Carbon Baseline Assessment of the Envirofit G3300 and JikoPoa Improved Cookstoves in Kenya

Carbon Baseline Assessment of the Envirofit G3300 and JikoPoa Improved Cookstoves in Kenya for The Paradigm Project Berkeley Air Monitoring Group January 2011

Table of Contents 1 EXECUTIVE SUMMARY... 4 2 BACKGROUND... 4 3 METHODS AND APPROACH... 4 3.1 LEAKAGE EFFECTS... 4 3.2 KITCHEN SURVEY... 5 3.3 KITCHEN PERFORMANCE TEST... 5 4 RESULTS... 6 4.1 KITCHEN SURVEY... 6 4.2 KITCHEN PERFORMANCE TEST... 7 4.2.1 Fuel Use by Stove Type... 7 4.2.2 Final Clustering and Fuel Savings... 9 4.2.3 Other Fuel Use... 12 APPENDIX 1: KPT FIELD DATA COLLECTION FORM... 14 APPENDIX 2: SCALE CALIBRATION RECORD... 20 APPENDIX 3: SCALE CALIBRATION CERTIFICATE... 21 Traditional 3-stone Fire Traditional Wood Stove Envirofit Wood Stove Field Team JikoPoa Wood Stove January 2011 2

1 Executive Summary Berkeley Air Monitoring Group and Green Futures performed field-based assessments to determine the effect of the Envirofit G3300 and JikoPoa improved wood cookstoves on fuel consumption in households in Nairobi, Meru, and Marsabit in Kenya. A subsumed approach was used in which all fuels used throughout the three-day Kitchen Performance Test were weighed. Results from the monitoring assessments can be considered representative of actual household fuel use and user practice. Households were randomly selected using Kitchen Surveys, input from local partner organizations, and an expected future customer profile. There was no difference in fuel use and fuel savings between Envirofit and JikoPoa stove users, and therefore assessing fuel savings based on stove type was not a valid clustering method. The most accurate clustering criterion for fuel savings among this population of users was baseline fuel use. Among the KPT households, there was a group that did not use wood before purchasing their improved stove (Non-wood Households), as is typical of a proportion of real customers and projected future customers. This group had significantly different fuel use and fuel savings patterns than all other households (All Other Households) and was therefore analyzed as a separate cluster. Among the 194 households sampled, little LP gas and kerosene was used, and there was no change in either fuels use before and after purchase of the improved stove. The Summary Table below shows the wood and charcoal savings among the Non-wood Households and All Other Households sub-groups. Among the Non-wood Households sub-group wood savings were -0.37 kg/person-day. Among the All Other Households sub-group, average wood savings were 0.64 kg/person-day. Average charcoal savings did not meet the 90/30 rule and therefore the lower limit 90% confidence interval charcoal savings need to be used. The lower limit 90% confidence interval charcoal savings among the Non-wood Households were 0.10 kg/person-day and among All Other Households were 0.022 kg/person-day. January 2011 3

Summary Table. Average wood and lower limit 90% confidence interval charcoal savings among non-wood using households and all other households (kg/person-day). Non-wood HHs All Other HHs N Average Wood Savings Lower Limit 90% CI Charcoal Savings 22-0.37 0.10 172 0.64 0.022 2 Background From September to November 2010, Berkeley Air Monitoring Group in collaboration with Green Futures (Nairobi, Kenya) performed several field-based assessments to determine the effect of improved cookstoves on fuel consumption in households in Nairobi, Meru, and Marsabit in Kenya. Fuel savings were measured for the Envirofit G3300 and JikoPoa improved wood cookstoves distributed by the Paradigm Project through their local partners, including Food for the Hungry (FHI), Explore Kenya, and TIST. Both improved stoves are rocket-type, wood-burning technologies. Only one size of each stove was distributed. Berkeley Air Monitoring Group is a social venture that provides independent third-party program evaluation and rigorous, scientifically-based measurement of indicators related to solid fuel and biomass energy throughout the world. Green Futures is a Kenyan company that provides technical expertise for energy projects in Kenya. Neither organization has any commercial interests in the Envirofit or JikoPoa stoves or in the Paradigm Project, and the outcomes of this study in no way impact either organization s compensation. The Paradigm Project also collaborated with Impact Carbon (San Francisco, CA, USA) on carbon project development. 3 Methods and Approach 3.1 Leakage Effects The Gold Standard Methodology requires that specific forms of leakage be reviewed during the baseline and periodic monitoring processes. Significant sources of leakage were subsumed in the January 2011 4

