Health and Heating in the City of Temuco (Chile). Monetary Savings of Replacing Biomass with PV System in the Residential Sector

The paper conducts a comprehensive analysis of replacing residential use of wood burning stoves for heating with photovoltaic systems for the generation of electricity using storage batteries (PV + storage systems). The research focuses on the city of Temuco (Southern Chile) as a case study, since this city has a high use of firewood for heating (80% of households) and also because this city has been declared by authorities as an area saturated with suspended particles. The total cost corresponds to the acquisition of systems. The reduction of monetary value of the impact of polluting emissions, resulting from the combustion of firewood, on health and traffic accidents is calculated. The interactive tool GDB Compare has been used to calculate the impact of pollution on health both in terms of attributable deaths and disability-adjusted life years. The monetary value of the impact on health has been calculated using two alternative approaches: the value of statistical life and the human capital approach. To identify firewood use requirements, heating degree–days has been used for temperatures ≤15 ◦C and ≤18 ◦C. The emissions avoided calculations have been refined, including emissions associated with the manufacture and transport of systems through the life cycle analysis. For all scenarios, the main results show that the savings outweigh the costs.


Introduction
The residential sector in Chile is highly intensive in the use of firewood for heating. In the mid-1990s, 85% of urban households consumed firewood [1]. Currently, 80% of households in the city of Temuco, in which this article is focused, use wood for heating [2]. In 2013 (latest data available), the total consumption of firewood in Chile was 56,219.5 teracalories (Tcal), of which 62.6% corresponded to the residential sector [3]. To meet this demand, 10,074.1 tons of firewood were required, causing environmental stress on the Chilean forest area, acting against a sustainable model to meet a comfortable standard of living in the residential sector. This area represents 22.9% of the total area of the country (17.3 million hectares) and acts as a large greenhouse gas (GHG) emission sink [4] to sustainability. However, it might be borne in mind that the pressure on the forest is also due to its use as fodder and not only the extraction of firewood.
illustrates how a PV + storage system replaces the use of a wood burning stove, thus, reducing the emission of pollutants. The aim of this paper is to assess monetary saving on illness and traffic accidents due to replacing domestic wood burning stoves for heating with domestic pollutant emission-free PV + storage systems. Monetary savings derive from the mitigation of pollutant emissions due to the reduced burning of wood, the reduced deforestation, and the decreases in both health damage and traffic accidents. When calculating benefits, CO2 emissions stemming from manufacturing and transport might be considered [15].
Therefore, the impact of air pollution from fossil fuels and ambient particulate matter on health is assessed taking DALYs and attributable deaths as the central topics from the output provided by the interactive tool GDB Compare [16]. The monetary value is calculated by following two complementary approaches; the value of statistical life and the human capital approach. Due to availability of data, 2013 is used as the reference year.
Despite the use of DALYs in the field of Health Economics, its use in fields such as Environmental Economics or Energy Economics has been limited; thus, this is one of the novelties of this paper. To the best of our knowledge, no previous research exists conducting a similar analysis of introducing clean nonconventional technology (PV + storage systems) replacing the conventional use of biomass for heating based on the health impact and the absence of CO2. This paper contributes to filling this gap in the literature. Additionally, this paper does not adopt the assumption of extensive literature establishing biomass combustion as neutral in terms of CO2. Alternatively, we use a specific emissions factor for firewood combustion and a lower factor for establishing CO2 through photosynthesis.

Materials and Methods
A description of the PV + storage system is provided in Section 2.1. After this, a method used for the monetary valuation of health damage is described. This section ends by explaining the monetary valuation of emissions avoided. Data sources are also detailed before the results.

