Interpreting data - effects of the heatwave

In my last post on the effects of the heatwave, I demonstrated the short-term impact of the heatwave on the numbers of records and numbers of species recorded. To do this, I used the data assembled from social media; primarily Facebook, but also to a much smaller degree Flickr. The big advantage of this dataset is that people almost invariably post their observations within a day or two of the observation; thus giving as near to a 'real-time' picture of what is going on.

It is certainly true that a small number of records will arrive over the following weeks, but approximately 98% arrive within two days of the record data. Thus, any small difference between data secured two days after the event and two weeks after the event is around 2%. When you bear in mind that the numbers of records for each day can be in the range of 120 to 180 per day in mid-summer, any slight lag in receiving records is of comparatively little consequence. I should add that all data posted are extracted within 24 hours of the post, so there is a nominal lag there too. Furthermore, the data are aggregated into weekly blocks, so there would need to be a huge influx of data for any one week to make a significant change to the shape of the graphs.

In the case of the graphs I presented on 16 August, the dip in record numbers between weeks 25 and 26 is 191 records (611 down to 420) (31.2%)  and the dip in the numbers of species recorded at the same time is 21 (75 down to 54) or 28%. Thus, there would have to have been a huge volume of records withheld for those dates but no records withheld for other dates. That seems unlikely and, indeed it is clearly not the case as the graphs look much the same now using updated data (figures 1 & 2).

Thus, I think we can be fairly safe in saying that there was a significant drop in both the abundance and diversity of hoverflies recorded during the most influential part of the drought. We can also see that as the season has progressed, the numbers of records arriving has revived to approximately the same levels as might be expected from the average of the preceding 3 years.

Figure 1. Weekly numbers of records in 2018 compared with the avaerage for the preceding 3 years.

Figure 2. Numbers of species recorded each week in 2018 compared with the average for the preceding 3 years.