by Matt Bowser1, Anya Bronowski, and Dom Watts2
Pollinating insects provide important ecosystem services in Alaska (Fulkerson et al. 2021) and the pollinators themselves are wildlife that the Kenai National Wildlife Refuge (KNWR) was established in part to conserve (Kenai National Wildlife Refuge and US Fish & Wildlife Service, Alaska Regional Office, Division of Conservation Planning & Policy 2010). Because many pollinators appear to be generally declining (Potts et al. 2010, Cameron et al. 2011, Koh et al. 2016), our objective was to begin documenting pollinator diversity on KNWR and surrounding lands.
The Alaska Bee Atlas (Fulkerson et al. 2021, https://accs.uaa.alaska.edu/wildlife/ak-bee-atlas) is a sampling program designed to provide information on the biodiveristy of pollinators throughout Alaska. In 2022, KNWR biologists participated in the Alaska Bee Atlas effort.
We followed the sampling plan guidance of Fulkerson et al. (2021). Most of KNWR lies within lowest priority areas mapped in Fulkerson et al. (2021), but the southernmost part of the Refuge lies within a medium priority area. We prioritized sampling in this area, but access in this area is difficult. We surveyed only at Emerald Lake in this medium priority area.
We surveyed for insect pollinators at a variety of other sites on the Refuge, trying to sample in diverse habitats (Figure 1). We sampled dry, rocky slopes off of Skilak Lake Road following the advice of Justin Fulkerson (Alaska Center for Conservation Science, Anchorage, Alaska).
We accessed sites by road and floatplane.
We sampled pollinators using bee bowl traps (Figure 2), blue vane traps (Figure 3), and aerial nets (Figure 4), generally following the field methods of Fulkerson et al. (2021) with the exception that we collected specimens into SK picglobal 99.9% pure propylene glycol. Field notes are available from Bowser (2022c) and Bronowski (2022).
Samples were stored in a -23°C freezer except when samples were being processed. Invertebrates were separated from debris by hand under a dissecting microscope. Care was taken to reduce possible cross-contamination of DNA among samples.
We separated samples that were all or mostly bees from samples that were mostly flies and other invertebrates. We shipped 12 samples of bees to the Alaska Center for Conservation Science3, University of Alaska Anchorage, Anchorage, Alaska to be processed by methods described by Fulkerson et al. (2021).
We homogenized the remaining 19 samples plus one legacy bulk pollinator sample from a previous project (Bowser 2012) using a blender and cleaning between samples with DIY-DS cleaning solution as described by Buchner et al. (2021). Our sample homogenization protocol is included below.
We homogenized samples using a Nutri Ninja QB3000SS blender.
DIY-DS recipe
- 20 g NaOH
- 20 g Alconox
- 15.1 g NaHCO3
- 267 ml 4.5% bleach
- deionized water to fill to 2 l
Preparation
- 120 ml plastic cups should be washed with DIY-DS and rinsed before sampling. Finish by rinsing inside the 120 ml cup with deionized water. Hand dry 120 ml cup with paper towel.
Homogenize samples
- Before running samples, rinse blender by running 100 ml of deionized water for 20 s.
- Pre-label a 10 ml plastic vial with the specimen GUID and add a barcode vial label. Also pre-label and add a barcode label to a 120 ml specimen cup.
- Clean forceps with DIY-DS.
- Take the label out of the original container with the cleaned forceps and place into the new 120 ml sample container.
- Add the contents of the sample vial to the blender.
- Rinse original sample vial with cold, clean propylene glycol and pour rinsate in the blender with the rest of the sample.
- Fill blender to 100 ml with cold, clean propylene glycol.
- Blend for 90 s.
- Using a new disposable pipette, fill the pre-labelled 10 ml plastic vial with about 9.5 ml of homogenate.
- Pour the rest of the sample into the pre-labeled 120 ml specimen cup.
- Rinse blender by running 100 ml tap water for 10 s.
- Wash blender by running 100 ml of DIY-DS for 10 s.
- Rinse this out in the lab sink with tap water.