paired Kitchen Performance Test, which assessed all stove usage and fuel mixing within households. 3.2 Kitchen Survey The Kitchen Survey (KS) is designed to characterize the typical improved stove customer. The KS was designed, conducted, analyzed, and reported by the Paradigm Project. In total, the Paradigm Project completed 37 surveys of typical customers in Nairobi, 86 in Meru, and 98 in Marsabit. Berkeley Air and Impact Carbon used the KS results to create the household selection criteria for the Kitchen Performance Tests. 3.3 Kitchen Performance Test Kitchen Performance Tests (KPTs) were performed in 40 households (HHs) in Nairobi, 80 HHs in Meru, and 80 HHs in Marsabit (200 HHs in total) by a field team of four plus one supervisor from September through November 2010. One hundred tests were performed in HHs with the Envirofit G3300 stove and 100 in HHs with the JikoPoa stove. A Berkeley Air representative worked on site to train and guide the team during the first half of the project. The KPT was conducted over three full days (72 hours), requiring daily household visits for four days. Wood, charcoal, kerosene, LP gas, and any other fuels used were weighed daily using Salter Brecknell (Fairmont, MN) ElectroSamson digital hand-held scales (25 kg x 0.02 kg). Wood moisture was measured daily in each HH using a JT-2G digital moisture meter (HQRP, Harrison, NJ). A KPT survey was also administered daily to record information about cooking stove and fuel usage, the number and type of meals prepared, and the number of people cooked for. The KPT was performed using a Before and After (paired) study design. The KPT was first performed in the HHs with their existing stove(s) prior to the introduction of the Envirofit or JikoPoa improved wood stove ( Before ). After these baseline measurements were taken, the households were sold the improved stove -- either the Envirofit or JikoPoa -- by the staff of the Paradigm Project s local partner organizations. The Paradigm Project set a uniform price for the stoves that is sustainable within their program parameters. The HHs were given at least 14 days to become accustomed to their improved stove, and finally the After KPT was performed. HHs were asked to use whatever stove(s) and fuel(s) they preferred during each of the days of testing. January 2011 5

The KPT HHs were selected using screening criteria based on the baseline Kitchen Surveys, input from local partner organizations, and a profile of the program s expected future customer population, so as to be representative of the typical, projected Envirofit and JikoPoa stove purchaser. Efforts were taken to minimize the impact on HHs during the course of the monitoring. All participants were requested not to modify their typical patterns of fuel usage, fuel type, or stove type. No criteria limited the inclusion of multi-fuel or multi-stove users in the cohort. The subsumed approach resulting from the paired KPT was applied as all fuels used by the HHs per their typical cooking patterns were weighed. The raw KPT data was analyzed to yield fuel use and fuel savings in terms of kilograms per person per day using the total reported HH size. All wood measurements are reported in terms of wet mass. The KPTs were conducted on a representative sample of current customers and the expected future customer population nationwide. Therefore, the values presented are the best estimate of fuel use for this KPT group. Adhering to the Gold Standard Methodology, the customer base will continue to be monitored for changes in fuel use patterns before and after adoption of the improved cookstove. 4 Results 4.1 Kitchen Survey Using the Paradigm Project s KS results, input from local partner organizations, and a profile of the program s expected future customer population, Berkeley Air and Impact Carbon created the KPT HH selection criteria. The criteria were created to be representative of the expected future customer population and were used to select HHs randomly. Below is a list of the criteria by region. 1. Nairobi (among the 20 Envirofit KPT HHs and 20 JikoPoa HHs): a. Charcoal or kerosene should be their primary fuel. Wood can be their secondary (or alternative) fuel, but not their primary fuel. b. HH size must be within 2 and 12 people. January 2011 6

