Improved Stove Evaluation in Ayaviri, Peru

Improved Stove Evaluation in Ayaviri, Peru University of Colorado at Boulder Mortenson Center in Engineering for Developing Communities Practicum Report Kristen Matsumura November 15, 2011

Executive Summary Most families outside of the city of Ayaviri cook with cow dung and use unimproved clay stoves. These high altitude homes tend to use their stoves for both cooking and heating, resulting in an increased exposure to unhealthy levels of indoor air pollution which can cause problems such as respiratory infections, tuberculosis, and cataracts, among others. Despite these problems few families in the prelature of Ayaviri have the knowledge or financial capacity to improve their stoves. Caritas Ayaviri has established an ongoing goal of helping families gain access to improved stoves while building household capacity and knowledge. While the ultimate goal is to develop demand for improved stoves, there is little evidence of adequate capacity or motivation to build or purchase stove improvements. In the cycle of appropriate technology implementation there are three major phases including appraisal, design, and monitoring/evaluating. Caritas has assessed the need for improved cook stoves in the rural communities around Ayaviri. Working with the CU Law School, they selected a design that appears to fit the communities needs. In May of 2011, Caritas and students from the University of Colorado s Law School constructed improved stoves in 15 key households in 3 rural communities outside of Ayaviri. It was expected that these families would learn about the technology and promote the benefits of having an improved stove to others in their communities. In August of 2011 engineers from CU Boulder, SENCICO and GIZ completed an evaluation of the implemented stoves by looking at technical performance, indoor air pollution, and acceptance by the users. A group of families who possess traditional (unimproved) stoves were also included in the evaluation in order to provide a quantitative baseline from which to compare stove characteristics. This report summarizes the findings of the stove evaluations including the performance improvements that can be expected from the improved stove which includes decreased cook time, decreased fuel consumption and significantly reduced indoor air pollution. It was found that despite performance improvements, most households were not using their improved stoves because they had not finished constructing their kitchens or because they were unaware of the benefits. It was also found that although reduction in smoke was significant in the improved stoves, emissions still exceeded World Health Organization standards. Background Caritas International is a Catholic organization operating in 165 countries where they focus on the reduction of poverty and social injustices. In Peru, Caritas manages a network 49 local office, also known as diocesans. Each branch manages their own programs in the fields of economic development, family health, microfinance, and emergency response with support from the regional and international levels. Caritas operates a branch of the organization out of the city of Ayaviri in southeastern mountains of Peru. Ayaviri is both the name of the district and the capital city of the Peruvian province of Melgar. The city has an approximate population of 21,000 people while the surrounding communities are more sparsely population and significantly less developed. Infrastructure such as paved roads and piped water does not reach even the closest communities lying immediately beyond the city limits.

Caritas Ayaviri maintains multiple active projects in the local area focusing on assisting the populations in these surrounding communities. In May of 2011 Caritas and members of the University of Colorado Law School began a project to introduce improved cookstoves in the communities outside of Ayaviri. Figure 1 Traditional stove Figure 2 Dried cow dung In these rural area outside of Ayaviri, families cook with dried cow dung on small traditional earthen stoves (see Figure 1 and Figure 2). Families tend to have a dedicated room for cooking where they also often store potatoes and other foods. Women and other family members spend hours in these kitchens throughout the day and are exposed to harmful levels of smoke. Figure 3 Combustion chamber Figure 4 Map of communities and households There are a limited number of income generating activities for the families in these communities. The majority of the population relies on pastoral subsistence living or on finding jobs in the city of Ayaviri. However, traveling to Ayaviri becomes difficult for those living far away as few families can afford cars and the roads are not well maintained. Given the limited financial resources, isolation from the city s infrastructure, and absence of a function economy the

