Saturday, 18 November 2017

Changes in recorder techniques - can we detect differences in the dataset?

I have long felt that the shift away from records based largely upon specimens to records based on non-lethal methods such as photography was likely to be influencing the outputs of analyses to investigate trends in species' abundance. Whilst revising the text for the Provisional Atlas, I became acutely aware that some of the trends did not seem to fit my perceptions from field work and from monitoring the UK Hoverflies Facebook page. I therefore suggested to Stuart that it would be helpful to run two separate analyses; one for all data and the other for a subset of the data that excluded known photographic records. The results are really very interesting and can be summed up as a table (Figure 1).

Figure 1. Permutations of possible trends in the HRS dataset and in a subset that excludes photographic records. The final column highlights whether the permutations were found in analysis.
I think the overall results need to be published in the peer-reviewed press because they show how research teams must treat trends with caution. I expect that all of the graphs will become available at some point once we have decided how we might use them in the provisional atlas but probably not readily apparent in any printed version because of the cost of printing in colour. We can of course do so as a pdf without any problem. Here are a couple of examples:
Figure 2. Trends for Cheilosia proxima, All records in blue and with photographic records excluded in red.
Figure 3. Trend for Cheilosia impressa. All records in blue and with photographic records excluded in red.
Figure 4. Trend for Epistrophe diaphana All records in blue and with photographic records excluded in red.

Wednesday, 15 November 2017

Interpreting Sphaerophoria - or can we?

Today, I went through the species accounts for the genus Sphaerophoria, which is one of the more challenging genera for the student of hoverflies. Most records have to be based on males, because there remains quite a lot of uncertainty about the identity of females.

In my experience, most Sphaerophoria fall into two groups. One is largely associated with ericacious heaths and is substantially northern and western in distribution, or is confined to the major heathlands of Dorset, Hampshire and Surrey. It is chacaterised by S. batava, S. fatarum, S. philanthus and S. virgata. The main exceptions are S. scripta, S. interrupta, S. loewi, S. rueppellii and S. taeniata which are mainly grassland and wetland species. Some have clear distributions (e.g. S. tarniata and S. ruppellii) and others such as S. interrupta and S. scripta are widespread and more difficult to fit to a habitat.

So, what is happening to this genus? Quite a few appear to be declining, especially those that are associated with ericaceous habitats. Can we be sure that there have been declines? Anecdotally, I certainly see far fewer S. fatarum and S. philanthus than I used to, and I don't see the numbers of S. batava and S. taeniata that I used to. Part of the reason for my experience is that I no longer work the Surrey heaths where so many of these species occur. BUT I do spend a lot of time in Scotland and I am always pleased to even find a Sphaerophoria. Perhaps I go too early in the year? or perhaps something is happening?

So here are a few maps. Each is as yet un-edited for doubtful records so there may be changes, especially in the square spots which are NBN records which are often very doubtful:

Each map presents a few problems, but the main one is whether the decline in coverage is real or follows a general drift away from difficult taxa. The surprise, therefore is S. taeniata, which seems to be holding its own, at least in terms of coverage. Meanwhile coverage for the other four is clearly declining. There might be a range of reasons, however:

In the case of S. batava, S. fatarum and S. philanthus, these are species that generally occur in ericaceus communities, which in southern England are most widespread in Dorset, Hampshire and Surrey. Both Dorset and Surrey no longer get the attention they once did from recorders who looked seriously at Sphaerophoria. Dave and Ted Levy did a huge amount of work for their Dorset and Somerset atlases, Whilst Graham Collins and I did a great deal in Surrey at the same time. This year, we learned that Ted Levy was retiring from recording, having been unwell for several years, whilst I left Surrey many years ago. So, it is likely that these combined factors may have affected records from southern England - lots of records in the 1980s, but very few recently. BUT, what about Scotland? I make an annual pilgrimage and do a fair bit of recording there. I see very few Sphaerophoria but then perhaps I am going too early or too late in the year?

Then, what about S. rueppellii? It is a classic 'Thames Estuary' species but is not turning up in the numbers it used to. Unlike many of this genus, it can be identified from good photographs but is not often reported (we do see mis-identifications quite frequently). Perhaps it is declining? I suspect something is happening because the map for S. interrupta seems to indicate quite a significant reduction in records from south-east England. As this is the commonest species after S. scripta, and it does seem to have a significant northern and western distribution, I wonder whether SE England is becoming unsuitable for it. If so, my instincts are that this is a result of climate change and in particular periods of intense hot weather and drought.

So, do the maps and trends tell us something about how reliant we are on a small cohort of specialist recorders, or is something more insidious happening? I really don't know!

Tuesday, 14 November 2017

Interpreting data: Portevinia maculata

One of the big benefits of a developing network of new and enthusiastic recorders is that it is possible to cover new ground and look for species that are relatively easy to identify and find.

