A Comparative Study of Methods Recording Beekeeping Flora

Abstract

The knowledge of beekeeping flora of an area and of each plant’s provision is crucial for beekeepers to plan their hive transfers when practicing nomadic beekeeping. Thus, in the present study, we evaluated the use of pollen traps as a means of identifying beekeeping plants in target areas, assessing their pollen percentage contributions and estimating their flowering seasons. The results were compared with the classical field observation method, widely used for flora recording. In total, 98.8% of the plants were recognized by using pollen traps and 89.4% from observations in the field, while for 73% there was found an agreement between population size (Wittig Scale) and Pollen Contribution Scale (PCS). The results showed that pollen traps can be helpful tools not only for defining the presence or absence of a beekeeping plant, but also for providing important information regarding the population size of a taxon of major beekeeping importance in the area surrounding the apiary. Finally, the estimation of the flowering season and its maximum point can be accurately predicted by using pollen traps on plants of beekeeping interest.

1. Introduction

The use of natural resources (nectar for energy and pollen for growth) to feed honeybee colonies and produce honey is one of the main factors in successful beekeeping at a professional level [1,2]. For this reason, beekeepers schedule the growth of their colonies and seasonal management (e.g., transfers of hives, dividing colonies, production of hive products, etc.) based on the beekeeping flora of an area. Honeybees forage a large number of plants, but it still remains under discussion that there is a probable relation between bees’ preferences for a pollen source and its availability linked to flower features, such as architecture, color and odor [3,4,5], with Liolios et al. [6] highlighting the population size of beekeeping plants surrounding an apiary as the major factor affecting this relation.

In addition, climate change and its impact on the flowering of beekeeping plants creates problems and leads to failures in the moving schedule of the hives. Especially in water-limited ecosystems, variation in precipitation may lead to divergence of flowering time over the season [7]. Rafferty et al. [8], studying the flowering observations of 590 species of plants in five communities, observed linkages between temperature and phenology in colder ecosystems, while in drier, water-limited ecosystems, the interactive effects of both temperature and rainfall on the phenological changes of beekeeping plants are less understood.

Greece, due to its Mediterranean climate, provides high floristic heterogeneity and diverse physiognomies, with some areas containing more trees, while others are home to a larger number of shrubs or herbs, offering a rich biodiversity for bee foraging. Nevertheless, there is little information on the peak blooms of beekeeping plants in the different areas, while the transfer of hives is mainly based on beekeepers’ previous experiences or rumors, with the result that they are often incorrect. Also, the increase in the number of bee colonies and the limitation of the beekeeping flora due to a multitude of factors (extensive fires, degradation of ecosystems, etc.) make the transfer of hives in different areas a necessary condition for satisfactory productivity. Indeed, currently, there are few areas blooming all seasons that can sustain stable apiaries and probably mainly at an amateur level.

Thus, the knowledge of bee flora of an area as well as the estimation of the flowering duration of target beekeeping plants seems to be vital for beekeepers to optimize production [9], contributing remarkably to the planning of hive transfers and consequently to the reduction of production costs. Pollen traps, as a flora-recording tool, have been used in the past with considerable success [10,11,12,13], given that the composition of pollen loads can vary according to the region or season, indicating patterns and variations of the local flora [14]. Dimou et al. [15] also noted that it is possible to capture the beekeeping flora of an area using pollen traps, harvesting samples of a three-day trapping duration collected at six- to nine-day intervals through one year.

Considering all the above, in the present study, we applied and compared two methods to record the beekeeping plants in two target areas, by determining their population size through field observations and their pollen percentage contributions using pollen traps. We also inserted a new index and scale in order to identify the pollen contribution in the area surrounding the apiary, and compared the results to those from field recording. We finally investigated the flowering season of selected beekeeping plants, using both methods, with the ultimate goal of developing a reliable tool for the future study of climate change effects on the alteration of flowering time in plants of major beekeeping importance.

