It was a cool but bright day in October, and I took a few steps back from my equipment to take a photograph. You can see from the picture here the variation in grass quality from the near to middle distance. The equipment had been fenced off during the cattle grazing period of the farm calendar. Ironic really, because if you look close at the photograph you can see that the grass inside the markers on the closest side is dark green. The grass here is enriched with cattle urine, that myself and my PhD supervisor had sprinkled down on the grass inside the markers with watering cans whilst bemused cattle looked on from over the electric fence.

The white boxes allowed me to measure the release of greenhouse gases from the field surface. Next to the boxes, some porous tubes were buried in the soil. By measuring the greenhouse gas concentrations inside those tubes I was able to determine some of the key biological production and physical transport processes and see how these changed when the soil had been subject to addition of the cattle urine. Along with measurements of soil chemistry, moisture and temperature, an overall picture could be developed of what heppened in that soil to create the observed greenhouse gases at the surface and how these things changed over time.

Controlled field experiments like these are crucial if we are to understand the effects that changing land use can have on the natural environment. Today, computers become ever more powerful. Processors and memory that once filled whole rooms can now fit into the palm of your hand. And scientists have capitalised on this new technology; something that was once a dream has become a feasible reality: modelling the Earth’s complex environmental system and making inferrences about how this system might change into the future. Now we move ever closer to being able to enter into a computer how many cattle we put on our field and being told the amount of fertilisation that will occur, a value for how much nitrogen might make it into nearby waterways and the effect that the cattle will have on the greenhouse gas balance of your piece of land.

This might sound highly ambitious, and it is. Yet these models continue to appear and are in continuous development and improvement; researchers validate these models using measurements from the field and look at improving understanding of processes and process interaction so that the amount of uncertainty in our estimates may be reduced (JULES land surface model, ECOSSE carbon model). Models are known by scientists to be a tool in development, a simplification of complex systems to allow us to search for missing information in the knowledge base and give us clues as to how a known process might change in the future as human activity or natural variation alters land use, biodiversity or climate. Issues arise when model data is misreported or uncertainty in the models is not adequately described. It’s important to always remember that models (particularly the largest and most complex) can have missing or inadequately described processes (Stephens and Bony, Science, 2013 (paytoview) ; Carbon Brief, 2013 (free commentary)). In order to ensure that models are telling us the right things, we need field studies (validation studies) and in order to know what to put in to the models in the first place, we need field studies (parameterisation studies). Models only tell us what they know about, which is what we input into them. A complex computer model can give us far more detail on complex processes than we would ever be able to get without them, but without field measurements to provide references and comparisons, we will never know if we’re going wrong.

In the next blog article, I’ll return to the technical articles and write a little more about city environments and their influence on the lower atmosphere. In the next insight article I will write about the importance of literature in a scientist’s life and the perils that are associated with under and over exposure to the enormous piles of reading that accumulate in the scientist’s in-tray.