Greenhouse gas emissions (industry and household): Quarterly

The quarterly greenhouse gas emissions account produces timely estimates of New Zealand’s gross production-based emissions, combining data from the annual greenhouse gas emissions by industry and household account and indicator data to estimate emissions on a quarterly basis.

This is done within the System of Environmental-Economic Accounting (SEEA) framework in order to track emissions in relation to economic activity.

^en-NZ

Quarterly emissions data provides more timely information on New Zealand’s greenhouse gas profile. This data can also be useful to examine the impact of economic shocks, such as that resulting from COVID-19. Measuring GHG emissions on a quarterly basis, using an industry classification consistent with those used in producing gross domestic product (GDP), allows us to quickly identify what emissions path the economy is on, and whether emissions are decoupling from economic growth. Quarterly information also allows turning points to be identified with greater accuracy which may be useful for decision makers.

^en-NZ

Quarterly greenhouse gas emissions (industry and household): sources and methods, introduces the methodology and indicators used to produce New Zealand’s quarterly greenhouse gas production-based emissions estimates, within the System of Environmental-Economic Accounting (SEEA) framework.

About the quarterly greenhouse gas emissions estimates

Quarterly emissions data provides more timely information on New Zealand’s greenhouse gas profile. Measuring GHG emissions on a quarterly basis, using an industry classification consistent with those used in producing gross domestic product (GDP), allows us to quickly identify what emissions path the economy is on, and whether emissions are decoupling from economic growth. Quarterly information also allows turning points to be identified with greater accuracy and timeliness, and can provide a more reliable estimation of the impact of sudden shocks and technological changes.

Initially released as an experimental series in 2021, the quarterly GHG emissions (industry and household) estimates were further developed, internationally peer reviewed, and improved over the course of a year, to become an official statistical series in July 2022. This was done within the System of Environmental-Economic Accounting (SEEA) framework to track emissions in relation to economic activity. Quarterly emissions are released 3 - 7 weeks after quarterly gross domestic product, averaging around 4.5 months after the reference quarter.

SEEA suite of emissions statistics

The quarterly emissions statistics are part of the suite of SEEA emissions statistics produced by Stats NZ. The current suite includes (on an annual basis):

• national estimates from production and consumption perspectives
• regional production emissions
• tourism emissions (both production and consumption).

All SEEA emissions statistics have the same underlying measurement framework and are based on concepts and classifications consistent with economic statistics, allowing for joint presentation. The value of joint presentations (that is, displaying emissions data alongside economic data) is it allows for analysis of emissions intensity and decoupling indicators. The quarterly series uses the national annual production estimates as benchmark data and is used to derive provisional estimates for the regional emissions series.

Scope of quarterly greenhouse gas emissions

In order to track emissions with gross domestic product (GDP), the same principles, concepts, and classifications as the annual SEEA greenhouse gas emissions by industry and household account are applied (Stats NZ, 2020b). This defines the scope of the quarterly emissions series in terms of its perspective, coverage of emissions and population, and its compilation and presentation.

Gross emissions

The annual greenhouse gas emissions by industry and household account records the gross emissions of greenhouse gases relating to the energy, industrial processes and product use (IPPU), agriculture, and waste sectors, but excludes the land use, land use change, and forestry (LULUCF) sector. The exclusion of LULUCF follows Eurostat (2015) guidance. More detail on how LULUCF is accounted for in different measurement frameworks is presented in Stats NZ (2020a). This focus on gross emissions allows for a closer conceptual comparison between economic statistics and emissions.

Residency principle

The production boundary of quarterly emissions on a SEEA basis is that of economic residents (residency principle). This contrasts with the GHG Inventory and the Ministry of Business, Innovation and Employment’s (MBIE) New Zealand energy quarterly, which measures emissions on a territorial basis, or all the emissions that occur within New Zealand’s exclusive economic zone (EEZ). The residency principle means that all emissions from New Zealand’s residents are recorded, regardless of where the emission occurs. More discussion on the residency principle can be found in Stats NZ (2020a).

The application of the residency principle means that emissions by economic residents abroad are included, while emissions from non-residents within New Zealand’s territory are excluded. This is to enable comparisons with other economic statistics, such as GDP. This leads to an observable difference with the Inventory — with the SEEA annual production estimate being higher in total, with most of the adjustments relating to carbon dioxide. The magnitude of this difference varies from year to year. For example, in 2007, the first year of the time series, the difference between the Inventory and the SEEA estimate was 3.6 percent. In 2020, this difference fell to 1.6 percent as emissions from economic residents abroad decreased due to COVID-19 border restrictions.

Industry classification

The use of an industry classification enables the aggregation of multiple emissions processes to a given economic unit in the population. For consistency with economic statistics, units for estimating emissions by industry are allocated using a standard industrial classification. For New Zealand, this is the Australian and New Zealand Standard Industrial Classification 2006 (ANZSIC06) . The ANZSIC06 classification groups producing units according to the predominant economic activity of the operator and assumes all units within an industry convert inputs to outputs in the same way.

The industry classification (which groups similar producing units) differs from the process classification used in the Greenhouse Gas Inventory and facilitates the integration of emissions data with other economic statistics. The industry approach aligns the physical processes that create emissions with the economic value created, as captured in GDP. For example, agricultural outputs utilise energy and agricultural processes to produce primary goods, the output of the agricultural industries. This relationship between processes and production is discussed further in Benchmarking, interpolation, and extrapolation on indicator types.

Industry breakdown

As quarterly emissions are estimated using indicators (see Estimation of quarterly emissions), the level of industry detail will be much lower than the annual production emissions estimates, which are available for 116 industries. At this point in its development, quarterly greenhouse gas emissions (industry and households) are available for seven industry groups, which are aggregations of industries based on the ANZSIC06 classification.

The seven industry groups are:

• agriculture, forestry, and fishing
• mining
• manufacturing
• electricity, gas, water, and waste services
• construction
• transport, postal, and warehousing services
• services excluding transport, postal, and warehousing.

As with the annual production emissions estimates, there are also estimates for households’ direct emissions for the following categories: transport; heating/cooling; and other.

Gas coverage

The scope of these estimates are all greenhouse gases as covered in the GHG Inventory and the SEEA annual estimates.

Greenhouse gases currently included are:

• carbon dioxide
• methane
• nitrous oxide
• sulphur
• hexafluoride
• perfluorocarbons
• hydrofluorocarbons.

All gases are converted and aggregated into carbon dioxide equivalents (CO₂-e) using the global warming potentials from the Fifth Assessment Report (IPCC, 2014) based on a 100-year time frame. Data for the seven industry groups are presented in CO₂-e, while data for broad groups (primary, goods-producing, services, households) are available by gas.

Actual and seasonally adjusted estimates

Quarterly emissions by industry and household have been produced on an actual and seasonally adjusted basis.

Actual emissions refer to estimates that indicate the actual amount of emissions that occurred within a period, in this case a quarter. The data is useful for associating emissions with the economic activity that occurred over the same period.

Seasonally adjusted estimates remove the seasonal aspect of emissions, leaving the trend and irregular aspects of emissions. This makes any quarter-to-quarter comparisons more meaningful and removes the regular seasonal pattern from any assessment of decoupling.

Time series

The quarterly emissions series begins from the March 2010 quarter. Estimates are only available from 2010 onwards due to the availability of data sources for complete coverage.