2. Materials and Methods

2.1. Installation of Experimental Apiaries/Pollen Collection

Six bee colonies of equal population (10 frames in each colony of Langstroth pattern) were transferred to two semi-mountainous areas (three in each area) with native beekeeping plants in the prefecture of Thessaloniki, Greece (Area 1: 40.553059, 23.050231, Area 2: 40.520636, 23.153601). The straight distance between the two apiaries was 10 km, while there was a high mountain mass between them. The experiment took place from early March to middle November 2022 in order to cover the whole beekeeping period. These areas were selected as targets, as they are rich in beekeeping flora, mainly herbaceous and bushy vegetation, and thus attract many beekeepers of the region.

After the transfer, plastic pollen traps with a separation grid were fitted to the entrance of the hives, and bee pollen was collected every 5 days, with the removal of the collection drawer. The samples after their harvest were placed in the freezer until the time of their separation, to avoid any alteration in texture and color.

2.2. Identification and Estimation of Population Size of Surrounding Flora

For the estimation of the population size of the surrounding flora, we recorded the plant species in the target areas every second week, in a distance about 3 km from the experimental apiaries, as bees usually collect within this radius around the perimeter of their hive. Having drawn an imaginary circle with a radius of 3 km on the map and within this area, the sampling took place. We sampled the entire perimeter and the transects were demarcated from the experimental apiaries. In each field observation, there was a team of 4 people, recording from 8 to 12 in the morning. For the calculations, we used a semiquantitative scale with four classes, according to the Wittig scale [16], where Class I contained scarcely-located taxa and class IV contained the dominant taxa (taxa covering at least 30% of the collection site).

2.3. Pollen Separation and Identification

The pollen loads were separated mainly based on color, shape and size. For the identification of pollen grains, the method of Louveaux [17] was used, according to which a small amount of pollen was placed on a slide and the pellets were dissolved with 2–3 drops of diethyl ether. After the evaporation of the solvent, one drop of aqueous isoglucose solution (2:1) was added on the slide to hydrate the pollen grains and one drop of aqueous fuchsine to color them. The slides were placed on a heating plate to remove the moisture and covered with coverglass containing a drop of Entellan. For the identification of the pollen grains, the database of the Laboratory of Apiculture-Sericulture, AUTH, was used. At the same time, slides were also prepared from flower stems collected from the fields around the bee colonies, followed by microscopic examination for their identification.

2.4. Estimation of Pollen Percentage Contribution of Polliniferous Plants

To evaluate the contribution of each plant taxon in the total amount of bee pollen collected from the traps, we analyzed a representative pollen sample in each sampling. Specifically, the total amount of collected pollen was weighed, recorded and then the quarter sampling method was repeatedly applied to ensure a random sample of 10% [13]. Visual separation and microscopic identification were followed, and the percentage of each taxon was calculated based on the formula:

$${\mathrm{P}}_{\mathrm{i}}\left(\%\right)=\frac{{\mathsf{\alpha }}_1{\mathsf{{\rm X}}}_1{+\mathsf{\alpha }}_2{\mathsf{{\rm X}}}_2+\dots {\mathsf{\alpha }}_{\mathrm{n}}{\mathrm{X}}_{\mathrm{n}}}{{\mathrm{X}}_1{+\mathrm{X}}_2+\dots {\mathrm{X}}_{\mathrm{n}}}$$

where

• P_i(%): The percentage of i plant taxon found in the pollen trap during its flowering period
• α_1…n: The amount of pollen of i taxon in the total amount of pollen in each sampling
• Χ_1…n: Total amount of pollen collected from the traps in each sampling

In order to facilitate the visualization of the percentage contribution of each species on the total amount collected during flowering season, as well as to be able to compare the results to the Wittig scale used for the estimation of population size in field observations, we created a scale of 4 classes (Pollen Contribution Scale, PCS); class IV corresponded to percentage participation in traps > 20% (dominant species), class III to 10%–20% (important species), class II to 2%–10% (moderately important species) and class I in 0%–2% (species of minor importance).

