Today’s forests for tomorrow’s world; Stable Isotope Ratio Analysis of timber to protect two forest concessions in Gabon

Authors: G.O. Rees1, C.J Watkinson1, M. Boner2, S. Hofem2, L. Michely2

1 Agroisolab UK Ltd, Greets House Road, Welburn, YO60 7EP. 2 Agroisolab GmbH, Prof.-Rehm-Str. 6, D-52428 Jülich.

Correspondence: [email protected]

Front cover image credit: Cynel Moundounga, 2019.

exporting nations to provide scientific evidence Executive summary with respect to the origin and species of traded Context timber. Moreover, through the dissemination of Traceability of timber has come under the spotlight science-based techniques and implementation of during recent decades with consumers becoming capacity building in timber exporting nations, increasingly aware of the damage caused by efforts to combat deforestation using scientific deforestation and illegal logging to forest techniques can be implemented at both ends of the ecosystems. Tropical rainforests possess some of supply chain. the greatest biodiversity on earth, being home to many critically endangered species ranging from Relevance to policy Since the beginning of the 21st Century, many insects to mammals. They also act as significant CO2 sinks and their loss is considered a significant governments have implemented statutory contributor to climate change. Besides ecological instruments as means to control the origin and impacts, the economic development of many poor species of imported timber. The US Lacey Act was nations is at stake. Illegally harvested timber leads amended in 2008 to include timber, making it the to a loss of revenue for many developing world’s first ban on trade in illegally sourced wood economies and leads to hardship for many products. In the EU, regulations took the form of indigenous forest communities who depend on the European union Timber Regulation (EUTR) with forest resources for their livelihoods and wellbeing. the UK set to adopt a parallel version (UKTR) in the Misdeclarations, with respect to origin and species year 2021. Several organisations including the of timber also pose a risk to importing nations due International Tropical Timber Organisation (ITTO), to the variety of plant health pathogens and insects Environmental Investigation Agency (EIA), Global that are able to proliferate outside of their natural Witness, US Department of Justice (DOJ), range. Ash Dieback (Hymenoscyphus fraxineus) and Worldwide Fund for Nature (WWF), International the Asian Long Horn beetle (Anoplophora Union for Conservation of Nature (IUCN) and World glabripennisis) are just a few examples where Resource Institute (WRI) work to expose illegal ecological damage has been accelerated due to practices with respect to harvesting of timber and imports of mislabelled timber. Timber export encourage governments and corporations to take markets compound these problems by trading actions against illegal logging. Several certification timber and wood products without ensuring schemes including Forestry Stewardship Council responsible sourcing, or by failing to carry out (FSC) and Programme for the Endorsement of enough due diligence in their supply chains. Forest Certification (PFEC) provide the timber Research suggests that between 20% and 40% of all industry with a means of carrying out due diligence globally traded timber may be mislabelled with checks within supply chains, providing paper-based respect to origin and species declarations, equating evidence linked to forest and supplier audits. to a volume of 650 million cubic meters of timber. However, certification bodies are only able to audit The Worldwide fund for Nature (WWF) estimates a fraction of the world’s forests at a fraction of the that €3bn of illegally harvested timber is placed on time, and participation by the timber industry is European markets each year with most being only voluntary. sourced from the tropical forests in the Far East, Method statement and the Congo Basin of West Africa. During the last Stable isotope analysis is a widely accepted decade, scientific methods have increasingly analytical technique that has become well played a part in role in verifying the origin and established around the world, and it is often used species of traded timber. Recent advancements in to verify the origin of biological samples. To verify scientific techniques including wood anatomy, if the origin claim for a piece of timber (test sample) genetics (DNA), stable isotope ratio analysis (SIRA) is correct, the method of analysis typically involves and ambient mass spectrometry (DART) now comparison of the stable isotopes (C,H,N,S,O) enable regulatory bodies in both importing and

i

measured in the test sample against a timber Project objectives in Gabon: reference dataset of stable isotopes for the same • Collaborate with Royal Botanic Gardens, Kew (RBG, origin and species. Despite the growing number of Kew) and FSC to obtain authentic georeferenced research studies utilising authentication analyses samples of endangered, commonly traded or to verify origin of timber, many have involved species of interest of timber from two forest insufficient numbers of samples and have focused concessions in Gabon. on single tree species. To address these shortfalls, • Collect and perform stable isotope analysis of World Forest ID (WFID) was conceived from the authentic geo-referenced timber such as: Okoumé (Aucoumea klaineana), Azobé (Lophira alata), collaborative efforts of several key organisations Okan (Cylicodiscus gabonensis), Padouk including DEFRA, Royal Botanic Gardens at Kew, (Pterocarpus soyauxii), Sipo (Entandophragma FSC, Agroisolab and the US Forestry Service, with utile), Tiama (Entandophragma angolense), Tali the aim of co-ordinating the collection and analysis (Erythrophleum ivorense), Bilinga (Nauclea of timber reference samples from some of the most diderrichii), Kévazingo (Guibourtia tessmanii), endangered forests around the world. Bubinga (Guibourtia pellegriniana), Ovèngkol (Guibourtia ehie), Beli (Paraberlinia bifoliolate), Aims of the World Forest ID project Niangon (Heritiera densiflor), Douka (Tieghemella World Forest ID aims to: Africana), and Movingui (Distemonanthus benthamianus). • Facilitate the co-ordination and development of • Establish if differences occur in the stable isotope timber tracking technologies. ratios of trees from different forest concessions

within and outside Gabon. • Provide governments with a scientific resource to • Assess what differences occur in the stable isotope carry out enforcement activities which are based ratios of single tree taxa from the same location on evidence established through laboratory within Gabon. analysis of seized timber. • Utilise isotopic methods to verify if the declared

origin of the Gabonese timber is consistent with • Carry out large-scale collections of timber data from Gabon. reference samples from around the world, including ‘at risk’ categories of tree species.

• In collaboration with the Royal Botanic Gardens, Kew, provide the wider scientific community with a go-to resource of authentic georeferenced timber samples covering over 100 commonly traded taxa.

• Disseminate research findings, including all data outputs, to the wider scientific community.

Okoume fretboard on a Ukulele. Image Credit: Charlie Watkinson, 2020.

ii

Looking up towards the canopy of a Pterocarpus soyauxii tree in Gabon. Image credit: Cynel Moundounga, 2019 Results address this much needed comparison as it may enable • Significant differences in the sulfur isotope ratios of more efficient allocation of reference sampling resources. Okoume (Aucoumea klaineana), from PWG and CBG Further sampling of Burseraceae family timbers in the concessions were observed, suggesting that concession tropics may permit a global model for verifying their level classification in Gabon may be achieved with geographic origin. extensive sampling. Future work • Discrimination in the sulfur isotope ratios was evident between samples of Dacryodes taken from CBG and Expanding the collection reference samples will be PWG concessions. necessary to investigate regional classification further. It is hypothesised that having greater quantities of sample data • Stable isotope ratios measured in samples of Aucoumea will improve the discrimination between concessions or klaineana and Dacryodes buttneri are not directly regions, as well as allowing for comparison of stable comparable despite belonging to the Burseraceae isotopes both within a single sampling site and single taxa. family. A higher frequency of reference sample collection will also • Aucoumea spp. and Dacryodes spp. do not follow enable assessment of the suitability of specific taxa to act as expected patterns with respect to correlation between isotopic proxies for other species. Natural variability of δ18O and δD in cellulose and precipitation. isotopic distribution within a site is still not fully understood, Conclusions and it is anticipated that large scale sample collections for Despite the quantity of samples and species being relatively specific taxa will also enable a better understanding of this. low, the data acquired establishes a basis of evaluation for So far, the study has focused on whole wood and cellulose assessing geographic origin claims of certain forest products components of timber reference samples. Several including plywood and laminates from Gabon. The alternative analytical methods detail procedures for differences in the sulfur isotope ratios of Okoume and analysing alternative metabolic fractions including lignin Dacryodes reference samples from Precious Woods Group and proteins in the form of amino acids. Furthermore, (PWG) and Compagnie des Bois du Gabon (CBG) forest isotope methods are already being used to measure the concessions suggest that regional scale origin classification stable isotopes of position-specific atoms within a selection may be realised with high enough frequency of refence of molecules such as ethanol, providing higher resolution sample collection. This study helped to establish the information on the source water incorporated during minimum number of reference samples (five per species per metabolism. Analysis of alternative fractions and position- site) required in order to account for variance of isotopic specific isotope ratios within a molecule such as glucose distribution at a single sampling site. Low frequency of may yield higher quality results and aid the discrimination sampling per site and limited sample numbers mean that of reference samples taken from concessions within the investigation of the Dunbar line hypothesis was not possible same country. on this occasion. Further, higher resolution sampling of target species including Okoume and Dacryodes will

iii

Abstract The aim of this report was to address the differences in stable isotope ratio profile of bulk, homogenised wood samples collected from living or recently felled trees, between two FSC concessions in Gabon, which are approximately 240km apart, for the purposes of origin classification and protecting valuable forest commodities.

Forty-seven timber samples comprising ten genera of tropical trees were obtained using a Pickering Punch sampling device or chainsaw from two forest concessions in Gabon (PWG and CBG) during July 2019. Samples were subject to 18O/16O, D/H, 13C/12C, 15N/14N and 34S/32S stable isotope analysis using elemental analysis- isotope ratio mass spectrometry (EA-IRMS). Additional data from relevant taxa and geographic origins (altitude, soil type, geographic co-ordinates) were used to assess potential higher-order spatial patterns in the data.

Results show that significant differences are evident in the stable isotope ratios of Aucoumea klaineana between PWG (Precious Woods Group) and CBG forest concessions. Relationships are evident between climatic and geological variables and the stable isotope ratios of the samples suggesting that further degrees of origin classification may be achievable in Gabon. For other species that were sampled, insufficient numbers meant the possibility to determine discriminating factors between the two concessions was limited.

The data presented establishes a basis for evaluating origin claims of forest products and timber from the CBG and PWG concessions and lays a foundation for future development of timber tracking technologies in Gabon. The technique can be used for purposes of due diligence or forensic investigation by law enforcement as part of demand-side regulations such as EUTR, ILPA, or the Lacey Act.

Introduction Gabon is a West African nation located on the Gulf of Guinea and Atlantic Ocean, bordering Cameroon, Equatorial Guinea, and the Republic of Congo. Its land mass is approximately 85% forested with deforestation rates of approximately 0.1% occurring annually [NEPcon, 2017]. The timber industry in Gabon accounts for nearly 20,000 jobs [EIA, 2019] and is the country’s most important export (5% of GDP) after oil [NEPCon, 2017]. Of the 22-23 million hectares of forested area, about 4 million ha is protected and 14 million ha are allocated for forestry [EIA, 2019]. All Gabon’s forests are managed by the Ministry of Forest, Environment and Natural Resource Protection who oversees the monitoring of forest resources, including the allocation of forest concessions. Two types of permits are issued by the ministry: Concession Forestière sous Aménagement Durable (CFAD), and Permis Forestier Associé (PFA). CFAD, allows for logging by corporations in land areas of between 50 000 and 200 000 hectares, whereas PFA - (which is reserved exclusively for Gabonese nationals) has a maximum size of 50 000 hectares [NEPcon, 2017]. As of 2017, China, followed by France, Belgium, and Italy were the largest export markets for timber sourced from Gabon [ITC, 2018]. The most important species for the Gabonese timber industry is Okoumé (Aucoumea klaineana) which accounts for most of the country's timber exports. The timber is commonly manufactured into plywood and veneers at plants within Gabon and Asia [Timber Trade Portal, 2020]. A four-year investigation by the EIA established that illegally sourced timber from Gabon has routinely entered the United States (US) for over a decade and made its way to thousands of US consumers. An in-depth analysis of the okoume (Aucoumea klaineana) veneer imported directly into the US from Gabon indicates the complicity of the US importers [EIA, 2019]. Following a corruption scandal in 2010, the country moved to ban the export of logs and switched to the export of sawn logs and plywood as a means of stimulating the domestic economy [Kersenty, 2019].Other economically important species for the Gabonese timber industry include Azobé, Bongossi (Lophira alata), Okan (Cylicodiscus gabunensis), Padouk d'Afrique (Pterocarpus soyauxii), Beli (Julbernardia pellegriniana), Tali (Erythrophleum ivorense), Missanda

