Interpreting phenology histograms

Since the great changes brought about by the UK Hoverflies Facebook page I have had growing concerns about a small part of the HRS dataset. Most of it is probably fine, but as time goes by it is becoming increasingly clear that many more species than we thought hitherto have relatively short emergence periods. There can be regional variation which means that for species that occur widely, but not at high densities, the aggregated phenology histograms are misleading. In many cases they show a very elongate tail into the Autumn. I don’t think this interpretation is correct, although there is always the possibility of occasional aberrations.

I’ll use four species as examples. Two are widespread and regularly recorded both by traditional net and microscope recorders and by photographers; the third is a recent arrival but is almost entirely recorded by photographers and the final one is a very early spring species. The important separation is that since we have had the Facebook group I have done the bulk of the determinations that have found their way onto the dataset (until this year). In that time, all three species have been found to have very tight emergence patterns.

Epistrophe nitidicollis generally occurs in May and June, perhaps into July. It is quite infrequently seen and I suspect is regularly reported as a misidentification of a Syrphus. I am pretty sure the tail on the graph (Figure 1) arises from such misidentifications.

Figure 1. Phenology of Epistrophe nitidicollis

Epistrophe melanostoma arrived in the UK in 1986 or thereabouts and has since become established in south-east England. It is very similar to E. nitidicollis and this is one reason why it may not have been recognised prior to 1990 when a Dutch specialist (Paul Beuk) was living in the UK and found it on my local patch. Remarkably, we get most records from photographs and, as can be seen from Figure 2, there is no extended tail of records. More importantly, whilst I think there is growing evidence that E. nitidicollis is declining in abundance, E. melanostoma is becoming far commoner.

Figure 2. Phenology of Epistrophe melanostoma

Platycheirus tarsalis is another spring species. It is quite difficult to separate from P. manicatus and I suspect that there have been numerous mis-identifications in the past. Unlike P. tarsalis, P. manicatus has at least two broods and continues well into the Autumn. I think this probably explains the tail in the graph for P. tarsalis (Figure 3), although there may also be confusion with female P. peltatus amongst recorders who don’t use keys.

Figure 3. Phenology of Platycheirus tarsalis

Melangyna quadrimaculata flies from late February to April, perhaps into May in Scotland but I have never seen it there despite many trips in late May and early June. Nevertheless, there is a tail in the graph (Figure 4), which I think can be explained by misidentification of very dark male Leucozona laternaria; I have seen this mistake made on several occasions and suspect that in my novice days I too might have made this mistake!

Figure 4. Phenology of Melangyna quadrimaculata

How do we resolve this issue? I am at a loss, because there will probably be some genuine records in the tails. However, there are almost certainly a lot of misidentifications. I suspect the best we can do is to provide a written interpretation to accompany graphs but there remains the danger that the graphs will be mis-interpreted, and the data glitches will be reinforced by further misidentifications. Alternatively, should we clean up the graphs and adopt a cut-off date after which we don’t accept records? That runs another set of risks because we might then be accused of massaging the data to fit our interpretation.

We are faced with a conundrum! I think we need to develop a far more comprehensive analysis of the phenology of those species that are regularly recorded. I suspect it will prove to be illuminating!

All at sea!

We know that some hoverflies cover considerable distances over land and sea. Moreover, there is some evidence of northward movement through mountain passes in the spring and southward movement in the autumn. In the UK, we occasionally get vast influxes of hoverflies but in most years there is probably more of a gradual trickle. It is an area of research that offers all sorts of questions and perhaps opportunities.

One of the regular contributors to the UK Hoverflies Facebook group works on the Floating Production Storage and Offtake (FPSO) vessel Curlew, which lies 130 miles East of Montrose. From time-to-time, he reports sightings at this offshore location. The hoverflies involved must be part of a larger 'migration' but the big question is 'where are they coming from?' Many will doubtless be continental in origin, but perhaps some also come from the UK and are heading north-eastwards?

These conundrums got me thinking: how big a proportion of the records gathered for hoverflies in July and August are 'home-grown' and how many are continental visitors? I doubt we can say with any certainty, but perhaps we can do some basic investigation by packaging together the records for the main migratory species (Episyrphus balteatus, Eupeodes corollae, Helophilus trivittatus and Scaeva pyrastri) and see how much of a contribution they make to the data (Figures 1 & 2).

