I’ve had a quick play with some data – mostly pulled from the FAO and DEFRA – in an attempt to explore the relationship between agricultural ‘intensification’ and farmland bird declines (in the UK). This brief bit of procrastination (whilst half-listening to the Commons grouse shooting debate) was inspired by this paper, which I really liked, but which is now almost 20 years out of date.
Agricultural intensification entails a suite of changes in the way that farmland is managed. In my mind, it covers things like farm specialisation, simplified crop rotations, loss of ‘semi-natural’ habitats (hedgerows, field margins, fallows, stubble) and an increase in the use of agro-chemicals. Most of this can be summarised as a loss of ‘structural heterogeneity’ (in both time and space).
Panel a, below, shows the aggregated population index of 12 ‘specialist’ farmland birds (corn bunting, goldfinch, grey partridge, lapwing, linnet, skylark, starling, stock dove, tree sparrow, turtle dove, whitethroat & yellowhammer), which has more-or-less halved in the last 40 years. This decline started in the late 1970s, and was shockingly rapid during the 1980s, but has since slowed.
What’s happened to agriculture over this 40 year period? In the absence of freely and immediately available agricultural data (covering 1970-present), I’ve summarised 4 key variables: area of permanent pasture, area of oilseed rape, wheat yield and nitrogen use. High values of my index correlate with more OSR, higher wheat yields and less pasture. The intensification index increased pretty linearly through the 1970s, with the rate of increase slowing down (but not stopping) in the late 1980s (panel b, below).
How do these two indices match up (panel c)?
- There’s a negative association between the intensification index and the farmland bird index. Correlation ≠ causation though, and looking for associations in time-series data is particularly dangerous – lots of things co-vary with ‘Year’. But I think it’s safe to say that there’s something concrete behind this correlation – lots and lots of research has basically confirmed this.
- As the intensification index increased in the 1970s (purple), the farmland bird index didn’t decrease immediately – a lag effect?
- As the intensification index continued to increase, the bird index started to down-turn; in the middle of the graph, the association is pretty striking. It’s worth noting that the correlation between the two indices is (slightly) stronger than the correlation between either index and ‘Year’.
- As the increase in the intensification index slowed (greens and yellows), so too did the decrease in the bird index.
I’m not sure what my conclusions from this are – I’d definitely like to delve into this analysis in more detail – particularly with some additional data on pesticide use, crop diversity, etc.
I think it’s clear that this quote from the NFU’s Vice President isn’t quite true (though obviously ‘agricultural intensification’ means different things to different people):
logically, you cannot attribute biodiversity loss in the past quarter of a century to agricultural intensification, because there hasn’t been any
But, to be fair, the ‘nitrogen use’ variable didn’t correlate with the other variables (or with year), and the wheat yield variable has levelled off in recent years (so the continued increase in my intensification index is driven by OSR area increasing and permanent pasture area decreasing).
I guess the point I want to make is that intensification (or whatever we want to call it) does *not* need to be sustained for farmland bird declines to continue. It can take a while for populations to ‘catch up’ with environmental change. Agricultural intensification probably isn’t killing birds directly; but if it slightly diminishes their breeding productivity (or winter survival) then population size will slowly decrease, even in the absence of continued environmental change.
I’ve done a couple more graphs, exploring the relationship between my crude index of agricultural intensification, and some other bird population indices.
Panel a shows the population index of generalist farmland birds (greenfinch, jackdaw, kestrel, reed bunting, rook, woodpigeon and yellow wagtail). The association is much weaker for these species, as might be expected given that they don’t depend entirely on farmland habitat (unlike the specialists, above).
Panel b is for woodland specialists (22 species in total). The association here is a bit concerning – it’s hard to think of a direct causal link between, say, wheat yields and willow tits. Instead, this looks like a temporal confound – woodland birds show a declining trend, intensification shows an increasing trend, so the two indices correlate. So, based on this crude analysis alone, we can’t be sure that the negative association in c (above) is not also driven by a confound like this (though the association between intensification and farmland specialists is more convincing).