Forest habitats and plant communities strongly predicts Megachilidae bee biodiversity

Introduction

Insect pollinators provide key ecosystem services; over 86% of all flowering plant species depend on insects for pollination (Klein et al., 2006; Ollerton, Winfree & Tarrant, 2011), and bees alone pollinate roughly one-third of the world’s staple food crops (Potts et al., 2010). Reported declines in bee species abundance may affect both economically important crop yield and wild plant populations worldwide (Brown & Paxton, 2009; Potts et al., 2010; Sahli & Conner, 2006). Bee declines can be attributed to habitat loss, pesticides, invasive species, and climate change (Klein et al., 2007; Memmott et al., 2010; Potts et al., 2016). Despite the importance of bees, most bee species distributions are unknown, and even less is known about the factors that drive their distributions.

Bees are highly diverse and have a wide range of life history strategies including sociality, nesting strategies, parasitism and diets (Michener, 2007; Westrich, 2007; Danforth et al., 2019; Mikát, Matoušková & Straka, 2021; Bossert et al., 2022). Many of these traits have evolved and/or been lost over evolutionary history and geographic space, contributing to the diversification of bees (Danforth et al., 2013; Hedtke, Patiny & Danforth, 2013). Climate, seasonality and shifting plants have all contributed to species and trait diversification for this group (Toussaint et al., 2012; Kergoat et al., 2018). One unifying characteristic of bees, however, is that bees have strong biotic interactions, such as relying on pollen and nectar from plants for food (Bartomeus et al., 2013). Despite these strong ties, floral richness has not been well associated with bee species richness globally; the tropics are a hot bed of diversity for plant species but have the lowest bee species diversity (Michener, 2007; Orr et al., 2020). This suggests that other life history traits are strongly influencing geographic distribution for this taxa.

Megachilidae is globally distributed and comprises 35% of the known 4,000 bee species (Michener, 2007) in the United States. Megachilid bees have important roles in pollination of a diverse range of ecosystems, including natural, urban, and agricultural environments (Gonzalez et al., 2012). Osmia and Megachile are successfully managed for crop and wildflower pollination; and Osmia lignaria is the key pollinator reared for apple, cherries, and other fruit trees Megachile rotundata is reared as the primary pollinator of alfalfa (Bosch & Kemp, 2002; Boyle & Pitts-Singer, 2019; Pitts-Singer et al., 2018; Pitts-Singer & Cane, 2011; Sampson & Cane, 2000).

Many megachilid bees nest in pre-existing cavities in decaying wood, hollow plant stems, or hollow twigs, and use a range of materials to create their nest cells, including masticated leaves, mud, plant resins, or flower petals (Bosch & Kemp, 2002; Litman et al., 2011; Young et al., 2016). The wide use of foreign materials for nest building is a likely reason for evolutionary range expansion and high diversification in Megachilidae, (Gonzalez et al., 2012; Litman et al., 2011). About 75% of the genera for this family have above ground/cavity-nesting traits (Eickwort, Matthews & Carpenter, 1981), however primitive forms of nesting (i.e., ground borrowing) still exist in eight of the 19 genera (Bosch, Maeta & Rust, 2001). It is likely that the above ground nesting behavior was repeatedly evolved and lost numerous times throughout the evolution of this group, since very closely related species can have drastically different nesting behaviors (Sedivy, Dorn & Müller, 2013).

Bees’ habitat, shelter, and food may all be impacted by climate change events (Corbet et al., 1993; Tews et al., 2006; Bartomeus et al., 2011; McCabe, Aslan & Cobb, 2022). When assessing how climate change and other anthropogenic disturbances will impact these species, research should not just look at how increasing temperatures will directly affect bees. Understanding where species are distributed can help to understand what the most limiting resources are for the success of a population.

Recent aggregation of databases and occurrence records have made it possible to assess spatial and temporal questions about insect species diversity (Cavender-Bares et al., 2009). This data can provide insight on how species and communities will respond to future disturbance (i.e., climate, habitat loss, etc.), but it is first critical to understand what is driving patterns of biodiversity and how these species respond to different environmental parameters (Weiher et al., 2011).

While it is widely accepted that Megachilidae have strong ties to biotic associations (e.g., host plants and cavity nesting resources: trees and pithy stems) in North America (Martins & de Almeida, 1994; Michener, 2007; Bosch, Maeta & Rust, 2001; Cane, Griswold & Parker, 2007), no study has systematically assessed the relationship between species richness and nesting substrate. We reviewed the biogeographic patterns of United States Megachilidae species, with two primary questions. (1) How can biogeographic patterns be explained by climate, nesting resources, and floral resources? (2) Are there specific relationships between Megachilidae biodiversity hotspots and taxonomic and/or morphological structure of woody plants? We predicted that for Megachilidae in particular they would be distributed in areas where cavity-nesting resources were also abundant, and in areas with high plant diversity.

