The new integrated model BEEHAVE: a tool for exploring multiple stressors of honeybee colonies

The new integrated model BEEHAVE: a tool for exploring multiple stressors of honeybee colonies Volker Grimm 1, Matthias Becher 2, Peter Kennedy 2, Juliane Horn 1, Pernille Thorbek 3, Juliet Osborne 2 1 Helmholtz Centre for Environmental Research UFZ, Leipzig, Germany 2 Environment & Sustainability Institute, University of Exeter, Cornwall, UK 3 Syngenta, Jealott s Hill, UK

Presentation based on project: "Honeybee population dynamics: integrating the effects of factors within the hive and in the landscape" (Rothamsted Research, UK, 2009-2013). Co-funded by BBSRC (88%) and Syngenta (12%) Matthias Becher*, Pete Kennedy*, Jenny Swain #, Judy Pell, Juliet Osborne*: ESI, University of Exeter, UK * Previous affiliation: Rothamsted Research, UK # Rothamsted Research, UK Current affiliation: J.K. Pell Consulting Dave Chandler, Gillian Prince, Sally Hilton: University of Warwick Pernille Thorbek: Syngenta Volker Grimm, Juliane Horn: UFZ

BEEHAVE team The modeller: Matthias Becher The experimenter: Pete Kennedy The PI: Juliet Osborne

Existing honeybee models: review

Existing honeybee models: review Three categories of models: 1. Within-hive colony dynamics (8) Khoury et al. (2011): EFSA guidance 2. Varroa mite population dynamics within hives (11) 3. Foraging (12)

Conclusions from review Model testing, validation, and analysis of most models was very limited No clear separation of imposed and emergent dynamics No clear indication of how much calibration was involved Limited or no sensitivity analysis For foraging models, a benchmark test exists: the Seeley et al. (1991) feeder experiment

Conclusions from review Well-tested building blocks exist in published models A model that allows stressors to be integrated within and outside the hive does not yet exist Colony structure and important feedback loops need to be included (e.g., "age of first foraging") Egg-laying rate, weather, colony structure, and availability of nectar and pollen should drive the dynamics

BEEHAVE: developed by Matthias Becher

affects egg laying brood care (max. ratio: 1:3) Colony module eggs (workers) eggs (drones): 4% 3d 4d larvae (workers) 6d pupae (workers) 12d in-hive workers 5-65d foragers (agents) => FORAGING MODULE larvae (drones) 6d pupae (drones) 14d adult drones

affects egg laying brood care (max. ratio: 1:3) Colony module eggs (workers) eggs (drones): 4% 3d 4d larvae (workers) 6d pupae (workers) 12d in-hive workers 5-65d foragers (agents) => FORAGING MODULE NECTAR & POLLEN CONSUMPTION! larvae (drones) 6d pupae (drones) 14d adult drones

affects egg laying brood care (max. ratio: 1:3) Colony module eggs (workers) eggs (drones): 4% SPECIFIC MORTALITIES: 3d larvae (workers) 4d larvae (drones) - for developmental stages 6d 6d - foraging risk pupae (workers) 12d pupae (drones) 14d - lack of honey & pollen in-hive workers 5-65d foragers (agents) => FORAGING MODULE NECTAR & POLLEN CONSUMPTION! adult drones - lack of nurse bees - effects of virus infection

recruited foraging found a patch? Foragers: resting searching collect nectar/pollen dying? Foraging module leaving hive? experienced? good patch? dancing unloading abandon patch? stop foraging? # repetitions depending on weather conditions

based on Martin 1999, 2001: # phoretic mites Varroa module REPRODUCTION random distribution of phoretic mites on brood cells: +3 +1 suitable drone cells (preferred) suitable worker cells VIRUS TRANSMISSION Viruses: Deformed wing virus (DWV) or Acute paralysis virus (APV) => reduction of life span via pupae....or via phoretic mites - adult workers

Landscape module Automated calculation of: - number of patches - distance to apiary - area of patch - chance to find the patch - crop type (colour) nectar: pollen: area apiary distance time time

Documentation, testing, validation Implemented in NetLogo (free software platform) Documented in ODD format (ca. 40 pages) User manual and guided tour exist (ca. 60 pages) Extensive testing (debug code, consistency tests, visual output) Validation: Age of first foraging, lifespan Number of reproductive cycles of varroa in a year Seeley's feeder experiment

Validation: life span and age of first foraging Becher et al., submitted to J. Appl. Ecol.

