Thursday, 23 November 2017

Difficult taxa - should we believe the trends?

In my previous posts I have hopefully shown some of the challenges that we face when trying to interpret trends. Simply taking statistical outputs and accepting the results is generally unwise. One needs to understand what is happening within the data. As we have progressed with the revised atlas, I have therefore gone back to Stuart with a series of questions. The Frescalo outputs tell an important story but, as we have seen already, there are factors that may be having a profound effect on the results. Changing recorder methods are one factor that we can detect by separating off records derived from photography but here is one that confounds the basic analysis.

In my experience, Pipiza noctiluca is a very common species. It does visit flowers, but is far more commonly seen sun-basking on leaves - I favour elm, sycamore and lime for monitoring. On the right day in my local woods I can see maybe a dozen or more, mostly males. I usually retain a broad sample of specimens because they are so similar that it is quite possible to get two or even three species in the mix. It was in this way that I picked up the first British record of Pipiza quadrimaculata this spring (still to be written up and published). Taking a broad spectrum of Pipizines often adds species within Heringia and Pipizella as well. I find many of them to be pretty common, but often have a lot of difficulty being sure that I have actually got P. notata (bimaculata) and not another form of P. noctiluca!

So, what do the outputs tell us about Pipiza noctiluca? It all looks pretty clear (Figure 1)  - there has been a significant decline! But has there? Figure 2 where the same Frescalo analysis has been run on the whole dataset and then with photographic data excluded changes matters, as might be expected (Figure 2).

Figure 1. Trend for Pipiza noctiluca based on Frescalo analysis of all data.
Figure 2. Trend for Pipiza noctiluca based on Frescalo analysis for all data (blue) and with photographic data excluded (red).
The problem I find with both of these outputs is that one can roughly eye in a regression line but it is clear that there have been a number of very marked peaks. Unlike data for wetland species that has a spike associated with the major NCC surveys of Welsh Peatlands and East Anglian Fens in the late 1980s, the peaks here are a bit different. Is there something happening with recorder activity and perhaps changes in who is active and where they are active?

Figure 3 shows the current state of our knowledge of distribution of Pipiza noctiluca. It tells an interesting story. For example, in the 1980s and 1990s there was tremendous activity by a small number of very active recorders who covered all taxa. Dorset and Somerset were extremely well covered, as was Surrey and so too was the Sorby area. Several of those really active recorders have died or become too unwell to continue recording; others, such as me, have moved. There is also one recorder who is still very active, but whose records we don't have for the past 20 years (although this may be rectified soon).
Figure 3. Current distribution map for Pipiza noctiluca. Black symbols indicate the latest record dates from 2000 to date, grey - 1980 to 1999 and white - before 1980
In this instance, if we really want to understand what is happening, we probably need to look at the peaks and how they relate to activity by individual recorders; or whether in fact there is any correlation? Peaks do seem to be somewhat repetitive, so perhaps there is another environmental factor at work? Alternatively, perhaps we are just seeing bursts of enthusiasm followed by the inevitable misery of struggling to be sure one has correctly identified yet another form of P. noctiluca!

No comments:

Post a Comment