Page 1 of 55

(Erythrophleum suaveolens). These timbers, and forest products derivatives, are of great importance for protection. Relevance of stable isotope analysis to other timber tracking technologies Population genetics is an excellent method for timber tracking [Jolivet & Degen 2012], however, its application as a tool to verify the origin of timber-based products is dependent on the availability and accessibility of genetic material within a sample. For example, okoume is most commonly converted into plywood and is often used for boatbuilding due to its excellent properties [Negro et al., 2011]. Plywood is produced from the layering of multiple sheets (veneers) of rotary-peeled logs which are held together with thermosetting formaldehyde- urea-based resins [Desch & Dinwoodie 1996; Negro et al., 2011]. Typically, heartwood is used for veneer making. Heartwood can be defined as “the inner layers of the wood, which, in the growing tree, have ceased to contain living cells, and in which the reserve materials (e.g., starch) have been removed or converted into heartwood substance” [Hillis 1987]. The process of heartwood formation does not always lend itself to the preservation of genetic material either in plant cell nuclei, mitochondria or chloroplasts. The process of veneer- making involves boiling the wood, applying steam and hot-pressing at over 100°C to cure the thermosetting resin. This process also does not lend itself to the preservation of genetic material for later analysis by population geneticists. As such, there needs to be a method capable of evaluating the origin of the veneers in plywood so supply-chain stakeholders, enforcement and concerned parties can be assured of its legal origin as part of demand-side authentication. Nevertheless, population genetics is a vital technique and is suitable for addressing the origin of timber within Gabon where genetic material can be accessed for analysis. Ecological importance of Gabonese trees Aucoumea klaineana pierre (okoume) is a long-lived pioneer species capable of converting savannah into rainforest [White et al., 1996; Born et al., 2006] and is the only species of its genus. It grows relatively quickly and makes up a large proportion of the trees within Gabon. The natural range of okoume includes Gabon but also extends into Equatorial Guinea, southern Cameroon and parts of the Republic of Congo [Born et al., 2011]. Despite its relatively large coverage of Gabon, okoume can be difficult to grow in plantations; one study into the silviculture of young okoume trees found that saplings grew best in the soil of their original population. The growth of the saplings could be described in terms of a function of the distance from the sites where their seeds were harvested to the research site [Koumba Zaou et al., 1998]. This means that if okoume is harvested to the point of exhaustion in one region of Gabon, it will not be a straightforward task to simply replant the trees from another region. Genotypic mechanisms were proposed as a potential explanation for this phenomenon. However, phenotypes are ultimately the product of genetic expression, of which, the available genetic sequences are only part of the story; the mechanism phenotypic variation can also be considered in terms of epigenetics or other genetic regulatory mechanisms. Even so, there are clear examples of distinct genetic populations of okoume within Gabon; Muloko-Ntoutoume [et al., 2000] revealed population differences in chloroplast DNA (cDNA), and differences in polymorphic microsatellites have also been identified [Born et al., 2006, Born et al., 2008].

Humans are not the only great apes that rely on okoume for survival; western lowland Gorillas have been observed eating the flowers of okoume. Though it is not a main food source for Gorillas, the availability of okoume during times of famine caused by shortages of other food sources may ensure that Gorillas survive [Williamson et al., 1990]. Nevertheless, in research where Gorillas have been observed, effect of logging did not appear to have affected the local population at the time meaning that logging the trees can still be acceptable if done so in a sustainable manner. Although Gabon constitutes a fraction of the Congo Basin, the country also shelters approximately 45,000 forest elephants, representing nearly 60 percent of Africa’s remaining population [EIA, 2019]. Natural forest is vital for the survival of forest elephants.

Page 2 of 55

Project aims The ambitions of this project were to define the ranges of stable isotope ratios from multiple species of trees from two forest concessions in Gabon by analysing timber samples extracted from living trees. Analysis of the analytical data should enable:

1. Assess what differences occur in the stable isotope ratios of trees from different forest concessions within Gabon. 2. Assess what differences occur in the stable isotope ratios of single tree taxa from the same location within Gabon. 3. Utilise isotopic methods to verify if the declared origin of the Gabonese timber is consistent with data from Gabon. 4. Collect and perform stable isotope analysis of authentic geo-referenced timber such as: Okoumé (Aucoumea klaineana), Azobé (Lophira alata) Okan (Cylicodiscus gabonensis) Padouk (Pterocarpus soyauxii) Sipo (Entandophragma utile), Tiama (Entandophragma angolense), Tali (Erythrophleum ivorense), Bilinga (Nauclea diderrichii), Kévazingo (Guibourtia tessmanii), Bubinga (Guibourtia pellegriniana), Ovèngkol (Guibourtia ehie), Beli (Paraberlinia bifoliolate), Niangon (Heritiera densiflor), Douka (Tieghemella Africana), and Movingui (Distemonanthus benthamianus).

If significant differences in the stable isotope ratios of trees from within different concessions in Gabon are evident, this may be of great benefit to audit teams wanting to meet due diligence requirements and demonstrate sustainable practices. If the isotope ratios of trees from within different forest concessions within Gabon are relatively homogenous, this would mean that it may not be necessary to collect reference samples from all concessions to verify that the timber is from the Gabon.

Stable isotope analysis for provenance verification Stable isotope analysis is widely accepted analytical discipline and has a long history, with many laboratories adopting the technology around the world. Since the beginning of the 21st century, the technique has become well established as a means of verifying the origin of food and drink [Kelly et al., 2002; Boner & Förstel, 2004; Heaton et al., 2008; Pilgrim et al., 2010; Li et al., 2014]. The same principles used to authenticate food were later applied to timber provenance research [Boner et al., 2007; Keppler et al., 2007; Kagawa et al., 2008; Horacek et al., 2009; Kagawa et al., 2010; Gori et al., 2013; Rees, 2015]. Verifying the origin of timber typically depends on comparing an unknown sample against an authentic reference database for a region or territory. The technique is used routinely to assess legality, compliance with labelling legislation, and its use to conduct due diligence is advocated by EUTR [Regulation (EU) No 995/2010].

Methodology Sampling Agroisolab liaised with several organisations including FSC, Royal Botanic Gardens at Kew, US Fish and Wildlife Service, World Resource Institute (WRI) and US Forestry Service to establish a taxa priority list and reach a consensus on which locations within Gabon should be sampled. It was agreed that two forest concessions, Compagnie des Bois du Gabon (CBG) and Precious Woods Group (PWG) would be selected for refence sample collection. The concessions are separated by approximately 240km and presented an opportunity to assess the variability in stable isotope ratios at higher resolution and the viability of stable isotope data to differentiate several taxa at concession level resolution.

Page 3 of 55

15 taxa including a mix of high (green) and medium (yellow) priority species were selected for sampling:

Table 1. List of priority timber species in Gabon.

Trade name Scientific name Okoumé Aucoumea klaineana Azobé Lophira alata Okan Cylicodiscus gabonensis Padouk Pterocarpus soyauxii Sipo Entandophragma utile Tiama Entandophragma angolense Erythrophleum ivorense Erythrophleum Tali suaveolens Bilinga Nauclea diderrichii Kévazingo Guibourtia tessmanii Bubinga (Kévazingo) Guibourtia pellegriniana, demeusii Ovèngkol Guibourtia ehie (Not available at Prescious Woods) Beli Paraberlinia bifoliolata Niangon Heritiera densiflora, utilis (Not available at Prescious Woods) Douka Tieghemella africana Movingui Distemonanthus benthamianus

Samples were taken from forty-seven trees covering two forest concessions (CBG, n= 33 and PWG, n=14) in Gabon during June 2019. Between 2 and 4 pieces of timber (heartwood and sapwood) were collected from each tree. In most cases, three samples were taken per tree as well as leaf and twig samples. Samples of heartwood/sapwood were submitted to Agroisolab for stable isotope ratio analysis whereas the remaining material was distributed between Conservation of Biodiversity at IRET/CENAREST in Gabon and the World Forest ID collection at the Royal Botanic Gardens at Kew. A Pickering Punch (Agroisolab UK, Welburn, UK), a type of hammer-driven bore, was used to collect 45 of the 47 cores of timber 9 to 11 cm in length and 1.5 cm wide. One sample (Berlinia confusa) was smaller than the recommended size due to the hardness of the tree. Samples were then transferred from the Pickering Punch into cardboard tubes which were sealed inside 500 mL evacuated bags and silica gel to aid drying and protect the sample from the humid air in the local environment. 2 of the 47 samples were collected with a chainsaw. Samples were dried in-transit to Conservation of Biodiversity at IRET/CENAREST using 30g silica gel per plug/core. No perforations were made in the acid-free cardboard tubes; field tests found this did not adequately facilitate drying and some cores presented with mould growth upon receipt at Plant Quarantine Unit, Jodrell Laboratory of the Royal Botanic Gardens, Kew. To eliminate the risk of pathogenic fungi entering the UK, samples were subject to 121°C heat at 15psi for 30 minutes before they were released into the WFID collection for analysis and storage.

GPS data, photographs of the trees and leaves, descriptions and comments about the sampled trees were recorded in a mobile phone application named TreeSnap [Staton, Condon & Almsaleed, University of Kentucky 2016]. The data from this collection has since been moved to the World Forest ID app [app.worldforestid.org Staton, Condon & Almsaleed, University of Kentucky 2019].

Data from Watkinson [et al. 2020] was also included in this study in order to evaluate the relationship between the quantity of wood samples collected from an area and the range of stable isotope ratios that can be observed from that collection. This was done with the aim to create a better understanding of limitations in stable isotope data and to define a minimum quantity that should be collected in future projects.

Page 4 of 55

There are clear environmental differences between the two concessions (table 2) that may influence isotope ratios in timber. These are in addition to differences in elevation (figures 1 and 2) and distance to the sea. The CBG concession is situated next to a lagoon and approximately 20km from the Atlantic Ocean at its closest point. PWG concession is approximately 380km from the ocean at its nearest. The two concessions are approximately 240km apart.

Table 2. Environmental differences between the PWG and CBG concessions.

Environmental variable Spatial Compagnie des Precious Woods resolution Bois du Gabon Group (PWG) (CBG) concession concession 18O/16O Annual precipitation 0.25° x 0.25° -3.6‰ -4.4‰ [Bowen & Revenaugh 2003] D/H Annual precipitation 0.25° x 0.25° -16‰ -22‰ [Bowen & Revenaugh 2003] 0.5° x 0.625° 6.5mg/m2 7.8mg/m2 SO4 tropospheric deposition, December (mean 1980-2018) [GMAO 2015] Precipitation, mean February 1983- 0.5° x 0.5° 200-250mm/day 130-150mm/day 2016 (multi-satellite method) [Huffman and Bolvin 2019] Soil types • Dystric fluvisol • Ferralic cambisol [Jones et al., 2013] • Ferralic arenosol • Xanthic ferralsol • Ferralic cambisol • Xanthic ferralsol

Page 5 of 55

Figure 1. Locations and taxa of samples collected in the CBG concession in Gabon. Data sources: Global Administrative Dataset (GADM), ASTER GDEM (ASTER GDEM is a product of METI and NASA), Interactive Forest Atlas Forest Management [WRI 2009], DIVA GIS.

Page 6 of 55

Figure 2. Locations and taxa of samples collected in the Precious Woods Group (P.W.G) concession in Gabon. Data sources: Global Administrative Dataset (GADM), ASTER GDEM (ASTER GDEM is a product of METI and NASA), Interactive Forest Atlas Forest Management [WRI 2009], DIVA GIS.

Page 7 of 55

Measurement Samples were initially dried at 103°C before being coarsely ground by hand and placed into a Retsch MM400 ball-mill. The resulting fine powder was extracted in a soxhlet apparatus over 6 hours with non-polar (Dichloromethane) and polar (methanol) solvents which were then dried in a laboratory-type drying cabinet for at least 1 hour. Finally, the samples were stored in air-tight sample vials and weighed for analysis.

The method isolate non-exchangeable hydrogen in cellulose is outlined by Cheung [2014]. Homogenised samples were nitrated using 4ml HNO3 (90%) and 4ml of glacial H2SO4 (96%) in falcon tubes with 35ml distilled water. As the reaction is exothermic the acid/sample solution was refrigerated for 2 hours during the digest. The sample solution was agitated using an automatic shaker for two hours allowing the cellulose to precipitate. Precipitated cellulose was separated from the solution with an initial centrifugation for 1 minute at 3000rpm. The supernatant was then discarded, and the precipitate was resuspended using 40ml distilled water and subjected to another centrifugation for 1 minute at 3000rpm. This process was repeated until a pH of 6-7 was achieved. Finally, precipitated cellulose was resuspended in 2-3ml of distilled water, transferred to a glass vial and water decanted following centrifugation. Residual water was removed by gentle heating.