Validation: Foraging, Seeley's feeder experiment Becher et al., submitted to J. Appl. Ecol.

Forage and varroa infestation - One food patch in distance of 250, 500, and 1000 m from the hive - Varroa infestation in the first year Becher et al., submitted to J. Appl. Ecol.

Availability of nectar and pollen Oilseed rape food flow: high availability of pollen: 30 day period -> based on real data of flower phenology of oilseed rape Variable Value Literature Flowers/plant 60 Wikipedia (14.04.2013) Oilseed rape (Brassica napus L.) Oliver Dixcon: Wall-to-wall oilseed rape % open flowers at observation sites day B10 B50 B100 B150 B200 Mean 1 76 86 90 90 95 87,4 2 68 90 76 88 78 80 3 79 75 64 79 68 73 4 72 56 58 56 70 62,4 5 34 34 35 24 32 31,8 6 32 28 28 21 29 27,6 7 37 28 26 40 30 32,2 8 14 14 13 14 14 13,8 9 10 12 14 12 22,5 14 10 10 12 14 12 22,5 14 -> from Illies (2005) lifespan/flower 3 days Fruwirth (1922) Plants/m² 60 Feiffer Consult Nectar/flower n 600 µg Illies (2005) Pollen/flower p 292 µg Von der Ohe et. al (1999) Pollen/ha 90-170kg Von der Ohe et al. (1999) Sugar content 44 59 % Maurizio & Schaper (1994) Numbers of flowers / ha = c = 60 * 60 * 10000 Nectar & pollen per observation day: pollen / ha [kg] = c * mean * p / 1000000000 nectar / ha [kg] = c * mean * n / 1000000000 Horn et al., in preparation 20

Henry et al. (2012; Science) Scenario No varroa Nectar feeder at 1000 m; sugar concentration 1.5 mol/l Hypothetical scenario of pesticide-induced increase in forager mortality Henry et al: increase in forager mortality from ca. 15 to 30% in Khoury et al. model Equivalent in BEEHAVE: double mortality per foraging trip for 30 days

Effect on average colony size Becher et al., submitted to J. Appl. Ecol.

Effect on colony survival Becher et al., submitted to J. Appl. Ecol.

BEEHAVE: current state - Model submitted in July - Once accepted, model, computer program, and manual will be freely available (www.beehave-model.net) Designed so that others can test and use it Offer training courses, workshops - 2 PhD students currently working with BEEHAVE Multiple stressors, landscape structure and dynamics (Juliane Horn, UFZ) Specific pesticide module (Jack Rumkee, Univ. Exeter/Syngenta) - To do: From one to many colonies SEITE 24

Summary BEEHAVE: first attempt to link within-hive dynamics to foraging in heterogeneous and dynamic landscape BEEHAVE is comprehensive and realistic, and therefore relatively complex BEEHAVE is still a model but better validated than any other existing model Exploring resilience mechanisms will take some time BEEHAVE (or refinements) will be suitable for regulatory risk assessment Please contact me, Matthias Becher, or Juliet Osborne if you have interest in using or developing BEEHAVE SEITE 25

Omholt 1986 degrandi-hoffman et al. 1986 Martin 2001 AlGhamdi & Hoopingarner 2004 Thompson et al. 2005/2007 Schmickl & Crailsheim 2007 Becher et al. 2010 Khoury et al. 2011 Omholt & Crailsheim 1991 Calis et al. 1999a Calis et al. 1999b Boot et al. 1995 Wilkinson & Smith 2002 degrandi-hoffman & Curry 2004 Sumpter & Martin 2004 Vetharaniam & Barlow 2006 Vetharaniam 2012 All forager models BEEHAVE Representation of stressors Factors Genetic diversity Varroa mites + + (+) + + + (+) + + + (+) (+) + Viruses + + + Bacterial pathogens Nosema spp. (+) + Loss of forage quantity + + Forage nutritional quality Beekeeping practice Pesticides inside hive Pesticides outside hive Forager death unknown cause + (+) (+) (+) (+) (+) (+) + (+) (+) + + + + (+) + (+) (+) + + (+) (+) +