This was a piece I wrote for The Conversation in early May and thought I’d reproduce it here. Hope you find it interesting and thought-provoking.
It could be testament to our creativity and ingenuity that we can make an attempt at managing water in such a hugely complex ecosystem like the Murray-Darling Basin. But perhaps it’s also a demonstration of our arrogance that we actually think we can do it. Or perhaps it’s just sheer desperation that we have to do something, anything, to attempt to rectify all the poor decisions of the past. In the end, I reckon it is probably a combination of all three.
But one thing is very clear: as with most things that humans do when we attempt to predict or manage vastly complex systems, whatever happens, we will be flying by the seat of our pants. And we will almost certainly produce a less-than-perfect outcome. But that’s OK, despite the angst it might cause.
Why will we make mistakes and why is it OK to do so?
We cannot help make mistakes, because of the very complexity and size of the Murray-Darling Basin and because we don’t know how factors like climate and river flow influence water quality, individual animals and plants, populations, communities and ecosystem function (such as the movement of energy from the floodplain to the main channel of rivers).
Plans can be based on rough guesses, conceptual models or even mathematical models. But they are all built on data, and interpretation of the biological or ecological meaning of that data. Collecting, interpreting and forecasting data on rainfall and river flows, despite their complexity, is like a walk in the park compared to collecting, interpreting and forecasting data on even the simplest of ecological levels – the individual – let alone whole communities.
This is because the existence of particular physical conditions (like a flood) or chemical conditions (say, high salinity) will only tell you about what ecological outcomes are possible, not what actually will happen.
If we were only interested in bacteria or algae, perhaps, predictions could be made about outcomes based on specific physical and chemical conditions, because there is a close connection between inputs and outputs. But even with these relatively simple organisms, there is competition between individuals and species for space, nutrients and light, and so even there outcomes can be uncertain.
But when it comes to the complex behaviour of, say, fish, all bets are off. Sure, you can predict with a reasonable level of certainty that if you provide the right ranges of temperature and flow, a species will breed. But translating that simply to survival of the eggs is hard enough. Translating breeding to survival, growth, movement, the outcomes of competition and predation, maturation, attainment of breeding territories and successful mating, is nigh impossible.
There are approaches to help us realise our goals of obtaining improved outcomes. They include Adaptive Management and Bayesian Modelling. In both cases, there is inbuilt flexibility and potential to include improved understanding in subsequent iterations of predictions and on-ground actions. Progress will still be very slow, even with these approaches, because they still rely on data, and data collection on river ecosystems is costly and time-consuming. So, patience and persistence are important commodities with all of this.
Secondly, why is it OK to make mistakes in management? The answer to this relates to the first part of the discussion.
It is OK, because we can often learn more from mistakes than from successes. But, and it’s a big BUT, the only way to learn from mistakes is to set up the management decisions as experiments.
It is certainly a start to have good, sound, ecological bases – not to mention hypotheses – for giving more water back to the environment, such as flooding a wetland that has previously been dry. But it is absolutely critical to do this in a way that tells us something meaningful and allows us to improve – or rectify mistakes – next time.
This requires not only an appreciation of how experiments are done, but also intestinal fortitude on the part of natural resource management agencies involved. Intestinal fortitude is needed because a good experiment must include control or reference systems (that is, systems that are not manipulated). It must have a period prior to the intervention to gather baseline data to compare with after the intervention.
In most cases, this may mean not intervening with some systems because you need them to act as references, and delaying remedial action to collect “before” data, both of which can incur the wrath or at the very least, antipathy and scepticism of stakeholders. Most natural resource management agencies find this approach unpalatable, because:
- they understandably want to fix things as soon as possible, and
- discontented stakeholders typically blame politicians and politicians don’t want to be seen in a bad light.
Sure, making mistakes is not something we strive for, and clearly it can cause financial suffering as well as poor environmental outcomes. I am not advocating that we deliberately make poor decisions. What I am saying is that we should recognise our profound ignorance of how ecosystems, like the Murray-Darling Basin, function. We need to accept that we are largely flying by the seat of our pants when making decisions and invest heavily in targeted data-gathering to improve our models. But most importantly, when making decisions, we have to conduct large-scale, ecological experiments to learn from the mistakes that we do make. Not to do so, risks making the same mistakes over and over again.
To paraphrase Oscar Wilde: to make one mistake may be regarded as a misfortune, to make the same mistake twice looks like carelessness.
Previous articles in this series
Paul Humphries receives funding from the Murray-Darling Basin Authority