Energy Use Survey Data Collection

Data source The New Zealand Energy Use Statistics Programme resulted from the Domain plan for energy sector 2006–2016 that was published in 2006. The energy domain plan was produced by Statistics NZ in collaboration with the Energy Efficiency and Conservation Authority (EECA), and the Ministry of Business, Innovation and Employment (MBIE). The domain plan identified energy use statistics as a key gap in energy information and prioritised a suite of energy use surveys. The Energy Use Survey delivers information to help fill the gap and to provide a benchmark of energy use for New Zealand's economy, excluding households. Data from the survey also feeds into modelling systems that give current energy-use estimations and future demand forecasts. Modelling assumptions can then be updated, which improves the accuracy of modelled information. Non-response and imputation Unit non-response Unit (or complete) non-response occurs when units in the sample did not return the questionnaire. We then adjust the initial selection weight of the remaining units in the stratum to account for the unit non-response (item non-response imputation does not occur for units that did not return the questionnaire). Item non-response Item (or partial) non-response occurs when units return the questionnaire but some questions are not answered. We carry out item non-response imputation for units that answered some but not all the questions they were required to (based on questionnaire routing rules). We classify respondents who did not answer any of the questionnaire as unit non-responses and the weights are adjusted accordingly. We impute for item non-response as follows. Imputation of numeric variables We use random donor imputation to impute for numeric variables. In this method, the responses of a randomly selected donor from within the same imputation cell as the non-respondent are imputed in the recipient unit. Donor imputation is used so the distribution is maintained. Imputation of categorical variables We use random donor imputation to impute for categorical variables. The donor supplies responses for all categorical variables requiring imputation. If the donor unit does not respond to any of the variables requiring a response, then the next-best donor is selected to supply this information. This is continued until all the variables have a response. Special treatment In 2018, six respondents (0.2 percent) to the NZ Energy Use Survey received ‘special treatment’. Special treatment is applied in rare cases where response has an undue influence on survey results. For example, if an enterprise provides an extreme outlier which would unduly impact the survey results, their enterprise is assumed to be unique and for that reason its survey weight is reduced so that they only represent themselves. The removed weight is then redistributed evenly over confirmed results for similar enterprises. Energy units standardised We collect information on energy usage in the unit that applies to each commodity – for example, litres for petrol and kilowatt hours (kWh) for electricity. We convert these units to a standard unit (terajoules, TJ) for reporting. This conversion enables the energy contained in different forms to be compared directly. We applied a calorific value (enthalpy value) to each energy type and form for the conversion. We source or derive calorific values from MBIE's Energy in New Zealand File 2013. See the table below for the calorific values for each energy type. Energy types and their calorific values Energy Type Details Calorific value Electricity Electricity's standard universal unit, the watt, is defined as one joule per second 3.6 MJ/kWh Petrol Two main forms of petrol: regular and premium, and each has a slightly different conversion factor. We use a weighted average of the two values, according to their current prevalence in the market 35.08 MJ/L Diesel The value used is that of regular diesel 38.45 MJ/L Fuel oil Two types of fuel oil: light fuel oil and heavy fuel oil. We derived the conversion factor using a weighted average of the two, according to their current prevalence in the market 40.70 MJ/L LPG Liquid petroleum gas figures were provided in both litres and kilograms 49.51 MJ/kg 26.44 MJ/L Aviation fuel Two major forms: jet fuel and aviation gasoline, conversion factor is a weighted average of the two according to current prevalence in the market 34.55 MJ/L Natural gas Most natural gas figures were provided in joules; some in kilowatt hours which is converted to joules 3.6 MJ/kWh Coal Bituminous 29,250 MJ/tonne Sub-bituminous 20,120 MJ/tonne Lignite 15,340 MJ/tonne Where the type was not known, the conversion factor was a weighted average of the three above 23,432 MJ/tonne Wood and wood waste Hog fuel or bark 9,060 MJ/tonne or 7,701 MJ/cm3 Sawmill residues of fuel wood 12,080 MJ/tonne or 10,268 MJ/cm3 Black liquor 10,500 MJ/tonne or 8,925 MJ/cm3 Joinery, building, or furniture residues 17,790 MJ/tonne or 15,122 MJ/cm3 Oven-dried wood 20,550 MJ/tonne or 17,468 MJ/cm3 Where the wood type was not known the conversion factor was a weighted average of the other types 13,996 MJ/tonne or 11,897 MJ/cm3 en-NZ