Global Bioenergy Potential from High-Lignin Agricultural Residue

Global bioenergy potential from high-lignin agricultural residue

Venugopal Mendua, Tom Shearina, J. Elliott Campbell, Jr.b, Jozsef Storka, Jungho Jaec, Mark Crockerd, George Huberc, and Seth DeBolta,1

aDepartment of Horticulture and dCenter for Applied Energy Research, University of Kentucky, Lexington, KY 40546; bSierra Nevada Research Institute and School of Engineering, University of California, Merced, CA 95343; and cDepartment of Chemical Engineering, University of Massachusetts, Amherst, MA 01003-9303

Edited by Chris R. Somerville, University of California, Berkeley, CA, and approved January 17, 2012 (received for review August 4, 2011) Almost one-quarter of the world’s population has basic energy Despite obvious benefits of using agricultural residues as a waste needs that are not being met. Efforts to increase renewable en- stream (10), removal of some types of agricultural residues raises ergy resources in developing countries where per capita energy concerns (4). For instance, repeated harvesting of total above- availability is low are needed. Herein, we examine integrated dual ground biomass from annual cereal crops will ultimately reduce use farming for sustained food security and agro-bioenergy de- soil organic matter, leading to long-term loss of soil fertility and velopment. Many nonedible crop residues are used for animal feed accelerate CO2 emissions (11). However, an example of removing or reincorporated into the soil to maintain fertility. By contrast, partial residues has been demonstrated for rice (Oryza sativa) drupe endocarp biomass represents a high-lignin feedstock that is grain husks in India (12), which are gasified in small-scale eco- a waste stream from food crops, such as coconut (Cocos nucifera) friendly units to produce electricity for users spending approxi- shell, which is nonedible, not of use for livestock feed, and not mately $2 a month for energy (12). Such a model for renewable reintegrated into soil in an agricultural setting. Because of high- energy could serve globally as an inexpensive decentralized energy lignin content, endocarp biomass has optimal energy-to-weight mechanism. In looking for parallel scenarios, environmental fac- returns, applicable to small-scale gasification for bioelectricity. Us- tors such as temperature, rainfall, and altitude dictate the type of ing spatial datasets for 12 principal drupe commodity groups that crop produced in a given ecozone. Hence, identification of source have notable endocarp byproduct, we examine both their poten- feedstocks suitable for dual-use cropping and that are available in tial energy contribution by decentralized gasification and relation- regions with energy scarcity is needed. ship to regions of energy poverty. Globally, between 24 million An existing dual-use feedstock that is underused is endocarp and 31 million tons of drupe endocarp biomass is available per tissues from horticultural fruit crops, particularly the drupes. The year, primarily driven by coconut production. Endocarp biomass endocarp of a drupe fruit is the hardened inedible portion of the used in small-scale decentralized gasification systems (15–40% ef- fruit that encases the seed and is discarded, whereas the flesh ficiency) could contribute to the total energy requirement of sev- (mesocarp) is edible. The hardened drupe endocarp represents eral countries, the highest being Sri Lanka (8–30%) followed by the highest lignin content of any woody biomass source produced Philippines (7–25%), Indonesia (4–13%), and India (1–3%). While in appreciable amounts, commonly up to 50% wt/wt (13, 14). As representing a modest gain in global energy resources, mitigating a biofuel, lignin has much higher energy density (approximately energy poverty via decentralized renewable energy sources is pro- double) than cellulose (15, 16). Collectively, these crops repre- posed for rural communities in developing countries, where the sent some of the most abundant horticultural fruit crops in the greatest disparity between societal allowances exist. world. Herein, we explore the geographical distribution of several perennial endocarp commodity crops—coconut (Cocos ccording to the World Health Organization, United Nations nucifera), mango (Mangifera indica), olive (Olea europaea), wal- ADevelopment Program (UNDP), 1.5 billion people, repre- nut (Juglans spp), pistachio (Pistacia vera), cherry (Prunus cera- senting approximately one-quarter of the world’s population, lack sus, P. avium), peach (P. persica), plum (P. domestica, P. salicina), — basic access to electricity (1). Two billion people need modern apricot (P. armeniaca), and almond (P. dulcis) and quantify energy services to meet the UNDP millennium development goals their potential for decentralized bioelectricity production. We fi (1, 2). Lack of stable access to electricity and liquid transportation focus on the potential of endocarp biomass for energy speci - fi fuels disproportionally impacts undeveloped and developing cally and to nd the spatial relationship between the availability countries, where population density is high and access to resources of endocarp and the prevalence of energy poverty. ≈ is low. Strikingly, 2 billion people are reliant on solid fuels, such Results and Discussion as crude burning of biomass or coal, which is a primary cause of the brown cloud over South Asia (3). Health costs, predominantly Geospatial Examination of Global Yield and Production Zones for Drupe Endocarp Biomass. Herein, we estimated the potential global affecting woman and children, are attributed to the burning of yield of drupe endocarp biomass by using crop yield data derived solid fuels in poorly ventilated housing (1, 2). This motivation from published data (17, 18) and cross-referenced with Food and contrasts with developed countries, which are seeking renewable Agriculture Organization (FAO) world commodity yield data (19) energy from biomass as a means to reduce CO 2 emissions and for the year 2000 (Table 1 and Table S1). Several limitations exist ensure domestic energy security (4). Dedicated energy crops are in generating an accurate yield estimate that accounts for global being sought that produce high-yielding lignocellulosic biomass, such as Miscanthus (5). In undeveloped and developing countries, dedicated energy crops could displace food crops and confer im- – Author contributions: V.M., J.E.C., G.H., and S.D. designed research; V.M., T.S., J.S., J.J., balance in food and fuel security (6 8). Further, clearing forest and M.C. performed research; J.E.C. contributed new reagents/analytic tools; V.M., T.S., land for bioenergy crops (directly or indirectly) will enhance CO2 and S.D. analyzed data; and V.M., T.S., M.C., G.H., and S.D. wrote the paper. emissions resulting from land-use change (9) unless crops are The authors declare no conflict of interest. grown on marginal (degraded) agricultural land or use waste This article is a PNAS Direct Submission. biomass. Therefore, dual-use cropping scenarios may provide 1To whom correspondence should be addressed. E-mail: [email protected]. opportunities to improve agricultural productivity by producing This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. bioenergy from agricultural waste while maintaining food security. 1073/pnas.1112757109/-/DCSupplemental.

