Available online at www.sciencedirect.com Landscape and Urban Planning 85 (2008) 123–132 Africanized honey bees in urban environments: A spatio-temporal analysis Kristen A. Baum a,∗ , Maria D. Tchakerian b , Steven C. Thoenes c , Robert N. Coulson b a b Department of Zoology, 430 Life Sciences West, Oklahoma State University, Stillwater, OK 74078, USA Knowledge Engineering Laboratory, Department of Entomology, Texas A&M University, College Station, TX 77843-2475, USA c BeeMaster, Inc., 11358 N. Mandarin Lane, Tucson, AZ 85737, USA Received 22 May 2007; received in revised form 30 July 2007; accepted 17 October 2007 Available online 4 December 2007 Abstract For honey bees in the desert southwest, urban environments may provide abundant cavities and a more spatially and temporally continuous supply of nectar, pollen and water than would be available in surrounding natural desert areas. The presence of abundant cavities and food resources in urban environments places honey bees in close proximity to humans, creating concerns over public health and safety, particularly in areas dominated by Africanized honey bees. Africanized honey bee colonies are abundant in the greater Tucson metropolitan area, and requests for colony and swarm removals increased from 14 in 1994 to 1613 in 2001. We obtained invoices with data on honey bee colony and swarm removals from 1994 to 2001 from a private company in Tucson, Arizona, which specializes in the removal and control of Africanized honey bees. We evaluated spatio-temporal patterns in the distribution of Africanized honey bee colonies and swarms and evaluated the role of precipitation in generating the observed patterns. Colonies and swarms showed a shift from no spatio-temporal clustering in the initial years following the arrival of Africanized honey bees to significant spatio-temporal clustering in later years. Precipitation was a good predictor of honey bee abundance, with more colony and swarm removals following wet seasons and fewer following dry seasons. These patterns suggest the greatest likelihood of human–honey bee interactions in urban areas in the desert southwest will occur with high honey bee abundances following wet winters. © 2007 Elsevier B.V. All rights reserved. Keywords: Apis mellifera; Feral colonies; Swarms; Cross-correlation; Mantel test; Sonoran desert 1. Introduction Urban environments provide suitable habitat for many organisms, including native and nonnative species. In some cases, urban environments may provide ideal habitat compared to surrounding natural areas, with some species thriving in cities, such as European starlings (Clergeau and Quenot, 2007) and mosquitoes (Leisnham et al., 2006). Species that thrive in urban environments are often widespread and may be considered pests because of their overlap in resource use with humans. Honey bee colonies nest in natural, wildlife- and human-made cavities, such as tree hollows and buildings, and will exploit urban sources of nectar, pollen and water. For honey bees in the desert southwest, urban environments may provide abunCorresponding author. Tel.: +1 405 744 7424; fax: +1 405 744 7824. E-mail addresses: kristen.baum@okstate.edu (K.A. Baum), mtchakerian@tamu.edu (M.D. Tchakerian), beemaster@dakotacom.net (S.C. Thoenes), r-coulson@tamu.edu (R.N. Coulson). ∗ 0169-2046/$ – see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.landurbplan.2007.10.005 dant cavities and a more spatially and temporally continuous supply of nectar, pollen and water than would be available in surrounding natural areas (Buchmann, 1996; Rabe et al., 2005; Shochat et al., 2006; Stuart et al., 2006). The presence of abundant cavities and food resources in urban environments places honey bees in close proximity to humans (Pereira and Chaud-Netto, 2005), creating concerns over public health and safety, particularly in areas dominated by Africanized honey bees (Schmidt and Boyer-Hassen, 1996; Johnston and Schmidt, 2001). Honey bees are not native to North America, but generally are considered an important component of many ecosystems because of the pollination services they provide (Allen-Wardell et al., 1998; Morse and Calderone, 2000). Feral and managed honey bee colonies extract large amounts of pollen and nectar from all habitats where they live and may compete for limited floral resources with native bees and other pollinating animals (Buchmann, 2000). Colonies reproduce by swarming when part of the colony leaves with the queen in search of a new nest site. 124 K.A. Baum et al. / Landscape and Urban Planning 85 (2008) 123–132 Thus, colonies occupy relatively permanent locations, whereas swarms occupy transient locations until they select a new nest site. Feral, unmanaged honey bee colonies can be found in many different settings, including urban/suburban (Morse et al., 1990; de Mello et al., 2003), forest/woodland (Galton, 1971; Kerr, 1974; Visscher and Seeley, 1982; Oldroyd et al., 1994; Coulson et al., 2005), agricultural (Ratnieks et al., 1991), coastal prairie (Baum et al., 2005), and semi-desert and desert habitats (Taber, 1979; Boreham and Roubik, 1987; Schneider and Blyther, 1988; McNally and Schneider, 1996; Loper et al., 2006). Few studies have evaluated spatial patterns in honey bee colony distributions (Oldroyd et al., 1995, 1997; McNally and Schneider, 1996) or spatial patterns through time (Baum et al., 2005). However, those studies that have evaluated spatial distributions suggest feral honey bee colonies tend to form aggregations (Oldroyd et al., 1995, 1997; McNally and Schneider, 1996; Baum et al., 2005). Possible explanations for colony aggregations include aggregated resources, short swarm dispersal distances, attraction of swarms to existing colonies, increased predator defenses and increased mating efficiency (Seeley and Morse, 1977, 1978; Jaycox and Parise, 1980, 1981; Seeley et al., 1982; Oldroyd et al., 1995; Schmidt, 1995; Baum et al., 2005). Africanized honey bees, hybrids between African honey bees (Apis mellifera scutellata) and European honey bees (Clarke et al., 2002; Pinto et al., 2005), first arrived in the United States from South America in 1990 (Hunter et al., 1993; Rubink et al., 1996; but see Pinto et al., 2007) and Arizona in 1993 (GuzmanNovoa and Page, 1994; Loper, 1997). Africanized honey bees are characterized by stronger defensive behavior, higher reproductive rates, smaller colony sizes, and less selectivity in nest sites than colonies of European origin (Winston et al., 1983; Winston, 1992; Schneider et al., 2004b). Africanized colonies will utilize smaller cavities than European colonies (Schmidt and Hurley, 1995), expanding the range of suitable nest sites to include flower pots, underground concrete water meter boxes, tires, cement blocks, garbage cans, etc., thereby increasing the proximity of Africanized honey bees to humans. In Africanized areas in the United States the feral honey bee population is predominantly Africanized (Pinto et al., 2004; Schneider et al., 2004b), including Arizona (Rabe et al., 2005; Harrison et al., 2006). Temperature and precipitation have been identified as potential factors contributing to the invasion pattern of Africanized honey bees in North America (Schneider et al., 2004b). For areas below the 34◦ N latitude line proposed as the potential northern limit for Africanized honey bees based on similar limits in South America (Taylor and Spivak, 1984), winter temperatures and winter survival are unlikely to limit their distribution. However, precipitation and associated patterns of floral resource availability could influence the distribution and abundance of honey bee colonies in the desert southwest. African colonies are adapted to seasonally arid habitats and may respond more strongly to precipitation patterns than photoperiod, which is linked to annual cycles in European honey bee colonies. This response to precipitation reflects the ability of Africanized honey bees to rapidly respond and exploit transient “flushes” in floral resource availability. Feral honey bee colonies are abundant in the greater Tucson metropolitan area, generating demand for pest control companies that specialize in removing honey bee colonies and swarms. Requests for honey bee colony and swarm removals increased from less than 20 in 1994 to more than 1,800 in 2001 (unpubl. data, S.C. Thoenes), representing a dramatic increase in removal requests with the arrival and establishment of Africanized honey bees in the area. We obtained invoices with data on honey bee colony and swarm removals from 1994 to 2001 from BeeMaster, Inc., a private company in Tucson, Arizona, which specializes in the removal and control of Africanized honey bees. The goal of this study was to evaluate spatio-temporal patterns in the distribution of feral honey bee colonies and swarms in the greater Tucson metropolitan area based on removal records. We predicted that colony and swarm removals would show a strong spatio-temporal correlation, with removed colonies and swarms located close together in space also occurring close together in time. We also evaluated the role of precipitation in generating the observed patterns of colony and swarm abundance based on the removal data. We predicted swarm abundance should be correlated with precipitation at lag times reflecting faster colony growth during years with early and/or abundant pollen production as indicated by fall/winter precipitation and precipitation during the summer monsoon period (July to mid-September). Lag times reflecting slower colony growth should occur during years with late and/or limited pollen production due to little precipitation during the fall and/or winter. Colonies should show similar correlations with precipitation as swarms because flight activity levels are high for colonies producing swarms and for newly founded colonies, and high activity levels would increase the likelihood of a colony being detected and removed. Identifying patterns in the spatio-temporal distribution of feral honey bee colonies in urban environments will provide information needed to develop strategies to control Africanized honey bees and minimize human–honey bee interactions in urban areas. 