Loading...
HomeMy WebLinkAboutMethodologyMethods This report is organized according to the items in the approved scope 1. Specifically, the report presents residential and nonresidential energy end-use data by Climate Zone for more urban communities in the Railbelt and Southeast Alaska Regions. End-use energy estimates for rural Alaskans are presented separately, and are based on the energy end use for Bethel (a representative rural hub community) and New Stuyahok, Savoonga and Selawik (representative rural communities). Criteria for rural community selection were availability of data, proximity to major hub cities and Climate Zone dispersion. Energy end-use data is also presented for infrastructure systems, including streetlights and traffic lights, water and wastewater, and non-residential buildings (specifically, rural community buildings). Finally, the report summarizes residential and non-residential end-use energy information throughout Alaska, and describes assumptions used to extrapolate the data region- and state-wide. The Alaska Retrofit Information System (ARIS), developed by the Alaska Housing Finance Corporation (AHFC) to compile residential energy information on homes served by AHFC energy programs, provided source data for Alaska residential energy consumption for heating and ventilation. The research team performed an analysis of the ARIS data’s generalizability. This analysis suggests that energy intensity in the Alaskan residential building sector sample approximates that estimated nationwide, based on the U.S. Census. For an analysis of generalizability, refer to the appendices and the AEA Energy End-use Study Methodology resource document. Finally, the Methods section of this report is only a summary of the methods employed, as the methods are described in greater detail in the AEA End-use Study Methodology resource document. Regions There were three primary regions identified as the focal point for the AEA End-use Study: the Railbelt, Southeast Alaska and the Rural North/Northwest. A description and discussion of the regions and how they were selected is provided below. Railbelt & Southeast Alaska The AEA/EUS original intent was to focus on Railbelt and Southeast Alaska as shown in Figure 1. 1 AEA End-Use Study- Implementation plan, Final Draft, WHPacific, September 1, 2011. Figure 1 Alaska Regional Zones These regions were further subdivided into three climate regions primarily to account for temperature variations, and due to the tendency to aggregate by geographic regions, including Southeast Alaska, South Central, and the interior. This is because energy use, particularly for heating and lighting in residential and non-residential buildings, is strongly influenced by the Climate Zone. Climate Zones are based on temperature variations using heating degree days (HDD) and cooling degree days (CDD). Alaska’s Climate Zones are shown in Figure 2, a map of Alaska Climate Zones2. Figure 2 Climate Zones by Census Area Climate Zones have been organized by census areas to facilitate the correlation of population and housing data. The relationship between Climate Zones and census areas is shown in Table 1. The table identifies one major climate region in Southeast Alaska (Zone 6), and divides the Railbelt region into two 2 Alaska Specific Amendments to the IECC 2009, Alaska Housing Finance Corporation, March 9, 2011, Chapter 3, p. 2, Table A301.1(1) climate regions (Zones 7 and 8). There were no study communities in Climate Zone 9. For the purposes of this study, Valdez and Cordova census areas were included in Climate Zone 6. Table 1 Climate Zones by Census Area Zone 6 Zone 7 Zone 8 Zone 9 Juneau Aleutians East Bethel North Slope Ketchikan Gateway Aleutians West Denali Prince of Wales Anchorage Fairbanks North Star Sitka Bristol Bay Nome Skagway-Hoonah-Angoon Dillingham Northwest Arctic Wrangell-Petersburg Haines Southeast Fairbanks Yakutat Kenai Peninsula Wade Hampton Valdez Kodiak island Yukon-Koyukuk Cordova Lake and Peninsula Matanuska-Susitna Rural North/Northwest The high cost of data collection in the region required a unique data collection approach. With assistance from ISER, a cluster sampling methodology was employed3 for residential and non-residential buildings in a single hub community and three smaller outlying communities to characterize energy use in the rural North and Northwest. Bethel All Alaska rural hub communities (Barrow, Kotzebue, Nome, Dillingham, and Bethel) were considered as potential representative Hub Communities. After consultation with AEA, Bethel was chosen as a representative Hub Community and targeted for additional investigations. Selection criteria included relative size, geographic proximity to a large contingent of outlying communities, and its status as a past, present, and future recipient of State of Alaska-sponsored energy programs. Bethel is a second-class city of approximately 6000 residents, located at the mouth of the Kuskokwim River, 40 miles inland from the Bering Sea, and included in Climate Zone 8. It serves as the regional center for 56 villages in the Yukon-Kuskokwim Delta. There