|
Food and Ecological Systems Modelling Journal :
Formal Model Article Format
|
|
Corresponding author: Elżbieta Ziółkowska (e.ziolkowska@uj.edu.pl)
Academic editor: Francesco Nazzi
Received: 16 Feb 2023 | Accepted: 04 May 2023 | Published: 06 Jun 2023
© 2023 Elżbieta Ziółkowska, Agnieszka Bednarska, Ryszard Laskowski, Christopher Topping
This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Citation:
Ziółkowska E, Bednarska AJ, Laskowski R, Topping CJ (2023) The Formal Model for the solitary bee Osmia bicornis L. agent‑based model. Food and Ecological Systems Modelling Journal 4: e102102. https://doi.org/10.3897/fmj.4.102102
|
|
Solitary bees provide an important ecological and agricultural service by pollinating both wild plants and crops, often more effectively than honey bees. In the context of worldwide pollinators' declines, it is important to better understand the functioning of populations under multiple stressors at larger spatial and temporal scales. Here we propose building a detailed, spatially-explicit agent-based model of one of the best-studied species of solitary bees, Osmia bicornis L. In this Formal Model, we review various aspects of O. bicornis biology and ecology in detail and provide descriptions of their planned implementations in the model. We also discuss the model gaps and limitations, as well as inclusions and exclusions, allowing a dialogue with the reviewers about the model's design.
The ALMaSS model of O. bicornis aims to provide a realistic and detailed representation of O. bicornis populations in space and time in European agricultural landscapes. The model will be a part of the Animal, Landscape and Man Simulation System (ALMaSS); thus will be able to utilise a highly detailed, dynamic ALMaSS landscape model. It will consider the behaviour of all bee life stages daily and use state transitions to allow each individual to decide their behaviour. The development of egg-to-pupa stages in the nest will be temperature-driven. Adult bees, after they emerge from the nest in spring, will interact with the environment. They will be able to search for suitable nesting locations, provision their brood cells with pollen and reproduce. Modelled females will balance offspring size and number following the optimal allocation theory, but local environmental factors will modify their actual parental investment decisions. The model will include the daily mortality rate for the egg-to-pupa stages, overwintering mortality, and background mortality outside the nest. We will also consider the risk of open-cell parasitism as increasing with the time the brood cell is open.
With the level of detail suggested, the model will be able to simulate population-level dynamics in response to multiple factors at the landscape scale over long periods. The European Food Safety Authority (EFSA) has suggested O. bicornis as a model organism for non-Apis solitary bees in the pesticide risk assessment scheme. Therefore, we hope our model will be a first step in building future landscape risk assessments for solitary bees.
agroecosystem, pollinators, population dynamics, spatially-explicit model, wild bees
Over the last few decades, a dramatic decrease in the abundance of arthropods has been observed worldwide, especially in agricultural landscapes (
Here, we present a Formal Model of a spatially-explicit agent-based model for the solitary red mason bee Osmia bicornis (Linnaeus, 1758, Hymenoptera: Megachilidae, formerly Osmia rufa L.). The description follows the “Formal Model” format proposed by
In contrast to these existing models, our model will be the first fully spatially-explicit one to model population-level dynamics in response to multiple factors at the landscape scale (up to hundreds of km2) over long periods (decades). The behaviour of individual bees will emerge from bees’ interactions with the environment, which in ALMaSS is represented by a highly detailed, dynamic landscape model (
In the context of worldwide declines in pollinator populations, most scientific efforts, legislation and conservation practices have focused on a single species, the western honey bee Apis mellifera. For many years, the USA and Europe have also focused on the honey bee for assessing environmental risks for pollinators. Recently, the European Food Safety Authority (EFSA) has promoted the development of a system-based approach to the environmental risk assessment of multiple stressors in honey bees (
The ALMaSS model of O. bicornis aims to provide a realistic and detailed representation of O. bicornis populations in space and time in European agricultural landscapes. The model should be able to generate emergent population patterns, based on landscape and management context. The model will be used to:
Here, we use an agent-based modelling approach (
Our model is spatially explicit; the position of a bee and its nest are both simulated. The ALMaSS modelling environment also provides a detailed spatio-temporal representation of the landscape from which individual bees obtain the information necessary to simulate their behaviour. This representation describes spatial landscape heterogeneity through a detailed raster land-cover map with a spatial resolution of 1 m2. Farmed areas are represented as accurate maps of fields grouped into farm units of different types (e.g. cattle or arable farms). The temporal component of agricultural landscape heterogeneity refers to crop management throughout a year, described through individually-tailored management plans for each crop. The cropping system is understood as a pluri-annual crop rotation. Crop management plans consist of combinations of farm activities (including pesticide treatments), time windows and probabilities of carrying out these activities. The temporal component includes weather conditions and vegetation growth models for all modelled vegetation types and crops; this is all updated daily. Such an approach gives a highly realistic dynamic landscape simulation with vegetation growing in response to the weather and the pattern of farming activities related to each specific crop, farm and field (
The model will consider the behaviour of all bee life stages daily. The model's parameters are based on field and laboratory data on O. bicornis available in literature and from our own experiments (performed by the Terrestrial Ecosystems and Ecotoxicology Group, Institute of Environmental Sciences, Jagiellonian University, Kraków, Poland). When the necessary data were not available for O. bicornis, we estimated parameters, based on published data from other species of the Osmia genus with similar life history, particularly O. cornuta (European orchard bee, slightly larger than O. bicornis), O. cornifrons (Asian sister species) and O. lignaria (North American sister species).
The red mason bee O. bicornis is an important pollinator in agricultural landscapes. It is often reared on a commercial scale to be used as a managed alternative pollinator, especially in orchards (
Some limitations must also be considered related to the nature of the available data. Laboratory studies investigating the developmental time of O. bicornis in various temperature regimes were carried out on populations reared in Poland and Germany and, thus, may not represent populations from other climatic regions. Similarly, following
Populations of wild bees depend on the spatial distribution of nesting and foraging habitats and it has been suggested that O. bicornis is strongly constrained by nesting site availability (e.g.