2. Methods 2.1. Study site The Tucson metropolitan area is located in Pima County in southeastern Arizona in the northern Sonoran Desert. The city of Tucson is a highly urbanized area with a population of 946,362 (U.S. Census Bureau, 2007). The climate is arid with an average annual rainfall of 29.7 cm. Most of the precipitation occurs during the summer monsoon period (July to mid-September), with additional precipitation in the winter (December to early March). April, May and June are often without rain. Maximum temperature regularly exceeds 38 ◦ C during the summer, with an average annual high of 27 ◦ C and a low of 12 ◦ C. The vegetation in the study area is characteristic of the Arizona Upland subdivision of the Sonoran Desert as described by Shreve (1951). Succulent cacti, drought tolerant shrubs and wildflowers are common plants found in this subdivision. The flowering patterns in the northern Sonoran Desert have been well documented with many species flowering when seasonal soil K.A. Baum et al. / Landscape and Urban Planning 85 (2008) 123–132 moisture is greatest (Bowers and Dimmitt, 1994). The Arizona Upland subdivision desert has two flowering periods: the first flowering season extends from mid-February to mid-June with a peak from early March to late April and the second flowering season begins near the end of the summer rains and continues into late fall (O’Neal and Waller, 1984; Dimmit, 2000). These two periods represent the major natural flushes of nectar and pollen availability for honey bees in the area. Within the greater Tucson metropolitan area, native vegetation is mostly intact in riparian, low-density housing and natural open space areas (Shaw et al., 1998; Livingston et al., 2003). However, rapid urbanization has fragmented patches of native vegetation with commercial and residential development (Cane et al., 2006), and much of the natural vegetation community has been replaced by exotic species in these areas. The presence of exotic species combined with irrigation and other landscaping practices have likely extended the flowering periods in the Tucson area beyond those encountered under natural conditions. Natural cavity sources in the area are probably relatively rare, with rock crevices being used by colonies in nearby natural areas (Taber, 1979; Loper et al., 2006). However, human-made sources of cavities are abundant in the greater Tucson metropolitan area, including openings in buildings, water meter boxes, tires, cement blocks, garbage cans, flower pots and a variety of other hollow spaces of adequate size. Therefore, the urban environment may represent ideal habitat for Africanized honey bees, with abundant cavities and a more spatially and temporally continu- 125 ous supply of nectar, pollen and water compared to surrounding natural desert areas. 2.2. Bee colony data Invoices with data on honey bee colony and swarm removals from 1994 to 2001 were obtained from BeeMaster, Inc., a private company in Tucson, Arizona, which specializes in the removal and control of Africanized honey bees. For each removed colony or swarm, recorded information included date, address, problem type (i.e., colony, swarm or individual honey bees), and specific location (i.e., building, tree, swarm trap, ground or miscellaneous). The miscellaneous category included flower pots, water meter boxes, tires, cement blocks, garbage cans, etc. Only data on colonies and swarms were included in the analyses. Colony and swarm removal records were assigned geographic coordinates by linking street addresses for each colony to line segments contained in the Pima County Department of Transportation street network coverage (1999) using a commercial GIS product (ArcGIS® v. 9.2; Fig. 1). An address was considered “matched” if the street address of the removal had the same number and street name as the street network coverage. When an exact match could not be made, interactive matching was used to refine misspelled addresses, directions, and street type errors. An “unmatched” address occurred if an address was not found in the street network coverage or Fig. 1. Shaded relief map of the greater Tucson metropolitan area with point locations for all colonies and swarms removed in 2001. 126 K.A. Baum et al. / Landscape and Urban Planning 85 (2008) 123–132 necessary information was lacking (no street number or name, P.O. Box listing, etc.) and these were excluded from all analyses. 2.3. Data analysis We used a Mantel test to evaluate whether or not honey bee colony and swarm removals located close together in space were also located close together in time. We conducted two separate analyses (one for colonies only and one for colonies and swarms) because colonies occupy relatively permanent locations and swarms occupy transient locations until they select a nest site. The Mantel test compares two distance matrices (for our study, distance in space between colony removal locations and distance in time between colony removal dates; Fortin and Dale, 2005). The units in space (m) and time (days) were not comparable, so we ranked the values in each data matrix prior to analysis (Dietz, 1983; Fortin and Dale, 2005). We used a randomization test with 100,000 permutations to evaluate significance, with significant values occurring when the Mantel statistic is more extreme than the reference distribution as ascertained using the Spearman rank correlation coefficient. We used zt to conduct the analysis (Bonnet and Van de Peer, 2002). For all analyses the significance level was set at α = 0.05. We used a cross-correlation analysis to evaluate if weekly precipitation was correlated with weekly swarm or colony removal numbers. Daily precipitation data were obtained from the National Climatic Data Center (http://www.ncdc.noaa.gov/ oa/ncdc.html, accessed 20 October 2006) using data inventories from a first order weather station located in Tucson, Arizona. We evaluated weekly lags from 0 to 50 and used 95% upper and lower confidence limits to evaluate significance. Crosscorrelation values range from 1 to −1, with 1 indicating strong positive correlation, 0 indicating no correlation and −1 indicating strong negative correlation. The cross-correlation analysis Table 1 Mantel test results for the detection of spatio-temporal clustering of honey bee colonies and honey bee colonies and swarms in the greater Tucson metropolitan area Year # colonies R p # colonies + swarms R p 1994 1995 1996 1997 1998 1999 2000 2001 14 323 474 445 1012 871 353 1035 −0.071 0.009 0.011 0.022 0.019 0.027 0.039 −0.029 0.376 0.286 0.187 0.096 0.061 0.036 0.021 0.009 14 458 665 617 1407 1225 525 1613 −0.071 0.004 0.009 0.021 0.017 0.025 0.046 −0.033 0.376 0.385 0.203 0.075 0.051 0.024 0.005 0.000 was performed using SPSS for Windows, Release Version 15.0 (©SPSS, Inc., 2006, Chicago, IL). 