Evaluating the Potential of Boswellia rivae to Provide Sustainable Livelihood Benefits in Eastern Ethiopia

Frankincense is an oleo-gum-resin collected from wild Boswellia spp. trees, and widely used in perfumery, cosmetics, aromatherapy, incense, and other industries. Boswellia rivae, growing in Ethiopia, Somalia, and Kenya, is one source of frankincense, but is little-commercialized compared to species such as B. sacra, B. frereana, and B. papyrifera. In this study, we examine the resin essential oil chemistry and harvesting systems of B. rivae in order to evaluate its potential for increased trade and potential positive livelihood benefits. Boswellia rivae produces an essential oil rich in α-thujene (0.1–12.4%), α-pinene (5.5–56.4%), β-pinene (0.3–13.0%), δ-3-carene (0.1–31.5%), p-cymene (1.4–31.2%), limonene (1.8–37.3%), β-phellandrene (tr-5.6%), trans-pinocarveol (0.1–5.0%), trans-verbenol (0.1–11.2%), and trans-β-elemene (0–5.7%), similar to major commercial species, although it is difficult to detect mixing of B. rivae and Commiphora africana resins from chemistry alone. The B. rivae trees are not actively tapped, so resin collection has a neutral impact on the health of the trees, and resin production is unaffected by drought. Consequently, collecting resins acts as a key income supplementing livestock herding, as well as a safety net protecting pastoral communities from the severe negative effects of climate change-exacerbated drought on livestock. Therefore, Boswellia rivae is well positioned chemically, ecologically, and socially to support expanded trade.


Harvester Perceptions and Collection System
The harvesters interviewed unanimously stated that the resins are collected only when they naturally exude from the tree, with no active tapping taking place. This was confirmed using field observations, where no active tapping was apparent even near harvesting villages. Interestingly, the harvesters reported that collection of resin resources is often segmented by gender, with women collecting B. rivae resins and men collecting resins of Commiphora myrrha and C. guidottii, reportedly as the prominent spines of the latter species make collection difficult for women wearing the standard long, flowing dresses. However, during periods of higher demand, these gender divides do not apply as strictly and both genders collect all types of resin. Additionally, children may accompany their adult family members on collecting trips, although this was not said to be common. In addition to being collected for trade, resins were reportedly burned locally to purify spaces and heal sick children, chewed, and applied topically to heal wounds.
All harvesters interviewed agreed that only a small amount of the resin produced is currently collected and traded, with the opportunity for significantly expanded collection if there is a market. No harvesters perceived issues with sustainability or declines in tree health or abundance. Although drought has been impacting the region for several years, informants did not perceive a negative impact on the health of the B. rivae trees; to the contrary, they perceived the drought to increase resin production. Resin collection was also reported to be one of the only good sources of income during the drought, due to the negative effects of drought on livestock, and collection seems to be expanded during periods of drought. However, the response to drought was not uniform, as harvesters from one village reported a decrease in resin collection during drought as a result of most people spending increased time trying to save their livestock.
Resins are collected opportunistically, predominantly by women, with harvesters sometimes making dedicated collecting trips and sometimes collecting resins as they herd livestock, and then brought back to the harvesters' village for storage. Harvesters from a given village reported pooling their resins for sale to local traders, in practice forming an informal sales cooperative. These traders in turn sell the resins to national exporters who export the resins to processors, largely based in Europe (Figure 2). Harvesters reported that the sales price of the resin fluctuates seasonally and annually, due to market changes, exchange rate changes, and the quality (size, color, percentage foreign bodies in the resin) of the resin but is normally USD 1-1.25 per kg of resin. As a focused harvester can collect

