Geospatial mapping of biomass supply and demand for household energy management in Nepal

Adhikari, N.P.; Adhikari, R.C. Article Geospatial mapping of biomass supply and demand for household energy management in Nepal Development Engineering Provided in Cooperation with: Elsevier Suggested Citation: Adhikari, N.P.; Adhikari, R.C. (2021) : Geospatial mapping of biomass supply and demand for household energy management in Nepal, Development Engineering, ISSN 2352-7285, Elsevier, Amsterdam, Vol. 6, pp. 1-12, https://doi.org/10.1016/j.deveng.2021.100070 This Version is available at: https://hdl.handle.net/10419/299103 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/4.0/ Development Engineering 6 (2021) 100070 Available online 3 August 2021 2352-7285/© 2021 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Geospatial mapping of biomass supply and demand for household energy management in Nepal N.P. Adhikari a , * , R.C. Adhikari b a Alternative Energy Promotion Centre, Government of Nepal, Kathmandu, Nepal b Department of Mechanical and Manufacturing Engineering, University of Calgary, AB, T2N 1N4, Canada ARTICLE INFO Keywords: Biomass Geospatial mapping Energy supply and demand Fuelwood Crop residues Animal dung ABSTRACT This paper presents a geospatial mapping model for assessing spatial distribution and demand of biomass sources for household energy use in Nepal. In the context of rural households, correlation between supply and demand of biomass is crucial for designing effective rural energy programs. Three districts were considered to represent the country’s main topographical regions: lowlands, hills, and mountains, where geospatial distribution and demand of biomass are different. The supply potential of fuelwood was assessed using Geographical Information System (GIS) tool, and the potential of crop residues and dung and household energy demands were determined by field surveys and experiments. The results showed that households with secure access to biomass sources in lowlands, hills and mountains were 57%, 50% and 3% respectively. In lowlands, crop residues and dung were extensively used due to lack of forest biomass, whereas forest biomass was extensively used in hills and mountains, with negligible use of crop residues and animal dung. The results indicate that use of improved cooking stoves and biogas was negligible and thus cleaner biomass energy conversion and cooking technologies are needed to achieve universal target of clean cooking for all. The GIS model provided better estimation of biomass energy supply potential in the communities, which is crucial in the design of energy policies for sustainable clean cooking solutions. It is anticipated that this geospatial mapping model is also applicable to the cases of other developing countries, which have dominant biomass consumption for household energy use. 1. Introduction Biomass shares a significant portion of household energy use in the developing countries (Saygin et al., 2015; Cantarero, 2020; Kindermann et al., 2008). Under the framework of sustainable development goal 2030, biomass is an important energy source to support the clean energy generation target of the United Nations (Saygin et al., 2015; Thofern, 2011; Iverson et al., 1994; V´ avrov´ a et al., 2017). As biomass is used mostly as a traditional fuel in developing countries, there is a lack of scientific information on biomass sources and their spatial distribution, which has hindered the development of new technologies and appropriate policies for supply and demand management of household energy (Long et al., 2013; Rosillo-Calle et al., 2015; Ullah et al., 2015). The Geographic Information System (GIS) has been used as an appropriate computational tool to calculate the comprehensive spatial distribution of biomass resources (Kindermann et al., 2008; Iverson et al., 1994; V´ avrov´ a et al., 2017; B˘ anes¸ et al., 2010). The technical potential of biomass energy could be estimated by considering distance, means of transportation, and relevant landscape details in GIS (B˘ anes¸ et al., 2010; Milbrandt and Overend, 2008; Van Hoesen and Letendre, 2010; Chauhan, 2010). The modeling of energy resources in GIS provides spatial distribution, which is crucial to develop future energy technologies and systems. The GIS method helps to create useful maps with spatial distribution of different biomass resources and compare different energy options, and environmental and economic constraints. In this paper, we have presented a model for spatial mapping of biomass sources: fuelwood, crop residues, and animal dung considering the case of Nepal. Nepal relies heavily on biomass fuels to meet its energy needs, which contributes about 85% of the total energy use (about 77 million barrels of oil equivalent). The residential sector consumes about 90% of this biomass energy for cooking and space heating. The fuelwood, agricultural residues, and livestock dung are the sources of biomass with contributions of 88%, 5% and 7% respectively (WECS, 2014). Among three major geographically different regions of the country, the hills and mountains have relatively good forest