are approximately 1,900 occupied housing units in Bethel Climate Zone. Rural communities The intent of the cluster sampling approach was to leverage existing community data and to supplement this data with field investigations, energy modeling of a community’s water and sewer infrastructure, and residential and non-residential energy end-use calculations. Through consultation with AEA, RurAL 3 For additional details on the Rural North & NW methodology, refer to the EUS Implementation plan. CAP 4, and ANTHC’s Alaska Rural Utility Cooperative 5 (ARUC), the three communities identified below were selected as the representative communities in the cluster sampling approach 6. • New Stuyahok is a second-class city of just over 500 residents in the Bristol Bay census area. There are approximately 114 occupied housing units in the community. New Stuyahok is in Climate Zone 7, although it is often considered to be in a climatic transition zone. The primary climate influences are maritime, influenced by the continental climate zone. • Savoonga, located on the northern coast of St. Lawrence Island in the Bering Sea, is a second- class city of approximately 670 residents. The community is in Climate Zone 8, and has a sub- arctic maritime climate with some continental influences during the winter. The community is iced- in much of the winter. It is part of the Nome census area. There are 166 occupied housing units in the community. The economy of Savoonga is largely based on subsistence hunting and fishing. • Selawik is an Inupiat Eskimo community of 830 residents located approximately 90 miles east of Kotzebue. Part of the Northwest Arctic borough, it is in Climate Zone 8. Most of the inhabitants rely on subsistence. There were 186 occupied housing units in the community. Residential& Non-Residential Buildings by Building Type Both existing and new data were used to estimate total residential energy use. Researchers accessed existing historical data as found in the ARIS database 7. The ARIS database, however, did not contain the level of detailed appliance, electrical, and other energy end-use data necessary for the scope of EUS. Therefore, an in-depth “electrical, appliance, and other energy” survey was developed to capture the required level of data, and combine the new electrical energy data with that derived from ARIS. Residential Building Type Categorization & Definitions Residences were divided into four basic categories 8: • Single family detached dwellings are those that are typically occupied by a single household or family, and consist of a single, stand-alone unit. A single family detached house does not share an inside wall with any other house or dwelling. Most single family homes are built on lots larger than the structure itself. In this study, it was defined as a stand-alone home. • Single family attached dwellings consist of houses built side-by-side as units sharing a “party wall.” They are typically constructed so that each houses lay out is the mirror 4 RurAL CAP’s Energy Wise program collected extensive residential energy data through an on-site data programmatic collection process. The existing data was supplemented with a simplified phone survey that collected additional structural data that enabled the end use calculations and analysis. 5 ARUC has one year of water and sewer operations data, including fuel data, available for participating communities. 6 This information was obtained from the state of Alaska Department of Commerce and Economic Development Community Database, accessed at http://www.dced.state.ak.us/dca/commdb/ 7 AHFC enabled full access to project researchers to the ARIS data base. 8 Much of this information was taken from Wikipedia, accessed December 2010. image of its twin. In this study, it was defined as a duplex, one story (i.e. no units above or below) town home or one story condominium. • Multifamily residential units include multiple separate housing units for residential inhabitants within one building, or several buildings within one complex. A common form is an apartment building. In this study, it was defined as an apartment complex, multi-story, or condo. • Mobile homes are pre-fabricated homes built in factories rather than on site and then taken to the place where they will be occupied. Mobile homes are usually placed in one location and left there permanently, but they often retain the ability to be moved. They usually contain center walls and less insulation than other types of housing. Non-residential Building Type Categorization and Definition There is little consistent data on non-residential building in Alaska and its Climate Zones. The research team reviewed available tax assessor data from Alaska boroughs and municipalities across Alaska. The level of detail varied by jurisdiction, with some jurisdictions maintaining very detailed data on each parcel, and others collecting minimal data. As noted in the AEA End-use Study Methodology9, the Municipality of Anchorage (MOA) parcel data appeared to be the most detailed in the state. The MOA parcel data was used to develop the list of building types