According to many studies, food is not a limiting factor for O. bicornis in the field (
In the first version of the O. bicornis model, we will assume that adult bees are not limited by food availability for themselves and that the nest-building material is unlimited. The impact of pollen quantity on maternal investment will be considered. However, we will not account for either O. bicornis preference towards certain types of pollen or the influence of pollen quality on the offspring. We will not model energy intake and consumption by individual bees during foraging and dispersal. Nevertheless, the pollen search algorithm will include the distance from the nest to favour closely-available pollen resources (see section "Use of resources"). Competition from other pollinators will be modelled as a decrease in the floral resources available to O. bicornis (controlled by the OsmiaDensityDependentPollenRemovalConst parameter). The assessment of parasitoids will be simplified by not explicitly modelling parasitoid populations. Instead, the risk of open-cell parasitism will increase with the time the cell is open (
Although the model is aimed towards use for pesticide risk assessment, the first version of the model will only consider that pesticides are present in the environment and we will not include specific pesticide handling processes. This is because we assume that these will be specific to any particular pesticide scenario and that exposure procedures developed for ApisRAM (
The red mason bee is a univoltine (one generation per year), polylactic/polylege solitary species of the Osmia genus, common in central and northern Europe (
O. bicornis females are monandrous and males are polygamous (
The development of the red mason bee is similar to that of other spring species of the Osmia genus. It can be categorised into six main life stages: egg, larva, prepupa, pupa, cocooned adult and adult. Life stages from egg to cocooned adult occur inside the nest. Egg laying and larval development occur in spring, while prepupal and pupal stages occur in summer. The imago instar appears in the cocoon at the end of summer/beginning of autumn. The cocooned adult is the overwintering form of the red mason bee (
Each year, a new generation of bees appears in spring. After finding a suitable nesting place, bees enter the reproduction phase. A mother bee builds a separate brood cell for each egg, which she provisions with pollen and a low proportion of nectar. The further development can be summarised as follows (the details of development and overwintering drivers with references are described in sections "Development in the nest" and "Overwintering", respectively):
The model will include all six-life stages of O. bicornis female individuals. It is assumed that all adult females are being fertilised, therefore, we will not model mating activity. The males will not be included in the model, except for the nesting and provisioning activities of the mother (further information is provided in sections "Nesting" and "Provisioning"). The development within the nest (from egg to cocooned adult) will be modelled according to the in-nest life-stage state machine and transition path (Fig.
Bees of all life stages will have some common attributes, such as the duration of the life stage measured in both days and accumulated degree days (Table
State variables with units common for all life stages of O. bicornis. These will be explained in greater detail in the text that follows.
|
Variable name |
Units |
Description |
|
Stage age |
days |
How many days in this life stage |
|
Sex |
female/male |
Allocation of sex. For eggs: female = fertilised, male = unfertilised. After the egg stage, the unfertilised eggs will not be tracked in the model |
|
Mass |
mg |
The mass of provision gathered in the brood cell for a larva or mass of an adult bee being a derivative of the provision mass |
|
Accumulated degree days |
degree days |
How many degree days are accumulated |
|
Target degree days |
degree days |
How many degree days are needed to transition to the next stage or trigger some type of behaviour |
|
Daily mortality |
daily probability of death |
The mortality experienced in a given life stage besides overwintering |
The early and late flying populations of the same Osmia species have different durations of the developmental cycle. For example,
Duration of in-nest development is temperature driven, such that higher temperatures lead to faster development of eggs, larvae and pupa (Table
Summary of studies investigating temperature-driven development of O. bicornis in the nest. Only data on females are shown.
| Reference | Temperature (°C) | Duration (days) | |||||
| Egg | Larva | Prepupa | Pupa | Prepupa + pupa | Total development | ||
| Natural conditions | |||||||
|
|
- | 7.6 | 39.1 | - | - | 54.2 | 100.9 |
|
|
- | 7 | 34 | 25 | 36 | 61 | 102 |
| Laboratory conditions | |||||||
|
|
17.5 | 8 | 32 | 32 | 54 | 86 | 126 |
|
|
20 | 3 | 36 | 28 | 32 | 60 | 99 |
|
|
22.5 | 5 | 32 | 25 | 30 | 55 | 92 |
|
|
25 | 3.2 | 28 | 24 | 29 | 53 | 84.2 |
|
|
27.5 | 4 | 20 | 42 | 20 | 62 | 86 |
|
|
28 | 3.1 | 19.7 | 36.3 | 10.9 | 47.2 | 70 |
|
|
30 | 2 | 27 | 31.2 | 23.1 | 54.3 | 83.3 |
Relationship between duration of O. bicornis female developmental stages and temperature. For egg, larva and pupa, we combined the results obtained by
In the model, the transition to the next developmental stage for the eggs, larvae and pupae (Fig.
Parameters for temperature-driven development of O. bicornis in the nest (based on data from Table
|
Osmia life stage |
Lower developmental threshold (LDT) in °C |
Sum of effective temperatures (SET) in degree days |
|
Egg |
13.8 |
37.0 |
|
Larva |
8.5 |
422.4 |
|
Pupa |
13.2 |
272.3 |
The duration of the prepupal stage will be related to the temperature using a quadratic function (a mean of the functions presented in Fig.
During each developmental stage there is a probability of dying (see section "Mortality"), generally dependent on the temperature profiles. O. bicornis will also be affected by open-nest parasitism (see section "Parasitism").
The overwintering period of O. bicornis can be divided into three parts: pre-wintering, diapause (wintering) and post-diapause quiescence. In central Europe, pre-wintering occurs in September – November, diapause in November – January and post-diapause quiescence in February – March. The photoperiod has not been implicated in overwintering, since development from the egg to the adult in this species takes place inside a sealed nest in complete darkness. The initiation of pre-wintering does not require a temperature signal either; rather, it is a fixed component of the ontogenic bee programme (
Although no studies on triggers for diapause have been conducted on O. bicornis, important results can be found for O. lignaria and O. cornuta. It was shown that respiration rates for overwintered bees under both natural and experimental conditions follow the same pattern. There is an abrupt drop in respiration shortly after adult eclosion, then a sharp increase in coincidence with a strong decline in ambient temperatures, followed by a further slow, steady increase throughout the winter (
A proper definition of the transition between pre-wintering and diapause is important, as the duration of pre-wintering has important consequences for the diapause development, winter survival and the bee’s vigour at emergence in spring. Increased duration of pre-wintering leads to an increase in body-fat depletion and decreases the post-emergence longevity without feeding (
During diapause, the value of the supercooling point (i.e. the temperature threshold above which the body of the bee does not freeze) decreases. Diapause lasts about 100 days and seems to be independent of temperature variation. After this period, in post-diapause, bees develop normally, but their development is inhibited by the temperature (higher temperatures shorten the duration of post-diapause). At the beginning of post-diapause, their supercooling point increases gradually until spring.
In the model, the pre-wintering period will be defined as the period from adult eclosion (end of the pupa stage) to the time threshold D, determined, based on the temperature regime as defined below.
For every day d after 1 September, we will be checking the average daily temperature Tavg, in order to detect a sharp and stable temperature drop:
If
\((T_{avg}(d-5) - T_{avg}(d-4) ≥ 1°C)\) and \((T_{avg}(d-4) - T_{avg}(d - 3) ≥ 1°C)\)
and \(T_{avg}(d-2)\) and \(T_{avg}(d-1)\) and \(T_{avg}(d) < 13°C\) (Eq. 1)
Or
\((T_{avg}(d-5) - T_{avg}(d-4) ≥ 3°C)\) and \(T_{avg}(d-3)\) and \(T_{avg}(d-2)\) and
\(T_{avg}(d-1)\) and \(T_{avg}(d)<13°C\) (Eq. 2)
\(Then threshold \space D=d\)
In
In our model, the temperature conditions during pre-wintering will influence the bee’s overwintering mortality (see section "Mortality"), while temperature conditions during diapause – the timing of bee’s emergence from the nest on spring (see section "Emergence from the nest"). The impact of the duration of the pre-wintering on the longevity of bees will not be considered as there is evidence that prolonged pre-wintering duration does not affect post-emergence longevity in nature, probably because emerging females could rapidly replenish their metabolic reserves through feeding (
The start of emergence from the nest for bees of the Osmia genus depends on the overwintering temperatures (
\(noDaysToEmerge=39.4819-1.70001 \times mean\_temp \qquad \qquad\) (Eq. 3)
Not all bees emerge from the nest/nests on the same day. The curves of the cumulative percentage of emerged females are sigmoid-shaped (
In the model, we will relate the beginning of emergence to the accumulated degree days during the wintering period, according to
\(noDaysToEmerge=39.48-0.0147 \times \sum(DD\_winter) \qquad \qquad\)(Eq. 4)
where noDaysToEmerge is the number of days to the beginning of emergence after 1 March with a mean daily temperature above a certain threshold and Ʃ(DD_winter) is the number of accumulated degree days with baseline temperature T0 = 0°C, during the wintering period as defined in section "Overwintering".
