First Flowerings 2020 – interim analysis

Is this really such an exceptional year?

We’ve been recording first flowering dates (FFDs) for ten weeks now, and with lockdown beginning to be eased—and with me taking a one-week sabbatical over the Whit half-term week—this seems like as good a moment as any to look back and take stock.

I think we all feel that this year’s march of spring has, so far, been an exceptionally rapid and early one. But is this perception borne out by the data we’ve been collecting?

In Somerset we have two datasets against which we can usefully compare this year’s FFDs.

• The first is a Table of ‘average flowering times’ produced by eminent local botanist Walter Watson in the first half of the last century. This includes both FFDs and flowering period. So, for example, the average FFD for Bluebell, Hyacinthoides non-scripta, he calculated to be 12th April, while the flowering period was noted as usually being April-May, but with occasional early starts in March and late finishes in June. His Table, which runs to eighteen-and-a-half pages and more than 800 species, was published in the Proceedings of the Somerset Archaeological and Natural History Society (SANHS) for 1947. To anyone even vaguely interested in flowering times, Watson’s ‘Big Table’ is like gold dust. His dates were, for the most part, based on at least ten years’ observations. Much of his fieldwork was carried out in the Taunton area.

• The second dataset, modest in comparison, is from my own survey—again based mainly on observations in the Taunton area—beginning in 2008. My initial aim was to emulate Watson’s earlier survey by recording at least a decade’s-worth of FFDs for each species. By 2017 this had been achieved for 339 species, along with a further 216 species more patchily recorded. 2008-17 average FFDs for the first group—the 339—were calculated, and similarities and differences between these and Watson’s dates were summarised in the SANHS Proceedings for 2019. For all species combined, FFDs for the decade 2008-17 were found to be, on average, about two weeks earlier than in Watson’s day.

So, how do FFDs for the spring of 2020 compare with Watson’s dates, and with my own 2008-17 average FFDs? It would be astonishing if this year’s FFDs weren’t early at least by Watson’s standards, but are they really much different to those of other, more recent, springs?

  1. Comparing our ‘combined’ FFDs with Watson’s dates

First, let’s look at this year’s SRPG records against Watson’s dates. Of the 339 target species, we have already recorded FFDs for 266 of them. (Note this includes my own dates for species coming into flower between the start of January and mid-March, before our lockdown project began.) In Fig. 1 each point is a species, and its location is determined by Watson’s date (along the x axis) and our date (up the y). The numbers along each axis refer to Day Number, with Day 1 being 1st Jan., Day 32 being 1st Feb., etc, etc. The diagonal red line marks the line along which the data-points would lie if this year’s FFDs were identical to those recorded by Watson. Above the line, the 2020 FFD is later than Watson’s date, below the line is earlier.

I’m no statistician—never have been, never will be—but I think any sane person would be forced to conclude from Fig. 1 that FFDs in 2020 have (so far) been exceptionally early in comparison with Watson’s dates almost a century ago; indeed, many of our dates are more than a month earlier. Only two species, Blackgrass, Alopecurus myosuroides, and Gooseberry, Ribes uva-crispa, had later dates than Watson’s.

  1. Comparing our ‘combined’ FFDs with the 2008-17 average FFDs for the Taunton area

How do our dates compare with FFDs from my 2008-17 survey centred on the Taunton area? Fig. 2 is constructed in the same way as Fig. 1, but for each species the value along the x axis is the average FFD for the period 2008-17 rather than Watson’s date. The red line marks the line along which the data-points would lie if our FFDs this year were identical to the 2008-17 average FFDs – above the line is later than the decadal average, below the line is earlier.

There are, once again, very few species above the red line, although the gap between the mass of points below the red line and the red line itself is perhaps not quite as large as that in Fig. 1.

One reasonable objection to the above comparisons is that both Watson’s dates and my own 2008-17 FFDs are based on the observations of one pair of eyes operating across a fairly restricted geographical area, whereas our ‘combined’ first-flowering records for 2020 have had the benefit of many pairs of eyes and much wider geographical scope. One might expect the latter to produce an earlier crop of dates than the former. So, just to make sure, let’s make one more comparison…

  1. Comparing my own 2020 FFDs with 2008-17 average FFDs

Another graph, then, but this time comparing like with like: i.e. one person’s observations in 2020 against that same person’s observations between 2008 and 2017. And still you’ll notice that the data-points are, for the most part, below the line, indicating that my own FFDs this year are generally early in comparison with dates from the period 2008-2019—and this, despite the logistical and geographical constraints imposed by lockdown.

One further observation, and then I’ll shut up. As already noted, of the 339 species for which we have 2008-17 average FFDs, we have, between us, so far spotted 266 in flower. This spring, 94 of these—so more than one-third of the total—have had their earliest FFD on record, i.e. since 2008. Even within my own records, 55 species (so far) have had their ‘earliest ever’ FFDs. That’s a higher figure, at this stage in the year, than in any other year of the study. And yet the flip side to this statistic is also true, that for the vast majority of species the ‘earliest ever’ FFD wasn’t in 2020.

Many thanks to everyone for their records so far. It has been the strangest of springs, in ways both botanical and not-so-botanical, but it’s not over yet! We still have many species to search for, so it would be good to keep going a little longer if we can. Is it too fanciful to imagine that when we spot our first blossoming Ivy, Hedera helix—in the first week of September (ish)—we could be together, as a group, rather than each of us still isolated on our own patch?

Now there’s a thought.

With all good wishes for the coming week,