Relationship between Household Dynamics, Biomass Consumption, and Carbon Emissions in Pakistan

Abstract

Over the years, the household sector has become an important energy consumer and the main source of greenhouse gas (GHG) emissions. The rural household sector has significant potential for emission reduction due to its heavy reliance on traditional fuels and technologies. A great number of academic studies have been undertaken to analyze patterns of household energy and their determinants around the globe, particularly in developing countries. However, little is known about the association between household dynamics and patterns of energy (biomass vs. non-renewable) use. This study aims to analyze the relationship between different household dynamics, such as household size, income, climate, availability of resources, markets, awareness, consumption of energy, and carbon emissions. The study uses the STIRPAT model to investigate the impact of income, household size, housing dimensions, clean energy, and market accessibility on energy consumption. The findings of the study reveal that biomass energy accounts for the majority of household energy consumption and dung has the highest share in total household energy consumption (39.11%) The consumption of biomass increased with the size of the household and decreased with the level of income. A 1 kgoe increase in biomass consumption resulted in a 15.355 kg increase in CO₂ emissions; on the other hand, a 1 kgoe increase in non-renewable-energy consumption resulted in just a 0.8675 kg increase in CO₂ emissions. The coefficients of housing unit size, distance from the LPG market, and livestock were the primary determinants for choosing any fuel. Having knowledge of modern cookstoves, clean energy, and the environmental impact of fuels reduced the consumption of both energy sources. Furthermore, it was found that households with a greater reliance on biomass emitted higher quantities of carbon compared to those with a low reliance on biomass. Based on the results of the study, it can be stated that a reduction in the use of biomass and non-renewable energy is possible with adequate interventions and knowledge.

1. Introduction

The single most prominent challenge of the 21st century is global warming induced by surging levels of greenhouse gas (GHG) emissions. Carbon dioxide, hydrofluorocarbons, nitrous oxide, perfluorocarbons, methane, and sulfur hexafluoride are the six major greenhouse gases that significantly influence the environment through global warming, and reducing the quantity of these greenhouse gases is the foremost concern of the global community [1]. Among these six greenhouse gases, carbon dioxide has a particularly great influence on climate change [2]. Therefore, it is important to understand the factors responsible for increasing GHG emissions and what policy options can be taken into consideration to counter this eminent threat. In the last century, human activities have played a significant role in the emissions of GHGs [3]. Evidence suggests that aggregate pollution levels and shifts in world temperature have a strong relation to household activities, and the housing sector is one of the leading contributors to GHGs [4]. Abeydeera et al. [2] pointed out that household activities contribute 80 percent of global carbon emissions. Transportation, construction, power generation, cooking, and heating are the main household activities contributing to GHG emissions [5,6].

The household sector is a major consumer of energy, and energy consumption has a significant share in total GHG emissions. It is predicted that household energy demand will rise in the future, and this will increase the emissions of greenhouse gases [7]. The intensity of carbon emissions depends on energy sources. Traditional fuels are the main source of energy in most parts of the world and considerably contribute to carbon emissions due to incomplete combustion. According to estimates, the combustion of traditional fuels by households contributes about 21 percent of carbon dioxide emissions [8,9]. Traditional fuels, including firewood, coal, charcoal, crop residues, and dung; renewable energies, such as biogas and solar; and modern energy, comprising LPG, natural gas, and electricity, are the three major categories for household energy. A significant majority of the population around the globe consumes traditional biomass for domestic purposes [10]. In developing countries, about 90 percent of the residents depend on traditional biomass, which is the main contributor to carbon emissions [11]. Carbon emissions are increasing rapidly in underdeveloped nations, in comparison to developed economies [12]. Roughly 2.8 billion low-income households use biomass, which comprises 9 percent of the primary energy supply and 55 percent of the wood harvest. Traditional biomass is an important source of energy for low-income households in developing countries [6].