3. Results After excluding records without accurate spatial information, BeeMaster, Inc., removed 6, 524 colonies and swarms from the greater Tucson metropolitan area from 1994 through 2001 (Table 1). The number of colony and swarm removals increased from 14 in 1994 to 458 in 1995 following the arrival of Africanized honey bees in 1993. Removals increased 228% from 1997 to 1998 for colonies + swarms, decreased by 43% from 1999 to 2000 and increased by 307% from 2000 to 2001 (Table 1). A majority of the colonies were located in buildings (55%), followed by the miscellaneous category (27.9%), trees (10.1%) and in the ground (7%). Swarms were primarily located in trees (55%), followed by buildings (22.5%), the miscellaneous category (13%) and in the ground (9.5%). The annual temporal distribution of colony and swarm removals did not exhibit any strong trends, but some years (i.e., Table 2 Cross-correlation results testing for correlation between weekly swarm and colony removal numbers and weekly precipitation Year Swarms (% change) 1995 Precipitation (mm)a Correlation Swarms Colonies 245 Pos Neg ns ns ns ns 1996 +41 120 Pos Neg ns ns ns 0 1997 −10 138 Pos Neg 31, 32, 34, 35 ns 31, 32, 34 ns 1998 +230 187 Pos Neg 12 0 15 4 1999 −10 35 Pos Neg ns ns ns ns 2000 −49 35 Pos Neg 35, 36 ns ns ns 2001 +336 227 Pos Neg 24, 26, 27, 28 ns 28, 29, 30 ns Significant weekly lags are listed separately for lags with positive and negative correlation. The percent change in the number of swarms removed from the previous year and fall and winter precipitation is also provided. a Precipitation is the total amount during the fall of the previous year (beginning 22–23 September) and the winter of the current and previous years (ending 19–20 March). K.A. Baum et al. / Landscape and Urban Planning 85 (2008) 123–132 127 Fig. 2. Weekly precipitation (mm), number of colonies removed and number of swarms removed from 1995 through 2001. 1995, 1996, and 2001) showed a bimodal distribution with an early peak from March through May and a late peak in October and November (Fig. 2). Overall precipitation was highest in 1998 and lowest in 2001 (Fig. 2). Fall and winter precipitation was the lowest during 1998/1999 and 1999/2000 and highest during 1994/1995 and 2000/2001 (Table 2). Colonies and colonies + swarms showed a shift from no significant spatiotemporal clustering from 1995 through 1998 to significant spatio-temporal clustering from 1999 through 2001 (Table 1). Spatio-temporal clustering was positive in 1999 and 2000 and negative in 2001 for both colonies and colonies + swarms. The number of colonies removed was positively crosscorrelated with precipitation in 1997 (31, 32, and 34 week lags), 1998 (15 week lag) and 2001 (28–30 week lags) and negatively cross-correlated with precipitation in 1996 (0 week lag) and 1998 (4 week lag; Table 2, Fig. 3). The number of swarms removed was positively cross-correlated with precipitation in 1997 (31, 32, 34, and 35 week lags), 1998 (12 week lag), 2000 (35 and 36 week lags) and 2001 (24 and 26–28 week lags) and negatively cross-correlated with precipitation in 1998 (0 week lag; Table 2, Fig. 4). 4. Discussion Several factors could potentially bias the spatio-temporal interpretation of the honey bee colony and swarm removal data. First, socio-economic considerations, in particular the ability to pay for the removal service, could influence the completeness of the dataset. The detectability of colonies and associated requests for removals also could differ based on the location of the colony, proximity to human activity, frequency of human activity, etc. Tolerance of stinging insects and insects in general could influence the likelihood of someone choosing to have a colony removed once detected. These factors also could influence the removal options considered, such as an individual attempting to remove a colony or swarm on their own. However, given the extensive media coverage given to Africanized honey bees (aka “killer bees”), the bias of these factors is probably much less than would be expected for similar data obtained for other pest species. Honey bee colony densities based on colony and swarm removals ranged from 0.36 colonies/km2 (or 0.51 colonies and swarms) in 1995 to 1.12 colonies/km2 (or 1.80 colonies and swarms) in 2001. These densities fall within the range of previously reported densities for aggregations of honey bee colonies in other areas (Ratnieks et al., 1991; Oldroyd et al., 1994, 1997; McNally and Schneider, 1996; Baum et al., 2005), although only Morse et al. (1990) examined feral colonies in an urban area. This study represents the largest number of colonies and/or swarms reported for any published study, even considering the data represent removed (killed) colonies and swarms which no longer contribute to population growth (Table 1). From 1995 to 1998, colonies and colonies + swarms showed no significant spatio-temporal clustering. However, from 1999 to 128 K.A. Baum et al. / Landscape and Urban Planning 85 (2008) 123–132 Fig. 3. Cross-correlation coefficients for precipitation and colony removals for years with significant lags. Lags from 0 to 50 weeks are shown with 95% upper and lower confidence limits indicating significance. 