PV + Storage System Characteristics
The PV systems installed in houses feed electrically driven heat pumps. The system is optimised when batteries are incorporated to store the electricity generated but not consumed instantaneously ( Figure 1). When this replacement is made in a house, emission of pollutants derived from the combustion of firewood is prevented. However, due to installation requirements, not all homes enable the installation of a standard PV + storage system. The aim of this paper is to assess monetary saving on illness and traffic accidents due to replacing domestic wood burning stoves for heating with domestic pollutant emission-free PV + storage systems. Monetary savings derive from the mitigation of pollutant emissions due to the reduced burning of wood, the reduced deforestation, and the decreases in both health damage and traffic accidents. When calculating benefits, CO 2 emissions stemming from manufacturing and transport might be considered [15].
Therefore, the impact of air pollution from fossil fuels and ambient particulate matter on health is assessed taking DALYs and attributable deaths as the central topics from the output provided by the interactive tool GDB Compare [16]. The monetary value is calculated by following two complementary approaches; the value of statistical life and the human capital approach. Due to availability of data, 2013 is used as the reference year.
Despite the use of DALYs in the field of Health Economics, its use in fields such as Environmental Economics or Energy Economics has been limited; thus, this is one of the novelties of this paper. To the best of our knowledge, no previous research exists conducting a similar analysis of introducing clean nonconventional technology (PV + storage systems) replacing the conventional use of biomass for heating based on the health impact and the absence of CO 2 . This paper contributes to filling this gap in the literature. Additionally, this paper does not adopt the assumption of extensive literature establishing biomass combustion as neutral in terms of CO 2 . Alternatively, we use a specific emissions factor for firewood combustion and a lower factor for establishing CO 2 through photosynthesis.

Materials and Methods
A description of the PV + storage system is provided in Section 2.1. After this, a method used for the monetary valuation of health damage is described. This section ends by explaining the monetary valuation of emissions avoided. Data sources are also detailed before the results.

PV + Storage System Characteristics
The PV systems installed in houses feed electrically driven heat pumps. The system is optimised when batteries are incorporated to store the electricity generated but not consumed instantaneously ( Figure 1). When this replacement is made in a house, emission of pollutants derived from the combustion of firewood is prevented. However, due to installation requirements, not all homes enable the installation of a standard PV + storage system. Additionally, it must be taken into account that when the surface area of the house is very large, the heating capacity of a standard PV + storage system may be insufficient. Currently the most widely used is Solar Kit 5000 W, 48 V, with a capacity to generate 16,000 kWh/day. The main characteristics are summarized in Table 1. A lifetime of 20 years is assumed. The PV + storage system chosen for this research took into account specific characteristics of the city of Temuco. The optimal number of panels is determined by applying Equation (1): Week days in f orce 7 PSH × µ × Installed capacity per panel (1) where energy consumption is the total electricity consumption required by the heating splits. We assume 2 splits per home working 6 hours per day during the whole week. PSH represents the peak solar hours. Following the tool provided by the Chilean Department of Energy [15], the global climbed annual radiation for the city of Temuco is 4.71 kWh/m 2 /day. As colder months register low values, the storage system allows for bridging of this issue. The PSH for autumn and winter months varies from 4.81 (March) to 1.56 (June). Table A1 in Appendix A detailed the PSH for the city of Temuco per month and per hour of the day. µ is the loss coefficient. We assume a value of 0.6, meaning 40% of losses because of panels, inverter device, batteries, and accumulators. The optimal number of batteries is determined by Equation (2): where 5 days of use of batteries are considered. Tables A2 and A3 contain data of shadow and clouds in the city of Temuco for months and daily hours. DoD is the depth of discharge for batteries (50% in the case of lead batteries), the tense assumed is 12 V, and a 15% loss coefficient is taken into account. The PV + storage system is switched on when outside temperature falls under 18 • C or 15 • C. These cold days are the so-called heating degree-days (HDDs). The technical requirements for the installation and its heating capacity make it optimal for homes whose habitable surface area varies between 51 to 70 m 2 . The number of Temuco homes included within this range is 12,388, representing 41.7% of total housing. The HDD is a measurement designed to quantify the demand for energy needed to heat a building. The HDD is derived from measurements of outside air temperature. We assume that standard comfortable heating is necessary when the outside air temperature falls below 18 or 15 degrees Celsius. Both levels of temperature are considered in order to provide a sensitive analysis.