2. Meru (among the 40 Envirofit KPT HHs and 40 JikoPoa KPT HHs): a. 40 HHs should use an open 3-stone fire as their "primary stove." These HHs can also use another wood stove, charcoal, kerosene, and/or LP gas at any frequency. b. 40 HHs should use charcoal at least one time per day. These HHs can also use an open fire stove, other wood stove, kerosene, and/or LP gas at any frequency. Ideally, these 40 HHs should be further broken down into approximately 30 that use an unimproved charcoal stove at least one time per day and approximately 10 that use an improved charcoal stove at least one time per day. c. HH size must be within 1 and 9 people. 3. Marsabit (among the 40 Envirofit KPT HHs and 40 JikoPoa KPT HHs): a. 5 HHs should use charcoal as their primary fuel (use charcoal 2 times per day or more) and use little or no wood (use wood 1 time per day or less). b. 5 HHs should use an open 3-stone fire as their primary stove (use 3-stone fire 2 times per day or more) and use charcoal 1 time per day or more. c. 30 HHs should use an open 3-stone fire as their primary stove (use 3-stone fire 2 times per day or more) and use little or no charcoal (use charcoal less than 1 time per day). d. LP gas and kerosene use should be less than 1 time per day for all selected households. e. HH size must be within 2 and 14 people. f. In all cases of charcoal use, the charcoal stove type is not important (use of either a traditional or improved charcoal stove, or both, is fine). A list of potential customers created by the Paradigm Project and local partner organizations was initially used to find participating HHs. The field team then applied the criteria above to this list to randomly select the KPT HHs. 4.2 Kitchen Performance Test 4.2.1 Fuel Use by Stove Type The average wood, charcoal, and kerosene use (kg/person-day) among Envirofit and JikoPoa stove users are shown in Tables 1, 2, and 3 below. January 2011 7

Table 1. Wood use among Envirofit and JikoPoa stove users (kg/person-day). Values in parentheses represent one standard deviation. 1 Stove N Before After Savings Envirofit 96 1.43 (1.16) 0.84 (0.61) 0.59 (0.84) JikoPoa 98 1.35 (1.16) 0.83 (0.49) 0.52 (1.01) p-values* -- 0.62 0.93 0.58 * Paired t-test comparing Envirofit and JikoPoa values. Table 1 above compares Envirofit and JikoPoa stove users wood use. The p-values reveal that there was no significant difference in wood use (Before, After, and Savings) between Envirofit and JikoPoa stove users. Table 2. Charcoal use among Envirofit and JikoPoa stove users (kg/person-day). Values in parentheses represent one standard deviation. Stove N Before After Savings Envirofit 96 0.080 (0.16) 0.035 (0.10) 0.045 (0.15) JikoPoa 98 0.075 (0.19) 0.039 (0.13) 0.036 (0.22) p-values* -- 0.92 0.84 0.83 * Paired t-test comparing Envirofit and JikoPoa values. Table 2 above compares Envirofit and JikoPoa stove users charcoal use. There was no significant difference between Envirofit and JikoPoa stove users day charcoal use (Before, After, and Savings). Table 3. Kerosene use among Envirofit and JikoPoa stove users (kg/person-day). Values in parentheses represent one standard deviation. Stove N Before After Savings Envirofit 96 0.009 (0.03) 0.00 (0.02) 0.009 (0.03) JikoPoa 98 0.010 (0.02) 0.010 (0.03) 0.00 (0.02) p-values* -- 0.92 0.16 0.18 * Paired t-test comparing Envirofit and JikoPoa values. 1 All fuelwood values are reported on a wet basis, as directly measured in the KPT. The average wood moisture fraction (on the wet basis) was 14% in the Before KPT and 18% in the After KPT. January 2011 8

