James Taylor, Brett Whelan (& a wee bit of help from Alex)
Until now the ACPA has concentrated most of its effort on broadacre crops (grains and cotton). This year sees a shift as the ACPA broadens its research field and begins to look at the application of precision agriculture technology to horticultural crops. Such work has previously been impractical due to the absence of satisfactory yield monitors. The last year however has seen the release of yield monitors for a variety of machine-harvested horticultural crops including potato, tomato and wine grapes.
Using a HM-570 grape yield monitor kindly supplied by Ron Campbell HarvestMaster, Utah, USA, James Taylor of the ACPA has begun monitoring the vintage at Orlando-Wyndhams Richmond Grove vineyard at Cowra. While the majority of the vineyard is yet to be harvested some varieties have been completed and preliminary analysis begun. Below are two yield maps of part of the verdehlo crop. Map 1 shows the raw yield data as retrieved from the yield monitor. The yield data has been divided into 25% quantiles and mapped (Red = 0-25%, Yellow = 25-50%, Light Green = 50-75%, and Dark Green = 75-100%). While the raw data is noisy it can be seen that Block A has a greater percentage of red compared with Blocks B and C indicating lower yield.
Map 1. Shows raw yield data as retrieved from the yield
monitor and plotted in Jmp. The data is divided into 25% quantiles.
(Legend: Red <14 t/ha, Yellow
14-18 t/ha, Light green 18-22 t/ha, Dark
green >22 t/ha)
Grape yield can be better estimated by the use of local block
kriging. In this instance we have applied ordinary 5 metre by 5 metre
kriging function with local exponential variogram to the data to produce
a more coherent map on a 2metre raster. (Map 2). This clearly
shows the lower yield in Block A compared with Block's B and C. Map
2 also indicates a low yielding area in the centre of Block B. The
difference in yield between Block A and Blocks B/C can possibly be explained
by their location. Blocks B and C are located on fertile river flats
while Block A is situated on a gentle slope. A change in topography
is often associated with a change in soil type which in turn affects yield.
However soil testing will be needed to confirm such a hypothesis.
The low patch in the centre of Block B may also be due to a localised soil
difference. Alternatively it may be a result of a localized pest/disease
outbreak or management error. Liasing with the farm manager and staff
will help to identify such problems.
Collection of yield data is the first step toward precision viticulture however it is not the most important step. It is the ability of the farmer/manager to interpret the cause of the yield variation that will be determine how valuable the yield data is. Accurate identification of yield determining characteristics and effective remediation will allow the farmer to improve both productivity and profitability.
To do this we then need to understand the specific plant-soil-climate interactions at each site in the field. It is our aim to try and understand these site-specific interactions through the use of several data layers including yield maps, soil maps, satellite imagery, plant tissue analysis, management decisions, disease/pest maps etc. By analysing and intergrating these data layers we hope to be able to derive yield-determining factors at each site in the field and then a method to differentially apply inputs to satisfy each individual sites requirements. Through such research the huge potential for precision agricultural technology will be realised in Australian viticulture.
For further information please contact
James Taylor or Brett Whelan
Australian Centre for Precision Agriculture
McMillan Building, A05
University of Sydney
NSW 2006
Australia
ph. +61 2 9351 5813
e-mail: j.taylor@agec.usyd.edu.au or acpa@acss.usyd.edu.au
Thanks to Ron Campbell, CEO, HarvesterMaster US for kindly providing the yield monitor and Orlando-Wyndham for the provision of a harvester and test site.
THE UNIVERSITY OF SYDNEY
© 2008 - Australian Centre for Precision Agriculture