Regional scale simulation of reduced fertilization scenarios

In this study a regional simulation in Flanders, Belgium was performed with a Roth-C model to study the effect of reduced mineral fertilization on the soil organic carbon (SOC) stock. For this study, a model environment developed for a European scale simulation was used (Pape 2024) to determine the loss of SOC resulting from a reduced fertilization with a reduction of 20% N for seven crops (barley, grain maize, silage maize, potato, rapeseed, sugar beet and wheat). Data for the crop yield and area specific for Flanders were used to perform the simulation. The model environment uses yield reductions to simulate the impact on SOC of a reduction of 20% mineral N fertilization by simulating monocultures and using the share of agricultural area to find the resulting difference in SOC for each crop. The difference in C input is proportional to the yield reduction due to a reduction in fertilisation of -20% N, the yield and the crop’s share of agricultural area. In the case study for Flanders a total of 0.84 Mg C/ha SOC loss over a period of 30 years was simulated. Most of this SOC loss was linked to cultivation of rapeseed and wheat, followed by grain and silage maize. In the simulation for Flanders higher SOC losses were found for the crops silage maize, grain maize, potatoes and sugar beet in comparison to the European scale simulation, while for the crops wheat, grain maize and barley a lower SOC loss was found. In Flanders maize is the most commonly grown crop, followed by cereals, potatoes and sugar beet. Due to a high share of area for these crops in Flanders high SOC losses were found, some higher than found in the European scale simulation. In the European scale and regional scale simulations, rapeseed has a high contribution to the simulated SOC losses, but this is rather linked to the high reduction in yield as a consequence of the reduction in mineral fertilization than due to the crop’s share of total arable land. With the introduction of the regional specific data for yield and area, a more realistic simulation of the SOC losses were obtained due to the inclusion of detailed data on the crop’s share of cultivated area. In Flanders, the agricultural area is very fragmented and scattered over all of Flanders. In European datasets this leads to very rough estimates of agricultural land which does not coincide with the reality. To further develop this scenario additional regional specific data such as climate, clay content and SOC content should be incorporated in the regional scale simulation.