Presentation

The quarterly series are released in excel tables showing the distribution of emissions by industry and household. Two analytical series that provide insight into the drivers of emissions are also presented with the data:

• Emissions per household – expressed as tonnes of CO₂-e per household. This is calculated as tonnes of emissions directly attributed to households divided by the number of households in private occupied dwellings (mean quarter ended)
• Emissions intensity of the economy - expressed as tonnes of CO₂-e per $million of GDP in 2009/10 prices (constant prices)

The emissions intensity indicator is consistent with the UNECE climate change indicator framework for measuring the emissions intensity of production activities. Details on the formulation of this indicator can be found here. Households are explicitly excluded from this comparison.

Sources and methods

Approach to compiling quarterly emissions estimates

Producing comprehensive and timely estimates of quarterly emissions can be challenging, given emissions data is compiled annually and often without recording quarterly movements. For this reason, indicators and statistical methods are used to best represent quarterly estimates of emissions in a similar process to that used to compile quarterly GDP, rather than through new data collection (for example, a survey). This section describes these methods and indicator selection in detail.

Most of the general principles outlined in Quarterly national accounts manual: Concepts, data sources, and compilation (Bloem, Dippelsman, & Maehle, 2001) apply equally to economic accounts and emissions estimates, and underpin the methodology discussed in this section. The following concepts and principles from the above manual have been applied in estimating quarterly greenhouse gas emissions (QGHG):

• quarterly accounts should be made consistent with their annual equivalents, partly for the convenience of users and partly, and more fundamentally, because the benchmarking process incorporates the information content of the annual data into the quarterly estimates
• quarterly accounts should be built on a foundation of timely and accurate quarterly source data that directly cover a high proportion of the totals
• quarterly accounts data should be presented as a consistent time series
• seasonally adjusted data, trend data, and unadjusted data all provide useful perspectives, but the unadjusted data should be the foundation
• revisions are needed to allow timely release of data and to allow incorporation of new data – possible inconvenience of revisions can best be dealt with by openness about the process.

Annual estimates of greenhouse gas emissions (industry and households) are compiled using an inventory-first approach (Eurostat, 2015). This involves converting and aligning estimates of emissions from the processes recorded in the inventory into estimates by industry and households (the relationship between processes and production is discussed in Stats NZ (2020b). These annual estimates are used as a base, with indicators and other techniques used to create quarterly estimates.

Compiling QGHG estimates by accounting for processes by industry recognises that an industry’s emissions profile can consist of multiple processes. This is important because not all processes will always change uniformly with economic activity, as practices and technologies change. This provides an emissions estimate that is independent of GDP, which allows for meaningful comparisons.

Estimation of quarterly emissions

There are three major transformation processes when converting annual industry and household emissions estimates into quarterly seasonally adjusted SEEA-based emissions estimates:

1. benchmarking indicators to annual estimates, interpolating, and extrapolating
2. converting process-based estimates into industry estimates (allocating to industry)
3. seasonally adjusting the industry estimates.

Benchmarking, interpolation, and extrapolation

The benchmarking, interpolation, and extrapolation processes involves two important steps:

1. choice of quarterly indicators
2. benchmarking the quarterly indicator series to annual emissions estimates.

The choice of indicator will depend on the nature of the emission process and the availability of data, whereas the benchmarking process involves statistical methods as developed for quarterly GDP. The different factors and types of indicators are discussed in the sections below, as is the process of benchmarking.

Types of indicators

Indicators generally represent a simplification of a more complex phenomenon, by focusing on a dominant factor. For the purposes of explaining the nature of the indicators used for quarterly emissions, it is useful to differentiate between direct and indirect indicators (Eurostat, 2014). A direct indicator attempts to measure the actual phenomenon, whereas an indirect indicator uses a related phenomenon as a proxy.

The purpose of indicators is to provide a representation of movements in emissions over a quarter. As such, indicators do not need to be expressed in emissions units. The use of benchmarking (discussed in Quarterly breakdown of annual benchmarks) implicitly converts movements in an indicator series into emissions units. The benchmarks capture the emissions factor (emissions per unit of activity) and level of activity in a given year, and the indicators provide the information on movements in the activity data within the year. Changes in emissions factors are therefore captured when the benchmarks are updated.

The choice of indicator type has implications for the quality of the quarterly series (see Substances, processes, and production). In selecting the type of indicator, the conceptual accuracy of the indicator will need to take into consideration the nature of the emitting process (see Stocks and flows). Conceptual accuracy refers to the conceptual alignment between the indicator series and the emissions source.

Substances, processes, and production

To help understand emissions, it is helpful to know the distinction between substances used, the processes that utilise the substances, and the production by economic units who control the emitting processes, to meet economic needs. In this context, substances can refer to the material (for example, petrol) that is utilised in a process. A process refers to the use/transformation of the substance (for example, an internal combustion engine), and production may refer to economic activity, such as agricultural production, where multiple processes (biological, internal combustion, waste) are combined to produce an output (for example, milk). Understanding this distinction along with the nature of the emitting process (stock or flow based) can help inform how well an indicator will capture this phenomenon.

Stocks and flows

Emissions processes generally result from the transformation of energy and matter. While all emissions are recorded as flows, the underlying process by which they are generated will have both stock and flow attributes. Identifying whether the emission is primarily driven by the stock or flow dimension ultimately informs the choice and adequacy of the quarterly indicator selected. Some emissions result from material and energy throughput (for example, energy use), whereas others can result from gradual breakdown of a stock of material over time (for example, waste) or depend primarily on the amount of stock used in production (for example, livestock).

Direct indicators have higher conceptual accuracy, as the indicator can capture the dominant attribute in the emission process and therefore reflect actual emissions. Indirect indicators are likely to have a weaker conceptual alignment with the emissions process and therefore, under certain conditions, may deviate from actual emissions, despite having a strong statistical relationship. The level of alignment, and any weaknesses, are important for understanding the performance of the indicator.

Direct indicators

Direct indicators are indicators that relate directly to the substance or process that is central to the emission. The indicator may be the use of a substance, such as coal, which has a known conversion rate into greenhouse gases. An indicator based on the process may capture a dimension of a process which transforms energy and matter, creating emissions, such as the number of animals by type and class, which on average will produce a certain amount of methane as a function of their metabolism.

Indicators based on either substances or processes are consistent with the principles of material flow accounting (European Commission and Eurostat, 2001), and are therefore considered to be more accurate and more likely to capture turning points.

Indirect indicators

Indirect indicators (often termed proxy indicators) are used where data for direct measures are not available. Generally, indirect indicators will have a strong positive correlation with the phenomena they are attempting to capture but can also have a conceptual relationship based on an understanding of the nature of the activity or production processes. Indirect indicators can include the use of related substances, measures of economic activity (adjusted for inflation), population numbers, and upstream or downstream activity. The use of indirect indicators assumes that the relationship between the indicator and the emissions remains stable, and therefore is at risk of missing turning points when the relationship breaks down. The relationship may break down due to reasons such as changes in technology, changes of inputs to production, or changes in behaviour.

For this reason, monitoring industry developments will be extremely important where proxies are used.

Estimation in the absence of indicators

To ensure complete coverage of industries and emissions sources, quarterly estimates are required when no indicator is available. For emissions, quarterly estimates can be made in the absence of indicators using interpolation for benchmark years or forecasting for later quarters.