Table 3 above compares Envirofit and JikoPoa stove users kerosene use. The p-values show that Envirofit and JikoPoa stove users kerosene use (Before, After, and Savings) were not significantly different. Hence, we conclude that all Envirofit and JikoPoa stove users that fit this customer profile are not different and may be treated as one cluster. 4.2.2 Final Clustering and Fuel Savings Fuel Use Among Non-wood HHs and All Other HHs Twenty-two HHs did not use wood before purchasing the improved stove (Non-wood HHs), as was typical of a proportion of real customers and is projected to be typical of a proportion of future customers. The fuel use within this group was different from the rest of the KPT HHs (All Other HHs), which used wood before purchasing the improved stove. Within the Non-wood HHs sub-group (N=22), 18 HHs were in Nairobi, 3 HHs were in Meru, and 1 HH was in Marsabit. Within the All Other HHs sub-group (N=172), 19 HHs were in Nairobi, 75 HHs were in Meru, and 78 HHs were in Marsabit. Table 4 below reveals that there is a significant difference between the Non-wood HHs and All Other HHs sub-groups wood and charcoal use before and after purchase of the improved stove. The changes in wood (p=0.000) and charcoal (p=0.029) use were also significant. Therefore, these sub-groups are analyzed separately. Table 4. P-values of paired t-test comparing Non-wood HHs and All Other HHs kilogram per person-day wood and charcoal use. Fuel Before After Savings Wood 0.000 0.000 0.000 Charcoal 0.000 0.006 0.029 Tables 5, 6, and 7 below compare the Non-wood HHs and All Other HHs sub-groups average wood and charcoal use. Table 8 below compares the sub-groups average kerosene use. January 2011 9

Table 5. Wood use among non-wood using households and all other households (kg/person-day). Values in parentheses represent one standard deviation. Group N Before After Savings p-value* Savings Precision Non-wood HHs 22 0.00 (0.00) 0.37 (0.26) -0.37 (0.26) 0.000 90% CI 0.00 0.092 0.092 -- 25% 30% of Mean 0.00 0.11-0.11 -- All Other HHs 172 1.54 (1.07) 0.89 (0.55) 0.64 (0.91) 0.000 90% CI 0.14 0.070 0.12 -- 18% 30% of Mean 0.46 0.27 0.19 -- * Paired t-test comparing Before and After values. Table 5 above shows that the Non-wood HHs did not use wood before purchasing the improved stove, but started using wood after purchasing the improved stove. As expected, this resulted in an increase in wood use within this sub-group (-0.37 kg/person-day). Alternatively, the All Other HHs sub-group used 1.54 kg/person-day of wood before purchase of the improved stove and 0.89 kg/person-day of wood after purchase of the improved stove, resulting in a savings of 0.64 kg/person-day of wood. As required by v3 of the Gold Standard Methodology, the 90/30 rule was applied to the wood fuel savings values, testing whether or not there is 90% confidence that the observed mean is within 30% precision of the true population value. To determine whether the sample size was large enough to meet the 90/30 rule, the 90% confidence interval was calculated using the following formula: 90% CI = [1.66 * s/ n] where s is the standard deviation and n is the sample size. If the 90% confidence interval is less than or equal to 30% of the sample mean, the 90/30 rule is satisfied and the actual fuel use sample means are used in quantifying carbon credits. Table 5 above shows the 90% confidence interval and 30% of the sample mean values for wood. Table 5 demonstrates that the 90% confidence interval savings values are less than the absolute value of 30% of the mean fuel savings. The 90% confidence interval for the Non-wood HHs was 0.092 kg/person-day and 30% of the mean was -0.11 (absolute value of 0.11) kg/person-day, yielding 25% precision. The 90% confidence interval for the All Other HHs was 0.12 kg/person-day and 30% of the mean was 0.19 kg/person-day, yielding 18% precision. The 90/30 rule was satisfied for wood January 2011 10

