In ecology Big data, like the Prime Minister’s much derided ‘Big society’, is neither the whole story nor the whole answer…
Ecology has a data collection obsession; I could even be persuaded to say addiction. Perhaps that is unfair. All research is about data collection and analysis, but today, ecology is a bit like the cat that got the cream. Study upon study involves monitoring data from different tracking technologies. As the size of trackers increases, the technology costs reduce, and we get better at working out complicated behavioral patterns from accelerometer data or salinity recordings and acoustic tags, there has been a monitoring frenzy1.
The vast quantities of data now involved in ecological research, and the impact this has on methods, is reflected in the skill set now required to be a researcher in the field. There is a huge focus on techniques for managing this data. The Natural Environment Research Council’s (NERC) most wanted skills list includes modeling, data management and numeracy and not fieldwork(!) in its top five. Despite this apparent surfeit of data, ecologists and conservation biologists still seem to spend a lot of time talking about how little we are absolutely sure of, and all the things we don’t know2.
Some context – what are we actually talking about?
According to the Convention on Biological Diversity, biodiversity is, “the variability among living organisms from all sources… including diversity within species, between species and of ecosystems.”
We are thought to be experiencing a global trend of biodiversity loss3, a fact that has been seen as important even by the national newspapers here, here and here. In 2002 governments across the world committed to achieve “a significant reduction in biodiversity loss by 2010”. This target was not met, but it kick started efforts to design mechanisms for measuring biodiversity, or monitoring it – a thing that had not really been done before4.
Monitoring is about gathering information on how the variables in a system work together over time, in order to infer something about how the whole system is working5. For example, using something easy to measure like blood pressure or body temperature as signposts for human health. Practically it extends from counting numbers of species in an ecosystem, and counting individual members of a species, all the way to applying tracking technology to different animals to see where they go and recording the movements they make.
Ok… But why do we do it?
The data from these different monitoring experiments can then be used to say something about the organism of interest, or the system it lives, in or both. Such as, is it at risk of extinction or experiencing a population explosion? Does its movement patterns show anything out of the ordinary? And from this can we infer anything about the overall state of the system? Is it functioning well, is it on the verge of collapse, or somewhere in between? Ultimately we monitor to understand more about the world; logically it follows that if we understand more about the world then we can look after it better.
There are two main drivers for monitoring and it is important to remember that they are subtly different; the first is knowledge focused and has an academic bias. The second is action focused and has a conservation bias. While it might seem that you get the answer to the action part of the problem by addressing the knowledge part of the problem and vice versa this isn’t always the case5.
Well that sounds sensible to me, what is the problem?
To be blunt, the problem is monitoring for its own sake. Evidence from monitoring is really important, and can tell us lots of things that are both interesting and important for the future state of the world, but only if it has been thought about properly and has clear goals5. Which brings me to my political analogy, the big society sounded like a great idea, everyone works together to help each other and make the world a better place, but that is all, it was a sound-bite. In reality, it had no clear ambitions and no plans for delivery or evaluation – it has failed to motivate anyone to do anything or effect any change, while making the government feel as though something was in fact being achieved.
The goal of monitoring should not be to go around spending money on attaching an electronic tag to every animal on the planet to see what happens. This is data-mining, it is not hypothesis driven research. Monitoring programmes need clear objectives and thought-through experimental designs that achieve these ambitions. Without critical consideration of the true objective and how it will be achieved, monitoring does not add value. If monitoring becomes the default position, universally accepted as the way to gather information or react to environmental change, then to some extent it is just a distraction, and an expensive one at that.
Equipment is getting cheaper but it will never be free. Conservation and research will always have budgetary constraints. Simply collecting vast amounts of data is not the point, we do not want an exact replica of the world that mirrors the real one, so we can say: ‘oh look how good we are at building mini-worlds.’ The point is to understand the data we have so that we can add power to our predictions. For example can we use bird data to tell us about all terrestrial mammals? If not – and I agree its unlikely – do other mammals behave differently in reliable ways?
There is a long established scientific history of wanting to be exactly sure about everything we have. But you can’t “stock take” the planet. If we wait to classify and count everything before we can make decisions nothing will ever be done. We will never be able to know everything. Like climate change modeling, the point isn’t to get bogged down in the detail; it is about having enough information to make a good decision.
Find me on twitter @theurban_fairy
- Legg, C. J. & Nagy, L. Why most conservation monitoring is, but need not be, a waste of time. J. Environ. Manage. 78, 194–9 (2006).
- Nicholson, E. et al. Making robust policy decisions using global biodiversity indicators. PLoS One 7, e41128 (2012).
- Dirzo, R. et al. Defaunation in the Anthropocene. Science 345, 401–6 (2014).
- Butchart, S. H. M. et al. Global biodiversity: indicators of recent declines. Science 328, 1164–8 (2010).
- Yoccoz, N. G., Nichols, J. D. & Boulinier, T. Monitoring of biological diversity in space and time. Trends Ecol. Evol. 16, 446–453 (2001).