Figure 1. Total records for 2017 extracted from UK Hoverflies Facebook page broken down to include the contributions of four known migratory species. Note that Syrphus ribesii, Syrphus vitripennis and Sphaerophoria scripta are also regular migrants but there are also a substantial resident populations.

Figure 2. Records for 2017 data in Figure 1 with migratory species excluded.

I think Figure 2 demonstrates how a substantial dip in records occurred in 2017 if migratory species are excluded. Then, there was nothing like the heatwave that we have seen in 2018, but even so there seems to have been a pronounced drop in numbers. It will be interesting to see the differences for 2018 at the end of the season.

Oh such wonderful weather - or is it?

The press are loving the recent hot spell, as are many of the public. Memories flood back to the ‘wonderful’ summer of 1976, and I expect that in due course another generation will look back to the summer of 2018 as one of those ‘exquisite’ summers. As a biologist I sit and fret about the heat and its effects on our insect fauna.

The 1976 drought had a devastating effect on a lot of insect populations and some, such as the craneflies, probably have never fully recovered. I’m pretty convinced that this event and other exceptional droughts, such as the one in the early 1990s and again in the early 2000s, have been a major factor in the catastrophic decline in our insect life and in corresponding declines in many bird species. Most insects actually dislike the heat! Why might this be so?

We need to think of adult insects as the dispersal stage, but for most of the time insects are either eggs, larvae or pupae/puparia. Most of these stages are poorly adapted to excess heat and to drought. Those in the UK, where the climate is normally mild and comparatively wet, are especially poorly adapted to periods of prolonged drought. They occur in the UK because it is mild and wet. If our fauna was adapted to hot and dry, we would have one more akin to the Mediterranean!

Our fauna IS changing. We have seen the arrival of numerous species from Europe in recent years, and some species at the edge of range have expanded their range markedly. Last night I watched a colony of perhaps 200 burrows of the bee-wolf Philanthus triangulum in Mitcham Common. Yet, I recall that when the first insect Red Data Book was published in 1987 there was concern that it was on the brink of extinction on its few localities on the Isle of Wight. How things have changed! A similar situation obtains with the spectacular hoverfly Volucella zonaria. Once it was the prized find for a few entomologists based in the London area and the south coast. Now it has almost reached Scotland!

Tracking species declines is far harder than watching species expansions. Yet there is no doubt that species are disappearing from many parts of lowland southern England. In Mitcham I noticed that the last record I have for the common and easily recognised Leucozona lucorum was in 2002. I have 6 records for the 1980s and 2 from the 1990s when I probably did very little recording in Mitcham but lots elsewhere in Surrey. Now I work Mitcham Common almost daily when at home, but L. lucorum has gone! The habitat has not changed in any wrong way, but we have had some horrendous droughts in those 35 years.

Drought can have unforeseen consequences. For example, the hoverfly Rhingia campestris lays its eggs on grass blades overhanging cow pats. When the larvae emerge, they drop off the grass blade and into the nice moist cowpat. But during extreme heat, the cowpat develops a hard crust far more quickly and the larvae fail to penetrate this. They die, and there is a failure of the second generation of the hoverfly that year. Numbers quickly pick up, however, once wetter conditions are re-established.

The problem comes when there are repeated events before populations can recover. I think this is the key reason why we have seen such serious declines in insects. Yes, land management has changed, and we have lost a lot of suitable habitat, but the biggest impact comes from droughts. The most devastating land management changes happened between the 1950s and 1970s and yet we cannot show its effects because there was very little biological recording. Since the mid-1970s, biological recording has been revolutionised and we are witnessing the impact of a range of factors. Even so, data for the 1980s is far less robust than data for the current day, so we have to rely on occupancy models to explore the changes. Sadly, I think we greatly under-estimate the simple impact of drought, higher temperatures and a more Mediterranean climate.

The one 'upside' of this latest drought is that with luck we will have a lot more data than in the past, so perhaps we will be in a better position to explore the impact of drought!

Square-bashing in Scotland - hoverfly records

Having already reported on the fungus gnat results of my trip, I really ought to provide some information on the hoverflies.