Biomass is used by 105 million individuals in Pakistan for cooking, and the incomplete combustion of biomass results in 28,000 deaths and 40 million respiratory illness cases per annum in Pakistan [11,13]. Almost 86 percent of individuals in Pakistan consume biomass (fuelwood, crop residues, and animal dung) for cooking and heating, and fuelwood is a major biomass source that is used by 54 percent of households. A high concentration of indoor air pollution is the outcome of burning with open fires or traditional stoves [14]. Moreover, it has been found that, globally, 3 million deaths per annum are caused by indoor air pollution. Pakistan is the fifth most populous country in the world, and it is estimated that its population will reach 403 million in 2050. The increasing population and per capita income have increased the energy demand. On the other hand, poor energy infrastructure, heavy reliance on thermal power, the low affordability of expensive imported fuels, high transmission and distribution losses, and other supply side factors have widened the gap between supply and demand. More than 60 percent of the population live in rural areas and, due to both demand-side and supply side factors, the majority of this population relies on traditional fuels for their domestic energy needs. Another reason for the heavy reliance on traditional fuels is easy accessibility. Pakistan is an agricultural country and the countryside is rich in traditional energy sources, such as crop residues, animal dung, and fuelwood.

However, Pakistan is not among the major contributors to CO₂ emissions. Compared to its Asian neighbors, China and India, who contribute 33% and 7.1%, respectively, to global CO₂ emissions, Pakistan’s share in global emissions is negligible (0.5%). However, since its independence, Pakistan has experienced a sharp increase in per capita CO₂ emissions, increasing from 0.02 tons in 1947 to 1.13 tons in 2017 [15]. The uncontrolled increase in CO₂ emissions is a serious threat to the environmental sustainability of Pakistan. Pakistan has been ranked among the top ten countries most affected by climate change. It is necessary for Pakistan to prepare itself for climate change adaptation as well as investigate sources that contribute to climate change domestically. Therefore, the purpose of this study is to estimate biomass and non-renewable-energy consumption in households and carbon emissions from the combustion of traditional biomass fuels and non-renewable energy.

2. Theoretical or Conceptual Framework

Household energy accounts for a significant proportion of the total global energy consumption. The energy-use patterns of households show variance because energy-use decisions are influenced by several factors [16]. Household energy consumption behaviors change over time, placing strain on both energy efficiency and the environment. Behavioral factors are considered more important in achieving energy conservation [17]. Thus, understanding how to change energy-use behavior can be an effective way to improve energy efficiency and environmental quality. The energy consumption behavior of households can be described in three dimensions [16]. The first dimension is time, which describes how the household’s energy-consumption behavior can vary over an hour, day, month, or year. It means that household energy-consumption behavior can be described in different time scales. For example, energy consumption in a day may differ from the energy consumption profile of a month. Second, the user dimension explains why household energy-consumption behavior greatly varies. It states that various factors, both external and internal, influence household energy consumption. The external factors include demographics, housing features and characteristics, income, and a variety of other factors. Similarly, internal factors include household members’ habits and environmental awareness. The third and final dimension is the spatial dimension, which describes how household energy consumption differs depending on location. The geographical, environmental, climatic, and development levels of a region all influence the energy-consumption behavior of households.

Moreover, household energy-consumption behavior has been described by different paradigms, such as economic and behavioral-oriented paradigms. The rational choice theory is highly applicable to energy-consumption behavior in the economic paradigm. This theory explains why consumers act rationally and seek the greatest utility at the lowest possible cost [18,19]. Rational consumers base their energy-consumption decisions on the benefits and costs of various energy types, as well as all available internal and external information. Therefore, the intellectual burden of information processing always deteriorates the ability of energy consumers to take purposeful action, and consequently they are restricted to the realm of rationality [20]. Similarly, the behavior-oriented paradigm states that energy-consumption behavior is influenced by a variety of intrapersonal factors (e.g., awareness, information, and habits), interpersonal factors (e.g., norms and social environment), and external factors (e.g., incentives, awards, and punishment).