4014–4019 | PNAS | March 6, 2012 | vol. 109 | no. 10 www.pnas.org/cgi/doi/10.1073/pnas.1112757109 Downloaded by guest on September 26, 2021 Drupe endocarp yeild (metric tonnes per hectare) Fig. 1. Global projection of drupe endocarp biomass 0.003 production. Here, we show yield of all crops x percentage 0.003 - 0.006 0.006 - 0.009 endocarp biomass as an estimate for the drupe endocarp 0.009 - 0.012 biomass. Grid density for this Earthwide composite map 0.012 - 0.016 of drupe endocarp production represents 5 min of lati- 0.016 - 0.022 0.022 - 0.028 tude by 5 min of longitude per hectare per year (as in- 0.028 - 0.038 dicated in scale bar). The geospatial grid display uses 10 0.038 - 0.067 quantiles, providing a deﬁned range linked to a color 0.067 - 0.820 coded scale among producing areas. The units are derived from the sum of all commodities as the proportion of each grid cell.

reporting inadequacies and variation. For instance, production Hence, as a biofuel, endocarp biomass is comparable to coal − metrics do not take into account regional, national, and in- based on the heating value (18.0–25.5 MJ·kg 1) (13, 20). The ternational trade representing movement of a commodity away average drupe endocarp energy content is higher than that of from where it is produced. Noting these limitations, we quantified dedicated bioenergy crops [reed canarygrass (Phalaris arundina- the amount of endocarp biomass via the ratio of drupe to flesh and cea), 17.7 MJ/kg; switchgrass (Panicum virgatum), 18.5 MJ/kg; energy content for individual commodities (Table S1). Each crop and hybrid poplar (Populus sp.) 19.3 MJ/kg] (21, 22). It is possible total and the sum total were then examined based on annual en- to examine the conversion of endocarp biomass to a fungible docarp production. The total endocarp yield was 2.4 × 107 tons liquid fuel by catalytic pyrolysis (23). Via pyrolysis, a 60–70 (mean = (Table 1, Table S1, and Dataset S1), which was consistent with the 65) energy % yield would be generated for bio-oil production value of 3.1 × 107 tons based on FAO sourced data (29% varia- (23). To project bioelectricity potential, yield was multiplied by an tion). To visualize these data we used 10 quantiles providing a endocarp to fruit ratio (Table 1, Table S1, and Dataset S1) and defined range linked to a color-coded scale to illustrate contrast the energy content (presented as a range: low and high) for each among producing areas (Fig. 1). Results demonstrate the greatest individual feedstock. Gasification units differ in their efficiency density of drupe endocarp production occurred in developing based on scale and type of units, for instance, large scale (>500 countries throughout South Asia, and broad lower-density distri- kWh) routinely get up to 50% efficiency (24), whereas a small- bution across Southern and Northern Europe and the Middle East scale gasification system that does not include cogeneration has (Fig. 1 and Fig. S1). Only isolated production zones occurred in ≈20–40% (herein a range of 15–40% was projected) conversion the United States of America (USA), Africa, China, Australia, efficiency (24–26). It is noteworthy that if cogeneration was used Central America, and South America. as combined heat and power, gasification efficiency can be up- Energy content of the drupe endocarp tissue under in- wards of 80% (25, 26). Results demonstrate that exploiting high- vestigation (Table 1 and Table S1) ranged between 16.2 (lowest) lignin feedstocks from existing production systems has a global − − and 22.8 (highest) MJ·kg 1 with an average of 19.5 MJ·kg 1. bioenergy production potential of 4.1–5.2 × 108 GJ (gridded) to

Table 1. Global yield and bioenergy potential derived from drupe endocarp based feedstocks Equivalent conversions

MWh, GJ × 0.28 × efﬁciency

Endocarp Energy, MJ/kg GJ, tons × MJ/kg 15% efﬁciency 40% efﬁciency Crop product, Name Crop tons % Tons Low High Low High Low High Low High