- Rinse blender by running 100 ml deionized water for 10 s.
We shipped 9 ml of homogenate from each of the 20 homogenized samples to Molecular Research Laboratory4, Shallowater, Texas for metabarcoding.
We chose to use the mlCOIintF/jgHCO2198 (GGWACWGGWT GAACWGTWTA YCCYCC / TAIACYTCIG GRTGICCRAA RAAYCA) primer set of Leray et al. (2013) for PCR, targeting a 313 bp region of the COI DNA barcoding region.
The mlCOIintF/jgHCO2198 primer pair was used with barcodes on the forward primer in 30–35 PCR cycles using the HotStarTaq Plus Master Mix Kit (Qiagen, USA) under the following conditions: 94°C for 3 minutes, followed by 30–35 cycles of 94°C for 30s, 53°C for 40 seconds and 72°C for 1 minute, after which a final elongation step at 72°C for 5minutes was performed. After amplification, PCR products were checked in 2% agarose gel to determine the success of amplification and the relative intensity of bands. Multiple samples were pooled together in equal proportions based on their molecular weight and DNA concentrations. Pooled samples were purified using calibrated Ampure XP beads. The pooled and purified PCR product was used to prepare an illumina DNA library. Sequencing was performed at MR DNA on a MiSeq following the manufacturer’s guidelines.
The bioinformatics pipeline was run on the Yeti supercomputer (USGS Advanced Research Computing 2021). We used the MetaWorks pipeline, version 1.11.3 (Porter and Hajibabaei 2022) with the RDP classifier (Wang et al. 2007) and the Eukaryote CO1 reference set for the RDP Classifier, version 4.0.1 (Porter and Hajibabaei 2018). We processed data in R, version 4.2.2 and 4.2.3 (R Core Team 2022, 2023) using the R packages ape, version 5.7-1 (Paradis and Schliep 2019); Biostrings, version 2.66.0 (Pagès et al. 2022); bold, version 1.2.0 (Chamberlain 2021a); curl, version 5.0.0 (Ooms 2023a); ips, version 0.0.11 (Heibl 2008); msa, version 1.30.1 (Bodenhofer et al. 2015); openssl, version 2.0.6 (Ooms 2023b); reshape2, version 1.4.4 (Wickham 2007); ritis, version 1.0.0 (Chamberlain 2021b); and uuid, version 1.1-0 (Urbanek and Ts’o 2022).
We compared our sequences to sequences from a local reference library (Bowser 2022a) using the vsearch --usearch_global
command of vsearch, version 2.21.1 (Rognes et al. 2016).
In order to exclude potential false positive detections as defined by MacKenzie et al. (2006) due to demultiplexing errors (see Deiner et al. 2017), we conservatively removed from the Exact Sequence Variant (ESV) table all occurrences that represented less than 0.4% of the total number of reads for any ESV, based the experience of (Bowser 2023b), where an apparent rate of mis-assignment of up to 0.36% was found. We also removed all occurrences represented by only one or two reads.
Complete methods including all configurations, commands, and scripts used for processing data are available from Bowser (2023a).
Identfications of Bombus species were conformed to the names provided by Sikes and Rykken (2020).
Project data and project photos are available on Arctos via an Arctos project record (https://arctos.database.museum/project/10003917) and specimen records can be viewed via an Arctos search5. Project information is also available from a project record on ServCat (https://ecos.fws.gov/ServCat/Reference/Profile/148742). Raw sequence data from this project are available from Bowser (2022b). Resulting occurrence data have been published as an occurrence dataset (Bowser et al. 2023). Results from specimens sent to the Alaska Center for Conservation Science are available from the 2018 to 2022 Results Map at https://arcg.is/1myveP.
The single legacy sample from 2011 yielded 71 species and 18 BINs (Table 1). The 17 samples collected in 2022 yielded 206 species and 85 BINs (Table 2).