Methods

Results

There was one hotspot of biodiversity identified in the United States for Megachilidae bees. The southwestern United States by far had the greatest species richness over any other area in the US (Fig. 1). Average richness in the southwest was 43.5 +/− 26.12 species, while in the southeastern US it was 19.03 +/− 8.64 species, northeast was 19.36 +/− 8.12, Midwest was 16.81 +/− 7.1 species and northwest was 21.59 +/− 12.43 species per 60 × 60 km area. The greatest species richness was in southern California near Los Angeles with 253 species for the 60 × 60 km area. Higher elevation mountain regions, including the Sierra Nevada Mountain Range, the Colorado Plateau and the Wasatch Mountains were the most diverse areas within the western region and overall United States.

Richness was predicted by three factors: land cover (i.e., tee cover), climate, and flowering plants (Table 2). Land cover type was the most correlated environmental factor in explaining bee richness (f = 28.57 p < 0.001). Areas that were classified by forested habitats (trees) had twice as much diversity than any other habitat with 44.75 +/− 25.53 species per given pixel (Fig. 2). Shrub and herb dominated habitats had the second highest number of species rich pixels with roughly 25.03 +/− 17.56 and 27.10 +/− 23.48 average species per pixel, respectively. Herb and shrub habitats were not significantly different from one another (p = 0.185). Agriculture, developed, barren and sparse habitats had averages between 10–15 species per pixel; none of these land cover types were significantly difference from one another (Table S2). Despite land cover type being highly correlated with Megachilidae richness, overall tree diversity did not have a significant effect in predicting Megachilidae richness (z = 1.321, p = 0.187).

The southwestern United States drove the pattern for having the most species in treed/forested habitats. In the midwest, the northwest and southeast none of the landcover types were different from one another (Table S2, Fig. 2). All habitats had an equal number of species per pixel. However, the southwestern United States had at least double the species in treed habitats than it did in agriculture, barren, developed, and shrub habitat (F = 30.702, p < 0.001, Fig. 2). Additionally, shrub and herbaceous habitats had more richness than agriculture, barren and developed lands.

All eight climatic variables selected showed a significant effect in predicting richness except for maximum temperature of the warmest quarter (z = 0.883, p = 0.337). Mean temperature of the driest quarter explained 4x as much variance than any other climatic variable (z = 23.15, p < 0.001). While mean temperature of the warmest quarter and precipitation seasonality were significant in the model, they explained less than 1% of the variance combined (z = −11.33, p < 0.001 & z = 8.774, p < 0.001 respectively). Herbaceous flowering plant diversity strongly predicted Megachilidae richness; as flowering plant diversity increased, so did Megachilidae diversity (z = 4.742, p < 0.001). Flowering plant richness also peaked in the southwestern United States. We did not examine specific associations for this analysis.

Discussion

Our results indicate that Megachilidae biodiversity has its hotspot in the southwestern United States, this is consistent with geographic patterns found by Michener (2007). Additionally, this is true for many bee species outside of Megachilidae, as well, since the southwestern United States is also a hotspot for bee biodiversity on a global scale (Griswold et al., 2018; Orr et al., 2020). The southwest is typically associated with dry air climates that are more suitable for ground nesting bees and thus hypothesized to harbor the greatest bee diversity globally, with around 80% of those being ground nesters (Danforth et al., 2019). However for Megachilidae, where 70% of the genera are cavity nesters, the same environmental traits may not be as crucial (Cane, Griswold & Parker, 2007). Importantly, the majority of cavity nesting species in tribe Osmiini are concentrated in North America (Praz et al., 2008), while many highly related Megachile species in the Palearctic still nest in the ground or have diversified nesting strategies (Danforth et al., 2004; Patiny, Michez & Danforth, 2008). It is hypothesized that cavity nesters were better able to survive trips across long water barriers, potentially explaining why it is mostly these Osmiini species inhabiting the new world (Michener, 2007). Therefore, some of the climate preferences and/or floral dependencies may be carried over from this lineage of soil nesting bees.

