An insect isoscape of UK and Ireland

Rationale: The study of insect migration is problematic due to the small size of insects. Stable isotope analysis can be used to elucidate movement, either by geographic assignment of location of a species, or by simply distinguishing migrant from resident populations. There are few isoscapes of any kind in the UK/Ireland available for interrogation. Thus, I have measured stable isotope ratios (of H, C, N and S) of 299 individuals of the non-migratory Brimstone moth ( Opisthograptis luteolata ) collected from 93 locations around the UK and Ireland by citizen scientists. Methods: After removing lipids, stable isotope ratios were measured by continuous flow isotope ratio mass spectrometry, using either a conventional elemental analyser (C, N and S) or a high-temperature, thermal conversion elemental analyser in reductive mode. Results: Maps (isoscapes) were constructed that illustrate the stable isotope spatial distribution of this insect. These are the first isoscapes of H, C, N and S of biological samples covering both UK and Ireland. Conclusions: The insect isoscape patterns can be explained from what we know of moth diet, climate and geology. Sulfur isotopes may be of particular use for distinguishing individuals from areas of unique geology. Isoscape patterns may (with care) predict isotope compositions of other, herbivorous, non-aquatic, chitinous taxa. Such isoscapes, when extended beyond the UK and Ireland, would provide a useful tool to elucidate insect migration.


| INTRODUCTION
Migration in insects is widespread 1 and in some species exceeds well-known mammalian and avian migrations in terms of biomass. 2 It is crucial to link locations used by individuals to understand their ecology, to conserve threatened species and to understand how climate change may change their distribution and numbers. Insect migration also has important implications for humanssome of these are negative, since some migratory insects are crop pests or act as disease vectors; conversely, other migrants are important pollinators and therefore provide positive impacts. Despite this, our understanding of insect migration is surpassed by that of vertebrates simply because of the small size of insects. Size is the main barrier to tracking individual insects over long distances; this is unfortunate, since many insect migrations may be multigenerational where each individual participates in only a section of the migratory circuit.
Tracking migration has involved many innovations, resulting in a multitude of techniques appropriate to the taxon in question.
Extrinsic markers (e.g. tags) require both capture and recapture but this has been popular for insects given the inherent challenges of their small size. For example, the monarch butterfly migration of North America has been extensively studied using numbered adhesive tags. 3 However, this does require considerable effort since the recovery rate of tagged individuals in such a large population is quite low, 0.046 to 1.27%. 4 Recent technological advances have allowed the attachment of miniaturised radiofrequency transmitters to taxa as small as 0.2 g. 5,6 Such innovations have expanded our ability to track the migration of small-sized taxa; however the cost and possibility of adverse effects on flight behaviour remain problematic.
Intrinsic markers such as stable isotope analysis require only one capture to reveal potential information about origin and are therefore less costly. Despite their typically lower resolution than that of extrinsic methods, stable isotopes comprise a convenient tool with which to investigate migration. 7 15 and Globe Skimmer dragonflies (Pantala flavescens, δ 2 H values). 16,17 In addition, emergent Red Admiral butterflies (Vanessa atalanta) and Silver Y moths (Autographa gamma) from several locations in western/ southern Europe have been discriminated using δ 2 H values, 18 and six species of North American butterfly were also provenanced using δ 2 H values. 19 To assign location based on stable isotope ratios, one requires a spatial landscape of environmental isotopic variation or "isoscape". 20 There are essentially two approaches to this: (i) for hydrogen and oxygen isotopes, using precipitation isoscapes derived from the Global Network of Isotopes in Precipitation (GNIP) 21,22 and a linear transfer function which relates the stable isotope ratio of tissues synthesised at known locations to that of precipitation 8 ; and (ii) a direct approach using the isotope ratio of a non-migratory analogue.
For insect migration, moths are ideal candidates, since moth trapping is a popular pastime and a well-chosen resident species can provide many geographically disparate points from which to interpolate an isoscape.
Only two GNIP stations exist in UK/Irelandin Wallingford, England, and Valencia, Ireland, which limits hydrogen isoscape modelling. No other country-wide isoscapes exist, although there are regional published isoscapes for different matrices, e.g. a Scotlandwide freshwater lake hydrogen/oxygen isoscape, 23 and a soil sulfur isoscape in Northern Ireland. 24 Thus, this study represents the first large-scale isoscapes of biological samples for the four elements H, N, O and S.

| METHODS
The first question when considering which part of the insect to analyse, is whether the tissue is metabolically active. When adult butterflies emerge from pupae, their wings are considered inert, 10 with no further tissue turnover, and so an individual's wings retain the isotopic signature of their larval origins. However, since the wings are composed primarily of the polysaccharide chitin ((C 8 H 13 O 5 N) n ), and therefore essentially free of S (bar small amounts of protein), the wings are unsuitable for δ 34 S analysis. Since the aim was also to investigate the sulfur isoscape of UK/Ireland, combined head/thorax were analysed for S (and N and C) isotope ratios. Although these tissues are at least partially metabolically active, the Brimstone moth is a resident, so it was assumed that the S, N and C compositions of the moth body would represent an accurate assessment of their environment.