To avoid equilibration or humidity effects in the oxygen and hydrogen analysis, samples were equilibrated overnight in desiccators with a humidity of 10.6 %. Afterwards, the samples were vacuum dried for at least 2 hours. Three-point calibration was used to ensure the robustness of the measurements. Samples were compared at the beginning, end, and between each measurement run to a laboratory internal reference standard. Measurements are also reported relative to an internationally defined standard; for hydrogen and oxygen isotope ratio analysis, Vienna Standard Mean Ocean Water (VSMOW) is used. For carbon isotope ratio analysis, Pee Dee Belemnite (PDB) is used. For nitrogen isotope ratio analysis air is used as the standard. For sulfur isotope ratio analysis Canyon Diablo Troilite (CDT) was is used. Measurements are expressed in delta notation in accordance with the following equation:

(ℎ푒푎푣푦 𝑖푠표푡표푝푒/푙𝑖𝑔ℎ푡 𝑖푠표푡표푝푒)푆푎푚푝푙푒 − (ℎ푒푎푣푦 𝑖푠표푡표푝푒/푙𝑖𝑔ℎ푡 𝑖푠표푡표푝푒)푆푡푎푛푑푎푟푑 훿 = ∙ 1000 (ℎ푒푎푣푦 𝑖푠표푡표푝푒/푙𝑖𝑔ℎ푡 𝑖푠표푡표푝푒)푆푡푎푛푑푎푟푑 Measurements are reported in permille (‰) and were made in accordance with processes outlined in Boner et al., 2007.

D/H and 18O/16O measurement: The high-temperature application uses HT-PyrOH with a covalently bonded silicon carbide tube (patented by Agroisolab) filled with glassy carbon chips and coal powder. Working temperature for pyrolysis is typically >1530°C. To gain a higher precision the isotopes are measured in a master / slave configuration with two Isotope Ratio Mass Spectrometers/IRMS (Isoprime, Elementar - Cheadle, UK). Each IRMS only measures one isotope ratio: D/H or 18O/16O. This configuration provides excellent stability because the magnetic field and accelerating voltage remain constant on each IRMS.

13C/12C measurement: Elemental Analysis (EA3000, Eurovector – Milano, Italy) in combination with IRMS (Nu Horizon, NU-Instruments – Wrexham, Wales). Working temperature: 1021°C (oxidation), 600°C (reduction, Cu).

15N/14N measurement: Elemental Analysis (Carlo Erba NA 1500 series 2 elemental analyser (Thermo Fisher Scientific, Dreieich, Germany) in combination with a Nu Horizon IRMS (NU-Instruments – Wrexham, Wales).

34S/32S measurement: Elemental Analyser (EA3000, Eurovector – Milano, Italy) with IRMS (Isoprime - Cheadle, UK). A one tube combustion (oxidation and reduction in one tube) is used to solve issues caused by SO3. Combustion water is directly trapped with magnesium perchlorate at the end of the tube. Working temperature: 1000°C.

Page 8 of 55

Data analysis As many of the species analysed in this project had not been previously analysed by Agroisolab or other laboratories, and the quantities of samples were relatively low, multi-variate analysis methods were not considered ideal to draw conclusions from the data but were still performed to at least make inferences. Univariate analyses were performed to show differences between distributions of means such as Student’s T- test and ANOVA, or to assess co-variance such as regression as well as boxplot visualisations (Orange 3.24, University of Ljubljana, Slovenia; Microsoft Excel 2019 ver.16.0.6742.2048, Redmond WA). Aucoumea klaineana and Dacryodes buettneri were assessed in SAGA GIS 2.3.2 (Departments for Physical Geography, Hamburg and Göttingen, Germany) for spatially related patterns using Inverse Distance Weighting to the 2nd order power and weighting was applied using all sampling points in the area.

Results 18O/16O

Bilinga (Nauclea diderrichii) is remarkably consistent in terms of its δ18O isotopic composition relative to other sampled taxa, and it also has the most depleted 18O/16O isotope ratio. Lophira has the most enriched and widest range of 18O/16O isotope ratios. This seems unusual because figure 3 shows the range of isotope ratios for Aucoumea spp. and Dacryodes spp. which cover two regions within Gabon as opposed to Lophira which was only collected from the CBG concession. Figures 4 and 5 (boxplots of oxygen and hydrogen stable isotopes of different species) show that the range in results is unexpectedly high for the quantity of samples collected. The reason for this range can be attributed to a single sample with a highly depleted 18O/16O isotope ratio. Of the four samples of Lophira alata that were collected, the sample in question was the only one collected with a chainsaw as opposed to a Pickering Punch. The samplers were contacted about the result and were able to provide further corroborating information about the authenticity of the sample. It is not clear at the time of writing what the explanation is to this outlier result, other than outliers exist in dataset and must be included.

There are significant differences between the 18O/16O isotope ratios of the different taxa (p=0.002). For the interpretation of this result, it should be considered that the Aucoumea and Dacryodes were collected in two sites, whereas the other taxa were collected in just one.

Based on the data from samples collected in this study, there is no strong relationship between the oxygen stable isotope ratios and the elevation of each sample. Aucoumea klaineana has the strongest relationship with elevation (R2 = 0.2) (figure 9), although this value lacks explanatory power to be used as a sole predictor of δ18O isotopic composition in Aucoumea klaineana.

D/H

Bilinga (Nauclea diderrichii) has the narrowest range of D/H isotope ratios in comparison to other taxa that were analysed in this study. Contrary to the trend between taxa having the most depleted 18O/16O isotope ratios, Nauclea shows the most enriched D/H isotope ratios (figure 4). The differences in the mean D/H isotope ratios of the taxa are significant (P=0.009) however, there is a lot of overlap in the ranges.

The range of all taxa is approximately 20‰ (between -62 and -42‰). The typical range of D/H ratios is around 10‰, however, Pterocarpus (5 samples collected in in CBG concession, 1 sample collected in PWG concession) shows a range of 18‰ (table 3).

Nauclea diderrichii and Pterocarpus soyauxii appear to have strong, positive relationships with elevation (R2 s of 0.79 and 0.55 respectively). This positive proportionality suggests that their D/H isotope ratios should

Page 9 of 55

increase with increasing elevation if this trend persists. Dacryodes buettneri, Lophira alata, and Aucoumea klaineana do not show strong relationships between their D/H isotope ratios and elevation (figure 10).

Non-exchangeable D/H from cellulose

Relative to the D/H ratios of the extracted wood, analysis of the non-exchangeable D/H isotope ratios show some interesting patterns: Aucoumea klaineana has the most negative D/H isotope ratios in both cases, the remainder of the sampled taxa seemed to have switched places entirely. Table 3 shows that the typical ranges observed in the taxa are approximately 10‰ for D/H from extracted wood, whereas the ranges in the D/H isotope ratios for the non-exchangeable D/H are all greater than 10‰ for all well-sampled taxa.

Nauclea diderrichii, Pterocarpus soyauxii and Lophira alata appear to have positive relationships between their non-exchangeable D/H from cellulose and elevation with R2 s of 0.26, 0.73 and 0.31 respectively. This positive proportionality suggests that their D/H isotope ratios should increase with increasing elevation if this trend persists. Aucoumea klaineana has a negative relationship between its non-exchangeable D/H from cellulose and elevation with an R of -0.43. Dacryodes buettneri was the only taxa without a strong relationship between its D/H isotope ratios and elevation (figure 11).

13C/12C

Carbon isotope ratios of the sampled species have a range that varies between 1.5 and 4.0‰. The carbon isotope ratios of all species vary between -27 and -30.2‰ (table 3). The widest range evident in this study exists in Pterocarpus (4‰ range). It should be known that 5 samples of Pterocarpus were collected in the CBG concession and 1 in the PWG concession.

15N/14N

The range of nitrogen isotope ratios in all sampled taxa were approximately between -2‰ to 8.5‰. The most negative nitrogen isotope ratios belong to the Pterocarpus samples, whereas the most positive nitrogen isotope ratios belong to the Nauclea samples (figure 7). The nitrogen isotope ratios of each taxa had an individual range of around 1 to 4‰. There are significant differences in the mean nitrogen isotope ratios of the taxa sampled (P=0.000) implying a strong taxonomically related effect in the data.

34S/32S

All taxa show significantly enriched sulfur isotope ratios that are in excess of 6‰. The range of sulfur isotope ratios in all taxa range between 6 and 13.5‰ (figure 8) and have individual ranges of approximately 2-3‰ in each concession (table 3). The broadest range per taxa was evident in the Aucoumea samples which were collected in two concessions and also happen to be the most well-sampled timbers in this project. There are significant differences between the taxa (P=0.001) suggesting a strong taxonomically related effect in sulfur stable isotope ratio profile.

All analysed taxa show negative relationships between sulfur isotope ratio and elevation meaning that, with increasing elevation, a decreasing sulfur isotope ratio would be expected if this trend persists. The strongest relationship was evident in Aucoumea klaineana with an R of -0.79 (figure 14).

Page 10 of 55

Table 3. Stable isotope data from taxa sampled from the two concessions in Gabon. Ranges are given when n>1, standard deviations (σ) are given when n>2.

Origin Taxa N 18O/16O D/H D/H 13C/12C 15N/14N 34S/32S samples vs. VSMOW vs. VSMOW non-exchangeable vs. PDB vs. Air vs. CDT (‰) (‰) vs. VSMOW (‰) (‰) (‰) (‰)

Mean σ Range Mean σ Range Mean σ Range Mean σ Range Mean σ Range Mean σ Range CBG Concession, Gabon Aucoumea klaineana 9 23.6 0.3 1.1 -55.2 3.3 10.1 -16.3 4.4 10.5 -27.5 0.7 2.4 2.8 1.6 5.1 10.3 1.4 3.8 Baillonella toxisperma 2 23.6 1.2 -44.7 4.2 -2.2 0.3 -28.3 0.8 4.9 10.1 0.6 Cylicodiscus gabonensis 2 22.9 0.4 -45.4 3.5 -4.7 1.2 -26.7 0.3 5.0 11.3 0.3 Dacryodes buettneri 5 24.0 0.3 0.7 -50.7 6.8 14.4 -8.0 5.8 14.3 -29.2 1.4 3.3 4.5 1.4 3.8 11.5 1.2 3.0

Guibourtia tessmannii 1 25.1 -58.8 -10.8 -28.0 4.5 8.1 Lophira alata 4 23.5 1.2 2.6 -54.6 3.9 8.8 -10.1 7.5 16.3 -28.2 0.7 1.5 5.7 0.4 0.9 9.7 0.9 1.9 Nauclea diderrichii 5 22.9 0.2 0.5 -48.6 2.7 6.5 -10.3 4.3 11.2 -27.7 0.6 1.5 7.0 1.6 3.7 9.7 0.4 1.0 Pterocarpus soyauxii 5 23.3 0.4 0.9 -51.1 7.0 18.0 -16.1 6.5 16.0 -28.2 1.5 4.0 1.5 2.7 6.2 12.1 0.9 2.5 Precious Woods Concession, Aucoumea klaineana 7 23.8 0.3 0.8 -57.6 4.3 10.6 -21.5 4.7 11.9 -27.3 0.8 2.0 1.9 1.3 3.7 7.3 0.9 2.3 Gabon Berlinia confusa 1 25.1 -44.8 -6.0 -28.2 6.2 Cylicodiscus gabonensis 1 23.4 -42.7 3.3 -26.8 7.4 8.7 Dacryodes buettneri 3 24.0 0.3 0.5 -53.0 1.9 3.7 -7.6 5.9 11.0 -30.2 0.1 0.1 5.4 1.0 2.0 10.5 0.7 1.2

Didelotia africana 1 23.2 -55.8 -26.2 -28.9 4.7 8.5 Pterocarpus soyauxii 1 22.9 -43.2 -5.4 -27.0 2.5 10.6

Page 11 of 55

Figure 3. Boxplots of 18O/16O ratios (‰.) Aucoumea (n=16), Dacryodes (n=8), Lophira (n=4), Nauclea (n=5), and Pterocarpus (n=6).

