An Investigation into the Factors Influencing the Accuracy of Yield Maps

Keywords: Yield mapping, Mapping errors, Yield meter, Precision farming, GPS.

Yield mapping is the continuous measurement and recording of grain yield, and its position within the field, on-board the combine while the machine is harvesting. A yield map is interpolated from the raw data which indicates, by contour lines, yield variations across the field. As a result of yield mapping a number of farmers are already using yield maps to target existing resources in areas of low yield to maximise both yield and gross margin. However, the potential to gain a greater financial return and associated environmental benefits exists by varying crop inputs to match the yield potential of different parts of the field. Yield maps are therefore, becoming an increasingly important source of information within Precision Farming.

The aim of this study has been to investigate the magnitude of the errors associated with:

(i) raw data in the yield map,

(ii) the effect of the errors on commercial yield maps.

To achieve this aim, a Massey Ferguson combine, equipped with a yield mapping system, was used to yield map a number of trial fields between 1992 and 1997. While harvesting the trial fields, grain samples were taken at regular intervals and analysed for grain density and moisture content. In addition, a test rig was constructed to enable the yield sensor to be validated for its ability to measure instantaneous grain flow – the primary measurement required for yield mapping. Data gathered from the trial fields, analysis of grain samples and the test rig were used to identify four areas of error:

(i) ability of the yield meter to measure instantaneous grain flow,

(ii) other crop variables not measured by the yield mapping system,

(iii) positioning of yield readings within the field, and

(iv) the operation of the combine within the field.

It was concluded that the Massey Ferguson “Flow Control” yield sensor provided a typical error range of +/-6% when measuring instantaneous grain flow and +/-0.5% when measuring a batch weight. The within field variation of grain density was found to have a significant affect on volumetric yield measurement with errors of up to 16%, despite regular calibration of the sensor. Operational errors, as a result of incomplete filling of the cutter bar, incorrect lead and lag timings, were found to introduce average errors of -38%, -20% and -16% respectively into the raw data. These errors were reduced significantly in the yield map after data interpolation.

To conclude, yield maps depict yield levels within an acceptable error provided the yield meter is:

(i) of a mass flow type,

(ii) calibrated correctly,

(iii) the combine operator takes precautions to drive the machine to ensure the quality of the raw data, and

(iv) errors within the raw data are deleted before interpolation.

The maximum resolution that a yield map is able to indicate yield levels is 20m x 20m to 25m x 25m. This is limited by the smoothing effect of the crop flow characteristics through the grain transport mechanism of the combine and the parameters specified during data interpolation.

M R Moore, AGCO Ltd, Coventry, UK

P R Kremmer, Dronningborg Industries a/s, Randers, Denmark

60 pages, 17 figures, 16 tables, 21 refs.