2001, colonies and colonies + swarms showed significant spatiotemporal clustering (Table 1). The number of colony and swarm removals was relatively low from 1995 through 1997, with fewer than 700 removals per year, but more than doubled to 1407 removals in 1998. 1998 may represent a transition year where Africanized honey bees were becoming established in the area and the colonies and swarms were responding to the observed increase in population size, resulting in strong spatiotemporal clustering in the following years. Spatio-temporal clustering was positive in 1999 and 2000 and negative in 2001 (Table 1). 1999 and 2000 were both dry years with reduced swarming activity throughout the year except during the spring peak which typically occurs in April and May. 2001 was a very wet year with the largest peak in swarming activity recorded during this study (Table 2, Fig. 2). During dry years, swarm removals tended to be spatially and temporally restricted, with activity occurring within specific areas of Tucson. However, during wet years removals were broadly distributed in space and time (Schmidt and Edwards, 1998) and tended to occur throughout the greater Tucson metropolitan area. The significant negative spatio-temporal clustering in 2001 may reflect the 40 and 49% decline in the number of colonies and swarms removed in 2000, respectively, and the subsequent 293 and 336% increase in the number of colonies and swarms removed in 2001. Several potential explanations have been proposed for why honey bee colonies may be spatially aggregated, including dispersal behavior, resource distributions, predator defenses and mating efficiency (Seeley et al., 1982; Oldroyd et al., 1995). Existing studies on swarm dispersal distances have produced conflicting results, with some studies reporting short dispersal distances and others reporting longer dispersal distances (Seeley and Morse, 1977; Jaycox and Parise, 1980, 1981; Schmidt and Thoenes, 1990; Oldroyd et al., 1995; Schmidt, 1995; Schneider, 1995). Genetic differences among study populations may explain these conflicting results, such as if Africanized honey bee swarms travel longer distances than European swarms (Winston, 1987; Schneider, 1995). Differences in resource availability among study areas are another possible factor influencing swarm dispersal distances (Winston, 1987; Schneider, 1995). Several studies have suggested swarms may select nearby cavities when cavities are abundant (Jaycox and Parise, 1980, 1981; Seeley and Morse, 1977, 1978). Cavity availability is probably not limiting in our study area, especially considering the wide K.A. Baum et al. / Landscape and Urban Planning 85 (2008) 123–132 129 Fig. 4. Cross-correlation coefficients for precipitation and swarm removals for years with significant lags. Lags from 0 to 50 weeks are shown with 95% upper and lower confidence limits indicating significance. range of cavities utilized by Africanized honey bee colonies. Nectar and pollen sources are probably not limiting either, and may be more consistently available in urban environments than in surrounding natural areas (Schneider et al., 2004a). Aggregations could increase predator detection and defenses, but potential predators other than humans are rare in urban environments and aggregations may serve to increase human detection and removal. Aggregations also could serve to increase mating efficiency, but this potential explanation is difficult to evaluate without information on drone congregation areas or the relatedness of colonies. Thus, the mechanisms behind spatiotemporal clustering in the colony and swarm removal data is not clear, but likely reflects a combination of resource availability, honey bee biology and the likelihood of detection and removal. Spatio-temporal clustering suggests efforts to control and remove Africanized honey bees should be spatially and temporally focused in areas near where colonies and/or swarms are found. Precipitation is directly tied to pollen availability, with more rain resulting in more pollen. The timing of the precipitation also matters, with colony survival being positively correlated with winter rainfall (Loper et al., 2006). Heavy rainfall during October initiates flower production in mid-February and swarming in the spring. The summer monsoon period (July to mid-September) initiates flowering that begins a few weeks after the first summer rains and continues into late fall, leading to swarming activity in the fall. The development period from egg to adult worker is approximately 21 days and from egg to adult queen is approximately 18 days, with Africanized honey bees emerging one to 2 days earlier than European honey bees. Thus, peak swarming would be expected to occur approximately 6–8 weeks after the initiation of flowering in mid-February or 24–26 weeks after precipitation in October. Precipitation during the monsoon period generates a more immediate floral response within several weeks. Thus, monsoon rains in July and August would lead to swarming in October and November, with a lag of approximately 10–12 weeks. Any negative correlations would have very short lags because precipitation would suppress honey bee activity levels and also decrease the likelihood of detection and removal. Seasonal absconding also may contribute to the distribution and abundance of swarms and colonies. Seasonal absconding is the abandonment of a nest site due to low resource availability and the relocation of the colony to another nest site in an area with presumably higher resource availability (Winston et al., 1979; Schneider and McNally, 1992; Schneider et al., 2004a). 