The trip was exceptionally poor for hoverflies, with numbers extremely low at the majority of stops. I suspect that part of the reason lay in the availability of nectar sources. Apart from the buttercup verges of the more southerly areas of the west coast, the main nectar source was pignut Conopodium majus, and even this was limited. Rowan Sorbus aucuparia was almost entirely over; markedly reducing the effectiveness of my activities. So, a large part of my haul arose from sweeping.

These factors meant that the catch was dominated by smaller species, especially within the Bacchini and Chrysogastrini, with Platycheirus, Melanostoma and Sphegina making up a lot of the sample (Figure 1). The Syrphini were noteworthy by their absence, as was the genus Eristalis, although I did see E. rupium on several occasions. In total I recorded 89 species (Figure 2), which I think was substantially below what might have been possible. Amazingly, Dasysyrphus were completely absent and Syrphus were rare.

Figure 1. Numbers of records within each Tribe of hoverfly

The numbers of records look to be high, but it needs to be borne in mind that I record males and females separately, so the actual number of species-level records will be  a fair bit lower (perhaps 25%).

The high point for me was finding Platycheirus amplus, a species I've never previously found. In addition, I think I can say a bit more about the preferred habitat of Lejogaster tarsata, which seems to favour yellow flag iris Iris pseudacorus flushes close to the sea shore; especially where silverweed Potentilla anserina is flowering. The biggest surprise was Brachypalpoides lentus at a patch of Cotoneaster horizontalis at the edge of an old oak woodland. There is plenty of this habitat in the Oban/Loch Lonond area so I suspect this and other Saproxylics are probably more widely distributed than the records suggest. I have one specimen of Sphaerophoria that needs further investigation, as it does not readily fall out in the key in Stubbs & Falk; it might be interesting, but then I might be mistaken and it will prove to be something relatively common.

Figure 2. Hoverflies recorded from the north and west coasts of Scotland between 26 May and 17 June 2018

Square-bashing in Scotland - first results

I have previously written about 'parataxonomy' as a way of improving geographic coverage in less well-supported recording schemes. Whilst on field trips I retain a wide range of insects across many Orders for relevant schemes as a matter of course. The two schemes that benefit most are craneflies and fungus gnats, but I also hold on to specimens for many other Diptera schemes. In addition, I retain specimens for several beetle schemes, Neuroptera and Hemiptera, all of which are effectively by-catch whilst sweeping for flies.

This year, I headed to Scotland on 25 May and focussed on the north and west coasts, which are relatively under-recorded. I got home on 18th June after covering well over 2000 miles. I've still to send samples to many of the recording schemes but have passed on my gnats and craneflies. Results from the gnat sample came back just six days later! Goodness knows how he did it, but Peter Chandler worked through the 92 packets of gnats in less than a week! The resulting records comprised nearly 1,000 new records:

Family	Records
Bolitophilidae	19
Diadocidiidae	7
Drosophilidae	5
Keroplatidae	112
Mycetophilidae	841
Sciaridae	2

The majority of the species were relatively well-known but there were a few that were considered noteworthy and one species new to Britain. So, it looks like my efforts were worthwhile. Breaking down the list into frequency of occurrence it is clear that there is a vast amount still to do. Of the 160 species recorded, all but a handful were seen on fewer than 10 occasions (figure 1). Bearing in mind that a lot of western Scotland is comparatively uniform coverage of Birch woodland with deep stream gullies, I suspect that many of the species I found will be much more widespread than records suggest. There is a lot more to do, and I suspect a repeat trip at the same time of year would yield far more. Change the dates a little bit and the differences are likely to be even more pronounced.

Figure 1. Numbers of records against the species list for gnats and related families recorded from western and northern Scotland in May-June 2018

The trip was extremely hard work, often meaning that I worked into the early hours of the morning - sorting samples and attending to the UK Hoverflies Facebook page. My target group (hoverflies) was extremely thin on the ground, but I do have a massive sample of Empids and Dolichopodids plus numerous packages of Cantharidae (Coleoptera), Hemiptera and Neuroptera. There are also lots of pinned Chrysomelidae, Elateridae and Curculionidae (Coleoptera) and numerous Symphyta (Sawflies) to identify. I will probably do a fair block of these samples but inevitably some will go directly to their respective recording schemes. I hope that they all make a useful contribution but the real big hits are inevitably the gnats and craneflies. The results encourage me to go back! Perhaps others will think about developing as parataxonomists too!