Monetary Valuation of Health Damage
The interactive tool GDB Compare provides information on two types of health damaging pollution. These are household air pollution from fossil fuels and ambient particulate matter pollution. We assume that firewood combustion causes indoor CO 2 emissions that must be considered as part of "household air pollution from fossil fuels"; despite the fact that firewood is not a fossil fuel, it acts as an energy resource similar to other fossil fuels. Together with firewood combustion, it causes ambient particulate matter pollution.
The disability-adjusted life year (DALY) together with attributable deaths are two extensive health gap measures. The DALY concept was jointly developed for the Global Burden of Disease (GBD) study by the World Bank, the Harvard School of Public Health, and the World Health Organization (WHO) in the late 1980s [19][20][21][22][23]. DALY measures both mortality and morbidity and combines them into one single figure providing an evidence-based tool for healthcare policy prioritization and for monitoring intervention effects [24]. The interactive tool GDB Compare provides information for an extensive list of risk disease factors in terms of DALYs and attributable deaths [16].
Several studies [25][26][27] detail the methods for calculating the monetary valuation of health impacts from risk factors. The literature available offers two main methods that can be used to value health impact in monetary terms. The first is based on the willingness to pay (WTP) for a statistical life year. This relies on asking for the maximum amount of money an individual would be willing to give up in order to avert the loss of a year of life. The WTP approach is also named as the criterion of the value of statistical life (hereafter, VSL). VSL is a measure of the cost of mortality derived from the aggregation of the individual willingness to pay (WTP) in a society [10,28].
This takes into account the number of deaths attributable to a risk factor. VSL is a narrow approach to the consequences of a risk factor on health. The interactive tool GDB Compare provides its output not only in terms of DALYs by risk factor, but also in terms of deaths attributable to different risks. Consequently, deaths attributable to ambient particulate matter pollution and household air pollution from solid fuels are also valued in economic terms by gender by multiplying them by the VSL value for Chile up to 2013, as Equation (3) details: where GBD i measures the GBD attributable to i (ischemic heart disease, stroke, tracheobronchial and lung cancer, lower respiratory infections, chronic obstructive pulmonary disease, and diabetes mellitus. R j is the risk factor of pollutant j (ambient particulate matter pollution and household air pollution from fossil fuels). P is the population of the city of Temuco in 2013 and VSL is measured in USD 2013. The second method of calculating the monetary valuation is lost output or the human capital approach (HCA). By measuring an individual's yearly economic contribution as the GDP per capita, years lost due to disability or death are years that are not productive. Following the HCA approach, the monetary value of DALYs is calculated by multiplying a standardised GDP value per capita by the total number of DALYs. DALYs are the sum of the years of life lost due to premature mortality in the population and the years lost due to disability for incidents of disease or health conditions [29]. This approach was criticised in the literature [26]; thus, a refined version of HCA considering health expenditure provides more specific findings. This is our second approach estimating the monetary value of diseases caused by ambient particulate matter pollution and household air pollution from fossil fuels. In this case, we multiply the number of DALYs per GDP net health expenditure per capita in Chile (both current and capital) measured in USD 2013, as Equation (4) shows [27].

Monetary valuation HCA
where GDP pc being the Chilean gross domestic product per capita in 2013, CHE pc the Chilean current health expenditure per capita, and KHE pc the Chilean capital health expenditure per capita. All of the monetary variables are measured in USD 2013. Both criteria described above must be used with caution, not attributing all fossil fuel emissions and ambient particulate matter to the combustion of firewood. To determine the portion of energy consumption of Chilean households corresponding to firewood, the climatic diversity of the country must be taken into account. An exhaustive analysis was based on a sample of urban households corresponding to three cities: Metropolitan Area (zone 2), Valparaiso (zone 3), and Concepción (zone 4) [30]. This last city is the most similar to the climate of Temuco (zone 5) [14]. For the city of Concepción, 51% of energy consumption in a typical dwelling was for heating (4776 kWh/year) [30]. The source was firewood. We conservatively assumed that 51% of the impact on health caused by household air pollution from fossil fuels is due to combustion of firewood. This corresponds with ω 2 in Equations (1) and (2).
Regarding the portion of impact on health that corresponds to ambient particulate matter emitted by the combustion of firewood, the averages for PM2.5 emissions due to this combustion (94%) and PM10 emissions (82%) [2] have been considered. This corresponds with ω 1 in Equations (1) and (2).
Finally, as ambient particulate matter pollution also increases traffic accidents in cities like Temuco due to thermal inversion, a monetary valuation of victims is carried out. The authorities of Temuco requested that this city be declared as an area saturated with ambient particulate matter and, thus, PM10 was recognised in 2005 and PM2.5 in 2013 [2,28]. In 2018, the Air Decontamination Plan for PM 2.5 took effect.
The monetary values of car accidents was estimated by multiplying the number of victims in the city of Temuco attributable to ambient particulate matter pollution by internal and external cost [30][31][32]. The authors distinguished between fatalities, serious injuries, and minor injuries. They also differentiated between internal and external costs. Internal costs include medical expenses, material damage, and also WTP for reducing the risk of accidents. External costs include property damage to third parties, health care expenses, police expenses, administrative and judicial expenses, and the current net value cost to the state treasury for loss of contributions of the victims. Equation (5) was used to calculate the monetary values of car accidents The number of damaged in Equation (5) includes fatalities, serious injuries, and minor injuries. Both internal and external costs are measured in USD 2013 from [33].