130 K.A. Baum et al. / Landscape and Urban Planning 85 (2008) 123–132 Seasonal absconding in the Tucson area is associated with low resource availability in the mountains surrounding Tucson in the fall and winter months and higher resource availability in the Tucson basin where agriculture and horticultural practices increase floral availability (Schneider et al., 2004a). The significant precipitation lags for swarms in 2000 (35–36 weeks) and 2001 (24 and 26–28 weeks) suggest pollen availability was low in 2000 and high in 2001 (Table 2, Fig. 4). When heavy rainfall occurs during October, flowering begins in mid-February and pollen becomes available to the honey bees. If precipitation is shifted later, pollen production would be delayed and swarming activity will exhibit a corresponding shift with longer lags reflecting lower or later pollen availability. Precipitation was extremely low in the fall/winter of 1999/2000, representing the driest winter during this study, whereas the fall/winter of 2000/2001 was one of the wettest (Table 2, Fig. 2). In 1997 significant precipitation lags for swarms occurred at 31, 32, 34 and 35 weeks, representing an intermediate year in terms of pollen availability (Table 2, Fig. 4). The precipitation lag for 1998 was short (12 weeks), but can be explained by 1998 representing an El Niño year and being the wettest year in terms of overall precipitation and number of days with rain (Table 2, Figs. 2 and 4). Colonies showed significant precipitation lags very similar to those for swarms, which is not surprising because colony activity levels are high during periods of pollen abundance and swarm production (Table 2, Fig. 3). Also, newly founded colonies would exhibit high levels of worker activity, which could increase the likelihood of detection depending on the cavity selected. Colonies and swarms in 1998 and colonies in 1996 showed short negative lags with precipitation, which would be expected based on immediate and temporary decrease in honey bee activity due to rainfall (Table 2, Fig. 3). Precipitation and associated changes in productivity also have been shown to influence the distribution and abundance of other species in the Sonoran Desert region, including vertebrates and other invertebrates (Marshal et al., 2002; Shochat et al., 2004). For example, spider abundance was five times higher following an El Niño winter (1997/1998) compared to a dry winter (1999/2000; Shochat et al., 2004). Overall, these data suggest precipitation may be a good predictor of honey bee abundance in the greater Tucson metropolitan area, with more colonies and swarms expected after wet winters. Significant spatio-temporal clustering in colony and swarm removals suggests control efforts should be concentrated in areas where colonies and swarms are found, although data for additional years could clarify these patterns given the significant negative spatio-temporal clustering observed in 2001. Information on additional factors, such as disease and parasite loads, could provide additional insights and further help target control efforts. For example, Loper et al. (2006) documented an initial drastic decline in colony survival in a feral honey bee population after the arrival of the Varroa mite in 1995. Their study area was located approximately 80 km northeast of Tucson, suggesting infestation by Varroa mites also could have influenced the patterns observed in this study. Furthermore, this study defines methods for studying pest species in urban environments using data from pest control companies. Address matching within the context of a GIS can be used to create point data which can be analyzed using spatial or spatio-temporal statistics. Acknowledgements A. Bunting, J. Yu and S. Kim provided valuable technical assistance throughout this project. S. Buchmann and three anonymous reviewers provided valuable comments on the manuscript. Funding for this project was provided by the Texas Legislative Initiative: Protection and Management of Honey Bees—Pollinators of Agricultural Crops, Orchards, and Natural Landscapes. References Allen-Wardell, G., Bernhardt, P., Bitner, R., Burquez, A., Buchmann, S., Cane, J., Cox, P.A., Dalton, V., Feinsinger, P., Ingram, M., Inouye, D., Jones, C.E., Kennedy, K., Kevan, P., Koopowitz, H., Medellin, R., Medellin-Morales, S., Nabhan, G.P., Pavlik, B., Tepedino, V., Torchio, P., Walker, S., 1998. The potential consequences of pollinator declines on the conservation of biodiversity and stability of food crop yields. Conserv. Biol. 12, 8–17. Baum, K.A., Rubink, W.L., Pinto, M.A., Coulson, R.N., 2005. Spatial and temporal distribution and nest site characteristics of feral honey bee (Hymenoptera: Apidae) colonies in a coastal prairie landscape. Environ. Entomol. 34, 610–618. Bonnet, E., Van de Peer, Y., 2002. zt: a software tool for simple and partial Mantel tests. J. Stat. Softw. 