By Alan Jones
As an idealistic PhD student, I was once told that ecologists are constrained by the limits of our techniques. Within the vagaries of nature, every experiment we undertake is essentially flawed and the best we can do is to try to work within this. Nevertheless, a tendency exists for some experimental outcomes to be presented as definitive, even by scientists themselves. Are we always mindful of these limitations, particularly when interpreting field manipulation studies?
Plant science has used field manipulation techniques for over 150 years. In fact the birth of plant science as we know it probably began at Rothamstead in the 1850s, with the original fertiliser and crop rotation experiments that continue to this day. Modern challenges mean ecologists are now tasked with predicting the responses of plant communities to a range of Global Change issues, including climate change – and controlled field manipulation of select environmental factors remains a core technique in this research.
Any review of papers from manipulation experiments on Global Change will reveal a typical narrative. Initially, dramatic responses are reported to experimental treatments, but these eventually fade and can become subsumed over time. Various ecophysiological mechanisms are usually invoked to explain this, but it may simply be that uncontrolled environmental variables overwhelm those few experimental factors, as time passes.
In nature, plant communities are influenced by dozens of long-term environmental drivers. Plants, fauna, climate, soils and hydrology interact mutually in uncertain and potentially chaotic ways. At its most basic level, climate warming for example, occurs with precipitation changes, but it may also affect insect numbers and alter rates of herbivory. Field manipulation experiments attempt to address this using combined treatments. However, even a relatively complex experiment might only involve warming, precipitation and nitrogen addition treatments, so at best, three abiotic factors might be all that is tested. Nature is never this tidy. So how can simplified short-term field manipulation experiments ever be realistic?
The simplified experimental approach is useful, because it delivers clear-cut outcomes, usually within a short timeframe. But evidence suggests that real world conditions often cancel out the effects of single drivers. A tendency exists therefore, for Global Change experiments to generally overestimate the impacts of their few chosen factors. This bias is further amplified by the short duration and limited scale of field manipulation experiments. Global Change effects involve tens of environmental drivers, which will affect plant communities in complex ways for several decades. In the real world, multiple influences vary simultaneously across a large space, so overall effects are damped. If outcomes from field manipulation studies are ever proclaimed as definitive, it is likely that they fall far short of this. So, are these experiments that test only a few select factors, essentially meaningless?
Like all ecological approaches, manipulation studies are limited in what they can tell us. It remains impractical however, to try to manipulate all potential conditions in these experiments. Besides the logistical constraints, our imaginations could never entertain the actual complexity of these. Computer controlled and enclosed growth cabinets were once used in attempts to create simplified and idealised experimental conditions, but these methods are now considered obsolete. Meaningful ecological outcomes, it seems, only exist when the random whimsy of their natural environment is also allowed to play a part. Field manipulation experiments have a strength therefore, because this natural variation provides the principal threshold over which any experimental responses must first cross.
Ecological processes remain difficult to predict because we can only ever understand a small part of the system. The challenge of Global Change is that we need to understand how the entire mechanism works. The pragmatic point once put to me is a useful one. We need to acknowledge that these efforts are deeply flawed, but sometimes they are all we have.
Dr Alan G. Jones is a NERC-funded postdoctoral research associate based at Aberystwyth University. The views represented above are the author’s own.
Follow Alan on twitter: @alanjones_eCO2
Academic profile: http://www.aber.ac.uk/en/ibers/staff/staff_profiles/agj5/
Leuzinger, S., Luo, Y., Beier, C., Dieleman, W., Vicca, S. and Koerner, C. (2011) ‘Do global change experiments overestimate impacts on terrestrial ecosystems?’ Trends in Ecology & Evolution, 26 (5)