families in these outlying communities lack the capacity to improve their stoves. Many are also simply unaware of the dangers related to prolonged exposure to indoor air pollution. In May fifteen improved stoves were constructed in three rural communities outside of Ayaviri; Sunimarca, Condormilla Bajo and Condormilla Alto (See Figure 4). A stove developed by the German Agency for International Cooperation (GIZ) was used that can efficiently burn both firewood and cow dung. Households were selected based on their influence in their communities, with the hope that they would promote stove construction and educate their neighbors about the benefits of the new stoves. This improved stove design incorporates a metal reinforced concrete surface with two standard 26 cm pot holes. The combustion chamber is an L-shaped cavity that is surrounded by ceramic and contains a metal grate for holding fuel and a space below where ash can fall and air can enter (Figure 3). It also features a metal chimney that draws smoke outside the kitchen. Table 1 Field Team Name Organization Email Percy Quispe Caritas Ayaviri perohaq199@hotmail.com Kristen Matsumura CU Boulder kristen.matsumura@colorado.edu Veronica Pilco Mamani GIZ veronica.pilco@giz.de Rafael Torres Rojas SENSICO rtorresr17@hotmail.com In August of 2011, three months after the stoves were constructed, an evaluation team returned to each household to determine if the stoves were being used and how well they were performing. The team included a member from Caritas Ayaviri and three engineers from CU Boulder, GIZ and SENSICO (listed in Table 1). The representatives from GIZ and SENSICO were experts in the design and operation of the stoves, while the CU Boulder engineer provided expertise in indoor air pollution and efficiency measurements. Methodology Two sample groups that were identified for inclusion in the evaluation; households who were using the improved GIZ stove and those who were using traditional stoves. The sample size was determined by how many of the fifteen improved stoves were being used by the households regularly which was defined by if they had two pots to use with the stove and the stove was indoors. The evaluation took two weeks and included measurement of stove performance, emissions and adoption. Table 2 illustrates which households were included in different aspects of the evaluation including the water boil test, emission measurements, and the survey. To evaluate the appropriateness of the improved GIZ stoves, three characteristics were surveyed including performance, emissions, and adoption. The performance of the stoves was measured using the latest edition of the Water Boil Test, version 4.1.2, which evaluates the amount of fuel and time used to boil water (ETHOS 2009). Emissions of particulate matter (PM2.5) and carbon monoxide were measured during a 24 hour span of normal cooking in each household using UCB Particle and Temperature Sensors (PATS) and Aprovecho Indoor Air Pollution Meters. Adoption was measured with a survey that asked about use, acceptance, and appeal.

Table 2 Household Sample Community Name GIZ Use PATS- IAQ- IAQ- Survey WBT stove GIZ PM PM CO Sunimarca Pedro x x x x x Comunidad de Sunimarca x x x x Leónidas x x x Plácido x x Víctor x Santos x x x Adolfo x x Condormilla Alto Leoncio x x x x x x x Román x x x x x Valentín x x x x x Lucas x x x x x x Ignacio x x x Ermogenes x x x x Condormilla Bajo Juan x x x x x x x Abel x x x x x x Pablo x Idelfonso x x x x x Vicente x x x x x x Gambino x x x 15 7 12 16 12 8 8 Performance Testing While the improved stoves are expected to reduce indoor air pollution associated with cooking, it is less clear how they affect the speed and efficiency of cooking. While less smoke is desirable and should positively affect the adoption of the stoves, the consumption of time and fuel are equally, if not more important, in the eyes of the users. The water boil test is intended to simulate the cooking process and provides a systematic method of measuring and comparing how efficiently different stoves use fuel. This test was utilized to compare the performance of the GIZ stoves to traditional stoves in the same communities. The water boil test includes three phases; the cold-start test, the hot-start test, and the simmer test (see Table 3). The cold-start test begins with the cool stove and measures the time and amount of fuel used to boil five liters of water. The hot-start tests begin immediately after with the still warm stove and five liters of new room temperature water. The simmer test is conducted following the hot-start test and measured the amount of fuel used to maintain a boil for 45 minutes. Table 3 Water Boil Test Protocol Test Name Metric Measure 1 Cold-start Test Boil five liters Time, Fuel 2 Hot-start Test Boil five liters Time, Fuel 3 Simmer Test Maintain boil for 45 minute Fuel The water boil test was conducted with 16 different homes, which included7 homes using the GIZ stoves, 7 that are using traditional stoves, one home that uses a different model of improved stove, and a home with a traditional stove with a chimney. The sample is listed in Table 2. The improved stove that was not installed by Caritas was implemented by another NGO, and includes a similar design with a concrete surface and a chimney, but also features a component

