Insect pollination is one of the most important keys for ecosystem services (Bartholomée and Lavorel, 2019). Bees, in particular, are efficient pollinators due to their ability to carry more pollen than all non-bee taxa (Bernauer et al., 2022). The giant honey bee, Apis dorsata Fabricius 1793, is one of the most common pollinators found in subtropical Asia, i.e., China (Sakagami et al., 1980) and Nepal (Thapa, 2001), to the tropical Asia, covers India (Reddy, 1980), Thailand (Wongsiri et al., 1996), Sri Lanka (Koeniger and Koeniger, 1980), Philippine (Ruttner, 1988), including Indonesia (Ruttner, 1988; Nagir et al., 2016; Dyahastuti et al., 2022; Zahara et al., 2022; Kahono et al., 2023). Apis dorsata colonies have an essential ecological role in the ecosystem as pollinators for crops and natural plant communities (Rattanawannee et al., 2023). Almost 40 plant species were known to interact with A. dorsata in Pakistan (Sajjad et al., 2017). They also pollinate at least 15 plant species in Malaysia (Momose et al., 1998), six and 17 plant species in Thailand (Suwannapong et al., 2013; Stewart et al., 2018), and eight plant species in Indonesia (Bramasta et al., 2023). Considering the large number of workers per colony, hairy body of workers, generalized visitation pattern, floral constancy, and higher flower visitation rate make them effective pollinators (Layek et al., 2023). They also have a high flight range and efficient communication when foraging (Ruttner, 1988). Apis dorsata is important for honey hunters in Indonesia as the bees produce economically valuable honey (Schouten et al., 2020).

This giant honey bee is also a keystone species in dipterocarp forests (Rattanawannee et al., 2023). Indonesian archipelagoes were dominated by tall dipterocarp trees of more than 60 m. These forests are characterized by seasonal flowering and the so-called “general flowering,” most canopy trees mass-flower within several months but only every 4-5 years (Appanah, 1993; Sakai et al., 1999). When these events occurred, one of the main flower visitors was found to be A. dorsata (Momose et al., 1998). Under natural conditions, in the highly diverse lowland tropical rainforest of Sumatra, A. dorsata can select more nutrient-rich flowers and change the target plant following seasonal changes or mass flowering events (Rosmarlinasiah et al., 2015). In Sumatra, the nesting tree of A. dorsata is known by local people as the Sialang tree, which refers to more than one tree species, such as Koompasia excelsa (Shwetha et al., 2023) and Gluta renghas (Gussuwana et al., 2015; Dyahastuti et al., 2022). However, since the mid-20th century, the rainforests in Sumatra, particularly in the lowland, have been logged on a large scale and heavily converted into monoculture plantations of acacia (Acacia sp.), rubber (Hevea brasiliensis) and oil palm (Elaeis guineensis) (Drescher et al., 2016).

Several studies have shown that the conversion of rainforest to a transformation system affects the foraging behavior of bee pollinators (Gervais et al., 2020; Pulungan et al., 2023). Due to the plantation-dominated landscape, limited support for tree nesting (personal observation), and nectarine flower trees, as shown in the pollen in the honey in this study, might have affected the low number of bee colonies that migrated to Kampar in the past years. Besides that, the high temperature in the studied area reached 40°C, which is unfavorable for this important pollinator in the forest. The high temperature of 40-45°C also triggered the bees in Bengal to migrate in May and June (Singh et al., 2007). In addition to the temperature, rainfall and wind velocity highly influence the migration of A. dorsata colonies to safe places (Abrol, 1992). An extensive haze from the forest fire occurred in Riau Sumatra in 2015 and increased the temperature (Lee et al., 2016; Kozan, 2019; Sze et al., 2019). Before 2015, each nesting tree in Kampar, Riau, was home to hundreds of A. dorsata (Hotma Barinah, personal communication). After the haze, only a few colonies of A. dorsata migrate to the same Sialang trees (Raffiudin, personal observation).

