Year 2 of the project saw additional blocks added. Most attention remains at the MicroFarm, but three commercial paddocks in Hawke’s Bay have been tracked through the season. Three Auckland paddocks were also included.
The purpose of this exercise was to test the concept of Management Action Zones. We identified zones in these paddocks and confirmed them after establishment as being of lower population and or plant size.
In each we set up 15 monitoring plots. All up we had 180 plots in 12 zones. We followed their development paying particular attention at 3 leaf, 5 leaf and bulbing then counted and are weighing all plots at harvest.
We also used the MicroFarm crop to test the identification of zones using ground cover mapping. For this we used the 2016 version of CoverMap and high accuracy (sub metre) GPS.
The cover map was converted to a predicted yield map. While the app can determine relative canopy cover, it does not count plants so the population is unknown. This is a key piece of information for predicting yields so our map values are quite likely incorrect. However, the estimate is not dissimilar to experienced grower estimates made at the same time.
Until we can automatically count plants as the paddock is surveyed, sample plots remain the best way to collect that information. However, we can now map the canopy cover which seems a good way to identify areas likely to be Management Action Zones.
We also looked at how a grower might apply variable rate inputs. Assuming equipment can only apply a uniform rate per swath, we tried 10 and 20m grids.
The 20m grid seems too coarse for a high value crop like onions. The 10m grid seems to show reasonable similarity with the detailed yield map.
What would you do?