The emergence distribution was implemented according to data provided by A. Bednarska (personal communication; data come from the lab experiment where bees were overwintered at 4 degrees and, on 11 April, were moved to 20 degrees; Fig.
In solitary bees, progeny body size is largely determined by maternal provisions and microclimate (
Cocooned adult mass increases with provision mass (
Relationship between pollen provisions and cocooned adult mass of O. bicornis (based on data from
|
Temperature |
Duration of larva to cocooned adult stages (according to |
Relationship between cocooned adult mass [mg] (y) and mass of consumed pollen (x) for females [mg] |
Mean mass of pollen remnants [mg] |
|
|
Days |
Accumulated degree days |
|||
|
20 |
96 |
1920 |
y = 0.23*x +57 |
0 |
|
25 |
81 |
2025 |
y = 0.34*x-1 |
3.85 |
|
30 |
81.3 |
2439 |
y = 0.29*x-6 |
119.23 |
The fresh mass of adult females after emergence can vary considerably from 25 to 150 mg (Fig.
We will assume that the larva consumes all provided pollen. This is because, under natural conditions, the larva does not experience extreme temperatures > 30°C for a long enough time to cause substantial pollen remnants as reported by
The following formulae will be implemented to calculate the mass of cocooned adults from the provision mass derived, based on
\(femaleCocoonedMass={provisionMass \over 3.247} \qquad \qquad\)(Eq. 5)
\(\\maleCocoonedMass={provisionMass \over 3.597} \qquad \qquad\)(Eq. 6)
where provisionMass is the mass of pollen available in the nesting cell and all masses are in [mg].
It is noted that the mass of the cocooned male is not used directly as a life stage attribute (as males are not tracked in the modelling past the egg stage), but Equations 5 and 6 are both used by the mother bee in the provisioning plan (see section "Provisioning").
Since, according to
\(femaleAdultMass=0.80 \times femaleCocoonedMass+4.00 \qquad \qquad\) (Eq. 7)
In the code, each bee after emergence from the nest will obtain a mass property as a direct function of the mass of provision available in the nest cell from which the bee originates. This function will be a combination of two linear relationships, one predicting the mass of the cocooned adult from the mass of provision (Equation 5) and the second predicting the mass of the adult bee from the mass of cocooned adult (Equation 7):
\(femaleAdultMass=0.25 \times provisionMass+4.00 \qquad \qquad \)(Eq. 8)
We are aware that body mass decreases during the flight season. However, this will not be tracked during the simulation. Instead, the model will use a provisioning efficiency variable related to the age of the bee (see section "Provisioning").
In the model, two classifications of adult female bees, based on their mass, will be used:
\(BeeSizeScore1=int(floor({adultFemaleMass-OsmiaFemaleMassMin \over(OsmiaFemaleMassmax-OsmiaFemaleMassMin)/3.0}+0.5))\) (Eq. 9)
BeeSizeScore2 – will use smaller classes of bee sizes with size class controlled by the OsmiaAdultMassCategoryStep variable and is calculated according to Equation 10. BeeSizeScore2 will be used in the calculation of foraging and dispersal distances (see section "Foraging and dispersal"), the planned sex ratio of female vs. male progeny in a given nest (see section "Reproduction) and the planned mass of provision needed for each egg (see section "Provisioning").
\(BeeSizeScore2=int(floor({adultFemaleMass-OsmiaFemaleMassMin \over OsmiaAdultMassCategoryStep}+0.5))\qquad\)(Eq. 10)
We can differentiate between the bee’s homing distance and foraging distance. Homing distance is the maximum distance from which a bee can find its nest. Homing distance differs for individuals of the same species, indicating that some bees have been further from the nest in their life than other bees (
\(\log _{10}(r90)=-1.363+3.366 \times \log _{10}(IT\_span) \qquad \qquad\) (Eq. 11)
\(\log _{10}(r50)=-1.643+3.242 \times \log _{10}(IT\_span) \qquad \qquad\)(Eq. 12)
where r90 is the maximum homing distance (i.e. the distance at which 10% of bees are able to return to the nest) in [km], r50 is the typical homing distance (i.e. the distance at which 50% of bees are able to return to the nest) in [km] and IT_span is the intertegular span in [mm].
The foraging range includes inbound and outbound flights from the nest to the resources and varies with the spatial and temporal availability of resources in the landscape (
According to the general Equations 11 and 12, O. bicornis female weighing 92.63 mg (mean body mass according to the data provided by A. Bednarska; see Fig. 7; sd = 17.36) has r50 of 0.97 km and r90 of 2.13 km. These values are much higher than those provided by
In the model, two different types of movement will be considered: long-range (dispersal) and short-range. The long-range movement will be used when looking for a new nesting site if there are no nesting sites available in the surroundings of the previous nest or the resources available around the nest are too low to allow for nesting. In the model, the dispersal distance will be related to the maximum homing distance r90, i.e. r90 will be used as a rescaling parameter to the general movement distribution defined as the beta distribution X ~ Beta(α, β) with assumed parameters α = 10 and β = 5 (Figure 8). r90 will be calculated according to Equation 11 and, therefore, will be related to the bee mass through the intertegular span, IT_span. The intertegular span will not be directly introduced in the model, but we will calculate it from the body mass of the bee, based on the equations provided by
\(IT\_span=0.77 \times dryBodyMass^{0.405} \qquad \qquad\)(Eq. 13)
\(dryBodyMass=1.219 \times 10^{-4}+0.358 \times femaleAdultMass \qquad \qquad\)(Eq. 14)
The short-range movement will be used to evaluate food resources around the nest (i.e. food resource availability will be assessed in the vicinity of the nest defined by the bee’s foraging distance; see section "Use of resources"). As no studies show foraging histograms for O. bicornis (no capture-mark-recapture studies with enough observations), in the model, the foraging distance will be related to the typical homing distance r50, that is, r50 will be used as a rescaling parameter to the general movement distribution defined as the beta distribution X ~ Beta(α, β) with assumed parameters α = 10 and β = 5 (Fig.