that allows for the user to adjust the flow of air through the chimney. The users of this stove were extremely happy with their stove. Emissions Testing Fine particle pollution includes microscopic solid and liquid particles suspended in the air that are 2.5 micrometers in diameter and smaller. This particulate matter (PM 2.5) can also be described as the respirable fraction of dust and smoke and has been linked to several adverse respiratory and cardiovascular health effects including acute respiratory infections, tuberculosis, low birth weight and cataracts(seaton et al. 1995; Ezzati and Kammen 2002). Table 4 World Health Organization Recommended Limits Pollutant Recommended Limit Time Period PM 2.5 25 µg/m 3 24 hours CO 9 ppm 8 hours Carbon monoxide (CO) is a toxic gas that is released during inefficient combustion that is common in open fires and unimproved cookstoves. Exposure to CO has many acute and long term health effects that can range from delirium and flu-like symptoms to memory loss, central nervous system damage, or heart failure(prockop and Chichkova 2007). The World Health Organization recommends a PM 2.5 limit of 25 micrograms per cubic meter over a 24 hour period and a carbon monoxide limit of 9 ppm over an 8 hour period (WHO 2004; WHO 2005). Figure 5 Emissions Monitors Particulate matter was measured using two different instruments, the Particle and Temperature Sensor (PATS) developed by the University of California Berkeley, and the Aprovecho Indoor Air Pollution (IAP) meter as shown in Figure 5. As displayed in Table 2, PM was measured in twelve households using the PATS, and eight households using the IAP meter. Carbon monoxide was also documented in the same eight households as the Aprovecho IAP meter also measured CO.

In order to measure emissions associated with normal cooking behavior, monitors were left in households for at least 24 hour. Measurement of emissions did not overlap any water boil test since only normal cooking emissions were desired. In order to later compare data from both instruments, it is important to ensure that their values are highly correlated. Redundant measurements were therefore taken in the same households with both types of instruments. According to the standard operating procedure of the UC Berkeley PATS, instruments were to be placed 100 cm from the edge of the stove, 145 cm above the floor, and 150 cm horizontally away from any open windows or doors(berkeley Air 2005). Instruments were suspended from the ceiling of the kitchen using string and placed as close to the proper location as possible (See Figure 5). Adoption Survey A verbal survey was given to twelve households who received an improved stove. The survey was developed in conjunction with the CU Boulder Law students who implemented the stoves, and built on the general energy survey Proyecto Energía, Desarrollo Y Vida written by GTZ. The survey included five observations and fifteen questions which can be viewed in Appendix A. The purpose of the survey was to understand how often households were using their stoves, how they maintained it, what characteristics they liked, and what characteristics they thought could be improved upon. Also, since households were originally selected based on their influence in the community with the hope that they would disseminate positive feedback to their neighbors, the survey included a question asking if they have shown their stove to their family or friends. Results It was found that the improved GIZ stoves provide an improvement in performance and emissions, however adoption and consistent use has been slow. Of the fifteen stoves implemented, only seven appeared to be used with any regularity, and only a handful of these seem to use the stoves consistently for the majority of their meals. While this gives a pessimistic impression households were enthusiastic about their stoves and were assuring that they will soon be using their stove more often once they complete the walls and roof around the stove. Table 5 GIZ stove use Community Name Complete Using GIZ Kitchen Stove Sunimarca Pedro Figueroa Bautista x x Comunidad de Sunimarca x x Leónidas Huayta Machaca Plácido Vilca Atamari Víctor Valeriano Tuni Condormilla Alto Leoncio Quispe Ccama x x Román Mamani Quispe x x Valentín Choque Farfán x x Lucas Morales Mamani Ignacio Huayta Aguilar Condormilla Bajo Juan Huisa Quispe x x Abel Filomeno Cáceres Suca (Dolores) x x Pablo Julián Vilca Quispe Idelfonso Clemente Quispe Huisa x Vicente Suca Choque (Florencio)

