Institution/institution type: Augustana College (residential liberal arts college)
Course/Course format: General Ecology (Junior and senior biology and environmental studies majors, does not count towards general education requirements so it is a majors only course). The course had 36 enrolled students. Lab met once a week for three hours.
Module synopsis: In this module, students collect demographic data for a local plant population, and then use a stochastic model in R to look at the impact of changing survivorship and fecundity on the population.
Instructor notes:
What I hoped to accomplish with this module was to give students an opportunity to apply demographic models in a more real world setting, and see how those models can inform management decisions. As a secondary objective, I wanted to give students a chance to work with a stochastic model, rather than the deterministic models I have traditionally used in this class.
I introduced the module during a broader unit on population ecology. By the time students were working on this module, we had already covered survivorship curves and basic life table parameters in lecture (Molles 6th edition, chapter 10). Students had already worked simple, deterministic, problems in class, for example, determining which survivorship curve a population has given lx values, and given sx, mx, and nx, values predict nt+1. I had not used matrix-based models in this class before, so I also spent about a half hour at the beginning of lab working through the example on page 6 of the Charney and Record module.
I also used the cemetery demography module, which I have used before, in this same unit. As Charney and Record note, these two modules together serve as a good unit on populations.
The lab report was modified from the set of questions in the TIEE module. I used questions 3-6, and 9 as the basis for the lab report.
Unfortunately, my experience with the viability analysis module was mixed. The module states that no prior experience with R is needed, however, my lack of experience with R proved to be a barrier to effective implementation of this activity.
The students loved the fieldwork aspects of the lab, and collecting the data and getting students to enter the data on google docs was straightforward, with only the minor and expected difficulty of a few students miscoding things or entering redundant data. Further, the students noticed lots of variation in petal number in the population we were working on, and this got them thinking about the possible fitness impact of petal number as well as asking good ecological questions about this trait, for example, do large plants also have more petals? The module was very successful on that level.
However, I was not able to get the R script I was supposed to run to simulate next year’s data to run (PVA_instructor_script-singleyear.R). The sample data the module authors provided ran just fine, but I just could not get my file to work. More familiarity with R would have helped here, as it is likely that some small format issue was the problem.
The students were still excited about the project, though, so I came up with some simulated data for next year based on crude back-of-the–envelope calculations assuming that the current data represented a stable stage distribution. I practiced using the student R scripts in my office, and the simulations ran perfectly. I scheduled our next class meeting for one of the computer labs on campus, after making sure R was installed on these machines.
Unfortunately, during class, only two or three stations were able to run the student scripts successfully. Error messages popped up across the room, and without R experience, I was not able to troubleshoot the simulation. Sometimes if students just exited R and started over, the script worked, which suggested user error on our part, but mostly we would just hit the same errors repeatedly.
Wanting to salvage more from the experience, I created an excel spreadsheet that used the data we collected in a deterministic matrix model (Chapter 14 in Donovan and Welden Spreadsheet exercises in Ecology and Evolution, 2002 Sinauer and associates press), and I asked students to prepare a lab report using our data and entering different parameter values in this excel based model. I did not ask students to create the spreadsheet. Rather, I used a template, and write-protected the formulas, so students could only change the parameters I wanted them to manipulate. This worked reasonably well, but lacked the more realistic element of the stochastic model in the R based module. I was pleased with most of the lab reports, but a few students took a very cursory approach to the simulation (changing very few parameters in the model and changing them only a little). If I were to do this again, I would be more explicit about this aspect of the report.
The module does not give details about how to collect fecundity data, admittedly a big task that would have added another layer of complexity to the student directions, but this is a critical step if the module is to be successfully implemented across years. One issue I did not anticipate here is setting aside class time for seed collection/germination.
On the plus side, the variation in petal numbers became the basis for a good, engaging, student project. Further, the lab reports based on the excel-based, deterministic, model showed that students successfully got the connection between stage based survival, fecundity, and the viability of the population. While students were initially surprised that changing the fecundity of different stages by the same amount had different population impacts, most reasoned out the connection with survivorship and the potential conservation applications of this connection.
