Serotiny, undergrads, and generality in fire ecology

Cape Conran, August 2012
(Photo: John Morgan)

I've just spent a delightful four days in coastal vegetation at Cape Conran (about 5 hrs east of Melbourne) with my undergraduate ecology students. Here, there are warm temperate rainforests in deep gullies, tall forests on very old sand dunes, and coastal heaths fringing the margin of the coast. All are subject to fire, but have very different regimes. And fire is being used as a management tool to maintain (perhaps even maximise) plant diversity. This provides an excellent opportunity to ask questions about how plants and plant communities respond to fire.


I find teaching in the field, in small groups, a really satisfying way to train budding plant ecologists. I'm of the opinion that most undergrads learn through their eyes, and by getting their hands dirty! So, last week, I split the students into groups of six, and posed each a research question related to fire ecology. The aim was to make some observations that inform a question, make some predictions, design a study, collect data, collate and interpret that data, and make some statements about what had been found. This is something many of us do every day, but clearly it has to be taught (and learnt!) by students new to the field.

Happy students in recently burned sandplain heath
(Photo: James Shannon)

Surprise, surprise, the students quickly found that ecology isn't as easy as they had thought it might be! And interestingly, the first explanation from an undergrad is "we must have done something wrong" or "because we are undergrads, we are not very good". Nothing could be further from the truth. Undergrads and post-grads often collect meticulous data (down to three decimals points) when the bucket and spade approach would have been fine. Rather than this being an issue about the veracity of the data, it hints about the complexity of nature and our ability to accurately predict/forecast that complexity.

One of the biggest challenges our discipline faces is to develop generalities that might apply across taxa, communities, ecosystems. We've all conducted research where the results are contingent on the study system under examination. Or where opposing results are found across sites due to unknown site history factors. This might be something we accept as the 'norm' and keeps us endlessly asking questions about the natural world. But, it is a major challenge for students trying to understand the contribution they might make to the field, and to managers needing advice about the conservation management and restoration concerns that occupy them daily.

As an example of the challenges faced by my undergrads, one project looked at serotiny in the genus Banksia (Proteaceae) with an eye to thinking about the implications for fire management. Serotiny is the term used for seeds that are stored in fruits for extended periods of time (usually years) and then released spontaneously by an event. Triggers for this release include:

• Wetting (hygriscence)
• Warming by the sun (soliscence)
• Drying atmospheric conditions (xeriscence)
• Fire (pyriscence)
• Fire followed by wetting (pyrohydriscence)

Fire is the most common and best studied case of seed release in Banksia and, at Cape Conran, is  likely the trigger for seed release from follicles in three species that exist within close proximity (B. marginata, B. spinulosa, B. serrata). I challenged my students to design an experiment to test the idea that Banksia species have a similar responses to fire with respect to seed release.



This is what happens when you heat Banksia
infructescences at 500 deg C for 2 mins.
From top: B. serrata, B. spinulosa, B. marginata.
(Photo: Susan Hoebee)

Using a muffle furnace, they did a terrific job of teasing out the effects of duration of heating vs. maximumum temperature of heating. What was clear from their impressive data on follicle opening and seed viability was that each species responded in an individualistic manner. Banksia serrata folicles remain 100% closed when heated at 150 deg C for 10 mins while 80% of B. marginata follicles opened under this regime. B. spinulosa was somewhere in between. Hence, while these Banksia are all serotinous, they seem to differ in the type of fire event that will be necessary to ensure follicle opening and hence, seedling recruitment. The implications of this work are clear: three Banksia species at Cape Conran will respond to fire in subtley different ways and hence, we can't assume that one fire type will benefit/disadvantage these species in the same way. Perhaps this is the generality that we can draw from this study?


To me, one way to bridge this 'uncertainty' about data and, why it varies, is to emphasise and adopt proper sampling design in all our studies. This is not a new idea but rather, a call to remember what constitutes sound, evidence-based research. This may reduce some of the uncertainty that is generated by conflicting ecological studies - are the differences due to poor experimental design or real population differences?  I recommend you read the essay by Downes (2010) to brush up on sampling design and its importance in ecology, as well as how to deal with questions that aren't always amenable to experimentation.

Downes, B. (2010) Back to the future: little-used tools and principles of scientific inference can help disentangle effects of multiple stressors on freshwater ecosystems. Freswater Biology 55, Supplement s1, 60-79.