Performance The water boil test was performed in sixteen homes, eight of which were with improved stoves. The local boiling point was consistently found to be 86 Celsius where the average elevation of the houses was 12,997 feet (3961 meters). When evaluating the efficiency of stoves we look at the time to boil, equivalent dry fuel consumed, and total fuel used which are all listed in Table 6. The equivalent dry fuel consumed is equal to the weight of the fuel before it is burned minus the amount of ash left after the test, while the total fuel used is equal to the equivalent dry fuel consumed divided by the effective mass of water boiled. The total fuel used is calculated across the three phases of the water boil test using Equation 1. It was found that on average the improved stoves use about 400 g less equivalent dry fuel over the whole water boil test. When disaggregated it was found that the improved stoves use less fuel for both the cold-start boil and the hot-start boil, but about the same amount of fuel for the simmer. When incorporating the mass of water boiled, we find the same pattern for total fuel use where the improved stoves use les total fuel over the whole water boil test by an average of 120 g/liter. Total Fuel Use = 5 (SC + SC ) + SC 2 Equation 1 Specific Consumption (SC) = Equivalent Dry Fuel Consumed Effective Mass of Water Boiled Equation 2 It was also found that the improved stoves on average boiled water faster for both the cold and hot start phases. On average the improved stoves boiled water in 35 minutes and 22 minutes for the cold and hot starts respectively, while the traditional stoves took an average of 40 and 29 minutes respectively. Community Name Table 6 Water Boil Test Results Equivalent Stove Dry Fuel Type Consumed (g) Time to Boil - Cold Start (min) Time to Boil - Hot Start (min) Total Fuel Use for WBT (g/liter) Sunimarca Pedro Figueroa Bautista Improved 2,159 43 22 618 Comunidad de Sunimarca Improved 3,350 27 23 1,088 Santos Vilca (Improved) Improved 4,795 35 26 1,226 Adolfo Arapa Traditional 1,713 28 18 947 Leónidas Huayta Machaca Traditional 2,850 53 40 756 Plácido Vilca Atamari Traditional 5,950 53 34 1,904 Condormilla Leoncio Quispe Ccama Improved 3,600 48 25 1,164 Alto Román Mamani Quispe Improved 3,888 29 16 1,315 Valentín Choque Farfán Improved 2,778 41 19 829 Lucas Morales Mamani (Chimney) Traditional 3,425 31 28 1,082 Ermogenes Quispe Traditional 4,590 50 39 1,240 Ignacio Huayta Aguilar Traditional 3,620 26 23 1,009 Condormilla Juan Huisa Quispe Improved 3,385 29 18 1,090 Bajo Abel Filomeno Cáceres Suca Improved 2,913 29 28 942 Vicente Suca Choque Traditional 3,423 38 27 929 Idelfonso Clemente Quispe Huisa Traditional 4,550 42 22 1,364 ALL Improved 8 3358 35 22 1034 Traditional 8 3765 40 29 1154

It can be seen in Figure 6 and Figure 7 that the improved stoves provide an improvement on the time to boil regardless of whether the stove is cold or hot to start. The box plots display five important values of the distribution of values including the median as a thick line, the lower and upper quartile as the limits of the box, and the minimum and maximum values as whisker lines. Without more data the decrease in time to boil cannot be conclusively quantified, but we can express that there is a non-negligible difference between the time it takes to boil between the two types of stoves. Time to Boil - Cold Start (Minutes) 25 30 35 40 45 50 Traditional Improved Figure 6 Box Plot of Time to Boil Cold Start Time to Boil - Hot Start (Minutes) 20 25 30 35 40 Traditional Improved Figure 7 Box Plot of Time to Boil Hot Start Emissions The twenty-four hour particulate matter concentration was measured in a total of 14 homes, six of which were measured simultaneously by both instruments as explained in Table 7.While it was found that the PM2.5 concentrations were not consistent between the two instruments, they are highly correlated with a correlation coefficient of 0.93. The reason for this discrepancy can be explained by the placement of the instruments in the kitchen and the difference between the maximum limits that each instrument can record. Table 7 Twenty-four hour average of PM2.5 concentration Results Community Name Stove Type PATS APROVECHO (mg/m 3 ) (mg/m 3 ) Sunimarca Pedro Improved 0.54 Santos Improved 0.17 Comunidad de Sunimarca Improved 0.08 Adolfo Traditional 0.98 Condormilla Alto Román Improved 0.26 Valentín Improved 0.07 Leoncio Improved 0.21 0.95 Ermogenes Traditional 0.66 0.49 Lucas Traditional 0.07 0.08 Condormilla Bajo Juan Improved 0.23 1.90 Abel Improved 0.15 Florencio Traditional 0.79 0.28 Gambino Traditional 2.52 13.27 Idelfonso Traditional 5.56 All Improved 0.22 1.00 Traditional 1.00 3.93 Omitting Santos and Lucas Improved 0.23 1.00 Traditional 1.24 4.90