We decided to use Equations 11 and 12 to calculate r90 and r50, even if they are based on data from homing studies for 16 different bee species rather than for O. bicornis, as these are the only available data that allow relating movement to bee size. Although
Immediately after mating, females are rarely seen around their natal nesting site and they use this time for feeding and to complete ovary maturation. The pre-nesting period is usually 2-5 days, but it may be longer due to bad weather conditions (
After the pre-nesting period, bees return to their natal nesting site to check for empty nesting places in the surroundings (within a distance defined by the typical homing distance r50, see section "Foraging and dispersal"). If free nesting sites are unavailable in the vicinity closest to the natal nesting site, then dispersal occurs to look for new nesting sites outside the foraging distance (within maximum homing distance range r90). If there are still some eggs left to lay after constructing a first nest, the bee will look again for a new nesting site.
When building their nests, females of O. bicornis use pre-existing cavities, which are usually tube-shaped. The bees are readily attracted to holes in dry stems of hollow plants (e.g. the common reed), dead wood or even cavities in building walls or wooden fences. The cells in the nest are arranged linearly in series with transverse partitions between, built by the female who collects mud and mixes it with saliva. According to
Therefore, whenever a bee finds a patch with available nesting places, a nesting material (a reed, cavity etc.) of certain dimensions (length and diameter) is selected, which corresponds with a certain number of constructed cells. In general, the longer and wider the nesting material, the more cells can be constructed (
In the model, the minimum duration of pre-nesting will be set to 2 days with good weather conditions allowing for O. bicornis flying (flying conditions), that is, almost no rain (daily sum of precipitation < 0.1 mm) and no strong wind (i.e. wind speed < 8 m/s) and mean daily temperatures > 13°C (
In the model, the nest will be defined as a linear structure of suitable diameter to allow the construction of O. bicornis cells. For simplicity, instead of describing nest size by two dimensions (length and diameter), we will characterise nest size through a possible number of cells that can be built within it.
After emergence, when a female bee finds a suitable nesting patch with free nesting places (see section "Use of resources"), it will select a nesting cavity of a certain size (allowing to build a certain number of cells). The number of planned cells (in optimal conditions) in the first nest will be randomly selected from a beta distribution X ~ Beta (α, β) with assumed parameters α = 1.8 and β = 5 with minimum and maximum values defined by the variables OsmiaMinNoEggsInNest and OsmiaMaxNoEggsInNest (Fig.
The number of cells in consecutive nests will decrease by two cells per nest (DecreaseStepNestSize), following the study by
The principle of risk spreading as well as the unpredictability of life expectancy, weather conditions and floral resources, favour a mixed-sex allocation in O. bicornis. The advantages of mixed nests force mother bees to shift the proportion of daughters and sons between different nests, as observed in the field, instead of a sex-assorted investment (
The probability of an egg being fertilised (female egg) depends on its position in the nest, i.e. non-fertilised eggs (sons) are laid closer to the nest entrance to lower the probability of female eggs (daughters) being attacked by parasites. The sex ratio (proportion of daughters) per nest depends on:
In the model, the reproductive potential of the female will be characterised as the total number of eggs the bee can produce in her lifetime (property given to an adult female after emergence) and will depend on the bee mass. Therefore, we will assume that a given female of a mass femaleAdultMass can produce, on average, the following number of eggs per nest (see also Fig.
\(noEggsNest=0.0371 \times femaleAdultMass+ 2.8399 (+/-3 \space eggs)\) (Eq. 15)
If we assume that a bee can colonise a maximum of totalNestsPossible nests in her lifetime, the maximum total number of eggs to lay will be defined as:
\(totalEggsPossible=totalNestsPossible \times noEggsNest \qquad \qquad\)(Eq. 16)
For example, if totalNestsPossible = 5, then the maximum total number of eggs to lay increases with the mass of the mother bee, as shown in Fig.
In the model, we will combine the progeny sex ratio dependence on both the mother bee mass and reproductive age. The dependence on reproductive age (Fig.
\(Logistic \space (x \space | \space x0, \space min, \space max, \space k)={min+(max-min) \over{1+exp(-k \times (x-x0))}} \qquad \qquad\)(Eq. 17)
We will consider that the dependence on body mass (Figure 13), showing that larger females produce more daughters, is valid for reproductive age zero. We will use a linear equation being a mean of those provided by
\(sexRatio \space (age=0, \space femaleAdultMass)=\\ 0.0055 \times femaleAdultMass-0.1025\) (Eq. 18)
We will then create a family of logistic curves (Fig.
\(sexRatio \space (age, \space femaleAdultMass)=\\0.0914+{{0.0055 \times femaleAdultMass-0.1939} \over {1+exp(0.3921 \times (x-14.9026))}} \qquad \qquad\)(Eq. 19)
In the model, the family of logistic curves for the sex ratio will be calculated for bee sizes with a step defined by the OsmiaAdultMassCategoryStep variable. Then each mother bee will be linked with the proper curve using the BeeSizeScore2 parameter (see section "Osmia mass").
Females provision brood cells with pollen and a comparatively low proportion of nectar (2-4%) (
The provisioning level differs between male and female progeny and is significantly higher for females. Provision mass is highly correlated with the cocoon mass of the offspring (and, therefore, the body mass of the emerging bee). Under natural conditions, incompletely used provisions rarely occur in O. bicornis nests (
There is an optimal body size in both sexes of O. bicornis, i.e. bees scatter the size of their offspring around this optimum, irrespective of their own body size (
Therefore, under optimal/favourable conditions (unlimited pollen resources close by), there is a maximum amount of provision that a female is willing to provide to the nest and this depends on bee size. As bee provisioning efficiency decreases with bee age (
However, suppose the conditions are unfavourable (pollen resources are scarce, travel costs are high or weather conditions are bad for a longer time). In that case, food gathering is interrupted by the time norm and the egg is laid in an underloaded cell (
Females of O. bicornis spend about 80% of the total time for cell preparation in collecting food (
In O. bicornis, the rate of provisioning of brood cells decreases with the age of the female (
After finding a suitable nesting location (see section "Nesting"), a bee will start to build a nest. Each bee will have a 'plan' to build n cells (based on their nesting plan; see section "Nesting") with a given sex ratio (based on its age and mass; see section "Reproduction"), with a planned (or target) cocoon mass for each cell in the nest. Next, given the target cocoon mass, the amount of provision required will be given by Equations 5 and 6 in section "Osmia mass".
What is important, due to landscape (pollen resource availability), weather conditions, and other factors (e.g. death of bee or nest abandonment), the execution of the provisioning plan will not be guaranteed. Brood cells may be provisioned less than planned, the sex of the egg may be changed if provisions are too low and the total number of planned eggs per nest might be reduced.
Provisioning plan for the first female cell in the nest
The planned (target) cocoon mass for the first female in the nest will depend on the age and mass of the mother bee and we explain this next. The planned (target) cocoon mass of the following female and/or male cocoons in the nest will follow another pattern.
The dependence on the reproductive age of the mother bee will be a fitted logistic function, according to
The mass dependence will be based on data from
\(avgFemaleCocoonedMass=0.46 \times femaleAdultMass+\\63.85 (+/- \space 5-7 \space mg)\qquad \qquad\) (Eq. 20)
Based on
\(firstFemaleCocoonedMass=avgFemaleCocoonedMass+\\OsmiaLifetimeCocoonMassLoss/2 \qquad \qquad\) (Eq. 21)
By combining Equations 20 and 21, we have:
\(firstFemaleCocoonedMass (age=0, femaleAdultMass)=\\0.46 \times femaleAdultMass+63.85+\\OsmiaLifetimeCocoonMassLoss/2 (+/− 5−7 mg)\) (Eq. 22)
The mass of cocooned offspring for the first female cell in the first nest will be calculated, based on the family of curves defined by Equation 23 (Fig.
