Elsevier

Science of The Total Environment

Volume 532, 1 November 2015, Pages 1-13
Science of The Total Environment

Correlations between land covers and honey bee colony losses in a country with industrialized and rural regions

https://doi.org/10.1016/j.scitotenv.2015.05.128Get rights and content

Highlights

  • 60 out of 133 land cover classes were repeatedly correlated with bee colony losses.

  • 17 land cover classes correlated with bee losses were related to agriculture.

  • 43 land cover classes related with bee losses served other human activities.

  • Human activities besides agriculture need to be considered more carefully.

Abstract

High levels of honey bee colony losses were recently reported from Canada, China, Europe, Israel, Turkey and the United States, raising concerns of a global pollinator decline and questioning current land use practices, in particular intense agricultural cropping systems. Sixty-seven crops (data from the years 2010–2012) and 66 mid-term stable land cover classes (data from 2007) were analysed for statistical relationships with the honey bee colony losses experienced over the winters 2010/11–2012/13 in Luxembourg (Western Europe). The area covered by each land cover class, the shortest distance between each land cover class and the respective apiary, the number of plots covered by each land use class and the size of the biggest plot of each land cover class within radii of 2 km and 5 km around 166 apiaries (2010), 184 apiaries (2011) and 188 apiaries (2012) were tested for correlations with honey bee colony losses (% per apiary) experienced in the winter following the season when the crops were grown. Artificial water bodies, open urban areas, large industrial facilities including heavy industry, railways and associated installations, buildings and installations with socio-cultural purpose, camping-, sports-, playgrounds, golf courts, oilseed crops other than oilseed rape like sunflower or linseed, some spring cereals and former forest clearcuts or windthrows were the land cover classes most frequently associated with high honey bee colony losses. Grain maize, mixed forest and mixed coniferous forest were the land cover classes most frequently associated with low honey bee colony losses. The present data suggest that land covers related to transport, industry and leisure may have made a more substantial contribution to winter honey bee colony losses in developed countries than anticipated so far. Recommendations for the positioning of apiaries are discussed.

Introduction

Pollination is an important ecosystem service with an estimated global value of 153 × 109 US $ (Gallai et al., 2009). Honeybees (Apis mellifera L.) accounted for > 25% of crop visits in 21 studies (Garibaldi et al., 2011) and thus significantly contribute to satisfactory pollination of many high-value crops and wild plant species (Potts et al., 2010, Lautenbach et al., 2012). Furthermore, they are essential for the production of honey, wax and pollen based products. High levels of honey bee colony losses over winter were recently reported from Canada, China, Europe, Israel, Turkey (van der Zee et al., 2012) and the United States (Ellis et al., 2010, vanEngelsdorp et al., 2012). Several reasons related to the decline of bees such as the parasitic mite Varroa destructor (De la Rúa et al., 2009, Rosenkranz et al., 2010), fungal pathogens (Genersch et al., 2010), viruses (de Miranda and Genersch, 2010, Clermont et al., in press) and modified land-use (Nicholls and Altieri, 2013) were identified. In the case of plants providing pollen, the time lag between pollen production in summer and honey bee colony death in winter was discussed to be caused by the storage of contaminated pollen and bee bread until winter (Škerl et al., 2009). Pretty et al. (2000) estimated that honey bee colonies having a value between one and two £ million were lost each year in the UK due to agriculture within the period 1990–1996.

Land cover types differ with respect to quality and quantity of nutrients (Donkersley et al., 2014), risk for exposure to toxins, suitability for pathogens and parasites and may thus directly or indirectly affect bee health. Due to the recent advances in remote sensing and classification techniques (Foody, 2002), accurate land cover maps became available. The position and size of for instance rivers, lakes, railways, most settlements, parks or forests do not significantly change each year, whereas the locations and areas covered by crops are more variable due to crop rotation. Therefore, at least annual land cover data are required particularly for agricultural crops, while temporal resolutions larger than one year may still be suitable for investigating effects of land covers other than annual crops.