resources as compared to lowlands (Chakraborty, 2001; Gautam et al., 2004; FAO, 1999; WECS, * Corresponding author. E-mail addresses: [email protected] (N.P. Adhikari), [email protected] (R.C. Adhikari). Contents lists available at ScienceDirect Development Engineering journal homepage: www.elsevier.com/locate/deveng https://doi.org/10.1016/j.deveng.2021.100070 Received 1 February 2021; Received in revised form 19 July 2021; Accepted 28 July 2021 Development Engineering 6 (2021) 100070 2 2010). In lowlands, crop residues and animal dung are extensively used for households energy needs due to deficit of fuelwood, whereas such uses are quite limited in other two regions (WECS, 2010; Pokharel and Chandrashekar, 1994; Pokharel, 2004; Pant, 2013). Direct burning of biomass has caused significant indoor air pollution, damage of forest ecology, and agricultural productivity (Pant, 2013; AEPC, 2014; AEPC, 2010; Singh et al., 2012; Metz, 1991; Melsom et al., 2001; Kurmi et al., 2013; Devakumar et al., 2014; Gami et al., 2009; Upadhyay et al., 2005). Over the past few decades, the government and various international agencies have initiated some programs to provide clean energy sources for cooking and reduce biomass consumption, mainly the dissemination of biogas and improved cooking stoves (ICS) with different types of subsidies (Gami et al., 2009; Upadhyay et al., 2005). The dissemination of biogas technologies can play a significant role in reducing traditional biomass use but presently, it is used by less than 3% of households. Improved cooking stoves (ICS) are used by a similarly low percentage of households (AEPC, 2015). These results indicate that the household energy use is highly inefficient as they mostly rely on traditional biomass burning. The health impact is severe to inhabitants, especially during cooking hours due to high indoor emissions of CO 2 from inefficient burning of biomass in non-ventilated buildings (Pokharel and Rijal, 2020). Furthermore, the use of ICS and biogas is negligible although they are more efficient and cleaner than traditional biomass burning. Hence, for the long run, the evaluation of sustainable biomass supply potential for cleaner energy production is the key step for long term energy planning based on which the need of intervention of other energy technologies can be explored. Given the need of technological intervention for utilizing biomass as clean fuels, biomass mapping would provide important information. In Nepal, about 50% people live in the lowlands, followed by 43% in the hills and 7% in the mountainous regions (CBS, 2011). In mountainous regions, fuelwood is the main energy source for almost all households, whereas this share is about 67% in hills and 57% in lowlands (CBS, 2011). In lowlands, the households use other fuel types such as dung cake, liquefied petroleum gas (LPG), kerosene and biogas. About 22% of these households burn dung cake in lowlands, whereas no such households exist in the other two regions. The fuelwood utilization is more in rural households than in urban, whereas the monthly national average use of fuelwood is higher than that of other developing countries (Shahi et al., 2020). Even though crop residues are used for cooking and space heating especially in lowlands, this information is not incorporated in national energy statistics. As such, other sources such as electricity, waste, coal, briquettes, etc are marked as “others”. Furthermore, there are no specific building designs that incorporate space heating in households. About 22% households have concrete walls, 29% have cement-bonded stone walls, and the remaining have walls made from local materials such as thatch, straw, wood, and mud (CBS, 2011). Due to direct burning of biomass with improper ventilation and lack of designs for space heating and cooling, heat is wasted in winter whereas excessive heating takes place in summer (Pokharel et al., 2020). National biomass energy strategy of Nepal revealed that local biomass could not be deployed effectively for energy generation because of inadequacy of resource information (MoPE, 2017). Biomass assessment is the key to manage sustainable biomass supply for various alternative uses including energy generation in a society which also provides an opportunity to promote local bio-economy (Hoang et al., 2020). The current energy policies and programs are based mostly on households’ demand due to self-awareness created by various community-level organizations, governmental institutions, non-governmental agencies, and private energy companies (AEPC, 2016). In some cases, various government and non-government agencies have directly intervened improved cooking stoves to poor and socially-excluded households through subsidies and incentives (REDP, 2009; SNV, 2013). Due to lack of reliable information on biomass sources, there are challenges in designing appropriate programs for rural households. Availability, accessibility and calorific value of biomass energy sources are the key aspects that determine the actual potential for energy generation. This is important, particularly for Nepal where the country’s topographical