Figure 4. Boxplots of D/H isotope ratios (‰.) Aucoumea (n=16), Dacryodes (n=8), Lophira (n=4), Nauclea (n=5), and Pterocarpus (n=6). Page 12 of 55

Figure 5. Boxplots of D/H nitrated isotope ratios (‰.) Aucoumea (n=16), Dacryodes (n=8), Lophira (n=4), Nauclea (n=5), and Pterocarpus (n=6).

Figure 6. Boxplots of 13C/12C isotope ratios (‰.) Aucoumea (n=16), Dacryodes (n=8), Lophira (n=4), Nauclea (n=5), and Pterocarpus (n=6). Page 13 of 55

Figure 7. Boxplots of 15N/14N isotope ratios (‰.) Aucoumea (n=16), Dacryodes (n=8), Lophira (n=4), Nauclea (n=5), and Pterocarpus (n=6).

Figure 8. Boxplots of 34S/32S isotope ratios (‰.) Aucoumea (n=16), Dacryodes (n=8), Lophira (n=4), Nauclea (n=5), and Pterocarpus (n=6). Page 14 of 55

Figure 9. Relationship between Elevation (metres) and the 18O/16O (d18O) isotope ratios (‰) of genera sampled in the PWG and CBG concessions from Gabon.

Page 15 of 55

Figure 10. Relationship between Elevation (metres) and the D/H (dD) isotope ratios (‰) of genera sampled in the PWG and CBG concessions from Gabon.

Page 16 of 55

Figure 11. Relationship between Elevation (metres) and the non-exchangeable D/H (dD) isotope ratios (‰) from cellulose of genera sampled in the PWG and CBG concessions from Gabon.

Page 17 of 55

Figure 12. Relationship between Elevation (metres) and the 13C/12C (d13C) isotope ratios (‰) of genera sampled in the PWG and CBG concessions from Gabon.

Page 18 of 55

Figure 13. Relationship between Elevation (metres) and the 15N/14N (d15N) isotope ratios (‰) of genera sampled in the PWG and CBG concessions from Gabon. Page 19 of 55

Figure 14. Relationship between Elevation (metres) and the 34S/32S (d34S) isotope ratios (‰) of genera sampled in the PWG and CBG concessions from Gabon. Page 20 of 55

Aucoumea klaineana (Okoume) Significant differences are evident in the mean values of non-exchangeable figure 23.1 shows insignificant differentiation in the oxygen isotope ratios, D/H isotope ratios (P=0.036), and the 34S/32S isotope ratios (P=0.000) of the however, the scales of the isotope ratios of carbon (2), extracted hydrogen Aucoumea klaineana samples between the CBG and PWG concessions. (3), non-exchangeable D/H from cellulose (4), nitrogen (5), and sulfur (6) Figure 17 shows there is a lot of overlap in the non-exchangeable D/H all show good ranges in their respective values. Figure 23.2 shows that isotope ratios between the two concessions limiting its applicability as a carbon isotope ratios may have been a more useful classifier to each concession classifier. However, figure 20 shows there is very little overlap concession were it not for a single sample in the north west of the main between the two concessions in the sulfur isotope ratios of the samples portion of the PWG concession having such an enriched carbon isotope from the two concessions. ratio. Figure 23.3 and figure 23.4 do not show comparable spatial patterns in hydrogen isotope ratios, figure 23.3 shows that, generally, the hydrogen The FreeViz (figure 21) shows that the most useful measurements to isotope ratios of Aucoumea are enriched in the CBG concession and classify Aucoumea to the Precious Woods concession are the oxygen and depleted in the Precious Woods Group concession, whereas Figure 23.4 carbon stable isotope ratios. The most useful measurements to classify seems to show the opposite save for one sample in the CBG concession Aucoumea to the CBG concession are the non-exchangeable D/H isotope that has an enriched hydrogen isotope ratio in both situations. Figure 23.5 ratios from cellulose (dDnit). The Principle Component Analysis (PCA) shows that the CBG concession has some populations of δ15N enriched highlights that, while there is good separation in the data between the two Aucoumea close to δ15N depleted Aucoumea, whereas most of the samples concessions, there are also similarities (figure 22). The k-means silhouette collected in the Precious Woods Group concession are relatively depleted scores show that it is most likely that the data has two clusters (48% in δ15N with the exception of a sample in the north west of the main body probability), however, there are also reasonable possibilities that there are of the concession. This same sample had an enriched carbon isotope ratio. more clusters to the data (3 clusters = 36.5%, 4 clusters 37.3%). Figure 23.6 shows that the sulfur isotope ratios of Aucoumea within each Figure 23 highlights the spatial patterns in the stable isotope ratios in concession are consistent. The CBG concession shows enriched sulfur Aucoumea between and within the CBG and Precious Woods Group isotope ratios whereas the PWG shows depleted. Only one sample in the concessions. The colour scales of the figures are set to maximise the visual CBG concession shows a relatively depleted sulfur isotope ratio. differentiation in values rather than to reflect whether differences are Significant differences are evident between the mean values of non- significant or not. For example, there are no significant differences exchangeable D/H (P=0.020), 15N/14N (P=0.004), 34S/32S (P=0.046) of between the oxygen, hydrogen (extracted), carbon and nitrogen stable Aucoumea by soil type. Significant differences are not evident in the isotope ratios of Aucoumea between the two concessions. The scale for oxygen, hydrogen and carbon isotopes by soil type (figures 24 – 29).

Page 21 of 55

Figure 15. Boxplots of 18O/16O stable isotope ratios (‰) of Aucoumea klaineana from the CBG (n=9) and PWG (n=7) concessions.

Figure 16. Boxplots of D/H stable isotope ratios (‰) of Aucoumea klaineana from the CBG (n=9) and PWG (n=7) concessions.

Page 22 of 55

Figure 17. Boxplots of non-exchangeable D/H from cellulose stable isotope ratios (‰) of Aucoumea klaineana from the CBG (n=9) and PWG (n=7) concessions.

Figure 18. Boxplots of 13C/12C stable isotope ratios (‰) of Aucoumea klaineana from the CBG (n=9) and PWG (n=7) concessions.

Page 23 of 55

Figure 19. Boxplots of 15N/14N stable isotope ratios (‰) of Aucoumea klaineana from the CBG (n=9) and PWG (n=7) concessions.

Figure 20. Boxplots of 34S/32S stable isotope ratios (‰) of Aucoumea klaineana from the CBG (n=9) and PWG (n=7) concessions.

Page 24 of 55

Figure 21. FreeViz of the analysed stable isotope ratios of Aucoumea klaineana samples from the CBG (n=9) and Precious Woods Group (n=7) showing which measurements are more relevant at identifying each concession. d18O = 18O/16O, dD = D/H, dD nit = non-exchangeable D/H from cellulose, d13C = 13C/12C, d15N = 15N/14N and d34S = 34S/32S isotope ratios.

Page 25 of 55

Figure 22. Principle Component Analysis of Okoume (Aucoumea klaineana) from the Precious Woods and CBG concessions. Data used to generate the PCA were the 18O/16O, D/H, non-exchangeable D/H from cellulose, 13C/12C, 15N/14N and 34S/32S isotope ratios.

Page 26 of 55

Figure 23. Inverse Distance Weighting (global) isoscapes of Aucoumea klaineana stable isotope data from samples collected in the CBG and PWG concessions. 1. 18O/16O, 2. 13C/12C, 3. D/H, 4. Non-exchangeable D/H from cellulose, 5. 15N/14N, 6. 34S/32S.

Page 27 of 55

Figure 24. 18O/16O (‰) of Aucoumea klaineana compared to soil type of the sampling location [Jones et al, 2013].

Figure 25. D/H (‰) of Aucoumea klaineana compared to soil type of the sampling location [Jones et al, 2013].

Page 28 of 55

Figure 26. non-exchangeable D/H of cellulose (‰) of Aucoumea klaineana compared to soil type of the sampling location [Jones et al, 2013].

Figure 27. 13C/12C (‰) of Aucoumea klaineana compared to soil type of the sampling location [Jones et al, 2013].

Page 29 of 55

Figure 28. 15N/14N (‰) of Aucoumea klaineana compared to soil type of the sampling location [Jones et al, 2013].

Figure 29. 34S/32S (‰) of Aucoumea klaineana compared to soil type of the sampling location [Jones et al, 2013].

Page 30 of 55

Dacryodes buettneri results Significant differences in the oxygen, hydrogen, non-exchangeable patterns in the hydrogen stable isotopes of the extracted wood and of the hydrogen, and nitrogen stable isotope ratios were not evident between the non-exchangeable hydrogen from cellulose of the Dacryodes. The CBG and Precious Woods concessions. Differences between the mean relatively high variation of hydrogen isotope ratios in the CBG concession values of the carbon and sulfur stable isotope ratios between the two limit the statistical separation of Dacryodes using the hydrogen isotope concessions were observed. These differences were close to being ratios alone. Figure 37.5 shows that Dacryodes buettneri have lower significant (P=0.143 for carbon isotope ratios, P=0.127 for sulfur isotope nitrogen isotope ratios in the CBG concession relative to the Precious ratios), however, as the data overlapped clear segregation was not evident Woods Group concession save for a single sample in CBG that was between the two concessions based on the eight samples of Dacryodes particularly enriched in δ15N. Figure 37.6 shows that Dacryodes follows a buettneri sampled (figures 30 – 35). similar pattern to Aucoumea in its sulfur isotope ratios; the samples from the CBG concession are more enriched in δ34S than in the Precious Woods The first three principle components of the data show that there is good concession save for one sample in CBG that had a depleted sulfur isotope segregation in the data, though some datapoints from CBG and Precious ratio. Woods are more similar than they are different. This is also evident when observing the k-means silhouette scores of the PCA, there is a 24% Significant differences in the mean values of the oxygen, hydrogen, and probability that there are two clusters in the data, however, there is 24.7% carbon stable isotope ratios were not observed between Dacryodes probability that there are 3 clusters or 4 clusters within the data. The best samples classified to different soil types. However, significant differences multivariate model for the dataset tested so far, Naïve Bayes using the first between the mean values of non-exchangeable D/H from three Principle components, is only able of achieving 6/8 correct cellulose(P=0.016), nitrogen (P=0.029) and sulfur (P=0.042) stable isotope assignment of the Dacryodes to the correct concession (figure 36). ratios were observed classifying the Dacryodes samples by soil type. Dacryodes samples grown in Ferralic cambisol were notably different to The inverse distance weighted (IDW) isoscapes (figure 37) give insight into those grown in Xanthic ferralsol and Ferralic arenosol in terms of the non- the spatial patterns of the Dacryodes data between and within the two exchangeable D/H from cellulose and 15N/14N isotope ratios (figures 40 concessions. Figure 37.1 shows that there is more variation of the oxygen and 42). Dacryodes in Ferralic arenosol had significantly different sulfur stable isotopes of the Dacryodes within each concession than there is isotope ratios to those grown in Xanthic ferralsol and Ferralic cambisol between the two. Figure 37.2 shows that there would be good segregation (figure 43). of the two concessions by the carbon isotope ratios of the Dacryodes were it not for the high variance in the CBG concession perhaps explaining why There is no relationship between the D/H and 18O/16O isotope ratios of the Student’s t-test shows a nearly significant difference in carbon isotope Aucoumea klaineana and Dacryodes buettneri (figure 44) which are both ratios (P=0.127). Figure 37.3 and 4 show that there are similar spatial of the Burseraceae family.

Page 31 of 55

Figure 30. Boxplots of 18O/16O stable isotope ratios (‰) of Dacryodes buettneri from the CBG (n=5) and Precious Woods Group (n=3) concessions.

Figure 31. Boxplots of D/H stable isotope ratios (‰) of Dacryodes buettneri from the CBG (n=5) and Precious Woods Group (n=3) concessions.

Page 32 of 55

Figure 32. Boxplots of nitrated D/H stable isotope ratios (‰) of Dacryodes buettneri from the CBG (n=5) and Precious Woods Group (n=3) concessions.

Figure 33. Boxplots of 13C/12C stable isotope ratios (‰) of Dacryodes buettneri from the CBG (n=5) and Precious Woods Group (n=3) concessions.

Page 33 of 55

Figure 34. Boxplots of 15N/14N stable isotope ratios (‰) of Dacryodes buettneri from the CBG (n=5) and Precious Woods Group (n=3) concessions.