While an extensive rainforest conversion arises in Sumatra, how A. dorsata adapts to this land cover is still unknown. Therefore, more studies are needed to improve our understanding of the important role of ecological functions of bee pollinators linked with forest conversion in Indonesia. To understand the effects of land cover on the ecological behavior of A. dorsata, our study aimed to (1) investigate the foraging behavior of A. dorsata and (2) compare the floral composition of the honey produced by colonies of A. dorsata in two land cover types in the transformation systems of plantations landscape (PL) in Kampar (Riau) and the mixed forest-agriculture landscape (FL) in Kerinci (Jambi), Sumatra.

This study attempted to answer the question of how A. dorsata adapts to different land covers. Overall, the lowland rainforests in Sumatra converted into plantations of oil palm, rubber, and acacia (Drescher et al., 2016). Our study revealed that the land cover transformation in Kampar affected A. dorsata behavior; the giant honey bee flew mainly to the oil palm plantation ( Figure 3A ) to reach the pollen, as shown in the melissopalynological results ( Figures 6A, B ). Even though the distance between the nest of A. dorsata was approximately 1.9 km to the nearest oil palm plantation ( Figures 1C , 3C ). This long-distance flight to collect the oil palm pollen was also shown by A. dorsata in Kampar Island flying across the strait to reach the oil palm plantation in Belitung Regency (Bramasta et al., 2023). Indeed, Apis dorsata can also have long-distance flights during migration across continents (Robinson, 2012, 2021). Oil palm pollen is the typical source of A. dorsata in Marang, Terengganu, Malaysia, even though it is in a tropical rainforest (Ibrahim et al., 2012). The experiment of A. dorsata foraging distance in Bangalore, India, showed a shorter distance according to the habitat. The bees preferred to forage in the garden with a distance of 800-900 m due to the various cultivated flowering plants (Young et al., 2021b). In accordance, A. mellifera foraged at a shorter distance during the major flowering seasons compared to the non-flowering seasons (Couvillon et al., 2015).

Our observations of two colonies of A. dorsata in each location of the landscape are nesting in the same aggregation of a tree. Although both colonies are foraging in the same direction, colonies within aggregation might not be related to mother and daughter colonies; this was shown in the A. dorsata aggregated colonies in the same tree in Assam, India. By using microsatellite DNA, Paar et al. (2004) revealed that the colonies apparently are not genetically related. Due to the high genetic differentiation among the bee colonies, the management of the conservation should be carefully taken, especially since these giant honey bees migrate to their home at the same site in the coming year (Paar et al., 2000).

Both A. dorsata colonies in the same aggregation in the plantation-dominated landscape (PL) in Kampar were dominated by oil palm pollen, although in low concentrations ( Figure 4A ). The low concentration also indicates a lack of pollen availability in the surrounding nest area, supported by a moderate plant diversity index ( Table 1 ). Given the absence of a nectarine gland in oil palm (Silberbauer-Gottsberger, 1990), A. dorsata and other pollinators might suffer from the lack of important resources in the vicinity of the nest. The nectar resource in Kampar Riau remains unknown. However, it is possibly represented by the Myrtaceae species (Freitas et al., 2016) due to it being the second most abundant pollen grain found in the honey samples ( Figures 6A, B ). In the studied area of plantation-dominated landscape (Kampar), a hundred hectares of monoculture, Eucalyptus (Myrtaceae) is the nectar source. Myrtaceae is one of the nectar sources foraged by A. dorsata in West Bengal, India (Layek and Karmakar, 2018) and also in Kampak Island, Belitung, Indonesia (Bramasta et al., 2023).