While relationships between land cover and wild bee abundance, their pollination efficacy or pollinator species richness were studied on several occasions (Morandin et al., 2007, Garibaldi et al., 2011, Le Féon et al., 2010, Le Féon et al., 2013, Xie and An, 2014), studies on potential relationships between land cover and honey bee colony losses are scarce. Naug (2009) reported a decline in cropland, pasture and rangeland and an increase of developed land in the US during the period 1982–2003, over which bee colony numbers declined as well and hypothesized that nutritional stress due to habitat loss may be partly responsible for the honey bee decline. In contrast to wild bees, visitation rates of honey bees in fields did not change with isolation from natural areas (Garibaldi et al., 2011) or increasing proportion of natural habitats in the landscape (Xie and An, 2014). A meta-analysis by Winfree et al. (2009) demonstrated that honey bee abundance was not significantly associated with human disturbance. The present study aims to identify land cover classes and potentially spatial land use patterns that were associated with honey bee colony losses and to derive recommendations for choosing locations for setting up apiaries with a reduced risk of winter losses.

Section snippets

Data sources and characteristics

Data on colony losses per apiary were available from a previous study for the winters 2010/11 and 2011/12 (Clermont et al., 2014). Colony death was confirmed by visual inspection by the beekeeper owning the colony and communicated to the national veterinary administration in the spring following the winter when the colony loss occurred. Over the period 2010–2013, the honeybee genotypes Buckfast and Carnica were used in Luxembourg with Buckfast being by far more popular. In 2013, beekeepers

General observations

Land use differed between Northern and Southern Luxembourg (Fig. 1). Northern Luxembourg is characterized by more softwood forest, fewer meadows, more crops, fewer urban areas and open environments compared to Southern Luxembourg. In the period 2010–2012, permanent meadows plus pastures covered about 26% of the land, field crops 23%, vineyards 0.1%, urban and industrial areas 12%, deciduous and mixed forest 29%, softwood forest 8%, uncultivated open environments 3% and water surfaces and

Agriculture, industry, transport, leisure and honey bee colony losses

Activities related to transport (major roads, railways, parkings), industry (brownfields, production facilities, huge buildings, landfills, military areas, installations of the electrical grid, installations for the storage of gas), leisure (sports-, camping-, and playgrounds, garden patches, buildings with socio-cultural purposes) as well as artificial water surfaces were frequently associated with high honey bee colony losses. Potentially damaging crops were several spring cereals, feed

Conclusions

The present data suggest that more attention needs to be paid to land use classes not related to agriculture than assumed so far to avoid high honey bee colony losses over winter. This conclusion is in general agreement with the fact that the increase in developed land was accompanied by a decline in honey bee colony numbers in the US during the period 1982–2003 (Naug, 2009) as well as with the report by Smith et al. (2014) who found increasing honey bee colony numbers world-wide except for

Acknowledgements

We thank the ‘Ministère de l'Agriculture, de la Viticulture et de la Protection des Consommateurs’ (BeeFirst) of Luxembourg for financial support, Laura Giustarini for advice on data extraction with ArcGIS, Ivonne Trebs for critical comments on an early version of the manuscript, Vanessa Peardon for language editing, Carlo Georges (Administration des Services Vétérinaires) and the members of the ‘Lëtzebuerger Beienziichter’ for their help during the gathering of the honey bee colony loss data.

References (32)

  • M. Beekman et al.

    Long-range foraging by the honey-bee, Apis mellifera L

    Funct. Ecol.

    (2000)
  • M. Beekman et al.

    Comparing foraging behaviour of small and large honey-bee colonies by decoding waggle dances made by foragers

    Funct. Ecol.

    (2004)
  • A. Clermont et al.

    A survey on some factors potentially affecting losses of managed honey bee colonies in Luxembourg over the winters 2010/2011 and 2011/2012

    J. Apic. Res.

    (2014)
  • A. Clermont et al.

    Virus status, varroa levels and survival of 20 managed honey bee colonies monitored in Luxembourg between summer 2011 and spring 2013

    J. Apic. Sci.

    (2015)
  • M.J. Couvillon et al.

    Waggle dance distances as integrative indicators of seasonal foraging challenges

    PLoS One

    (2014)
  • P. De la Rúa et al.

    Biodiversity, conservation and current threats to European honey bees

    Apidologie

    (2009)
  • Cited by (0)

    View full text