Figure 35. Boxplots of 34S/32S stable isotope ratios (‰) of Dacryodes buettneri from the CBG (n=5) and Precious Woods Group (n=3) concessions.

Page 34 of 55

Figure 36. Principle components 1, 2 and 3 of the Dacryodes spp. data (variables used: 18O/16O, D/H, non-exchangeable D/H from cellulose, 13C/12C, 15N/14N, 34S/32S).

Page 35 of 55

Figure 37. Inverse distance weighting (global) isoscapes of Dacryodes buettneri stable isotope data 1. 18O/16O, 2. 13C/12C, 3. D/H, 4. Non-exchangeable D/H from cellulose, 5. 15N/14N, 6. 34S/32S.

Page 36 of 55

Figure 38. 18O/16O (‰) of Dacryodes buettneri compared to soil type of the sampling location [Jones et al , 2013].

Figure 39. D/H (‰) of Dacryodes buettneri compared to soil type of the sampling location [Jones et al, 2013]

Page 37 of 55

Figure 40. Non-exchangeable D/H from cellulose (‰) of Dacryodes buettneri compared to soil type of the sampling location [Jones et al, 2013].

Figure 41. 13C/12C (‰) of Dacryodes buettneri compared to soil type of the sampling location [Jones et al, 2013].

Page 38 of 55

Figure 42. 15N/14N (‰) of Dacryodes buettneri compared to soil type of the sampling location [Jones et al, 2013].

Figure 43. 34S/32S (‰) of Dacryodes buettneri compared to soil type of the sampling location [Jones et al, 2013].

Page 39 of 55

Figure 44. Relationship between the D/H and 18O/16O isotope ratios of Aucoumea klaineana and Dacryodes buettneri.

Page 40 of 55

How many samples are required to describe natural variation on a site? ‘Sampling area’ is a term that is currently loosely defined. Some have evaluated highly defined areas [Mayer et al., 1995; Leavitt 2010; van der Sleen et al., 2017; and Gori et al., 2018] but none have used explicitly defined areas of land. Though the community would certainly benefit from defining a resolute area, there is no doubt this would be subject to controversy and does not lend itself to the practical limitations of sampling real forests. In this evaluation, sample areas of islands (Kolombangara, Guadalcanal), forestry management units (Taiwan) and concessions (CBG and Precious Woods) defined a ‘site’ category [Watkinson et al., 2020].

Oxygen isotope range vs. samples 1.8 y = 0.3176ln(x) + 0.3436 1.6 R² = 0.3095 1.4 1.2 1.0 0.8

0.6 O range in range (‰) data O in

16 0.4 O/

18 0.2 0.0 0 5 10 15 20 25 30 Samples collected per 'site'

Figure 45. Ranges of oxygen isotope ratios relative to how many samples were collected per site for multiple different taxa.

The majority of the natural range of oxygen isotope ratios on a site can be defined by approximately 5-10 samples as 90% of samples have a range of 1.2‰ irrespective of species (median range = 0.8‰). This is in contrast to some literature [Van der Sleen et al., 2017] which shows oxygen variation to be as great as 9‰ at 1500m altitude, though this seems to be an exceptional case that has not been encountered at all commonly. Leavitt [2010] mostly reviewed studies containing 4-6 samples per site, inter-tree variability ranged from 1-4‰ for oxygen stable isotope ratios.

Hydrogen isotope range vs. samples 30.0

25.0 y = 4.8431ln(x) + 3.8825 R² = 0.2552 20.0

15.0

10.0

D/H range in (‰) data in D/H range 5.0

0.0 0 5 10 15 20 25 30 Samples collected per 'site'

Figure 46. Ranges of hydrogen isotope ratios relative to how many samples were collected per site for multiple different taxa.

Page 41 of 55

Between 5 and 10 samples collected and analysed from a site is sufficient to define the range of D/H ratios on a site as judged by figure 46. 90% of site collections had ranges of up to 19.7‰ irrespective of species (median range = 10.1‰). Leavitt [2010] shows that hydrogen isotope ratios may vary from 5-30‰ between trees on the same site. The maximum range observed thus far is 25.3‰ with 9 samples collected from one site.

Carbon isotope range vs. samples 4.5 4.0 3.5 y = 1.0762ln(x) - 0.0118 R² = 0.4759 3.0 2.5 2.0

1.5 C range in (‰) data in range C

12 1.0 C/

13 0.5 0.0 0 5 10 15 20 25 30 Samples collected per 'site'

Figure 47. Ranges of carbon isotope ratios relative to how many samples were collected per site for multiple different taxa.

Relative to their natural range of approximately -24 to -34‰ in higher C3 plants [Smith and Epstein 1971], carbon stable isotopes vary significantly on a site. There is a strong logarithmic relationship between the number of samples collected and the range of carbon ratios defined by the collection. In some collections the range was as high as 4‰ with only 5 samples. 90% of sites had ranges of carbon isotope ratios up to 2.7‰ irrespective of species. The maximum site variation in Leavitt 2010 for carbon ratios was 3‰, Van der Sleen [et al., 2017] also reviews that tropical tree rings have carbon stable isotope ratios that vary by up to 3‰ on a given site. According to figure 47, more than 10 samples per site would be needed to define the range of carbon ratios. It must be considered that site size (e.g. Kolombangara Island and Taiwan) and bias towards sampling in rainy tropical regions may influence this value for site carbon isotope variability. Regions with variable aridity may have greater carbon isotope variability as trees respond to changes in water stress and vapour pressure by adjusting stomatal conductance, thus, more enriched carbon isotope ratios may be more common [Farquhar et al., 1989].

Page 42 of 55

Sulfur isotope range vs. samples 9.0 8.0 7.0 6.0 y = 1.7944ln(x) - 0.4462 5.0 R² = 0.4627 4.0

3.0 S range in range (‰) S data in

32 2.0 S/

34 1.0 0.0 0 5 10 15 20 25 30 Samples collected per 'site'

Figure 48. Ranges of sulfur isotope ratios relative to how many samples were collected per site for multiple different taxa.

Sulfur isotope ratio variability ranged between 0.3‰ (CBG concession, Gabon) to 8.4‰ (Kolombangara Island, an extinct stratovolcano in the Solomon Islands archipelago) on a site. 90% of site collections had a sulfur isotope range of 3.8‰ irrespective of taxonomic classification. In order to capture natural variation in future sample datasets, up to 10 samples must be collected per ‘site’. It is challenging to put this result into context as it does not appear that there is a publication that has investigated sulfur isotope variance between trees on the same site despite the fact sulfur is one of the strongest geographic discriminators [Thompson et al., 2010; Rummel et al., 2010; Pianezze et al., 2019]. This evaluation represents a first attempt at defining the relationship between sulfur isotope ratios ranges in tropical trees and the quantity of samples collected on a site.

Discussion Baillonella toxisperma, Cylicodiscus gabonensis, Guibourtia tessmannii, Lophira alata, Nauclea diderrichii, Pterocarpus soyauxii, Berlinia confusa, and Didelotia africana Data presented in this report establishes a basis for evaluating geographic origin claims of the aforementioned species from the concessions where they were collected from within Gabon. However, little can be stated about what exactly the data means until a point where collections are large enough to permit thorough analysis and interpretation. There are clear, significant differences between these species even though they were collected mostly on the same sites. This means that data from one species cannot easily be applied to another at this stage. Watkinson et al., 2020] suggest that as collections of samples and data grow, it is likely that higher- order patterns will become evident and this may permit interpretation of the origin of various species using data from another species such as using Dunbar lines [Dunbar & Wilson, 1983] to convert between datasets, or using mechanistic models to forecast data [Roden et al., 2000; Cueni et al., 2019].

Of all sampled timbers, Pterocarpus soyauxii showed the most negative nitrogen stable isotope ratios. This was expected as Pterocarpus soyauxii is a nitrogen-fixing tree and perhaps acts as a primary nitrogen source in the areas where it was sampled. The results are supported by Heitz [2011] who demonstrated that leaves of leguminous trees had more negative 15N/14N isotope ratios relative to non-leguminous trees. Variability in nitrogen isotope composition was also demonstrated by Heitz [2011] to be influenced by sun/shade. Results from nitrogen fixers are important from an ecological perspective as they are one of the nitrogen sources for other nearby trees due to the nitrogen they fix in the soil and the distribution of the nitrogen by mycorrhiza. Mycorrhizal type varies with climate, such as arbuscular mycorrhiza which are more abundant in tropical areas

Page 43 of 55

such as Gabon. Forecast arbuscular mycorrhiza populations in Gabon by Steidinger [2019] show that varying proportions of this mycorrhiza exist across Gabon and between the two concessions that were sampled in this project. This variable may give rise to geographic variation in nitrogen isotope ratios of Pterocarpus and other flora and fauna that source nitrogen through mycorrhiza networks [Williamson et al., 1990]. Aucoumea klaineana Aucoumea Klaineana (Okoume) is a fast-growing, light-demanding pioneer tree [Zaou et al., 1998]. In growing conditions with favourable light, okoume can develop rapidly. In terms of photosynthesis, it can be assumed that the main source of water for incorporation into sugars is precipitation [Ohashi et al., 2016]. It is therefore difficult to understand why there is no strong relationship between the hydrogen and oxygen isotope ratios of okoume, and very little trend with respect to elevation and its water hydrogen and oxygen isotope ratios. The lack of a trend may suggest that the forces of nature may be conspiring against each other, such as the rainout effect and the continental effect [Ohashi et al., 2016], or that despite the fact there are differences in rainfall and elevation, perhaps there is little meaningful difference in relative humidity between the Precious Woods Group and CBG concessions [Roden et al., 2000; Fritts et al., 1971]. One of the most important activities of rainforests is regulating the temperature of the atmosphere by removing CO2 and moderating rainfall. Provided that excessive evaporation does not occur, precipitation can be expected to follow the Global Meteoric Water Line [Craig and Boato, 1955; Craig 1961]. Evapotranspiration in rainforests produces rain for the rainforest and contributes to atmospheric cooling. However, this cyclical evapotranspiration may be considered as ‘excessive evaporation’ meaning that the precipitation tropical trees receive is effectively not meteoric precipitation [Marryanna et al., 2017]. One idea to investigate this further would be to sample Aucoumea klaineana under very different conditions (i.e. not just within Gabon, but elsewhere in the world), however, this is marred by the fact that Okoume chiefly grows in Gabon and surrounding countries. Okoume and Dacryodes are both of the Burseraceae family, this study offers the opportunity to see if their data is compatible. Based on figures 3, 4 and 44 it does not appear to be. There are clear differences in the oxygen and hydrogen isotope ratios of Dacryodes and Okoume meaning that their data cannot be applied to one another.

Differences are evident in the nitrogen isotope ratios of the okoume samples from the PWG and CBG concessions. Nitrogen isotope ratios are expected to increase with the age of a tree. Heitz et al. [2011] showed that a 2‰ increase in nitrogen isotope ratio was evident over an 80-year chronology of three topical timbers. Variation in age of trees that were sampled may have played a factor in the nitrogen isotope ratio that was measured. Nitrogen deposition is increasing in the tropics [Heitz et al., 2010; Hogberg and Johannisson 1993], one of the suggested means is by tropospheric NOx deposition. It is not inconceivable that the differences between the two concessions may be partially explained by this, however, there are insufficient samples and conditions to be able to assess if this was the case in this study. Adding nitrogen to trees is associated with 2- 15 losses of NO3 from a tree and is a mechanism by which trees can become more enriched in N [Hogberg and Johannisson 1993], this phenomenon is also related to cation loss from soil and increase in soil acidity. Different soils can be more resistant to change perhaps explaining why significant differences were observed in the nitrogen isotope ratios okoume in the three different soil types. It also must be considered that segregating the okoume samples by soil type rather than by concession may, on one hand, suggest that there is a predictable overall pattern within Gabon that may aid the origin classification of unknown samples of okoume; on the other, it may be an arbitrary segregation and the significant differences that were observed are the result of the fact that very few samples of okoume have been examined in this study (n=16). Though okoume was the best-sampled species the data presented only portrays a snapshot of reality. Greater sampling of okoume in the Precious Woods concession would give spatial models more certainty. Sampling other concessions and countries will give a better idea of the differentiation and classification that is possible with okoume. One challenge to using nitrogen isotope ratios as an origin classifier is that Okoume exists as plywood in western markets. Veneers in plywood are typically glued using formaldehyde-urea-based resins [Desch &

Page 44 of 55

Dinwoodie 1996; Negro et al., 2011], some also include melamine. Urea and melamine contain nitrogen, this exogenous nitrogen must be removed for the natural nitrogen to be accessed for comparison. The method of sample preparation posed in this study is also intended to mitigate the effect of resins and other secondary metabolic compounds and therefore is a practical solution to the problems that glues present in plywood.