for representative samples of non-residential facilities. The method for combining building sub-types into categories is shown in Table 2. The building sub-type provides the research team’s definition of these building types. Table 2: Anchorage Building Type and Building Distribution 9 AEA End use Study- Implementation plan, Final Draft, WHPacific, September 1, 2011, page 38-39. Anchorage Percent Distribution of Building Types Building Type Building Sub-type & Definition Anchorage Parcel Data Building Type Totals Building Type Percent Food Services Restaurant 133 214 4.5% Food Stand 0 Fast Food 48 Bar/Lounge 33 Night Club 0 Warehouse and Storage Hangar 173 1704 35.9% Refrigerated Warehouse 0 Warehouse—General 1531 Institutional Education 103 277 5.8% Public Assembly 0 Public Order & Safety 0 Religious Worship 174 Library 0 Cemetery 0 Institutional—Other 0 Health Care Health Care—Inpatient 0 69 1.5% Health Care—Outpatient 69 Nursing Home 0 Lodging Hotel/Motel 74 113 2.4% Dormitories 39 Home for elderly 0 Lodging—Other 0 Residential & Non Residential Data Use & Data Collection Methods This section describes how data was collected for residential and non-residential data, including how: • Existing ARIS data were utilized for residential DHW and heating end-use analysis; • Survey methodology was employed to collect supplemental residential electrical/appliance data and perform the energy end-use analysis; and • Survey methodology was employed to collect supplemental non-residential electrical/appliance, domestic hot water (DHW), and heating, ventilation and air conditioning (HVAC) data, and perform the energy end-use analysis. Residential Heating and Domestic Hot Water (DHW) ARIS Data The ARIS 10 data set was used as the principal source of information for residential and non-residential public facility energy use for sources other than electricity, including domestic hot water and heating. The ARIS system, which collects, manages, and reports information relating to AHFC’s Home Energy Rebate, Weatherization, and Public Facility programs, was developed by CCHRC, and is funded by AHFC. The Alaska Home Energy Rebate Program requires a certified home energy rater to conduct a detailed building energy rating. The rater must use AkWarm©11 to analyze building energy performance. The ARIS database contains information on residential energy use collected between 2008 and 2011. It includes calculations of energy end-use, by fuel type, for building HVAC and DHW systems. However, ARIS does not contain detailed end-use data on interior/exterior lighting, appliances, cooking and other end-use data for residential buildings. For this reason, a supplemental survey was conducted to obtain information on these end uses. The end-use energy calculations from the two data sets were 10 Alaska Retrofit Information System (ARIS) developed by the Cold Climate Housing Research Center, accessed at http://www.cchrc.org/aris- development 11AkWarm© is building and energy simulation model using software tools that analyze building energy use on an hourly basis using detailed climactic data, and detailed building envelope and HVAC system data obtained through the energy audit. Office 701 701 14.8% Mercantile and Retail Strip Malls 56 681 14.4% Enclosed Malls 201 Food Retail 35 Retail—Other 389 Service Cinema/Theater 0 260 5.5% Automotive Oriented Services 232 Spa/Salon 0 Communication 0 Service—other 28 Other Parking 0 722 15.2% Sports facilities 32 Multipurpose 0 Miscellaneous--Other 690 Totals 4741 4741 100.0% combined for each region and Climate Zone. The initial data set of “As Is” residential energy ratings contained 29,466 records12 and the regional distribution of data is shown in Table 3. Table 3 Regional Breakdown of Number of Home Energy Ratings, per ARIS Region Frequency Percent Aleutians 2 .0 Bristol Bay 10 .0 Copper River- Chugach 2 .0 Lower Yukon- Kuskokwim 5 .0 North Slope 35 .1 Railbelt 25275 85.8 Rural North-West 536 1.8 Southeast 3572 12.1 Yukon Koyukuk- Upper Tanana 7 .0 Location Information Missing 2 .0 Total 29446 100.0 The ARIS data were also identified by Climate Zone. Table 4 shows that most of the ARIS observations were in households in Climate Zones 6, 7 and 8. Information on the location of the residence or Climate Zone was only missing in three of the nearly 30,000 cases. Table 4 shows the distribution of only those cases that had the data. Table 4 Climate Zone breakdown of Number of Home Energy Ratings, per ARIS Climate Zone Frequency Percent 6 3,466 11.8 7 21,369 72.6 8 4,563 15.5 9 45 .2 Location Information Missing 3 .0 Total 29446 100.0 ARIS “As Is” data on homes within Climate Zones 6, 7 and 8, and in the Railbelt and Southeast region were used for domestic hot water and heating end-use calculations. The final dataset is shown in Table 5. Table 5 Final ARIS Dataset for Number of Home Energy Ratings within the Railbelt and Southeast Alaska Region Climate Zone 6 7 8 Railbelt 0 20944 4331 Southeast 3465 106 0 Total 3465 21050 4331 12 “As Is” records were chosen as the principal measure because this database does not include post rating ARIS data