Graham Centre agriculture business management researchers have crunched the numbers on precision site-specific nitrogen applications in dryland farming systems in northern Victoria – and found no clear economic benefit from the practice.
The research was carried out in support of a project investigating ‘in-paddock variability’ led by our farming systems group partner Riverine Plains Inc, funded by the Grains Research and Development Corporation (GRDC).
Charles Sturt University Adjunct Associate Professor, Dr Tom Nordblom said, “In precision variable-rate nitrogen applications (VRN), the idea is to optimise placement of inputs at the right rates; neither too little nor too much in each target area according to selected criteria at the right time for known and expected conditions.”
The research made use of farm physical and financial data from the Yarrawonga and Dookie areas .
Impact in the paddock
Dr Nordblom said the research made use of detailed historic geo-referenced field data on crop yields from 2000 to 2017, growing season rainfalls from the past 56 years and EM38 (electro-magnetic conductance) soil maps, often used to define zones for targeting precision N rates.
“Two farms provided detailed wheat and canola yield resords over an 18-year period,” Dr Nordblom said.
“The average yield per hectare from all wheat paddocks on one farm in one year, paired with the growing season rainfall of that year, provided one of 19 data points in Figure 1a. For canola yields in Figure 1b we found 13 such data points.
“In-paddock variability is illustrated in the case of a paddock for which we had detailed yield data that could be expressed as geo-referenced 90 x 90m grid areas over six years including dry and wet seasons (Figures 2 and 3); two grid areas are marked which exhibit opposite yield responses to low and high growing season rainfalls.“We found some of the lowest-yielding grid areas in dry years became the highest in the wet years and vice versa, apparently due to differences in soils and relief (lateral flow and waterlogging).
“Repeatability of the yield rankings was very low, with fewer than 5 per cent of grid areas remaining in the same 25-percentile range of yield in four out of five years,” he said.
“Because yields are most limited by growing season rainfall, which cannot fully be known until late in the season, EM38 zoning alone will not always be a reliable guide to yield potentials.”
Impact on the bottom line
The study also simulated the long-term effects of VRN on the financial equity of hypothetical farms, with analysis of both the price and weather risks over time.
“We wanted to take the analysis of VNR a step further in a whole farm risk analysis accounting for variations in weather and prices, as well as spatial variations in farm soils. We also considered all costs including the cumulative effects of interest over time,” Dr Nordblom said.
“Based on our initial results for the Yarrawonga andDookie area of northern Victoria, VRN needs to increase crop yields by less than one per cent, or reduce costs of N applied by at least 70 per cent, to break-even over time.
“We found that because rainfall was the key driver of productivity in the dryland farming system of the study area, there was no clear economic benefit of VRN.”
The analysis also compared two modelling scenarios, a low-cost farm and a high-cost farm.
“A low-cost farm, beginning with no debt and assuming no further benefits of VRN, showed no risk of loss over random decades drawn from the past 56 years of weather records.
“However, a high-cost farm already facing a 41 per cent risk of loss, could find its risk of loss increased to 44 per cent with the additional costs of VRN,” Dr Nordblom said.
The study titled ‘Financial analysis of variable-rate Nitrogen (VRN) in the Riverine Plains Region’ is by Dr Tom Nordblom and Dr Sosheel Godfrey from Charles Sturt University, Dr Tim Hutchings from Meridian Agriculture and Dr Cassandra Schefe, Research Co-ordinator, Riverine Plains Inc.
It was a selected paper at the annual conference of agricultural economists and policymakers in Australasia, AARES 2019 in Melbourne. Read it here: https://ageconsearch.umn.edu/record/285048
The researchers also want to acknowledge the help of Mr Craig Poynter from the Charles Sturt GIS & Remote Sensing, Spatial Data Analysis Network (SPAN).