7, 1–12. Boreham, M.M., Roubik, D.W., 1987. Population changes and control of Africanized honey bees (Hymenoptera: Apidae) in the Panama Canal area. Bull. Entomol. Soc. Am. 33, 34–39. Bowers, J.E., Dimmitt, M.A., 1994. Flowering phenology of six woody plants in the northern Sonoran Desert. Bull. Torrey Bot. Club 121, 215–229. Buchmann, S.L., 1996. Competition between honey bees and native bees in the Sonoran Desert and global bee conservation issue. In: Matheson, A., Buchmann, S.L., O’Toole, C., Westrich, P., Williams, I.H. (Eds.), The Conservation of Bees. Academic Press, New York, pp. 125–142. Buchmann, S.L., 2000. Bees. In: Phillips, S.J., Comus, P.W. (Eds.), A Natural History of the Sonoran Desert. Arizona-Sonora Desert Museum Press, Tucson, and Univ. of California Press, Berkeley. Cane, J.H., Minckley, R.L., Kervin, L.J., Roulston, T.H., Williams, N.M., 2006. Complex responses within a desert bee guild (Hymenoptera: Apiformes) to urban habitat fragmentation. Ecol. Appl. 16, 623–644. Clarke, K.E., Rinderer, T.E., Franck, P., Quezada-Euán, J.G., Oldroyd, B.P., 2002. The Africanization of honeybees (Apis mellifera L.) of the Yucatan: a study of a massive hybridization event across time. Evolution 56, 1462–1474. Clergeau, P., Quenot, F., 2007. Roost selection flexibility of European starlings aids invasion of urban landscape. Landscape Urban Plann. 80, 56–62. Coulson, R.N., Pinto, M.A., Tchakerian, M.D., Baum, K.A., Rubink, W.L., Johnston, J.S., 2005. Feral honey bees in pine forest landscapes of east Texas. For. Ecol. Mange. 215, 91–102. de Mello, M.H.S.H., da Silva, E.A., Natal, D., 2003. Africanized bees in a metropolitan area of Brazil: shelters and climatic influences. Rev Saúde Pública 37, 237–241. Dietz, E.J., 1983. Permutation tests for association between two distance matrices. Syst. Zool. 32, 21–26. Dimmit, M.A., 2000. Plant ecology of the Sonoran Desert region. In: Phillips, S.J., Comus, P.W. (Eds.), A Natural History of the Sonoran Desert. ArizonaSonoran Desert Museum Press, Tucson, and University of California Press, Berkeley. Fortin, M.-J., Dale, M.R.T., 2005. Spatial Analysis: A Guide For Ecologists. Cambridge University Press, Cambridge. Galton, D., 1971. Survey of a Thousand Years of Beekeeping in Russia. Bee Research Association, London, England. Guzman-Novoa, E., Page Jr., R.E., 1994. The impact of Africanized bees on Mexican beekeeping. Am. Bee J. 134, 101–106. K.A. Baum et al. / Landscape and Urban Planning 85 (2008) 123–132 Harrison, J.F., Fewell, J.H., Anderson, K.E., Loper, G.M., 2006. Environmental physiology of the invasion of the Americas by Africanized honeybees. Integr. Comp. Biol. 46, 1110–1122. Hunter, L.A., Jackman, J.A., Sugden, E.A., 1993. Detection records of Africanized honey bees in Texas during 1990, 1991 and 1992 Southwest. Entomology 18, 79–89. Jaycox, E.R., Parise, S.G., 1980. Homesite selection by Italian honey bee swarms Apis mellifera ligustica (Hymenoptera: Apidae). J. Kans. Entomol. Soc. 53, 171–178. Jaycox, E.R., Parise, S.G., 1981. Homesite selection by swarms of black-bodied honeybees Apis mellifera caucasica and A. m. carnica (Hymenoptera: Apidae). J. Kans. Entomol. Soc. 54, 697–703. Johnston, A.N., Schmidt, J.O., 2001. The effect of Africanized honey bees (Hymenoptera: Apidae) on the pet population of Tucson: a case study. Am. Entomol. 47, 98–103. Kerr, W.E., 1974. Genetik des polymorphismus bei bienen. In: Schmidt, G.H. (Ed.), Sozial Polymorphismus bei Insekten. Wissenschaftliche Verlagsgesellschaft MBH, Stuttgart, Germany, pp. 94–109. Leisnham, P.T., Lester, P.J., Slaney, D.P., Weinstein, P., 2006. Relationships between mosquito densities in artificial container habitats, land use and temperature in the Kapiti-Horowhenua region New Zealand. N. Z. J. Mar. Freshwater Res. 40, 285–297. Livingston, M., Shaw, W.W., Harris, L.K., 2003. A model for assessing wildlife habitats in urban landscapes of eastern Pima County Arizona (USA). Landscape Urban Plann. 64, 131–144. Loper, G.M., 1997. Genetic evidence of the Africanization of feral colonies in S Arizona between 1993 and 1995. Am. Bee J. 137, 669–671. Loper, G.M., Sammataro, D., Finley, J., Cole, J., 2006. Feral honey bees in southern Arizona 10 years after Varroa infestation. Am. Bee J. 146, 521– 524. Marshal, J.P., Krausman, R.P., Bleich, V.C., Ballard, W.B., McKeever, J.S., 2002. Rainfall, El Nino, and dynamics of mule deer in the Sonoran Desert. J. Wildl. Manage. 66, 1283–1289. McNally, L.C., Schneider, S.S., 1996. Spatial distribution and nesting biology of colonies of the African honey bee Apis mellifera scutellata (Hymenoptera: Apidae) in Botswana. Africa. Environ. Entomol. 25, 643–652. Morse, R.A., Calderone, N.W., 2000. The value of honey bees as pollinators of U. S. crops in 2000. Bee Cult., 128. Morse, R.A., Camazine, S., Ferracane, M., Minacci, P., Nowogrodzki, R., Ratnieks, F.L.W., Spielholz, J., Underwood, B.A., 1990. The population density of feral colonies of honey bees (Hymenoptera: Apidae) in a city in upstate New York. J. Econ. Entomol. 