variations impact the distribution of forest biomass to determine the availability and utilization sources. Geospatial mapping of biomass is important because the collectable biomass is influenced by many factors such as, distance between the sources and the end-use locations, the extent of resource-protected areas, transportability and corresponding economic aspects. These parameters can be modeled in GIS to determine the spatial distribution of biomass sources (Long et al., 2013; Yousefi et al., 2017; Fernandes and Costa, 2010; Viana et al., 2010). The spatial aggregation and query tools in GIS can be used to evaluate biomass assessment where different layers or themes can be combined and perform analysis between different objects on those layers or within a single layer (Stanbury and Starr, 1999). The literature survey shows that previous studies on biomass assessment in Nepal have analyzed supply and demand status of fuelwood at the community level without considering all types of biomass sources, mainly fuelwood, crop residues, and animal dung (Schreier et al., 1991; Marzoli and Drigo, 2014). There is a significant lack of literature in geospatial distribution of these biomass fuels. Therefore, this study presents a geospatial mapping model to determine the spatial distribution or supply potential and demands of key biomass sources that are commonly used for household energy use. This is the first study in biomass resource mapping that presents an integrated information on key biomass sources (fuelwood, crop residues and dung). The fuelwood supply potential was evaluated using GIS tool. The crop residues and dung were measured through field surveys and lab experiments respectively, which provide better estimation of biomass sources. The results are useful in prioritizing different energy technologies to achieve clean cooking target of the United Nations. Furthermore, we anticipate that the biomass mapping model is also applicable in the context of other developing countries, where rural communities predominantly rely on biomass for household energy needs, and a complete replacement is not practically feasible in the foreseeable future (UN, 2018). The remainder of the paper is organized as follows. Section 2 describes the materials and methods of the study, mainly the data collection, survey and field experiments and measurements, and sampling methodology used in biomass mapping. Section 3 presents the GIS model for determining the geospatial distribution of forest biomass. Section 4 describes the results and discussion, and finally Section 5 summarizes the main findings of the study. 2. Materials and methods 2.1. Study areas The study was designed considering three topographically different districts, which are characterized by different types of biomass sources and variations in spatial distribution, agricultural production, and livestock management. To represent these variations, Bajhang, Lamjung, and Morang were selected, which represent mountains, hills, and lowlands respectively as shown in Fig. 1. For convenience, the names of districts are replaced by the corresponding region, i.e. mountains, hills and lowlands throughout the paper. Covering the area of 3,422 km 2 and the altitude variation from 900 to 7,077 m AMSL, Bajhang holds forest and arable lands of 43.7% and 7.9% respectively. Lamjung is a hilly district with altitude variation from 385 to 8,162 m AMSL and covering an area of 1,691 km 2 . The forest and arable lands are 55.2% and 36.8% respectively. Similarly, Morang covers an area of 1,855 km 2 , which is categorized as a lowland district with an altitude variation from 60 to 2,410 m AMSL. As compared to other two districts, Morang has relatively higher share of arable land (56.7%) but lower share of forest land (24.3%) (CBS, 2014; DFRS, 2018). Fuelwood is the main energy source for almost all households in the N.P. Adhikari and R.C. Adhikari Development Engineering 6 (2021) 100070 3 mountains, whereas this share is about 70% in hills and 45% in lowlands where the households use other fuel types such as dung cake, LPG, kerosene and biogas. About 25% of these households burn dung cake in lowlands whereas no such households exist in hills and mountains (CBS, 2011). The seasonal variations of temperature and weather are summarized in Table 1, which shows significant differences in climate in three regions. The mountains are characterized by high elevation and cold climate, where the temperature is usually below 10 ◦C. The hills are characterized by mid elevation and mild climate, and the lowlands are characterized by the lowest elevation with high temperature and humid climate. The temperature ranges in the lowlands are 25–40 ◦C. 2.2. Data collection and methodology The data used in the study were collected from household surveys conducted in each geographical region and complemented by GIS tool and secondary data from the literature. The methodology adopted for biomass resource assessment is shown in Fig. 2. The fuelwood supply potential was assessed using spatial GIS techniques. We used Woodfuel Integrated Supply-Demand Overview Mapping (WISDOM) method to