Modelling water and nitrogen dynamics at three different spatial scales – influence of different data aggregation levels on simulation results

The effect of different data qualities (different data aggregation levels) on the results of an integrated simulation model on soil water and nitrogen dynamics is investigated for an area of 3 communal districts which were part of a former large scale study on nitrogen losses. Disaggregated data from this regional scale study are compared to the real situation within the regarded area. Results of groundwater recharge and nitrogen leaching indicate that the aggregation of soil map units has a relative small effect compared to the averaging of weather data and to errors of land use. The use of average data derived from long term weather observations lead to a clear underestimation of groundwater recharge. Nitrogen uptake by plants in the regional study is often less than simulated with the real weather data and the observed rotation. Both result in higher nitrogen losses by leaching in the regional study although the water flux is less.     

Response of wheat to nitrogen fertilization,a data set to validate simulation models for nitrogen dynamics in crop and soil

A data set originating from winter wheat experiments at three locations during two years is described. The purpose is to provide sufficient data for testing simulation models for soil nitrogen dynamics, crop growth and nitrogen uptake. Each experiment comprised three different nitrogen treatments, and observations were made at intervals of two or three weeks. The observations included measurements of soil mineral nitrogen content, soil water content, groundwater table, dry matter production and dry matter distribution, nitrogen uptake, nitrogen distribution and root length density.     

Development of nitrogen fertilizer recommendations for arable crops in the Netherlands in relation to nitrate leaching

In the Netherlands, current nitrogen fertilizer recommendations for arable crops are based on the amount of soil mineral nitrogen in early spring. The larger the amount of soil mineral nitrogen, the lower the recommended application rate of fertilizer nitrogen. A more refined method is to draw up a balance sheet in which the nitrogen requirement of the crop is given on the one side and the contributions of fertilizer nitrogen, soil mineral nitrogen, and the amount of nitrogen mineralized during the growing period on the other. The most refined method of nitrogen fertilizer recommendation is the use of a simulation model that predicts the daily crop nitrogen requirement and nitrogen supply to the crop from various pools during the growing period. A simulation model thus adds the time element to nitrogen fertilizer recommendations. Moreover, in contrast with the other two methods, a simulation model allows identification of environmental side-effects of nitrogen fertilizer application.The current Dutch nitrogen fertilizer recommendations aim at predicting the economically optimum application rate of fertilizer nitrogen. From the environmental point of view it is interesting to know how much soil mineral nitrogen has accumulated in the soil at harvest, because this nitrogen is a potential loss to the environment through nitrate leaching during the subsequent winter period. If the economically optimum application rate of fertilizer nitrogen is applied to arable crops, it is unlikely that soil mineral nitrogen accumulates, except in the case of potatoes. Model calculations have shown that accumulation of soil mineral nitrogen after potatoes can be prevented when the recommended nitrogen application rate is reduced by 25%. In that case tuber yield is reduced by only 2%.     

Some results of nitrogen simulations with the model ANIMO

Simulation of the nitrogen behavior in the soil and the nitrogen uptake by winter wheat was performed using the model ANIMO. As input for the model ANIMO simulations of the hydrological conditions in the soil crop ecosystem were executed with the model SWATRE. Compared with measured data the simulation of nitrogen uptake by the crop was satisfactory. The simulation of mineral nitrogen in the soil agreed reasonably well with measured data for one of the experiments used for the analysis. The agreement was less for experiments with additional fertilizer applications in May and June.     

Simulation of soil nitrogen dynamics using the SOILN model

A model dealing with transport and transformations of nitrogen in soil is briefly described. The model has a one-dimensional layered structure and considers processes such as plant uptake, mineralization/immobilization, leaching and denitrification. A soil water and heat model provides daily values for abiotic conditions, which are used as driving variables in the nitrogen simulation. In this study, the model was run with data from a polder-soil area in the Netherlands, with winter wheat as the crop. The simulation results showed that if a measured time course of crop nitrogen uptake throughout the growing season is available, mineral-N dynamics in soil can be satisfactorily described with this model. The main problems identified in the simulations were related to the partitioning between above- and below-ground plant-N, and supplying the crop with sufficient N, as given by the measurements.     