The UCB PATS instrument can only record PM2.5 concentrations less than 50 mg/m 3 while the Aprovecho has a limit of 60 mg/m 3. While none of the houses had average concentrations greater 15 mg/m 3, peaks during cooking may have exceeded the maximum value that the instruments can read, therefore the Aprovecho instrument would be expected to yield greater estimates than the PATS. Figure 8 Example of PM2.5 Emissions over 24 hour span Similarly the instruments were suspended from the ceiling by string as seen in Figure 5, which resulted in one instrument being closer to the stove than the other. This specific placement could have significantly affected the values recorded by each of the instruments. Despite these differences, a trend still emerges illustrating that the improved stoves emit significantly less PM2.5 during a 24-hour period of normal cooking behavior. Households tended to cook twice per day and peaks of PM2.5 emissions can be seen during these cooking episodes as show in Figure 8. 24 hr PM2.5 average (mg/m3) 0.0 0.5 1.0 1.5 2.0 2.5 24 hr PM2.5 average (mg/m3) 0 2 4 6 8 12 Traditional Mejorada Figure 9 Box Plot of PATS PM2.5 (Omitting Lucas) Traditional Mejorada Figure 10 Box Plot of Aprovecho PM2.5 (Omitting Lucas)

Taken over the twenty-four hour span, we arrive at an average concentration of PM2.5 according to the UCB PATS instruments. Collectively, the improved stoves emit an average of.22 mg/m 3, while the traditional stoves emitted an average of 1.00 mg/m 3. This average was also calculated without Lucas Morales Mamani in the sample since his stove includes a chimney and could be considered a significant improvement over normal traditional stoves. Omitting Lucas, the average emissions associated with traditional stoves is 1.24 mg/m 3 as explained in Table 7. A box plot of the average of PM2.5 emissions measured by the UCB PATS instrument is shown in Figure 9. Since the box plots don t overlap, it can confidently be said that the improved stoves emit significantly less particulate matter in a twenty four hour span. Despite the fact that the PM exposure was reduced by a factor of 5 on average, the improved stoves still exceed the World Health Organization s recommendation of.025 mg/m 3 by almost 9 times. The Aprovecho IAP meters provide a similar trend between the improved and traditional stoves. There is only data from eight households with the Aprovecho meters, but a decrease from 3.93 to 1.00 mg/m 3 is observed between the traditional and improved stoves. When the data form Lucas Morales Mamani s household is omitted, as explained above, the average emissions associated with traditional stoves jumps to 4.90 mg/m 3 over a twenty-four hour span. A box plot representing the distribution of 24 hour emissions measured by the IAP meters can be seen in Figure 10. Carbon monoxide emissions were measured in eight households using the Aprovecho IAP meters as well. The average CO concentration over twenty four hour spans is displayed in Table 8 in parts per million (ppm). The households with improved stoves averaged 5.68 ppm, while the homes with traditional stoves averaged 14.68 ppm. Despite having a chimney, Lucas Morales Mamani s home saw a high level of CO even though they saw a relatively low level of particulate matter. This can be explained due to the fact that CO is present when fuel does not combust completely, as is common with traditional stoves and is therefore expected to be higher in Lucas stove. However the presence of a chimney is what affects the level of PM 2.5 in households, and would therefore also expect to be lower in Lucas home. Table 8 Carbon Monoxide Emissions Results Community Name Stove Type CO (ppm) Condormilla Alto Leoncio Quispe Ccama Improved 14.80 Ermogenes Quispe Traditional 12.50 Lucas Morales Mamani (Chimney) Traditional 11.30 Condormilla Bajo Juan Huisa Quispe Improved 1.30 Abel Filomeno Cáceres Suca Improved 0.95 Florencio Suca Choque Traditional 12.30 Gambino Cáceres Huaman Traditional 19.60 Idelfonso Quispe Huisa Traditional 17.70 ALL Improved 5.68 Traditional 14.68 It is important to note that while trends emerge when the data for the water boil test and indoor air emissions are presented as box plots, when evaluated using a Welch Two Sample t-test to measure if the two groups different significantly, no comparisons are found to have a p-value less than 0.05. That is to say, without more data the improvements in boil time and indoor air pollution are only anecdotal.