This approach may be used in the following cases:

• there may be no known quarterly indicator that adequately represents the process conceptually
• a quarterly indicator may be available but may not relate closely to annual benchmarks (see discussion on benchmark-to-indicator ratio in Benchmarking time-series estimates section)
• the emissions source does not exhibit a seasonal pattern, in which case an interpolated series may approximate an actual and seasonally adjusted series
• the benchmark series exhibits a stable trend and is not prone to economic shocks
• the emissions process accounts for an insignificant proportion of an industry’s total emissions (although care is required to ensure interpolation of multiple emissions sources does not account for a significant share of emissions).

Estimation without indicators is used to provide complete coverage, but the contribution of this technique to the compilation process is expected to decline over time as further indicators are sought. However, this technique will continue to be used if it satisfies any of the exceptions listed above.

Benchmarking time-series estimates

The application of indicators, or estimation without an indicator, depends on whether the quarter being estimated is covered by the latest annual estimate (benchmark) or after it.

Quarterly breakdown of annual benchmarks

Where annual benchmarks exist, the primary purpose of the quarterly estimates is to breakdown annual estimates into quarters, to show the seasonal patterns and irregularities which exist. The annual benchmarks, which retain both an industry and process dimension, are sourced from data underlying the SEEA greenhouse gas emissions (industry and households) series.

For emissions estimated either with or without an indicator, the benchmarking process ensures that the sum of the quarters equals the annual estimates while retaining, to the greatest extent possible, the quarter-on-quarter movement suggested by the indicator.

When an indicator is available, the proportional Denton method (a technique designed for macroeconomic statistics) is used. In this process, an estimate of the benchmark-to-indicator ratio is obtained, which provides an indication of the quality of the indicator. This ratio should ideally be stable over time as this would show a strong correlation between the annual benchmark and indicator series. The proportional Denton method is also preferred by Eurostat (2023).

Pearson correlation coefficients between the annual movements in the benchmark series and quarterly indicator series are assessed to determine the accuracy of the indicator. This helps us identify how well the movements in the indicator are likely to predict the emissions movements on average. Pearson correlation coefficients lie between values -1 and 1, with 1 representing a perfect positive correlation. While it is ideal that the indicator and benchmark have a strong positive correlation, the emissions level and level of aggregation is also considered. If a series accounts for a small component of an industry then an indicator with a lower correlation value is not going to have much impact on published numbers and still may be used to ensure completeness (which helps to meet relevance and coherency principles).

The Pearson correlation tests often show that the indicator performs well in relation to the main gas emitted and carbon dioxide equivalents but may have weak correlation to secondary gases. This test is used to prioritise developments as if the correlation is low then the indicator is not robust and an alternative indicator may be required.

While this test gives insight into the benchmark-indicator relationship over the long-term there is an implicit assumption that the series are cointegrated, ie share a long-term trend. The test may not therefore detect ‘drift’ or structural breaks fully. There is also the possibility of spurious correlation between the indicator and benchmark – in this instance consideration needs to be made of the indicator typology (eg direct vs indirect indicators). In addition, the test may lead to acceptance of incorrect seasonal signals as the annual movements may align but the quarterly signals may not capture all relevant phenomena. Milk production, which may be considered to reflect dairy cattle emissions, is an example of this as emissions from ‘dry periods’ will not be represented.

When an indicator is not available, quarterly estimates are formed by interpolating between benchmarks. For series exhibiting a relatively constant trend, the interpolation process will result in a primarily linear quarterly series. If the series does not show a constant trend, or fluctuates, the interpolation process results in smoothed but non-linear quarterly movements reflecting the changing nature of the benchmark data over time.

Extending beyond the annual benchmarks

Estimates for quarters beyond the last available annual benchmark are obtained by using the movement on a quarter-on-quarter basis in the indicator, to extend the time series from the last benchmarked quarter (a process known as extrapolation). Extending the annual series into more recent quarter’s results in improved timeliness of data. However, estimates for these quarters require additional analysis given the potential for loss of accuracy.

Provided the benchmark-to-indicator ratio is stable over time and has a strong conceptual relationship with the emissions process (capturing seasonal patterns, irregularities, and turning points), then the indicator will provide a robust estimate for the quarters beyond the last annual benchmarked year. However, as the benchmark-to-indicator ratio is based on historical data points, there is no guarantee that the indicator will perform well during times of significant change, such as during COVID-19, or following rapid technological change. The ability to integrate and compare emissions with economic data, under the SEEA, is advantageous in this case as a means of assessing whether the estimated quarters are to be expected. The incorporation of new benchmark values as they become available, may lead to revisions to previous quarters estimates across the time series.

Where quarterly estimates are obtained by interpolation without an indicator, estimates for quarters after the last available benchmark are forecasted using exponential smoothing models. Exponential smoothing functions are used to smooth time series using exponentially decreasing weights over time, as opposed to moving averages which assigns weights equally. Estimates obtained from this approach weigh newer observations higher than older ones and account for the variability in the trend. Accordingly, the forward estimates are only driven by the existing time series pattern of the series and no behavioural (eg relationship with economic activity) assumptions are made. See LaViola Jr (2003) for further detail.

As with indicators, the use of forecasting will lead to revisions when future benchmarks are incorporated. The extent of the revision will depend on the deviation of the next benchmark value and the historical trend. For this reason, exponential smoothing functions are not robust to sudden economic shocks that may lead to sudden changes in emissions.

Converting process-based estimates into industry estimates

The Greenhouse Gas Inventory classifies emissions by process, and groups them by common reporting format (crf) class codes for United Nations Framework Convention on Climate Change (UNFCCC) reporting. For the annual benchmarks, industries are defined according to ANZSIC06 and compiled at the New Zealand Standard Industrial Output Classification (NZSIOC) level.

Emissions are allocated to industry in two ways:

• one-to-one matches when the crf class aligns solely to an industry
• one-to-many matches when the crf class is attributable to multiple industries and additional data or methods are used to split across industries.

Most crf classes can be allocated directly to an industry (that is, a one-to-one match) due to the level of detail in the GHG Inventory. To allocate a crf class to an industry, an assessment is made on which producers (firms) are engaged in the process-based emissions and where those firms were allocated to in the industry classification system. Additional information is used if crf classes need to be allocated to a finer industry level. Where a one-to-many approach is used, detailed economic information is often used. More information on the estimation of the annual greenhouse gas emissions (industry and households) approach can be found in Environmental-economic accounts: Sources and methods (third edition) (Stats NZ, 2020b).

Due to costs and feasibility the same level of detail cannot be collected on a quarterly basis. However, these same two approaches are utilised, just at a more aggregated level. The choice of which one is used is dependent on the level of detail of the available quarterly indicators. As with the annual series, estimates are more robust where allocations can be made on a one-to-one basis, as one-to-many matches require additional information to allocate across industries.

Quality of industry emissions estimates

When these different allocation approaches are considered alongside the different indicator types, it is possible to construct a classification of approaches that give some insight into the quality of the resulting industry estimates. Table 1 summarises the approaches and the quality perspective relating to each along with examples. The extent to which these are used varies across sources and industries, as discussed in application of methods.