\(firstFemaleCocoonedMass \space (age, \space femaleAdultMass) = \\104.1982+{{0.46 \times femaleAdultMass-168.0482}\over {1+exp(0.1769 \times (x-18.0409))}}\qquad \qquad\) (Eq. 23)
\(\)The mass of cocooned offspring will be recalculated to the mass of provisions, based on Equation 5 (see section "Osmia mass").
Provisioning plan for the following cells in the nest
In a given nest, the mass of female cocooned offspring in consecutive cells will decrease at a fixed rate. Furthermore, based on
The total number of cells and sex ratio planned in the nest will be known from the nesting plan (see section "Reproduction"). The mass of the first female cocooned offspring will be calculated from the equations corresponding to Figure 19. The mass of the following female cocooned offspring will be calculated as follows:
\(femaleCocoonedMass[i]=femaleCocoonedMass[i-1]\\-{OsmiaNestCocoonMassLoss \over noFemaleCocoons} \qquad \qquad\) (Eq. 24)
Regarding male cocooned offspring, we assume that there is a common minimum amount of provision needed for them (regardless of the order of nest and cell in a nest), defined by the variable MaleMinTargetProvissionMass.
Time and habitat conditioning – deviations from the provisioning plan
All the above will represent the 'plan' of the bee. It can be reached only in optimal (favourable) environmental conditions, i.e. if there are no restrictions on floral (pollen) resources around the nest and weather conditions allow for constant provisioning. Therefore, we assume there will be a certain minimum effective cell construction time, tmin and a maximum acceptable cell construction time, tmax. This maximum is because the longer the cell is open (while the bee is provisioning), the higher the risk of parasitism (
If weather conditions allow, the bee will gather a certain amount of provision per day depending on the availability of food resources around the nest (see section "Use of resources"). Under optimal environmental conditions (enough pollen available in the proximity of the nest, good weather), the bee can fulfil her provisioning plan in tmin, that is, a maximum provision can be gathered in minimum time. However, if the floral resources are suboptimal, the amount of provision possible to collect in tmin will be smaller than the planned maximum and the offspring produced will be smaller. In such a case, the bee will continue provisioning the same cell (but it will increase the risk of parasitism) until:
In the case of the female progeny, if the stopping rule (1) is triggered before the stopping rule (2), then even if the provisioning plan has not been achieved, the mother bee will still lay a female egg if the amount of provision is above a defined threshold for female eggs. If not, then there will be a change in plan and a male egg will be laid (and the bee will continue to lay only male eggs afterwards for this nest attempt). If the provision is too small, even for a male egg, the current nest location will be given up. The procedure will continue for the next cells, with the amount of provisioning decreasing slightly for consecutive cells of daughters and sons and will be finished after the planned number of cells for this nest is reached. The bee will then start building another nest, following the same rules, but with a new ‘plan’ (i.e. number of eggs, sex ratio, amount of provision in each cell).
Weather will influence this plan, but only in such a way that, when the weather is bad (too cold, too rainy or too windy), the bee will have to wait in the nest and cannot provision (but the risk of parasitism will not increase as the bee stays in the nest and protects it).
Mortality at different developmental stages of O. bicornis in both natural and laboratory conditions was investigated by
Mortality rates of different developmental stages investigated by
Both
Overwintering temperatures (temperatures during diapause) do not influence overwintering mortality (
Maximum survival is obtained for individuals pre-wintered for short to intermediate (15–30 days) periods (
Assessing the background mortality of O. bicornis outside the nest is difficult. According to
Mortality of the egg-to-pupa stages
As all studies suggest that mortalities are very low in the prepupa and pupa stages and independent of temperature treatment, we will assume a constant daily mortality rate for these stages in the model. These being the mean of the results obtained by
Regarding mortality at earlier developmental stages, we will not differentiate between egg and larva, but rather define mortality for the egg to cocoon. As there are not enough data for O. bicornis to assume temperature dependence, we will assume a constant daily mortality rate. This daily constant is defined as the mortality experienced by bees in a fluctuating temperature regime of 10-25°C in the study by
We do not plan to assume temperature-dependent egg-to-cocoon mortality, as this would require substantial extrapolation and manipulation of existing data (Fig.
Overwintering mortality
Although it has been shown that the duration of both the pre-wintering and the wintering periods can influence overwintering mortality in Osmia spp. (
\(overwinteringMortality=0.05\times \sum{(DD\_prewinter)}-4.63 \qquad \qquad\) (Eq. 25)
The duration of the pre-wintering period will be defined according to the section "Overwintering". The overwintering mortality test will be applied only once, at the end of the overwintering period (before emergence).
Background mortality outside the nest
We will define the daily background mortality rate for an adult O. bicornis female outside the nest equal to 0.02.
Parasites attacking O. bicornis and bees of the Osmia genus, in general, can be divided into cleptoparasites and parasitoids.
Cleptoparasites lay eggs in bee nests and their larvae consume the host bee’s food provision and sometimes the host egg or larva. Within this group, the most common in O. bicornis nests are:
Parasitoids lay eggs within or adjacent to bee larvae or eggs and their larvae develop and consume the bee larva as a host. The most common parasitoids of this group are the following:
Most cleptoparasites and parasitoids represent the type of open-cell parasitism. Only the cleptoparasite mite Ch. osmiae and the parasitoid wasp M. obscurus are important O. bicornis parasites that do not belong to this group.
Several factors affect parasitism rates in the genus Osmia.
In the model, we will consider only open-cell parasitism. Its probability will depend on the time the cell is open according to the equation provided by
\(openCellParasitismProb=0.022\times timeCellOpen \space [h] \qquad \qquad\)(Eq. 26)
If the cell is parasitised, it will be in the form of cleptoparasitism by flies and wasps (e.g. C. indagator drosophilid fly and Sapygid wasps) or parasitism by flies (e.g. A. anthrax bombylid fly), with a given probability defined by the OsmiaBombylidProb variable.
Parasitism by flies will influence the offspring in two ways:
In the case of cleptoparasitism, we know that the influence of cleptoparasites on offspring depends on the number of eggs laid in the host cell by the parasite and these do not always cause the death of the bee larvae, i.e. if only a few eggs are laid, the cleptoparasite larvae do not have to eat the entire bee provision. However, since, in most cases, the number of cleptoparasite eggs is large enough to cause the removal of the entire or most of the bee supply, we will assume that if the cleptoparasite attacks the cell, it always results in the death of the bee larva.
We will not consider the influence of habitat on the probability of parasitism but we plan to add this impact in the next versions of the model development.