This study used all three dimensions and both paradigms to predict the energy consumption behavior of homes. For example, the daily consumption (time dimension) of different types of energy by different households (user dimension) at home (spatial dimension) was considered in this study. In addition, the economic and behavior-oriented paradigms were also taken into account when choosing the different variables that have a considerable effect on energy consumption. Income has been extensively discussed in the literature as a significant determinant of household energy consumption. The role of income in energy consumption was initially explained using the energy ladder model, which states that, as income rises, households shift from traditional to more sophisticated energy sources. Most studies have used Gross Domestic Product (GDP) to determine the link between income and energy use on a national or state level [21]. However, we contend that the relationship may differ at the household level. Therefore, this study analyzed the relationship between annual household income, energy consumption, and carbon emissions at the household level. The income of households is supposed to have a positive effect on the consumption of non-renewable energy and a negative effect on biomass energy consumption. High-income families, for example, are expected to use more non-renewable-energy sources than biomass. According to the literature, household energy consumption and carbon emissions are influenced not only by income, but also by a variety of other factors, such as spatiality [22,23,24]. These spatial factors include the type of housing unit, its size, the number of rooms, the distance from the LPG market, and the presence of a nearby fuelwood market. Therefore, these variables were also used in the analytical model, along with the income of households. For example, families owning large houses with many rooms prefer a cheaper and easily available energy source because the consumption of expensive and rarely available energy sources is difficult in a large house. Furthermore, demographic factors also influence the choice of energy sources. As a result, age, education, knowledge of clean cooking technologies, knowledge of clean energy sources, information on the impact of fuels on the environment, and knowledge of the health impact of biomass were all taken into account in this study. These factors play a major role in the selection of energy sources and their consumption. For example, a highly educated person would prefer renewable energy sources, in comparison to non-renewable-energy sources. Similarly, having the proper information regarding the impact of different energy sources on the environment and health would enhance the selection and consumption of clean energy sources at the household level. Apart from the factors discussed above, factors such as own land, own animal husbandry, and own sources of fuelwood, influence household energy-consumption levels and choices.

3. Material and Method

3.1. Study Area and Sampling

The study was conducted in rural areas of four randomly selected districts in the Punjab province in Pakistan. These districts were Rawalpindi, Khushab, Faisalabad, and Muzaffargarh, as shown in Figure 1. For the selection of district and household, a multistage random sampling technique was used in this study. The Punjab province has 36 districts. In the first stage, these districts were clustered into two agro-climatic zones, highland and lowland, to capture the potential effects of agro-climatic-related factors on household energy use. In the second stage, two districts from each zone were randomly selected. Consequently, Rawalpindi and Khushab were selected from the highland category, and Faisalabad and Muzaffargarh were selected from the lowland category. In the third stage, ten villages were selected from each district. Finally, 8–12 households were selected from each village for the sample survey. A semi-structured questionnaire was used to collect data from 400 randomly selected households.

To check the validity of the questions included in the questionnaire, a pre-testing survey was conducted. Based on the observation during the pre-testing phase, modifications were made to the questionnaire. A household was considered as a unit of analysis in this study because household energy is an issue of concern for the entire household. Therefore, respondents of the survey were household heads. Information on the different aspects of energy, such as monthly consumption, energy expenditures, types of energy, and technologies used, and information related to household composition, income, wealth, housing unit characteristics, and information about the pros and cons of different energy sources and their impact on the environment and human health were collected from the respondents. A team of data collectors with knowledge of the different crops cultivated in the area and about energy was recruited. Before entering the field, training was given to the data collectors on how to handle an interview. The overall process of the survey was supervised by the first author of this study.

The following Cochran (1963:75) sample size was determined by Equation (1):

$$\begin{gathered} N=z^{2}PQ/e^2 \\ N={\left(1.96\right)}^2·\left(0.5\right) \left(1-0.5\right)/{0.05}^2 \\ N=3.8416·\left(0.5\right)·\left(0.5\right)/0.0025 \\ N=384.16 \end{gathered}$$

(1)

where N is the required sample size, p is the estimated proportion of an attribute that is present in the population (assuming maximum variability at p = 0.5), Z² is the abscissa of the normal curve that cuts off an area α at the tails (1 − α equals the desired confidence level, e.g., 95%). The value for Z is found in statistical tables, 1.96 is the critical value for a two-tailed hypothesis test at 5% significance level, Q is 1 − p, and e the desired level of precision (e = 0.05 at 5%). The total sample size was calculated as 384.