Almond Nut 1.5E+06 60 9.2E+05 18.2 19.5 1.7E+07 1.8E+07 7.0E+05 7.5E+05 1.9E+06 2.0E+06 SCIENCE Apricot Fruit 2.5E+06 10 2.5E+05 17.8 21.1 4.5E+06 5.3E+06 1.9E+05 2.2E+05 5.0E+05 5.9E+05 SUSTAINABILITY Cherry Fruit 1.7E+06 8 1.4E+05 18.9 22.5 2.6E+06 3.1E+06 1.1E+05 1.3E+05 2.9E+05 3.4E+05 Coconut Oil/nut 5.2E+07 25 1.3E+07 17.0 22.8 2.2E+08 3.0E+08 9.3E+06 1.2E+07 2.5E+07 3.3E+07 Mango Fruit 2.5E+07 16 4.0E+06 17.5 21.5 7.0E+07 8.6E+07 2.9E+06 3.6E+06 7.8E+06 9.5E+06 Olive Oil 1.4E+07 14 1.9E+06 16.2 21.8 3.1E+07 4.2E+07 1.3E+06 1.7E+06 3.4E+06 4.6E+06 Peach Fruit 1.3E+07 15 1.9E+06 19.4 20.5 3.8E+07 4.0E+07 1.6E+06 1.7E+06 4.2E+06 4.4E+06 Pistachio Nut 4.4E+05 57 2.5E+05 16.6 19.3 4.1E+06 4.8E+06 1.7E+05 2.0E+05 4.6E+05 5.3E+05 Plum Fruit 8.8E+06 4 3.5E+05 17.8 21.1 6.3E+06 7.4E+06 2.6E+05 3.1E+05 7.0E+05 8.2E+05 Sour cherry Fruit 1.0E+06 8 8.3E+04 18.9 22.5 1.6E+06 1.9E+06 6.5E+04 7.8E+04 1.7E+05 2.1E+05 Stone fruit Fruit 3.5E+05 10 3.5E+04 17.8 21.1 6.3E+05 7.5E+05 2.6E+04 3.1E+04 7.0E+04 8.3E+04 Walnut Nut 1.2E+06 50 6.2E+05 17.9 19.7 1.1E+07 1.2E+07 4.6E+05 5.1E+05 1.2E+06 1.4E+06 Totals 1.2E+08 2.4E+07 4.1E+08 5.2E+08 1.7E+07 2.2E+07 4.5E+07 5.8E+07

The global total production values were obtained from the Food and Agriculture Organization (19), and the endocarp and equivalent energy conversions were calculated. The percentage of endocarp biomass relative to commodity yield and energy content were obtained from the literature (summarized in Table S1). GJ were calculated based on the average (low and high ends) calorific value for drupe endocarp biomass multiplied by the yield of endocarp in tons. Electricity was calculated by using a range of conversion efficiency (15–40%) considering the efficiency variation of different bioelectricity units. Apricot, plum, and other stone fruits: Average of the other seven endocarp energy contents were used.