savings values. Therefore, the mean wood fuel savings (Table 5 above) can be used to calculate carbon credits generated from the use of the Envirofit and JikoPoa stoves. Table 6. Charcoal use among non-wood using households and all other households (kg/personday). Values in parentheses represent one standard deviation. Group N Before After Savings p-value* Non-wood HHs 22 0.23 (0.11) 0.10 (0.11) 0.13 (0.13) 0.0010 90% CI 0.039 0.040 0.046 -- 30% of Mean 0.069 0.031 0.038 -- All Other HHs 172 0.06 (0.17) 0.031 (0.11) 0.033 (0.19) 0.040 90% CI 0.022 0.014 0.025 -- 30% of Mean 0.019 0.0093 0.010 -- Savings Precision 36% 75% * Paired t-test comparing Before and After values. Charcoal use was greater among the Non-wood HHs sub-group before (0.23 kg/person-day compared to 0.06 kg/person-day among the All Other HHs sub-group) and after (0.10 kg/personday compared to 0.031 kg/person-day among the All Other HHs sub-group) purchase of the improved stove (Table 6). The greater charcoal savings within the Non-wood HHs sub-group per person-day (0.13 kg compared to 0.033 kg within the All Other HHs sub-group) were the result of fuel-switching from not using wood (and therefore using more charcoal) before purchase of the improved stove to using wood after purchase of the improved stove. Table 6 also reveals that the p-values for charcoal savings were statistically significant for both sub-groups. The table above also demonstrates that the 90% confidence interval savings values are not less than 30% of the mean fuel savings. The 90% confidence interval for the Non-wood HHs was 0.046 kg/person-day and 30% of the mean was 0.038 kg/person-day, yielding 36% precision. The 90% confidence interval for the All Other HHs was 0.025 kg/person-day and 30% of the mean was 0.010 kg/person-day, yielding 75% precision. The 90/30 rule was not satisfied for charcoal savings values. Therefore, the lower limit 90% confidence interval charcoal savings (Table 7 below) need to be used to calculate carbon credits generated from the use of the Envirofit and JikoPoa stoves. January 2011 11

Lower Limit of the 90% Confidence Interval Charcoal Savings The lower limit 90% confidence interval charcoal savings for the Non-wood HHs and All Other HHs sub-groups are shown in Table 7 below. The lower limit 90% confidence interval charcoal savings were 0.10 kg/person-day for the Non-wood HHs sub-group and 0.022 kg/person-day for the All Other HHs sub-group. These lower limit 90% confidence interval charcoal savings should be used to calculate carbon credits generated from the use of the Envirofit and JikoPoa stoves. Table 7. Lower limit 90% confidence interval charcoal savings for non-wood using households and all other households (kg/person-day). Group N Before After SE Adjusted After Non-wood HHs All Other HHs Adjusted Savings 22 0.23 0.100 0.024 0.13 0.100 172 0.064 0.031 0.009 0.042 0.022 4.2.3 Other Fuel Use A subsumed approach was used in the paired KPT. All fuels used by HHs as part of their typical cooking pattern during the Before and After KPTs were weighed. Below are results on corn husk, LP gas, and kerosene use. Nine HHs in Nairobi used corn husks as an alternative, seasonal fuel. For these HHs, the average corn husk use was 2.02 kg/hh-day before purchase of the improved stove and 0.57 kg/hh-day after purchase of the improved stove. Twenty HHs had LP gas cylinders during the KPT. Eight HHs used a small amount of LP gas (0.30 kg/hh-day) before purchasing the improved stove and six HHs used LP gas (0.30 kg/hhday) after purchasing the improved stove. Thus there was no change in LP gas use. There were, however, limitations in qualifying the amount of LP gas used because its heavy weight makes it awkward to weigh and small changes in mass, typical of the daily amounts used by these households, may be difficult to detect. Table 8 below compares kerosene use between the two sub-groups. While more kerosene was used in the Non-wood HHs sub-group, little kerosene was used in both sub-groups, and there was January 2011 12