2.5. Flowering Season Determination: Flowering Rate Calculation

For the determination of flowering season, we targeted 10 beekeeping plants of major importance. The beginning of flowering was signaled when the first pollen loads of a species appeared in the pollen traps and ended when no more pollen loads were found. At the same time, in order to cross-check the results, a visual recording in the field was carried out. For field observation, we marked the areas and counted every five days the number of blooming flowers of the target plants. In general, we considered the beginning of blooming when taxa had at least 10% of their flowers open, while full flowering corresponded to 80% of blooming. In plants with single flowers, we counted the number of them, while in the case of other plants (i.e., trees) we assessed inflorescences of individual marked branches.

The whole study approach is given in Figure 1.

3. Results and Discussion

3.1. Identification of Surrounding Flora

In the present study, we recorded the surrounding flora in two target areas, comparing two methods, the field observation and the use of pollen traps. The scientific names of the taxon, the families, the flowering period in the target areas, and their presence in the field (population size) and in the pollen traps (Pollen Contribution Scale, percentage of plant taxon found in the trap) are given in

Table 1. Taxa recorded in two target areas (flowering period, presence in field, population size, presence in pollen traps, PCS, P_i%).

No.	Scientific Name	Family	Flowering Period in the Target Areas	Presence in Field	Population Size (Wittig Scale)	Presence in Pollen Traps	Pollen Contribution Scale (PCS)	Percentage of Plant Taxon Found in the Pollen Trap (Pi%)
1	Acer sp.	Spindaceae	March–April	√	II	√	I	1.2
2	Asphodelus aestivus	Asphodelaceae	March–July	√	II	√	I	1.9
3	Brassica rapa	Brassicaceae	April–June	√	III	√	III	17.8
4	Calendula arvensis	Asteraceae	March–April	√	I	√	I	0.8
5	Carduus armatus	Asteraceae	May–July	√	I	√	I	1.3
6	Carduus marianus	Asteraceae	April–June	√	II	√	II	2.5
7	Carthamus lanatus	Asteraceae	June–July	×	×	√	I	0.3
8	Castanea sativa	Fagaceae	April–June	×	×	√	I	0.9
9	Centauria solstitialis	Asteraceae	May–June	√	I	√	I	1.4
10	Centauria sp.	Asteraceae	May–June	√	II	√	II	4.6
11	Cephalaria sp.	Caprifoliaceae	May–June	x	x	√	I	0.1
12	Chenopodium album	Chenopodiaceae	August–September	√	II	√	III	18.3
13	Cichorium intubus	Asteraceae	June–August	√	II	√	II	8.7
14	Cirsium sp.	Asteraceae	April–September	√	II	√	II	2.9
15	Cistus creticus	Cistaceae	May–July	√	IV	√	IV	25.8
16	Cistus parviflorus	Cistaceae	May–July	√	III	√	IV	22.0
17	Convolvulus arvensis	Convolvulaceae	May–November	√	I	√	II	2.3
18	Crataegus monogyna	Rosaceae	April–May	√	I	√	II	2.1
19	Daucus carota	Apiaceae	May–August	×	×	√	I	0.1
20	Echinops ritro	Asteraceae	May–July	×	×	√	I	0.2
21	Echium plantagineum	Boraginaceae	May–August	√	I	√	I	1.8
22	Eleagnus angustifolia	Eleagnaceae	May	√	I	√	I	0.2
23	Epilobium parviflorum	Onagraceae	April–June	√	I	√	I	0.2
24	Erica arborea	Ericaceae	February–April	√	II	√	I	1.4
25	Erica manipuliflora	Ericaceae	September–October	√	IV	√	IV	20.6
26	Eriobotrya japonica	Rosaceae	November	√	I	√	I	0.2
27	Eucalyptus sp.	Myrtaceae	June–August	×	×	√	I	0.3
