ALL VOLUMES SEARCH TIEE
VOLUME 18 TEACHING ISSUES AND EXPERIMENTS IN ECOLOGY
PRACTICE

Arrested succession? Assessing plant community recovery on reclaimed oil and natural gas well pads in Alberta's boreal forests using multivariate analyses

This is a photo of a pump jack in an Alberta Boreal Forest. The well pads in this dataset have been reclaimed but the photo of the pump jack will help to convey the type of disturbance that we are studying the recovery of plant communities from after they have been decommissioned and received a reclamation certificate.

Crude oil pump jack in Alberta Boreal Forest. Pump jacks have been removed from well pads studied in this dataset. Pump jacks of this type created the original disturbance of plant communities. All areas sampled were decommissioned well pads that had undergone reclamation and received a certificate of reclamation. (Photo by the Ecological Recovery Project)

AUTHORS

Michelle Rude, Chantal Ricard, Randi Lupardus, and Anne C.S. McIntosh

University of Alberta, Augustana Faculty, 4901 46 Ave, Camrose, Alberta, Canada, T4V 2R3

Corresponding author: Anne McIntosh, amcintos@ualberta.ca

THE ECOLOGICAL QUESTION

Do plant communities on reclaimed well pads have similar composition to adjacent reference forests? (i.e., have they recovered?)

ECOLOGICAL CONTENT

Succession, Disturbance ecology, Community ecology, Plant ecology, Forest ecology

WHAT STUDENTS DO

Students will use a dataset (two .csv files) that compares plant community composition on reclaimed well pads and adjacent forested areas to develop hypotheses about how plant communities will recover from this disturbance and then investigate plant community recovery on the well pads using multivariate statistical analyses. An accompanying step-by-step primer is provided that will enable students to conduct a non-metric multi-dimensional scaling (NMS) ordination analysis using RStudio, which is an open-source statistical software environment. In addition to the ordination, students will also statistically test for differences in plant community composition among the well pad and reference forests using permutational multivariate analysis of variance (perMANOVA). They will then use indicator species analysis to identify which plant species are associated with the well pads compared with the reference forest study units. Finally, they can report summary statistics for the well pad and reference plant species of interest. There is a student worksheet that will be completed alongside the data analysis activity. For instructors that are not familiar with NMS - this is a multivariate ordination technique that doesn't have any underlying assumptions about the data, and thus is very popularly used by ecologists who often have non-normal data. Points (which represent sampling sites – wellsites and reference sites) closer together in the ordination are more similar than those that are further apart. As the ordination just visualizes the separation among the sampling sites, perMANOVA, which is a multivariate equivalent of ANOVA that also does not require normality of the dataset to proceed, tests whether there are significant differences in grouping variables among study sites - in this case well pads vs adjacent reference.

SKILLS

  • Develop hypotheses related to how plant communities may recover after disturbance - in this case an anthropogenic well pad disturbance.
  • Application of existing ecological knowledge (e.g., disturbance ecology, community ecology, succession) to a plant community dataset that focuses on recovery after well pad disturbance.
  • Recognize components of RStudio and use RStudio to run analyses and extract results from it.
  • Use of RStudio statistical software for multivariate statistical analyses, including ordination to visualize data and perMANOVA to test for differences among a priori plant community groups.
  • Interpretation of data to evaluate research hypotheses about plant community composition differences between well pads and reference forests.

STUDENT-ACTIVE APPROACHES

Guided Inquiry, Cooperative Learning, Quantitative Learning

ASSESSABLE OUTCOMES

By the end of this activity, students should be able to:

  • Write a high-quality ecological hypothesis to be tested by multivariate analysis.
  • Explain what types of data are appropriate for multivariate analysis.
  • Briefly explain what an ordination shows the reader.
  • Run a Non-metric multidimensional scaling (NMS) ordination in R that visualizes plant community composition patterns between well pads and adjacent reference forests.
  • Apply permutational multivariate analysis of variance (perMANOVA) to distinguish between plant community groups on the well pad and in the adjacent forest reference sites.
  • Use indicator species analysis to identify which species are contributing to separation among a priori groups (well pad vs forest).
  • Describe and interpret the output of their multivariate statistical analyses, including their plant community ordination plot, perMANOVA, and indicator species analyses.