| Sample collection
Samples were collected from around the UK and Ireland by canvassing county recorders from the National Moth Recording Scheme, and then from volunteers contacted directly to fill in the "gaps" where possible. Instructions were to collect three Brimstone moths on disparate occasions between May 1st and August 31st, although in a small number of cases samples were collected outwith those dates. The number of samples, three, collected was based on an attempt to maximise spatial coverage, explore uncertainty within a location, and minimise analysis costs. In some cases more samples were collected than requested, and in other cases, notably towards the edges of Brimstone moth distribution, only one or two samples were collected. In one location a deliberate attempt was made to collect multiple samples throughout the entire flight period, to investigate any potential seasonal isotopic change.
Moths were all collected nocturnally by light traps of variable construction. Each volunteer was asked to store each moth in a polystyrene petri dish held within a sealed polyethylene sample bag which snugly held the petri dishes, in a domestic freezer. Once the required number of samples had been collected, these were mailed to the laboratory, accompanied by silica gel sachets to reduce humidity.
On arrival at the laboratory, the samples were logged in a database and freeze-dried, before storage in labelled 20-mL glass jars which were placed in an electric desiccator. When all samples had been returned, each sample was twice cleaned in 10 mL of 2:1 chloroform/ methanol for 1 h, the solvent decanted, and the sample then air-dried in a fume hood. This has the effect of removing lipids which can change the %H exchangeability, the non-exchangeable δ 2 H value and the δ 13 C value of the samples. Furthermore, this may remove pigments which can interfere with the geographic δ 2 H signal 25 and possibly the other isotopic signals, 26 although in the case of the uniformly yellow Brimstone moths this is less likely to be a major concern.

| Stable isotope measurement
For δ 2 H analysis, a core of wing material was extracted from the rear wing of each individual, with care being taken not to include any of the dark-pigmented wing. This was weighed (0.101 ± 0.029 mg) into a silver capsule (5 3.5 mm) and loaded into a UniPrep autosampler (Eurovector, Milan, Italy) 27

| Data exploration/analysis
It is general practice to record minimum overnight temperature for moth recording (UK Garden Moth Scheme instructions); however, only two-thirds of the samples had these temperatures recorded.
Instead, daytime temperature was extracted from the CEDA archive. 31 Linear relationships were explored between daytime temperature, date of trapping, latitude and longitude, and isotope ratios (

| RESULTS
There was a considerable range of stable isotope composition among moth samples ( In terms of latitude and longitude, several significant relationships were discovered (Table S1,  values. 12 The linear transfer function is significant, comparable in slope to, but slightly lower, than those of North American Lepidoptera 10,37 but with a lot of scatter. Whilst the δ 2 H wing isoscape follows that of δ 2 H p , one source of noise in the data revolves around Brimstone moth phenology. 38 In the north of the UK the larvae are found around summer/autumn and precipitation in these months would be reflected in the wing tissue δ 2 H values. However, in the south, caterpillars can be found in almost all months, which might cause extra "seasonal" noise in the more southern δ 2 H moth data. This may explain why the most obvious gradient in δ 2 H values is in the far NW where the caterpillar season is much shorter; δ 2 H patterns in the south are more likely to be masked by seasonal effects from the previous season when the larvae were feeding. Given the deciduous tree leaf diet of Brimstone moths, their δ 15 N values will reflect that of host plants, and trophic enrichment can be negated since it will be the same for all samples. Globally, leaf δ 15 N values largely reflect the soil δ 15 N values, 39 but also integrate factors such as fractionation due to differing mycorrhizal associations 40 and uptake of atmospheric reactive N deposition. 41 From large analyses of global leaf data, the δ 15 N value increases with mean annual precipitation and, above À0.5 C at least, increases with mean annual temperature 39,40 ; in other words, much of the δ 15 N variation in leaves has climatic controls. It appears F I G U R E 1 Locations where moths were collected impossible to disentangle these conflicting effects on δ 15 N values in a way that would explain the moth isoscape ( Figure 3). Although anthropogenic reactive nitrogen is increasing in the atmosphere, 42 anomalies due to the proximity of traps to anthropogenic sources such as roads or industrial sources would be difficult to unpick from the data here. There are very few published examples of sulfur isoscapes anywhere in the world. 24,51,52 The largest δ 34 S gradient is that between land and sea with δ 34 S values in land mammals reflecting distance to the sea 53 following the "sea-spray effect", 54 which deposits marine-derived sulfate (δ 34 S $ +21‰,) 55  In addition to (non-aquatic) insects, the isoscapes also, however, offer the opportunity of investigating the migration (or otherwise) of other taxa, for instance insectivorous birds. I advocate the use of citizen scientist projects of this nature for building isoscapes, as the data would be otherwise unattainable by a single researcher.

PEER REVIEW
The peer review history for this article is available at https://publons. com/publon/10.1002/rcm.9126.

DATA AVAILABILITY STATEMENT
The data that supports the findings of this study are available in the supplementary material of this article.