Comparison of methods for the estimation of inert carbon suitable for initialisation of the CANDY model

Almost all soil organic carbon turnover models rely on a partitioning of total organic carbon into an inert and a decomposable pool. The quantification of these pools has a large impact on modelling results. In this study several methods to estimate inert carbon in soils, based either on total soil organic matter or physical protection, were assessed with the objectives of (1) minimising errors in carbon and nitrogen dynamics and (2) ensuring usability for sites with marked differences in site conditions. CANDY simulations were carried out by varying solely the method for calculating the size of the inert carbon pool used to initialise the model. Experimental data from Bad Lauchstädt and Müncheberg were used for the simulation. The data were made available for modellers at a workshop held at Müncheberg (Germany) in 2004. The results concerning not only carbon but also nitrogen dynamics were analysed by applying selected statistical methods. It was shown that even in short-term simulations model initialisation procedure may influence the simulation results considerably. Three methods of estimating inert carbon were identified as being the most appropriate. These methods are either based on soil texture or pore-space classes and therefore account for the physical protection of soil organic matter. Thus, physical protection seems to be of major importance. By extending the scope of the investigation into nitrogen dynamics, additional support for the applicability of a selected method was obtained.     

Modelling nitrogen dynamics in soil–crop systems with HERMES

Model runs with HERMES were performed over entire crop rotation cycles for two experimental sites on loamy and sandy soils in Germany with differently managed plots. The model was able to simulate soil water and nitrogen contents on the sandy plots of Müncheberg with an index of agreement (IA)  >0.8 and  >0.69. Crop growth and N-uptake was simulated well with IA values  >0.89 and  >0.75, respectively. For the loess site in Bad Lauchstädt model results for above-ground biomass, storage organ and N-uptake agreed well with observations over a 4-year rotation with IA values of 0.93, 0.94 and 0.71, respectively. Soil mineral nitrogen was significantly overestimated on the cropped plot (IA = 0.45) as well as on the black fallow plot (IA = 0.65) using the default initialization of the decomposable nitrogen pools from the organic matter content. Equilibration of the pools, using data from a neighbouring long term experiment, improved soil mineral nitrogen simulation to an IA of 0.72 for the cropped and 0.91 for the fallow plot. The long term model performance was investigated using data from 1903 to 2002 of four differently managed plots in Bad Lauchstädt. Soil organic carbon, derived from simulated nitrogen pools, showed acceptable results for the unfertilized plot, but a distinct underestimation on plots with farmyard manure application. Simulated historical winter wheat and potato yields were distinctly overestimated during the initial 50 years. Therefore, an adoption of crop parameters for older varieties is necessary. The index of agreement of 0.9 indicates that the annual weather impact on yield fluctuations was correctly reflected.     

Modelling water flow,nitrogen uptake and production for wheat

Soil water and temperature conditions were simulated for three years at three sites in the Netherlands, using a model named SOIL. Observed water table depths from one site with a sandy loam soil indicated bypass flow in macropores. Nitrogen turnover was simulated using the output of SOIL as input to a nitrogen model. To improve the nitrogen model, a crop-growth submodel was introduced, and simulations were compared with measured data for two seasons and three fertilizer treatments at the three sites. Mineral-N in the soil after application of fertilizer was substantially higher in the simulation than indicated by measurements in 4 out of 18 simulations. Regression analyses showed that simulated mineral-N content in the uppermost metre explained 64% of the observed variation. The corresponding values for nitrogen content (N _ta) and biomass (W _ta) of aboveground tissues were 86 and 93%, respectively. With a few exceptions annual values ofW _ta andN _ta were simulated with an accuracy of approximately 20%. A sensitivity test showed that growth parameters and especially the light use efficiency parameter strongly influenced biomass production for fertilized treatments whereas the control of nitrogen uptake from soil was most important for non-fertilized treatments.     

An easy pot incubation method for measuring nitrogen mineralization from easily decomposable organic material under well defined conditions

A pot incubation method for measuring mineralization dynamics from fresh plant material was tested. The aim was to develop a method which under well-defined conditions could produce mineralization data suited for estimating model input parameters for nitrogen prediction models. The results showed that the water tension of the soil could be controlled easily and precisely by diffusion through porous ceramic cups, and that nitrogen mineralization or immobilization could be measured already after 15 days at 15°C. The results showed that for the incubated catch crop residues carbon, nitrogen and nitrate-N contents were the most important factors determining mineralization. No significant effects ould be ascribed to other parameters measured.     