Sulfur isotope ratios were the best discriminator of okoume between the two concessions. There are clear differences between CBG and PWG concessions in terms of distance from the sea, tropospheric sulfate deposition [Novak & Bottrell, 1996] and elevation that may explain the differences in observed values either individually or in combination. Soil type also appeared to be important in separating ranges of sulfur isotope ratios in okoume; different soils contain varying concentrations of sulfate. All these potential predictors suggest that there is a higher spatial structure of sulfur isotope ratios in okoume across Gabon that is likely to be useful at identifying the origin of okoume. Dacryodes buettneri Eight samples of Dacryodes buettneri were sampled in the Precious Woods (n=3) and CBG (n=5) concessions. As judged by the figures illustrating the range of isotope ratios relative to the quantity of samples collected from a ‘site’, there is insufficient sampling of Dacryodes buettneri to come to any resolute conclusions about the data. It is likely collecting more samples of Dacryodes from the CBG and Precious Woods concession would elucidate the true differentiation possibilities between the two origins. Though multiple multi-variate models were attempted with the Dacryodes data, it is statistically inappropriate to attempt to use three to six variables to classify eight samples to two origins. The Principle Component Analysis and the silhouette scores of the data show that there isn’t any real clustering of the data, this is because there aren’t enough samples to define clusters. Further interpretation of the differentiation possibilities of Dacryodes buettneri using stable isotope ratio measurements is beyond the scope of this paper.

Nonetheless, it is interesting that the sulfur isotope ratios of Dacryodes buettneri appear to follow a similar trend to that of the Aucoumea klaineana reference samples. It is not unreasonable to expect that further collection of this timber may also reveal a higher spatial structure within Gabon that can be used to classify the origin of Dacryodes timber. Conclusions Despite the quantity of samples and species being relatively low, the data acquired establishes a basis of evaluation for assessing geographic origin claims of certain forest products including plywood and laminates from Gabon. The differences in the sulfur isotope ratios of Okoume and Dacryodes reference samples from Precious Woods Group (PWG) and Compagnie des Bois du Gabon (CBG) forest concessions suggest that regional scale origin classification may be realised with high enough frequency of refence sample collection. This study helped to establish the minimum number of reference samples (five per species per site) required in order to account for variance of isotopic distribution at a single sampling site. Low frequency of sampling per site and limited sample numbers mean that investigation of the Dunbar line hypothesis was not possible on this occasion. Further, higher resolution sampling of target species including Okoume and Dacryodes will address this much needed comparison as it may enable more efficient allocation of reference sampling resources. Further sampling of Burseraceae family timbers in the tropics may permit a global model for verifying their geographic origin. Future work Expanding the collection reference samples will be necessary to investigate regional classification further. It is hypothesised that having greater quantities of sample data will improve the discrimination between concessions or regions, as well as allowing for comparison of stable isotopes both within a single sampling site and single taxa. A higher frequency of reference sample collection will also enable assessment of the suitability

Page 45 of 55

of specific taxa to act as isotopic proxies for other species (Dunbar lines). Natural variability of isotopic distribution within a site is still not fully understood, and it is anticipated that large scale sample collections for specific taxa will also enable a better understanding of this. So far, the study has focused on whole wood and cellulose components of timber reference samples. Several alternative analytical methods detail procedures for analysing alternative metabolic fractions including lignin and proteins in the form of amino acids. Furthermore, isotope methods are already being used to measure the stable isotopes of position-specific atoms within a selection of molecules such as ethanol, providing higher resolution information on the source water incorporated during metabolism. Analysis of alternative fractions and position-specific isotope ratios within a molecule such as glucose may yield higher quality results and aid the discrimination of reference samples taken from concessions within the same country.

Acknowledgements Thanks to

• U.S. Department of Agriculture’s Animal and Plant Health Inspection Service (APHIS), U.S. Forest Service, and U.S. Department of State for funding. • Beth Lebow, Alex Moad and their respective colleagues at the US Forest Service for overseeing the World Forest ID project. • FSC, Phil and Tatai Guillery for collecting samples in the Solomon Islands. • FSC, Emily Crumley and Cynel Moundounga for collecting samples in Gabon. • KFPL for collecting samples and supporting collection efforts in the Solomon Islands. • Meaghan Parker-Forney and her respective colleagues at the World Resources Institute for overseeing the World Forest ID project. • Jade Saunders of Forest Trends for her advisory and coordination support to the World Forest ID project. • Peter Gasson, Isabella Miles-Bunch and Sara Redstone at the Royal Botanic Gardens at Kew for assisting with sample shipment documentation, receipt of samples in quarantine and their expert help in verifying declared information of field collectors. • Tim Fox, Jackie Borrows, and Matt Watkinson for spending 100s of collective hours assembling sample collection packs. • Roger Young for type-checking the document throughout its various editions. • Agroisolab GmbH staff for spending 100s of collective hours preparing and measuring samples.

References [Orange] Demsar J, Curk T, Erjavec A, Gorup C, Hocevar T, Milutinovic M, Mozina M, Polajnar M, Toplak M, Staric A, Stajdohar M, Umek L, Zagar L, Zbontar J, Zitnik M, Zupan B (2013) Orange: Data Mining Toolbox in Python, Journal of Machine Learning Research 14(Aug): 2349−2353.

[SAGA] Conrad, O., Bechtel, B., Bock, M., Dietrich, H., Fischer, E., Gerlitz, L., Wehberg, J., Wichmann, V., and Böhner, J. (2015): System for Automated Geoscientific Analyses (SAGA) v. 2.1.4, Geosci. Model Dev., 8, 1991- 2007, doi:10.5194/gmd-8-1991-2015.

Boner, M., & Förstel, H. (2004). Stable isotope variation as a tool to trace the authenticity of beef. Analytical and bioanalytical chemistry, 378(2), 301-310.

Boner. M., Somner, Th., Erven. C., Forstel. H. Stable isotopes as a tool to trace back the origin of wood. Proceedings of the international workshop "Fingerprinting methods for the identification of timber origins", October 8-9 2007, Bonn/Germany.

Page 46 of 55

Born C,. Vignes H., Muloko N., Wickings J., Hossaert-McKey M., Chevallier M.H. Isolation and characterization of polymorphic microsatellite loci from Aucoumea klaineana Pierre (Burseraceae), a tropical rainforest tree of Central Africa. Molecular Ecology Notes (2006) 6, 1054–1056. doi: 10.1111/j.1471-8286.2006.01431.x

Born C., Alvarez N., McKey D., Ossari S., Wickings E.J., Hossaert-McKey M., Chevallier M. Insights into the biogeographical history of the Lower Guinea Forest Domain: evidence for the role of refugia in the intraspecific differentiation of Aucoumea klaineana. Molecular Ecology (2011) 20, 131–142. doi: 10.1111/j.1365- 294X.2010.04919.x

Born, C., Hardy, O. J., Chevallier, M.H., Ossari, S., Attéké, C., Wickings, E. J., & Hossaert-McKey, M. Small-scale spatial genetic structure in the Central African rainforest tree speciesAucoumea klaineana: a stepwise approach to infer the impact of limited gene dispersal, population history and habitat fragmentation. 2008. Molecular Ecology, 17(8), 2041–2050. doi:10.1111/j.1365-294x.2007.03685.x

Bowen, G.J., Revenaugh, J. Interpolating the isotopic composition of modern meteoric precipitation. Water resources research, Vol. 39, No.10, 1299. doi:10.1029/2003WR002086, 2003.

Camin, F., Boner, M., Bontempo, L., Fauhl-Hassek, C., Kelly, S., Riedl, J., Rossmann, A., Stable isotope techniques for verifying the declared geographical origin of food in legal cases, Trends in Food Science & Technology (2017), doi: 10.1016/j.tifs.2016.12.007.

Cheung, C., Studies of the nitration of cellulose - application in new membrane materials, T, University of British Columbia, 2014, doi : http://dx.doi.org/10.14288/1.0072160.

Conrad, O., Bechtel, B., Bock, M., Dietrich, H., Fischer, E., Gerlitz, L., Wehberg, J., Wichmann, V., and Böhner, J. (2015): System for Automated Geoscientific Analyses (SAGA) v. 2.1.4, Geosci. Model Dev., 8, 1991-2007, doi:10.5194/gmd-8-1991-2015.

Craig H., Boato. Isotopes. Annu. Rev. Phys. Chem. 1955.6:403-432

Craig, H. Isotopic Variations in Meteoric Waters. Science, New Series, Vol. 133, No. 3465 (May 26, 1961), 1702- 1703.

Cueni F., Nelson D.B., Boner M., Ansgar K. Mechanistic oxygen and hydrogen isotope models can predict the geographic origin of fruit. Geophysical Research Abstracts . 2019, Vol. 21, p1-1. 1p

Demsar J, Curk T, Erjavec A, Gorup C, Hocevar T, Milutinovic M, Mozina M, Polajnar M, Toplak M, Staric A, Stajdohar M, Umek L, Zagar L, Zbontar J, Zitnik M, Zupan B (2013) Orange: Data Mining Toolbox in Python, Journal of Machine Learning Research 14(Aug): 2349−2353.

DeNiro, M.J., Cooper, L.W. Post-photosynthetic modification of oxygen isotope ratios of carbohydrates in the potato: Implications for paleoclimatic reconstruction based upon isotopic analysis of wood cellulose. Geochimica et Cosmochimica Acta Vol. 53, pp2573-2580. 1989.

Desch H.E., Dinwoodie J.M. Timber Structure, Properties, Conversion and Use. 1996, 7th Edition. Macmillan Press Ltd. ISBN 978-0-333-60905-7. DOI 10.1007/978-1-349-13427-4

Dunbar J, Wilson AT, Oxygen and Hydrogen Isotopes in Fruit and Vegetable Juices. Plant Physiology Jul 1983, 72 (3) 725-727. 1983.

EIA, 2019. https://content.eia-global.org/assets/2019/03/Toxic_Trade_Executive_Summary-web.pdf. Last accessed 21-04-2020

Page 47 of 55

EIA. Toxic Trade - Forest Crimin in Gabon and the Republic of Congo and contamination of the US Market. 2019. https://content.eia-global.org/posts/documents/000/000/830/original/Toxic_Trade_EIA- web.pdf?1553480150

FAO, 2015 http://www.fao.org/forest-resources-assessment/current-assessment/country-reports/en/. Last accessed 21-4-2020.

FAO. (2002). Etude de cas d’aménagement forestier exemplaire en Afrique centrale: la concession forestière sous aménagement durable (CFAD), Gabon. Par Nicolas Bayol, octobre 2002. Document de travail FM/15F. Service de la mise en valeur des ressources forestières, Division des ressources forestières. FAO, Rome (non publié).

Farquhar G.D., Ehleringer J.R., Hubick K.T. Carbon Isotope Discrimination and Photosynthesis. Annu. Rev. Plant Physiol. Plant Mol. Biol. 1989. 40:503-37

Farquhar, G.D., Ehleringer, J.R., Hubick, K.T. Carbon Isotope Discrimination and Photosynthesis. Annu. Rev. Plant Physiol. Plant Mol. Biol. 1989. 40:503-37.

Fritts H.C., Blasing T.J., Hayden B.P. Kutzback J.E. 1971. Multivariate techniques for specifying tree-growth and climate relationships and for reconstructing anomalies in paleoclimate. Journal of applied meteorology. Vol. 10, No. 5. 845-864.

Global Modeling and Assimilation Office (GMAO) (2015), MERRA-2 tavgM_2d_aer_Nx: 2d,Monthly mean,Time- averaged,Single-Level,Assimilation,Aerosol Diagnostics V5.12.4, Greenbelt, MD, USA, Goddard Earth Sciences Data and Information Services Center (GES DISC), Accessed: [10.04.2020], 10.5067/FH9A0MLJPC7N

Gori Y, Stradiotti A, Camin F (2018) Timber isoscapes. A case study in a mountain area in the Italian Alps. PLoS ONE 13(2): e0192970. https://doi.org/10.1371/journal.pone.0192970.

