Abstract
– Using the FarmLab zone sampling tools, Precision Pastures saved the farmer 20% ($1,655.70) on lime application when compared to a single rate lime application. Further savings are anticipated through the farmers continued use of the zones and soil data over the course of the soils and pasture production management plan.
– The seamless integration of the FarmLab mobile and web-based software reduced administration costs and time for Precision Pastures agronomists, in addition to strengthening their relationship with the client through added transparency and sharing of results.
– The use of the FarmLab software, combined with Precision Pastures’ local expertise, resulted in a meaningful balance between total sampling costs and accuracy of recommendations. Whilst other sampling strategies such as grid-based sampling would yield more accurate recommendations, the balance of cost and accuracy offered by the zones gave the agronomist confidence in the recommendations whilst still offering a cost-effective solution for the client.
– The long-term storage of soil information on FarmLab provides a tool for the farmer to monitor soil health over time as an integral part of farm production and sustainability management.
Background
On 13 and 17 June, FarmLab and Precision Pastures agronomists undertook a sampling activity at Myanbah, a 1,273 hectare (3,145 acres) mixed livestock and cropping operation near Yarrowyck, New England. The goals of the activity were as follows:
– To identify key production paddocks on the new property and help make recommendations on how to improve across those designated areas.
– To generate a soil map over Myanbah to baseline and inform production strategies across the property.
– To assist Precision Pastures in identifying and communicating their recommendations.
– To test and validate FarmLab’s soil zoning features.
– To test and validate the FarmLab mobile app in geo-locating each soil sample and streamlining the sample recording and submission process.
Previous samples had been taken on the property, however they did not contain GPS coordinates, sampling methods and other important metadata. This rendered the information useless, and it was decided that the paddocks previously sampled would need to be resampled in order to gather appropriate insights.
Method
Using the FarmLab™ zoning tool, which combines publicly available gamma-radiometric data, Digital Elevation Maps and Landsat NDVI aggregated over several years, Precision Pastures created 5 zones that were likely to represent similar soil properties across Myanbah (Fig 1). The basis of these datasets were various studies from GRDC and CSIRO showing the relation between gamma-radiometrics and soil properties. This was combined with expertise from University of Sydney in the creation of zones as part of the FarmLab involvement in the Soiltech Project.
Precision Pastures weighted each dataset in the K-means algorithm to best reflect the soil types across the area. This was done using a combination of available satellite imagery and ground truthing.
Following the creation of the 5 zones using FarmLab, sample points were designated by the agronomists who believed these would accurately reflect each zone. Non-contiguous zones being sampled over multiple areas to ensure potential variation in soil properties was captured. In total, 4 sites were chosen in the northern section of the property for sampling at depths of 0-10cm, 10-50cm and 50-100cm. These sites were chosen as they reflected known high production areas. Another 11 sites were sampled over the southern zones at a depth of 0-10cm, with each 0-10cm sample consisting of 4-6 sub-samples, bulked from a nearby area within the zone. Figure 1 reflects the final sample sites chosen and sampled against each zone.
When planned sample sites were not accessible or suitable for sampling (for example, those close to fence lines, or in dense vegetation), sites were adjusted in the field based on the agronomist’s experience and knowledge of the region. These ‘actual’ sites were recorded using the FarmLab mobile app on a Samsung Galaxy 9 and a standalone GPS locater accurate to 50cm provided by Precision Pastures.
At each site, samples were bagged and the barcoded bags were scanned using the FarmLab mobile app against the depth and sample site recorded (Figure 2). Upon completion of sampling on 17 June, all bags (23 samples in total) were submitted to EAL (Southern Cross University) for testing. Tests conducted were ‘Agricultural Standard 3’ for 0-10cm samples, and ‘Agricultural Standard 1’ for all other samples. Further details on the nutrients and properties analysed under each test can be found at https://www.farmlab.com.au/tests. The primarily testing being for soil properties, Cation Exchange Capacity and pH levels.
Sample results were returned via the FarmLab app, and provided to Precision Pastures agronomists for interpretation. Precision Pastures compared the GPS locations and sample barcodes recorded via the app against their own paper records, and once validated, developed a report with recommendations to the farmer.
Precision Pastures were able to use the results to make recommendations according to which zones were suitable for grazing, and those that were suitable for cropping. As the zones often extended over multiple paddocks, the farmer was able to use the recommendations for multiple paddocks rather than a single paddock. As the entire farm contains 22 paddocks, this was an significantly valuable outcome by reducing sampling costs whilst maintaining the required level of accuracy.