3.2. Model Specification

Various models have been used to test the effect of different factors on energy consumption at the household level, for example, linear approximate almost ideal demand system [25], multinomial logit [26,27,28], the energy demand of utility maximization [29], dynamic panel regression model (Han et al., 2018) Engel curve [23], and Regression model [30]. Wang and Yang [31] used the STIRPAT model for the aforementioned purpose. In this study, the Stochastic impacts of regression on the Population (P), Affluence (A), and Technology (T) (STIRPAT) model was preferred due to its many advantages, such as simple model modification and inclusion of the variable, easy interpretation of the findings [32], and high flexibility [32]. Another reason for choosing the STIRPAT model was to explore the causal effects portrayed by the different drivers (income, household size, and clean energy) of energy consumption. The STIRPAT model is an extension of the IPAT model [33]. The IPAT (Impact = Population + Affluence + Technology) model was first proposed by Ehrlich and Holdren [34]. The basic IPAT model is expressed in the following Equation (2):

I = PAT

(2)

The above equation can be modified as follows:

$$I_i={\gamma }_o P_i^{{\beta }_1} A_i^{{\beta }_2}{T}_i^{{\beta }_3}$$

(3)

By augmenting the stochastic variables in Equation (3), the IPAT model can be converted into the STIRPAT model. Equation (4) depicts the STIRPAT model.

$$I_i={\gamma }_o P_i^{{\beta }_1} A_i^{{\beta }_2}{T}_i^{{\beta }_3}{e}^{{\epsilon }_i}$$

(4)

By taking the logarithmic on both sides of Equation (4), the impact of various factors on dependent variable was explored [33,35]. Finally, the energy-consumption STIRPAT model was developed as presented in Equation (5):

$$\ln I_i={\beta }_o+{\beta }_1\ln P_i+{\beta }_2\ln A_i+{\beta }_3\ln T_i+{\epsilon }_i$$

(5)

where ${\beta }_o \mathrm{is} \mathrm{equal} \mathrm{to}\ln {\gamma }_o$ and lne is equal to 1. This re-specification of the initial STIRPAT model is known as the elasticity model due to the coefficients associated with the variables in the model describing the elasticities. In the current study, we used the quantities of non-renewable and biomass energy (in Kg of oil equivalent) at the household level as dependent variables, and population (household size), affluence (income, number of animals, land owned, housing unit size, and number of rooms), and technology (use of clean energy sources, such as biogas or solar) and information about energy sources and their impacts, age, and education as independent variables. Therefore, ${\beta }_s$ are the coefficient values of the variables and their description is provided in

Table 1. Description of variables.

LnAge	Age of the household head was measured in years and then logarithm was performed according to STIRPAT modeling
Education	Education of household head was measured in years then logarithm was performed according to STIRPAT modeling
LnIncome	Total income (rupees) was incorporated for income effect upon energy-consumption choices, and logarithm was prformed for income elasticity and STIRPAT modeling
House type	Ordinal: Single family detached = 1, single family attached = 2, and other = 3
Ln housing unit size	Housing unit size was measured in Sq. ft and then logarithm was performed according to STIRPAT modeling
Ln no. of rooms	Rooms in a house were measured in numbers and then logarithm was performed according to STIRPAT modeling
Ln distance from LPG Market	Distance form LPG market was measured in Km and then logarithm was performed according to STIRPAT modeling
Fuelwood market in village	Dummy: 1 for availability of fuelwood market in village, otherwise 0
Info. about modernized cookstoves	Dummy: 1 for having information about the modern cookstove, otherwise 0
Info. about clean energy	Dummy: 1 for having information about clean energy, otherwise 0
Info. about fuel’s impact on environment	Dummy: 1 for having information about the impact of fuels on environment, otherwise 0
Info. about health impact of biomass	Dummy: 1 for having information about the health impact of biomass, otherwise 0
Use of any clean enrgy source	Dummy: 1 for using clean energy in the house, otherwise 0
Ln own land	Land was measured in Acres and then logarithm was performed according to STIRPAT modeling
Lnlivestock	Livestock was measured in numbers and then logarithm was performed according to STIRPAT modeling
Primary factor for choosing any fuel type	1 for convenience, 2 for efficiency, and 3 for price
Source of fuelwood	1 for “collecting from own land”, 2 for “collecting form others’ land”, and 3 for “purchased form market”