collected after a homeowner completed projects to improve energy efficiency. Residential & Non-Residential Supplemental Data Collection Methods Survey data were used to supplement available data, especially regarding residential appliance and electrical use. The research team employed a multi-dimensional data collection process to ensure survey participation by respondents. This process included CATI (computer assisted telephone interviewing) technology and The Survey System 13 software, which allowed for CATI and Web survey integration. The CATI system allowed for data collection of randomly selected Alaska residences and non-residential buildings, supplemented with on-site surveys and facility walkthroughs of selected non-residential facilities in the various Climate Zones, in particular for more complex non-residential buildings. On-site methodology was also used in some rural locations where access to telecom and broadband technology was limited. Interviewers used laptop computers and went to locations throughout Alaska to complete surveys. The on-site methodology was employed in 37% of the buildings. Residential & Non-Residential Survey Instrument Development The residential survey instrument was designed to supplement the ARIS data set by collecting additional electrical and appliance end-use information. The questions asked were driven by the specific needs of the end-use analysis and AEA directives. The survey process employed a screening strategy to identify the right person to answer salient questions. The residential survey instrument included a variety of energy end-use questions, including refrigeration, cooking, office equipment, energy consuming appliances, and lighting. The electrical energy use for each category includes a number of sub- categories, outlined in Table 6. The survey instrument directed the interviewer to inquire from building owners/managers/maintenance people basic information on facility characteristics and energy use to calculate a baseline for energy end uses for the non-residential building. For example, under “lighting,” researchers collected data that specified what types of lighting were currently in use, the wattage of the bulb(s), and the hours the lighting was in use. The data obtained from this survey included building types, fuel types, and end uses. The final Residential Supplemental Survey instrument (“residential instrument”) had a total of 218 questions, and is provided in the appendices. Participation was enhanced by offering cash incentives and the chance to be a part of the “State of Alaska’s first-ever study to understand energy use in buildings.” Many respondents indicated that they participated in the survey to improve the state’s understanding of end-use energy consumption. 13 Trademark registered. Creative Research Systems, Petaluma California. Table 6 Electrical Energy End-use Categories and Sub-Categories Electrical end-use category End-use sub-category Interior lighting Fluorescent, Incandescent, CFL, Small halogen, Large halogen, LED, Other Exterior lighting Fluorescent, Incandescent, CFL, Small halogen, Large halogen, LED, Other Major appliances Refrigerators, Freezers, Dishwashers, Washer/Dryer Cooking Primary cooking, Microwave Other kitchen equipment Coffee maker (Total) (Drip, Percolator, Espresso); Electric deep fryer, Electric frying pan , Electric kettle , Slow cooker, Toaster, Toaster oven Entertainment Television, Gaming, DVD player, VCR, Digital video recorder (DVR), Stand-alone cable box, Cable box with DVR, Music system, Satellite dish, Other Information technology Desk top computers, Stand-alone monitors, Laptop computers Combined inkjet printer, Combined laser, Router/DSL/Cable Modem, Stand-alone copy machine, Stand-alone fax machine, Stand-alone printer, Other network equipment, Other computer peripherals, Other small electronics Seasonal decorative lighting No sub-categories Miscellaneous appliances Garage door opener, Electric waterbed, Hot tub, Water well pump, Sewage lift pump, Sump pump, RV trickle charger, Engine block heaters, Heat trace/heat tape, Electrical vehicle charging Detailed survey data provided usage information for items in each sub-category. This information was used to calculate the total number of kilowatt hours (kWh) consumed within each sub-category. Sub- category consumption data were aggregated into the major use categories, and then converted into MMBTUs to allow comparisons between different energy sources and dimension The final Non-Residential Supplemental Survey instrument (“non-residential instrument”) included additional questions to estimate DHW, HVAC, and other end-use energy practices specific to non- residential buildings. The final non-Residential Supplemental Survey instrument included 667 questions designed to integrate with programmed use on the CATI and WEB systems. Interview logic was embedded in the instrument to control the manner in which question and answer choices were presented to each respondent. This entailed the use of “question skip” patterns, algorithms, randomization of question sequences, and the presentation order