In total, the activity took approximately 2 weeks. Most of this time was spent waiting for the sample results to be analysed by the lab (including transit time from Armidale to Lismore where EAL are based).
Outcomes
All intentions of the sampling activity were met for both Farmer and Agronomist. Technically, the trial served as an important opportunity for FarmLab to trial the accuracy of the zoning tool and stability and offline functionality of the mobile app – all of which was deemed a success. The mobile app providing efficient in capturing GPS coordinates accurately even when out of mobile network range.
The main benefits of the activity for both Farmer and Agronomist can be summarised as follows:
Farmer’s benefits
– Using zones to determine sampling strategies rather than by paddocks meant that recommendations could be made over several paddocks at once. In this case, the soil sample results were mapped back to the zones and a clear picture was developed around where the farmer should concentrate his efforts. Previous sampling methods would have provided the farmer recommendations per paddock, making interpolation across multiple paddocks challenging and reducing the accuracy of the recommendations. By identifying zones based on common soil properties there was a reduction in total samples taken whilst accuracy of the recommendations was still improved. When compared to sampling a single paddock, the ability to apply a recommended variable rate lime application across various paddocks put the farmer’s savings at 20%, or $1,655.70. This was based on a single rate application across sampled paddocks, estimated at a total of $8,441 over the 3 paddocks identified for improvement, compared to $6,785.31 estimated expenditure for variable rate application of lime across those same paddocks using FarmLab zones. Lime costs including freight and spreading was estimated at $100 per tonne.
– By providing the farmer with soil sample data recorded against GPS locations, he now has a baseline to measure ongoing soil health improvements over time. This data, made available via FarmLab for the life of the farm, will be used to provide continued discussions and history with Precision Pastures agronomists in the improvement and management of prescribed areas over Myanbah. As this data is added to and improved, the trends shown can be used to generate a picture around the farmer’s soil and land management – not only for the farmer, but for future buyers, insurers and investors.
– The farmer provided the following feedback on the process; “FarmLab and Precision Pastures provided me with a soil testing plan that used digital and satellite technology to identify separate zones. This immediately saved me the cost of extra soil tests which would’ve been done using traditional transecting soil test. Then once the results came back, they were able to further reduce my costs by providing mediation zones rather than remediating the whole paddock. Also with the soil tests GPS located, we can retest the same zones in future and demonstrate the improvement we’ve had and ultimately the increase in land value.”
Agronomist’s benefits
– Because the FarmLab app was used to reliably and accurately record barcodes against sample sites in the field, Precision Pastures agronomists saved time in matching written records with GPS points. Whilst the estimated time saved was only 2 hours, the reduction in errors that this provided in transcribing barcodes using pen and paper improved accuracy by about 10%.
– The average difference in accuracy between the Precision Pastures GPS points (recorded via a dedicated GPS locater), and FarmLab GPS points recorded with a Samsung Galaxy 9, was 3m. In this case, due to the size of the zones and sample areas, the difference was deemed insignificant. The benefit resulting from the improved accuracy in matching sample barcodes to GPS points outweighed the difference in accuracy.
– Precision Pastures were able to use the soil zones to demonstrate to the farmer the differences over his farm. Not only did they aid in the sampling process, but they provided a tangible insight for the farmer to support Precision Pastures’ recommendations. The ‘soft’ outcome of this is that it worked to further strengthen the relationship between agronomist and client by providing transparency and trust. Total consulting costs were comparable to costs normally associated with a transect sampling activity over a range of paddocks. In this case the entire farm was mapped, and so the ‘value’ offered by the agronomist was significantly increased, with further savings provided to the user at the end of the day based on variable rate lime recommendations.
Conclusion
When combined with the local knowledge and expertise of the agronomist, the FarmLab zoning and mobile app was an effective land management tool with reaches beyond just improving soil sampling logistics. The areas identified by the zones were accurate reflections of soil types and were used to indicate new strategies around which paddocks could be cropped and grazed. The integration of soil sampling through a mobile app, alongside zoning created using scientifically proven concepts provided a distinct advantage in accuracy and user experience. FarmLab are continuing to work with Precision Pastures to improve the accuracy of zones and the useability of the tool and feedback from the farmer will be gathered over the coming season as he begins to apply the recommendations.
The zoning tool is scheduled to be made publicly available in the 4th quarter 2019, whilst the FarmLab mobile app will be available via the Android Play and iOS App store on 01 Sep 19.