Arthropod Abundance and Shrub Cover and Height on the Seward Peninsula, Alaska, 2015−2016

Study plots were located at one interior study site (n = 7 plots) and one coastal study site (n = 6 plots) and were each 600 m square. All plots were established in pairs (<4 km apart) that contrasted in dominance of tundra vegetation vs. taller shrubs, except for one added tundra-dominated plot at the interior study site. Each plot had 2−5 sampling points, depending upon heterogeneity of the vegetation, systematically separated 200 m apart. Invertebrates were collected at the sampling points using pitfall traps and/or sweep nets approximately weekly throughout each summer season (June-July 2015 and 2016). Mean dry biomass was measured on a subsample of the arthropods. Vegetation was measured at each sampling point one time during July of either 2015 (12 plots) or 2016 (1 plot).

A pitfall array was centered on each sampling point within a given plot. Each pitfall array consisted of three cups (7.6 cm diameter; 16.5 cm tall) placed in a line 1 m apart, buried flush with the soil surface, filled with 2 cm of propylene glycol, and covered with a plywood square 3–4 cm above the rim to protect the cup from rainwater. Contents of the three cups (subsamples) in each array were combined into a single sample for identification and enumeration. Samples at most pitfall arrays were collected every 7 days (some intervals were 5‒9 days) for a total of 7 weeks each summer (8 June–28 July 2015; 3 June–27 July 2016). In 2015, one additional weekly sample was collected (27–30 July 2015) but not analyzed for the associated publication (McDermott et al. 2021). Counts represent the total number of invertebrates collected in each 3-cup array on each sampling date.

Sweep-net transects consisted of 50 side-to-side sweeps over 25 m in a random direction from a given sampling point within a plot. Sweeps were aimed at the top 25% of vegetation within a 2 m swath along each transect. Sweep-net samples were collected at each sampling point every 6 to 8 days across 7 weeks each summer (8 June - 28 July 2015; 3 June - 27 July 2016) for a total of 8 samples per season. Weekly counts were then summed for a seasonal total for each point.

We characterized vegetation within five 0.5 m x 0.5 m quadrats, 5 m apart along one 20-m, randomly oriented transect that was centered on each invertebrate sampling point (where pitfall arrays were located and all sweep-net transects started). Shrub height and cover were measured separately for total shrubs (Families Betulaceae, Salicaceae, Ericaceae, Rosaceae), dwarf birch (Betula nana ssp. exilis, B. glandulosa), ericaceous shrubs (Family Ericaceae), short willow species (Salix arctica, S. fuscescens, S. phlebophylla, S. polaris, S. reticulata), and tall willow species (S. pulchra, S. alaxensis, S. glauca, S. richardsonii). Mean percent cover and mean modal height of shrubs were recorded within five 0.5 m x 0.5 m quadrats, spaced 5 m apart along one 20-m transect randomly oriented and centered on the point. Vegetation was measured during a 3-week period in mid to late July 2015 (12 original plots) or 2016 (1 added plot), after most tundra plants had reached peak greenness. Plant cover was estimated for each quadrat using modified Daubenmire cover classes (0, 1-5, 6-25, 26-50, 51-75, 76-95, or 96-100%), and the midpoint of the cover class was recorded (0, 2.5, 15, 37.5, 62.5, 85, or 97.5%). Shrubs from five families (Betulaceae, Salicaceae, Ericaceae, Rosaceae) were recorded during vegetation sampling. Individual plants were not identified to species. We recorded 'dwarf birch' (Betulaceae) for Betula nana exilis or B.glandulosa and 'Tall Willow' for Salix arctica, S. fuscescens, S. phlebophylla, S.polaris, S. reticulata. The designation 'Tall Willow' indicates that other willow species were present and were recorded as 'Short Willow', but our focus in the paper was on those willows that form tall shrubs and are implicated in shrub expansion. Short willows were not analyzed.

For pitfall samples or Acari and Collembola, each pitfall sample was placed in a gridded petri dish and agitated in alcohol to evenly distribute the specimens. The number of individuals in a grid cell (or row, if there were <20 individuals total) was counted, then multiplied by the total number of grid cells (or rows) to obtain the estimated total number of individuals in the sample.

To estimate arthropod biomass, we dried as many individuals as possible of each taxon of arthropods for 48 hours at 40°C, calculated mean individual dry mass, and multiplied this by counts of individuals to estimate biomass of each sample. When size was highly variable within a taxon, we separated individuals by length and obtained mean dry mass for each size class. Individuals within each taxon were counted and weighed separately by life stage (larva, nymph, adult), but life stages were combined for analyses of community composition. Minute taxa (e.g., Chalcidoidea, <2 mm) were not weighed as these represent a very small fraction of overall biomass. Arthropod voucher specimens are in the University of Alaska Museum of the North (accession number UAM-2014.24-USGS-Ento).