Development of an online decision-support infrastructure for optimized fertilizer management using a probabilistic approach

Determining an optimalfertilizer application rate involves several factors, including past managementpractices, soilcharacteristics, weather, commodity prices, input materialcosts, and risk preferences. Spatial and temporalvariations in these factorsconstitute sources of uncertainty in selecting the most profitable applicationrate.Therefore, a decision support system (DSS) is needed to help minimizeproduction risks amid uncertain cropperformance.

This paper presentsadvancements to the newly developed dynamic DSS, NumericAg, which estimates theprobability of achieving expected profits under specific growth conditions. Theproposed system includes adatabase, a user interface, and a numerical enginefor computing the profit space. The web interface(www.numericag.com) allows a user to specify production conditions(e.g., previous crop, tillage system, soil type,organic matter content,rainfall, and crop heat units) and to accept or modify grain and fertilizerprices. The databasestores the results of previously recorded fertilitytrials. The profit space computation engine was designed toestimate theprobability of achieving different net return levels relative to nitrogenfertilizer costs for each potentialapplication rate. The profit space engineconsiders over 20,000 potential quadratic-plateau fertilizer responsefunctionsand 49 cost scenarios.

The DSS evaluates eachcombination of potential scenarios against each fertility trial, weighted bythe user-specifiedgrowing conditions. Consequently, the probability ofdifferent levels of potential net return relative to fertilizer costwasestimated using fertilizer response observations that closely match theuser-specified growing conditions. Asensitivity analysis was used to show DSSresponse to changes in specified growing conditions.

When comparing differentgrowing conditions, a lower probability of achieving relatively high profitswas observedwith sandy soils, relatively low crop heat units and wateravailability, or low nitrogen contribution from previouscrops. Although therate that maximized the expected net return relative to nitrogen cost did notchangesubstantially, the rate of profit decline due to under-application ofnitrogen fertilizer varied across growing conditions.