28	Ferula communis	Apiaceae	June–July	×	×	√	I	1.2
29	Genista acanthoclada	Fabaceae	May–June	√	II	√	II	2.3
30	Geranium macrostylum	Geraniaceae	May–June	√	I	√	I	0.6
31	Hedera helix	Araliaceae	August–October	√	II	√	II	6.3
32	Helianthus annuus	Asteraceae	June–August	√	I	√	I	1.3
33	Heliotropium europeum	Heliotropiaceae	July–September	√	II	√	I	0.6
34	Hypericum triquetrifolium	Hypericaceae	June–July	√	III	√	III	12.3
35	Inula viscosa	Asteraceae	August–October	√	II	√	II	7.3
36	Iridaceae	Iridaceae	March–April	√	II	√	II	2.9
37	Juglans nigra	Juglandaceae	May–June	√	I	√	I	1.8
38	Lamium sp.	Lamiaceae	March–April	√	III	√	II	4.6
39	Laurus nobilis	Lauraceae	April	√	I	√	I	1.1
40	Ligustrum japonicum	Oleaceae	June–July	√	II	√	II	6.8
41	Lonicera japonicus	Caprifoliaceae	May–June	√	I	√	I	1.3
42	Marticaria chamomilla	Asteraceae	March–April	√	III	√	II	7.4
43	Olea europaea	Oleaceae	April–May	√	IV	√	II	13.3
44	Onagraceae	Onagraceae	April–May	√	I	√	I	1.5
45	Onopordum acanthium	Asteraceae	May–July	√	II	√	II	2.2
46	Opuntia Ficus	Cactaceae	May–June	×	×	√	I	0.3
47	Ornithogalum pannonicum	Asparagaceae	June–July	√	II	√	II	2.7
48	Paliurus spina-christi	Rhamnaceae	May–June	√	III	√	III	19.2
49	Papaver rhoeas	Papaveraceae	March–June	√	IV	√	IV	21.4
50	Parthenocissus quinquefolia	Vitaceae	April–June	√	I	√	I	1.2
51	Pastinaca sativa	Apiaceae	April–May	√	I	√	I	1.3
52	Pinus halepensis	Pinaceae	April	√	III	√	I	0.4
53	Pimpinella peregrina	Apiaceae	May	√	I	√	I	0.1
54	Pitosporum tobira	Pitosporaceae	May–June	√	I	√	I	1.6
55	Polygonum aviculare	Polygonaceae	June–October	√	II	√	II	7.3
56	Portulaca oleracea	Portulacaceae	July–September	√	III	√	III	12.1
57	Prunus amygdalus	Rosaceae	March	√	II	√	II	7.6
58	Pyracantha coccinea	Rosaceae	April–May	√	I	√	I	0.4
59	Pyrus pyraster	Rosaceae	April	√	II	√	II	3.7
60	Quercus sp	Fagaceae	April–June	√	II	√	III	14.5
61	Ranunculus sp.	Ranunculaceae	April–May	×	×	√	II	4.6
62	Robinia pseudoacacia	Fabaceae	May	√	II	×	×	×
63	Rubus fruticosus	Rosaceae	May–September	√	II	√	II	2.6
64	Rubus ulmifolius	Rosaceae	June–August	√	III	√	III	10.1
65	Rumex crispus	Polygonaceae	April–June	√	II	√	II	6.5
66	Salix sp.	Salicaceae	April	√	I	√	I	1.6
67	Silybum marianum	Asteraceae	April–May	√	III	√	III	16.8
68	Sinapis arvensis	Brassicaceae	March–June, August–November	√	IV	√	IV	21.6
69	Sisymbrium irio	Brassicaceae	March–June, August–November	√	IV	√	IV	20.8
70	Sonchus asper	Asteraceae	March–June	√	III	√	III	10.1
71	Sorghum halepensis	Poaceae	June–August	√	III	√	I	0.2
72	Syringa vulgaris	Oleaceae	April–May	√	I	√	I	0.3
73	Tamarix sp.	Tamaricaceae	April–August	√	I	√	I	0.5
74	Taraxacum officinale	Asteraceae	March–June	√	III	√	III	17.3
75	Thymus sp.	Lamiaceae	March–June	√	I	√	I	0.6
76	Tilia intermedia	Malvaceae	June	√	I	√	III	12.2
77	Tribulus terrestris	Zygophyllaceae	June–October	√	III	√	III	11.6
78	Trifolium pratense	Fabaceae	May–June	√	III	√	III	11.5
79	Trifolium sp.	Fabaceae	April–July	√	III	√	III	19.2
80	Urtica dioica	Urticaceae	May–October	√	II	√	II	2.7
81	Verbascum sp.	Scrophulariaceae	May–July	√	II	√	II	2.3
82	Vicia villosa	Fabaceae	March–April	√	III	√	III	12.0
83	Vitex-agnus castus	Lamiaceae	June–September	√	II	√	II	3.3
84	Xanthium strumarium	Asteraceae	July–October	√	II	√	II	9.8
85	Zea mays	Poaceae	July–September	√	I	√	I	0.8