Phylum | Class | Order | Species | BINs |
---|---|---|---|---|
Arthropoda | Arachnida | Araneae | 1 | 0 |
Arthropoda | Insecta | Coleoptera | 1 | 0 |
Arthropoda | Insecta | Diptera | 61 | 14 |
Arthropoda | Insecta | Hemiptera | 1 | 2 |
Arthropoda | Insecta | Hymenoptera | 2 | 2 |
Arthropoda | Insecta | Lepidoptera | 3 | 0 |
Arthropoda | Insecta | Odonata | 1 | 0 |
Ascomycota | Dothideomycetes | Capnodiales | 1 | 0 |
Phylum | Class | Order | Species | BINs |
---|---|---|---|---|
Annelida | Clitellata | Crassiclitellata | 1 | 0 |
Arthropoda | Arachnida | Araneae | 11 | 0 |
Arthropoda | Arachnida | Sarcoptiformes | 2 | 0 |
Arthropoda | Arachnida | Trombidiformes | 0 | 1 |
Arthropoda | Collembola | Collembola | 3 | 0 |
Arthropoda | Collembola | Entomobryomorpha | 1 | 1 |
Arthropoda | Collembola | Symphypleona | 0 | 1 |
Arthropoda | Insecta | Coleoptera | 22 | 3 |
Arthropoda | Insecta | Diptera | 99 | 60 |
Arthropoda | Insecta | Ephemeroptera | 1 | 0 |
Arthropoda | Insecta | Hemiptera | 16 | 9 |
Arthropoda | Insecta | Hymenoptera | 28 | 10 |
Arthropoda | Insecta | Lepidoptera | 9 | 0 |
Arthropoda | Insecta | Orthoptera | 1 | 0 |
Arthropoda | Insecta | Psocodea | 1 | 0 |
Arthropoda | Insecta | Thysanoptera | 2 | 0 |
Arthropoda | Insecta | Trichoptera | 1 | 0 |
Chordata | Aves | Galliformes | 1 | 0 |
Chordata | Mammalia | Rodentia | 1 | 0 |
Mollusca | Gastropoda | Stylommatophora | 3 | 0 |
Ascomycota | Dothideomycetes | Capnodiales | 1 | 0 |
Ascomycota | Sordariomycetes | Hypocreales | 2 | 0 |
In the single sample from 2011, the most abundant species in terms of read abundances was 16,848 reads of Ctenicera angusticollis (Figure 5). Other abundant identifications were flies in the families Muscidae, Anthomyiidae, and Fanniidae. In 2022, the highest read abundance was for Speyeria mormonia (Boisduval, 1869) (Lepidoptera: Nymphalidae), which we detected in three samples (Figure 6). Reads of Helina species (Diptera: Muscidae), Rhadiurgus variabilis (Zetterstedt, 1838) (Diptera: Asilidae), and Xylota subfasciata Loew, 1866 were also abundant.
The most frequently observed identifications were four Helina species that were detected in 5–10 out of the 17 samples (Figure 7). The fungus Cladosporium allicinum (Fr.) Bensch, U.Braun & Crous (Capnodiales: Cladosporiaceae) was detected in 6 samples. The bee parasite Apocephalus borealis Brues, 1924 (Diptera: Phoridae) was detected in four samples.
We detected no bees in the sample from 2011 and 12 species of bees in 2022 (Table 3). All of these bee species are widespread in Alaska based on occurrence records available through the Global Biodiversity Information Facility (GBIF, https://www.gbif.org/), but we are not aware of other records of Andrena milwaukeensis or Halictus rubicundus from the Kenai Peninsula. All of the Bombus species we found are known to be abundant or common in our area (Rykken 2022).