Table 1. Summary of techniques used to estimate quarterly industry emissions

Technique	Quality perspective	Examples of where the technique was applied
Direct calculation	Most robust approach	Coal production, international aviation
Direct indicator, applied directly to an industry	Most robust indicator approach conceptually and empirically	Jet kerosene emissions applied to the air transport industry
Direct indicator, apportioned across industries and constrained to control total	Robust at process level, possible uncertainty introduced in industry allocation	Coal use apportioned across mining and manufacturing
Indirect indicator, applied directly to an industry	Attention to conceptual and statistical performance required	Steel production, allocated to metal product manufacturing
Indirect indicator, apportioned across industries and constrained to control total	Performance of control total is priority for quality assessment, attention to statistical performance required	GDP for road transport by industry, constrained to a control total for road transport
Benchmark extension, estimate remaining quarters without an indicator	Least common approach, useful to minimise uncertainty when only annual information available	Manufacturing of solid fuels
Estimation without an indicator, applied directly to industry	Useful for ensuring completeness and valid in several situations	F-gases used in refrigeration and air conditioning; landfill emissions
Source: Stats NZ

A direct indicator that is applied directly to industry is the most robust indicator type, as it directly relates the emissions process to the economic unit. This approach enables any changes in emissions, due to shocks such as COVID-19, to be traced back to the economic unit, and provides more robust estimates at aggregated and disaggregated levels.

When a direct indicator is apportioned across industries, the quality of the industry estimates is dependent on the proportions used to allocate across industries. The total movement across industries, however, is still considered robust.

To allocate across multiple industries, we derived quarterly proportions based on interpolated annual series at the component industry level — which gives variable proportions across quarters in a year but assumes similar seasonal patterns across industries.

An indirect indicator applied directly to an industry can provide robust quarterly estimates if the indicator is considered to have good conceptual and empirical properties. The ability to relate the indirect indicator to the industry allows for confrontation against other statistics for that industry (including other emissions sources), which provides information on the quality of the estimates.

In the case of road transport at the industry level, indirect indicators were applied but needed to be benchmarked across industries by a control total based on a direct indicator (for example, fuel use). We applied the indirect indicator to the industry by source benchmark data using the proportional Denton method, and then formed quarterly proportions for allocating across relevant industries. This approach captures different seasonal patterns but assumes a constant relationship between the indicator variable and process. The process of balancing the industry road transport data against a direct indicator provided a constraint for the total industry movements and therefore relaxed the assumption of a constant relationship between the indicator and the process.

The techniques for both direct and indirect indicators that are allocated across industries were applied at the lowest level of industry detail available but aggregated for publication to a level that minimised the distortions and uncertainties arising from the allocation process.

Seasonal adjustment

When measuring emissions on a sub-annual timeframe, the presence of seasonal patterns needs to be assessed. For emissions, a seasonal pattern can result from real-world phenomena such as breeding seasons, use of fuels for heating during winter, and increased use of private vehicles for tourism purposes during summer.

There are two motivations for producing seasonally adjusted estimates of GHG emissions. Firstly, by removing any regular seasonal pattern it is possible to make meaningful quarter-on-quarter comparisons. For example, if emissions increased by 10 percent in a quarter but 6 percent of this change was expected because of a regular seasonal pattern (for example, fuel for heating) then it is the 4 percent change that is of interest to show changes in technology, trend, or other irregular contributors.

Secondly, one of the main purposes for producing quarterly GHG emissions on a SEEA basis is to show how emissions are changing as GDP and other quarterly economic indicators change. As economic statistics are seasonally adjusted, emissions statistics also need to be seasonally adjusted so that comparisons can be made between the two.

Identifying and accounting for seasonality requires identifying the trend, seasonal, and irregular components of the time series.

• The trend reveals the smooth, relatively slow-changing features in a time series. They are usually estimated by applying repeated moving averages.
• The seasonal component shows the seasonal patterns found in many sub-annual series. It is reasonably stable in terms of annual timing, direction, and magnitude and can be caused by natural factors, administrative measures, and fixed social traditions or behaviours.
• The irregular component is the part of the observed value that is not included in the trend cycle or the seasonal effects. Its values are unpredictable regarding timing, impact, and duration. Irregular movements arise from a combination of factors such as sampling error, non-sampling error, unseasonable weather, natural disasters, and strikes. Random fluctuations are the main cause. The irregular component is estimated as a residual.

The seasonal adjustment method used by Stats NZ for quarterly GHG emissions is the X-13 variant of the seasonal adjustment programme developed by the United States Census Bureau. This technique uses all observations in the time series to statistically determine the seasonal pattern.

For more information regarding seasonal adjustment see Eurostat (2018).

Points to consider when applying and interpreting seasonal adjustments:

• The quality of the seasonal pattern is generally considered to improve with the length of the times series. Adding new data points (quarters) can cause the statistically derived seasonal pattern to change and lead to revisions.
• Stats NZ tends to use five years as a rough guide for the adequacy of a times series for deriving a robust seasonal pattern. The quarterly GHG emissions estimates have an adequate time series length at over a decade.
• As the irregular component of the times series is calculated residually, it will be impacted by the quality and changes to the seasonal component.

Fuel use for heating is an example of how irregular components are affected by the seasonal estimate. While most of the estimate of emissions from fuel use for heating will be seasonal, there may also be an irregular component due to temperature changes occurring outside of the regular seasonal pattern. As climate change increasingly impacts weather and climate, the seasonal component may become more variable over time. This means the seasonal component observed in the earlier part of the time series may be unrepresentative for more recent years. It is unclear what the impact will be, if any, and will be assessed over time.

The results of the seasonal adjustment can be found in Seasonal adjustment – industry and household.

Sources of revisions

Quarterly emissions estimates are based on information available at the time of compilation. Revisions to the time series will be inevitable, due to the nature of benchmarked quarterly series.

Revisions to the ‘actual’ emissions time series may result from:

• revisions to the annual benchmarks
• revisions to indicator series
• changes to methodology
• revisions to the data used or changes in the process for allocating emissions to industries.

Any revision to the actual emissions series will also result in revisions to the seasonally adjusted series. However, revisions caused by new data points in the time series can also lead to adjustments to the statistically derived seasonal pattern.

Application of methods

Quarterly emissions estimates are compiled from two perspectives: SEEA industry and inventory sector. This dual perspective enables a wider range of data sources and modelling options to be utilised.

Stats NZ (2020b) shows how sector-based emissions from the Greenhouse Gas Inventory are allocated to industry (where industry is defined by ANZSIC). Emissions in the inventory are classified by source category and grouped by crf codes for UNFCCC reporting. The crf processes are retained in the underlying calculations of the annual industry and household estimates, which provide further methodological options for quarterly measurement. A summary of how sectors are spread across the published level of the quarterly emissions series is presented in table 2.