Mendu et al. PNAS | March 6, 2012 | vol. 109 | no. 10 | 4015 Downloaded by guest on September 26, 2021 5.5–6.7 × 108 GJ per annum (FAO). These energy yield estimates (852.9 KWh/year) for the year 2000 (27–29). Broadly, energy corresponded to the production of between 17 and 58 (gridded) consumption among developed countries is ≈9 times greater and 23 and 75 (FAO) million MWh of electricity per annum than the developing countries, although this disparity can be based on a conversion efficiency (15–40%) of the bioenergy into larger in specific regions within developing countries (27–29). actual electricity. In contrast, a theoretical projection of a pyrol- Strikingly, per capita energy consumption of the USA (13,656.10 ysis-based biofuel resulted in an approximate yield of between 67 KWh/year) is 34 times higher than the per capita consumption of and 85 (gridded) and 91 and 110 (FAO) million barrels of oil India (402 KWh/year) and Indonesia (400.4 KWh/year) for the energy equivalent per year (Table S2). year 2000 (27, 28), linked to insufficient and/or lack of power Crops that afforded the largest yield of endocarp were coconut supply to the rural areas in these countries (30, 31). Indeed, the (1.31 × 107 tons), followed by mango (3.99 × 106 tons) (Table 1, Indonesian government has initiated the “Village Independent Dataset S1, and Fig. 2). Combined, coconut and mango production Program” to secure the supply of electricity for lighting, cooking, generated 72% of total global drupe endocarp. Spatial analysis and other productive activities in remote areas and isolated is- allowed a visual perspective of the relationships between avail- lands (31). This program aims at meeting 60% of the total energy ability of endocarp and energy poverty. The maximum density of demand from renewable energy, including biomass. Indonesian coconut endocarp production was observed across developing rural areas and isolated islands are rich in biomass waste gen- countries in South Asia, focused in the south of the Indian sub- erated from oil palm and coconut plantations (31). The amount continent, Sri Lanka, Bangladesh, Philippines, Myanmar, Thai- of bioelectricity from drupe endocarp utilization could create land, Malaysia, Vietnam, Laos, and Indonesia (Fig. S1 and Dataset a framework to sustainably increase energy consumption in de- S1). Here, based on land area, Indonesia and India displayed the veloping countries because of it being a dual-use agricultural greatest potential for production. When visualizing the world residue that will not divert land used for food production. It gridded for coconut production beyond South Asia, it was evident should be recognized that in terms of renewable bioenergy that minor production potential existed in tropical Africa and sources that maintain energy, environmental, and food security, Central and South America, but with more discreet and spatially there is a cost to all potential solutions (4). constrained production zones. Similarly, the densest production of The majority of drupe endocarp was produced in the Southeast mango was identified in South Asia, but production maxima were Asian region spanning India, Indonesia, China, Philippines, Sri documented across equatorial tropical zones in Central America, Lanka, and Thailand (Fig. 1 and Dataset S3). India is the largest ≈ North Asia, and Africa (Fig. S1). Accounting for the other 28% of producer of drupe endocarp, sharing 17.3% of the world total, global endocarp waste stream was a scattered range of commodi- followed by Indonesia (17.0%) and Philippines (15.4%). To- ties. For instance, in Poland, Turkey, Japan, Spain, Italy, and gether India, Indonesia, and the Philippines produced 50% of the Greece, production was linked to stonefruit (peach, cherry, plum, world endocarp production for the year 2000. To establish a spa- nectarines), olive, and nut (pistachio, almond) (Fig. S1). African tial relationship between the endocarp production and energy production was also scattered, being most concentrated in Tan- poverty, we examined global per capita electricity consumption zania in the east and ranging from Senegal to Nigeria in the west (32) at country-level resolution (Fig. 3 and Dataset S4). Drupe (mango and coconut). A noteworthy cell density was also observed endocarp production was coincident with regions where per in Morocco (plums, olives, and nuts) and California’s central valley capita electricity consumption was low, particularly in South Asia (Figs. 1 and 3 and Dataset S4). Although we could visualize a (walnut, almond, pistachio, olive, plum, peach, cherry, and apri- fi cots). Central and South American production was identified in signi cant concentration of endocarp production and low per Mexico, Guatemala, Honduras, Nicaragua, Costa Rica, Chile, and capita energy availability in South Asia, we were unable to allo- Argentina (coconut, olive, mango, plum, walnut, and peach) cate data to individual countries based on spatial data alone. To (Fig. S1). further examine the capacity for endocarp biomass to mitigate energy poverty, the relationship between country level annual Modeling Bioenergy Based on Geospatial Data Shows Marked Potential bioelectricity production from endocarp biomass (Dataset S3) in Developing Countries in Southeast Asia. Globally, a wide gap and electricity consumption (Dataset S5) was compared (Fig. 4 exists in the per capita consumption of electricity (Dataset S2) and Table S3) for the year 2000. Percentage contribution of drupe between developed (7,620.6 KWh/year) and developing countries endocarp-based bioelectricity to the total national consumption was calculated by using bioelectricity production from endocarp with 15–40% efficiency and the total electricity consumption data (27–29) of selected countries (Table S3). Data document that Pistachios Plums Misc. Almonds Apricots – 1% 2% 1% 4% 1% endocarp-based bioelectricity could provide 8 30% of the total Peaches Walnuts consumption in Sri Lanka and 17–25% in the Philippines because 8% 3% of their low energy consumption and substantial production Olives 8% (coconut driven). India and Indonesia produced the greatest amount of endocarp biomass, which could contribute 0.8–3% and 4–13% of their energy consumption, respectively. By contrast, developed countries such as Italy and the USA could obtain, respectively, 0.16–0.51% and 0.02–0.05% from drupe endocarp Mangoes 17% biomass. We note that these data encompassed a national de- Coconuts mographic of energy consumption, which does not accurately 55% reflect the disparity of resource allocation within a given country. The potential for small-scale bioelectricity generation from biomass should be targeted toward rural communities where severe energy poverty exists (12), where it is most needed, and where the greatest disparity between societal allowances exists (33).

Fig. 2. Percent contribution of individual endocarp commodities relative to Economic, Technical, Social, and Environmental Challenges in Decentral- Economic. the global drupe endocarp production. Peach and nectarine were combined. ized Endocarp Utilization. Cost of developing decentralized Misc., cherries, sour cherries, and other stonefruits. Data derived from yields energy infrastructure requires capital investment. Although inex- as of year 2000 (17, 18). pensive relative to other infrastructure options, the cost barrier of

4016 | www.pnas.org/cgi/doi/10.1073/pnas.1112757109 Mendu et al. Downloaded by guest on September 26, 2021 Electricity consumption (KWh per capita) Fig. 3. Global per capita electricity consumption. 0.12 - 370.47 3 Per capita consumption values were determined by 370.47 - 1.11 x 10 33 dividing the annual electricity consumption by the 1.11 x 10 - 1.91 x 10 1.91 x 10 33 - 2.98 x 10 population of individual countries, and these data 2.98 x 10 33 - 4.49 x 10 were globally expressed spatially at a country-level 33 4.49 x 10 - 6.31 x 10 resolution. The data represents annual electricity 6.31 x 10 33 - 9.27 x 10 9.27 x 10 34 - 1.45 x 10 consumption in kWh per person. The geospatial grid 1.45 x 10 44 - 2.46 x 10 display uses 10 quantiles, providing a deﬁned range 44 2.46 x 10 - 5.29 x 10 linked to a color coded scale.