O. bicornis depends on both nesting and foraging habitats. It may use a wide range of natural nesting cavities, such as dry stems of hollow plants (e.g. the common reed), dead wood or even cavities in building walls or wooden fences. A five-year experimental study by
On the other hand, it has been suggested that food limitation is not common in O. bicornis at natural densities, which are far from their carrying capacity (
Evaluation of nesting resources
Nesting density will be provided as the number of nests available per 1 m2 of each landscape element in the land-cover map. For each type of landscape element, the minimum and maximum nesting density will be defined and the actual nesting density of a landscape element will be driven from this range of values at the beginning of each simulation.
The minimum and maximum nesting density will be calculated, based on the cavity nesting suitability provided by
Minimum and maximum nesting density defined for different types of landscape elements (TOLE) of the land-cover map in ALMaSS, based on nesting suitability values for cavity-nesting bees provided by
|
Type of landscape element in ALMaSS (TOLE) |
Habitat type (according to
|
Nesting suitability (according to Koh et al. (2016)) for cavity-nesting bees |
Min nesting density [no of nests/m2] |
Max nesting density [no of nests/m2] |
||
|
Mean |
Low25 |
Upp75 |
||||
|
Building |
- |
- | - | - |
0.00 |
0.00 |
|
UrbanNoVeg |
Developed / High Intensity |
0.14 |
0.00 |
0.15 |
0.01 |
15.39 |
|
UrbanVeg |
Developed / Low Intensity |
0.21 |
0.00 |
0.34 |
0.25 |
34.43 |
|
Garden |
Developed / Open Space |
0.24 |
0.01 |
0.41 |
0.67 |
41.42 |
|
AmenityGrass |
- |
- | - | - |
0.00 |
0.00 |
|
RoadsideVerge |
Grass |
0.20 |
0.00 |
0.32 |
0.12 |
31.58 |
|
Parkland |
Developed / Open Space |
0.24 |
0.01 |
0.41 |
0.67 |
41.42 |
|
StoneWall |
- |
- | - | - |
0.00 |
0.00 |
|
BuiltUpWithParkland |
Developed / Low Intensity |
0.21 |
0.00 |
0.34 |
0.25 |
34.43 |
|
UrbanPark |
Developed / Open Space |
0.24 |
0.01 |
0.41 |
0.67 |
41.42 |
|
Field |
- |
- |
- |
- |
0.00 |
0.00 |
|
PermPastureTussocky |
Pasture / Hay |
0.25 |
0.02 |
0.41 |
2.13 |
41.48 |
|
PermPastureLowYield |
Pasture / Hay |
0.25 |
0.02 |
0.41 |
2.13 |
41.48 |
|
UnsprayedFieldMargin |
Grass |
0.20 |
0.00 |
0.32 |
0.12 |
31.58 |
|
PermanentSetaside |
Grass |
0.20 |
0.00 |
0.32 |
0.12 |
31.58 |
|
PermPasture |
Pasture / Hay |
0.25 |
0.02 |
0.41 |
2.13 |
41.48 |
|
DeciduousForest |
Deciduous Forest |
0.72 |
0.51 |
0.97 |
51.34 |
97.28 |
|
Copse |
Shrubland |
0.67 |
0.45 |
0.94 |
45.22 |
93.87 |
|
ConiferousForest |
Christmas Trees |
0.32 |
0.01 |
0.61 |
0.99 |
61.02 |
|
YoungForest |
Shrubland |
0.67 |
0.45 |
0.94 |
45.22 |
93.87 |
|
Orchard |
Orchard |
0.25 |
0.00 |
0.44 |
0.18 |
44.33 |
|
BareRock |
- |
- | - | - |
0.00 |
0.00 |
|
OrchardBand |
Orchard |
0.25 |
0.00 |
0.44 |
0.18 |
44.33 |
|
MownGrass |
Grass |
0.20 |
0.00 |
0.32 |
0.12 |
31.58 |
|
MixedForest |
Mixed Forest |
0.68 |
0.45 |
0.96 |
45.28 |
95.62 |
|
Scrub |
Shrubland |
0.67 |
0.45 |
0.94 |
45.22 |
93.87 |
|
PitDisused |
Barren |
0.08 |
0.00 |
0.09 |
0.05 |
9.40 |
|
Saltwater |
Open Water |
0.00 |
0.00 |
0.00 |
0.00 |
0.00 |
|
Freshwater |
Open Water |
0.00 |
0.00 |
0.00 |
0.00 |
0.00 |
|
Heath |
Herbaceous Wetlands |
0.15 |
0.00 |
0.19 |
0.04 |
18.64 |
|
Marsh |
Wetlands |
0.25 |
0.00 |
0.46 |
0.17 |
45.93 |
|
River |
Open Water |
0.00 |
0.00 |
0.00 |
0.00 |
0.00 |
|
RiversideTrees |
Woody Wetlands |
0.34 |
0.01 |
0.69 |
0.94 |
68.89 |
|
RiversidePlants |
Wetlands |
0.25 |
0.00 |
0.46 |
0.17 |
45.93 |
|
Coast |
- |
- | - | - |
0.00 |
0.00 |
|
SandDune |
- |
- | - | - |
0.00 |
0.00 |
|
NaturalGrassDry |
Grassland Herbaceous |
0.58 |
0.31 |
0.88 |
30.78 |
87.98 |
|
ActivePit |
- |
- | - | - |
0.00 |
0.00 |
|
Railway |
- |
- | - | - |
0.00 |
0.00 |
|
LargeRoad |
- |
- | - | - |
0.00 |
0.00 |
|
SmallRoad |
- |
- | - | - |
0.00 |
0.00 |
|
Track |
- |
- | - | - |
0.00 |
0.00 |
|
Hedges |
Shrubland |
0.67 |
0.45 |
0.94 |
45.22 |
93.87 |
|
HedgeBank |
Grass |
0.20 |
0.00 |
0.32 |
0.12 |
31.58 |
|
BeetleBank |
Grass |
0.20 |
0.00 |
0.32 |
0.12 |
31.58 |
|
FieldBoundary |
Grass |
0.20 |
0.00 |
0.32 |
0.12 |
31.58 |
|
RoadsideSlope |
Grass |
0.20 |
0.00 |
0.32 |
0.12 |
31.58 |
|
MetalledPath |
- |
- | - | - |
0.00 |
0.00 |
|
Carpark |
- |
- | - | - |
0.00 |
0.00 |
|
Churchyard |
Developed / Open Space |
0.24 |
0.01 |
0.41 |
0.67 |
41.42 |
|
NaturalGrassWet |
Grassland Herbaceous |
0.58 |
0.31 |
0.88 |
30.78 |
87.98 |
|
Saltmarsh |
- |
- | - | - |
0.00 |
0.00 |
|
Stream |
- |
- | - | - |
0.00 |
0.00 |
|
HeritageSite |
- |
- | - | - |
0.00 |
0.00 |
|
Wasteland |
Barren |
- | - | - |
0.05 |
9.40 |
|
UnknownGrass |
Grass |
0.20 |
0.00 |
0.32 |
0.12 |
31.58 |
|
WindTurbine |
- |
- | - | - |
0.00 |
0.00 |
|
Pylon |
- |
- | - | - |
0.00 |
0.00 |
|
IndividualTree |
Mixed Forest |
0.68 |
0.45 |
0.96 |
45.28 |
95.62 |
|
PlantNursery |
- |
- | - | - |
0.00 |
0.00 |
|
Vildtager |
Flowers |
0.09 |
0.00 |
0.08 |
0.01 |
8.31 |
|
WoodyEnergyCrop |
Tree Crops |
0.37 |
0.02 |
0.73 |
1.54 |
72.78 |
|
WoodlandMargin |
Shrubland |
0.67 |
0.45 |
0.94 |
45.22 |
93.87 |
|
PermPastureTussockyWet |
Pasture / Hay |
0.25 |
0.02 |
0.41 |
2.13 |
41.48 |
|
Pond |
Open Water |
0.00 |
0.00 |
0.00 |
0.00 |
0.00 |
|
FishFarm |
Open Water |
0.00 |
0.00 |
0.00 |
0.00 |
0.00 |
|
RiverBed |
Open Water |
0.00 |
0.00 |
0.00 |
0.00 |
0.00 |
|
DrainageDitch |
Grass |
0.20 |
0.00 |
0.32 |
0.12 |
0.00 |
|
Canal |
Open Water |
0.00 |
0.00 |
0.00 |
0.00 |
0.00 |
|
RefuseSite |
- |
- | - | - |
0.00 |
0.00 |
|
Fence |
- |
- | - | - |
0.00 |
0.00 |
|
WaterBufferZone |
Wetlands |
0.25 |
0.00 |
0.46 |
0.17 |
45.93 |
|
Missing |
- |
- | - | - |
0.00 |
0.00 |
Evaluation of food resources
We assume that adult bees are not limited by their food requirements; therefore, we will focus only on assessing pollen resources for offspring. We also assume that the model providing a spatially and temporally changing pattern of pollen resource density will be available as a part of the ALMaSS landscape model (it is currently under development). Thus, on each day of the simulation, a bee will be able to sample the landscape to obtain information on the amount of pollen produced in a given resource patch (in mg per m2), taking into account only pollen produced by plants visited by O. bicornis. The plants that compose each habitat type will be divided into those used and not used by O. bicornis (i.e. qualitative approach based on literature and expert knowledge). In addition, competition from other pollinators will be modelled as a global decrease in floral resources available to O. bicornis.
