremecochores and ant communities. Ecological Applications. 12(5): 1364-1377.
Turner, M. G., D.N. Wear,
and R.O. Flamm.
Wedin, D. A.; Tilman, D. 1996. Influence of nitrogen loading and species composition on the carbon balance of grasslands. Science 274:1720-1723.
Report
first prepared:
Updated
: