Contingencies ....in time

Everything happens somewhere. And in nature, the somewhere matters.

One of the most obvious things about the discipline I work in is that the results of one experiment or observation are not always replicated elsewhere across the globe, the nation, or even the state that I live in. Ecologists often talk about the fact that their findings are "contingent" on the organisms being studied (for example, short-lived perennial herbs versus long-lived trees will probably have very different demographies), the setting of the study (e.g. alpine versus lowland grasslands probably cycle nutrients in different ways) and the history of the system under investigation (e.g. evolutionary history and recent landuse legacies will affect successional pathways after disturbance).

This can make it all very confusing to come up with simple, widely applicable and generalisable predictions about the way in which natural ecosystems respond to climate perturbations and the impacts of utilisation. I guess that is why there has been a rapid growth in the use of meta-analyses to tease out ecological principles from spatial contingency, as well as the growth in global collaborative networks. I, myself, am loosely affliated with three such networks - NutNet, ITEX and Shrub Hub - whose chief aim is to study questions pertinent to ecologists working in similar systems across the world, using standardised protocols, to see if ecological patterns emerge beyond the local-scale.

One of the things that is also true about nature is that different things happen in different years. As someone who studies seedling regeneration processes, I know this only too well. But I think we tend to under-estimate the importance of temporal contingencies in our ecological thinking.

Just how much does the timing of an experiment or observation cloud our view of key ecological processes? If we study seedling recruitment in a decade of drought, for example, are the outcomes of such a study sufficiently robust to propagate models that forecast extinction risk? The answer is probably no. But collecting data is a hard excercise. Ask any student at the completion of their doctoral thesis about whether they want to do it all again and you'll get a stern look no doubt.

The Button Wrinklewort in remnant native grassland, Victoria
(Photo: Donna Burns)

Returning to seedlings as a case in point. I've been following the seedling recruitment of the endangered grassland daisy Button Wrinklework for almost 20 yrs now. In 1991/2 (crikey, that makes me feel old), I observed a large seedling cohort in the two populations I have followed ever since. I even wrote up a paper on this (see here) and concluded that "....... seedling recruitment is not restricted to climatically favourable years (i.e. is not episodic) but rather, is potentially on-going provided suitable microsites are available for seedling survival."

How wrong was I! My long-term observations of recruitment in the Button Wrinklewort has revealed that there has NEVER been another recruitment year like 1991/2. This probably has a lot to do with the fact that south-eastern Australia experienced a 13-yr drought from 1996-2009 where rainfall was below-average in just about every year. Rather than observe on-going recruitment, seedlings have been entirely episodic. The number of adult plants have declined in some populations and not been replaced by seedlings (whereas I had suggested declines might occur where microsite limitation limits survival; while this is certainly true, you've got to have seedlings before you have this problem). Interestingly, the recruitment event I observed in 1991/2 was very successful. Many of the seedlings that emerged in that year went on to become reproductive. A couple are even still alive today! This is what Peter Chesson calls the "storage effect" - infrequent recruits are stored in the population and these buffer the population in the intervening years / decades when recruitment is far less successful.

This simple example highlights that temporal studies are crucial if we are to adequately understand the dynamic nature of species populations and their interactions with climate and other species. Of course, this is incredibly difficult to do. Permanent plots, data archiving, and good old fashioned persistence are the keys here. These three things are also terribly unfashionable in ecology today. We need answers now to the pressing conservation questions that face our society. But we also need to acknowledge that what data we do have will be imperfect and model outputs are not final nor definitive. And we need scientists (like me) to champion the need to conduct long-term, old fashioned studies. Otherwise we potentially face a future where decisions are made on data that is temporally contingent and this could have serious implications for the future of the species we are interested in preserving. Particularly in a world undergoing a changing climate. Something to ponder!