assess the supply potential of fuelwood where accessibility was determined based on cost-distance algorithm (Masera et al., 2006). The WISDOM model uses geo-referenced socio-demographic and natural resource data, which integrates relevance data of fuelwood from multiple sources (Masera et al., 2006). In case of crop residues and dung, the supply potential was assessed through household surveys and field experiments. The field experiments were conducted to measure dung, crop residues and biomass utilization for household energy uses. Similarly, the biomass demand was determined by assessing the prevailing energy use patterns of households. All findings were synthesized in terms of annual per capita and mapped at the lowest administrative unit of the district, defined as a community unit (CU) in this study. The CUs manage community forests in most cases by introducing local forest regulations, and in some cases, provide technical assistance to community groups for forest management. Hence, biomass information is crucial to formulate regulations on fuelwood usage and thus contributes toward secure supply and distribution of fuelwood in the communities. 2.2.1. Sample households and survey The probability proportional to size (PPS) sampling method was used to select sample households in different CUs in each district. The households using at least one type of biomass for cooking and space heating were selected. For the sample size of households in each district, we followed a rule of thumb of 10–15 observations as per the United Nations’ recommendation for household surveys (Babyak, 2004). As there are five combinations of fuel usage (fuelwood only, fuelwood and biogas, fuelwood and LPG, fuelwood and dung cake, and fuelwood, dung cake and crop residues), we assumed that a sample size of 80 households is adequate to characterize the variations of households’ energy use in each CU. Household surveys were conducted using general and closed modes. Fig. 1. Research areas for biomass spatial mapping. Table 1 Average temperature and precipitation in different research areas (Adhikari and Adhikari, 2021). Season Parameter Mountains Hills Lowlands Pre-monsoon (April–May) Maximum temperature (◦C) 10–15 15–20 30–35 Minimum temperature (◦C) 0–5 5–10 15–20 Precipitation (mm) 100–200 200–400 200–400 Monsoon (August–September) Maximum temperature (◦C) 15–20 20–25 30–35 Minimum temperature (◦C) 5–10 10–15 20–25 Precipitation (mm) 1,000–1,500 1,500–2,000 1,000–1,500 Winter (December–January) Maximum temperature (◦C) 5–10 10.1–15 20–25 Minimum temperature (◦C) −5 to 0 0–5 10–15 Precipitation (mm) 100–200 50–100 30–50 N.P. Adhikari and R.C. Adhikari Development Engineering 6 (2021) 100070 4 The purpose of the two modes is to acquire information both on current biomass utilization status as well as measure biomass supply potential for different seasons through field experiments. The questionnaire for general survey was structured to collect information about household demographics, cooking and space heating energy uses, fuelwood collection, crop harvesting methods and livestock management. Similarly, the closed survey was conducted to measure the consumption of fuel and to evaluate the supply potential of crop residues and livestock dung for energy uses. Further details of the methods are discussed in the next section. 2.2.2. Lab experiment for moisture content and dry mass An important part of this study was the experimental measurement of moisture content of biomass fuels (fuelwood, crop residues, and dung) and oven-dried (OD) weight or dry mass. The experimental method adopted to determine dry mass is shown in Fig. 3. The moisture content was measured in two stages, one with reference to sun-dried mass and other oven-dried in a temperature-controlled muffle furnace. We measured moisture content and oven-dried weight of selected fuelwood, crop residues and dung. First the samples of biomass were collected at field and brought into lab by placing them in tightly packed plastic bags. After a few days, the samples were sun-dried for 40–45 days. Then the samples were heated in the muffle furnace at a uniform temperature of 103 ◦C ±5 ◦C up to 14 h until dry mass was achieved. The weights of samples were measured every 4 h, and the oven –dried weights of sample and the moisture content were determined. The moisture contents of fuelwood and dung were determined from the average of three samples. Three samples were taken for each type of crop residues. The uncertainty in the measurements of dry weight and moisture content is ±0.1%. 