Conservation Assessment and Chemistry of Boswellia ogadensis, a Critically Endangered Frankincense Tree

Boswellia ogadensis is a critically endangered species of frankincense tree, restricted to a small area of the Shabelle river valley in southern Ethiopia. It has only been recorded from two botanical collecting trips, in 1972 and 2006, with no indication of the abundance, threats, or population status of the trees, and it was listed on the IUCN Red List of Endangered Species as “Critically Endangered” in 2018. More recent expeditions, in 2019 and 2021, were not able to locate the species, raising concerns about its continued survival. We carried out a field survey in June 2022 to re-locate the species, assess the threat level it is facing, and collect samples of resin for analysis. This survey revealed that B. ogadensis is present in more locations than previously recorded, and is more abundant than thought. While it is facing multiple threats, including grazing, cutting for firewood, and insect attacks, these threats vary geographically, and there are populations that appear to be healthy and regenerating well. While more research is needed, the current survey indicates that downlisting to “Endangered” status may be appropriate. Samples of resin were also collected and analyzed using gas chromatographic techniques, revealing that while the essential oil profile is similar to that of other Boswellia species (dominated by α-thujene, α-pinene, p-cymene, and terpenin-4-ol), there are chemical markers that can distinguish it from other sympatric Boswellia species, indicating the potential for this to be used as a tool to monitor whether B. ogadensis is being harvested alongside other more common Boswellia species.


Introduction
The genus Boswellia Roxb. ex Colebr. (Burseraceae: Sapindales) consists of approximately 24 species of small to medium trees, typically characterized by papery, exfoliating bark, imparipinnate leaves, and the production of aromatic resin via a deep red resiniferous layer of bark [1]. The genus is widely distributed across west and east Africa, southern Arabia, and the Indian subcontinent, with the Horn of Africa region featuring the highest species concentration, particularly Socotra Island, which hosts almost half (11 of 24) of known Boswellia species [1]. The taxonomy of Boswellia is still dynamic, with recent years seeing multiple species added or occasionally removed; genetic work is ongoing [1][2][3].
The genus is best known for its production of a highly aromatic, terpenoid oleo-gumresin, called frankincense, which is produced and stored in resin canals in the bark [4,5]. The resin exudes whenever the bark is broken, either by an animal or insect, or intentionally by humans cutting the bark to extract the resin. Most frankincense has been widely used and traded around the world for thousands of years, and is considered to be one of the oldest internationally traded commodities [5,6]. It is prized for its use in traditional medicinal systems, such as Ayurveda and Traditional Chinese Medicine, as well as its use in religious ceremonies, cosmetics, and perfumery. Additionally, essential oils and extracts Plants 2022, 11, 3381 2 of 11 from frankincense resins have become increasingly popular, with several million kilograms of resin processed annually to meet these demands [5,7].
Many Boswellia species are facing significant sustainability challenges and probable or confirmed population declines due to a variety of factors [7][8][9][10][11][12]. Key threats to many species include grazing by ungulates (goats, cattle, camels), fire, land conversion for agriculture, improper or excessive harvesting of resin, and attacks by insects [7,8]. Grazing and fire kill seedlings or saplings, and in some cases can completely block the regeneration of the species [13]. Improper resin harvesting, often in combination with insect attacks, or land conversion kills adult trees and can result in the complete conversion of Boswellia woodlands. The combination of these threats can be significant; studies on Boswellia papyrifera in northern Ethiopia, Eritrea, and Sudan have projected wide-scale population reductions of more than 70% on average within 25 years [8]. Other studies have noted sustainability concerns in B. sacra and B. frereana in Somaliland [9] and Oman [14], B. serrata in India [12], and B. elongata and other Boswellia species on Socotra island [10,11]. While studies have primarily focused on the major commercial species-with the exception of several studies conducted on the endemic Boswellia on Socotra island-rarer species likely face similar pressures, but at a greater threat level due to their small population sizes and geographically restricted ranges.
Boswellia ogadensis Vollesen is one such unique frankincense species restricted only to gypsaceous hillsides in the Shabelle river valley in southeastern Ethiopia, where it grows with species of Commiphora, Vachellia, and Senegalia at elevations of 280-350 m above sea level [1]. It is known from only three locations, close together, on the road between Gode and Kelafo; the first specimen (used to describe the species) was taken in 1972 [15], while an expedition in 2006 found two additional sites where the tree was growing [1]. The species was assessed as "Critically Endangered B1ab(iii)" in 2018 [16], but subsequent collecting trips were unable to locate it in 2019 and 2021, raising further concern about its current status (personal communication with Boris Vrskovy and Sebsebe Demmisew). It is known to occur in mixed-use forests, where it could face a combination of pressures such as grazing and cutting for firewood or construction material. Additionally, resins from Boswellia rivae and other species are collected around the same area, and B. ogadensis resin may be collected and mixed in with the more common B. rivae, as has happened to other rare and endemic Boswellia species in Somaliland (B. occulta mixed in with B. sacra and B. frereana) [17] and India (B. ovalifoliolata substituted for Commiphora wightii) [18].
As a result of these concerns, a conservation survey was carried out in June 2022 to attempt to re-locate the species, confirm it is extant in the Shabelle River Valley, and assess the level of threat the species is facing. The study also aimed to collect a sample of resin, if possible, to identify potential chemical markers in B. ogadensis that could help determine if the species is being harvested and mixed in with other resins from the same region.

