Graphical Abstract


Whole-farm spatial appraisals are rare and often without yield data validation.



  • Whole-farm spatial appraisals are rare and often without yield data validation.

  • We processed, validated and analysed a ten year whole-farm yield map data set.

  • Wheat yield map data had a RMSE of 1.0 t/ha with a mean error of 8.1%.

  • 33% of the farm was losing >£100/ha compared to the best performing zone per field.

  • Yield data validation is required when developing precision agriculture techniques.



Statistical methods used for delineation of field management zones and yield stability are frequently only applied to relatively small areas, with few studies performing rotational, whole-farm economic spatiotemporal appraisals. To enable accurate economic analysis, yield map datasets must contain minimal errors while cleaning procedures are often used to remove errors, it is rare that cleaned data is validated before its application.


The objective of this study was to process, validate and combine spatial statistical approaches for a rotational yield map dataset from a whole-farm across 7 crops in a winter wheat based rotation. Developing a framework for using validated yield map datasets to support precision agriculture techniques that are applicable for farm-level decision making.


The rotational completeness of a 10 year combine yield map dataset for a 435 ha farm in Eastern England was assessed. The dataset was cleaned statistically, and its accuracy assessed by comparison with recorded yields from trailer weigh cells. The cleaned, validated, and corrected yield map dataset was used to identify management zones across the whole farm using fuzzy clustering. The temporal stability of management zones and economic performance across the rotation was also assessed.


Data cleaning methods removed 16% of data points, improving the degree of spatial correlation within the individual yield maps. Independent validation demonstrated varied accuracy of yield maps from combine harvester data and errors in wheat ranged from 0.53 to 1.53 t/ha RMSE. These errors have implications for researchers using combine yield data to develop and validate precision agriculture technologies. This data set required correction before yield data can be applied with confidence for on-farm decision making. Compared to the zones with the highest margin in each field, 34% of zones had an average annual margin loss of >£100 ha. The temporal stability of the resulting management zones also varied. Areas with the lowest economic performance and greatest yield stability across years will potentially see the greatest economic and environmental benefits from precision agriculture techniques.


The accuracy of combine yield map data should not be assumed. The application of these datasets, including for the identification of management zones or in developing precision agriculture techniques should attempt to address this through data cleaning and validation procedures. Only then should it be used for on farm decision making, such as identifying areas with the most economic benefit by applying precision agriculture tools such as variable rate nutrient applications.

David E. Clarke, Elizabeth A. Stockdale, Jacqueline A. Hannam, Benjamin P. Marchant, Stephen H. Hallett,
Whole-farm yield map datasets – Data validation for exploring spatiotemporal yield and economic stability,
Agricultural Systems,
Volume 218,
ISSN 0308-521X,

Download the full article below.

Related Organisations

Connected Content

Agriculture now generates vast amounts of data of different types and at different scales. The ownership, access, control, sharing, value, permissioning, intergration, security, safeguarding and sovereignty of this data represent multiple challenges and opportunities for the sector.

The challenges of food security, climate change and sustainable development present exciting opportunities for agricultural research and innovation. The NIAB Group is the UK’s fastest growing crop science organisation, having trebled in size over the past decade through a strategic programme of investment, merger and acquisition.

The agri-tech sector is vibrant and growing, with many exciting companies, organisations and networks are working to develop agri-tech solutions.

As a specialist postgraduate university, Cranfield’s world-class expertise, large-scale facilities and unrivalled industry partnerships are creating leaders in technology and management globally

We are a world-leading independent research organisation providing objective, expert geoscientific data, information and knowledge.

Crop yield has been a major subject of research and industrial improvement for decades.  The YENs have set out a framework for understoanding yield in terms of capture and conversion of the major resources light and water. Temperature plays a role in crop development and can also have a substantial impact on growth and yields.