3. Description of the GIS model The GIS sheets with different layers of topography, administrative boundary, population settlements and road networks were collected from the Department of Survey, Government of Nepal (DoS, 2017) as inputs in the geospatial model. Similarly, the land cover raster maps of the districts were obtained from the work of (Uddin et al., 2015), which were prepared from 30 m Landsat TM data for the year 2010. These maps have classified land into twelve different uses and cover four types of forests (needle leaved open, needle leave closed, broadleaved open and broadleaved closed), shrub-land, agricultural land, bare area, built-up area, river, lake and snow/glacier as shown in Table 2. Likewise, the Digital Elevation Map (DEM) of 30 m spatial resolution was obtained from ASTER Global DEM data, which is a joint product of the Ministry of Economy, Trade, and Industry of Japan, and the US National Aeronautics and Space Administration (NASA). Arc GIS 10.1 was used to model the biomass spatial distribution or the supply potential of each Fig. 2. Methodology for biomass resource assessment. Fig. 3. Schematic of the experimental measurement procedure for moisture content and dry mass of biomass fuels. N.P. Adhikari and R.C. Adhikari Development Engineering 6 (2021) 100070 5 region (ESRI, 2015). The model was developed with the following procedures and assumptions. 3.1. Modeling of biomass supply potential 3.1.1. Fuelwood The major sources for fuelwood collection in Nepal are national or community forests, private forests, and homestead surroundings and terraced agricultural lands (Mahat, 1987). The distributions of private and homestead surroundings and terraced agricultural lands are uneven for different households and in some cases, the households have no private land to get fuelwood. In many cases, there are specific rules developed by the local community. This study did not consider such limitations of accessibility that preclude households to collect fuelwood while assessing potential. Therefore, the data reported in this study should be used by considering specific accessibility issues in local areas. The methods to estimate fuelwood supply were based on the work of (Marzoli and Drigo, 2014) in which three key distinct modules for demand, supply and integration were developed for defined minimum spatial boundary. The data from previous studies (Marzoli and Drigo, 2014; Baral et al., 2009; FAO, 1999; DFRS, 2014b; DFRS, 1999; FAO, 1999; DFRS, 2014a) were used to determine mean annual fuelwood increment in terms of dry mass (DM) from different land covers as no single systematic study was available with such information. Mean annual increment is the total biomass produced in a particular area divided by the number of years required to produce it (Amatya and Shrestha, 2010). The results of mean annual increment of biomass are presented in Table 2. 3.1.2. Assessment of crop residues As the crop residues share a significant portion of household energy use, five major cereal crops: paddy, wheat, maize, millet and barley were evaluated to assess the contribution of crop residues in household energy use. The share of these cereal crops is more than 90% of total agricultural production in mountains, hills, and lowlands (CBS, 2014). Closed mode survey was carried out in the sample households for three different seasons (post-monsoon, pre-monsoon and monsoon) to determine residue-to-product ratio (RPR) for each crop residue. RPR refers to the ratio of weight of crop residues in air dried form that is available after processing certain amount of harvested crops to the weight of main crop obtained from the same process (Ayamga et al., 2015). The surveys were conducted in these seasons considering crop harvesting times to determine RPR. Nine households in each district were requested to separate all crop residues collected from a predefined land area and process them as usual. The land area was different in each household and varied from 50 to 100 m 2 . The grains and residues were weighed on air-dry basis from 25 to 30 days of crop harvesting, based on which corresponding RPR values were calculated. The crop residues remaining in the field were not considered because of their essential role in maintaining soil nutrients. The RPR values thus obtained were used to quantify crop residues used for household’s energy use. The estimation of crop residues for energy uses were based on following assumptions: a. The cultivated lands were classified into four broad ranges based on the sea-level altitude, which are low (up to 500 m), Moderate (500–1,200 m), High (1,200–1,800 m) and very high (above 1,800 m). Based on this, cropping patterns were labeled according to the information synthesized from survey. b. The RPR values for respective crop residues as obtained from field experiment are presented in Table 3. c. The crop productivity in each CU was derived by averaging yearly productivity of ten year data (2004–2014) which was obtained from the Ministry of Agriculture and Livestock Development, Nepal (MoALD, 2019). d. The potential crop residues for energy uses was evaluated by analyzing four major uses of crop residues, such as building material, mulching/burning, selling and livestock fodder. e. The weight of net crop residues was converted from air-dried to oven-dried form on the basis of corresponding moisture contents. 