of answer choices. Interviewer Training- Residential & Non-Residential Interviewers were trained on relevant energy end-use language and topics in order to relate to study participants. This included instruction on insulation, furring, furnace and boiler, CFL vs LED lighting etc. Training also included topics specific to regional differences (e.g. Interior vs. Southeast; fuel oil vs. natural gas). Interviewers were prepared to capture unique Alaska energy uses both formally through the survey, and via informal conversations and queries with respondents. Standard protocol for the web surveys included monitoring tools and numerous (up to 25) call-back procedures. These procedures were designed to maximize contact with respondents, to maximize participation, and yield a high degree of survey completeness. Sampling Method Residential Supplemental Energy Data In order to make statistically credible inferences or generalizations about the state’s current building stock, it was essential that scientifically sound, random sampling be employed. At times this required the use of Municipal tax rolls, geographic or cluster stratification, and sampling quotas. All of these tools were used to assure statistical validity. Random sampling was critical to counter the “self-selection bias 14” that prevails when respondents prefer to participate in the web survey, rather than via interview, thus removing researcher controls. With rigorous control over the sampling process, a telephone/web/on-site methodology produced findings at a 90% confidence level.15 For the overall sample size, if researchers sought to interview building owners/managers in Alaska using the same questionnaire, the findings would differ from the overall survey results by no more than 20 percentage points in either direction. Thus, the margin of error is +/- 20%; for sub-groups the sampling error would be larger. Information on residential electrical energy use was obtained through a stratified random sample of households in Climate Zones 6, 7 and 8, corresponding to both the Railbelt and Southeast Alaska regions. Sample sizes for each of four different housing types were calculated, to ensure adequate representation of each type of housing. The sample sizes are shown in Table 7. Table 7 Residential Projected and Actual Samples Housing Type Climate Zone 6 Climate Zone 7 Climate Zone 8 Total Sample Units 16 Sample Size Units Sample Size Units Sample Units Sample Single Family Detached 13,611 16 82829 21 17452 14 113,892 51 Single Family Attached 1,389 16 12745 22 2685 12 16,819 50 Multifamily 9,999 16 54160 24 11411 11 75,570 51 Mobile Home 2,778 17 9717 19 2047 15 14,542 51 Total 27,777 65 159451 86 33595 52 220,823 203 14 Choice to complete the survey was self-motivated based on factors outside researcher controls. 15 This study provides a 90% confidence that the actual population statistic is within +/- 20% of the sample finding. 16 2010 US Census, Housing data. Sampling Method- Non-Residential Supplemental Energy Data The non-residential sample size for this study was based on the distribution of non-residential facilities taken from the Anchorage tax rolls. A detailed description of these categories is shown previously in Table 2. The resource limitations for this study resulted in a fairly substantial margin of error. For most of the building types and climate zones, the margin of error was set at 20%. The number of buildings of each type and within each climate zone is shown in Table 8. Table 8 Non-Residential Projected and Actual Samples Non-Residential Building Type Zone 6 Zone 7 Zone 8 Total Units Sample Units Sample Units Sample Units Sample Food Service 82 21 612 17 136 7 830 45 Warehouse 647 17 4728 17 997 25 6372 59 Institutional 87 10 979 13 162 22 1228 45 Health Care 28 11 243 16 40 10 311 37 Lodging 62 17 456 19 81 9 599 45 Office 256 25 1663 21 406 15 2325 61 Mercantile/ retail 258 16 1878 24 398 18 2534 58 Service 99 18 1017 16 152 17 1268 51 Other 268 10 1737 16 415 15 2420 41 Total 1,787 145 13,313 159 2,787 138 17,887 442 Survey Administration Supplemental data collection for residential and non-residential buildings used a telephone survey. Responses were documented using CATI technology. Complex algorithms and skip patterns were used to present a customized set of questions to each respondent. This interactive adaptability of the survey question patterns based on real-time replies keeps the interview relevant to the respondent and enhances an informal atmosphere conducive to respondent participation. CATI techniques were required to gather the depth of information necessary to meet the objectives of the end-use study. Initial calls were made to building owners to secure their participation. Upon securing permission and identifying the survey participant (building manager, owner or property manager), the survey respondents were offered a CATI, web-based, or hybrid combination approach. With CATI, data is entered only once, thus increasing accuracy and reducing the turn-around time it takes to complete manual telephone surveys with the necessary transcription. This computer