.

3.3. Energy Equivalents of Non-Renewable and Biomass Energy

To proceed with the empirical investigation, the physical consumption quantities of biomass and non-renewable energy were converted into standard units (kilogram of oil equivalents (Kgoe)). The energy equivalents of different energy sources presented in

Table 2. Energy equivalents and emission coefficients of different energy sources.

Biomass Fuel Type	Energy Equivalents (Kgoe)	Emission Coefficients (t CO2/t Fuel)	Non-Renewable-Energy Type	Energy Equivalents (KGOE)	Emission Coefficients (t CO2/t Fuel)
Crop residues (kg)	0.023	1.174 *	LPG	11.7	0.401
Fuelwood (kg)	0.3	0.03
Dung cake (kg)	2.16	0.787 *

Sources: [36,37]. * Emission coefficient for dung cake is kg CO₂/kg from Baul et al. [37].

were used to convert physical quantities into standard unit Kgoe. Biomass and non-renewable-energy consumption of households were calculated as follows:

Non-Renewable-Energy Consumption = LPG (Physical quantity) * 11.7 (Kgoe)

Biomass Energy Consumption = Fuelwood (Physical quantity) * 0.3 (Kgoe) + Crop Residuals (Physical quantity) * 0.023 (Kgoe) + Dung Cake (Physical quantity) * 2.16 (Kgoe)

Similarly, carbon emissions from different energy sources were estimated by using carbon emission coefficients of energy sources and described below:

LPG _(CO2) = LPG × CEC _LPG × 44/12

Fuelwood _(CO2) = Fuelwood × CEC _Fuelwood × 44/12

Crop-Residuals _(CO2) = Crop-residual × CEC _{Crop-Residual} × 44/12

Dung cake _(CO2) = Dung cake × CEC _{(Dung cake)} × 44/12

$$\mathrm{Total} \mathrm{Carbon} \mathrm{emission}={{\displaystyle \sum }}_{i=1}^{4}e{s}_i\times CE{C}_i\times \raisebox{1ex}{$44$}\!\left/ \!\raisebox{-1ex}{$12$}\right.$$

where CEC denotes the carbon emission coefficient, es refers to different types of energy sources, and i shows these five energy sources used in the study area (i = 1, 2, …, 5). The information about the carbon emission coefficient for non-renewable- and biomass energy sources is presented in Table 2. Before calculating the total carbon emissions, the units of crop residues and dung cake were first converted into t CO₂/t fuel.

4. Results and Discussion

4.1. Energy-Consumption Quantities

The consumption quantities of different energy sources in the study area are presented in

Table 3. Fuel types, physical quantities, and share in total consumption.

Energy Source	Physical Quantity of Consumption per Month	Coefficient of Kilogram of Oil Equivalent (Kgoe)	Consumption (Kgoe/Month)	Percent of Total Consumption
LPG (kg)	10.72	11.70	125.42	24.22
Fuelwood (kg)	296.93	0.30	89.08	17.20
Crop residual (kg)	446.25	0.02	10.26	1.98
Dung cake (kg)	93.76	2.16	202.52	39.11
Natural gas (MMBTU)	3.27	25.00	81.75	15.79
Electricity (kwh)	103.42	0.09	8.79	1.70

