Decline of bumble bees (Bombus) in the North American Midwest

Abstract

Declines in many bumble bee species have been documented in Europe raising several ecological and economic concerns. The nature and extent of bumble bee decline in North America is poorly understood due mainly to a lack of baseline and long-term data. Museum collections provide excellent sources of information on past and current species distributions, which can be used to infer changes in the composition of insect communities. Using the Illinois Natural History Survey’s electronic database of Hymenoptera and a recent biodiversity survey of historically sampled localities, we were able to examine changes in the richness and distribution of the bumble bee fauna of Illinois over the last century. We found that bumble bee species richness declined substantially during the middle of the century (1940–1960). Four species were locally extirpated: Bombus borealis, Bombus ternarius, Bombus terricola and Bombus variabilis. The ranges of Bombus affinis, Bombus fraternus, Bombus pensylvanicus and Bombus vagans have also decreased in Illinois. Our analyses also indicated that current bumble bee diversity is highest in northern Illinois, where conservation efforts would be most productive. Our study demonstrates that half of the bumble bee species found historically in Illinois have been locally extirpated or have suffered declines, supporting observations of broader declines in North America. Major declines in the bumble bee fauna coincided with large-scale agricultural intensification in Illinois between 1940 and 1960. Attempts to conserve bumble bees in Illinois should involve wildlife-friendly approaches to agriculture, such as increasing agricultural land set-asides and hedgerows, and employing integrated pest management.

Introduction

Bumble bees (Bombus) provide the vital ecosystem service of pollination in both natural and managed systems (Corbet et al., 1991, Kevan, 1991, Buchmann and Nabhan, 1996, Goulson, 2003, Memmott et al., 2004, Pywell et al., 2006, Goulson et al., 2008), and declines in their abundance and distribution have serious ecological and economic ramifications (Corbet et al., 1991, Biesmeijer et al., 2006, Carvell et al., 2006). Numerous studies from Europe have documented recent declines in many species of Bombus (Williams, 1982, Williams, 1986, Carvell, 2002, Biesmeijer et al., 2006, Goulson, 2006, Rasmont et al., 2006, Fitzpatrick et al., 2007, Kosior et al., 2007), and such declines were often observed in areas where anthropogenic changes in habitats have occurred, such as agricultural intensification and urbanization (Pywell et al., 2006).

In North America, little is known about the status of native bumble bee species, hampering conservation efforts to protect these valuable pollinators (Committee on the Status of Pollinators in North America, 2007). Over the last decade, various North American entomologists have observed (some anecdotally) that some species of once-common bumble bees have declined in abundance and distribution (Thorp, 2005, Thorp and Shepherd, 2005, Colla and Packer, 2008). This prompted the Xerces Society for Insect Conservation to add several presumably declining bumble bees to their Red List of Pollinator Insects of North America (Thorp, 2005, Thorp and Shepherd, 2005). These species belong to a single subgenus (Bombus sensu stricto Latreille) (Williams, 1998, Cameron et al., 2007) and include the eastern Bombus affinis Cresson and Bombus terricola Kirby, and western Bombus franklini Frison and Bombus occidentalis Greene. In addition, the Committee on the Status of Pollinators in North America, (2007) also added two species from the subgenus Thoracobombus Dalla Torre (Williams et al., 2008), the eastern Bombus pensylvanicus (DeGeer) and western Bombus sonorus Say, to their list of bee species in decline. Potential causes cited for the decline of North American bumble bees include land-use changes in the form of urbanization and agricultural conversion (Kremen et al., 2002a, Kremen et al., 2002b, Greenleaf and Kremen, 2006, McFrederick and LeBuhn, 2006), extensive pesticide use (Gels et al., 2002, Marletto et al., 2003), and pathogen spillover from commercial bumble bee colonies that contain many parasites (Colla et al., 2006, Otterstatter and Thomson, 2008). A recent study has substantiated the decline in B. affinis in eastern North America, and documented declines in the diversity of the bumble bee fauna during the past 35 years over a small spatial scale (approximately 100 km²) in southern Ontario, Canada (Colla and Packer, 2008). More data, especially over larger spatial and temporal scales, are needed to substantiate a decline of bumble bees in North America.

In Illinois, intensive farming and urban development over the last century have resulted in the loss of most of the state’s once vast prairie, forest and wetland habitats (Iverson, 1988, Wang and Moskovits, 2001, Duram et al., 2004). As a result, Illinois ranks second to last of American states in the percentage of natural areas surviving (Jeffords et al., 1995); and nearly 95% of the northern two-thirds of the land area of Illinois is currently used for agricultural purposes (DeWalt et al., 2005). Given our knowledge of land-use changes and bumble bee decline in Europe, it would be expected that intensive farming and urban development in Illinois would have had a negative impact on the bumble bee fauna as a result of the loss of nesting habitat and continuous pollen and nectar sources. A study conducted in Iowa (Hines and Hendrix, 2005), a state with similar land-use changes to Illinois, found bumble bee diversity in prairie remnants to be strongly influenced by the availability of resources in the surrounding area, particularly in grasslands, which provide abundant pollen sources and nesting habitat for bumble bees (Svennson et al., 2000). Hines and Hendrix (2005) suggest that the resources that grassland habitats provide may be limited in agricultural landscapes as a result of mechanical disturbances over large areas. Bumble bees are susceptible to habitat loss due to their relatively limited flight range (Brian, 1954, Walther-Hellwig and Frankl, 2000), long colony cycle (commonly several months), and specific food and nesting requirements (Alford, 1975, Sakagami, 1976, Richards, 1978). Furthermore, bumble bees, like many other Hymenoptera, have complementary sex determination (van Wilgenburg et al., 2006, Heimpel and de Boer, 2008) which, in small populations, leads to the production of costly diploid males instead of females and this can substantially increase the risk of population extinction (Zayed, 2004, Zayed et al., 2004, Zayed and Packer, 2005).

Documenting the extent and possible causes of species declines requires both historical and current distribution and species richness data. Museum collections provide an excellent source of historical data, but most collections are not electronically catalogued, thus extracting specimen data is extremely time-consuming (Suarez and Tsutsui, 2004). The Illinois Natural History Survey (INHS, Champaign, Illinois) has captured its entire Hymenoptera (ants, bees and wasps) collection (approximately 360,000 specimens) into an electronic database and made them publicly available. A major strength of the Hymenoptera database is its bumble bee database, providing an invaluable source of information on the historical distribution of Bombus in Illinois. In this study, we use the state of Illinois as a model for assessing bumble bee decline over the last century across a state-wide scale (145,934 km²). Illinois is ideal for such a study because it exemplifies a worst case scenario for loss of native habitat to farming and urban development. Further, Illinois has one of the best North American historical records of bumble bee diversity and distribution, dating back to the nineteenth century (e.g., Frison, 1919, Frison, 1921, Frison, 1923, Frison, 1926, Robertson, 1928, Waldbauer et al., 1977). These two factors allow us to examine whether and how the bumble bee fauna has changed in response to changes in habitat over the last century. We achieve this by integrating historical distributions and species richness from the INHS bumble bee database with current bumble bee distribution and species richness data obtained from our recent biodiversity survey.