Gori Y., Wehrens R., Greule M., Keppler F., Ziller L., La Porta N., & Camin F. (2013). Carbon, hydrogen and oxygen stable isotope ratios of whole wood, cellulose and lignin methoxyl groups of Picea abiesas climate proxies. Rapid Communications in Mass Spectrometry, 27(1), 265-275.

Heaton, K., Kelly, S. D., Hoogewerff, J., & Woolfe, M. (2008). Verifying the geographical origin of beef: The application of multi-element isotope and trace element analysis. Food Chemistry, 107(1), 506-515.

Heitz P., Dunisch O., Wanek W. Long-Term Trends in Nitrogen Isotope Composition and Nitrogen Concentration in Brazilian Rainforest Trees Suggest Changes in Nitrogen Cycle. Environ. Sci. Technol. 2010, 44, 1191–1196

Heitz P., Turner B.J., Wanek W., Richter A., Nock C.A., Wright S.J. Long-Term Change in the Nitrogen Cycle of Tropical Forests. Science 334, 664 (2011); DOI: 10.1126/science.1211979

Hillis W.E. Heartwood and Tree Exudates. 1st Edition. Springer-Verlag Berlin Heidelberg 1987. Page 21. DOI 10.007/978-3-642-72534-0

Hogberg P., Johannisson C. 15N abundance of forests is correlated with losses of nitrogen. Plant and Soil 157: 147-150, 1993.

Horacek, M., Jakusch, M., & Krehan, H. (2009). Control of origin of larch wood: discrimination between European (Austrian) and Siberian origin by stable isotope analysis. Rapid Communications in Mass Spectrometry, 23(23), 3688-3692.

Page 48 of 55

Huffman G.J, Bolvin D.T. (2019), GPCP Precipitation Level 3 Monthly 0.5-Degree V3.0 beta, Edited by George J. Huffman and David T. Bolvin, Greenbelt, Maryland, USA, NASA GES DISC, Accessed: [Data Access Date], 10.5067/TTO0VJF2FSYR

ITC (2018). Main markets, in terms of export value, for the product ’44 wood and articles of wood, wood charcoal’ exported by Gabon in 2017.

Jolivet, C., & Degen, B. (2012). Use of DNA fingerprints to control the origin of sapelli timber (Entandrophragma cylindricum) at the forest concession level in Cameroon. Forensic Science International: Genetics, 6(4), 487– 493. doi:10.1016/j.fsigen.2011.11.002

Jones, A., Breuning-Madsen, H., Brossard, M., Dampha, A., Deckers, J., Dewitte, O., Gallali, T., Hallett, S., Jones, R., Kilasara, M., Le Roux, P., Micheli, E., Montanarella, L., Spaargaren, O., Thiombiano, L., Van Ranst, E., Yemefack, M. , Zougmoré R., (eds.), 2013, Soil Atlas of Africa. European Commission, Publications Office of the European Union, Luxembourg. 176 pp.

Kagawa, A., Abe, H., Fujii, T., & Itoh, Y. (2008). Stable isotopes and inorganic elements as potential indicators of timber geographic origin.

Kagawa, A., Leavitt, S.W. Stable carbon isotopes of tree rings as a tool to pinpoint the geographic origin of timber. J Wood Sci (2010) 56:175–183. DOI 10.1007/s10086-009-1085-6.

Kelly, S., Baxter, M., Chapman, S., Rhodes, C., Dennis, J., & Brereton, P. (2002). The application of isotopic and elemental analysis to determine the geographical origin of premium long grain rice. European Food Research and Technology, 214(1), 72-78.

Kelly, S., Marigheto, N., Baxter, M., Brereton, N., Rose, M., Homer, D., Matthews, K., Hoelzl, S., Rummel, S., Dekannt, D. THE BRITISH BEEF ORIGIN PROJECT (BBOP). UK Food Standards Agency, Q01123. 2007

Keppler, F., Harper, D. B., Kalin, R. M., Meier‐Augenstein, W., Farmer, N., Davis, S., & Hamilton, J. T. (2007). Stable hydrogen isotope ratios of lignin methoxyl groups as a paleoclimate proxy and constraint of the geographical origin of wood. New Phytologist, 176(3), 600-609.

Kersenty, A (2019). Certification of tropical forests: a private instrument of public interest? A focus on the Congo Basin

Koumba Zaou, P., Mapaga, D., & Verkaar, H. . (1998). Effect of shade on young Aucoumea klaineana Pierre trees of various provenance under field conditions. Forest Ecology and Management, 106(2-3), 107–114. doi:10.1016/s0378-1127(97)00301-0

Leavitt, S.W. Tree Ring C-H-O isotope variability and sampling. Science of the Total Environment 408 (2010) 5244-5253.

Li, A., Keely, B., Chan, S. H., Baxter, M., Rees, G., & Kelly, S. (2014). Verifying the provenance of rice using stable isotope ratio and multi-element analyses: a feasibility study. Quality Assurance andSafety of Crops & Foods, 7(3), 343-354.

Loader, N.J., Robertson, I., McCarroll. Comparison of stable carbon isotope ratios in the whole wood, cellulose and lignin of oak-tree rings. Palaeography, Palaeoclimatology, Palaeocology 196 (2003) 395 – 407.

Marryanna, L., Kosugi, Y., Itoh, M., Noguchi, S., Takanashi, S., Katsuyama, M., . . . Siti-Aisah, S. (2017). Temporal variation in the stable isotopes in precipitation related to the rainfall pattern in a tropical rainforest in peninsular Malaysia. Journal of Tropical Forest Science, 29(3), 349-362. Retrieved April 29, 2020, from www.jstor.org/stable/44272913

Page 49 of 55

Mayer B., Fritz P., Prietzel J., Krouse H.R. The use of stable sulfur and oxygen isotope ratios for interpreting the mobility of sulfate in aerobic forest soils. Applied Geochemistry. Vol 10, pp 161-173, 1995.

Muloko-Ntoutoume N., Petit R.J., White L., Abernathy K. Chloroplast DNA variation in a rainforest tree (Aucoumea klaineana, Burseraceae) in Gabon. Molecular Ecology (2000) 9, 359–363

NASA/METI/AIST/Japan Spacesystems, and U.S./Japan ASTER Science Team. ASTER Global Digital Elevation Model V003. 2019, distributed by NASA EOSDIS Land Processes DAAC, https://doi.org/10.5067/ASTER/ASTGTM.003. Accessed 2020-04-16.

Negro F., Cremonini C., Zanuttini R. A new wood-based lightweight composite for boatbuilding. WOOD RESEARCH. 56 (2): 2011. 257-266

NEPcon, 2017. https://www.nepcon.org/sourcinghub/timber/timber-gabon last accessed 21-4-2020

Novak, M., Bottrell, S.H. Sulfur Isotope Signals in Forest Soils of Central Europe along an Air Pollution Gradient. Environ. Sci. Technol. 1996, 30, 3473-3476.

Ohashi S., Durgante F.M., Kagawa A., Kajimoto T., Trumbore S.E., Xiaomei X., Ishizuka M., Higuchi N. Seasonal variations in the stable oxygen isotope ratio of wood cellulose reveal annual rings of trees in a Central Amazon terra firme forest. Oecologia (2016) 180:685–696. DOI 10.1007/s00442-015-3509-x

Pianezze S., Bontempo L., Perini M., Tonon A., Ziller L., Pietro F., Camin F. δ34S for tracing the origin of cheese and detecting its authenticity. 2019. Journal of Mass Spectrometry, Special issue research article. https://doi.org/10.1002/jms.4451

Pilgrim, T. S., Watling, R. J., & Grice, K. (2010). Application of trace element and stable isotope signatures to determine the provenance of tea (Camelliasinensis) samples. Food Chemistry, 118(4), 921-926.

Rees, G.O. Verifying the declared origin of timber using stable isotope ratio and multi-element analyses. 2015. Thesis, University of York, UK. http://etheses.whiterose.ac.uk/9522/, last accessed 23.04.2018.

Regulation (EU) No 995/2010. Guidance document for The EU Timber Regulation. Brussels, 12.2.2016 C(2016) 755 final.

Roden, J.S., Lin, G., Ehleringer, J.R. A mechanistic model for interpretation of hydrogen and oxygen isotope ratios in tree-ring cellulose. Geochimica et Cosmochimica Acta, Vol. 64, No. 1, pp. 21–35, 2000

RStudio Team. RStudio: Integrated Development Environment for R. Year: 2016. RStudio, Inc. Boston, MA

Rummel S., Hoelzl S., Horn P., Rossmann A., Schlicht C. The combination of stable isotope abundance ratios of H, C, N and S with 87Sr/86Sr for geographical origin assignment of orange juices. Food Chemistry 118 (2010) 890–900

Smith, B. N., & Epstein, S. (1971). Two Categories of 13C/12C Ratios for Higher Plants. PLANT PHYSIOLOGY, 47(3), 380–384. doi:10.1104/pp.47.3.380

Steidinger, B.S., Crowther, T.W., Liang, J. et al. Climatic controls of decomposition drive the global biogeography of forest-tree symbioses. Nature 569, 404–408 (2019). https://doi.org/10.1038/s41586-019-1128-0

Thompson, A. H., Chesson, L. A., Podlesak, D. W., Bowen, G. J., Cerling, T. E., & Ehleringer, J. R. (2009). Stable isotope analysis of modern human hair collected from Asia (China, India, Mongolia, and Pakistan). American Journal of Physical Anthropology, NA–NA. doi:10.1002/ajpa.21162

Timber Trade Portal, 2020. https://www.timbertradeportal.com/countries/gabon/ last accessed 21-4-2020.

Page 50 of 55

U.S. Fish and Wildlife Service. Office of Law Enforcement. 18 USC 42-43. 16 USC 3371-3378. Lacey Act

UN Comtrade, https://comtrade.un.org/, looking at data for logs (HS code 4403) for non-coniferous species from tropical countries.

Van der Sleen, P., Zuidema, P. A., & Pons, T. L. (2017). Stable isotopes in tropical tree rings: theory, methods and applications. Functional Ecology, 31(9), 1674–1689. doi:10.1111/1365-2435.12889

Watkinson C.J., Rees G.O., Boner M., Hofem S., Erven C. Establishing a basis for evaluating geographic origin claims of timber from the Solomon Islands using stable isotope ratio analysis. [In publication]

White L., Abernethy K., Oslisly R., Maley Jean. (1996). L'okoumé (Aucoumea klaineana) : expansion et déclin d'un arbre pionnier en Afrique Centrale au cours de l'Holocène. In : Dynamique à long terme des écosystèmes forestiers intertropicaux : résumés. Paris : ORSTOM, 195-198. Dynamique à Long Terme des Ecosystèmes Forestiers Intertropicaux : Symposium, Paris (FRA), 1996/03/20-22.

Williamson E.A., Tutin C.E.G., Rogers M.E. Fernandez M. Composition of the diet of lowland Gorillas at Lope in Gabon. American Journal of Primatology 21;265-277 (1990)

WRI. 2009 Aménagement de la forêt du Gabon en Novembre 2016. https://gis.forest- atlas.org/server/rest/services/gab/amenagement_forestier_fr/MapServer/4

List of tables Table 1. List of priority timber species in Gabon.

Table 2. Environmental differences between the Precious Woods Group concession and CBG concession.

Table 3. Stable isotope data from taxa sampled from the two concessions in Gabon. Ranges are given when n>1, standard deviations (σ) are given when n>2.

List of figures Figure 1. Locations and taxa of samples collected in the CBG concession in Gabon. Data sources: Global Administrative Dataset (GADM), ASTER GDEM (ASTER GDEM is a product of METI and NASA), Interactive Forest Atlas Forest Management [WRI 2009], DIVA GIS.

Figure 2. Locations and taxa of samples collected in the Precious Woods Group (P.W.G) concession in Gabon. Data sources: Global Administrative Dataset (GADM), ASTER GDEM (ASTER GDEM is a product of METI and NASA), Interactive Forest Atlas Forest Management [WRI 2009], DIVA GIS.

Figure 3. Boxplots of 18O/16O isotope ratios (‰.) Aucoumea (n=16), Clylicodiscus (n=3), Dacryodes (n=8), Lophira (n=4), Nauclea (n=5), and Pterocarpus (n=6).

Figure 4. Boxplots of D/H isotope ratios (‰.) Aucoumea (n=16), Clylicodiscus (n=3), Dacryodes (n=8), Lophira (n=4), Nauclea (n=5), and Pterocarpus (n=6).