Since the foraging patterns are not well known for O. bicornis, there is little information on which to base a search algorithm. It is clear, however, that O. bicornis makes an active choice of forage locations and it seems sensible to assume that they explore the area around the nest, likely with the efficiency decreasing with distance. To implement this, we suggest that we use a radial-spoke pattern of searching from the nest location. When the bee needs to find a forage location, it will search in eight directions (N, NE, E, SE, S; SW, W, NW) and select the best forage location encountered within the typical homing distance r50 (see section "Foraging and dispersal"). Since the pattern is radial with the spokes evenly distributed, the chance of encountering a patch of pollen decreases with distance from the nest and the likelihood of encounters near the nest is very high. The bee will continue to forage from the patch found as long as the pollen return rate is high enough (threshold parameter, for example, 75%) and a new nest is not started. If either condition is triggered, a new search will be initiated. This algorithm does not map to any measurable biological parameters, except forage distance; hence the step size, acceptance level and frequency of search will need to be fitted.
The Osmia model is based on a relatively large body of literature; however, many processes have uncertain parameter values and mechanisms. Information comes primarily from laboratory studies or studies using artificial nests to manage studied populations. In some cases, the data come from species other than O. bicornis. This means that some crucial information about the species is still missing. Important missing knowledge includes the actual population densities found in the field in different habitats and under other conditions, adult emergence time and reproductive success under natural conditions.
The problem expected to be encountered in the calibration is that the development times measured under constant laboratory conditions are very different from those experienced by the bees in the wild. In the field, temperatures fluctuate and are experienced differently by different nests with different locations (aspect, height, exposure). It is also still to be determined to what extent a nest itself may buffer natural ambient temperature fluctuations. The fluctuating temperatures are considered in the model through the input weather file. The weather input file consists of real daily weather data; thus, the development will be assessed, based on the mean daily temperature. However, there is a variation in temperature within the day, which is not considered in this version of the model. In future versions, we plan to implement hourly weather data, which are now available for Europe (
The timing of adult emergence in Osmia spp. depends on the overwintering conditions. Longer overwintering periods result in shorter emergence periods (
Bee mass loss during overwintering is not included in this version of the model. As
Our model assumes a positive relationship between bee mass and fecundity (total number of offspring produced and the sex ratio), following studies on O. bicornis by
As the available data for Osmia spp. are mainly from laboratory studies or artificial nests, we need more information on population density regulation mechanisms. We know that relying on expert opinion to set the habitat-related nesting densities is the key driver since the model does not include other regulating feedback. We do not expect the food limitation or the weather to act that way. One key aspect of the density-dependence response of Osmia spp. in natural populations is related to parasitoids. The relationship is, however, complicated. Therefore, we decided to defer the development of a full parasitoid model until a later model version. This is because the response picture varies in time and space (see section "Parasitism") and would require a separate modelling study to untangle the relationships with the parasitoid community. Theoretically, the parasitism rate might be positively related to the density of nests. Still,
The foraging decisions are relatively simple in the first proposed version of the model. The foraging algorithm is a computational construct that must be calibrated in the final model. However, so far, we have assumed that bees forage for pollen and nectar at the same sites and that they are not constrained by nectar availability (i.e. there is always enough nectar to meet their energy needs). We recognise that this is an oversimplification as not all plants provide pollen and nectar; therefore, bees may have to make additional flights to collect nectar. This may increase foraging time, affect provisioning efficiency and increase the time required for nest construction, especially when floral resources are scarce. Unfortunately, there is a lack of data estimating food intake in Osmia spp. (
The foraging algorithm does not include pollen quality in its decision-making (except for a binary assessment). The degree of specialisation of the host plant amongst bees varies widely and O. bicornis is known to be a polylectic/polylege bee. Still, it also shows some preference towards trees, such as Quercus spp., Salix spp. and Rosaceae or herbaceous plants, such as Ranunculus spp. (e.g.
Experimental studies show that the availability of chemical elements in larval food shapes bee fitness. Suboptimal concentrations of certain nutrients in pollen produced by specific plant species reduce bee fitness (
In the first version of the model, we will not consider nectar as part of the larval diet. This is because only a few studies refer to the pollen/nectar ratio in Osmia spp. provisions and, of these, those on O. bicornis report a very low proportion of nectar (2-4%;
A new, but significant module to be added to the next version of the model will be the response to pesticides. This model will build on the mechanisms used for honey bees in the ALMaSS (
See summary of state variables planned to be implemented in the ALMaSS O. bicornis model in Table
Summary of state variables planned to be implemented in the ALMaSS O. bicornis model. Parameters with the type ‘Assumed’ are estimated.