Monetary Valuation of Emissions Avoided
Due to its greater accessibility, the type of firewood most used in the city of Temuco is eucalyptus firewood. Emission factors have been assigned to calculate CO 2 emissions and ambient particulate matter. The emission factors for eucalyptus firewood is 129.1 grs CO 2 /kg and 10.9 grs PM10/kg [34]. The operation condition considered in the paper to determine the emission factor assumes that the combustion is done with a half and closed chimney flue. The type of firewood considered was dry firewood (Hd. b.s. < 25%). To determine the heat capacity of the wood, its density (520 kg/m 3 ) [35,36] and electricity generation capacity (4.6 kWh/gr) [37] were considered.
The heating needs of households were calculated based on the HDD approach. Wood requirements for heating are different depending on the size of the home considered. This paper analyses the housing of Temuco, classifying it in a range from dwellings with an area of less than 35 m 2 to those with a surface area of more than 140 m 2 (see Table A9 in the annex).
Along with the polluting emissions derived from the combustion of firewood, the felling of trees required to supply the firewood reduces the capacity of the forest surface area to act as a sink. To calculate the amount of CO 2 not fixed due to the felling of eucalyptus, its fixation rate was used. This factor was 3.67 grs CO 2 /kg [34].
This approach is different from what the literature usually indicates. It is normally considered that the combustion of biomass is neutral in terms of emissions, given that the CO 2 emitted is equivalent to that fixed by photosynthesis [38]. In this paper, we used a burning emission factor (129.1 grs CO 2 /kg) greater than the fixation rate for photosynthesis (3.67 grs CO 2 /kg). This methodological approach is more stylised than that usually used in literature and better expresses the effect of environmental stress caused by the felling of trees for use as firewood for heating. Our approach is in line with the fact that in 2002 the region of Araucanía (region IX of Chile, to which the city of Temuco belongs) lost its status as a net sink due to the significant forest area lost as a result of fires [39].
The installation of PV + storage systems enables calculating emission reduction by comparing two scenarios. The baseline scenario considers that 80% of households included in the surface area range of 51 to 70 m 2 use wood burning stoves with eucalyptus (Eucalyptus globulus) as firewood [40]. In the final scenario, it is assumed that 100% of these households replace the stoves with PV + storage systems, avoiding pollutant emissions. By comparing the total emission volume of both scenarios, it is possible to calculate emission reduction coefficients (both household air pollution from fossil fuels and ambient particulate matter). These reduction coefficients will reduce the economic value of the impact of polluting emissions on health and traffic accidents.
PM Reduction coe f f icient = A stylised analysis of avoided CO 2 emissions requires a calculation of the CO 2 emissions associated with the manufacture, transport, and installation of PV + storage systems using life cycle analysis (LCA) [41][42][43]. The Results section details the LCA. Figure 2 summarises the main sequence of methodological steps described above.

Data
The data on the consumption of firewood in Chile were taken from the national energy balances that provided this data until 2013. As of that year, the series is interrupted, with no data provided for subsequent years despite having been requested. For the period 1991-2013, Figure 3 shows that the trend of wood consumption in Chile was growing and was linked to population growth.