3.1.3. Experimental measurement of dung supply By considering availability of crop residues for livestock fodder, the amount of dung supply was quantified by measuring daily dung yield for three different seasons under the following conditions: a. Fresh dung yield of livestock for 24 h and flow analysis of dung from stall to end use were measured for 70 livestock (cattle and buffalo) in each district from the sample households. In order to cope with district sample statistics, the livestock samples were proportionally divided into four categories: mature cattle (>3 yr), young cattle (≤3 yr), mature buffalo (>3 yr), and young buffalo (≤3 yr). b. The dung yield was measured for three different time periods of a year, in which fodder availability was different. Fodder availability was the highest in monsoon: June–September, and the lowest in winter: December–February. Likewise, fodder availability was moderate in pre-monsoon: March–May. The average dung yields in three different seasons are presented in Table 4, which were used to determine the potential energy use. c. To determine the energy potential of dung, household’s livestock management practice, dung collection method, and alternative uses of dung were systematically assessed in the survey. Table 2 Annual fuelwood increment (kg dry mass/ha) for different land covers at various altitudes (Marzoli and Drigo, 2014; Baral et al., 2009; FAO, 1999; DFRS, 1999; DFRS, 1999; FAO, 1999; DFRS, 2014a; Amatya and Shrestha, 2010). Land classification upto 500 m 500–1,200 m 1,200–2,200 m 2,200–3,500 m >3,500 m Needle leaved open forest 1,429 1,429 1,753 2,457 2,509 Needle leaved closed forest 3,039 2,541 2,612 3,889 3,889 Broadleaved open forest 2,535 2,535 2,191 3,451 2,502 Broadleaved closed forest 3,242 3,242 3,089 2,883 2,783 Grassland 549 549 549 549 549 Agriculture 612 612 612 612 – Built up area 550 550 550 – – Table 3 Residue to product ratio (RPR) values of crop residues. Crop residues No of samples RPR Paddy husk 18 0.363 ±0.14 Paddy straw 18 1.97 ±0.57 Wheat husk 11 0.82 ±0.17 Wheat straw 11 1.46 ±0.39 Maize stalk 11 2.12 ±0.45 Corn cob 11 0.28 ±0.05 Corn ear 11 0.29 ±0.06 Millet husk 11 0.14 ±0.04 Millet straw 11 1.89 ±0.53 Barley straw 9 1.52 ±0.43 N.P. Adhikari and R.C. Adhikari Development Engineering 6 (2021) 100070 6 3.2. Modeling of household energy demand The household surveys showed that all the sample households in the mountains exclusively used fuelwood, whereas multiple energy fuel sources were used in hills and lowlands. These sources were fuelwood, biogas, LPG, crop residues, dung cake, kerosene and electricity. There were eight such different combinations of fuel sources to households in hills, whereas there were fifteen in the lowlands. In order to align with the existing national energy statistics of CUs (CBS, 2011), fuel use was categorized into five types: fuelwood only, fuelwood and biogas, fuelwood and LPG, fuelwood and dung cake, and fuelwood, dung cake and crop residues. By determining the "fuelwood equivalent" of crop residues, dung and biogas (Adhikari, 2017), the annual biomass energy use in household was calculated on per capita basis, Table 5. The "fuelwood equivalent" refers to the conversion of other energy sources into equivalent weight of the fuelwood to meet household energy needs (Adhikari, 2017). The values for "fuelwood equivalent" of biogas, crop residues, LPG and dung were taken from (Adhikari, 2017). By utilizing the measured data of moisture contents from the experiment conducted in this study, the weight of biomass was converted into oven-dried form. The annual energy usage per capita was determined through experiments to cook a standard meal in each household considering different fuels (fuelwood, dung, biogas, crop residues and LPG). Since the cooking experiments were carried out in the community households, the moisture contents of fuelwood, crop residues and dung were assumed negligible when estimating the “fuelwood equivalent”. Then the annual biomass energy use per capita in each household was determined, Table 5. However, no detailed uncertainty analysis was conducted to determine whether 80 observations are sufficient to capture the degree of variations of biomass usage in various CUs. As energy usage per capita for cooking a standard meal is similar in different areas, we anticipate that the variation in distribution of biomass usage within a CU is small. 4. Results and discussion The contour maps of annual accessible supply of fuelwood in three different regions are plotted in Fig. 4. It was found that the net annual supplies of accessible fuelwood in lowland, hill, and mountain districts were estimated as 232,950, 43,025, and 44,534 metric tons respectively. On per capita basis, the hill district has the highest fuelwood supply with 257 kg/yr, and the lowland and the mountain districts have 240 kg/yr and 228 kg/yr respectively. The CUs with annual per capita fuelwood supply more than 800 kg/ha lie on the northern part of lowland, where only 5% of the district’s population lives. As compared to hill and mountain districts, the distribution of forest and human settlement is highly disproportionate in the lowland district. The weights of annual supply potential and demand of three types of biomass are summed in terms of fuelwood equivalent to obtain a single value for annual per capita supply and demand of biomass for each CU. Biomass energy of a particular CU was calculated based on the corresponding values