. Dung cake has the highest share of total energy consumption of rural households in the study area. On the other hand, when converted into standard energy units, electricity consumption has the least share (1.70%) in total household energy consumption. Another important source of biomass energy in the study area is fuelwood, which contributes 17.20% of total energy consumption. Households also use significant quantities of crop residue (446 kg per month). The higher use of biomass fuels (such as dung, fuelwood, and crop residue) can be attributed to the easy availability and inexpensive nature of these fuels in the study area. On the other hand, the frequent breakdown of electricity and high cost have led to just the essential use of electricity in the study area, such as cooling (mostly with fans) and lighting. Table 3 reveals that, in daily life, rural households rely more on biomass fuels, in comparison to modern (e.g., electricity and natural gas) or non-renewable fuels (LPG). A significant share of the population in the study area also used LPG for cooking. However, it was found that LPG was used as a secondary cooking fuel and the primary cooking fuel was fuelwood/crop residue in the study area. LPG is also an expensive fuel and was not affordable for low-income households in the study area. A natural gas (piped gas) infrastructure is almost non-existent in rural areas, with the exception of a few. Therefore, natural gas consumption in rural households is almost negligible.

A higher use of biomass energy by households was also found by Zhou et al. [38]. Wang et al. [39] explored the consumption of different sources of energy in 8 different counties of China and found that biomass energy sources have the highest share of total household energy consumption. The higher share of biomass energy in total energy consumption is widely observed worldwide [40,41]. Wu et al. [42] also determined that biomass is one of the mainly consumed energy sources in China.

4.2. Association between Income, Household Size, and Energy Consumption

The association between household size and energy consumption (source-wise) is illustrated in Figure 2. The figure shows the variations in energy consumption in relation to household size. The biomass consumption increases as the family size increases, and LPG share decreases as family size increases. Biomass is the main source of energy, contributing 43.04% of total energy consumption in a household with more than 12 members, whereas, on the other hand, LPG has the highest share (43.05%) in total energy consumption of households with small family sizes (< 5 members). It can be concluded that household size (population) has a negative association with non-renewable energy (LPG) and a positive association with biomass fuels.

Figure 3 represents the relationship between income and the share of different energy sources in total energy consumption. The figure reveals that low-income households rely more on biomass energy and less on non-renewable energy. Looking at other sources, we can observe that the increasing income share of modern energy sources (mainly electricity) increases in total energy consumption. The share of LPG in total household energy consumption increases until the 4th income quantile. The richest population replaces LPG with electricity for cooking. Therefore, in the 5th income quantile, the share of LPG is low compared to the 3rd and 4th income quantiles. Similar findings were presented by Cai and Jiang [43], where they stated that biomass consumption increases with a large family size and low-income levels.

4.3. Association between Energy Sources and CO₂ Emissions

The sources of electricity are complex in Pakistan. Therefore, our analysis of carbon emissions from different energy sources is limited to biomass fuels and non-renewables (LPGs). The biomass contribution to the emissions of CO₂ is 15.355 kg, according to the coefficient value of the trend line depicted in Figure 4. This means that an increase of 1 kg in biomass consumption generates 15.355 kg of CO₂. It may be concluded that biomass consumption is a major contributor to CO₂ emissions, which may lead to a harmful environment in the locality considering the concentration of biomass fuels in the study area.

This emission of CO₂ could be reduced by increasing the consumption of other energy sources that emit less CO₂. An alternative to biomass is LPG, which is being used in the study areas as secondary cooking fuel. The trend line of CO₂ and LPG (non-renewable energy) shown in Figure 5 shows that the increase in consumption of non-renewable energy by 1 kg produces only 0.8675 kg of CO₂, which is comparatively lower than biomass and 1 kg of LPG has more heat than 1 kg of biomass sources.

4.4. Association between Household Size, Income, and Carbon Emissions

Figure 6 shows the association between CO₂ emissions and household size. The results show that, as the household size increases, the CO₂ quantity also increases. The highest level of CO₂ emissions was observed in households with a size of more than 12 family members. The lowest quantity of CO₂ was observed in households with less than 5 family members. It may be that the largest family-sized households need more energy, and they use more biomass than other sources of energy, as depicted in Figure 2. This result is striking for Pakistan, considering the average family size in the country.

