Australian Agriculture: Horticulture methodology

The Australian Bureau of Statistics (ABS) is modernising the way official agricultural statistics are produced to better support Australian agriculture. These changes were outlined in early 2023 in Modernising ABS Agriculture Statistics which explained why the ABS is modernising its Agriculture Statistics Program and the guiding principles being followed.

A key part of the modernisation has been the development of partnerships with industry and government to identify new data sources and develop new statistical methods. This has reduced respondent burden on farmers, and created greater consistency between existing data sources which helps to build trust in the statistics.

ABS Horticulture statistics are being produced to align with horticulture statistics from the Australian Horticulture Statistics Handbook which is published annually by Hort Innovation, the research and development corporation for the horticulture industry. Australian and state/territory production and local value statistics in this publication are sourced from the handbook (with the exception of wine grapes).

The ABS has developed an experimental method to produce regional production, value, area and business count statistics that complement and are consistent with the national production and value figures produced by Hort Innovation. These statistics use additional crop mapping and industry data and have been produced for avocados and macadamias in this release. The intention is to incrementally expand this experimental method to a broader range of horticultural commodities in future releases.

This publication contains statistics on the production and local value of horticultural crops, and for avocado and macadamia crops, the area of crops grown, and the number of grower businesses (avocados only).

A summary of the key data source used to produce the horticultural statistics contained in this publication are presented below.

Local value and production estimates relate to crops produced (grown for sale).

Local value, also known as farm-gate value, is the value of the commodity at the first point of sale to packhouse or processor (local value excludes transport and marketing costs).

Production estimates refer to the amount of crops grown in Australia in tonnes.

Total crop area includes all land identified as growing a particular crop. Total crop area includes bearing and non-bearing crop area.

Bearing crop area includes all land identified with bearing crop area. Avocado trees 5 years of age and older and macadamia trees 6 years of age and older were considered bearing.

Business/grower count estimates were provided by industry. A business/grower may have one or multiple orchards at one or more locations.

Horticulture production and value estimates presented in this publication relate to the crops produced for sale in the 2022-23 growing season. Some sales may have occurred in the following financial year.

Statistics relate to crop production for commercial purposes and excludes production that is not harvested or used for domestic purposes (e.g. used on farm for animal feed).

Crop area estimates were produced using crop mapping derived from satellite imagery. These provide an estimate of all crop growing areas and may differ from the harvested area.

Bearing area includes all area identified as being of bearing age (avocados 5 years and older, macadamia 6 years and older). Any production area under this age threshold will not be considered bearing. Young trees will produce small amounts of fruit and nut depending on irrigation; however, this production is considerably lower than that of a fully mature tree.

These experimental regional horticulture estimates align with the national production and value statistics published by Hort Innovation. These statistics have been modelled to smaller regions using a combination of regional production, value, yield and bearing area data provided by industry groups. This data is combined with tree crop mapping produced by the AARSC to provide the additional regional detail about the area and locations, where different crops are grown. The intention is to incrementally apply this approach to additional crops as data becomes available in future years.

The modelling process is broken into a summary of the main steps below:

1. Validate regional industry production estimates. Industry region production amounts are validated against the national production data published by Hort Innovation to ensure consistency.
2. Define growing regions. Where required, industry specific growing regions are mapped to enable the use of supplementary industry data as regional benchmarking values together with the tree crop mapping. This is a one-off step for each commodity type.
3. Calculate the area of bearing vs non-bearing crops. Bearing and non-bearing crop areas are applied to the tree crop map at paddock level using a Geographic Information System. Bearing area is based on tree age information provided by industry groups and there is an assumption that bearing age is consistent across all farms.  
4. Allocate regional production to paddock area. The production amounts from each industry region are spatially allocated across paddock areas from the tree crop map based on the bearing area and estimated yields. There is an assumption that yields are consistent across all bearing trees within a region. 
5. Allocate value to paddock area. Value estimates were calculated for paddock areas by multiplying the allocated production (from step 4), with either the national or state price per tonne. Price per tonne information was sourced from the 2022-23 Australian Horticulture Statistics Handbook.
6. Aggregate data. Paddock level estimates for total area, bearing area, production, and value are aggregated to Australian Statistical Geography Standard (ASGS) Statistical Areas Level 2 to produce the experimental estimates. 