presented in Table 5. The annual per capita demand of a particular CU was deducted by the corresponding biomass supply to establish the relationship between biomass supply and demand. By deducting annual biomass demand from the corresponding biomass supply, the net availability of biomass in terms of either surplus or deficit was determined for each CU. It is noted that the positive values indicate biomass surplus, whereas the negative values indicate biomass deficit. The information of each type of biomass on each CU is useful to quantify the amount of particular biomass used. For example, one of the CUs in the lowland (Table 6, Code no. 10) has surplus annual biomass per capita as 92 kg. This shows that there is a potential for reducing fuelwood by 83 kg by utilizing all dung available for biogas production. Similarly, spatial information helps to evaluate the possibility for transporting biomass from abundant CUs to other CUs with biomass deficit. The GIS method is valuable in assessing fuelwood potential to those areas of mountains and hills with complex terrain where alternative methods of assessment require significant cost. From Table 6, more than 80% CUs have surplus biomass (54 CUs out of 66) in lowlands, where fuelwood can only provide surplus biomass to only 13% CUs (9 CUs). Those CUs with fuelwood surplus lie in northern part of the district and hold typical characteristics of hillswith a populationshare of less than 5% of the district population. Therefore, in lowlands, the higher production of crop residues and dung contributes to reliable supply of biomass for household energy (45 CUs: 70% of CUs). It is worth mentioning that these crop residues and dung as well as fuelwood are directly burned for household energy use, such as cooking and space heating. These results reveal that the currently used crop residues and dung for household energy need to be converted to cleaner biomass fuels, and cleaner cooking technologies should be developed. Similarly, the fuelwood supply and demand presented in Table 7 for hills show that only three CUs out of 64 have reliable supply of fuelwood. Because of easy availability of fuelwood, no crop residues were used for energy generation whereas the use of dung was limited in the form of biogas. Unlike in lowlands, dung was not burned directly for energy generation. We note that although there was reliable supply of crop residues and dung for about nearly half of total CUs, the fuelwood was over exploited. Despite significant increment in total forest areas in hills, the forest areas have been declining significantly on densely populated areas due to expansion of agricultural areas, settlements and infrastructure development. Hence, there have been uneven spatial distribution of forest areas for which the spatial analysis is crucial to take suitable policy measures and protect forest areas. Similarly, the results presented in Table 8 for mountains show that only 5 CUs (about 10% of total CUs) had potential to maintain reliable biomass supply and demand. The addition of crop residues and dung was not sufficient to maintain adequate supply in none of the CUs due to relatively lower production of crop residues and dung. The higher Table 4 Weight of daily fresh dung yield per livestock (kg). Categories Location No of samples Winter Young cattle Lowlands 20 9.5 ±4.6 Hills 11 5.5 ±3.0 Mountains 21 2.2 ±1.5 Mature cattle Lowlands 31 15.8 ±7.1 Hills 20 10.2 ±2.3 Mountains 32 4.3 ±1.9 Young buffalo Lowlands 7 11.3 ±4.5 Hills 14 10.3 ±4.9 Mountains 8 6.3 ±2.9 Mature buffalo Lowlands 12 22.5 ±6.7 Hills 25 24.3 ±6.3 Mountains 9 16.9 ±4.3 Table 5 Biomass energy use (oven-dry mass). Household energy mix Energy source Fuelwood equivalent (kg/yr per capita) Lowland district Hill district Mountain district Fuelwood only Fuelwood 467 ±183 779 ± 295 713 ±367 Fuelwood & biogas Fuelwood 314 ±195 430 ± 235 – Biogas 130 ±53 216 ± 78 – Fuelwood & LPG Fuelwood 405 ±155 533 ± 249 – LPG 70 38 – Fuelwood & dung cake Fuelwood 331 ±172 – – Dung cake 145 ±116 – – Fuelwood, dung cake & crop residues Fuelwood 235 ±128 - - – Dung cake 161 ±91 – – Crop residues 61 – – N.P. Adhikari and R.C. Adhikari Development Engineering 6 (2021) 100070 7 fodder deficiency and less cultivated lands with poor agricultural facilities were found to be the major causes of lower production of dung and crop residues respectively. Therefore, the use of crop residues and dung was quite negligible in mountains, and fuelwood was predominantly used by almost all households. From the results presented in Tables 6–8, we classified the supply potential of fuelwood, crop residues, and dung into six types with respect to the demand of household energy. The number of households at each CU was taken from national statistics (CBS, 2011). The results are summarized in Table 9. From Table 9, it is observed that only about 3% biomass adopting households have adequate supply of biomass energy in mountains for which fuelwood has the biggest role. The households with adequate biomass supply in hills and lowlands are 50% and 57% respectively. The forest resources in hills and mountains for most households