We found that the pollen concentration of the honey from FL Kerinci was ten times higher compared to the PL in Kampar ( Figure 4A ). This finding was supported by the high foraging activity of A. dorsata in Kerinci ( Figures 2C, D ), which might have influenced the pollen concentration in honey. An experimental study of honey bee A. mellifera revealed that the amount and rate of pollen and nectar collection were positively correlated with the individual foraging experience of the bees (Klein et al., 2019). In addition, relationship analysis between the pollen contained in honey and vegetation types in the West Coast (WC) and Western Ghats (WG), India, revealed that floral resources also affected pollen concentration in honey (Hegde et al., 2023). Apis indica honey from the more diverse vegetation of the WG, showed a higher pollen concentration compared to those from the WC, which has less diverse vegetation (Hegde et al., 2023). In contrast, A. dorsata and A. florea honey from the less diverse vegetation of WC contained a higher pollen concentration compared to those from more diverse vegetation of WG. Thus, we suggest that the higher pollen concentration in A. dorsata honey from Kerinci could be shaped by two factors, i.e., different flower (pollen) resources and the foraging activities of the bees. This result was supported by a high number of pollen taxa found in the honey from forest-agriculture-dominated landscapes (FL) in Kerinci, which revealed more diversified pollen collected by the foraging honey bees with a total of sixteen pollen types ( Figures 4B , 6C, D ). Our result is concordant with the honey of A. dorsata produced in the more diverse vegetation of deciduous forests in Western Ghats, India, which contains a more diverse pollen type compared to the honey produced in less diverse vegetation in West Coast India (Hegde et al., 2023). No predominant pollen taxon was found in the FL Kerinci. Thus, our results in FL Kerinci suggest an admixture of open grassland and forest resources with the agriculture of coffee plantations surrounding the forest of the nesting trees of A. dorsata ( Supplementary Table 3 ). We found the Coffee sp. pollen in a low percentage in A. dorsata honey in FL Kerinci ( Figures 6C, D ). Coffee flowers are also foraged by A. dorsata binghami in the agroecosystem near Lore Lindu National Park, Central Sulawesi (Klein et al., 2002). Besides A. dorsata, the flower of Coffee sp. was also pollinated by the native cavity-nesting Asian honey bee A. cerana (Saepudin, 2014; Sari and Putra, 2015). Apis dorsata might pollinate a lesser number of plant species compared to the other Asian honey bees, A. cerana and A. florea (Stewart et al., 2018). This is presumably due to the A. dorsata needs mass flowering plants despite the lower number of plant species. This giant honey bee needs much more pollen due to the larger colony size compared to A. cerana and A. florea (Stewart et al., 2018).

The maximum foraging activity of A. dorsata occurred during morning times in both landscapes, while the minimum activity occurred at noon ( Figure 2 ). Foraging observation of A. dorsata on blooming Eucalyptus in South Gujarat, India, revealed the same phenomenon: the maximum and minimum foraging activity of A. dorsata occurred in the morning and during midday, respectively (Behera et al., 2018). Moreover, an extensive observation of A. dorsata foraging activity in Bangalore, India, revealed that the foraging peaks occurred during the morning (before 9 am) and evening twilight (after 5 pm) and still exhibit the foraging activity during the night time (Young et al., 2021a). The nocturnal activities of A. dorsata are supported by their higher eye sensitivity compared to A. cerana and A. florea (Somanathan et al., 2009). In addition, the flight activity of A. dorsata in forest-agriculture-dominated landscapes was three times higher than in plantation-dominated landscapes at noon ( Figure 2 ). This is presumably due to the blooming flowering plants surrounding the nesting tree in FL Kerinci ( Supplementary Table 3 ), which was also supported by the higher plant diversity in Kerinci compared to Kampar ( Table 1 ). This phenomenon in Kerinci might be due to the high nectar-sugar concentration that positively influences the foraging activity of A. dorsata (Abrol, 1992).

Besides nectar-sugar concentration, our result revealed that the foraging activity of A. dorsata was also positively affected by temperature and light intensity and negatively affected by humidity. The same pattern of environmental influence also occurred in the foraging time of A. dorsata in India (Abrol, 1992). On the other hand, we found a lack of blooming flowering plants in the Kampar location due to the high domination of Fabaceae seedlings surrounding the nesting tree. The high domination of Fabaceae seedlings represented by the highest IVI value in Kampar ( Supplementary Table 2 ) is presumably due to the high usage of Fabaceae plants for rehabilitation of the transformed land in Indonesia, including Sumatra (Wiryono et al., 2022).

In our study, A. dorsata in Kerinci migrates in March, and the highest number of migrations is in April-May (unpublished data). The blooming flowering plants are important for the survival of the bees, thus attracting them to migrate to their previous nesting site (Paar et al., 2004). The migration of A. dorsata is from October to June, which depends on the source of nectar flowering plants (Singh et al., 2007). While in Borneo, A. dorsata migrate after one-month peak flowering (Itioka et al., 2001). Different flowerings in certain seasons were found in West Bengal, India, attracting giant bees to migrate (Layek and Karmakar, 2018). In summer, the predominant pollen type that attracts the most is the blooming of Eucalyptus. In winter, Brassica sp. is predominant, followed by Borassus sp. in spring (Layek and Karmakar, 2018). The plant resources for A. dorsata in Mount Tinanggo Kolaka Southeast Sulawesi found a total of 237 types of flowering plants, and September is the flowering peak (Rosmarlinasiah et al., 2015).