Portevinia maculata is one such species. Its larvae live in the bulbs of Allium ursinum (Ramsons or Wild Garlic) where they exist from late May through to early April. Adults emerge after a brief spell as a puparium and males (see photo) can be very abundant sitting on the Ramsons leaves for a short spell between late April and early June. We can therefore make use of this species biology to test the reliability of records and can also see how recording efforts have changes as people have become more interested in hoverflies.

Portevinia maculata - male (photo by John Bridges)
 The statistics tell us some interesting stories:

Figure 1. Trend in frequency of records

Figure 1 clearly shows how levels of recording remains substantially constant until around 2011 or 2012, after which they have changed dramatically. This cannot be a result of the plant spreading, as it is largely associated with older woodlands and does not move quickly. Nor is it likely that the insect has moved dramatically. It was always relatively easy to find and when looked for in new locations was usually located if sufficient plants were present and repeat visits made to coincide precisely with what is a very short emergence period. The map, Figure 2, shows the current situation.

Figure 2. Current distribution of Portevinia maculata (to 2017)
And then comes the phenology histogram (Figure 3). Most of the records sit tightly between early April and the middle of June. Experience with the UK hoverflies Facebook group suggests that the earliest dates are indeed around the end of the first week in April and that by the second week in June it has disappeared, even from Scotland. So, we must start to question outlying records in March and from mid-June onwards. We can probably discount the March records because the Ramsons won't have properly emerged and the insect is likely still to be a larva or a very early puparium. Beyond the middle of June it is highly unlikely that adults will be found (possibly the odd female into early July) and the larvae are first or second instar, so absolutely tiny and unlikely to have been found. So records within these timeframes can, with substantial confidence, be logged as erroneous.

Figure 3. Phenology histogram for Portevinia maculata
This is a nice example of the sort of problems one has to consider with all datasets, but is helped because the animal and its associated plant have very well-defined life cycles that make it relatively easy to interpret the data. Where we have less biological information it is much harder to make definitive statements and to question records.

Monday, 13 November 2017

Interpreting data: the problem of Platycheirus

Platycheirus  is one of the largest and more perplexing genera of hoverflies in Britain. There are often good and relatively straightforward male characters (on the front and middle legs) provided you know where and how to look (at high magnification). Females are often much more difficult, with characters that are somewhat subjective and open to misinterpretation: I know they give me problems, so I suspect they give others problems too! (The same applies to Melanostoma).

I have already highlighted the problem with female immarginatus/perpallidus but there are others such as the split between females within the clypeatus complex (clypeatus, europaeus, occultus, and  ramsarensis) and in the peltatus complex (nielseni/peltatus especially). These problems mean that one has to be careful interpreting the data. So, as an example I shall describe the problems with Platycheirus ransarensis.

A bit of history

P. ramsarensis was split from P. clypeatus by Speight & Goeldlin (1990) and Goeldlin et al. (1990). It occurs in oligotrophic (acid, base-poor) situations, most frequently beside moorland streams and lakes, and usually where there are small flushes with abundant sedges or rushes.

When the split was first announced, there was considerable interest amongst the active Dipterists of the day and many went back through their collections to see what they contained. As P. clypeatus was known to be a problem species I suspect many of had a better collection of specimens than we had for easier species. We also took a great interest in looking for this species and our summer field meetings happened to be substantially northern at the time; so lots of records were assembled. There was also considerable activity in the Sorby Natural History Society at the time and they too generated lots of records. Look in the uplands, and it seems that P. ramsarensis is almost a standard species to be expected! More recently, there has been a lot of activity amongst the Devon Fly Group and Dartmoor turns out to be a hot spot. I expect with similar activity both Exmoor and Bodmin Moor will turn out to be other south-western hotspots.

Figure 1. Trend for Platycheirus ramsarensis
Figure 2. Distribution of Platycheirus ransarensis

Interpreting the trend

The overall trend from 1980 says an increase, but I think that would be a misinterpretation. Equally, the trend from 2000 onwards is down, but that too is probably misleading. I think the steep rise prior to 2000 is increased awareness and interest in the Platycheirus splits, whilst in subsequent years there has been a decline in interest in these difficult species that gives an impression that they are occurring less frequently.

In reality we simply don't know what is going on with this one! My field experience suggests that where one looks for it there is a pretty strong chance of finding P. ramsarensis. It is likely to be substantially under-recorded and will probably be found to occur much more widely on the Pennines, in the Lake District, upland Wales, and on the Southern Uplands of Scotland. Further north, I would expect it to be quite widespread across the Western Isles and in the Highlands but probably absent from much of the lower ground of north-east Scotland.

So, will we define a reliable trend? I suspect not unless we start to see an absence from places such as Dartmoor. Whilst the Devon Fly group has many active specialists corroboration of any apparent trend should not be an issue. But, were we to see a decline in specialist activity we might see a negative trend that is a reflection of changing recorder activity rather than a loss of this species.