Simulating water and nitrogen behaviour in soils cropped with winter wheat

The SWATNIT model [26], predicting water and nitrogen transport in cropped soils, was evaluated on experimental data of winter wheat for different N treatments. The experiments were monitored at three different locations on different soil types in the Netherlands. Crop growth was simulated using the SUCROS model [11] which was integrated in the SWATNIT model. Both water and nitrogen stress were incorporated. Except for the soil hydraulic properties, all model parameters were taken from literature. The model performance was evaluated on its capability to predict soil moisture profiles, nitrate and ammonia profiles, the time course of simulated total dry matter production and LAI; and crop N-uptake. Results for the simulations of the soil moisture profile indicate that the soil hydraulic properties did not reflect the actual physical behaviour of the soil with respect to soil moisture. Good agreement is found between the measured and simulated nitrate and ammonia profiles. The simulation of the nitrate content of the top layer at Bouwing was improved by increasing the NH ₄ ⁺ -N-distribution coefficient thereby improving the simulation of the NH ₄ ⁺ -N-content in this layer. Deviations between simulated and measured nitrate concentrations also occurred in the bottom layers (60–100 cm) of the soil profile. The phreatic ground water might influence the nitrate concentrations in the bottom layers. Concerning crop growth modelling, improvements are needed with respect to the partitioning of total dry matter production over the different plant organs in function of the stress, the calculation of the nitrogen stress and the total nitrogen uptake of the crop through a better estimate of the N-demand of the different plant organs.     

Simulation of the nitrogen balance in the soil and a winter wheat crop

A simulation model for winter wheat growth, crop nitrogen dynamics and soil nitrogen supply was tested against experimental data. When simulations of dry matter production agreed with measurements, nitrogen uptake was simulated accurately. The total amount of soil mineral nitrogen as well as the distribution of mineral nitrogen over the various soil layers were generally simulated well, except for experiments in which fertilizer was applied late in spring. In these experiments, applied nitrogen disappeared because it could not be accounted for by the model. Some explanations for this disappearance are briefly discussed.     

The impact of crop growth sub-model choice on simulated water and nitrogen balances

It is the aim of this study to analyse how different crop growth model routines affect the simulation of water flow and nitrogen transport of a crop rotation in agricultural fields. The model system Expert-N is briefly described and used to test the crop growth sub-models against data of a six-year field experiment on sandy soils. Expert-N is a modular soil–plant–atmosphere model system, which comprises different sub-models to simulate one-dimensional vertical transport of water, solute and heat in the unsaturated zone. It includes several sub-models to describe organic matter turnover and has three generic crop growth sub-models. The latter are derived from the crop models CERES, SPASS and SUCROS. Simulations were performed using the different sub-models for each of the cereal crops in the sugar beet, winter wheat, winter barley, winter rye crop rotation. Results show the impact of crop model choice on simulated water balances and turnover of C and N. It is concluded that the simulation of root growth and plant residue mineralisation needs some improvement.     

Nitrate leaching and soil moisture prediction with the LEACHM model

The LEACHM model developed by Wagenet and Hutson [1989] was used to predict the mineral nitrogen and water content in the soil under a winter wheat crop from February to April in two years and three locations. The model grossly overestimated soil water content, probably due to the bad fitting of the assumed water retentivity function to the experimental data at high water contents, and to the presence of a relatively shallow water table (1.0–1.5m). Measured soil hydraulic conductivity varied with water content in a different manner than predicted by the model. By assuming a sandy or gravelly soil layer between the bottom of the measured soil profile and the water table, prediction of soil water content improved considerably. Simulation showed that, under the experimental conditions studied, soil mineral nitrogen varied mainly due to the fertilizer additions, mineralization and denitrification. Nitrogen uptake by plants and leaching were small. Low values of nitrate leaching were predicted by the model because of low drainage. Large differences between predicted and observed values in the mineral nitrogen in the soil occurred in some cases, both in the total amount and its profile distribution.     

