Simulation of variable-rate manure application under different application scenarios

Abstract

Although the variable-rate manure application (VRMA) offers a promising solution to overcome the shortcomings of the conventional uniform-rate (UR) application, effectively incorporating this technique into precision agriculture solutions presents a substantial challenge. This study entailed developing a quantitative analysis tool to calculate manure consumption, evaluate environmental risks, and compare different treatment schemes under different implementation scenarios. Firstly, a simulator was designed and developed in the LabVIEW development environment with MATLAB. Next, this tool was used to compare traditional UR application with VRMA, by which manure application rate was varied according to soil phosphorus (P) maps measured with an on-line visible and near infrared soil sensor, or according to data-fusion (DF)-based management zone (MZ) maps, developed by fusion of on-line measured soil properties and crop normalized difference vegetation index (NDVI). In the DF case, simulations included Robin Hood (RH) and Kings approaches, which meant to apply the largest manure application rate at the poorest fertility zones, or in the richest fertility zones, respectively. The implementation scenarios included both map-based (MB) without accounting for on-line manure sensing and map-sensor-based (MSB) with on-line manure sensing in three commercial fields in Flanders, Belgium. The results revealed that the P-based-VRMA scheme consumed 1–13% less nitrogen (N) and P2O5 in the case of ignoring legislative limits, compared to the UR and this was true for the RH-VRMA scheme in all three fields and for the Kings-VRMA scheme in two out of three fields. This was because the P-based maps have larger rich fertility areas (e.g., large P concentration) than the MZ maps that both the RH and Kings approaches used. Furthermore, imposing the “MAP6” legislation limits caused all VRMA schemes to consume less fertilizer than would be the case without restrictions, reducing environmental risks due to decreasing the amount of applied N and P2O5. According to the simulation results, the MSB scenario saved 6–9% manure compared to the MB scenario in the P-based-VRMA scheme when the mean value of real-time sensed P2O5 in the applied manure was greater than the nominal values measured in the laboratory. Opposite results were observed in the UR, RH-, and Kings-VRMA schemes. The MSB was more expensive than the MB by 5–7%, when the mean N content of manure used for MSB was lower than that for MB. We concluded that the manure consumed in VRMA depends on treatment schemes, measured manure quality, imposing legislative limits, and the proportion of rich, medium, and poor fertility areas in a field. It is essential to impose the legislative limits to prevent over-applications of N and P2O5, while further agronomy study should evaluate the impact of these limits on crop responses.