Natural forests are dominant landscape habitats that provide nesting trees and food resources for honey bee species (A. dorsata, A. mellifera, and A. cerana). Based on our results, the composition of land cover in Kerinci, which is still dominated by natural forest, supported more pollen sources from different plant species (16 pollen types) for A. dorsata than the composition of plantation-dominated landscape in Kampar (3 pollen types) ( Figure 6 ). A similar pattern also occurred in A. mellifera, where the forest landscape increased the diversity in honeybee diets, particularly trees, which were the dominant floral source of bee bread (Cannizaro et al., 2022). Apis cerana also collected more pollen types as a food diet in deciduous forests (16-28 pollen types) than in agricultural regions (9-16 pollen types) (Jhansi et al., 1994). Land cover transformation, such as natural forests to oil palm plantations, change pollination as the habitat function loss (Dislich et al., 2016) and the loss of biodiversity (Meijaard et al., 2018). Land cover change from forest to other land cover could change the diet of bees and might lead to loss of pollination services (Atmowidi et al., 2007; Pot et al., 2010) where honey bees are known as crucial pollinators (Kovács-Hostyánszki et al., 2018). A study on pollination services revealed that Apis dorsata has an essential role in compensating for stingless bee decline as critical pollinators during drought season in agricultural land in Bangalore, India (Mukherjee et al., 2019).

Our investigation concluded that the effect of land cover on the pollen diversity in the honey of A. dorsata related to pollen source with flight direction of the bees from a landscape perspective. The differences in environment can give empirical data to identify flower resources used by A. dorsata. Moreover, the information on floral resources used by A. dorsata is crucial for establishing a deterministic link between land cover, plant composition, and pollinator population development. The foraging activities of A. dorsata in the plantation-dominated landscape in Kampar showed low flight activities, particularly during the noon. The pollen results in the honey from Kampar show low diversity and concentration. While in the forest-agriculture-dominated landscape, this giant honey bee showed high flight activities. The pollen analysis from the honey harvested in this area revealed high pollen concentration and diversity of pollen types.

Although our research did not cover the natural habitat of A. dorsata, the result revealed a more diversified foraging behavior of A. dorsata in a more diverse heterogenetic landscape compared to the monoculture habitat. Our comprehensive study of foraging behavior, melissopalynology, and vegetation analysis in two land cover types proposes that enriching plant diversity around plantation habitats is needed to increase flower resources for A. dorsata. In this way, the sustainability of socioeconomic and ecological functions in plantation-dominated landscapes is expected to be enhanced.

The original contributions presented in the study are included in the article/ Supplementary Material , further inquiries can be directed to the corresponding author/s.

Ethical approval was not required for the study involving animals in accordance with the local legislation and institutional requirements because the research method employed involved observation of the foraging behavior of honey bee Apis dorsata.

RR: Conceptualization, Investigation, Methodology, Validation, Writing – original draft, Writing – review & editing. MD: Formal analysis, Investigation, Project administration, Writing – review & editing. RN: Formal analysis, Investigation, Methodology, Project administration, Writing – review & editing. TS: Formal analysis, Writing – original draft, Writing – review & editing. ND: Methodology, Writing – review & editing. ES: Formal analysis, Investigation, Visualization, Writing – review & editing. VA: Formal analysis, Investigation, Visualization, Writing – review & editing. RM: Formal analysis, Investigation, Visualization, Writing – review & editing. SB: Conceptualization, Formal analysis, Investigation, Methodology, Visualization, Writing – review & editing. CS: Formal analysis, Writing – review & editing. LP: Formal analysis, Methodology, Validation, Visualization, Writing – review & editing. WP: Resources, Writing – review & editing. TA: Supervision, Writing – review & editing. AS: Writing – review & editing. HB: Conceptualization, Validation, Writing – review & editing.