|
Parameter name (including symbol) |
Value |
Units |
Parameter type |
Source |
|
Development in the nest |
||||
|
Lower development threshold (LDT) |
Table 3 |
°C |
Estimated from literature |
|
|
Sum of effective temperatures (SET) |
Table 3 |
degree days |
||
|
Maximum developmental speed for prepupa (OsmiaPrepupaDevelTotalDays) |
24.3 |
days |
||
|
Variation around maximum developmental speed for prepupa |
+/- 10 |
% |
Assumed |
- |
|
Overwintering |
||||
|
Start of pre-wintering |
1 September |
- |
Assumed |
- |
|
End of pre-wintering/onset of overwintering |
Time threshold D; Equation 1 & 2 |
- |
Modelled |
- |
|
Temperature threshold for end of pre-wintering/onset of overwintering (OsmiaInCocoonPrewinteringTemp Threshold) |
15 |
°C |
Estimated from literature |
|
|
Baseline temperature to calculate sum of effective temperatures during overwintering (OsmiaInCocoonOverwinteringTempThreshold) |
0 |
°C |
Assumed |
- |
|
End of overwintering |
1 March |
- |
Assumed |
- |
|
Emergence from the nest |
||||
|
Emergence counter (noDaysToEmerge) |
Equation 4 |
- |
Estimated from literature |
|
|
Temperature threshold for emergence counter (noDaysToEmerge) |
12 |
°C |
||
|
Emergence distribution |
Discrete; Figure 6 |
- |
Estimated based on data from laboratory experiment |
A. Bednarska (personal communication) |
|
Osmia mass |
||||
|
Minimum possible male mass (OsmiaMaleMassMin) |
6.5 |
mg |
Assumed |
- |
|
Minimum possible female mass (OsmiaFemaleMassMin) |
50 |
mg |
Assumed |
- |
|
Maximum possible female mass (OsmiaFemaleMassMax) |
200 |
mg |
Assumed |
- |
|
Size class step for female mass (OsmiaAdultMassCategoryStep) |
10 |
mg |
Assumed |
- |
|
Female adult mass (femaleAdultMass) |
Equation 8 |
- |
Estimated from literature |
|
|
Foraging and dispersal |
||||
|
Flying weather mean daily wind threshold |
8 |
m/s |
Estimated from literature |
|
|
Flying weather rain daily threshold |
0.1 |
mm |
||
|
Flying weather mean daily temperature threshold |
13 |
°C |
||
|
Intertegular span (IT_span) |
Equations 13 and 14 |
mm |
Estimated from literature |
|
|
Typical homing distance (OsmiaFemaleR50distance) |
Equation 12 |
m |
||
|
Maximum homing distance (OsmiaFemaleR90distance) |
Equation 11 |
m |
||
|
General movement distribution |
Beta; α = 10; β = 5 |
- |
Assumed |
- |
|
Nesting |
||||
|
Duration of pre-nesting (OsmiaFemalePrenestingDuration) |
2 |
days |
Assumed |
- |
|
Minimum number of eggs planned for a nest (OsmiaMinNoEggsInNest) |
4 |
- |
Assumed |
- |
|
Maximum number of eggs planned for a nest (OsmiaMaxNoEggsInNest) |
25 |
- |
Assumed |
- |
|
Distribution for the planned number of eggs in the first nest |
Beta; α = 1.8; β = 5 |
- |
Assumed |
- |
|
Decrease step for number of eggs in consecutive nests (DecreaseStepNestSize) |
2 |
- |
Estimated from literature |
|
|
Maximum number of nests possible for a bee (TotalNestsPossible) |
5 |
- |
Assumed |
- |
|
Maximum number of eggs possible for a bee (TotalEggsPossible) |
Equation 16 |
- |
Assumed |
- |
|
Reproduction |
||||
|
Maximum total number of eggs per bee |
Equations 15 and 16 |
- |
Estimated from literature |
|
|
Sex ratio for a given nest |
Equation 19 |
- |
Estimated from literature |
|
|
Provisioning |
||||
|
Minimum time to construct a cell in days (tmin) |
1 |
days |
Assumed |
- |
|
Maximum time allowed to construct a cell in days (tmax) |
4 |
days |
Assumed |
- |
|
Multiplier to re-calculate from cocooned adult mass (femaleCocoonedMass / maleCocoonedMass) to provision mass (provisionMass) |
3.247 |
- |
Estimated from literature |
|
|
Minimum amount of pollen needed to provision a male cell (MaleMinTargetProvissionMass) |
10 |
mg |
Estimated based on data from laboratory experiment |
A. Bednarska (personal communication) |
|
Maximum difference in cocoon masses produced by a bee in its lifetime (OsmiaLifetimeCocoonMassLoss) |
30 (+/- 5) |
mg |
Estimated from literature |
|
|
Planned female cocoon mass for the first cell in the nest (firstFemaleCocoonedMass) |
Equation 23 |
mg |
Estimated from literature |
|
|
Difference between maximum and minimum female cocoon mass in a nest (OsmiaNestCocoonMassLoss) |
15 (+/- 5) |
mg |
Estimated from literature |
|
|
Planned female cocoon mass for the next cells in the nest (femaleCocoonedMass) |
Equation 22 |
mg |
Assumed |
- |
|
Mortality |
||||
|
Daily unspecified mortality for eggs |
0.0014 |
- |
Estimated from literature |
|
|
Daily unspecified mortality for larvae |
||||
|
Daily unspecified mortality for pre-pupae |
0.003 |
- |
Estimated from literature |
|
|
Daily unspecified mortality for pupae |
||||
|
Overwintering mortality |
Equation 23 |
- |
Estimated from literature |
|
|
Daily background mortality rate for females (OsmiaFemaleBckMort) |
0.02 |
- |
Assumed |
- |
|
Maximum lifespan (OsmiaFemaleLifespan) |
60 |
days |
Assumed |
- |
|
Parasitism |
||||
|
Probability of open cell parasitism |
Equation 24 |
- |
Estimated from literature |
|
|
Probability of being parasitised by bombylid fly (OsmiaBombylidProb) |
0.5 |
- |
Assumed |
- |
|
Use of resources |
||||
|
Minimum nesting density per landscape element type |
Table 5 |
- |
Estimated from literature |
|
|
Maximum nesting density per landscape element type |
||||
|
Decrease in available pollen before a new pollen patch search is initiated (OsmiaPollenGiveUpThreshold) |
75 |
% |
Assumed |
- |
|
Proportion of pollen available to O. bicornis individuals (OsmiaDensityDependentPollenRemovalConst) |
1 |
- |
Assumed |
- |
The model development was supported by the National Science Centre, Poland (project no. 2015/19/B/NZ8/01939) and the European Union's Horizon 2020 Research and Innovation Programme (EcoStack project; grant agreement no. 773554).
The model development was supported by the National Science Centre, Poland (project no. 2015/19/B/NZ8/01939) and the European Union's Horizon 2020 Research and Innovation Programme (EcoStack project; grant agreement no. 773554).
Elżbieta Ziółkowska: Conceptualisation, Methodology, Investigation, Data Curation, Validation, Formal analysis, Visualisation, Writing - Original Draft, Writing - Review & Editing
Christopher J. Topping: Conceptualisation, Methodology, Investigation, Formal analysis, Writing - Review & Editing
Agnieszka J. Bednarska: Conceptualisation, Methodology, Writing- Reviewing and Editing
Ryszard Laskowski: Conceptualisation, Writing - Review & Editing, Project leadership, Funding acquisition
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.