technology also facilitates call-back attempts or appointments and assures that hard-to-reach respondents are interviewed to eliminate potential bias from limiting interview to easily-reached respondents. The survey team also requested from respondents permission to access their utility records. This information may be available for use in verifying energy end-use consumption calculations. A survey took, on average, one hour and forty-five minutes to complete when it was done with a professionally trained interviewer on the telephone. On-site interviews took, on average, two and a half hours, as they usually included facility tours. Incentives were offered to encourage respondents to complete the survey. Those participating in the Residential survey received $50, and businesses that participated in the Non Residential/Commercial survey received $100. Data Analysis- Residential & Non-Residential Survey and ARIS data were prepared for analysis in Microsoft Excel. Extensive look-up tables were compiled to calculate the amount of energy used in each category and sub-category. All energy use was converted into MMBTUs to assure compatibility of energy measures across fuel types. Most analyses were completed in Excel or SPSS. Most responses were statistically analyzed utilizing simple frequency distributions and descriptive measures. Included in the presentation of each response is a summary or example of any significant findings, followed by relevant tables. All percentages in the narrative are rounded to the nearest whole percentage point. Where respondents failed to answer a question, the data point is not included in the total used to calculate the percentage, unless the number that failed to answer is statistically significant. Percentages in the tables occasionally do not add to exactly 100% due to rounding. Cross tabulations describe data that may be related in some way. In many cross tabulations, categories are combined or omitted because the numbers are too small to be statistically significant. This manipulation may change the totals on which percentages are based, but does not affect the relationships between percentages. Cross tabulations may be used to indicate differences (or lack of differences) between subgroups of data points. When a lack of difference is shown, a footnote is appended to the table indicating that the differences are not "statistically significant.17 Rural North & Northwest- Methods The Rural North & Northwest methodology is based upon a cluster sampling methodology, as recommended by ISER. The methodology relies upon the availability of existing residential energy end- use data 18, a water and sewer energy model, and on-site collection of non-residential building data. The communities of Bethel, New Stuyahok, Savoonga and Selawik were selected as representative rural communities to create a comprehensive community energy end-use assessment, excluding transportation. The general approach is outlined below: • Conduct additional analyses of existing energy data sources (Energy Wise & ARIS); • Represent energy use of three village communities and one Hub Community; • Leverage key informants within the various communities; • Stratify by size and class (Hub Community and village); 17 Statistical significance is determined by using a chi-square test, analysis of variance (ANOVA) or t tests of means, with a significance factor of less than .05. The chi-square, ANOVA and t- tests are used by researchers to determine whether a result may be due to random variation, it is sensitive to sample size, since large random variations may occur in small samples. 18 RurAL CAP collected residential energy use data for 2,300 existing homes through their Energy Wise Program. The data set appears to have detailed energy end-use data. Three of these communities were selected for supplemental data collection. • Use combined data sources to develop an energy end-use framework of existing residential and non-residential rural energy consumption; and • Collect non-residential field data on site in the targeted three communities. The proposed rural cluster sample is identified in Table 9. Table 9 Rural Communities and Available data Community Region Available Data Villages Selawik NW Arctic Borough; NANA Region Energy Wise Data; ANTHC Energy Audit Data; Duplex High School Addition; Service to 5 Plex; Tank Farm SVC; Teacher Housing; Savoonga Norton Sound; Kawerak Energy Wise; ANTHC Water & Sewer Data; Hogarth Kingeekuk Sr. Memorial School New Stuyahok Bristol Bay Region; SW Alaska Energy Wise; ANTHC Water & Sewer Data; and AHFC REALS Data; School building (electrical/appliance data only) HUB Bethel SW AK ARIS Data; AHFC REALS Data; Village Non-Residential & Residential Sample and Methods The stratification of the sample is based upon a cluster sampling approach 19. The method focused on the four communities above with a high sampling fraction in each. The methodology was employed largely due to the high cost of collecting data in rural Alaska. It is likely that the variation in use between individual buildings within a rural community is greater than