83, 81–83. Oldroyd, B.P., Lawler, S.H., Crozier, R.H., 1994. Do feral honey bees (Apis mellifera) and regent parrots (Polytelis anthopeplus) compete for nest sites? Aust. J. Ecol. 19, 444–450. Oldroyd, B., Smolenski, A., Lawler, S., Estoup, A., Crozier, R., 1995. Colony aggregations in Apis mellifera L. Apidologie 26, 119–130. Oldroyd, B.P., Thexton, E.G., Lawler, S.H., Crozier, R.H., 1997. Population demography of Australian feral bees (Apis mellifera). Oecologia 111, 381–387. O’Neal, R.J., Waller, G.D., 1984. On the pollen harvest by the honey bee (Apis mellifera L.) near Tucson Arizona (1976–1981). Desert Plants 6, 81– 109. Pereira, A.M., Chaud-Netto, J., 2005. Africanized honeybees: biological characteristics, urban nesting behavior and accidents caused in Brazilian cities (Hymenoptera: Apidae). Sociobiology 46, 535–550. Pinto, M.A., Rubink, W.L., Coulson, R.N., Patton, J.C., Johnston, J.S., 2004. Temporal pattern of Africanization in a feral honeybee population from Texas inferred from mitochondrial DNA. Evolution 58, 1047–1055. Pinto, M.A., Rubink, W.L., Patton, J.C., Coulson, R.N., Johnston, J.S., 2005. Africanization in the United States: replacement of feral European honey bees (Apis mellifera L.) by an African hybrid swarm. Genetics 170, 1653–1665. Pinto, M.A., Sheppard, W.S., Johnston, J.S., Rubink, W.L., Coulson, R.N., Schiff, N.M., Kandemir, I., Patton, J.C., 2007. Honey bees (Hymenoptera: Apidae) of African origin exist in non-Africanized areas of the Southern United States: Evidence from mitochondrial DNA. Ann. Entomol. Soc. Am. 100, 289–295. 131 Rabe, M.J., Rosenstock, S.S., Nielsen, D.I., 2005. Feral Africanized honey bees (Apis mellifera) in Sonoran desert habitats of southwestern Arizona. Southwest. Nat. 50, 307–311. Ratnieks, F.L.W., Piery, M.A., Cuadriello, I., 1991. The natural nest and nest density of the Africanized honey bee (Hymenoptera, Apidae) near Tapachula, Chiapas, Mexico. Can. Entomol. 123, 353–359. Rubink, W.L., Luévano-Martinez, P., Sugden, E.A., Wilson, W.T., Collins, A.M., 1996. Subtropical Apis mellifera (Hymenoptera: Apidae) swarming dynamics and Africanization rates in northeastern Mexico and southern Texas. Ann. Entomol. Soc. Am. 89, 243–251. Schmidt, J.O., 1995. Dispersal distance and direction of reproductive European honey bee swarms (Hymenoptera: Apidae). J. Kans. Entomol. Soc. 68, 320–325. Schmidt, J.O., Boyer-Hassen, L.V., 1996. When Africanized bees attack: what you and your clients should know. Vet. Med. 91, 923–928. Schmidt, J.O., Edwards, J.F., 1998. Ecology of feral and Africanized honey bees in Organ Pipe Cactus National Monument. In: Tibbitts, T.J., Maender, G.J. (Eds.), First Conference on Research and Resource Management in Southern Arizona National Park Areas: Extended Abstracts. Organ Pipe Cactus National Monument and Cooperative Park Studies Unit, The University of Arizona, Tucson. Schmidt, J.O., Hurley, R., 1995. Selection of nest cavities by Africanized and European honey bees. Apidologie 26, 467–475. Schmidt, J.O., Thoenes, S.C., 1990. The efficiency of swarm traps: what percent of swarms are captured and at what distance from the hive. Am. Bee J. 130, 811–812. Schneider, S.S., 1995. Swarm movement patterns inferred from waggle dance activity of the neotropical African honey bee in Costa Rica. Apidologie 26, 395–406. Schneider, S., Blyther, R., 1988. The habitat and nesting biology of the African honey bee Apis mellifera scutellata in the Okavango River Delta, Botswana Africa. Insectes Soc. 35, 167–181. Schneider, S.S., Deeby, T., Gilley, D.C., DeGrandi-Hoffman, G., 2004a. Seasonal nest usurpation of European colonies by African swarms in Arizona USA. Insectes Soc. 51, 359–364. Schneider, S.S., DeGrandi-Hoffman, G., Smith, D.R., 2004b. The African honey bee: Factors contributing to a successful biological invasion. Ann. Rev. Entomol. 49, 351–376. Schneider, S.S., McNally, L.C., 1992. Factors influencing seasonal absconding in colonies of the African honey bee Apis mellifera scutellata. Insectes Soc. 39, 403–423. Seeley, T.D., Morse, R.A., 1977. Dispersal behavior of honey bee swarms. Psyche 83, 19–209. Seeley, T.D., Morse, R.A., 1978. Nest site selection by the honey bee Apis mellifera. Insectes Soc. 25, 323–337. Seeley, T.D., Seeley, R.H., Akratanakul, P., 1982. Colony defense strategies of the honeybees in Thailand. Ecol. Monogr. 52, 4–63. Shaw, W.W., Harris, L.K., Livingston, M., 1998. Vegetative characteristics of urban land covers in metropolitan Tucson. Urban Ecosyst. 2, 65– 73. Shochat, E., Stefanov, W.L., Whitehouse, M.E.A., Faeth, S.H., 2004. Urbanization and spider diversity: influences of human modification and habitat structure and productivity. Ecol. Appl. 14, 268–280. Shochat, E., Warren, P.S., Faeth, S.H., McIntyre, N.E., Hope, D., 2006. From patterns to emerging processes in mechanistic urban ecology. Trends Ecol. Evol. 21, 186–191. Shreve, F., 1951. Vegetation of the Sonoran Desert. Carnegie Institution of Washington Publication, vol. 591, Washington DC. Stuart, G., Gries, C., Hope, D., 2006. The relationship between pollen and extant vegetation across an arid urban ecosystem and surrounding desert in Southwest USA. J. Biogeogr. 33, 573–591. Taber III, S., 1979. A population of feral honey bee colonies. Am. Bee J. 118, 842–847. Taylor, O.R., Spivak, M., 1984. Climatic limits of tropical African honeybees in the Americas. Bee World 65, 38–47. U.S. Census Bureau, 2007. Table 1: Annual Estimates of the Population of Metropolitan and Micropolitan Statistical Areas: April 1, 2000 to July 1, 2006 (CBSA-EST2006-01). Population Division. 132 K.A. Baum et al. / Landscape and Urban Planning 85 (2008) 123–132 Visscher, P.K., Seeley, T.D., 1982. Foraging strategy of honeybee colonies in a temperate deciduous forest. Ecology 63, 1790–1801. Winston, M.L., 1987. The Biology of the Honey Bee. Harvard University Press, Cambridge, MA. Winston, M.L., 1992. The biology and management of Africanized honey bees. Ann. Rev. Entomol. 37, 173–193. Winston, M.L., Otis, G.W., Taylor, O.R., 1979. Absconding behaviour of the Africanized honey bee in South America. J. Apic. Res. 18, 85–94. Winston, M.L., Taylor, O.R., Otis, G.W., 1983. Some differences between temperate European and tropical African and South American honey bees. Bee World 64, 12–21.