no significant change in kerosene use. The p-values also indicate that there was no significant difference in kerosene use during the Before and After KPTs. Table 8. Kerosene use among non-wood using households and all other households (kg/personday). Values in parentheses represent one standard deviation. Group N Before After Savings p-value* Non-wood HHs 22 0.030 (0.03) 0.039 (0.05) -0.009 (0.04) 0.49 All Other HHs 172 0.000 (0.02) 0.000 (0.01) 0.000 (0.02) 0.11 * Paired t-test comparing Before and After values. In total, twenty-nine HHs used kerosene during the KPT. Table 9 below shows the average kerosene use among these HHs. Kerosene use was small. There was no significant change in kerosene use and no statistical difference at the p=0.05 level in kerosene use before and after purchase of the improved stove. Table 9. Kerosene use among households using kerosene (kg/person-day) (N=29). Values in parentheses represent one standard deviation. Before After Savings p-value* 0.050 (0.05) 0.041 (0.05) 0.011 (0.07) 0.37 * Paired t-test comparing Before and After values. As a result of weighing all fuels used in these HHs, no change in other fuel use (corn husk, LP gas, and kerosene) was found. January 2011 13

Appendix 1: KPT Field Data Collection Form KITCHEN PERFORMANCE TEST 1. Project: 2. Project Location: 3. Project Stove: 4. Phase: Before or After new stove 5. Surveyor: 6. Household ID: 7. Name (Main Cook): 8. Date (dd-mon-yy): 9. Region: 10. City: 11. Neighborhood: 12. Household Address: 13. Participant Phone number: 14. How many people eat in this household normally per day? 15. What is the No. Children of age 14 or less? 16. What is the No. females of age 15 and above? 17. What is the No. of men age between 15-59 years? 18. What is the No. of men above 59 years? 19. a) Do you cook food for sale? i) YES; ii) NO (circle the answer given) 19 b) If YES, for how many do you normally cook for to sell per day? 20. How much fuel do you typically purchase/collect in a day? Types of fuel Quantity/volume in local units (e.g. Kgs, liters, headload, etc) Cost of the fuel in Ksh a) Firewood b) Charcoal c) Kerosene d) LPG (gas) 21. Stove Type a) Number of Stoves b) No of days stove used per A. Firewood B. Charcoal C. Kerosene D. LPG (gas) 22. What type of cooking does your household perform? (circle the answer given) c) Condition of stove (tick as appropriate) i) very good i) good ii) bad iii) very bad 1. Domestic 2. Commercial 3. Both Domestic & Commercial 4. Institutional (specify): Visit #1 23. Date (dd-mon-yy): 24. Time (hh:mm): 25. New Charcoal Total (kg): 26. New Wood Total (kg): 27. New LPG Total (kg): 28. New Kerosene Total (kg): 29. Wood Moisture Sample 1: (a)reading #1: (b)reading #2: (c)reading #3: 30. Wood Moisture Sample 2 (a)reading #1: (b)reading #2: (c)reading #3: 31. Wood Moisture Sample 3: (a)reading #1: (b)reading #2: (c)reading #3: Visit #2 (~24 hours later) 32. Date (dd-mon-yy): 33. Time (hh:mm): 34. Unused Charcoal Total (kg): 35. Unused Wood Total (kg): 36. Unused LPG Total (kg): 37. Unused Kerosene Total (kg): 38. New Charcoal Total (kg): 39. New Wood Total (kg): January 2011 14

40. New LPG Total (kg): 41. New Kerosene Total (kg): 42. Wood Moisture Sample 1: (a)reading #1: (b)reading #2: (c)reading #3: 43. Wood Moisture Sample 2: (a)reading #1: (b)reading #2: (c)reading #3:: 44. Wood Moisture Sample 3: (a)reading #1: (b)reading #2: (c)reading #3: 45. Breakfast (a) No. of people cooked for: (f) Food(s): (write the appropriate codes) (d) Men aged between 15-59 years? ------------- (h) No of stoves ----------------------------------- 46. Lunch (a) No. of people cooked for: (f) Food(s): (write the appropriate codes) (d) Men aged between 15-59 years? ------------- (h) No of stoves -------------------------------- 47. Dinner (a) No. of people cooked for: (f) Food(s): (write the appropriate codes) (d) Men aged between 15-59 years? ------------- 48. Tea/coffee/other (h) No of stoves --------------------------------- (a) No. of people cooked for: (d) Men aged between 15-59 years? ------------- (f) Food(s): (write the appropriate codes) January 2011 15