Family | Species |
---|---|
Andrenidae | Andrena milwaukeensis Graenicher, 1903 |
Apidae | Apis mellifera Linnaeus, 1758 |
Apidae | Bombus flavifrons Cresson, 1863 |
Apidae | Bombus frigidus Smith, 1854 |
Apidae | Bombus insularis (Smith, 1861) |
Apidae | Bombus lapponicus sylvicola Kirby, 1837 |
Apidae | Bombus melanopygus Nylander, 1848 |
Apidae | Bombus mixtus Cresson, 1879 |
Apidae | Bombus sitkensis Nylander, 1848 |
Halictidae | Halictus rubicundus (Christ, 1791) |
Megachilidae | Megachile melanophaea Smith, 1853 |
Megachilidae | Megachile relativa Cresson, 1878 |
Flies were by far the most speciose group collected by our sampling effort. In the single sample from 2011 we documented 61 species and 14 BINs representing 24 families (Table 4). In 2022 we found 99 species and 60 BINs of flies in 32 families (Table 5). In both years the most diverse families observed were Muscidae and Anthomyiidae.
Family | Species | BINs |
---|---|---|
Anisopodidae | 1 | 0 |
Anthomyiidae | 12 | 1 |
Bibionidae | 1 | 0 |
Calliphoridae | 2 | 0 |
Chloropidae | 0 | 1 |
Empididae | 1 | 2 |
Fanniidae | 1 | 0 |
Lauxaniidae | 1 | 0 |
Limoniidae | 1 | 1 |
Lonchaeidae | 0 | 2 |
Muscidae | 20 | 1 |
Mycetophilidae | 0 | 1 |
Phoridae | 1 | 0 |
Pipunculidae | 2 | 0 |
Psilidae | 1 | 0 |
Sarcophagidae | 1 | 0 |
Scathophagidae | 3 | 1 |
Sciaridae | 2 | 0 |
Sciomyzidae | 1 | 0 |
Sepsidae | 1 | 0 |
Stratiomyidae | 1 | 0 |
Syrphidae | 6 | 3 |
Therevidae | 0 | 1 |
Tipulidae | 2 | 0 |
Family | Species | BINs |
---|---|---|
Agromyzidae | 1 | 1 |
Anthomyiidae | 10 | 6 |
Anthomyzidae | 1 | 0 |
Asilidae | 1 | 0 |
Bibionidae | 1 | 0 |
Calliphoridae | 2 | 0 |
Ceratopogonidae | 0 | 1 |
Chironomidae | 4 | 4 |
Chloropidae | 4 | 3 |
Culicidae | 1 | 1 |
Dolichopodidae | 6 | 5 |
Empididae | 2 | 2 |
Ephydridae | 2 | 0 |
Fanniidae | 2 | 2 |
Heleomyzidae | 2 | 1 |
Hybotidae | 2 | 4 |
Lauxaniidae | 1 | 0 |
Muscidae | 27 | 6 |
Mycetophilidae | 6 | 3 |
Phoridae | 8 | 8 |
Pipunculidae | 1 | 0 |
Rhagionidae | 0 | 1 |
Sarcophagidae | 2 | 1 |
Scathophagidae | 1 | 0 |
Scatopsidae | 1 | 0 |
Sciaridae | 2 | 3 |
Sepsidae | 1 | 0 |
Simuliidae | 1 | 4 |
Sphaeroceridae | 1 | 1 |
Stratiomyidae | 2 | 0 |
Syrphidae | 3 | 2 |
Tachinidae | 1 | 1 |
Apocephalus borealis, a parasitoid of bees and vespine wasps (Tihelka et al. 2021) was both frequently observed in our samples from 2022 and abundant in terms of read counts. Adult A. borealis might have been collected, but it is more likely that these internal parasites were within their hosts at the time they were collected.
The fungus Cladosporium allicinum is common from environmental samples worldwide, collected from living and dead plants, air, water, and humans (Schubert et al. 2007, Bensch et al. 2012). We found no literature records of Cladosporium allicinum taken from insects, but Cladosporium species can function as symbionts or pathogens of insects (Liu et al. 2022).
We detected Wolbachia (Rickettsiales: Anaplasmataceae) sequences in the 2011 sample and in 15 out of the 17 samples from 2022. Wolbachia bacteria infect many insect species and alter their hosts’ reproductive systems (Werren 1997). Wolbachia can be present in high proportions of pollinator populations (Evison et al. 2012).