Population Status of Boswellia ogadensis
Boswellia ogadensis was re-located at two out of the three previously known locations, and multiple additional populations along the Gode-Kelafo road were identified ( Figure 1). A further population was found on the south bank of the Shabelle river valley, between the towns of Adadle and Gerrei (Jeerey), near the mountains known locally as Hul-Kujir. Although limited access prevented further exploration on the south bank, it is likely that additional populations exist in this area. The species was found only on gypsum hillsides, typically growing with Commiphora guidottii and other species of Commiphora, Vachellia, and Senegalia, but was abundant in all locations, with estimated densities of 50 trees or more per hectare commonplace. Although quantitative surveys will be required to definitively determine the total population size, these observations suggest it may be 10,000 mature individuals or more.
Although limited access prevented further exploration on the south bank, it is likely that additional populations exist in this area. The species was found only on gypsum hillsides, typically growing with Commiphora guidottii and other species of Commiphora, Vachellia, and Senegalia, but was abundant in all locations, with estimated densities of 50 trees or more per hectare commonplace. Although quantitative surveys will be required to definitively determine the total population size, these observations suggest it may be 10,000 mature individuals or more. While the south bank population appeared to experience minimal anthropogenic disturbance, several pressures on the north bank populations were observed. Grazing by goats and sheep is evident in all populations, and villagers confirmed that animals graze in the hills. Very few young trees (<5 cm basal diameter) were observed in all but the south bank and western-most populations. Many trees in grazed populations also showed damage to the trunks of the trees, likely by grazing animals consuming the bark (see Figure 2). The damage to the trunks of the trees could be attempted resin harvesting rather than grazing damage, but active tapping is uncommon in this region, and villagers denied tapping the trees.
Multiple trees in one population had been cut down by humans, and in other populations cutting of other tree species was evident, suggesting this is relatively common practice but does not target B. ogadensis specifically. Evidence of attacks by boring insects, most likely cerambycid or buprestid beetles, was present in many populations as well. This seemed to primarily affect branches, with adult tree mortality due to insects rare. Mistletoes were also seen parasitizing trees in multiple populations. Erosion is likely a cause of natural mortality. While the south bank population appeared to experience minimal anthropogenic disturbance, several pressures on the north bank populations were observed. Grazing by goats and sheep is evident in all populations, and villagers confirmed that animals graze in the hills. Very few young trees (<5 cm basal diameter) were observed in all but the south bank and western-most populations. Many trees in grazed populations also showed damage to the trunks of the trees, likely by grazing animals consuming the bark (see Figure 2). The damage to the trunks of the trees could be attempted resin harvesting rather than grazing damage, but active tapping is uncommon in this region, and villagers denied tapping the trees. Without exception, villagers claimed little knowledge of B. ogadensis. Although many knew the tree, they refer to it as mirafur, the same name used for B. rivae. Only one villager identified it by a modified name, mirafur silon ("similar to B. rivae but different"). Many also said that they rarely visit the steep hillsides where B. ogadensis grows, although animals graze there. Multiple trees in one population had been cut down by humans, and in other populations cutting of other tree species was evident, suggesting this is relatively common practice but does not target B. ogadensis specifically. Evidence of attacks by boring insects, most likely cerambycid or buprestid beetles, was present in many populations as well. This seemed to primarily affect branches, with adult tree mortality due to insects rare. Mistletoes were also seen parasitizing trees in multiple populations. Erosion is likely a cause of natural mortality.