establishing gasification in poor rural communities in developing Technical. Biomass is a complex conglomerate of biopolymers countries cannot be implemented without outside funds from (polysaccharides, lignin, and protein). Ash released from biomass government or organizations (34). For instance, establishing small- (or coal) poses problems because of its deposition on the heat scale industry capable of producing a 100 kW of power costs ≈1.5 transfer surfaces of boilers and the internal surfaces of gasifiers million rupees (India) (35), which may be cost prohibitive. Several (36). Hence, feedstocks with low ash content are preferred or fi developmental programs exist to alleviate this cost, such as Global regular cleaning is needed to maintain ef ciency. The ash content Environment Facility by UNDP, which together with support from of drupe endocarp is much lower than other agricultural byproducts fi such as rice straw and cotton stalks (37), which adds to the attrac- local government has established small-scale gasi ers in several fi rural communities (34). For instance, a 250 kWh capacity gasifier tiveness of drupe endocarp biomass for gasi cation to electricity. provides electricity to the Boregunte village, Karnataka, India. The However, in addition to ash, tar condensation arises from biomass gasification (38, 39), which imposes additional removal costs. same program has now established a 10,000 kWh capacity plant Moreover, removing the volatile tars by using water scrubbers and that provides bioelectricity for four villages in the same region (34). wet electrostatic precipitators results in the production of waste- Here, community management of biomass tree (coppicing) plan- fi water contaminated with organic and inorganic compounds (40). A tation and the gasi cation plant provides decentralized energy in- proper treatment of the wastewater, using physical, chemical, or frastructure and local employment. Similarly, the company Husk physiochemical methods, is necessary before releasing it to the Power Systems, which follows the same model, established 50 environment (40). To overcome these challenges, an indirect gas- fi small-scale gasi ers in rural Bihar, India, by using rice husks as the ification system has been developed to avoid quenching of the feedstock (12) and are projected to establish 2,014 small-scale units product gas and associated efficiency losses as well as tar formation by 2014. These decentralized solutions to energy poverty provide (41). Optimizing the biomass gasifiers with higher efficiencies and installation, expertise, and maintenance for gasification systems. lower environmental pollution is ultimately necessary for the sus- Overall, although some solutions are emerging for implementation tainable production of bioelectricity. It should be noted that critical of this technology, economic hurdles for scaling are prominent. assessment of any energy source has, to date, revealed inherent infrastructure and environmental costs. Therefore, social benefits of alleviating energy poverty must be weighed against such tech-

30 nical costs to fully gauge sustainability. Longer term, an ideal characteristic of an energy product is the ability for transportation and storage, such as a liquid transportation fuel and bioproducts 25 such as furfural, plastic ﬁller, abrasives, catalysts or resins (42), but high c.v. + 40% conv. effic. these products remain under development. high c.v. + 15% conv. effic. Socioenvironmental. Harvest, collection, transport, and storage lo- low c.v. + 40% conv. effic. 20 gistics will be prominent challenges in endocarp utilization in low c.v. + 15% conv. effic. both developing and developed countries. Primary challenges are SCIENCE 15 linked to the seasonality of harvest times. Coconut, the most abundant commodity examined, is produced year round (43), but SUSTAINABILITY others were deﬁned by a seasonal production window (for ex- 10 ample mango, peach, or olive) (44, 45). Year-round harvest is ideal, but a degree of collection and storage infrastructure will

Contribution to national requirement (%) national requirement to Contribution 5 likely be necessary regardless of the bioenergy system. Collection of drupe endocarp biomass will arise primarily from processing plants or urban consumption. The focus of geospatial analysis 0 revealed that developing countries were primarily producing USA China Brazil Italy Spain Thai. India Indo. Phill. Sri L. coconut and mango material (72% of global total). These com- Fig. 4. The potential for endocarp biomass to meet country specific electricity modities are processed for products like coconut milk, coconut requirements. Examination of country-specific electricity consumption versus chips/flakes, or canned fruit, and the endocarp material is dis- production of drupe endocarp biomass was calculated via endocarp yield and carded and not exported. The alterative use is direct consump- consumption data for the year 2000. Electricity consumption data and bio- tion, in which case postconsumer recycling of endocarp biomass electricity production data are derived from Dataset S2 (27–29) and Table 1, respectively. Ten countries were examined for comparative energy portfolio becomes an additional logistic challenge. An exemplary program contribution. Note: Production-based analysis was technically unable to account was trialed in war time New York, where endocarp material was for postproduction import/export dynamics of individual commodities across collected daily via community participation for gas mask prepa- national or international boundaries. conv, conversion; c.v., calorificvalue. ration (46), indicating the possibility of recovery and recycling;