Sunday, 12 November 2017

How might a Malaise trap network work?

In yesterday's post I raised the question of whether we might emulate the German Malaise Trap programme? The ensuing comments were largely positive and several people raised some really great ideas that helped to get me thinking further. So here is a bit of development thinking:

Does it have to be a centralised project?

Probably not; it could be a series of projects that ran independently but then pooled their data as and when the need arose. I think, however, that there would be great benefits to some sort of central oversight and promotion of the project to make people aware of what was going on and what the opportunities might be to get involved.

In the first instance, it would help to have somebody centrally to raise the funds needed to get the project started. At its simplest level, those funds would need to cover the cost of Malaise Traps and preservatives (degraded Alcohol). I think, however, that ther could be opportunities to link up with universities to develop basic identification skills. Even learning how to sort insects to Order would be useful and might help to open a few eyes. As such, maybe there are University Biology Societies that might like to get involved and have a long-term project? So, some resources to equip study centres with requisite microscopes and keys would be a possible draw on resources.

Do traps have to be run for the whole year?

The simple answer is no. What is needed is a programme that runs consistently from year to year. If it is focussed on 'pollinators' that are of commercial interest I think the traps might best be run from the start of May to the end of June. BUT, of course, running traps for only part of the year is only one option and there might be groups who would run traps throughout the year.

Where might traps be run?

The logical place to run such traps is in association with field centres and maybe the head offices of Wildlife Trusts that have a garden or nearby wild area. It seems to me that this is the sort of project that needs to be associated with stable recording locations where it is possible to have one or more people available to operate the trap and to empty the collecting bottle on a regular basis.

How would the samples be sorted?

My initial thinking was that this is the sort of project where one could develop a nice social group that met on a regular/irregular basis to sort samples together and to learn from one-another. It seems to me that the key to any project that needs to run over many years is to make it a social as well as a serious event. It won't appeal to everybody, but providing you can peer down a microscope for a couple of hours, it offers an opportunity to get to meet people and to do something communally.

I quite like the idea of setting up such a project within the regional and national Natural History Societies. In many ways, I think there is a need for societies to develop new relevance to today's world. Events that draw together people of similar interests can be immensely valuable, especially if they include an element of intellectual stimulation that at least a small proportion of society needs: we hear that this is essential to stave off dementia so there is a growing pool of possible participants!

To what level would one sort the specimens?

At its most basic, sorting to Order would at least mean that the insects collected could be investigated by relevant specialists. BUT, when you bear in mind that there are about 80 recording schemes, there is the potential for quite a large volume of material to be identified to species. That is not to say that all scheme organisers would want to participate, but if only 30% did so that would mean quite a large volume of material identified.

Beyond sorting to Order, counts of individual animals might be possible but I think I would desist from this and concentrate on counts at a more refined level – counts of bumblebees, solitary bees, solitary wasps, some Diptera families and some beetles would be a good start. The key is to choose the taxonomic level that is possible with the volunteers available.

Disseminating the results

No long-term project yields immediate results, and the minimum number of years needed to develop a 'trend' is three years! But in reality a ten year span is needed to start to see real changes. Also, one must be realistic that the first couple of years would be a big learning curve and that the data might not be as robust as it would be after the groups had gained skills.

This brings us back to the question of whether a centralised lead is needed? If the data are to be used, they do need rigorous scientific oversight, so some sort of project management is needed. Within that 'management' team one would need both organisers and analysts, so I think it is probably essential that at some point CEH should be involved. Alternatively, perhaps there is a University team who would get involved?

These are just a few initial musings, but as I ponder I think I can see the shape of such a project developing. Now all that is needed is the fire of enthusiasm somewhere to make it happen!

Saturday, 11 November 2017

Time to copy the Germans?

Recent publicity about the German Malaise Trap programme and the raw data it has created provides a lot of food for thought!

My last post showed how the UK dataset is highly open to modification by changing recorder practices. We are likely to get a lot more data on a sub-set of our fauna (most welcome), but unless we do something quite urgently we may lose track of a significant proportion of the fauna and end up with a politically weaker situation because the dataset is incomplete and can be chewed up by those vested interests who don't like the messages that the data convey.

I cannot see the UK Government ever funding long-term studies that are highly likely to expose even more weaknesses in their environmental policies, so it is down to the voluntary sector to take action. Could we ever get a network of Malaise Traps set up across the UK? Perhaps it is a long-shot but maybe this article is enough to get a few minds thinking about such a project?

Making it happen

It is one thing running Malaise Traps, and a very different matter if one is going to do something with the specimens collected. They have all got to be stored, and ideally something has to be done with them! Sorting Malaise Traps to Order is a major undertaking in its own right, but it might be possible using volunteers.