Table 2. How inventory sectors are split across published-level industries

Published industry	Inventory sector split
Agriculture, forestry, and fishing	The majority of emissions for the agriculture, forestry, and fishing industry are from the agriculture process inventory sector. The energy and waste sectors contributions are small in terms of overall proportions of total industry but significant in terms of absolute levels
Mining	Almost all emissions from mining are from the energy inventory sector, largely from stationary sources.
Manufacturing	Emission in the manufacturing industry are primarily from the energy inventory sector with the IPPU sector (primarily industrial processes) accounting for a significant contribution of emissions. Most of the energy emissions are from stationary sources.
Electricity, gas, water, and waste services	Contribution of different sectors can vary over time given the volatility in electricity generation emissions. Emissions in the electricity, gas, water, and waste services industry primarily come from the energy inventory sector with a significant contribution from the waste sector, and a modest contribution from the IPPU sector (main disposal of industrial products)
Construction	Almost all emissions from construction are from the energy inventory sector, comprising both stationary energy and road transport.
Transport, postal, and warehousing	Transport, postal, and warehousing emissions are virtually all from the energy inventory sector. Residency adjustments account for a large proportion of emissions from this industry.
Services excluding transport, postal, and warehousing	Services excluding transport, postal, and warehousing industry emissions are mainly from energy sources, with the IPPU sector contributing moderately, primarily through stationary air conditioning emissions.
Households	Primarily energy emissions (road transport and direct fuel use for heating), with minor contributions from waste.
Source: Stats NZ

Annual emissions benchmarks are sourced from Stats NZ’s GHG emissions by industry and household series. The underlying data contains information for 116 industries and households by Greenhouse Gas Inventory source category. Quarterly emissions data is compiled at a higher level, depending on the availability of indicators and their relevance to specific industries.

The following sections outline the sources and methods used by each main emissions source: energy, industrial processes and product use, agriculture, and waste.

Energy

Energy emissions span all industries in the economy and households. In broad terms, energy emissions cover stationary combustion processes (for example, industrial heating processes, electricity generation, and mining) and mobile combustion processes (for example, transportation by road, air, rail, and water).

For consistency with the SEEA framework, emissions from energy used by non-residents on the domestic territory are excluded and by residents overseas included. This affects the compilation of emissions from mobile energy processes, including international shipping and aviation, and tourism. This leads to noticeable differences to quarterly energy emissions produced by MBIE.

Quarterly energy and emissions statistics provide a rich source of information for compiling quarterly emissions on an SEEA basis. MBIE’s New Zealand Energy Quarterly provides oil, gas, and coal tables with information on energy use by high-level industry groups. Other sources include transport data, company data, electronic card transaction data, wood products production statistics, and quarterly GDP. The ability to incorporate multiple data sources is advantageous, given that energy sector emissions are the most prone to sudden changes in economic activity and the effects of such shocks may not be felt uniformly across industries. Estimation in the absence of an indicator is undertaken for just under 3 percent of energy-related emissions.

Stationary energy emissions

Stationary sources include non-moving industrial sources such as power plants, refineries, manufacturing plants, and other sources such as heating.

MBIE’s New Zealand energy sector greenhouse gas emissions statistics provide data on electricity generation and geothermal energy, which were applied directly to the electricity, gas, water, and waste services industry.

Quarterly coal, gas, and oil use estimates were used, where relevant, as indicators for stationary energy emissions for the following industries:

• agriculture, forestry, and fishing
• mining
• manufacturing
• construction
• service industries excluding transport, postal, and warehousing
• households.

As coal, gas, and oil use is a direct indicator of activity data (that is, directly represents the underlying fuel used), and there is a close alignment between energy sector groupings and broad industry groups defined by ANZSIC06, these estimates have high conceptual alignment and were closely correlated as shown in the benchmark-to-indicator ratios (see Benchmarking, interpolation, and extrapolation). Industrial energy use covers the mining, manufacturing, and construction industries. Estimates at lower industry levels were made using real quarterly GDP and constrained against the control total.

For oil use, indicators were applied by fuel type (petrol, diesel, and other) to relevant industries. Emissions from oil refining were estimated using movements in indigenous oil production.

Coal production emissions, allocated to mining, were derived using quarterly coal production statistics by coal type and annual proportions and emissions factors from the Greenhouse Gas Inventory. This direct estimation approach was preferred, as it aligns with the estimates in the Greenhouse Gas Inventory and accounts for the changes in coal types that is not possible using an aggregated indicator of coal production. Data from the energy balance tables was used to extend the annual benchmark for manufacture of solid fuel and other energy industries before the series was estimated in the absence of an indicator.

Further direct indicators from the MBIE quarterly gas tables were used for measuring quarterly emissions from gas production and flaring, with data provided directly by MBIE for estimating gas venting.

Further information on the sources and methods used for quarterly oil, gas, coal, and energy emissions statistics can be found in MBIE (2021).

Mobile emissions

Mobile emissions include road, rail, shipping, and air transportation. Mobile energy emissions are subject to adjustments for economic residency.

Road transport emissions

Road transport emissions are allocated to both households and industries and cover greenhouse gas emissions that occur because of car, motorcycle, heavy trucks and buses, and light-duty truck use. Carbon dioxide is the most prominent gas associated with road transport emissions.

As with the estimation of road transport emissions by industry and households at the annual level, quarterly road transport emissions are the most challenging to allocate to industry and households. Quarterly road transport emissions are allocated using multiple data sources that enable robust totals to be maintained while allowing for variation across industry and households, and vehicle type.

The process developed to calculate quarterly emissions from road transport is as follows:

• a total quarterly road transport estimate was derived by using quarterly fuel use data for land transport, by fuel type
• household and non-resident road transport was estimated using fuel spend card transaction data, adjusted for inflation, as an indicator
• industry road transport was estimated using economic activity data (GDP) as an indicator
• household, non-resident, and industry road transport emissions were then constrained to the total.

Although GDP was used in the road transport estimation process, this is mainly used to form relative industry movements rather than absolute movements. Its use assumes that the relative use of vehicles by industry, over quarters, is correlated with economic activity. Comparisons of emissions to GDP by industry are valid, as the movements in emissions and GDP are ultimately independent because of the final balancing process. In addition, the comparison is independent because of differences within industries in terms of the contribution to road transport emissions and GDP.

Rail transport emissions

Quarterly rail emissions were estimated using freight-tonne kilometres travelled (produced by Te Manatū Waka — Ministry of Transport as an indicator) and directly allocated to rail benchmarks in the transport, postal, and warehousing industry.

Shipping transport emissions

Quarterly shipping emissions were estimated using oil statistics data from MBIE and applied directly to domestic navigation benchmarks in the transport, postal, and warehousing industry.

Air transport emissions

Oil statistics produced by MBIE for jet kerosene were applied directly to the air transport industry. Quarterly aviation gas emissions were allocated across relevant industries based on the proportion derived from interpolated annual benchmarks. Additional information (provided directly) was used to adjust for emissions from resident economic operators.

Industrial processes and product use

Emissions from industrial processes and product use (IPPU) covers greenhouse gas emissions that result from industrial processes, from the use of greenhouse gases in products, and from non-energy uses of fossil fuels. Examples of processes include blast furnaces in the iron and steel industry, and manufacture of ammonia and other chemical products where fossil fuels are used as a feedstock. Product use refers to greenhouse gases used within products such as refrigerators, foams, or aerosol cans. In most cases there will be a significant lag between the manufacture of the product and the release of the gas. The lag can be as short as weeks, for example for aerosol cans, or as much as decades in the case of rigid foams (IPCC, 2006).

The IPPU sector accounts for a small part of New Zealand’s gross emissions, contributing between 5 to 6 percent between 2007–18 (Stats NZ, 2020). New Zealand’s Greenhouse Gas Inventory reports non-energy emissions from the:

• calcination of limestone in cement production
• calcination of limestone in burnt and slaked lime production
• production of ammonia, which is further processed into urea production of methanol
• production of hydrogen in oil refining and for making hydrogen peroxide
• production of steel, from iron sand and from scrap steel
• oxidation of anodes in aluminium smelting
• use of soda ash and limestone in glass making.

