A normalised difference vegetation index (NDVI) map was created from four-band satellite data captured in November, 2015. At the time the satellite image was captured the plants were small.
The map was used to identify areas where November biomass was high, medium and low, and in each we selected two areas to take samples at harvest. In each area we took eight samples, giving 48 in total.
Individual sample yields ranged from 51t/ha to 144t/ha showing just how variable onion crops are.
Comparing the weighed harvest samples with the satellite map values gave us a relationship that we used to create a yield map.
The yield map was converted to a profit map using gross margin analysis based on grower-provided costs and returns.
The profitability map showed gross margins averaged $3225/ha in the red area against $11,884/ha in the green.
The implication is the green part is more than three times as profitable as the red.
The breakeven yield with price and costs at levels used is 62.5t/ha. This indicates there were no negative profit areas in this paddock assuming the yields, costs and prices used.
However, if the onion price dropped to $370/t with costs staying the same, the breakeven yield would be 87t/ha and about one-third of this paddock would be below breakeven yield.
We wanted to know how profitability would be improved if the whole area performed well so compared achieved yield with potential yield. We assumed potential yield was in the 90 percentile yield.
That suggests the medium yield (yellow) zone lost 17t/ha or $8002 and the low yield (red) zone lost 37t/ha or $6667 after production and packing costs are removed. Against all zones this represented a potential overall loss of $27,945.
If losses result from some natural limiting factor the lower yield might have to be accepted. But in other cases management can change the outcome.
A satellite NDVI map with biomass zones selected for in-situ yield assessment.
In this case, blocked council drains caused total crop loss in the top corner of the paddock. We also looked at surface ponding using OptiSurface drainage software and found drainage patterns showed some correlation with the lower-yielding areas.
The OptiSurface programme allows us to model raised beds, a feature that significantly affects surface drainage flow paths and ponding. The drainage analysis indicated large areas of ponding in excess of 80mm deep.
OptiSurface creates a surface that ensures there is always a route available for drainage water to flow down.
In this case it effectively halves the volume and depth of soil movement compared to traditional laser grading.
Our design had very little soil cut or fill in excess of 100mm, and much would be less than 50mm.
The estimated contractor charge for earthworks on this site is $2000/ha, or about $14,000. This is about half the estimated cost of lost production from this season’s crop.