Sorting to Family is a rather bigger problem. Might we find volunteers? I don't know, but perhaps it is possible. Finding volunteers to extract particular families might be a possibility?

The logistics are frightening, so I take my hat off to the Germans. For the last couple of years I have been providing Axel Ssymank with a bit of help with proof-reading English abstracts for some of this work. He and his colleagues have done an amazing job and it is great to see some powerful messages emerging. I'd love to see something of a similar nature happening in the UK.

This sort of venture seems to me to require a collaborative project that involves all of the major entomological societies together with the Wildlife Trusts and perhaps others. Could it be pump-primed by a Heritage Lottery Fund project? I can certainly see a possible project, but is there an organisation that might take the lead?

Perhaps this is a step too far for volunteers? BUT, many societies need projects and long-term initiatives to give cohesion to their activities. Could this be one? In some ways, there might be parallels with 'Operation Wallacea' that looked at the fauna of Sulawesi in considerable detail and caused huge excitement at the time.

Any takers?

Interpreting trends in abundance

There is an increasing interest in the trends in the abundance of invertebrates, as illustrated by recent posts about 'pollinators', so it is perhaps apposite that I look at some of the trends in hoverfly abundance and think a bit about the reasons for such trends.

I am starting to get quite concerned at the messages that are emerging and begin to think that we probably need to develop two indices: one based on photographic records, which now make up the bulk of the data arriving at the HRS, and the other based on a small number of individuals who still retain specimens.

My rationale is driven by some startling trends that I don't think can be put down to environmental issues; or, at least, I don't think we can disentangle environmental and recording pressures sufficiently to provide a reliable account.

Here are a few examples extracted from the revised trends that should appear in the updated hoverfly atlas (when it finally emerges):

Neoascia podagrica

This is a relatively common species and we see lots of photographs of Neoascia that are probably N. podagrica, but we cannot be entirely sure from a photograph. What we can say is that it is likely to be either N. podagrica or N. obliqua, and unless there is butterbur Petasites hybridus close-by the chances are that it is N. podagrica. So a question-mark hangs over the data. We can also say that we very rarely see records of Neoascia that don't have infuscated outer cross-veins and yet our own field experience tells us that they are quite widespread and abundant (even super-abundant) in the right places.
Figure 1. Trend for Neoascia podagrica with marked downturn in abundance in the past 5 years.

Melanostoma mellinum

I have always found that this species occurs in lower numbers than M. scalare and it is quite plausible that its numbers have declined somewhat. But the data can be skewed by the numbers of people who record in the uplands where M. mellinum is far more prevalent. We see comparatively few records of this species as photographs, a phenomenon that is possibly complicated by the difficulty of doing Melanostoma from photographs.

Figure 2. Trend for Melanostoma mellinum in which a clear downward trend seems to have intensified in the past 4 years for no obvious reason.

Cheilosia illustrata

This is one of the traditional 'hogweed fauna' and as such might be expected to follow a similar trend to others such as Leucozona glaucia that has declined dramatically in south-east England. Yet, the trend appears to go in completely the opposite direction! My instincts are that because photographic recorders tend to concentrate on larger and more obvious species, this is one of the beneficiaries of such recording.

Figure 3. Trend for Cheilosia illustrata in which the upturn coincides quite closely with the arrival of iSpot and more data extraction from web-based sources.
This trend contrasts strongly with Cheilosia proxima, another member of the 'hogweed fauna' whose larvae are also associated with a very common plant (creeping thistle Circium vulgare).

Figure 4. Trend for Cheilosia proxima.


Do differences in recording technique matter?

Dramatic changes in the apparent fortunes of our insect fauna have already elicited sensational (correctly in my view) responses in the press. So we need to be clear about our interpretation of the results. There are parallels: the change in bathymetric readings on some estuaries was quite clear when traditional 'lead lining' was replaced by sonar readings. Without this change being recognised, it was possible that an incorrect geomorphological interpretation would be placed on the change. This in turn might have had important implications for our ports industry.

So, quite simply, yes we do need to think about the ways in which data are gathered and must think about our interpretations. With hoverflies, we are just about getting to the point where the data are so dominated by photographic recording that any change will be followed by a new asymptote, after which we can follow new trend lines. BUT, critically, we need to understand and take the changes into account. Thus, it is unwise to look at trends without undertaking critical evaluation taking into account the factors beyond simple environmental parameters that might affect the animal in question.

My feeling is that for the foreseeable future it will be necessary to develop parallel indicators using two different datasets - one based on traditional recording from retained specimens and the other from the photographic record. I suspect there is an element of inevitability that the photographic record will start to become the dominant indicator; in which case we need to be aware that trends for difficult taxa may not be wholly reliable and thus conservation organisations will need to develop new ways of thinking about species' status and conservation policy.