Mendu et al. PNAS | March 6, 2012 | vol. 109 | no. 10 | 4017 Downloaded by guest on September 26, 2021 however, no data exists for developing countries where municipal longitude, country, or missing data. In general, where reliable subnational infrastructure is unavailable. At the broader scale, we consider level agricultural statistics are available, the maps provide a better picture of the life cycle of the harvested feedstock (agricultural phase) in the where production actually occurs than in cases where an average is dis- two ways: Either the waste stream from agriculture results in tributed over a large geographic area (see Monfreda et al.; ref. 18). Data were analyzed as follows: For each commodity, we multiplied the a net zero agricultural phase or the life cycle costs for the agri- value of each cell of the area grid by the value of the corresponding cell of the − cultural phase (such as pruning, weed control, pesticide appli- yield grid to get an average production per hectare (tons.ha 1) for each cell. cation, and harvesting dating back to planting) would need to be This value, in turn was multiplied by an estimate of the percentage by halved. Given that this study looks at the current potential of weight of the endocarp by-product for the commodity in question (Fig. 1 existing crops, all of which are waste materials, we excluded the and Fig. S1). Tabular estimates were derived by multiplying each cell by its impact of the agricultural phase, acknowledging that its link to approximate area to get a production per cell (as opposed to a per-hectare food production is necessary to meet these criteria. value) and then computing the appropriate sums. In many rural and urban communities in developing countries with persistent unemployment and poverty, access to the basic Estimation of Drupe Endocarp Yield. The production estimation was for the water, sewage, transport, trade, education, and health services are total crop. To obtain the total endocarp production on a per-hectare basis for each commodity, we multiplied this map by an estimate of the ratio (%) of the severely limited, let alone access to electricity (33). Prior esti- – endocarp to gross weight of the harvested commodity (Table 1). Finally, this mates have indicated that 17 25% of agricultural residues in the metric was multiplied by a map containing the area estimates for each grid developing countries are burned in fields or homes as a crude cell to obtain a quantification of the total endocarp output associated with biofuel (47, 48). Burning biomass fuels results in degraded at- that commodity per cell. Cell areas were approximated by the areas of 5-min mospheric quality because of the particulate and ash emissions (3, surface patches on a sphere with a radius of 6,371 km. 49) as well as greenhouse gas emission (50). If decentralized It is important to note that, as a matter of convenience, we treated NO-DATA bioelectricity could service the income threshold of developing as a value of zero. This approach is unproblematic in the many cases in which the countries, the displacement of these emissions would have con- absence of data simply reflects the practice of many statistical agencies of only fi siderable health benefits. Even modest electricity supply has been reporting crops grown in signi cant quantities or otherwise locally important. In someimportant cases,however, theabsenceofdata is a consequenceofitsbeing shown to afford enormous social implication for improving edu- unavailable as a matter of national policies or resource limitations (18). This cational opportunities and overall community development (51). unavailability of data being the case, there are likely to be some important stocks that do not appear on the map or in the total production estimates. To Conclusions have a comparison, the map estimates were cross-referenced with the FAO With approximately one-quarter of the global human population production data (19) for the individual commodities for the year 2000. lacking basic access to electricity (1) and no emission-free renewable energy source available, we propose that gasification Estimation of Total Bioenergy and Its Equivalent Electricity and Liquid Fuel conversion of endocarp feedstocks to electricity represents a Production Potential. Total endocarp production was calculated by multiplying decentralized component of the complex renewable energy future the endocarp percentage to the global drupe production for different com- that could alleviate energy poverty. As a dual-use farming sce- modities. The total endocarp production was multiplied by low and high energy contents of individual drupe endocarp tissue of different crops to obtain the nario for food and bioenergy production, this feedstock poses less energy values (low and high) (Table 1 and Table S1). Energy content was mul- risk of threatening food security, but because of these sustain- tiplied by a conversion factor (0.28) and efficiency factor (15–40%) to convert GJ ability constraints, has lower projected capacity relative to dedi- energy into equivalent MWh bioelectricity. Barrels of oil equivalents were cal- cated energy crops in developed countries that also are not culated by dividing the energy in GJ by a conversion factor of 6.1 (23). without potential hurdles (4–6). Primary impediments to implementation are the cost of establishing and maintaining gasifica- Data Formats and Software. The grids are available for download in two tion infrastructure in rural communities and the seasonality of different formats, NetCDF and text files formatted for easy import into ArcGIS supply logistics, which are similar to challenges facing bioenergy (ArcGIS). The NetCDF format has several advantages. First, the source maps agriculture in developed countries. It should be further noted that (levels 3 and 4) are only available in the NetCDF format. More importantly, all endocarp drupe biomass sources defined herein are derived the text format reports three decimal places of precision, which leads to some very small numbers being zeroed out and can have consequences for a de- from perennial horticultural cropping systems and, therefore, an rived product. For instance, it appears that only national level data were envisioned risk is posed by projections of regional water scarcity, available for coconut production in China. In the absence of additional in- climate change (52, 53), or by inadequate supply of nutrients (54). formation, this was evenly distributed across the entire area of cropped land in China, resulting in very small values in the percent of area maps. When Materials and Methods these are rounded to zero, the net result is that China’s coconut production— Data Acquisition and Analysis. The data on drupe to endocarp ratio and energy and China is among the world’s top 10 producers—disappears from an es- content of different drupe endocarp tissues from various plant species was timate derived by multiplying yield by cropped area. This zeroing out does obtained from the published papers and referenced accordingly. The data for not occur with the NetCDF datasets. The presence of very small values in global production estimates of cultivated area and yields were obtained for some of the derived maps can also lead to problems of numerical instability. the year 2000 (18) and electricity consumption (Dataset S4) derived from the The area and yield map were initially imported from the text files into Central Intelligence Agency (32). Per capita consumption and total electricity ArcGIS and analyzed by using the Map Algebra toolbox from Spatial Analyst. consumption for the year 2000 were also derived (27–29) (Datasets S2 and S5). When the abovementioned problems of rounding and other questions of Production estimates were based on a series of maps representing the numerical instability due to the presence of very small values emerged, we global distribution of crop areas and yields for the year 2000 (18). Each map is performed the analysis in R (r-project.org). The ncdf library was used to in the form of a grid, or “raster,” covering the entire earth, each cell repre- extract the data from the NetCDF-formatted files. The map algebra oper- senting 5 min of latitude by 5 min of longitude. The authors combined the ations, i.e., cell-wise multiplication and addition, can be done as vector best available agricultural census estimates with a global grid of cropped arithmetic in R. Computation of commodity totals and the overall endocarp lands (17) to produce a set of four grids for each of 175 important agricultural total amount to summing the values of the corresponding vector. Because commodities (18). One grid estimates the percentage of the cell given over to floating point arithmetic in R is done in double precision (ArcGIS represents the production of the crop. If an area is cropped multiple times a year, it is real valued grids in single precision), we avoided numerical problems with- counted multiple times in determining this ratio. For example, if 50% of a cell out having recourse to logs or expedients like trying a reordering of the is devoted to a particularly commodity and it is cropped three times a year, operands. Results were exported from R and imported into ArcGIS, which the value for the cell will be 1.5 (150%), i.e., it is possible for a cell value to was used to create the maps and derive the by-country estimates by using exceed 100%. A second grid contains an estimate of the cell’s gross yield per the Zonal Statistics tool from Spatial Analyst (ArcGIS). harvest in tons per hectare. The third and fourth grids in the set indicate the administrative level of the source data in the first and second grids. These ACKNOWLEDGMENTS. This work was funded by National Science Founda- data may be county, state, interpolated from within 2 degrees latitude and tion Emerging Frontiers in Research and Innovation Grant 0937657.