Harvester Perceptions and Collection System
The harvesters interviewed unanimously stated that the resins are collected only when they naturally exude from the tree, with no active tapping taking place. This was confirmed using field observations, where no active tapping was apparent even near harvesting villages. Interestingly, the harvesters reported that collection of resin resources is often segmented by gender, with women collecting B. rivae resins and men collecting resins of Commiphora myrrha and C. guidottii, reportedly as the prominent spines of the latter species make collection difficult for women wearing the standard long, flowing dresses. However, during periods of higher demand, these gender divides do not apply as strictly and both genders collect all types of resin. Additionally, children may accompany their adult family members on collecting trips, although this was not said to be common. In addition to being collected for trade, resins were reportedly burned locally to purify spaces and heal sick children, chewed, and applied topically to heal wounds.
All harvesters interviewed agreed that only a small amount of the resin produced is currently collected and traded, with the opportunity for significantly expanded collection if there is a market. No harvesters perceived issues with sustainability or declines in tree health or abundance. Although drought has been impacting the region for several years, informants did not perceive a negative impact on the health of the B. rivae trees; to the contrary, they perceived the drought to increase resin production. Resin collection was also reported to be one of the only good sources of income during the drought, due to the negative effects of drought on livestock, and collection seems to be expanded during periods of drought. However, the response to drought was not uniform, as harvesters from one village reported a decrease in resin collection during drought as a result of most people spending increased time trying to save their livestock.
Resins are collected opportunistically, predominantly by women, with harvesters sometimes making dedicated collecting trips and sometimes collecting resins as they herd livestock, and then brought back to the harvesters' village for storage. Harvesters from a given village reported pooling their resins for sale to local traders, in practice forming an informal sales cooperative. These traders in turn sell the resins to national exporters who export the resins to processors, largely based in Europe (Figure 2). Harvesters reported that the sales price of the resin fluctuates seasonally and annually, due to market changes, exchange rate changes, and the quality (size, color, percentage foreign bodies in the resin) of the resin but is normally USD 1-1.25 per kg of resin. As a focused harvester can collect several hundred kilograms of resin per year, this represents a significant source of income for harvesting communities.
The B. rivae samples collected in this study have a similar composition to B. sacra, the most popular commercial essential oil. There are some exceptions, with one commercial sample showing an unusually high level of p-cymene and one tree sample showing a high level of δ-3-carene with very little α-pinene; the B. rivae samples are also typically higher in trans-verbenol than most B. sacra [27][28][29][30]. However, this is not necessarily a barrier to further commercialization, as individual deviant samples can be blended into batches of resin with the more common and preferred α-pinene/limonene-dominant profiles. Additionally, B. rivae can be distinguished from B. sacra, as it lacks incensole or incensyl acetate; it can also be distinguished from the much rarer conspecific B. ogadensis by the absence of 3,5-dimethoxytoluene and (Z)-and/or (E)-salvene [34]. These chemical markers would allow the authentication of B. rivae as a non-tapped, sustainably collected source of frankincense.
Mixing of resins from multiple species was acknowledged as common practice by many of the harvesters interviewed. Most commonly, this involved collection of both Boswellia rivae and Commiphora africana, as the two resins are highly similar in both appearance (see Figure 1) and scent. Indeed, the essential oil profiles of the C. africana samples were essentially indistinguishable from those of B. rivae. A number of minor components that may act as markers of C. africana were identified, although they would likely only be apparent if the percentage of C. africana in the resin were relatively high; this will need to be confirmed with further study. This difficulty in establishing clear species identification may be a barrier to expanded commercialization, as industries typically prefer clear, single-species product identification as a means of enhancing quality control and meeting regulatory obligations. However, from a local perspective, the logic of mixed species collection makes sense as long as traders do not raise an issue with it, as this allows expanded product collection with less effort. The chemical similarity between multiple species makes imposing postcollection, species-based quality checks difficult, and without this kind of control, there is little incentive for harvesters to adhere to single species collection.

Collection System and Role in Local Livelihoods
Boswellia rivae is unusual amongst frankincense species, and resin-bearing species in general, for being passively harvested. Most species are actively tapped to obtain the resin, necessarily causing damage to the harvested tree and presenting opportunities for pests and pathogens to attack the tree [13,35]. Active tapping also affects the carbon resources of the tree, reducing reproduction and growth [36,37]. Harvesting pressure driven by the expanding market for frankincense has affected other species of Boswellia, particularly B. sacra in Somaliland, where higher demand for resin combined with limited market incentives for sustainability have resulted in a wave of unsustainable harvesting practices [18]. Boswellia rivae, on the other hand, is relatively insulated from the effects of a potential market wave, as resin collection has a neutral impact on the trees. This positions the species well to support expanded market demand without negative ecological impacts. Boswellia neglecta is the only other non-tapped frankincense species, but its essential oil contains high levels of undesirable components such as terpinen-4-ol, limiting its market potential [38]. By contrast, B. rivae has relatively low levels of terpinen-4-ol, but often significant quantities of commonly desirable components such as α-pinene and limonene.
Although some harvesters reported collecting less resin during times of drought as a result of spending more time attempting to keep livestock alive, the trees' production of the resins was reportedly unaffected by drought conditions, representing a potential safety net of income even during livestock die-offs [24,39,40]. Few other alternative sources of income exist in these areas, so the collection of resins is a critical source both of supplementary income and stability, especially as climate change is likely to drive increasing frequency and severity of drought [41,42]. The collection of resins is generally considered culturally inferior to income from herding livestock, but it is a key source of cash resilient to the impacts of drought [18]. The involvement of women as direct collectors of the resins is also unusual in frankincense, where most species' collection systems follow a rigid traditional system of men harvesting and women sorting/cleaning the resins. This is particularly important given Ethiopia's low human development classification and pervasive gender income inequality [43,44]. The collection and sale of NTFPs by women has been effective at reducing gender-based income inequality elsewhere in Ethiopia, highlighting the opportunity for expanded commercialization in B. rivae to follow a similar pattern [45].