Industrial processes are allocated to the relevant manufacturing industry. Industrial product use occurs across all industries and households. Emissions from industrial product use includes hydrofluorocarbons (HFCs) and perfluorocarbons (PFCs) used in refrigeration and air conditioning equipment (MfE, 2020), and a small contribution from sulphur hexafluoride (SF6). In terms of gas type, carbon dioxide is the dominant gas type emitted from IPPU sources, primarily from the manufacturing of metals and cement. Industrial processes have typically been the main contributor to IPPU emissions , with its contribution becoming smaller over time. Note that approximate values are used here because the IPPU section of the inventory has not had two distinct chapters separating processes and product use since the 1996 guideline was in use.

Allocation of industrial process emissions

Industrial process emissions are allocated to industry through a combination of indirect indicators applied directly to industry and estimation in the absence of an indicator. As there are a relatively small number of industrial processes in the inventory and the main sources of these are often produced by single economic units, it is possible to apply indicators at a low level of detail.

Presently, indicators used for quarterly industrial process emissions estimation are indirect in nature. Direct indicators that attempt to measure the actual movements in greenhouse gas emissions have not been identified. The indirect indicators used for quarterly emissions estimation include those that relate to inputs and outputs.

Input indicators

Indicators that relate to inputs for a specific plant indirectly reflect emissions. In the IPPU sector, many large plants in the manufacturing industries run 24/7 operations and are dominated by a small number of units. In several cases, data that is available for the direct supply of a key feedstock or input to the production process, for example electricity or natural gas supply, is used to estimate emissions. Electricity use at the Tiwai aluminium smelter is an input indicator selected to estimate emissions from aluminium production.

Output indicators

Output-based indicators, such as production volume, can be directly related to activity at the plant and therefore indirectly its emissions. In these cases, the implicit assumption is that there is a constant emissions intensity over the quarters in a year. However, the regular introduction of new benchmarks will mean that these emissions intensities are eventually updated.

In identifying indicators for industrial processes, financial data from Stats NZ’s quarterly collections was explored but not used. Financial data should be avoided, as when expressed in current prices it may be influenced by local or world prices or other factors rather than production. If financial data is expressed in constant prices, any insights into decoupling quarter-on-quarter will be weakened as there will be less independence from measures of economic activity.

Output indicators selected for estimating industrial process emissions include steel volumes, refinery output, and cement production.

Neither input or output indicators could be identified for ammonia and carbide production. These emissions, which are minor sources of industrial processes, were estimated in the absence of an indicator.

Allocation of emissions from product use

Emissions from industrial product use are primarily air conditioning and refrigeration product use and are generally allocated across all industries and households. Product use emissions arise during the manufacturing process from stocks, or during its disposal. Indicators can therefore be difficult to identify when estimating quarterly movements. Hence, product use emissions were estimated without indicators. See table 3 for an indicator breakdown for the IPPU sector.

Table 3. Indicator breakdown for the industrial processes and product use sector

Common reporting format (crf) class codes	Common reporting format (crf) descriptors	Indicator
2.A.1 - 2.A.4	Cement, lime, glass, and other process uses of carbonates	Primarily cement production volumes
2.B.1 - 2.B.5	Ammonia production, carbide production	No indicator
2.B.8 - 2.B.10	Petrochemical and carbon black production, hyrdogen production	Oil and gas production data
2.C.1a	Steel	Primarily steel production
2.C.7	Aluminium	Electricity supply to Tiwai
2.D.1	Lubricant use	Road transport
2.D.2-2.D.3	Paraffin wax and urea catalyst in road transport	No indicator
2.F.1.a-2.F.4.b, except 2.F.1.e	Commercial, domestic, industrial and transport refrigeration, mobile and stationary air conditioning, closed cells, fire protection, and metered dose inhalers	No indicator
2.F.1.e	Mobile air conditioning	Road transport (stocks), no indicator (disposal)
2.G.1 - 2.G.3a	Electrical equipment, medical applications, and medical and other product use	No indicator
Source: Stats NZ

Agriculture

The agriculture sector is the largest emitting inventory sector and is predominantly made up of enteric methane and emissions from agricultural soils.

Gross emissions from the agriculture sector, as reported by the Greenhouse Gas Inventory, are allocated almost entirely to the agriculture, forestry, and fishing industry. Emissions from the agriculture sector account for most of those from this industry (91 percent), with emissions from energy, and industrial product use and waste sources accounting for a further 9 percent of emissions.

At this stage of development, methods have been developed to estimate quarterly emission numbers for the largest sources of agricultural emissions, namely methane from livestock enteric fermentation and manure, nitrous oxide from manure and nitrogen fertiliser, and carbon dioxide from soil applications (see table 4 for a summary of sources/indicators used). At the total level, all annual agricultural emissions are benchmarked to align with actual (to latest available benchmark year) and projected (post-latest available benchmark year) annual agricultural emissions inventory totals.

Table 4. Data sources/indicators for quarterly agricultural greenhouse gas emissions

Emission	Indicator for estimation	Frequency	Source	Data availability
Methane (CH4) - enteric fermentation and manure management	Methane research series (for selected livestock categories)	monthly	MPI	2009 to 2016 (updated intermittently)
	Agricultural emissions inventory projections for MfE	annual	MPI	1990 to 2050
	New Zealand's Greenhouse Gas Inventory	annual	MfE	1990 to latest benchmark year
	Total NZ Livestock slaughtering numbers, kill by animal type	quarterly	MPI (available via Stats NZ Infoshare)	1981 Q4 to present
	QGDP livestock models	quarterly	Stats NZ	1985 to present
Nitrous oxide (N2O) - manure management and soils	Agricultural emissions annual inventory projections for MfE	annual	MPI	1990 to 2050
	Nitrous oxide research series (for selected livestock categories)	monthly	MPI	2019 (Updated intermittently)
	New Zealand's Greenhouse Gas Inventory	annual	MfE	1990 to latest benchmark year
Carbon dioxide (CO2) - soil applications	Agricultural emissions inventory projections for MfE	annual	MPI	1990 to 2050
	New Zealand's Greenhouse Gas Inventory	annual	MfE	1990 to latest benchmark year
	Research and expert advice	ongoing	various
Note: Not all indicator series/models used are publicly available. As new research becomes available, indicators will be updated and/or added. MPI – Ministry of Primary Industries. MfE – Ministry for the Environment.
Source: Stats NZ

Estimating a quarterly time series for agriculture

Enteric methane and methane resulting from manure management accounted for 78 percent of agricultural emissions in 2018 and are related to the total amount of dry matter eaten by livestock. These two types of methane emissions are estimated together using a model that extends the logic presented in MPI’s detailed methodology document (MPI, 2020). The method is based on livestock numbers by livestock type and class, and their implied emissions factors as found in New Zealand’s Greenhouse Gas Inventory tables (MfE, 2020). Given that the emissions factors are largely stable, animal numbers are the main driver and therefore have been used as an indicator of methane emissions.