4018 | www.pnas.org/cgi/doi/10.1073/pnas.1112757109 Mendu et al. Downloaded by guest on September 26, 2021 1. Legros G, et al. (2009) The energy access situation in developing countries. A review fo- 29. International Energy Agency (2000) International energy statistics. Available at http:// cusing on the least developed countries and Sub-Saharan Africa. WHO-UNDP;http:// www.eia.gov/cfapps/ipdbproject/IEDIndex3.cfm?tid=2&pid=2&aid=2. Accessed Janu- content.undp.org/go/cms-service/stream/asset/?asset_id=2205620. Accessed January 2011. ary 2011. ’ 2. Gronewold N (2009) One-quarter of world s population lacks electricity. Sci Am 30. Modi V (2005) Improving electricity services in rural India. Working Paper No. 30. fi (Nov):4, http://www.scienti camerican.com/article.cfm?id=electricity-gap-developing- Working Papers Series. (The Earth Institute, Columbia University, New York). countries-energy-wood-charcoal. 31. Motoyama J, Nagai M, Monobe H, Niida K, Nakata N (2009) Preliminary Feasibility 3. Gustafsson Ö, et al. (2009) Brown clouds over South Asia: Biomass or fossil fuel Study on the Palm Oil Mill Wastes-Fired Power Generation Systems and CDM Project combustion? Science 323:495–498. for Rural Electrification in Sumatra, Indonesia (Engineer and Consulting Firms Assoc, 4. Tilman D, et al. (2009) Energy. Beneficial biofuels—the food, energy, and environ- Japan). ment trilemma. Science 325:270–271. 32. CIA (2009) The World Factbook. (Central Intell Agency, Washington, DC). Available at 5. Heaton EA, Dohleman FG, Long SP (2008) Meeting US biofuel goals with less land: The http://www.cia.gov/library/publications/the-world-factbook/rankorder/rankorderguide. potential of Miscanthus. Glob Change Biol 14:2000–2014. 6. Naylor RL, et al. (2007) The ripple effect: Biofuels, food security, and the environment. html. Accessed November 2011. Environment 49:30–43. 33. Coburn L, et al. (2008) Energy policy review of Indonesia. (Organisation for Economic 7. Battisti DS, Naylor RL (2009) Historical warnings of future food insecurity with un- Co-operation and Development/International Energy Agency, Paris). http://www.iea. precedented seasonal heat. Science 323:240–244. org/textbase/nppdf/free/2008/Indonesia2008.pdf. 8. Rajagopal D, Sexton SE, Roland-Holst D, Zilberman D (2007) Challenge of biofuel: 34. United Nations Development Programme (2008) Green power for rural areas in India. Filling the tank without emptying the stomach? Environ Res Lett 2:044004. http://content.undp.org/go/newsroom/2008/march/green-power-for-rural-areas-in- 9. Fargione J, Hill J, Tilman D, Polasky S, Hawthorne P (2008) Land clearing and the india.en. biofuel carbon debt. Science 319:1235–1238. 35. Pallav P (2009) Economic potential of biomass gasification projects under clean de- 10. Kim S, Dale BE (2004) Global potential bioethanol production from wasted crops and velopment mechanism in India. J Clean Prod 17:181–193. crop residues. Biomass Bioenergy 26:361–375. 36. Misra MK, Ragland KW, Baker AJ (1993) Wood ash composition as a function of 11. Lal R (2005) Soil erosion and carbon dynamics. Soil Tillage Res 81:137–142. furnace temperature. Biomass Bioenergy 4:103–116. 12. Bornstein D (January 10, 2011) A light in India. NY Times. Available at http://opinionator. 37. Jenkins BM, Ebeling JM (1985) Thermochemical properties of biomass fuels. Calif blogs.nytimes.com/2011/01/10/a-light-in-india. Agric 39:14–16. 13. Mendu V, et al. (2011) Identification and thermochemical analysis of high lignin 38. Balat M, Balat M, Kırtay E, Balat H (2009) Main routes for the thermo-conversion of feedstocks for biofuel and bio-chemical production. Biotech Biofuels 4:43. biomass into fuels and chemicals. Part 2: Gasification systems. Energy Convers Man- 14. Dardick CD, et al. (2010) Stone formation in peach fruit exhibits spatial coordination age 50:3158–3168. fl of the lignin and avonoid pathways and similarity to Arabidopsis dehiscence. BMC 39. Salam PA, Kumar S, Siriwardhana M (2010) Report on the Status of Biomass Gasifi- Biol 8:13. cation in Thailand and Cambodia. Prepared for: Energy Environment Partnership 15. Raveendran K, Ganesh A (1996) Heating value of biomass and biomass pyrolysis (EEP) (Mekong Region. Asian Inst Technol, Bangkok, Thailand). products. Fuel 75:1715–1720. 