Data
The data on the consumption of firewood in Chile were taken from the national energy balances that provided this data until 2013. As of that year, the series is interrupted, with no data provided for subsequent years despite having been requested. For the period 1991-2013, Figure 3 shows that the trend of wood consumption in Chile was growing and was linked to population growth. The interactive tool GDB Compare was used to calculate the impact on health from suspended particles and pollution from fossil fuels [16]. GBD Compare is an interactive tool to analyze updated estimates of the world's health for 359 diseases and injuries and 84 risk factors from 1990 to 2017. The GBD Study 2017, coordinated by the Institute for Health Metrics and Evaluation [16], estimated the burden of diseases, injuries, and risk factors for 195 countries and territories, and at the subnational level for a subset of countries. Measures available include attributable deaths and disability-adjusted life years (DALYs) used in this paper. Annual population estimates by single year of age and sex for 1950-2017 are available at the IHME [44]. Annual specific fertility rate (ASFR) estimates by 5-year age groups in the 10-54 year range for 1950-2017 are available at the IHME [45]. Via the "Files" tab above, select tables published by GBD [46] are available for downloading in this record.
The data taken in this paper, corresponding to 2013, are differentiated by sex and risk factors. Two types of data are used: (i) the number of DALYs according to disease and risk factor and (ii) the number of deaths attributable to each risk factor.
The number of deaths and injuries caused by traffic accidents came from the police (called Carabineros) registries' annual report for 2013 [47]. The inflation rates in terms of variation of the Consumer Price Index, population data, and exchange rates were taken from the National Statistics Institute and the Central Bank of Chile, respectively.
The specific Chilean VSL came from the VSL value estimated [48] for 2010, adjusting the inflation rates and applying the purchasing power parity specified in this report. It was assumed that the specific value for Chile's VSL could be a proxy for the value of the most similar country in terms of PPP GDP per capita. This was Croatia [49,50], with a value of 2.07 million USD. This value has been updated as of 2013 and is considered representative for Temuco.
The housing data were provided by a statistical unit associated with the Ministry of Housing and Urban Planning [51] and the HDD data were provided by the meteorological network of Chile (METEORED)[52].

Results
Major findings are offered in the next two subsections. Firstly, monetary impacts on health were calculated following the two alternative approaches described in Section 2 (WTP and HCA). Then, the monetary savings from emissions avoided were also obtained. The interactive tool GDB Compare was used to calculate the impact on health from suspended particles and pollution from fossil fuels [16]. GBD Compare is an interactive tool to analyze updated estimates of the world's health for 359 diseases and injuries and 84 risk factors from 1990 to 2017. The GBD Study 2017, coordinated by the Institute for Health Metrics and Evaluation [16], estimated the burden of diseases, injuries, and risk factors for 195 countries and territories, and at the subnational level for a subset of countries. Measures available include attributable deaths and disability-adjusted life years (DALYs) used in this paper. Annual population estimates by single year of age and sex for 1950-2017 are available at the IHME [44]. Annual specific fertility rate (ASFR) estimates by 5-year age groups in the 10-54 year range for 1950-2017 are available at the IHME [45]. Via the "Files" tab above, select tables published by GBD [46] are available for downloading in this record.

Impacts on Health
The data taken in this paper, corresponding to 2013, are differentiated by sex and risk factors. Two types of data are used: (i) the number of DALYs according to disease and risk factor and (ii) the number of deaths attributable to each risk factor.
The number of deaths and injuries caused by traffic accidents came from the police (called Carabineros) registries' annual report for 2013 [47]. The inflation rates in terms of variation of the Consumer Price Index, population data, and exchange rates were taken from the National Statistics Institute and the Central Bank of Chile, respectively.
The specific Chilean VSL came from the VSL value estimated [48] for 2010, adjusting the inflation rates and applying the purchasing power parity specified in this report. It was assumed that the specific value for Chile's VSL could be a proxy for the value of the most similar country in terms of PPP GDP per capita. This was Croatia [49,50], with a value of 2.07 million USD. This value has been updated as of 2013 and is considered representative for Temuco.
The housing data were provided by a statistical unit associated with the Ministry of Housing and Urban Planning [51] and the HDD data were provided by the meteorological network of Chile (METEORED) [52].

Results
Major findings are offered in the next two subsections. Firstly, monetary impacts on health were calculated following the two alternative approaches described in Section 2 (WTP and HCA). Then, the monetary savings from emissions avoided were also obtained.