. In total, 84 different polliniferous taxa belonging to 49 families were recorded in the pollen traps, giving a recognition capability of about 98.8%, while the recognition in field visits was lower (about 89.4%), as 76 taxa were spotted providing pollen and/or nectar. The flowering of most taxa occurred in mid-to-late spring, while some of them flowered in early autumn.

The observations in the field were carried out at regular intervals (every second week) within a radius of 3 km from the colonies; however, some plants could not be found. Indeed, Carthamus lanatus, Castanea sativa, Cephalaria sp., Daucus carota, Eucalyptus sp., Ferula communis, Opuntia ficus, Ranunculus sp. and Echinops ritro were only detected in the pollen traps. The difficulty of accessing some points within the measurement field, mainly because of natural obstacles, combined with the difficulty of spotting some plants found in very small populations (e.g., herbs), is probably responsible for the deviation of the results between the two recording methods. In the cases of chestnut (Castanea sativa) and eucalyptus (Eucalyptus sp.), despite the fact that these trees are easily recognizable, it was not possible to record them in the field, as they were outside the range of the 3 km radius defined as the observation area in the field. Indeed, the flight of bees when foraging can in some cases reach 10 km [18,19,20]. On the contrary, acacia trees (Robinia pseudoacacia) were observed in the field observations, but the corresponding pollen pellets were not detected in the traps. Despite the high bee foraging of the acacia tree, the plant offers abundant nectar but no pollen [21].

According to the results of Table 1, it seems that the recording of beekeeping plants of an area is better performed by the bees themselves, avoiding laborious and time-consuming observations in the field, but requires knowledge of palynology, so as to identify the plants under the microscope. Additionally, in order to obtain more reliable results, it is recommended that the use of pollen traps are combined with recording in the field [22]. In the microscopic analyses carried out on bee pollen for species-level identification, in several cases, field observation and collection of flower samples is considered particularly important to create a database (photographs of pollen grains) to facilitate the microscopic identification of the collected pellets.