40. Mehta V, Chavan A (2009) Physico-chemical treatment of tar-containing wastewater 16. White R (1987) Effect of lignin content and extractives on the higher heating value of generated from biomass gasification plants. World Acad Sci Eng Technol 57: Wood. Wood Fiber Sci 19:446–452. – 17. Ramankutty N, Evan AT, Monfreda C, Foley JA (2008) Farming the planet: 1. Geo- 161 168. fi graphic distribution of global agricultural lands in the year 2000. Global Biogeochem 41. BIOHPR (2004) Indirect gasi cation solves tar problem. Biomass heatpipe reformer Cycles 22:1–19. (BIOHPR). (Technical Univ Munich, Germany). http://ec.europa.eu/research/energy/ 18. Monfreda C, Ramankutty N, Foley JA (2008) Farming the planet: 2. Geographic dis- pdf/efchp_hydrogen7.pdf. tribution of crop areas, yields, physiological types, and net primary production in the 42. Rodríguez G, et al. (2008) Olive stone an attractive source of bioactive and valuable year 2000. Global Biogeochem Cycles 22:GB1022. compounds. Bioresour Technol 99:5261–5269. 19. FAO (2011) FAO statistical database. Available at http://faostat.fao.org/site/567/ 43. Isabelita MP, et al. (2008) Economic and environmental concerns in Philippine upland DesktopDefault.aspx?PageID=567. Accessed February 2011. coconut farms: An analysis of policy, farming systems and socio-economic issues. (Econ 20. Katalambula H, Gupta R (2009) Low-grade coals: A review of some prospective up- Environ Program Southeast Asia, University of the Philippines, Los Baños, Philippines). grading technologies. Energy Fuels 23:3392–3405. 44. Bally ISE (2004) Mangifera indica (mango). Species Profiles for Pacific Island Agro- 21. Adler PR, Del Grosso SJ, Parton WJ (2007) Life-cycle assessment of net greenhouse-gas forestry, ed Elevitch CR. Available at http://www.agroforestry.net/tti/Mangifera- flux for bioenergy cropping systems. Ecol Appl 17:675–691. mango.pdf. 22. Dien BS, et al. (2006) Chemical composition and response to dilute-acid pretreatment 45. Reginato GH, Victor Gd C, Robinson TL (2007) Predicted crop value for nectarines and fi and enzymatic sacchari cation of alfalfa, reed canarygrass, and switchgrass. Biomass cling peaches of different harvest season as a function of crop load. HortScience 42: – Bioenergy 30:880 891. 239–245. 23. Jae JH, et al. (2010) Depolymerization of lignocellulosic biomass to fuel precursors: 46. Anonymous (1918) War service by waste collection. Municipal J XLV:261–264. fi Maximizing carbon ef ciency by combining hydrolysis with pyrolysis. Energy Environ 47. Crutzen PJ, Andreae MO (1990) Biomass burning in the tropics: Impact on atmo- Sci 3:358–365. spheric chemistry and biogeochemical cycles. Science 250:1669–1678. 24. Krigmont HV (1999) Integrated biomass gasification combined cycle (IBGCC) power 48. Hao WM, Liu M-H (1994) Spatial and temporal distribution of tropical biomass generation concept: The gateway to a clearer future. A white paper. (Allied Environ burning. Global Biogeochem Cycles 8:495–503. Technol, Seal Beach, CA). Available at http://www.alentecinc.com/papers/IGCC/ 49. Lelieveld J, et al. (2002) Global air pollution crossroads over the Mediterranean. Sci- ADVGASIFICATION.pdf. ence 298:794–799. 25. IEA (2007) Biomass for power generations and CHP. Intl Ener Agency Energy Technol 50. Steinglass M (2009) Humble cooking fires contribute 18 percent of global greenhouse Essentials ETE03. Available at http://www.iea.org/techno/essentials3.pdf. gases. Asia Calling. Available at http://www.asiacalling.org/bn/news/vietnam/532- 26. Shipley A, Hampson A, Hedman B, Garland P, Bautista P (2008) Combined heat and fi power; Effective energy solutions for a sustainable future. (Oak Ridge Natl Lab, Oak humble-cooking- res-contribute-18-percent-of-global-greenhouse-gases. fi Ridge, TN). Available at http://www1.eere.energy.gov/industry/distributedenergy/ 51. Cecelski E (2000) Enabling equitable access to rural electri cation: Current thinking pdfs/chp_report_12-08.pdf. and major activities in energy, poverty and gender. (World Bank, Washington DC). 52. Lobell DB, et al. (2008) Prioritizing climate change adaptation needs for food security 27. International Energy Agency (2008) Energy balances of OECD countries: Beyond SCIENCE 2020 documentation. Available at http://earthtrends.wri.org/text/energy-resources/ in 2030. Science 319:607–610. SUSTAINABILITY variable-574.html. Accessed January 2011. 53. Parmesan C (2006) Ecological and evolutionary responses to recent climate change. 28. International Energy Agency (2007) Energy balances of non-OECD countries. Avail- Annu Rev Ecol Evol Syst 37:637–669. able at http://earthtrends.wri.org/text/energy-resources/variable-574.html. Accessed 54. MacDonald GK, Bennett EM, Potter PA, Ramankutty N (2011) Agronomic phosphorus January 2011. imbalances across the world’s croplands. Proc Natl Acad Sci USA 108:3086–3091.

Mendu et al. PNAS | March 6, 2012 | vol. 109 | no. 10 | 4019 Downloaded by guest on September 26, 2021