Missing jewels: the decline of a wood-nesting forest bee, Augochlora pura (Hymenoptera: Halictidae), in northern Georgia

1 Introduction

Evidence that native bees and other pollinating insects are in decline is mixed. While some studies report significant reductions in diversity (Barendregt et al. 2022; Forister et al. 2021; Grixti et al. 2009), others suggest no decline (Herrera 2019) or reach contrasting conclusions depending on how data are analyzed and on what length of time is considered (Aldercotte et al. 2022). Some of the strongest evidence comes from population trends of individual species, such as documented declines and even extinctions of certain well-studied bumble bee species over the past several decades (Cameron et al. 2011; Colla and Packer 2008). By contrast, populations of other species appear to be comparatively stable (Cameron et al. 2011; Roubik et al. 2021). Because bees vary widely in their resource requirements, seasonality, habitat associations, and other factors, it is not surprising that some are more sensitive to environmental stressors than others. Early indications that a species is in decline can help identify ongoing or emerging threats. However, there is a shortage of monitoring studies of sufficient duration to distinguish true population declines from normal inter-annual fluctuations (Aldercotte et al. 2022). Although recently initiated long-term monitoring campaigns will eventually provide much-needed information on the status of insect populations (Woodard et al. 2020), evaluations of existing data in the meantime can be informative. Even the impressions of scientists and other long-term observers of nature can be useful, though imperfect, early indicators of community-wide or species-specific changes (Janzen and Hallwachs 2021).

Semi-natural areas are understood to provide crucial refugia for numerous organisms in mixed-use landscapes, including pollinators. However, even some of these relatively undisturbed habitats are experiencing pollinator declines. For example, we recently documented significant declines in both bees and butterflies over 15 years at long-term monitoring sites in a National Forest in Georgia, U.S., even though the sampled forests experienced no major anthropogenic disturbances over that period (Ulyshen and Horn 2023). Interestingly, stronger declines were detected for above-ground nesting bees than for those that nest below ground. The current study was prompted by our observation that one of the most familiar forest bee species in our region, Augochlora pura (Say) (Hymenoptera: Halictidae), has become noticeably less common in recent trapping efforts compared to past studies (Ulyshen et al. 2010). To investigate this further, we conducted another year of sampling at the long-term monitoring plots mentioned above and explored changes in the abundance of A. pura relative to other bees over time based on past studies.

Augochlora pura is a strictly forest-dependent bee found throughout the deciduous forests of eastern North America. Although A. pura will visit flowers in non-forested habitats, it depends on hardwood forests for decomposing logs and stumps in which to nest and overwinter (Stockhammer 1966). It nests in abandoned insect galleries or other available openings in wood that has become soft from decay. Nest entrances are left open when females are foraging but are closed with a plug of sawdust when the bee returns for the night (Stockhammer 1966). While A. pura most commonly nests and overwinters within moist wood in close contact with the soil, it also has been observed nesting in dead limbs within the forest canopy (Urban-Mead et al. 2021). Although the dependence of A. pura on decomposing wood is shared by only a few other halictid species native to eastern North America, the species shares this resource with hundreds of other saproxylic invertebrates that also depend upon it for food, moisture, or shelter (Ulyshen 2018). Here we present evidence that A. pura populations have declined significantly in our study region. We also propose the hypothesis that an invasive species of wood-nesting ant, Brachyponera chinensis (Emery) (i.e., the Asian needle ant) (Hymenoptera: Formicidae), is driving the decline of this species.

2 Materials and methods

To document local changes in A. pura abundance, we synthesized data from nine separate bee sampling studies conducted over the past 20 years (2005–2024) in forests of Clarke County, Georgia (near the city of Athens), or about 30 km to the southeast on the Oconee National Forest in Oglethorpe and Greene counties (Table 1).

One of these studies involved 18 years (2007–2024) of sampling at three National Forest locations that originally served as reference sites in a Chinese privet eradication study (Ulyshen et al. 2022) but are now used for the long-term monitoring of pollinator populations (Ulyshen and Horn 2023). The same pan traps were used throughout the study and the forests experienced no noticeable changes in structure or composition over the entire study period. We established five subplots at each location and used a pair of colored bowls (blue and yellow) filled with soapy water to sample pollinators at ground level in each subplot for one week per month from March to October (September excluded). The 10 bowls per location were combined at the time of collection. We used the lme4 package (Bates et al. 2015) in R to test how the abundance of A. pura changed with time at these locations. We used a negative binomial generalized linear model (GLMM) with year (scaled) as the only fixed effect and location as a random term. A likelihood ratio test was performed to compare this model with a negative binomial model that did not include location as a random term.

All nine studies were used to describe how the abundance of A. pura has changed over time relative to all bees collected. Although initially designed to test different hypotheses, all of these studies involved sampling throughout the growing season within mature, closed-canopy hardwood forests. The studies either employed white bucket traps with plexiglass panels attached (WBT, see Traylor et al. (2024), blue vane traps (BVT, see Edelkind-Vealey et al. 2024), or pan traps (PT) consisting of colored plastic bowls (Ulyshen and Horn 2023) or cups coated with fluorescent paint (Fortuin and Gandhi 2021). Although WBT and BVT were consistent across studies, PT varied among studies in terms of size and color. Trap height also varied among studies, ranging from ≤1 m (ground) to 5–24 m (canopy). Although A. pura is usually more common in canopy traps than near the ground (Ulyshen et al. 2010), here we classify all traps ≥5 m as canopy traps as one study showed the abundance of A. pura to be similar between 5 and 15 m (Ulyshen et al. 2020a). The sampled forests also differed considerably, ranging from extensive floodplain forests adjacent to a major river to urban forests near smaller streams. Because trap type, height, and local forest conditions are all likely to influence the number of bees captured, it is not possible to isolate changes in A. pura numbers over time from these other parameters. We therefore did not attempt to analyze changes in the relative abundance of A. pura across studies. However, we can use these nine datasets to get a general sense of how the relative abundance of A. pura has changed over time. To do so, we calculated the proportion of bees consisting of A. pura in the different studies and made these calculations separately for canopy and ground traps.