were relatively greater than that of lowlands. Therefore, consideration of mean annual increment of fuelwood in view of fuelwood harvesting is not worthy from the perspectives of fuelwood users. As such, the households are able to collect fuelwood as per their demand. However, in view of annual supply potential presented in Table 9, the prevailing fuelwood consumption is inadequate in terms of reliable supply. Further, it is evident that despite higher fuelwood deficits in lowlands compared to hills and mountains, the relatively better production of crop residues and dung was the main reason for sufficient supply of biomass to more than half of biomass adopting households. The results revealed that the available but unutilized crop residues and dung could not play significant roles to meet biomass demand in mountains, whereas the same played a significant role in hills and lowlands. Therefore, despite having sufficient crop residues and dung for household energy use in hills, those were not utilized, resulting in overexploitation of fuelwood. The potential application of dung and crop residues is discussed in terms resource availability for direct burning. However, these biomass fuels should be converted into cleaner form by intervention of technologies such as biogas and briquette. The results of present study for annual biomass consumption per capita are compared with the previous studies in Table 10. The data for different biomass fuel supply are the corresponding average of data presented in Tables 7–9. The results for dung and crop residues at CU level are presented in terms of fuelwood equivalent by their respective equivalent coefficient (Adhikari, 2017). Except fuelwood consumption, Fig. 4. Contour maps of annual accessible supply of fuelwood in mountain, hill, and lowland districts. N.P. Adhikari and R.C. Adhikari Development Engineering 6 (2021) 100070 8 there are only a few limited studies to compare the results of present study. The results are compared with respective to geographical regions. The average fuelwood consumption data was obtained from Table 5. The average fuelwood consumption was found within the range of data reported by previous studies in all topographic regions. Similarly, we found that dung consumption in lowlands is significantly lower than that reported in reference (Behera et al., 2015). This difference could be due to variations in livestock per holding and fodder availability. The annual per capita supply potential of crop residues in hills from previous study (80 kg) is significantly lower than the present study (207 kg), which Table 6 Biomass supply and demand in CUs of Lowlands in oven-dried form (kg/yr per capita). Code no. CU Supply Demand Surplus/Deficit Fuelwood Crop res. Dung Total 1 Amahibariyati 188 83 221 492 370 122 2 Amardaha 220 97 194 511 404 107 3 Amgachhi 188 83 327 598 328 270 4 Babiyabirta 260 113 119 492 384 108 5 Bahuni 404 100 126 630 472 158 6 Banigama 217 112 169 498 430 68 7 Baradanga 281 80 96 457 352 105 8 Bayarban 359 59 178 596 479 117 9 Belbari 452 78 95 625 482 143 10 Bhathigachh 175 56 236 467 375 92 11 Bhaudaha 169 74 192 435 342 93 12 Bhogateni 3,981 54 133 4,168 500 3,668 13 Biratnagar 24 7 41 72 148 −76 14 Buddha Nagar 226 55 97 378 340 38 15 Dadarbairiya 261 75 93 429 341 88 16 Dainiya 168 74 243 485 371 114 17 Dangihat 151 30 43 224 487 −263 18 Dangraha 198 81 234 513 406 107 19 Darbesha 251 109 155 515 402 113 20 Dulari 97 71 187 355 415 −60 21 Gobindapur 238 105 155 498 384 114 22 Haraicha 215 85 279 579 463 116 23 Hasandaha 227 94 247 568 448 120 24 Hattimudha 207 77 221 505 392 113 25 Hoklabari 224 99 242 565 428 137 26 Indrapur 162 18 63 243 474 −231 27 Itahara 354 94 139 587 465 122 28 Jante 2,481 93 56 2,630 498 2132 29 Baijanathpur 142 83 397 622 392 230 30 Jhorahat 157 69 581 807 389 418 31 Jhurkiya 243 132 85 460 367 93 32 Kadamaha 183 97 184 464 367 97 33 Kaseni 352 111 119 582 446 136 34 Katahari 138 75 179 392 426 −34 35 Kerabari 1,133 62 39 1,234 494 740 36 Keraun 255 93 184 532 464 68 37 Lakhantari 228 101 140 469 350 119 38 Letang 699 47 41 787 489 298 39 Madhumalla 442 89 91 622 491 131 40 Mahadewa 206 85 165 456 352 104 41 Majhare 176 58 212 446 340 106 42 Motipur 328 144 99 571 438 133 43 Mrigauliya 396 55 141 592 468 124 44 Nocha 183 126 129 438 344 94 45 Pathari 126 32 59 217 486 −269 46 Patigaun 3,392 64 151 3,607 578 3,029 47 Pokhariya 197 56 195 448 345 103 48 Rajghat 234 103 252 589 470 119 49 Ramitekhola 3,788 45 321 4,154 583 3571 50 Rangeli 154 54 165 373 400 −27 51 Sanischare 127 41 70 238 485 −247 52 Sidraha 293 85 193 571 426 145 53 Sijuwa 2,46 158 175 579 453 126 54 Simhadevi 3,601 66 155 3,822 557 3,265 55 Sisabanijahada 217 52 86 355 385 −30 56 SisbaniBadahara 197 115 186 498 391 107 57 Sorabhag 251 89 146 486 377 109 58 Sundarpur 286 56 106 448 479 −31 59 Takuwa 214 89 142 445 358 87 60 Tandi 1,287 69 117 1,473 586 887 61 TankiSinuwari 83 36 56 175 425 −250 62 Tetariya 223 98 236 557 398 159 63 Thalaha 217 87 142 446 354 92 64 Urlabari 89 25 52 166 483 −317 65 Warangi 4,080 32 135 4,247 594 3,653 66 Yangshila 2,184 47 63 2,294 586 1,708 N.P. Adhikari and R.C. Adhikari