3.2. Population Size (Wittig Scale) and Percentage Contribution Scale (PCS) of Beekeeping Plants

The most dominant species according to the Wittig scale (class IV) were Cistus criticus, Erica manipuliflora, Olea europaea, Papaver rhoeas, Sinapis arvensis and Sisymbrium irio. These species were dominant in the PCS scale as well, with the addition of Cistus parviflorus, which in the Wittig scale was in level III. Between the population size of beekeeping plants in the field and their PCS, there was found an agreement of about 73% (62 taxa) and a deviation about 27% (Table 1). This discrepancy was mainly observed on the wind-pollinated species Pinus halepensis, Acer sp., Olea europaea and Sorghum halepensis, which were ranked as population size in the Wittig scale at levels III, II, IV and III, respectively, while their PCS were at levels I, I, II and I, respectively. These results could be attributed to the fact that bees prefer entomophilous species rather than anemophilous ones that usually provide pollen of poor nutritional value [6] (Liolios et al., 2016). Also, the species Lamium sp., Marticaria chamomilla, Heliotropium europeum, Erica arborea and Asphodelus aestivus were found in higher populations in the field than in the pollen traps. Lamium sp. and Asphodelus aestivus bloom very early in the spring, where adverse weather conditions probably inhibit bee flights and pollen collection. On the other hand, Marticaria chamomilla, Heliotropium europeum and Erica arborea bloom in spring at the same time with a variety of other plants, so bees are probably attracted to other plant sources, either with a stronger smell or better provisions. In the case of Chenopodium album, Cistus parviflorus, Convolvulus arvensis, Crataegus monogyna and Tilia intermedia, higher percentages were detected in the traps compared to the populations recorded in the field. The results highlight the use of pollen traps and the PCS as effective tools to identify not only the presence or absence of a beekeeping plant in an area, but also its population size in the area surrounding the apiary, with greater accuracy compared to field recording. The use of pollen traps and the color separation of pellets that can capture in several cases the extent of the population of the species in the field is also referred to in literature [13,23,24].

3.3. Flowering Season Determination: Flowering Rate Calculation

The flowering season of 10 taxa is depicted in Figure 2 based on the recording of the flowering stages of the target plants in the field and the percentage contributions of their pellets in the pollen traps.

In all cases, there was agreement of the data recorded from the field observations with those obtained from the analyses of the pollen samples collected from the traps [P_i(%)]. Indeed, on the dates when the beginning of flowering (at least 10% open flowers) was observed in the examined taxa, their first pollen loads were detected in the traps as well. Correspondingly, full flowering (80%) coincided with the maximum presence/percentage of the respective pollen loads in the traps. Although bees’ preferences for the amount of pollen collected is influenced by many parameters (e.g., plant abundance, plant supply, etc.), it seems that the estimation of the flowering period and its peak can be predicted accurately with the use of pollen traps on plants of beekeeping interest. A long-term study using pollen traps and P_i(%) could be further applied as an initial tool to record and better visualize the possible long-term alterations in the flowering of beekeeping plants in relation to climate change.

The results of the study confirm the strict linkage between honeybees, environment and biodiversity. The role of bees as bioindicators, as well as their importance in the conservation of biodiversity is also emphasized by other authors, highlighting the need for conservation measures to prevent the loss of honeybees and to preserve ecotypes and further world biodiversity [25,26,27,28,29].

The use of pollen traps to map beekeeping flora and estimate the suitable time for hive transfers in target areas is also suggested by other authors, such as Alves and dos Santos [24] for Sergipe (Brazil), Taha [30] for Al-Ahsa province (Saudi Arabia) and Taha et al. [31] for Kafrelsheikh province (Egypt), so that beekeepers move their apiaries to obtain high honey yield, supporting their colonies and/or economizing the cost of artificial feeding.

4. Conclusions

In the present study, we applied and compared two methods for beekeeping flora recording: the classical method of field observation and the use of pollen traps. Although the preference of bees in terms of the harvested quantity of pollen is influenced by many parameters (e.g., abundance of plants, plant supply, etc.), the estimation of the period and maximum flowering can be predicted with great accuracy with the use of pollen traps. Additionally, we created an index (P_i%) and a scale (PCS) to better visualise and compare the results from recording using pollen traps with the Wittig scale applied to determine the population size of plants in the field. The use of pollen traps for bee flora recording presents the advantage of ease of application, as bees travel long distances every day in order to collect pollen, while the removal of the collected pollen is easy, without additional beekeeping treatments. On the contrary, field recording requires time-consuming visits to the field, enabling the researcher to understand plant populations and their flowering stages, but does not provide access to information about bees’ feeding habits. Finally, the graphs exported from P_i(%), regarding the flowering season of 10 plants of major beeekeeping importance, could be further applied to target beekeeping plants in consecutive years to evaluate long-term alterations in their flowering affected by climate change. The results may be further used to design specific algorithms to find the appropriate beekeeping areas to transfer bee hives, minimizing costs and increasing honey yield.