The estimates in this publication have been developed in collaboration with industry. Data were sourced from the following organisations: 

• 2022-23 Australian Horticulture Statistics Handbook
• Applied Agricultural Remote Sensing Centre at the University of New England 
• Avocados Australia, Australian Macadamia Society
• Wine Australia.

By partnering with these industry bodies models were developed in a collaborative manner to ensure alignment between the estimates produced and industry data.

The data used to produce sub-state estimates were validated by industry experts and are considered high quality. The ABS verified the reliability of estimates from the new method by comparison with previously published Agricultural Census data. These are experimental estimates and although they are coherent with other published data at the national and regional level, should be used with caution.

Some detailed estimates in this publication may have been confidentialised to avoid potential identification of agricultural holdings in accordance with the Census and Statistics Act 1905 and the Census and Statistics Determination 2018.

Where detailed estimates have been suppressed for reasons of confidentiality, they have been included in relevant aggregated totals.

Where figures have been rounded, discrepancies may occur between sums of the component items and totals.

These statistics are produced through the integration of multiple data sources. This integration of multiple data sources can help to manage the strengths and weaknesses of different data sources some of which are developed for a range of purposes that are not purely statistical. A number of assumptions have been made in the integration of these data sources and these are highlighted below:

• Tree crop mapping data is assumed to be correct in identifying the accuracy of crop area, crop type and crop planting year. There are potential errors associated with each of these that are not measured.
• The bearing age of trees for each crop type is assumed to be consistent.
• Yield data is assumed to be consistent for each crop within each region, this does not consider crop variety, soil condition or irrigation levels.

Revisions are a change in a published value and may arise due to a variety of reasons. 

Revisions may be applied to horticultural crop estimates due to:

• revisions to source data
• refinements to decisions made around the treatment of data anomalies in the series
• implementation of methodological and process improvements.

The methods and data sources used to produce these estimates will be subject to ongoing review to improve the accuracy and timeliness of outputs.

There are conceptual differences between historical estimates produced from ABS’ Agricultural Census and agricultural surveys, and the data produced using the new data sources and methods appearing in this publication. These include:

• the scope of the survey based estimates was businesses with an estimated value of agricultural output (EVAO) of $40,000 or more. Non-survey data sources do not have this restriction and measure the full extent of Australian agriculture.
• survey based estimates were subject to sampling variability and may have differed from the estimates that would have been produced if information had been obtained from all businesses.
• survey based estimates for horticultural crops related to production and area of commercial crops grown for human consumption. Tree crop mapping area data may include areas of crops grown for other uses and not align with the crop harvested and sold for the region.
• survey based estimates for fruit trees reported number of total, bearing and non-bearing trees. The experimental method presented here, reports total and bearing hectares.

For more information on the historical survey-based estimates see Agricultural Commodities, Australia.

Production and local value statistics for a wide range of horticulture commodities as well as additional information on area and business counts for avocados and macadamias are available for Australia and states/territories in data cubes in this publication. 

Avocado and macadamia production, local value, area and the business count (avocados only) statistics are available on Statistical Area Level 2 (SA2) regions through the ABS' Data Explorer.

National, state/territory and SA2 boundaries used to publish statistics in this publication are from the Australian Statistical Geography Standard (ASGS) (Edition 3).

Other ABS publications containing agricultural data are listed on the ABS’ Agricultural Statistics page. The ABS also issues a daily Release Advice on the website which details products to be released in the week ahead.

The ABS Privacy Policy outlines how the ABS will handle any personal information provided to us.