the variation in average use between communities. This variation is minimized by collecting data from a high proportion of all buildings in a small number of communities. Non-residential data collection and analysis methods for this region were based upon the same methods used for non-residential buildings in Climate Zones 6-8. A field technician performed walk-throughs of all non-residential buildings in the three village communities, including the school, tribal buildings, power plant, street lights, and water and sewer system. The residential data collection sheet and supplemental data collection sheet are found in the appendices. ARIS data was used to supplement the thermal analysis at the home level. Residential data from RurAL CAP’s Energy Wise program were used by researchers for heating fuel use, major appliance use, small appliance use, and other data that provided the majority of inputs needed to perform a residential end-use analysis. This data had been previously reviewed by researchers at ISER and formatted to facilitate analysis. To supplement this data, an additional survey was developed and implemented with program participants. 19 Due to the nature of a cluster sampling approach, an estimated margin of error cannot be performed with the information that the research team has available. Bethel Residential Sample & Methods All residential data was collected on-site by a professionally trained interviewer. The Bethel Residential sample size, including estimated margin of error, is identified in Table 10. Table 10 City of Bethel Residential Sample Size Estimates Sampling Size Estimates, Bethel City Households 90% Confidence Level, Margin of error = 10% Household Characteristics Number Size Percent Families with children under 18 869 63 7.2% Non Family and Families without children under 18 1027 64 6.2% Total 1896 127 6.7% 2010 US Census Data, Bethel City, accessed on August 17, 2011 at http://live.laborstats.alaska.gov/cen/dp.cfm#h Existing data were compared with the expected values and are presented in the body of this report. In addition, the data was checked against the four housing categories used in Railbelt and Southeast Alaska, including: • single family detached • single family attached • multi-family • mobile homes The Bethel residential end-use analysis followed the same protocol as the residential analysis for Climate Zones 6-8. The Bethel end-use analysis results are located in the “Rural North & Northwest” section of this report, and include a database of Bethel residential end-use energy calculations and general household energy use characteristics. Bethel Non-residential Sample and Methods The non-residential building inventory and sampling framework was based upon the City of Bethel’s commercial water and sewer customer list and was used to develop the preliminary Bethel non- residential building sampling plan. There were a total of 195 units on the City of Bethel’s customer list. The non-residential data collection was performed by a Bethel-based building science technician and collected on-site through interviews with facility managers and facility walkthroughs. The non- residential end-use analysis as described in the AEA End-use Methodology was utilized for Bethel. The Bethel non-residential sample sizes were calculated based on an estimated 90% Confidence interval, a 50% prorated response distribution, and an estimated margin of error of 10%. The building type quotas are included in the Table 11. Table 1 Bethel Non-residential sample estimate Bldg Type Total Units Percent Category Units Sampled Food Service 8 4.1% 2 Warehouse 14 7.2% 2 Institutional 44 22.6% 10 Health Care 4 2.1% 2 Lodging 5 2.6% 3 office 40 20.5% 8 Mercantile/ retail 15 7.7% 6 Service 60 30.8% 11 Other 5 2.6% 5 Total 195 100.0% 49 Independent Energy End-Studies Methods The Alaska EUS also included independent energy end-use studies as summarized below. These studies often included analysis of other selected public services and infrastructure and are typically managed by local or state public utilities or organizations. These independent studies relied on data provided by individual cities, building owners, ANTHC/ARUC, and other participating entities. Each study was described in the approved implementation plan, is summarized here, and is further described in the AEA End-use Study Methodology. Statewide Street Lighting Energy Use AEA requested energy end-use data on street lighting to evaluate energy efficiency upgrade opportunities, such as conversion to LED lamps, hi/low dimming coupled with occupancy sensors, etc. Detailed street lighting data was collected from the Municipalities, DOT, and other agencies that own and maintain this data. The team integrated and analyzed existing street lighting data. To the extent possible within the limitations of the data available, this report provides a detailed energy end-use breakout by lamp type. The researchers also conducted a survey of street lighting (fixture counts, bulb types, and bulb wattages) in the rural villages and rural hub during the site visits for the rural sector field study. Water