Figure 7 presents the association between CO₂ emissions and income. It can be observed from the figure that wealthy households emit low quantities of CO₂, when compared to poor households. This may be because highest-income-quantile households consume more clean energy sources, as described in Figure 3; this is why the emissions of CO₂ were significantly reduced. The rich household used more clean energy sources, which led to low CO2 emissions, when compared to the low-income families. The higher use of biomass energy by households is a threat to the environmental sustainability of the country. Similar to its neighboring country (India), Pakistan can also roll out a program to replace the use of traditional biomass with LPGs or other modern/renewable sources of energy for cooking. It has been stated by the previous studies that biomass combustion emits several hazardous air pollutants [44] and majorly contributes to environment degradation [45].

4.5. Results of the STIRPAT Model

The results of the STIRPAT model are presented in

Table 4. Results of STIRPAT model.

Variables	Non-Renewable Energy	Biomass
	Coef. (Std. Error)	Coef. (Std. Error)
Constant	−53.065	48.621
	(26.146)	(36.650)
LnAge	−0.150	0.282
	(0.341)	(0.478)
Education	3.216 *	−8.231 *
	(1.093)	(2.345)
LnIncome	0.010 *	−0.012 *
	(0.002)	(0.001)
Ln housing unit size	−0.004 *	0.005 **
	(0.002)	(0.003)
Ln distance from LPG market	−1.621 *	1.082 ***
	(0.464)	(0.651)
Fuelwood market in village	6.477	17.964 *
	(8.281)	(6.567)
Info. about modernized cookstoves	−4.636 *	−4.628 *
	(1.365)	(1.098)
Info. about clean energy	−12.237 *	−4.112 *
	(3.276)	(0.987)
Info. about fuel’s impact on environment	−2.602 **	−7.071 *
	(1.004)	(0.789)
Info. about health impact of biomass	2.380 *	−4.702 *
	(0.684)	(1.009)
Use of any clean energy source	−26.096 *	−20.243 ***
	(8.062)	(11.301)
Ln own land	−0.801	1.597
	(0.790)	(1.108)
Lnlivestock	−0.105 **	0.053 *
	(0.057)	(0.012)
Primary factor for choosing any fuel type	−0.293 *	0.267 *
	(0.064)	(0.099)
Source of fuelwood	10.585 *	−5.267 *
	(2.586)	(1.056)

*, **, *** depicts the significance at 1%, 5%, and 10%, respectively. Figures in parenthesis are standard errors.

. The impact of different independent variables, such as population, affluence, and technology, on the households’ non-renewable and biomass energy consumption was analyzed. The results reveal that education has a significant positive impact on NRE consumption and a significant negative impact on biomass energy consumption. This explains how education plays an important role in the selection of different energy sources. Poortinga et al. [46] and Nair et al. [47] (2010) also found the significant effect of education on energy use.

The econometric results regarding the logarithmic income of the family describe a significant positive association with NRE sources. It shows that LPG (NRE source in the current study) is considered as a positive necessity by households, and if income increases by 1%, the consumption of NRE also increase by 0.010%. Similarly, the significant negative impact of income on biomass consumption determines that a 1% increase in income reduces the consumption of biomass energy (such as firewood, dung cakes, and crop residuals) by 0.012%. Cai and Jiang [43] also reported the same result that a high income reduces the consumption of biomass energy and increases the consumption of non-renewable energy, such as LPGs. Moreover, Yousaf et al. [6] also presented similar results regarding the consumption of LPGs.

Housing unit size also affects the choice and consumption of different energy sources. A 1% change in the size of a housing unit (sq. Ft) reduces the consumption of NRE by 0.004% and increases the consumption of biomass by 0.005%. A large housing unit means more room, more inhabitants, and higher energy needs for cooking, cooling, and heating. Moreover, a large household size relies more on biomass for cooking and heating. Zou and Luo [44] reported the positive, but insignificant, impact of dwelling size on non-renewable-energy consumption. However, their results describe a significant positive effect of dwelling size on the consumption of biomass energy.