Chemical Composition of Resin Samples
Without exception, villagers claimed little knowledge of B. ogadensis. Although many knew the tree, they refer to it as mirafur, the same name used for B. rivae. Only one villager identified it by a modified name, mirafur silon ("similar to B. rivae but different"). Many also said that they rarely visit the steep hillsides where B. ogadensis grows, although animals graze there.

Discussion
In this study, we aimed to re-locate and survey the previously recorded populations of B. ogadensis in the Shabelle river valley in southern Ethiopia, in order to assess the level of threat currently facing this unique species. Although the total range of the species is highly restricted, it was found to be locally abundant on the gypsaceous hills between Gode and Kelafo, and was found on the south bank of the river valley in addition to the north bank. The record from the south bank brings the total known Extent of Occurrence (EOO) to 405 km 2 , and the discovery of additional populations on the south bank-which seems highly likely-may expand this EOO potentially up to 1000 km 2 or more. Furthermore, the south bank population appeared to be relatively undisturbed, with good regeneration and no obvious signs of anthropogenic disturbance. Given the expanded EOO, with an Area of Occurrence (AOO) well over 10 km 2 , observed differences in the south versus north bank populations, and larger than expected total population size, B. ogadensis may not be as threatened as previously assumed. Down listing from "Critically Endangered" (CR) to "Endangered" (EN) under the IUCN Red List Criteria [24] is likely warranted, but further quantitative research is needed, particularly on the south bank populations.
Like many other species of Boswellia, B. ogadensis is facing several threats. The most prominent is the grazing pressure, primarily by goats, which suppresses new seedlings and has been shown to completely block regeneration in other Boswellia species [8,11]. In this case, the grazing pressure varies geographically, with some sites containing almost no young trees and others containing evidence of robust regeneration. Unsurprisingly, populations near human settlements showed more intense grazing pressure while more remote sites were less grazed. The grazing can also cause damage to the trees' bark, creating opportunities for boring insects to attack the tree. Boring beetles belonging to the Cerambycidae and Buprestidae families have been found to attack other Boswellia species, sometimes fatally [9,25,26]. However, few dead trees were observed, with most of the obvious insect damage occurring in branches or still-surviving trunks.
Cutting of trees, likely for firewood, is a threat. Cut B. ogadensis were observed in one population, and cutting of various other tree species was observed in other populations. Interestingly, local people did not identify the tree as having any distinct uses, and largely regarded it as equivalent to the far more common B. rivae. Although the cutting does not seem to target B. ogadensis specifically, the threat posed by general habitat degradation, particularly around villages, is still present.
Boswellia rivae, B. neglecta, and B. microphylla all occur in the same geographic area as B. ogadensis [1]. Boswellia rivae essential oil is most often dominated by α-pinene, with limonene, δ-3-carene, p-cymene, and β-pinene often present as major components [36][37][38]. Boswellia neglecta essential oil is very similar to that of B. ogadensis, with high levels of α-thujene, α-pinene, p-cymene, and terpinen-4-ol [36][37][38]. The essential oil of B. microphylla has not been characterized. While similar to B. neglecta essential oil, the essential oils of B. ogadensis can be distinguished by the lack of linalyl acetate and the presence of 3,5-dimethoxytoluene. Boswellia rivae resin is also collected in the same area where B. ogadensis grows; however, B. ogadensis essential oil can be identified by the presence of 3,5-dimethoxytoluene, and the presence of (Z)-salvene and/or (E)-salvene.
The presence of positive markers (3,5-dimethoxytoluene and (Z)-salvene) that are present in all analyzed samples of B. ogadensis, but not present in other species commercially harvested in the same area, indicate the potential for these to be used as monitoring tools, to determine if commercial batches of B. rivae essential oil include the far rarer B. ogadensis.

Field Surveying
Field surveys of possible B. ogadensis locations were conducted from the 5th-13th of June 2022. All previously known locations were visited and re-surveyed, and additional potential sites in the Shabelle river valley area that could host B. ogadensis were also visited. Where B. ogadensis was found, we estimated the number present, number of young trees (<5 cm basal diameter) present, health of trees, phenology, and any current or potential threats observed (cutting, resin harvesting, grazing, etc.). We also interviewed local communities in the area to determine whether they knew about the presence of B. ogadensis, if they had any specific name for it, and how they were using B. ogadensis and the broader ecosystem in which it is growing.

