Screening Potential Bioenergy Production of Tree Species in Degraded and Marginal Land in the Tropics

Bioenergy can produce at least 25% of the global energy demand to combat climate change through reducing emissions in the energy sector. However, information on the bioenergy production potential of woody species and their suitability for silviculture on various soils in the humid tropics is limited. This review aims to identify tree species suitable for bioenergy production under these conditions. Data were compiled from 241 publications and nine freely available databases to assess environmental and silvicultural information on tropical tree species. Energy outputs were derived from the estimated productivity of the reviewed species and ranged from 0.2 to 24.0 Mg biomass ha−1 yr−1, 0.1 to 9.0 Mg bio-oil ha−1 yr−1, and 0.2 to 20.0 Mg sugar ha−1 yr−1, equivalent to an energy yield between 2 and 444 GJ ha−1 yr−1. As such, these bioenergy yields are within the range reported for the lignocellulosic biomass of energy crops cultivated in Europe, the USA, and Brazil. Our review identified some high-yielding species (e.g., Dyera polyphylla (Miq.) Steenis, Metroxylon sagu (Rottb.), Pongamia pinnata (L.)) and leguminous species that could be beneficial in mixed stands (e.g., Elaeis oleifera (Kunth) and Pongamia pinnata) or are suitable species to grow on wet or re-wetted peatland (Dyera polyphylla). However, there are limitations to cultivate woody bioenergy species on wet peatland. Sustainable methods for managing and harvesting forests, particularly on wet or re-wetted peatland, need to be developed.


Introduction
It is predicted that global energy demand will increase by 45% by 2040 compared to consumption in 2013, resulting in a total CO 2 emission increase of 40% [1].At the same time, achieving the 2 • C limit on global warming requires new policies to reduce the energy sector's CO 2 emissions by replacing traditional and fossil fuels with renewable energies [1][2][3].Bioenergy, energy produced from biological sources, is one such renewable energy [2,3].Globally, bioenergy has the potential to produce 100-400 exajoules (EJ) yr −1 [2,4], which is equivalent to 25%-100% of the total energy consumed in 2014 [5].Despite such enormous potential, in 2014, only about 4% of the electricity and heat consumed were generated from biofuels, while traditional use of biomass (e.g., cooking) represented 9% of the energy consumed [5].Traditional use of biomass, although common in developing countries, remains inefficient and hazardous to health; bioenergy could provide clean and affordable energy to meet increasing demands in these countries [1,2,6].In the tropics, oil palm (Elaeis guineenis (Jacq.))dominates biofuel production from tree species [7,8].However, in comparison to forests, oil palm monoculture results in a loss of ecosystem functions [8].This is less severe in mixed systems [9]; however, the advantages of mixed systems are countered by a reduction in palm oil yields due to the reduced number of stems per hectare.Diversifying or replacing palm oil plantations by including other oil crops (e.g., Pangamia pinnata) may offset oil yield loss.This has yet to be assessed.
The aim of this study was to identify tropical tree species that could produce biological resources for bioenergy production and are able to grow on various type of soils.The potential biofuel and energy yields were estimated by assessing yields based on silvicultural information (e.g., stem density) and productivity (e.g., biomass per ha and year), which were converted into energetic values (e.g., GJ ha −1 yr −1 ).Due to the huge body of literature on species used and recommended for bioenergy production in the tropics, this study specifically focused on tree species for bioenergy production [10][11][12][13].We therefore excluded bamboo and other non-woody species to avoid repetition of recent results produced by Abel et al. [11], Darabant et al. [14], Pfister [15], and Wi et al. [16].