SOURCES

Data from: Lupardus, R., A.C.S. McIntosh, A.Janz, and D. Farr. 2019. Succession after reclamation: Identifying and assessing ecological indicators of forest recovery on reclaimed oil and natural gas well pads. Ecological Indicators. 106, 105515. https://doi.org/10.1016/j.ecolind.2019.105515.

DOWNLOADS

Description of Resource Files:

For faculty:

  • IntrotoMultivariateAnalysisofWellPads.pptx: This MS Powerpoint file contains some sample slides that should be used by the instructor to introduce multivariate analysis and this particular dataset prior to working with the datasets in the student assignment.
  • WellvsRef.R: This is the .R script file for the faculty member to have as a resource to troubleshoot issues with students running the code. It includes all of the R code that will be used to conduct the statistical analysis by students and notes to guide the instructor.
  • StudentAssignmentWorksheetSolutions.docx: This is a set of sample solutions for the worksheet that the students will use in conjunction with the primer and dataset files in the lab. This is what you could use for grading the assignment and assessing the learning outcomes.
  • DemoAdvancedAssignment.docx: This is the assignment that I use in my advanced biological analysis course for students to build on – this could be a next step activity for a semester-long project for students in the class using other datasets after they have done a demonstration of the analyses in this activity. I have extracted elements of this assignment to be the worksheet and activities for the current dataset submission.
  • Lupardus_etal_EcologicalIndicators.pdf: This is the publication that the data are taken from and provides more detailed information on the project and results. Note that it goes into additional types of analyses compared with what the students will do with the dataset.

For students:

  • Primarydataset.csv: This is the .csv file with the plant community data that will be imported into R. This is the same file as the instructor will use.
  • Seconddataset.csv: This is the .csv file with the additional environmental attribute data that were collected for each sample unit that will be imported into R. This is the same file as the instructor will use.
  • StudentAssignmentWorksheet.docx: This is a worksheet that the students will use in conjunction with the primer and dataset files in the lab.
  • WellvsRef_RPrimer.docx: This is the R primer for this dataset. Students will use it to walk through the steps and conduct the analyses of the demonstration well pad dataset.
  • WellvsRefMetadata.docx: This is the metadata file that describes background information about the project and sampling and detailed entity-level metadata for both primarydataset.csv and seconddataset.csv. It includes supplementary level of detail compared to the background information in StudentAssignmentWorksheet.docx and instructor may opt to include this.

ACKNOWLEDGMENTS

The idea for publishing this dataset originated during a University of Alberta Augustana Faculty STEP research assistantship that funded undergraduate student Michelle Rude in Summer 2018. This dataset was also used to develop a project for students in the initial offering of AUBIO 315: Advanced Biological Analysis. Development of this dataset was continued by summer research assistantship funding to Chantal Ricard from the University of Alberta Augustana Campus in Summer 2019. We are grateful to the funding sources who funded collection of the well pad data: This ecological recovery monitoring project was initiated and funded by Alberta Environment and Sustainable Resource Development's Land Monitoring Team (AESRD 2012-2015), the Alberta Environmental Monitoring Evaluation and Reporting Agency (AEMERA 2015-16), and Environmental Monitoring & Science Division, Alberta Environment and Parks (AEP 2016-21), including Environment and Parks grant funding to A. McIntosh. The ecological recovery monitoring project that collected these data has also been supported by the Alberta Biodiversity Monitoring Institute Application Centre and Alberta Innovates - Technology Futures (AITF - now Innotech Alberta). We also thank the field research assistants who collected this data including: Lee-Ann Bauman (Nelson), Victor Bachmann, Elise Martin, Andrew Underwood, Carissa Wasyliw, and Scott Wilson.

CITATION

Michelle Rude, Chantal Ricard, Randi Lupardus, Anne C.S. McIntosh. June 2022, posting date. Arrested succession? Assessing plant community recovery on reclaimed oil and natural gas well pads in Alberta's boreal forests using multivariate analyses. Teaching Issues and Experiments in Ecology, Vol. 18: Practice #1 [online]. https://tiee.esa.org/vol/v18/issues/data_sets/mcintosh/abstract.html

[ TOP ]