We also detected Steinernema (Rhabditida: Steinernematidae) in one sample6, a sample where most reads were of Bombus species. The Steinernema sequence was 97.34% identical to a sequence identified as Steinernema kraussei (Steiner, 1923) Travassos, 1927, a species known to be pathogenic to Bombus terrestris (Linnaeus, 1758) (Dutka et al. 2015).
Ero canionis Chamberlin & Ivie, 1935 (Araneae: Mimetidae); Atomaria testacea Stephens, 1830 (Coleoptera: Cryptophagidae); Liriomyza baptisiae (Frost, 1931) (Diptera: Agromyzidae); Fannia neopolychaeta Chillcott, 1961 (Diptera: Fanniidae); Tachypeza fenestrata (Say, 1823) (Diptera: Hybotidae); Phaonia protuberans Malloch, 1923 (Diptera: Muscidae); Phaonia serva (Meigen, 1826); Megaselia hirticrus (Schmitz, 1918) (Diptera: Phoridae); Megaselia lucifrons (Schmitz, 1918); Agria housei Shewell, 1971 (Diptera: Sarcophagidae); Boettcheria litorosa (Reinhard, 1947) (Diptera: Sarcophagidae); Olethreutes bipunctana (Fabricius, 1794) (Lepidoptera: Tortricidae); Coleophora quadruplex McDunnough, 1940 (Lepidoptera: Coleophoridae); Peristenus howardi Shaw, 1999 (Hymenoptera: Braconidae) and Cubocephalus anatorius (Gravenhorst, 1829) (Hymenoptera: Ichneumonidae) appear to be new records for Alaska. The the non-native species Odontothrips loti (Haliday, 1852) and Uroleucon taraxaci (Kaltenbach, 1843) also appear to be new for the state.
Coleophora quadruplex was known from multiple Canadian provinces (Pohl et al. 2018, GBIF.Org 2023e), but we found no previous records from Alaska. Olethreutes bipunctana had also been reported from Canada (Pohl et al. 2018, GBIF.Org 2023f), but apparently not from Alaska.
Fannia neopolychaeta had been known from as close to Alaska as British Columbia (Chillcott 1960) and Yukon Territory (GBIF.Org 2023a), but our record appears to be new for Alaska. Eudasyphora canadiana Cuny, 1980 had been reported from Alaska by Cuny (1980), but there were no georeferenced Aslakan records in GBIF. Phaonia protuberans was known from the Northwest Territories and Yukon Territory (Huckett 1965, GBIF.Org 2023b), but it had not been reported from Alaska. Phaonia serva (Meigen, 1826) occurs in the Northwest Territories (Huckett 1965), but ours appears to be the first record from Alaska. Tachypeza fenestrata appears to be a new record for Alaska, but this species is present nearby in Yukon Territory. Aspistes spathis had been reported from Alaska by Cook (1965), but we found no georeferenced Alaskan occurrences (GBIF.Org 2023d). We found no Alaska records of Megaselia lucifrons, but this species is known from Yukon Territory. Agria housei and Boettcheria litorosa appear to be new records for Alaska, but there are records of these species from Yukon Territory.
Peristenus howardi (Hymenoptera: Braconidae) had been reported from Idaho and Washington by Day et al. (1999) and it has since been found in Alberta (Zhang 2018, GBIF.Org 2023c), but this species had not been reported from Alaska.
We documented occurrences of seven non-native species. The European honey bee, Apis mellifera Linnaeus, 1758, was detected in bee bowls set out at Headquarters lake wetland near Soldotna7. The thrips Odontothrips loti (Haliday, 1852) (Thysanoptera: Thripidae) was detected in bee bowls deployed in a meadow off of Skilak Lake Road8. We found Uroleucon taraxaci in a disturbed clearing off of Ski Hill Road9. Pollenia vagabunda (Meigen, 1826) (Diptera: Polleniidae) was collected in bee bowls at Kenai National Wildlife Refuge headquarters in Soldotna10, where this species had been found previously by Bowser (2015). We detected the earthworm Dendrobaena octaedra (Savigny, 1826) (Crassiclitellata: Lumbricidae) in bee bowls set at Picnic Lake11. Sequences we obtained of the slug Deroceras agreste (Linnaeus, 1758) (Stylommatophora: Agriolimacidae) from bee bowls set at multiple locations12 were 99.35–99.68% identical to sequences identified by Zając and Stec (2020) as Deroceras agreste. We found the slug Arion fuscus (O.F.Müller, 1774) (Stylommatophora: Arionidae) off of the Vista Trail13.