Materials and Methods
The aim of this narrative review [17] was to identify tree species suitable for bioenergy systems in the tropics from literature [11][12][13][18][19][20] by combining their silvicultural information (Table 1) and potential energy yields per hectare per year (Table 2).A literature search using Google Scholar was conducted for silvicultural information using species names as keywords.
The search produced 241 documents and 9 freely available databases (Table S1).These were used to assess the following aspects of woody bioenergy crops: (i) botanical information (e.g., species and origin, synonyms, common name, typical use), (ii) ecological settings (e.g., temperature, mean annual precipitation, soil properties), and (iii) cropping and yields (e.g., stem density, biomass yield, bio-oil yield) (Table S2).Data extracted from original resources were taken directly from the publication; thus, our dataset represents original information without any conversion into a single system (e.g., FAO soil classification).Extracted soil pH values were mostly (i.e., 93%, Table S2) published without further clarification on solutions used (e.g., H 2 O, KCl, CaCl 2 ) or salt concentration, which affects the comparability of pH values [21,22].Thus, due to the lower accuracy of pH values and ranges presented, this review can provide only approximate information on the soil pH values tolerated.Yield data in mass or volume per unit area were used as presented in surveys or calculations, based on single tree productivity (e.g., dry biomass, fruit yield, oil content) and stand density per unit area (Table S2).Conversion factors used to derive energy yields (GJ ha −1 yr −1 ) were: (i) carbon density of 0.5 in dry mass of wood [23], (ii) energy of 37 MJ stored per kg carbon or 19 MJ per kg dry biomass [3], (iii) a bio-oil:biodiesel conversion rate of 90% (adapted from values published by Meher et al. [24], (iv) biodiesel density of 0.9 g cm −3 [24,25], (v) energy of 33 MJ stored per liter of biodiesel [24,25], (vi) a sugar:bioethanol conversion rate of 51% [26], (vii) bioethanol density of 0.8 g cm −3 [25], and (vii) energy of 21 MJ per liter of bioethanol [25].
Table 1.Potential bioenergy species that tolerate unfavorable soil conditions.Soil conditions that are most relevant for species selection are presented.Potential ecological adaptations are also shown to inform about tolerances toward to for instance droughts and flooding.Primary data and their corresponding references are shown in Table S2.  ) and energy values (GJ ha −1 yr −1 ).A 'Yes' indicates a promising species, but due to a lack of information in the literature, yield could not be estimated.Primary data and their corresponding references are shown in Table S2.

Species Biomass Bio-Oil and Biodiesel Sugar or Starch and Bioethanol
Mg ha

Results
Although numerous woody species are suitable for forest-based bioenergy systems in humid tropical regions, the estimation of potential bioenergy yields per unit of area (i.e., GJ ha −1 yr −1 ) was limited to 33 species due to the scarcity of silvicultural and biorefinery data (Tables 1 and 2).This study provides species-specific information on environments preferred by each species, silvicultural information (e.g., stem density per hectare), and yield data (Mg dry biomass ha −1 yr −1 ).About 50% of the species (n = 16) are adapted to mineral soils and able to tolerate acidic and nutrient-poor soils (e.g., eroded Acrisols) and droughts (Table 1).Trees that can tolerate drought include Aleuritis moluccana (L.), Calophyllum inophyllum (L.), and Pongamia pinnata.Although their cropping on terrestrial soils potentially produces high yields, such yields will be reduced by flooding and wet soil conditions.In addition, soil wetness, soil acidity, and low nutrient status may also limit plant productivity [27].In particular, biological nitrogen fixation by leguminous species that have been widely used to rehabilitate degraded land (e.g., Calophyllum inophyllum, Gliricidia sepium (Jacq.)) is drastically reduced in acidic soils.Based on tree productivity data, and information on their silvicultural recommendations, species suitable for growth on mineral soils and (re)-wetted peatland (Tables 1 and 2) can potentially produce between 0.2 and 24.0 Mg biomass ha −1 yr −1 , 0.1 and 9.0 Mg bio-oil ha −1 yr −1 , and between 0.2 and approximately 20.0 Mg sugar ha −1 yr −1 , which is equal to an energy yield between 2 and 444 GJ ha −1 yr −1 (Tables 2 and S2).
Seventeen species are potentially suitable for bioenergy activities on wet land and land which is regularly flooded (Table 1).Three tree species tolerate brackish environments, namely Cerbera manghas (L.), Nypa fructicans (Wurmb.),and Melaleuca cajuputi (Powell).The energy yield potential of these species ranges between 71 and 295 GJ ha −1 yr −1 (no data for Melaleuca cajuputi, Table 2).Calamus caesius (Blume) and Symphonia globulifera (L.f.) are adapted to wet soils rich in organic matter, while Combretocarpus rotundatus (Miq.)Danser, Dyera polyphylla, and Palaquium ridleyi (King & Gamble) can grow on permanently wet organic soils (i.e., peatland).Although peatland species produce raw material for bioenergy activities, data on productivity and energy yields are rarely reported, with productivity data found only for Dyera polyphylla (Table 2).The remaining nine tree species presented in Table 2 tolerate flooding and produce biomass, bio-oil, and sugar.Again, although information found on yields and productivity are minimal, the estimated energy output of some species may be too low for bioenergy activities (e.g., Euterpe oleracea (Mart.),Fleroya ledermannii (K.Krause), Spondias mombin (L.)), while the estimated productivity of Pentadesma butyracea Sabine and Sesbania bispinosa (Jacq.)seems to be promising for bioenergy activities (Table 2).Other species in this group are promising bioenergy crop candidates, but information on their productivity is not readily available (Tables 2 and S2).