To break annual methane estimates into quarters, a monthly methane series for key livestock types was used as a guide. A quarterly series was directly calculated for each methane source (enteric fermentation and manure management) and benchmarked to annual Greenhouse Gas Inventory estimates. Quarterly estimates were then proportionately allocated across industries based on the latest available industry greenhouse gas account. All agricultural sector methane emissions were directly estimated in this way.

Nitrous oxide emissions resulting from manure management and agricultural soils depends on the total amount of nitrogen going through a farm via feed and fertiliser. Total annual nitrous oxide emissions from manure management and agricultural soils are estimated within the Greenhouse Gas Inventory and projected out by MPI. A monthly research series provided monthly proportional splits for selected livestock categories. Using these monthly proportions and annual projections, we estimated quarterly nitrous oxide emissions for key livestock types. Residual nitrous oxide emissions (for non-key livestock types) were interpolated within the system and added to agricultural emissions. Total quarterly nitrous oxide emissions from agriculture were benchmarked to annual Greenhouse Gas Inventory estimates and projections. The quarterly estimates were then proportionately allocated across industries based on the latest available annual industry greenhouse gas account.

Carbon dioxide from soil applications (urea and lime) quarterly estimates are reliant on several factors. The timing of urea and lime applications on New Zealand farms depends on farm type, climate, soil condition, livestock birthing seasons, and location. No single quarterly indicator could be found for these emissions sources. Using research and expert opinion, quarterly series for carbon dioxide emissions from urea and from lime were separately calculated and benchmarked to annual Greenhouse Gas Inventory estimates. Because not all farm types apply urea and/or lime, the quarterly emissions estimates were further assigned to particular industries using the latest published annual greenhouse gas account as a guide. Directly estimated carbon dioxide accounted for 3 percent of all agricultural sector emissions.

Waste

Waste is the smallest of the sectors. Methane is the dominant gas emitted from waste sources.

Landfill emissions from the waste sector are unlike other gross emissions sources, as they reflect emissions resulting from past, as well as present, waste disposal. Landfill emissions in the inventory are estimated using first order decay methods, so that emissions each year are dependent on the total amount of accumulated waste, including that disposed in the current period. For example, large and sudden changes in waste to landfill will have a less-than-proportionate impact on emissions now but will have their effect spread over time. Another effect of this is that there may not be any significant seasonal pattern related to waste emissions. Non-landfill waste emissions, however, may be more directly related to current activity.

Given the nature of waste sector emissions, waste sector emissions are primarily interpolated between benchmark years and forecasted for quarters in post-benchmark years. Two indicators are used for wastewater emissions from pulp and paper manufacturing and meat manufacturing. Eurostat (2023) and Statistics Sweden (2016) also noted the difficulty in estimating waste sector emissions and used forecasting techniques for post-benchmark years.

Industry and household compilation

Quarterly estimates were compiled using information on both industry and source category and aggregated to industry for publication. Table 5 summarises the composition of industry estimates in terms of indicator types.

Table 5. Summary of estimation methods for each industry

Industry	Summary of estimation methods
Agriculture, forestry, and fishing	Biogenic methane estimated using direct methods.
	Direct indicator available for majority of emissions from soils and crops.
	Direct indicators used for energy component, no indicator for farm fills.
Mining	Primarily based on direct indicators.
Manufacturing	Direct indicators used for majority of emissions from energy sources.
	Key industrial process emissions estimated using both direct and indirect indicators.
	Estimation without an indicator used for the remaining industrial process and product use emissions, and waste.
Electricity, gas, water, and waste services	Combination of direct indicators (based on emissions) and estimation without indicators (primarily for waste component).
Construction	Direct and indirect indicators used that are apportioned to construction.
Services excluding transport, postal, and warehousing	Combination of direct indicators applied directly from energy sources and estimation. Without indicators for industrial product use.
Transport, postal, and warehousing	Utilises several indicators and accounts for main transport modes independently, and residency adjustments based on direct indicators. Minor use of estimation without indicators.
Households	Road transport (main component of household emissions) is estimated using a direct indicator.
	Direct indicators also used for other energy sources.
	No indicator available for industrial product use or waste emissions.
Note: For all industries, indirect indicators are used for road transport emissions.
Source: Stats NZ

Seasonal adjustment - industry and household

Table 6 shows the industries and aggregations which showed statistically significant seasonal patterns for carbon dioxide equivalents and were therefore seasonally adjusted. For those series that did not show a seasonal pattern, the seasonally adjusted series are equal to the actual series.

Table 6. The presence of seasonality across industry and households

Industry		Presence of seasonality	Peak quarter(s)
Primary industries		Yes	March/December
	Agriculture, forestry, and fishing	Yes	March/December
	Mining	No	–
Goods-producing industries		Yes	December/March
	Manufacturing	Yes	December
	Electricity, gas, water, and waste services	No	–
	Construction	No	–
Services industries		Yes	September
	Services excluding transport, postal, and warehousing	Yes	September
	Transport, postal, and warehousing	Yes	September/March
Total industry		Yes	March/December
Households(1)		Yes
	Transport	Yes	December
	Heating/cooling	Yes	September
	Other	No	–
Total (industry and households)		Yes	March/December
1. Household emissions are seasonally adjusted at the component level and then aggregated.
Symbol: – no peak quarters
Source: Stats NZ

Seasonal adjustment is also undertaken at the gas type level. The presence of seasonal patterns may change over time due to technology changes, changes in industry composition, or revisions.

Emissions from the electricity, gas, water, and waste services industry do not have a significant seasonal pattern. Non-seasonal variability in the emissions from this industry is due to changes in the generation mix for public electricity and heat production. Electricity demand (ie kilowatt hours), however, follows a seasonal pattern as this demand is met by both renewable and non-renewable sources.

Glossary

ANZSIC: Australian and New Zealand Standard Industrial Classification.

benchmarking: refers to the case where there are two sources of data for the same target variable, with different frequencies, and is concerned with correcting inconsistencies between the different estimates, for example, quarterly and annual estimates of value-added from different sources.

common reporting format (crf): a classification of anthropogenic greenhouse gas emissions by sources and removals by sinks. Used in the preparation of standardised emissions data tables for international inventory reporting.

decoupling: common analysis that entails examining the degree of decoupling between natural inputs or residual flows and economic variables. Decoupling occurs when the growth rate of an environmental pressure is less than that of its economic driving force (for example, real GDP) over a given period.

Denton method: method for benchmarking. Its aim is to achieve consistency between time series on the same target variables that are measured at different frequencies (for instance annual data with quarterly data) with a different reliability.

emissions intensity: measures the volume of emissions produced against some other relevant unit. Examples include emissions per person, or emissions per unit of economic activity/GDP.

enteric fermentation: the digestion process of multi-stomach ruminant animals (such as cattle, sheep, goats) that produces enteric methane (CH4) emissions.

extrapolation: statistical technique aimed at inferring the unknown from the known. It attempts to predict future data by relying on historical data, such as estimating the size of a population into the future based on the current population size and past rate of growth.

flow: statistical series presented as flow series/data are accumulated during the reference period, for example, passenger car registrations, where the figure for the reference period is the sum of daily registrations.

gross domestic product (GDP): a monetary measure of the market value of all the final goods and services produced in a country during one year. GDP can be measured using either the production, income, or expenditure method. The production method, which aligns to the production approach to measuring emissions, is based on the concept of value added. This is calculated as the value of output (the value of goods and services produced) less the value of intermediate consumption (the value of goods and services used to produce that output). GDP is measured as the value added from production by industry, where GDP equals the sum of value added for all producers, plus taxes on production and imports.

interpolation: use of a formula to estimate an intermediate data value.

irregular: irregular component of a time series is the residual time series after the trend-cycle and the seasonal components (including calendar effects) have been removed. It corresponds to the high frequency fluctuations of the series.

kilotonnes: 1 kilotonne or metric kiloton (unit of mass) is equal to 1,000 metric tons. A metric tonne is exactly 1,000 kilograms (SI base unit) making a kilotonne equal to 1,000,000 kilograms.