The distance from the LPG market and availability of the fuelwood market in the village also play vital roles in energy consumption. Near-the-market households consume more NRE and less biomass energy. A 1% decrease in the distance from the LPG market increases the consumption of LPGs by 1.621% and reduces the consumption of biomass energy by 1.082%. Behera et al. [48] also described similar results that determined that the easy access to the market increases the consumption of LPGs and the lengthy distance from the market increases the likelihood of the consumption of biomass. It was found that the market of fuelwood in the village did not affect the consumption of NRE, but it significantly affected the consumption of biomass. The fuelwood-market availability increased biomass consumption by 17.964%.

The awareness of and information about different technologies and the impact of fuels on the environment and health also affect the use of NRE and biomass energy. Information about improved cookstoves (ICSs) significantly reduced the consumption of NRE (4.636%) and biomass energy by 4.628%. The reduction in LPG use was because, after realizing that biomass can be used with clean and efficient technologies, households preferred biomass over LPG; whereas the reduction in biomass use was due to its efficient burning with ICS. Similarly, the information about clean energy sources also significantly reduce the consumption of NRE and biomass by 12.23% and 4.11%, respectively. Having the information of the negative impacts of different energy sources on the environment reduces the consumption of biomass by 7.071% and increases the use of NRE by 2.602%. Similarly, having information about the impact of biomass on health reduces the consumption of biomass by 4.702% and increases the consumption of LPGs by 2.380%.

The value associated with the use of clean energy sources is negative and it ensures that the adoption of clean energy sources reduces the consumption of LPGs by 26.096% and the consumption of biomass by 20.243%.

Livestock animals are the source of animal dung, which has highest share of total energy consumption in the study area. The number of livestock on the farm significantly reduces the consumption of NRE and increases the consumption of biomass. A 1 percent increase in livestock animals on the farm leads to a 0.105% reduction in NRE consumption, and a 0.053% increase in consumption of biomass. A household that considers the price of fuel as a primary factor for choosing the fuel type has an NRE consumption that is 0.293% lower and biomass that is 0.267% higher than the household for which price is not a determining factor for the choice of fuel. For households that purchase fuelwood from the market, their consumption of biomass is 5.267% less than the households that acquire it from their fields or collect it from others’ fields. Similarly, their consumption of NRE is 10.585% higher than others. This may be because, rather than purchasing fuelwood from the market, the household prefers to purchase NRE (LPG) and substitute wood with NRE.

5. Conclusions

Rural households in the study area relied on a variety of energy sources, such as crop residue, animal dung, fuelwood, LPG, natural gas, electricity, solar, and biogas. The results of the study reveal that biomass has the largest share of total household energy consumption. Due to easy access to biomass energy, such as fuelwood, crop residues, and animal dung, households in rural areas have a very low reliance on modern/commercial energy sources. Income and household size play an important role in the choice of energy source and consumption. Low-income and large households consume more biomass energy and produce higher carbon emissions. On the other hand, wealthy and small households consume more non-renewable energy and produce low carbon emissions.

Factors, namely, education, income, household size, dwelling size, number of livestock animals, distance from LPG market, availability of LPG and fuelwood in the locality, information about clean energy and negative impacts of biomass burning on the environment and health, and use of clean energy sources, have a significant impact on the consumption of NRE or biomass energy. The consumption of biomass can be reduced by increasing the awareness about the negative environmental and health impacts associated with the use of biomass and the benefits of using clean energy. A reduced family size has a substantial impact on the consumption of biomass energy and carbon emissions. Policies should be designed and vigorously executed to control the ever-increasing population. Moreover, ensuring the easy and cheap access to NRE, and modern or renewable energy in rural areas will significantly reduce the use of biomass, which will lead to environmental sustainability and improved health.