Impacts on Health
Following the WTP criteria approach through the VSL, total deaths in 2013 for males was 33.27 per 100,000 inhabitants with a range of 20.42-43.95 and 26.90 for females (15.54-36.94). The number of Temuco's inhabitants considered for 2013 was 265,101. The VSL value was taken from the WHO [48]. The VSL considered was 2.07 million USD 2010 per death. This value was updated as of 2013. It should be taken into account that only 51% of the impact on health from household air pollution from fossil fuels was considered due to the percentage of firewood consumption out of total energy consumption. Similarly, 88% of ambient particulate matter pollution impact on health was attributed to firewood combustion. The total monetary value of deaths for males was 13.7 M (USD 2013) and 16.3 M (USD 2013) for females. The highest impact of ambient particulate matter pollution on health was that for chronic obstructive pulmonary disease in the case of males, 2.7 M (USD 2013). For females, the highest impact of ambient particulate matter pollution was for ischemic heart disease, 3.3 M (USD 2013). When focusing on household air pollution from fossil fuels, the highest impact for males also came from chronic obstructive pulmonary disease, 0.18M (USD 2013), and again ischemic heart disease for females 0.38M (USD 2013). Tables A4 and A5 in the appendix detail the results obtained.
When using the HCA approach, the monetary value of DALYs lost due to ambient particulate matter pollution and household air pollution from fossil fuels was estimated by multiplying the estimated DALYs associated with different diseases by GDP net health expenditures per capita (including current and capital expenditures) [26]. When using HCA criteria, the total DALYs attributed to these risk factors in 2013 was 11,605.71 per 100,000 inhabitants for males with a range of 10,631.65-12,755.12 and 7971.75 for females (7092.28-8966.85). The GDP per capita considered was 15,790 USD, while current health expenditure per capita was 1173 USD and capital health expenditure per capita was 43 USD. The total monetary value of total DALYs for males was 12.3 M (USD 2013) and 7 M (USD 2013) for females. Compared with the VSL approach, the results obtained differed mainly in the case of females. The highest impact of ambient particulate matter pollution on health was on ischemic heart disease, 3.3 M (USD 2013) for males and diabetes mellitus for females, 2.1 M (USD 2013). When we focused on household air pollution from fossil fuels, the highest impact also appeared to be from ischemic heart disease in the case of males 0.2 M (USD 2013) and again from diabetes mellitus for females 0.2 M (USD 2013). Tables A6 and A7 in the annex detail the results obtained.
To calculate the monetary value of accidents, we grouped victim data from police reports into three categories (fatalities, serious injuries, and minor injuries). Internal and external costs per death or injury [33] were converted into 2013 USD. The authors used Unidad de Fomento (UF) as their monetary unit. The UF is an accounting unit used in Chile. The exchange rate for 2013 was 1 UF equal to 23,309.56 CLP (Chilean currency). The official exchange rate of CLP to USD was used (495 CLP/USD). As police reports offer no detailed information for accidents in the city of Temuco, data was estimated following previous studies [53], from data corresponding to the region of Cautín, to which the city of Temuco belongs. The authors estimated that, for the case of Chile, 43% of fatalities, serious injuries, and minor injuries occurred in urban areas. For the rest of the accidents they estimated that 84% of total accidents occurred in urban areas. Table A8 in the annex shows the results. Based on interviews with traffic authorities, it was assumed that accidents caused due to smog were the result of particulate matter emission. Following our assumption, 88% was attributable to firewood combustion. The total monetary value of accidents attributable to particulate matter pollution was 76.7 M (USD 2013) (72.3 M internal costs and 4.3 M external costs). Table A9 shows the emission of pollutants and unfixed CO 2 emissions due to logging for the use of eucalyptus firewood. Comparing the baseline scenario (Tables A4-A7) with the final scenario, as described in the methodology section, we obtained the reduction factors shown in Table 2. Higher values when considering HDD ≤ 18 • C derive from a reduction in firewood requirements. As Table A9 shows, the total number of HDD ≤ 18 • C was 17.64% higher than HDD ≤ 15 • C. From [39][40][41], it was assumed that the carbon footprint of the PV + storage system considered was equal to one year of total emissions avoided. Based on interviews with installation companies, it should be considered that most systems installed in Temuco are manufactured in China. The sea route considered was Shanghai-Valparaiso. The assumed useful life was 20 years. Our assumption was based on the meta-survey by [54].

