Friday, 14 October 2016

Yearly phenology signatures

A couple of days ago, I investigated the overall yearly data 'signatures' generated by records from photographers. I only went back over 3 years because it was not until 2014 that activity was sufficiently strong in the early and later parts of the year to present a comprehensive picture of hoverfly activity.  Prior to this, most of the data collected was opportunistic - gathered from various web sources such as iSpot and Flickr and the late-lamented Wild About Britain.

The emergence of the UK Hoverflies Facebook group has triggered much more detailed recording. There is now a group of 20-30 people who record on an almost daily basis. This sort of recording makes it possible to look at hoverfly abundance in a completely different way.

I took as a starting point the raw data for each day, plotted as a simple chart with a 30-day centred running mean superimposed. This approach has the obvious drawback that the dataset is bigger than the last one in each successive year (Figure 1). A second stage was needed to smooth out year on year differences in data volume - which I dealt with by turning daily plots into a proportion of the total records for that year. In this second stage I have simply presented the 30-day centred sunning mean (Figure 2).

So, what does this tell us?

In both 2014 and 2015 there was a steep rise in spring hoverfly activity that peaked in April before dropping off into May. This pattern looks to be consistent with recent years when a cold snap in late April and May has badly affected insect activity. In 2016 this pattern was not as obvious: activity started much earlier in the year, indeed it barely stopped at all throughout the winter. Numbers rose rather more gradually and the first signs of a stutter were in mid-May.

The peaks for hoverfly activity seem to have been in mid- to late-July in both 2014 and 2015, whereas there were two peaks in 2016. The first was in late June. A brief spell of fiercely hot weather in July clearly suppressed hoverfly numbers and many FB members commented on the apparent absence of hoverflies all over the country. A second peak followed in September, which was one of the warmest in several decades.

The autumn fall in hoverfly abundance also differs: in 2014 it was fairly precipitous and by the end of October there were very few incoming records. 2015 was far warmer and the decline was much more gradual.

The 3 year signatures also show how in 2015 and 2016 the season of high hoverfly activity was much broader, illustrating the effects of the warm winter of 2015-2016. Obviously it is too soon to comment in any detail about the Autumn of 2016 but the numbers of incoming records are still high.

Figure 1. Daily records of hoverflies based on the photographic dataset with a 30-day centred running mean

Figure 2. 30-day centred running mean of records presented as a % of the yearly totals

No comments:

Post a Comment