Adoption Survey A survey was given to twelve households who received an improved stove, however only six were able to give feedback on the use of the stove. As mentioned before only seven of the fifteen stoves appeared to be used with any frequency which was determined by reported use, the presence of pots that fit the stove, and the completion of the kitchen walls and roof. The community stove in Sunimarca was marked as a stove that is frequently used but since it is used by different people, there was no individual person who could give feedback on its use. The six households who did provide feedback described how their new GIZ stoves compare to their previous traditional stoves. The results of these answers can be found in Table 9. While all households agreed that the new stoves reduced the amount of smoke in the kitchen, there is disagreement as to whether the stove cooks faster and if it consumes more fuel. While our water boil tests indicate that on average the improved stoves utilize less fuel and boil water more quickly, this cannot be said for all cases; especially since the box plots for the improved and traditional stoves in Figure 6 and Figure 7 overlap one another. Community Name Table 9 Stove Use Survey Results How Time to Smoke Often Cook in Room Used Pots Fit Comfort to Use No. of Burners Fuel Used Sunimarca Pedro Figueroa Bautista sometimes same less better same same same Condormilla Leoncio Quispe Ccama always longer less worse better same more Alto Román Mamani Quispe sometimes same less better better same less Valentín Choque Farfán always longer less worse better less less Condormilla Juan Huisa Quispe sometimes same less worse better less less Bajo Abel Filomeno Cáceres Suca sometimes shorter less better better same less Worse 33% 0% 50% 0% 33% 17% Same 50% 0% 0% 17% 67% 17% Better 17% 100% 50% 83% 0% 67% According to feedback, the most common problems include getting the fire started and fitting pots to the burners. Users complained about the combustion chamber opening being too small making starting the fire and placing fuel difficult. Users also complained if their original pots did not fit the new burners. A couple households wanted more burners, or requested certain size burners. A less common complaint was that the combustion chamber fills with ash too quickly. Of the six families surveyed, only two said they had shown their stove to their friends or neighbors. Discussion Of the households who received an improved stove, only a handful of them are using them regularly even three months after they were built. While the rate of use may seem low, the main reason families are not using their stoves is because they have not completed the roof and walls around it yet. The homes that are using the stoves consistently had finished the walls and roof before the stove was even built. Future stove projects should require families to complete the kitchen prior to stove installation to ensure the stove will be usable immediately. Similarly households should be informed of their selection to receive a stove in the summer months since many households claimed they had not completed their kitchen because the earthen bricks are difficult to produce in the cold winter months.

However having a completed kitchen is not the only factor that affects stove use since there are families whose stove is indoors and are still not using it all the time. It is clear that many families are unaware of the benefits associated with the stoves, or don t prioritize them. Reduction of indoor smoke is not enough to convince a family to change; it must be clear that the stove consistently cooks more quickly, uses less fuel, and is just as convenient as their previous stove. Next Steps If families who received stoves are expected to use them, then more work must be done to educate families on the benefits of reduced indoor air pollution and on the expected increases in efficiency. It was apparent during this survey that many families did not know how to use the stoves properly, and had difficulty using them efficiently. If families are expected to educate their neighbors on the benefits of the stoves, and even how to build them, then significant work must be done to develop (1) education material regarding construction and proper operation, (2) the supply chain for important, non-locally produced components, and (3) a training program to ensure quality control and possible scalability. It is evident from this evaluation that although Caritas is helping families gain access to improved cookstoves, there has been little development of demand. It is clear that families lack the understanding of what improvement options exist for their cookstoves. For example, many are unaware that a chimney can be added to their stove in order to draw smoke out of the kitchen. Likewise families lack the understanding of the benefits associated with improving their stove; that is they are unaware that a kitchen with less smoke is significantly healthier for them. In order to encourage sustainable change, families need to create their own demand for changes. They need to be educated on the potential options for and benefits of improving their stoves. Caritas could organize workshops for community members in order to educate them of the dangers associated with their traditional stoves, and encourage discussion of possible changes they can make. Encouraging demand for improved stoves is only the first step in creating a long term solution. The infrastructure for providing family s with stoves will also have to be developed. Technicians will have to be trained in stove construction, different stove options will have to be designed to cater to different demands, and a supply chain of all necessary stove parts will have to be established. This process can be developed into a business that undertakes designing and building stoves in people s homes. Caritas can play a role in encouraging this sustainable process but will have to commit to focusing on creating demand and training stove builders rather than providing free stoves to selected families. Lessons Learned On a project level this evaluation has shown that it is important to implement programs with long-term adoption and practices in mind. While the few individual families who received a stove will see significant benefits, the communities may ultimately suffer if a more comprehensive program is not implemented to support a stove market.