Collection of Resins
Resin was collected opportunistically from natural exudations from B. ogadensis trees; no trees were tapped or otherwise harmed to collect the resin. Only small amounts were exuding from individual trees, so we pooled the resin from multiple individual trees in the same location for each sample (see Table 3). The exuded resins collected were of varying ages, but we focused on collecting more recently exuded resins and excluded old, dry resins that had been on the tree for a long time. Resins collected were sealed in plastic bags and shipped to the Aromatic Plant Research Center for analysis. A voucher specimen of B. ogadensis was collected and deposited in the Jigjiga Herbarium at the Somali Region Pastoral and Agro-Pastoral Research Institute (specimen no. 7204). The B. ogadensis trees were identified by S.J. and A.A., following the description from [1], based on the characteristics: bark smooth and sometimes flaking; imparipinnate leaves, 5-9 foliate, 4-15 cm long, sparsely puberulous, with leaflets ovate-elliptic to subcircular; 3-locular, narrowly pyriform, glabrous fruits, 12-16 × 4.5-6 mm, and pyrenes 4-5 × 1.5-2 mm, narrowly trullate with a long-acuminate tip and short basal horn, and trigonous without wings.

Hydrodistillation of Resins
Hydrodistillations of the resin samples of B. ogadensis were carried out using Likens-Nickerson apparatus for 6 h to give yellow essential oils (see Table 3).

Gas Chromatographic-Mass Spectrometry
The B. ogadensis resins were analyzed by GC-MS with a Shimadzu GCMS-QP2010 Ultra (Shimadzu Scientific Instruments, Columbia, MD, USA) with ZB-5ms capillary column (Phenomenex, Torrance, CA, USA) as previously described [27]. Identification of the chemical components was carried out by comparison of the retention indices determined with respect to a homologous series of normal alkanes and our comparison of their mass spectra with those reported in the literature [20][21][22] and the Aromatic Plant Research Center's inhouse library [23]. A representative chromatogram is shown in Supplementary Figure S1

Chiral Gas Chromatographic-Mass Spectrometry
The B. ogadensis essential oils were analyzed by chiral GC-MS as previously reported [41]: Shimadzu GCMS-QP2010S instrument (Shimadzu Scientific Instruments, Columbia, MD, USA), Restek B-Dex 325 capillary column (30 m × 0.25 mm × 0.25 µm film) (Restek Corporation, Bellefonte, PA, USA). Enantiomers of monoterpenoids identified by comparison of retention times with authentic samples (Sigma-Aldrich, St. Louis, MO, USA) and percentages determined based on peak areas. A representative chiral gas chromatogram is shown in Supplementary Figure S2.

Conclusions
While restricted to a small range in the Shabelle river valley in southern Ethiopia, Boswellia ogadensis is more abundant than previously thought. The species is facing multiple threats, but some populations are regenerating well, and it does not appear to be specifically targeted by local people for firewood or resin harvesting. Given the expanded EOO and AOO, newly identified populations, and larger than expected total population size, B. ogadensis likely qualifies for down listing from Critically Endangered B1ab(iii) to Endangered B1 and B2ab(iii). The essential oil of B. ogadensis oleo-gum-resin is similar to that of other Boswellia species, but it can be distinguished from other sympatric Boswellia by the presence of 3,5-dimethoxytoluene and (Z)-salvene, which indicates the potential for use as an ex-situ monitoring tool. Despite the importance of this area as part of the Horn of Africa Biodiversity Hotspot, it is still under-studied and under-collected, with further research on the biodiversity, socio-economic importance, and drivers of land management needed. Further research is needed on B. ogadensis to identify additional populations and quantify current population structure and trends.
Author Contributions: S.J. and A.A. designed the study, coordinated field surveying and sample collection, and analyzed the field data; W.N.S. carried out the hydrodistillations; P.S., A.P. and W.N.S. analyzed the GC-MS, GC-FID, and Chiral GC-MS data. All authors contributed to writing and editing the manuscript. All authors have read and agreed to the published version of the manuscript.