Figure 5. Boxplots of D/H nitrated isotope ratios (‰.) Aucoumea (n=16), Clylicodiscus (n=3), Dacryodes (n=8), Lophira (n=4), Nauclea (n=5), and Pterocarpus (n=6).

Page 51 of 55

Figure 6. Boxplots of 13C/12C isotope ratios (‰.) Aucoumea (n=16), Clylicodiscus (n=3), Dacryodes (n=8), Lophira (n=4), Nauclea (n=5), and Pterocarpus (n=6).

Figure 7. Boxplots of 15N/14N isotope ratios (‰.) Aucoumea (n=16), Clylicodiscus (n=3), Dacryodes (n=8), Lophira (n=4), Nauclea (n=5), and Pterocarpus (n=6).

Figure 8. Boxplots of 34S/32S isotope ratios (‰.) Aucoumea (n=16), Clylicodiscus (n=3), Dacryodes (n=8), Lophira (n=4), Nauclea (n=5), and Pterocarpus (n=6).

Figure 9. Relationship between Elevation (metres) and the 18O/16O (d18O) isotope ratios (‰) of genera sampled in the Precious Woods Group and CBG concessions from Gabon.

Figure 10. Relationship between Elevation (metres) and the D/H (dD) isotope ratios (‰) of genera sampled in the Precious Woods Group and CBG concessions from Gabon.

Figure 11. Relationship between Elevation (metres) and the non-exchangeable D/H (dD) isotope ratios (‰) from cellulose of genera sampled in the Precious Woods Group and CBG concessions from Gabon.

Figure 12. Relationship between Elevation (metres) and the 13C/12C (d13C) isotope ratios (‰) of genera sampled in the Precious Woods Group and CBG concessions from Gabon.

Figure 13. Relationship between Elevation (metres) and the 15N/14N (d15N) isotope ratios (‰) of genera sampled in the Precious Woods Group and CBG concessions from Gabon.

Figure 14. Relationship between Elevation (metres) and the 34S/32S (d34S) isotope ratios (‰) of genera sampled in the Precious Woods Group and CBG concessions from Gabon.

Figure 15. Boxplots of 18O/16O stable isotope ratios (‰) of Aucoumea klaineana from the CBG (n=9) and Precious Woods Group (n=7) concessions.

Figure 16. Boxplots of D/H stable isotope ratios (‰) of Aucoumea klaineana from the CBG (n=9) and Precious Woods Group (n=7) concessions.

Figure 17. Boxplots of non-exchangeable D/H from cellulose stable isotope ratios (‰) of Aucoumea klaineana from the CBG (n=9) and Precious Woods Group (n=7) concessions.

Figure 18. Boxplots of 13C/12C stable isotope ratios (‰) of Aucoumea klaineana from the CBG (n=9) and Precious Woods Group (n=7) concessions.

Figure 19. Boxplots of 15N/14N stable isotope ratios (‰) of Aucoumea klaineana from the CBG (n=9) and Precious Woods Group (n=7) concessions.

Figure 20. Boxplots of 34S/32S stable isotope ratios (‰) of Aucoumea klaineana from the CBG (n=9) and Precious Woods Group (n=7) concessions.

Figure 21. FreeViz of the analysed stable isotope ratios of Aucoumea klaineana samples from the CBG (n=9) and Precious Woods Group (n=7) showing which measurements are more relevant at identifying each concession. d18O = 18O/16O, dD = D/H, dD nit = non-exchangeable D/H from cellulose, d13C = 13C/12C, d15N = 15N/14N and d34S = 34S/32S isotope ratios.

Figure 22. Principle Component Analysis of Okoume (Aucoumea klaineana) from the Precious Woods and CBG concessions. Data used to generate the PCA were the 18O/16O, D/H, non-exchangeable D/H from cellulose, 13C/12C, 15N/14N and 34S/32S isotope ratios.

Page 52 of 55

Figure 23. Inverse Distance Weighting (global) isoscapes of Aucoumea klaineana stable isotope data from samples collected in the CBG and Precious Wood concessions. 1. 18O/16O, 2. 13C/12C, 3. D/H, 4. Non- exchangeable D/H from cellulose, 5. 15N/14N, 6. 34S/32S.

Figure 24. 18O/16O (‰) of Aucoumea klaineana compared to soil type of the sampling location [Jones et al, 2013].

Figure 25. D/H (‰) of Aucoumea klaineana compared to soil type of the sampling location [Jones et al, 2013].

Figure 26. non-exchangeable D/H of cellulose (‰) of Aucoumea klaineana compared to soil type of the sampling location [Jones et al, 2013].

Figure 27. 13C/12C (‰) of Aucoumea klaineana compared to soil type of the sampling location [Jones et al, 2013].

Figure 28. 15N/14N (‰) of Aucoumea klaineana compared to soil type of the sampling location [Jones et al, 2013].

Figure 29. 34S/32S (‰) of Aucoumea klaineana compared to soil type of the sampling location [Jones et al, 2013].

Figure 30. Boxplots of 18O/16O stable isotope ratios (‰) of Dacryodes buettneri from the CBG (n=5) and Precious Woods Group (n=3) Concessions.

Figure 31. Boxplots of D/H stable isotope ratios (‰) of Dacryodes buettneri from the CBG (n=5) and Precious Woods Group (n=3) concessions.

Figure 32. Boxplots of nitrated D/H stable isotope ratios (‰) of Dacryodes buettneri from the CBG (n=5) and Precious Woods Group (n=3) concessions.

Figure 34. Boxplots of 15N/14N stable isotope ratios (‰) of Dacryodes buettneri from the CBG (n=5) and Precious Woods Group (n=3) concessions.

Figure 35. Boxplots of 34S/32S stable isotope ratios (‰) of Dacryodes buettneri from the CBG (n=5) and Precious Woods Group (n=3) concessions.

Figure 36. Principle components 1, 2 and 3 of the Dacryodes spp. data (variables used: 18O/16O, D/H, non- exchangeable D/H from cellulose, 13C/12C, 15N/14N, 34S/32S).

Figure 37. Inverse distance weighting (global) isoscapes of Dacryodes buettneri stable isotope data 1. 18O/16O, 2. 13C/12C, 3. D/H, 4. Non-exchangeable D/H from cellulose, 5. 15N/14N, 6. 34S/32S.

Figure 38. 18O/16O (‰) of Dacryodes buettneri compared to soil type of the sampling location [Jones et al, 2013].

Figure 39. D/H (‰) of Dacryodes buettneri compared to soil type of the sampling location [Jones et al, 2013].

Figure 40. Non-exchangeable D/H from cellulose (‰) of Dacryodes buettneri compared to soil type of the sampling location [Jones et al, 2013].

Figure 41. 13C/12C (‰) of Dacryodes buettneri compared to soil type of the sampling location [Jones et al, 2013].

Page 53 of 55

Figure 42. 15N/14N (‰) of Dacryodes buettneri compared to soil type of the sampling location [Jones et al, 2013].

Figure 43. 34S/32S (‰) of Dacryodes buettneri compared to soil type of the sampling location [Jones et al, 2013].

Figure 44. Relationship between the D/H and 18O/16O isotope ratios of Aucoumea klaineana and Dacryodes buettneri.

Figure 45. Ranges of oxygen isotope ratios relative to how many samples were collected per site for multiple different taxa.

Figure 46. Ranges of hydrogen isotope ratios relative to how many samples were collected per site for multiple different taxa.

Figure 47. Ranges of carbon isotope ratios relative to how many samples were collected per site for multiple different taxa.

Figure 48. Ranges of sulfur isotope ratios relative to how many samples were collected per site for multiple different taxa.

Glossary ANOVA: Analysis of variance is a collection of statistical models and their associated estimation procedures used to analyse the differences among group means in a sample.

Atom Percent (Atom %): the absolute number of atoms of a given isotope in 100 atoms of total element. For example, the nitrogen-15 content of air nitrogen is 0.3663 Atom %. For calculation, At% = [Rs / (Rs + 1) * 100] where Rs is the ratio of the light isotope to the heavy isotope of the sample.

Combustion: a high-temperature exothermic redox chemical reaction between a fuel and an oxidant, usually atmospheric oxygen, that produces oxidized, often gaseous products.

Continuous-Flow (CF): refers to an automated preparation device and mass spectrometer in which sample analysis is conducted in a continuous stream of helium carrier gas.

Delta Units (d or δ): are expressed in molecules per thousand, or “per mil”. For example, δ15NAir = 12 per mil means that the sample was analysed against a reference material and found to be 12 molecules per thousand more than Air - the accepted zero point for expression of nitrogen-15 in per mil notation. For calculation, d = [ (Rs / Rr) - 1 ] * 1000 where Rs is the ratio of the heavy isotope to the light isotope of the sample and Rr is the ratio of the heavy isotope to the light isotope of the reference.

Elemental Analyser (EA): an automated sample preparation instrument in which samples are automatically converted into pure gases for isotope ratio analysis. An elemental analyser contains the following elements: (i) furnace for combustion, reduction or pyrolysis of sample material; (ii) chemical traps for analyte gas purification; (iii) gas chromatography for time separation of these analyte gases.

Equilibration: a sample preparation technique whereby an aqueous sample in a sealed container is allowed to exchange and equilibrate molecules and isotopes with a headspace gas.

Fractionation: the enrichment or depletion of a stable isotope caused by natural or artificial processes.

FSC: The Forest Stewardship Council is an international non-profit, multi stakeholder organization established in 1993 that promotes responsible management of the world's forests.

Genus: a taxonomic rank used in the biological classification of living and fossil organisms, as well as viruses, in biology. In the hierarchy of biological classification, genus comes above species and below family. In binomial nomenclature, the genus name forms the first part of the binomial species name for each species within the genus

Page 54 of 55

Isotopes: can be defined as atoms whose nuclei contain the same number of protons but a differing number of neutrons.

Isotope Ratio Mass Spectrometry (IRMS): a mass spectrometer is an instrument for separation of molecules based upon their mass-to-charge ratio. In IRMS the mass spectrometer used separates isotopes of different mass within a magnetic field and precisely measures the ratio of two, or more, isotopes.

Isotope Ratio: the ratio of the minor isotope over the major isotope. For example, nitrogen in air contains 0.3663 Atom % nitrogen-15 and 99.6337 Atom % nitrogen-14, giving an isotope ratio of 0.3663 / 99.6337 = 0.003676466).

METI: Ministry of Economy, Trade and industry.

NASA: The National Aeronautics and Space Administration is an independent agency of the United States Federal Government responsible for the civilian space program, as well as aeronautics and aerospace research. NASA was established in 1958, succeeding the National Advisory Committee for Aeronautics.

Natural Abundance: the concentration of isotopes as found in nature.

NGO: Non-governmental organization.

Per Mil (‰): see Delta Units

Pyrolysis: a high temperature means of converting sample material to pure gas in the absence of oxygen.

Site: The area where samples have been collected. In this report ‘site’ functions as a categorical label that may refer to a concession, island, small sampling area or forest management unit.

Species: the basic unit of classification and a taxonomic rank of an organism, as well as a unit of biodiversity.

Stable Isotope: a non-radioactive isotope in which the number of protons and neutrons in the atomic nucleus is constant through time. Stable isotopes pose no known physiological risk.

Taxa: a group of one or more populations of an organism or organisms seen by taxonomists to form a unit.

United States Forestry Service: The United States Forest Service is an agency of the U.S. Department of Agriculture that administers the nation's 154 national forests and 20 national grasslands.

USAID: The United States Agency for International Development is an independent agency of the United States federal government that is primarily responsible for administering civilian foreign aid and development assistance.

Vienna-Canyon Diablo Troilite (V-CDT): an FeS meteorite used as the accepted zero point for expression of sulfur-34 in delta units. V-CDT is thought to have a sulfur-34 value close to the universal mean.

Vienna-Pee Dee Belemnite (V-PDB): a belemnite from the Cretaceous Pee Dee formation of South Carolina, US which is used as the accepted zero-point standard for expression of carbon and oxygen isotope abundance in delta units.

Vienna-Standard Mean Ocean Water (V-SMOW): the accepted zero-point standard for expression of hydrogen and oxygen isotopes of water samples in delta units.

WRI: The World Resources Institute is a global research non-profit organization that was established in 1982 with funding from the MacArthur Foundation under the leadership of James Gustave Speth.

Page 55 of 55