(h) No of stoves ------------------------------------- 49. Other fuel usage / Notes: Visit #3 (~24 hours later) 50. Date (dd-mon-yy): 51. Time (hh:mm): 52. Unused Charcoal Total (kg): 53. Unused Wood Total (kg): 54. Unused LPG Total (kg): 55. Unused Kerosene Total (kg): 56. New Charcoal Total (kg): 57. New Wood Total (kg): 58. New LPG Total (kg): 59. New Kerosene Total (kg): 60. Wood Moisture Sample 1: (a)reading #1: (b)reading #2: (c)reading #3: 61. Wood Moisture Sample 2: (a)reading #1: (b)reading #2: (c)reading #3: 62. Wood Moisture Sample 3: (a)reading #1: (b)reading #2: (c)reading #3: 63. Breakfast (a) No. of people cooked for: (f) Food(s): (write the appropriate codes) (d) Men aged between 15-59 years? ------------ (h) No of stoves ------------------------------------- 64. Lunch (a) No. of people cooked for: (f) Food(s): (write the appropriate codes) (d) Men aged between 15-59 years? ---------- (h) No of stoves ------------------------------------- 65. Dinner (a) No. of people cooked for: (f) Food(s): (write the appropriate codes) (d) Men aged between 15-59 years? ----------- January 2011 16

66. Tea/coffee/other (h) No of stoves ------------------------------------- (a) No. of people cooked for: (d) Men aged between 15-59 years? ----------- (f) Food(s): (write the appropriate codes) (h) No of stoves ------------------------------------- 67. Other fuel usage / Notes: Visit #4 (~24 hours later) 68. Date (dd-mon-yy): 69. Time (hh:mm): 70. Unused Charcoal Total (kg): 71. Unused Wood Total (kg): 72. Unused LPG Total (kg): 73. Unused Kerosene Total (kg): 74. New Charcoal Total (kg): 75. New Wood Total (kg): 76. New LPG Total (kg): 77. New Kerosene Total (kg): 78. Wood Moisture Sample 1: (a)reading #1: (b)reading #2: (c)reading #3: 79. Wood Moisture Sample 2: (a)reading #1: (b)reading #2: (c)reading #3: 80. Wood Moisture Sample 3: (a)reading #1: (b)reading #2: (c)reading #3: (a) No. of people cooked for: (f) Food(s): (write the appropriate codes) 81. Breakfast (d) Men aged between 15-59 years? ----------- (h) No of stoves --------------------------------- 82. Lunch (a) No. of people cooked for: (f) Food(s): (write the appropriate codes) (d) Men aged between 15-59 years? ------------ January 2011 17

(h) No of stoves ---------------------------------- 83. Dinner (a) No. of people cooked for: (f) Food(s): (write the appropriate codes) (d) Men aged between 15-59 years? ----------- 84. Tea/coffee/other (h) No of stoves ------------------------------------ (a) No. of people cooked for: (d) Men aged between 15-59 years?------------ (h) No of stoves ------------------------------------- (f) Food(s): (write the appropriate codes) 85. Other fuel usage / Notes: Other Information on Visit #4 86. What kinds of pots are used for cooking (e.g. round or flat bottom, metal or ceramic, etc.)? 87. Are pot-lids usually used for cooking? 1. Yes 2. No The Questions (88-91) are applicable ONLY after supplying of the improved (EF/JP) stoves 88. Does the family perform maintenance on the improved stove? Type of maintenance: Frequency (circle appropriate response) (a) Cleaning stove of ashes 1.Never 2. Daily 3. Weekly 4. Monthly (b) Repairing cracks (c) Other task (specify): 1.Never 2. Daily 3. Weekly 4. Monthly 1.Never 2. Daily 3. Weekly 4. Monthly January 2011 18

89. What does the primary cook like about the stove (list replies)? 90. What does the primary cook dislike about the stove (list replies)? 91. Briefly describe the condition/appearance of the stoves. 92. Notes/Observations: **Take photos of the stove(s), kitchen, and home with HH ID. January 2011 19

Appendix 2: Scale Calibration Record January 2011 20

Appendix 3: Scale Calibration Certificate January 2011 21