It is also extremely important that enough data is collected during evaluations to draw statistically significant conclusions. While qualitative information is useful and can illuminate important information, the purpose of collecting quantitative data is often to identify statistical trends. It can be a waste of resources and time to collect data that can then not be used to draw significant conclusions from. That being said, all quantitative data should be coupled with qualitative descriptions of the situation on the ground in order to paint a full picture. Numbers alone cannot explain what problems exist and how they can be solved in a specific culture and context. A final key finding was that while expertise may come from international and other non-local workers, it is important to transfer technical knowledge to local leaders and NGO workers. While it is often easier to do measurements and surveys without simultaneously having to train a local expert, it will benefit the community in the long term by imparting a specific skill to someone who lives in the region. Local NGO workers often have fewer cultural barriers to overcome when interacting with local residents, therefore giving them a better perspective from which to report information and collect survey data.

References Berkeley Air (2005). Standard Operating Procedure - Guidelines for Instrument Placement. Berkeley, Berkeley Air. ETHOS (2009). Water Boiling Test Version 4.1.2. L. Fierce and T. Bond. Ames, Iowa, Engineers in Technical and Humanitarian Opportunities of Service Ezzati, M. and D. Kammen (2002). "The health impacts of exposure to indoor air pollution from solid fuels in developing countries: knowledge, gaps, and data needs." Environmental Health Perspectives: 1057-1068. Prockop, L. D. and R. I. Chichkova (2007). "Carbon monoxide intoxication: An updated review." Journal of the Neurological Sciences262(1-2): 122-130. Seaton, A., D. Godden, W. MacNee and K. Donaldson (1995). "Particulate air pollution and acute health effects." The Lancet345(8943): 176-178. WHO (2004). Environmental Health Criteria Carbon Monoxide. J. Raub. Geneva, World Health Organization. 213. WHO (2005). Air Quality Guidelines For Particulate Matter, Ozone, Nitrogen Dioxide And Sulfer Dioxide. World Health Organization Summary of Risk Assessment. Geneva, World Health Organization.

Appendix A Survey

Appendix B GPS Data Table 10GPS Data Community Name Stove Type Elevation (ft) Position (Degree Min) Sunimarca Pedro Improved 13002 S14 56.882 W70 44.092 Comunidad de Sunimarca Improved 13107 S14 58.181 W70 43.281 Santos Improved 13077 S14 57.222 W70 43.425 Adolfo Traditional 13099 S14 56.993 W70 43.469 Leónidas Traditional 13058 S14 57.477 W70 43.394 Plácido Traditional 13065 S14 58.087 W70 43.233 Condormilla Alto Leoncio Improved 13000 S14 57.361 W70 40.110 Román Improved 12992 S14 57.347 W70 40.197 Valentín Improved 12829 S14 54.655 W70 37.372 Lucas Traditional 12839 S14 55.621 W70 37.799 Ermogenes Traditional 13000 S14 57.361 W70 40.110 Ignacio Traditional 12894 S14 56.748 W70 40.336 Condormilla Bajo Juan Improved 13091 S14 58.983 W70 37.645 Abel) Improved 12882 S14 56.609 W70 36.677 Vicente Traditional 12887 S14 56.587 W70 36.883 Idelfonso Traditional 13135 S14 59.206 W70 37.498 Gambino Traditional 13118 S14 58.746 W70 37.485 Ayaviri Caritas 13075 S14 52.996 W70 35.448