Pollenia vagabunda had been found at the Kenai National Wildlife Refuge headquarters area previously (Bowser 2015). Uroleucon taraxaci (Kaltenbach, 1843) is believed to be introduced in North America (Foottit et al. 2006) and has been recorded from as close to Alaska as Yukon Territory (Maw et al. 2000). We detected this aphid in in bee bowls set in the back lawn of KNWR headquarters, where its host, Taraxacum officinale Weber ex Wiggins, is abundant. The epigeic earthworm Dendrobaena octaedra is almost ubiquitous near roads on KNWR (Saltmarsh et al. 2016) and this worm does climb (Römbke et al. 2017), so its presence in a bee bowl was not surprising. Deroceras agreste had previously been found in the Ski Hill Road area by Bowser et al. (2020); its occurrence at Picnic Lake in the Mystery Creek area was new. Our finding of Arion fuscus was the first record of an arionid slug on KNWR. Arion fuscus had previously been identified in Alaska from Sitka (Schade 2018).
We detected two vertebrate species: a single record of Willow Ptarmigan (Lagopus lagopus (Linnaeus, 1758)) at an alpine meadow above Twin Lakes14 and three records of northern red-backed voles (Myodes rutilus (Pallas, 1779)) in the vicinity of the Kenai National Wildlife Refuge’s headquarters in Soldotna15. At Twin Lakes we had seen and heard a family of Willow Ptarmigan within about 100 m of the area where we had sampled pollinators using aerial nets. In the sample from which the Willow Ptarmigan DNA was detected, no biting flies were detected, so the ptarmagin record was not from a blood meal of a fly. There might have been ptarmigan DNA in or on muscid or phorid flies in the sample. There were also no biting flies detected in any of the three bee bowl samples where vole DNA was found. Voles may have tasted the propylene glycol or otherwise explored the bee bowls.
Some of our reads were 100% similar to sequences both identified as Spilogona alticola (Malloch, 1920) (Diptera: Muscidae) and Spilogona contractifrons (Zetterstedt, 1837) in BOLD BIN BOLD:AAB527816. Huckett (1965) expressed his doubt that these two species were distinct. We assigned these to Spilogona contractifrons as we have done in previous work (Bowser et al. 2020). Other reads were 100% similar to sequences of Spilogona sp. 12AKR sensu Renaud (2012), which had also been documented locally by Bowser et al. (2020).
Complementing morphological identifications by metabarcoding enabled us to efficiently identify many more non-bee species than we would have been able to process and identify in a timely way and it also provided detections of some bee parasites. We believe that using multiple methods is an expedient way to improve our understanding of insect pollinators in Alaska.
We thank Justin Fulkerson (Alaska Center for Conservation Science) for sending us bee bowls, for providing the Alaska Bee Atlas hexagons as vector data, and for sharing helpful advice.
USFWS Kenai National Wildlife Refuge, Soldotna, Alaska, matt_bowser@fws.gov↩︎
USFWS Kenai National Wildlife Refuge, Soldotna, Alaska↩︎
https://arctos.database.museum/search.cfm?project_id=10003917↩︎
https://www.gbif.org/dataset/86875091-d166-4986-802a-343b341424c6/event/12127687↩︎
https://www.gbif.org/occurrence/4093716677, https://www.gbif.org/occurrence/4093716136, https://www.gbif.org/occurrence/4093717282↩︎
https://www.gbif.org/occurrence/4093718979, https://www.gbif.org/occurrence/4093718205, https://www.gbif.org/occurrence/4093716762↩︎