What makes United States Megachilidae so unique, however, is their strong ties to biotic nesting resources. Contrary to what Orr et al. (2020) found for all bee species, here we show that Megachilidae biodiversity was highly correlated with forested habitat, particularly in the southwestern United States, suggesting that these areas of more heterogeneous landscapes could be driving species diversification for this group (Hedtke, Patiny & Danforth, 2013). Groups such as Colletidae have shown stronger correlations with floral resources and climate than they have nesting requirements (Bystriakova et al., 2018). Other groups like Bombus also seem to be more strongly tied to floral resources for driving biogeographical patterns (Williams, Lonsdorf & Ward, 2014). Our study is not the first to show that bees may be restricted in distribution due to nesting resources or habitats (Gordon, 2000; Romero, Ornosa & Vargas, 2020). Additionally, traditional range models only factor in climate analyses however it is likely that landscape/ habitat play a large role in determining species geographic ranges (Graham & MacLean, 2018; Palma et al., 2016). McCabe et al. (2020) found that, along an elevation gradient in the southwest US Megachilidae were distributed in higher cooler elevations. These higher elevation environments also correspond with habitats that are forested. This pattern has only been documented in one other bee group, Bombus (Williams et al., 2014). However, Bombus is a social ground nesting species and has much lower diversity than the Megachilidae family.

Megachilidae showed a strong preference to floral diversity, as would be expected. Increased plant richness and diversity in an area, which likely also equates to increased floral abundance, can lead to greater floral resource availability (Blaauw & Isaacs, 2014; Crone & Williams, 2016). However, it was impossible to link specific flora types with drivers of Megachilidae biodiversity since many host plant associations are unknown. Even where host plant specialization has been well studied, it is unclear how the distribution of those plants plays a role in the bee species distributions (Michener, 2007). Although there are limitations for plant-insect interactions, open forest habitat with understory vegetation or open meadows with floral resources promotes bee diversity (Mola et al., 2021). Likewise, areas of increasing canopy cover in the southwest have been shown to decrease bee species richness, but Megachilidae species richness increases (McCabe et al., 2019).

Nationwide, climate analysis showed a trend towards warmer and drier environments. Megachilidae richness increased with increasing mean annual temperatures and decreasing mean annual precipitation, which is consistent with trends found in other bee species (Michener, 2007; Orr et al., 2020). However, analyses with the more nuanced climate variables revealed even more informative patterns for Megachilid species. Mean temperature of the wettest quarter was important in predicting richness; as mean temperatures decreased, Megachilidae richness increased. This is likely because species in the Megachilidae family often require a minimum number of cold days in order to complete diapause (Bosch, Sgolastra & Kemp, 2010) and the number of cool days during overwintering can directly influence their emergence timing (Sgolastra & Kemp, 2010). Unfortunately, future warming will likely change the distribution patterns of many Megachilidae species. McCabe, Aslan & Cobb (2022) found that for a community of cavity nesting species in Arizona, warming overwintering temperatures negatively affected emergence rates. Although the species were able withstand cooler temperatures in higher elevations, there were no nesting resources available in those environments.

Megachilidae may have more restrictions when adapting to climate change due to their associations with cavity nest availability (e.g., dead and down wood, pithy stems, etc.) and floral resources availability (Cane, Griswold & Parker, 2007). This dependency on nesting cavities adds an additional level of complexity when asking the question of how they will respond to climate change. In fact, it is estimated that 98% of Megachilidae depend on some type of non-floral, biotic resource (Requier & Leonhardt, 2020). Climate change is likely to only exacerbate the local levels of tree die-off in forested habitat, and in some instances, severe drought events can lead to over 90% tree die-off (Breshears et al., 2005). Additionally, an increase in wildfire frequency can lead to complete change in forest stands (Flannigan, Stocks & Wotton, 2000). It is unknown how these changes in forest structure will lead to subsequent changes in pollinator communities, especially within Megachilidae. Although little is known about climate change and soil composition, it is unlikely that ground nesting bees will have the same dramatic loss of nesting habitat under future climate scenarios (Allen, Singh & Dalal, 2011; Rinot et al., 2019). This uncertainty furthers the need to quantify the amount of nesting resources needed for Megachilidae within local communities.

Understanding what factors influence the distribution and biogeography of this taxa is important for future conservation efforts. With megachilids in particular, it may not only be the direct effect of warming temperature that causes shifts in species distributions. These species have strong ties to their host plants which could cause phenological mismatch in the future if the bees and plants are not tracking climate in the same manner (Bartomeus et al., 2018; Biesmeijer et al., 2006). Future conservation efforts for this family should consider altogether the direct effects of warming temperature, availability of floral resources, and availability of nesting resources.

Supplemental Information