Browsing by Subject "Model Kalibrierung"

Now showing 1 - 1 of 1

Validation and evaluation of the DNDC model to simulate soil water content, mineral N and N₂O emission in the North China Plain
(2009) Kröbel, Roland; Römheld, Volker
Using measured datasets (various soil properties, the soil water content, daily N₂O emissions, and different crop parameters) from a multi-factorial field experiment (N fertilisation, irrigation, and straw removal) in the years 1999-2002 on the experimental site Dong Bei Wang (DBW) in the North China Plain (NCP), the ability of the process-oriented model DNDC (DeNitrification-DeComposition) was tested to simulate soil processes, and especially N₂O trace gas emissions. The soil is classified as ?calcaric cambisol? (16 % clay content), while the site itself is further characterised by the regime of a continental monsoon climate. The central hypothesis in this work was that a thorough testing of the model (using a considerable range of different datasets) will allow the identification of shortcomings or discrepancies in the model, and that, given the linear succession of model calculation steps, the model calculation can be improved step by step, starting with improvements of initial calculation steps before continuing the improvement of following calculation steps. Due to increases in the N₂O atmospheric concentration, and a lifetime of 100 to 150 years for one molecule (as well as a global warming potential 32 times that of a CO₂ molecule), N₂O is estimated to account for 7.9 % of the global warming potential. 70 % ? 90 % of the anthropogenic N₂O emissions are thought to origin from agriculture. The formation of nitrous oxide is dependent on the availability of reactive nitrogen, and, therefore, mainly influenced by the N fertilisation rate, fertiliser type, application timing and method. China, and the main cropping area NCP, are expected to contribute considerably to the anthropogenic N₂O emissions. The DNDC model consists of two compartments, which first calculate soil temperature, moisture, pH, redox potential and substrate concentration profiles from climate, soil, vegetation and anthropogenic activity datasets, and in a second step NO, N₂O, CH4 and NH3 fluxes. In accordance with the data availability, the simulation of the soil water content, the mineral nitrogen concentration, and the N₂O fluxes were investigated. An automated parameter optimisation (using the software UCODE_2005) and programmed changes in the source code were conducted to improve the model simulations. In result, neither the automated parameter optimisations, nor the programmed changes, were able to improve the unsatisfying default simulations of the DNDC model. The results of the cascade model, employed by the DNDC model to simulate soil water dynamics, suggest that conceptual errors exist in the model calculation. Also the results of the mineral nitrogen and N₂O emissions simulations suggest shortcomings in the model calculation. The best agreement between measured and simulated total cumulative N₂O fluxes was achieved using an adapted (90 cm soil depth, adjusted SOC fractioning, and added atmospheric N deposition) default model version, despite unsatisfactory simulations of soil water content, mineral nitrogen, and daily N₂O fluxes. Thus, in conclusion, the investigated DNDC model version appears to be able to give an approximation of seasonal N₂O fluxes, without being able to simulate the underlying processes accurately in detail. Therefore, caution is suggested when modelling sites on the process level.