3 Results

The number of A. pura collected declined by nearly 70 % over 18 years of consistent sampling at our three long-term monitoring sites, a significant negative relationship in models with (estimate = −0.45, z = −2.92, p < 0.01) and without (estimate = −0.45, z = −2.88, p < 0.01) location included as a random effect (Figure 1A). The likelihood ratio test showed no significant difference in fit between these two models (χ ² = 0.21, p = 0.65).

Figure 1:

Changes in the number of Augochlora pura captured in pan traps over eighteen years (2007–2024) at three sites in Oconee National Forest, Georgia (A) and observed changes in the number of A. pura as a proportion of all bees captured over twenty years (2005–2024) near Athens, Georgia (B). The latter data come from nine studies using one of three trap types (WBT = white bucket traps, PT = pan traps, and BVT = blue vane traps) placed in the canopy (≥5 m, connected with a solid line) or near the ground (≤1 m, connected with a dashed line) (see Table 1 for more details).

The relative abundance of A. pura varied greatly among the nine studies (Table 1, Figure 1B) depending on trap type, height, and study location. Despite this variability, substantial reductions in the relative abundance of A. pura over time are apparent. Since 2020, A. pura has accounted for a particularly low proportion of bee individuals, although this pattern reversed somewhat in 2024.

4 Discussion

As reported previously, total bee richness and abundance declined significantly at our long-term sampling locations between 2007 and 2022 (Ulyshen and Horn 2023). That previous analysis found stronger declines among above-ground-nesting bees than those that nest below ground. These declines were somewhat surprising given that all three locations remained essentially undisturbed over the entire study period. The only significant changes known to us include significant increases in mean annual minimum temperature and invasion by B. chinensis, a non-native species of ant from Asia (Ulyshen and Horn 2023). Because the native range of A. pura extends far into Florida where average temperatures are naturally higher, we feel invasion by B. chinensis poses the greater threat to A. pura and other wood-nesting bee species.

Brachyponera chinensis was first detected in North America in 1932 (Smith 1934) and has since spread widely across the eastern US from Florida to Connecticut and west to Arkansas with a disjunct population introduced to Washington. It is not clear when the species first reached our study area, but we do know that it was present in the Oconee National Forest by 2011 (Vogt et al. 2022). Repeated surveys of thirty-six 2,000 m² plots within a nature preserve in Clarke County Georgia found the incidence of B. chinensis to increase from 8.3 % in 2011 to 91.6 % in 2022 (Doug Booher, unpublished data). Similarly, in another Clarke County survey from 2020, B. chinensis was detected at 31 out of 40 widely-separated forested sites (Clayton Traylor, unpublished data). Unlike most invasive ants, B. chinensis readily colonizes undisturbed forests in addition to anthropogenic habitats (Guénard and Dunn 2010; Nelder et al. 2006) and can be found nesting or foraging within practically every piece of dead wood encountered in invaded forests (Doug Booher, personal observation). It is known to feed on termites and other invertebrates encountered in dead wood and on the forest floor (Bednar and Silverman 2011).

A wide variety of invertebrates inhabit dead wood in forests, and interactions among species remain poorly studied. A co-occurrence analysis of ant and beetle species inhabiting decomposing pine logs in Mississippi suggests that most associations between these taxa are neutral or positive, with only a small fraction of interactions being negative (Ulyshen et al. 2020b). However, B. chinensis was not part of that study and the impact of that species on the abundance and diversity of saproxylic insects in dead wood remains unknown. However, A. pura may be particularly vulnerable considering that their nests are left open and exposed while females are out foraging (Stockhammer 1966). Our local observations suggest a low chance of such nests going undiscovered given the ubiquity of B. chinensis in dead wood.

The spread of B. chinensis throughout southeastern U.S. forests may represent yet another threat to native bee populations, particularly those that nest within dead wood. Although this hypothesis remains untested, ants have been shown to negatively impact bee populations in previous studies. For instance, Plentovich et al. (2021) reported lower reproductive success by a Hylaeus species (Hymenoptera: Colletidae) due to predation by a species of invasive ant in Hawaii. Studies specifically designed to assess the impact of B. chinensis on A. pura and other wood-nesting bees are needed.

If B. chinensis is indeed contributing to the decline of A. pura and perhaps other bee species, research on the spatial distribution of nests and foragers may help develop conservation strategies. For example, if B. chinensis primarily nests and forages near the ground, dead wood within the canopy may represent comparatively safe nesting substrates for A. pura. Urban-Mead et al. (2021) reported observing active Augochlorini nests (most likely A. pura) in dead oak branches about 25 m above the ground, highlighting the potential value of old trees as nesting habitat. However, recent reports of B. chinensis nesting in oak trees 9.5–18 m above the ground in North Carolina (Kirchner et al. 2025) suggest that A. pura is likely to encounter this species even high within the forest canopy.