Discussion
The species presented that tolerate acidic soils and droughts are known and often used to produce raw material for bioenergy in tropical countries [12,13,28].However, initiatives that aim to produce bioenergy require silvicultural information and yield data.The information presented here can be used to assess the economic feasibility of bioenergy projects and cropping system types [29][30][31].Silvicultural and yield data are scarce for tropical tree species adapted to permanently wet and regularly flooded environments, but such data are required to develop feasible bioenergy strategies for wetlands (e.g., peatland).Two reasons could explain this knowledge gap: (i) limited interest in most of these tree species, except for sugar-and starch-producing palm trees (Metroxylon sagu, Nypa fructicans) and (ii) a lack of machinery for harvesting [32].To avoid competition between food production and the production of raw materials for bioenergy, non-food crops should be cultivated on less productive land (e.g., eroded soil) [3,6,28].The simplest approach to rehabilitating eroded land is the establishment of plantations [33].Optimizing initial plant growth on eroded land for biomass production may require the application of fertilizer, which can cause the emission of N 2 O [3,34].A less-assessed, but promising, way to reduce the amount of N-fertilizer is to mix non-leguminous and leguminous crops (e.g., Elaeis oleifera (Kunth) Cortés and Pongamia pinnata).Rehabilitation may require initial site preparation by planting species that can shade out weeds, fix nitrogen, and improve soil organic matter [33].Trees suitable for site preparation are fast-growing, nitrogen-fixing species, e.g., Calliandra calothyrsus (Meisn.),Gliricidia sepium, and Zapoteca tetragona (Willd.).The cultivation of non-native tree species risks invasive competition [35][36][37].Thus, introducing for example species native to Africa (i.e., Croton megalocarpus (Hutch)) and America (i.e., Spondias mombin) to Southeast Asia and could have negative impacts on biodiversity and environmental services.
The rehabilitation of wet land requires selection of species that can tolerate wet soils and are adapted to natural conditions of peat swamp forests (e.g., Dyera polyphylla) [32], yet there is limited information available on suitable trees for peat-swamp rehabilitation activities.In this study, bioenergy yields are compared to those of palm oil trees (Elaeis guineenis), which produce 3-6 Mg bio-oil ha −1 yr −1 [38][39][40], equivalent to an energy output of 90-194 GJ ha −1 .Most of the assessed species have the potential to produce raw material (Palaquium ridleyi, Sandoricum koetjape (Burm.f.)) generating the same level of energy.For some species, very high yields have been reported (e.g., Dyera polyphylla, Metroxylon sagu, Pongamia pinnata) [19,41,42], potentially far above yields that are possible on degraded land.Other species with an estimated energy output of <90 GJ ha −1 yr −1 (i.e., the lowest energy output estimated for Elaeis oleifera) might not be feasible for bioenergy activities in tropical countries.

Conclusions
Tree species adapted to tropical wetlands and peatlands are potentially useful for bioenergy production, but published data are available only for a small number species.The estimated bioenergy yields of the reviewed woody species are in the range reported for lignocellulosic biomass of energy crops cultivated in Europe, the USA, and Brazil (110-370 GJ ha −1 yr −1 ) [2,3].However, the values and coefficients used to estimate energy yields per unit area may fail to reflect the real variability of caloric values of biomass from various species [43,44].Thus, this study provides initial estimations, which should be verified by experiments to test the impact of silviculture and biorefinery methods on energy yields.

Species that tolerate permanently wet and waterlogged or temporarily flooded soils
-/-no data available.

Table 2 .
Potential annual biomass, bio-oil, sugar (Su), and starch (St) productivity in Mg ha −1 yr −1 of species used/potentially suitable for forest-based bioenergy production in tropical regions.Biomass data were also converted into volumetric values (mL ha −1 yr −1

Species that tolerate continuously wet and waterlogged or temporarily flooded soils
-/-no data available.