Study Species
Boswellia rivae Engl. is one of 24 species of Boswellia, and one of six known from Ethiopia. It is distributed widely in eastern Ethiopia, Somalia, and northeastern Kenya, where it prefers Acacia-Commiphora woodland at elevations of 150-915 m above sea level. It was most recently assessed as least concern under the IUCN Red List Categories and Criteria in 2018 and has not been reported to be under significant threat. The species grows both on flat areas and on rocky slopes, in a variety of substrates including limestone, gypsum, and sandy soil. It grows as a small tree or spreading shrub, with dark yellow to pale grey exfoliating bark, imparipinnate leaves 4-18 cm long, with slightly serrated leaflets. Flowers are pink, occurring in pubescent racemes or panicles up to 6 cm long. The fruits are angular, pyriform, and pubescent, 3(-4)-locular, with pyrenes up to 10 × 7 mm [13].
The species is well known locally where it occurs, with both the tree and resin called Mirafur; sometimes other names are applied as well, such as beeyo or jawder, but these more properly refer to B. sacra and B. neglecta, respectively, and are rarely used to describe B. rivae. The resin is collected and used locally for incense, chewing, and as insect repellent. The wood is also reportedly used in construction and fencing, although the generally twisted branching architecture and softness of the wood do not make it ideal for this purpose [13].

Study System
The Somali Region of Ethiopia, also known as the Ogaden, is a large (~327,000 km 2 ) regional state consisting of an uplifted plateau sloping from about 1500 m in the northwest to 300 m in the south. The soils consist primarily of sedimentary limestone formed during the cretaceous and lower cretaceous periods. They are typically rich in potassium, phosphorus, and carbonates, but are low in nitrogen, and are fertile only with irrigation [46]. Annual precipitation varies from 200-400 mm, with a dual rainy season system that produces most of the precipitation during April-May and September-October [47]. Acacia-Commiphora woodland is dominant through most of the Ogaden region, grading into desert and semidesert scrubland at lower elevations in the south of the region [48]. Species of Vachellia spp., Senegalia spp., Commiphora spp., Boswellia spp., Balanites aegyptiaca, and Maytenus senegalensis are common [49].

Interviews with Harvesting Communities
Field surveys of B. rivae production sites in southeastern Ethiopia took place in June 2022, focusing on known harvesting villages. A total of 28 participants (8 women, 20 men) were interviewed using participatory semi-structured and narrative interviews, focusing on local peoples' perceptions of resource abundance, livelihood benefits, trade structure, local uses, and impacts of disturbances like drought. A semi-structured and narrative format was used to allow participants to raise key issues themselves and indicate the issues they found most important and impactful. Key pieces of information were authenticated using triangulation with at least three independent informants.

Collection of Resins
Five samples of resins said to be B. rivae were collected from harvesting villages and commercial stores, which served as examples of B. rivae resin currently in trade. As a control, three samples were also collected directly from B. rivae trees and two samples directly from Commiphora africana trees (locally called Geed Harag, and sometimes collected along with the B. rivae); these control samples allowed us to confirm the botanical identity of the trees, examine pure samples of these species' essential oils, and compare these pure samples to the village/commercial samples ( Figure 3). This allowed us to examine if components occurring in pure C. africana, but not pure B. rivae, also appeared in the commercial samples and thus indicated mixing of B. rivae and C. africana in the commercially-traded samples. Given the small amount of resin produced per tree (often only a few grams), we pooled resins from individual trees in a given area, such that each sample represents resins taken from multiple trees in a given sample location. The resins collected were naturally exuded, and thus of varying ages, but we focused on newly exuded resins and did not collect clearly old and excessively dry resins. of Vachellia spp., Senegalia spp., Commiphora spp., Boswellia spp., Balanites aegyptiaca, and Maytenus senegalensis are common [49].