and Wastewater Treatment Statewide Energy Use & Modeling AEA requested documentation on energy consumption for water and wastewater treatment plants. Initially, the data was expected to come from the Alaska Rural Utilities Collaborative (ARUC). Consistent data, however, proved to be incomplete, and a system of modeling the utilities was developed using information from ANTHC, ARUC, community maps, community profiles, and local sources when possible. A spreadsheet model of each utility was based on the information collected and supplemented by reasonable engineering assumptions. Energy requirements for systems operating temperatures between 35° and 60°F and ambient temperatures as low as -40°F. The community-specific energy requirement estimates for water and wastewater utilities generated using a final model were peer reviewed. Non-Residential Rural Community Buildings Benchmark Study AEA requested data that represent a broad dispersion of non-residential community buildings. The research team collected non-residential and community profile information from rural communities in order to support energy use estimates and extrapolations performed by ISER. Data secured includes facility square footage, year built, and other available information. Each community’s buildings have a variety of owners, including private, local government, school district and state of Alaska. The Non- Residential Rural Community Report found herein is facility benchmark data only. The study initially considered 390 communities for inclusion. The final number included was 219. Communities included in the study had to be: • Communities that are Rural. • Communities that are not considered hub communities. • Communities with a population of less than 1,000 residents. The team began by identifying contact information for non-residential building owners in each community through a search of the State of Alaska Community Information System (CIS) database, telephone books, and the internet. In the interest of time and budget, it was decided that the list of building types included in the study would focus on the following: • City Offices • Tribal Offices • Village Corporation Offices • Clinics • Schools • Grocery stores • Post Offices • Churches The buildings included in this list are common to most rural, non-hub communities in Alaska. Focusing on these buildings helped ensure that greater consistency in data results would be achieved. However, when information for other building types was readily available, data for those buildings was pursued and collected. Building owners/occupants were initially contacted by telephone. While the majority of data was collected through verbal contact, quite often respondents preferred to submit the information by facsimile or email. Attempts to collect information for each building or building owner (when multiple buildings were owned by the same entity) were limited to 3 attempts. Once the interviews were conducted, the data was entered into Microsoft Excel spreadsheets before being merged into a Microsoft Access database. Data Generalizability It was expected that the overall characteristics of residential households participating in this survey would be similar to the characteristics of all Alaskans households (US 2010 Census) and to households included in the ARIS database. Appendix C summarizes areas of comparison of these three data sources. There appear to be few material differences between the three sources, suggesting that the sample may be representative of Alaska, and that ARIS data can be merged with survey research data to give a better picture of overall energy consumption patterns. Our conclusion is that the residential ARIS data regarding heating and domestic hot water and supplemental survey data on electrical use can be merged to provide a reliable estimate of the use of energy by Climate Zone, type of residence, and end-use category. Non-residential data can also be generalized to the larger population, within the margin of error specified in the sampling plan. Rural data may be more difficult to generalize to other Alaskan communities. The decision to select a hub community to represent all Alaskan hub communities is threatened by the lack of information regarding energy use of similar communities in different Climate Zones. Additionally, the selection of three communities, each in a different climactic environment, may be difficult to generalize to communities across the state. Specifically, there is no small community in Southeast Alaska upon which to base energy use projections. Nonetheless, the calculations of residential and non-residential energy use estimates should be considered as the strongest data available upon which to estimate statewide energy use. The streetlight study includes energy use for street lighting in communities of various sizes. This level of aggregation may increase the reliability of the data for application to communities of similar size. The water and wastewater model appears to be applicable to most rural communities, assuming that basic operating information for each utility is available.