New Zealand’s exclusive economic zone (EEZ): within its exclusive economic zone, which extends 200 nautical miles from the coast, New Zealand has sovereign rights over the management of the resources of the sea, seabed and subsoil.

producing unit: New Zealand resident unit operating within the economy and classified to an industry or household.

production boundary: includes only emissions relating to New Zealand’s economic residents, for example, international tourists’ emissions are excluded; this differs from the ‘territorial’ approach used in the inventory.

seasonal: in time series, that part of the movement which is assigned to the effect of the seasons on the year, for example, seasonal variation in rainfall.

seasonal adjustment: a statistical technique to remove the effects of seasonal calendar influences operating on a series. Seasonal effects usually reflect the influence of the seasons themselves either directly or through production series related to them, or social conventions.

source category: source of emissions. In the inventory this refers to a process or industry and generally covers a grouping of common reporting format (crf) class codes.

statistical noise: random irregularity, or unexplained variability found in data.

stock: stock of an asset; the level of the resource available. Can be measured in both physical and monetary units.

System of Environmental-Economic Accounts (SEEA): developed by the United Nations Statistical Division as a satellite system to the System of National Accounts (SNA), to incorporate environmental concerns (costs, benefits, and assets) in the national accounts. SEEA is the international standard for measuring the links between the environment and the economy.

System of National Accounts (SNA): international accounting framework consisting of a coherent, consistent, and integrated set of macro-economic accounts, balance sheets, and tables. SNA is based on agreed concepts, definitions, classifications, and accounting rules. It provides a framework within which economic data can be compiled and presented for the purposes of economic analysis, and decision and policy making.

trend: steady underlying long-term movement and shorter-term movements in a series.

References

Bloem, AM, Dippelsman, RJ, & Maehle, NO (2001). Quarterly National Accounts Manual Concepts, Data Sources, and Compliation . Retrieved from www.imf.org.

European Commission and Eurostat (2001). Economy-wide material flow accounts and derived indicators: A methodological guide, 2000 ed. Retrieved from http://epp.eurostat.ec.europa.eu/portal/page/portal/environmental_accounts/documents/3.pdf

Eurostat (2014). Towards a harmonised methodology for statistical indicators - Part 1: Indicator typologies and terminologies. Retrieved from https://ec.europa.eu.

Eurostat (2015). Manual for air emissions accounts. Retrieved from https://ec.europa.eu.

Eurostat (2018). Handbook on Seasonal Adjustment. Retrieved from https://ec.europa.eu.

Eurostat (2023). Eurostat's estimates of quarterly greenhouse gas emissions accounts. Retrieved from https://ec.europa.eu/eurostat/documents/1798247/6191529/Methodological-note-on-quarterly-GHG-estimates.pdf/6bd54bde-4dd7-ebac-6326-f08c73eb9187?t=1644394935594.

Intergovernmental Panel on Climate Change (IPCC) (2006). IPCC Guidelines for National Greenhouse Gas Inventories. Retrieved from www.ipcc-nggip.iges.or.jp.

Intergovernmental Panel on Climate Change (IPCC) (2007). Fourth Assessment Report. Retrieved from www.ipcc.ch.

LaViola Jr, J (2003). Double exponential smoothing: An alternative to Kalman Filter-Based Predictive Tracking. Retrieved from http://cs.brown.edu.

Ministry of Business, Innovation and Employment (MBIE) (2021). Energy statistics sources and methods. Version 1.2. Retrieved from www.mbie.govt.nz.

Ministry for the Environment (MfE) (2023). New Zealand’s Greenhouse Gas Inventory (1990–2021). Retrieved from www.mfe.govt.nz.

Ministry for Primary Industries (MPI) (2020). Methodology for calculation of New Zealand’s agricultural greenhouse gas emissions. Version 7. Retrieved from www.mpi.govt.nz.

Stats NZ (2020a). Approaches to measuring New Zealand’s greenhouse gas emissions. Retrieved from www.stats.govt.nz.

Stats NZ (2020b). Environmental-economic accounts: Sources and methods (third edition). Retrieved from www.stats.govt.nz.

Stats NZ (2023). Greenhouse gas emissions (industry and household): Year ended 2021. Retrieved from www.stats.govt.nz.

Statistics Sweden (2016). New method for up-to-date environmental accounts: - quarterly emissions to air. Retrieved from www.scb.se.

Further reading

Department of the Environment and Energy (2019). Quarterly update of Australia’s national Greenhouse Gas Inventory: June 2019. Retrieved from www.industry.gov.au.

Office for National Statistics (2019). UK Greenhouse gas emissions: provisional estimates – Methodology summary. Retrieved from www.assets.publishing.service.gov.uk.

Statistics Netherlands (2010). CO₂ emissions on quarterly basis – final report. Retrieved from https://circabc.europa.eu/ui/explore.

Appendix 1 – Data quality declaration for the quarterly greenhouse gas emissions estimates

Quarterly emissions data are timelier, but there may be a loss of accuracy from using less robust data sources (for example, using indirect indicators as opposed to actual activity data) or utilising models and assumptions. There is also a loss of detail (and potentially relevance) as quarterly estimates can only be robustly estimated for seven industries and households at present.

The following summarises the quality of the quarterly emissions series in relation to the six dimensions of data quality.

Accuracy

While estimating quarterly emissions may result in a trade-off between accuracy and timeliness, Stats NZ prioritises the use of direct indicators and minimises imputations (except where the approach is sound), and aggregates to reduce statistical noise. Accuracy of the indicators are checked for by assessing the correlation between the indicator and benchmark series.

Relevance

The quarterly emissions estimates cover all sources of economic activity and households but are only available for seven industries at present. The series includes emissions from residents operating overseas and excludes non-residents on the territory. This residency approach differs from the territory approach used in the Greenhouse Gas Inventory but enables emissions data to be related to GDP.

Interpretability

Quarterly emissions estimates have the same scope as the annual estimates so can be interpreted in the same way as emissions associated with economic activity by New Zealand residents.

However, when interpreting the seasonally adjusted series it should be noted that regular seasonal patterns have been removed and the series reflects both trend and irregular components.

Coherence and consistency

The quarterly estimates utilise the principles, concepts, and definitions as specified in the SEEA Central Framework. Industries are defined according to ANZSIC. The estimates are calibrated to published annual industry and household greenhouse gas emissions estimates.

Consistent data sources are used across time to enable time series consistency.

Timeliness

Estimates are available with a lag of between four and five months which reflects the availability of input data and time required for processing, quality assurance, and dissemination.

Accessibility

Quarterly emissions data is presented in the same formats and using the same terminology as the annual industry and household estimates. The use of standard industrial classifications and statistical testing for seasonal adjustment also enhances accessibility of the data.

^en-NZ