Monetary Savings from Emissions Avoided
This assumption was a conservative one as part of the literature cited by [54] considers a 30 year lifetime period. As ambient particulate matter is a local pollution problem (different from CO 2 , which impacts on global warming), LCA does not modify its reduction factors.
By applying the emission reduction rate to the monetary values of health impact and accidents, it was possible to identify the reductions of these costs derived from replacing wood burning stoves with PV + storage systems. These reductions are the benefits of the cost-benefit analysis carried out. Tables 3-6 show the results.    When HDD ≤ 15 • C was used as reference, the monetary value of the impact on health was reduced by 5.5 M USD 2013 for males and 6.5 M USD 2013 for females. When HDD ≤ 18 • C was used as a reference, these figures were 9.4 and 11.2 M USD 2013.
Taking into account HCA criteria, when HDD ≤ 15 • C was used as reference, the monetary value of the impact on health was reduced by 4.9 M USD 2013 for males and 2.8 M USD 2013 for females. When HDD ≤ 18 • C was used as a reference, these were 8.4 and 4.8 M USD 2013.
When CO 2 emissions in PV + storage systems LCA were considered, the results in Tables 3 and 4 must be reduced. Table 5 summarizes the major findings.
Finally, ambient particulate matter pollution prevented due to the replacement of firewood reduced traffic accidents and their corresponding monetary value. The details can be seen in Table 6. When cool days were defined as ≤15 • C, total saving amounted to 30.93 M USD 2013, while 36.38 represented total saving for the case of HDD ≤ 18 • C. In this case, the database does not distinguish by sex; we assume males and females are equal.
When savings from reducing the impact of pollution on health and traffic accidents are compared with total costs of PV + storage purchases for homes between 51-70 m 2 , the benefits of replacing firewood for heating with this type of systems outweigh the total cost. The annualized total cost of PV + storage systems amounts to 8

Conclusions
Replacing wood burning stoves in homes with PV + storage systems reduces the environmental stress as the decrease of firewood requirements allows a more sustainable model to reach a good standard of living. This replacement would also contribute to enhancing health levels in the six types of diseases under evaluation both for males and for females. Last but not least, the number of traffic accidents in a city like Temuco, which suffers from the thermal inversion phenomenon, would also be significantly reduced.
Our results support the positive measures included in the PM 2.5 Air Decontamination Plan, aimed at improving the thermal insulation of homes, the efficiency of wood burning stoves, and the improvement of the quality of firewood used. However, they enable a potential improvement if we take into account our recommendation to include the replacement of wood burning stoves with PV + storage systems-a situation that would generate an increase in positive environmental and health effects.
Given that the size of the savings obtained is significant, we recommend that funds then not required for health care be used to finance partial subsidies for the purchase of these systems, and also to improve the health care of different diseases and road safety. Temuco (like the other cities in the South-Central area of Chile) has high levels of energy poverty. This aspect strongly affects the adoption of technology by users. Public subsidies are a correct measure to bridge this barrier.
The study has not included the decreased cost due to the reduction of logging activities oriented to the production of firewood. In certain ways, this negative impact may be offset by the increased activity of the installation and maintenance sector for the new PV + storage systems. The cost of dismantling stoves was considered null or is compensated by the income from its possible sale as scrap. It should be also considered that many houses in Temuco still use the wood stove simultaneously to cook, heat the house, and generate sanitary hot water with water containers placed around the cannons. This is not included in the analysis. Further research might deal with these limits.
The monetary assessment of traffic accidents is also limited in the analysis by the fact that information on the number of accident victims in Chile is only available up to 24 hours after the accident. There are no records for injuries or fatalities that occur beyond that period of time.
Another limit is derived from the fact that we assumed that smog was caused only by particulate matter emissions.
The results and conclusions obtained in this study have to be taken with caution. It has been assumed that the impact of pollution on health for the whole of Chile is the same for all cities. The GBD tool offers regional analysis for some countries, but not for Chile. This recommends taking the results as minimum values. It should also be borne in mind that the analysis focuses only on housing of 51 to 70 m 2 , with the possibility of extending it to housing with larger living areas. Further research might consider different operation conditions and managing stoves to spread out the set of firewood emission factors used.
Finally, it is important to point out that studies like this should delve deeper, taking into account that climate change, the growth of cities, and real estate development are resulting in more cities in Chile with pollution problems and saturation, which have highly damaging health effects. Thus, the search for new and better solutions to prevent the pollution produced by the heating of homes is a priority issue on the path to sustainable cities with a better quality of life.