Interviews with Harvesting Communities
Field surveys of B. rivae production sites in southeastern Ethiopia took place in June 2022, focusing on known harvesting villages. A total of 28 participants (8 women, 20 men) were interviewed using participatory semi-structured and narrative interviews, focusing on local peoples' perceptions of resource abundance, livelihood benefits, trade structure, local uses, and impacts of disturbances like drought. A semi-structured and narrative format was used to allow participants to raise key issues themselves and indicate the issues they found most important and impactful. Key pieces of information were authenticated using triangulation with at least three independent informants.

Collection of Resins
Five samples of resins said to be B. rivae were collected from harvesting villages and commercial stores, which served as examples of B. rivae resin currently in trade. As a control, three samples were also collected directly from B. rivae trees and two samples directly from Commiphora africana trees (locally called Geed Harag, and sometimes collected along with the B. rivae); these control samples allowed us to confirm the botanical identity of the trees, examine pure samples of these species' essential oils, and compare these pure samples to the village/commercial samples ( Figure 3). This allowed us to examine if components occurring in pure C. africana, but not pure B. rivae, also appeared in the commercial samples and thus indicated mixing of B. rivae and C. africana in the commercially-traded samples. Given the small amount of resin produced per tree (often only a few grams), we pooled resins from individual trees in a given area, such that each sample represents resins taken from multiple trees in a given sample location. The resins collected were naturally exuded, and thus of varying ages, but we focused on newly exuded resins and did not collect clearly old and excessively dry resins. Resins were sealed in plastic bags and shipped to the Aromatic Plant Research Center for analysis. Voucher specimens of both B. rivae and C. africana were deposited at the Jigjiga Herbarium at the Somali Region Pastoral and Agro-Pastoral Research Institute (B. rivae, specimen no. 7205; C. africana, specimen no. 7206) and identified by A.A.

Hydrodistillation of Resins
Hydrodistillation of the resin samples was carried out using a Likens-Nickerson apparatus [50] with continuous extraction with dichloromethane for 6 h each to provide yellow to pale yellow essential oils (Table 3). For each hydrodistillation, the resin was placed in a 500 mL flask with 200 mL of distilled water, the Likens-Nickerson apparatus suitable for solvents heavier than water was used [51], 25 mL of dichloromethane was placed in a 50 mL round-bottom flask for the continuous extraction of the hydrodistillate. The condenser was maintained at 10-15 • C with a recirculating refrigerated water bath. The hydrodistillation was carried out under normal atmosphere.

Gas Chromatography-Mass Spectrometry
The B. rivae and C. africana resins were analyzed using 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 [52] and the Aromatic Plant Research Center's inhouse library [53].

Gas Chromatography-Flame Ionization Detection
The B. rivae and C. africana oleogum resin essential oils were analyzed using GC-FID using a Shimadzu GC 2010 (Shimadzu Scientific Instruments, Columbia, MD, USA) equipped with flame ionization detector, a split/splitless injector, and Shimadzu autosampler AOC-20i (Shimadzu Scientific Instruments, Columbia, MD, USA), with a ZB-5 capillary column (Phenomenex, Torrance, CA, USA), as previously described [27]. Percent compositions were determined using peak integration without standardization.

Conclusions
Boswellia rivae produces an essential oil chemical composition comparable to that of other major commercial species, although it has not yet been commercialized to the same degree. The mixing of non-frankincense resins into the B. rivae collections is a concern, but the essential oil profiles of B. rivae and C. africana are similar, and commercial samples, regardless of mixing, are comparable to major traded frankincense species such as B. sacra. Given the neutral impact of resin collection on the species due to the passive collection system, and the least concern status of the species, trade in B. rivae could likely be expanded significantly without negative ecological consequences for the species. This would also have strong positive social benefits, as resin collection is a critical supplementary livelihood and bulwark against the increasing effects of climate change-driven drought. Therefore, expanding resin collection and trade would likely function as a critical support to pastoral communities under environmental pressure.
Author Contributions: A.D. and S.J. designed the study; S.J. and A.A. collected samples and field data; A.D., S.J. and A.A. analyzed 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 the writing and editing of the manuscript. All authors have read and agreed to the published version of the manuscript.