Evaluation and adjustment of description of denitrification in the CoupModel, DNDC and DeNi model based on N2 and N2O laboratory mesocosm incubation system measurements

Abstract

<p>Quantifying soil nitrogen processes – especially denitrification – are critical for the adequate prediction of the produced and emitted N<sub>2</sub>O and N<sub>2</sub> gasses and the production and consumption of NO<sub>3</sub><sup>-</sup> and NH<sub>3</sub>. Biogeochemical models are useful tools for the description of these N processes, but recent research progress is not considered on the denitrification sub-modules of these models. Denitrification typically occurs in hot-spots of the soils but the models describe the soils as a homogenized system. Another critical problem is the calibration of the decomposition sub-modules. Suitable soil N<sub>2</sub> flux data were not available during the development of the extensively used models but new measurement techniques provide appropriate N<sub>2</sub> gas flux data.</p><p>In this study we investigate the N<sub>2</sub> and N<sub>2</sub>O fluxes from mesocosm experiments of different complexity and use the measured data and experimental settings for testing the denitrification sub-module of existing biogeochemical models.</p><p>Two arable soils – a silty loam and a sandy soil – were used for the experiments and varied with N fertilization and organic matter amendment. The soils were incubated in laboratory incubation systems over 42 and 58 days, respectively. N<sub>2</sub>, N<sub>2</sub>O and CO<sub>2</sub> fluxes were quantified by gas chromatography and isotope-ratio mass spectrometry. Seven moisture and three NO<sub>3</sub><sup>-</sup> contents were set up to the loamy soil and only the temperature was manipulated during the experiment, while other factors were kept constant. In the experiment with the sandy soil, incubations were conducted with or without incorporation of organic litter (ryegrass) and initial water content was adjusted equivalent to 80% water-filled pore space. Temperature, water content and NO<sub>3</sub><sup>-</sup> content were manipulated during that experiment.</p><p>Three commonly used biogeochemical models – namely CoupModel, DNDC and DeNi (a self programmed early stage version of the nitrification and denitrification sub-model of the DailyDayCent) – were tested on the experimental data.</p><p>The average N<sub>2</sub>+N<sub>2</sub>O fluxes of the loamy soil as given by measurements, DNDC, DeNi and CoupModel calculations was 287.5±202.3, 1.8±0.5, 779.1±282.2, 67.9±8.4 gN ha<sup>-1</sup> day<sup>-1</sup>, respectively. For the sandy soil, these fluxes were 166.6±377.7, 23.7±34.7, 491.2±819.9 and 13.3±7.8 gN ha<sup>-1</sup> day<sup>-1</sup>, respectively. The results show that the models did not calculate the same magnitude of the measured values. The DeNi model overestimated and the DNDC and CoupModel underestimated the measured fluxes. However, in some cases the temporal patterns of the measured and the modeled emission were similar. Most cases of over- or underestimations by the models could be explained by certain deficiencies of the models or of the experimental data.</p>