Effect of nitrogen,phosphorus and soil and crop residues management practices on maize (Zea mays L.) yield in ultisol of eastern Cameroon

The shortening of fallow period in several areas in tropical Africa has reduced soil fertility and exposed soils to erosion and run-off. Fertilizer application and crop conservation practices are needeed to sustain high crop yield and to conserve the natural resource base for upland crop production in the continent. Field trials were carried out to evaluate the effect of fertilizer application and soil and crop residues management practices on yield of maize (Zea mays L.) planted on a Plinthudult soil at Bertoua, Eastern Cameroon. Maize yields increased significantly with nitrogen and phosphorus fertilizer application. Under the rainfall pattern prevailing in the area, the amount of nitrogen required for maximum yield was higher in the second season. On the other hand, the amount of phosphorus required for maximum yield appeared to decrease with time. The burning of crop residues and weeds prior to planting together with no-till practive gave higher yield of maize than other soil and crop residues management practices.     

Simulating phosphorus responses in annual crops using APSIM: model evaluation on contrasting soil types

Crop simulation models have been used successfully to evaluate many systems and the impact of change on these systems, e.g. for climatic risk and the use of alternative management options, including the use of nitrogen fertilisers. However, for low input systems in tropical and subtropical regions where organic inputs rather than fertilisers are the predominant nutrient management option and other nutrients besides nitrogen (particular phosphorus) constrain crop growth, these models are not up to the task. This paper describes progress towards developing a capability to simulate response to phosphorus (P) within the APSIM (Agricultural Production Systems Simulator) framework. It reports the development of the P routines based on maize crops grown in semi-arid eastern Kenya, and validation in contrasting soils in western Kenya and South-western Colombia to demonstrate the robustness of the routines. The creation of this capability required: (1) a new module (APSIM SoilP) that simulates the dynamics of P in soil and is able to account for effectiveness of alternative fertiliser management (i.e. water-soluble versus rock phosphate sources, placement effects); (2) a link to the modules simulating the dynamics of carbon and nitrogen in soil organic matter, crop residues, etc., in order that the P present in such materials can be accounted for; and (3) modification to crop modules to represent the P uptake process, estimation of the P stress in the crop, and consequent restrictions to the plant growth processes of photosynthesis, leaf expansion, phenology and grain filling. Modelling results show that the P routines in APSIM can be specified to produce output that matches multi-season rotations of different crops, on a contrasting soil type to previous evaluations, with very few changes to the parameterization files. Model performance in predicting the growth of maize and bean crops grown in rotation on an Andisol with different sources and rates of P was good (75–87% of variance could be explained). This is the first published example of extending APSIM P routines to another crop (beans) from maize. Dr R. J. Delve has recently left CIAT and joined Catholic Relief Services, Kenya.     

Phosphorus content in soil, uptake by plants and balance in three European long-term field experiments

The fate of phosphorus (P) derived from mineral fertilisers and organic manures, and the effective P balance, have been assessed in three long-term field experiments at Rothamsted (UK), Bad Lauchstaedt (Germany) and Skierniewice (Poland). This paper discusses the plant availability, uptake and overall utilisation of P over the last 30 years, based on soil test P availability indices and crop analyses determined by the standard methods used in each of the three countries. The data suggest that differences in soil type significantly influence the dynamics of P at the three locations, but most significantly between a loess Chernozem at Bad Lauchstaedt with a high organic matter content and the soils at the other two locations which have a low organic matter content. The application of P either as inorganic fertiliser or organic manure had a considerable influence on the availablity, uptake, leaching or fixing of P, but the crop recovery rate of P from mineral fertiliser did not exceed 35% with the smallest recovery (average 18%) occurring in the soil with the highest clay content at Rothamsted. At Bad Lauchstaedt and Rothamsted the most efficient utilisation of P (averages of 47% and 37%, respectively) was from soils treated with farmyard manure (FYM), with the greater quantity of P either leached or fixed (8 and 25 kg ha^-1 y^-1, respectively) occurring in soils treated with superphosphate. At Skierniewice, however, the reverse was true. Overall, the most efficient crop utilisation from mineral P (30% average) was from the loamy sand at Skierniewice. P balances for the three locations show that quantitatively, for the same P input, the amount of P either leached from or fixed in the plough layer of Broadbalk field, Rothamsted, was 2–3 times greater than at Skierniewice and 3–6 times greater than at Bad Lauchstaedt. The results suggest that differences in the soil physico-chemical properties, climate, the availability of other major nutrients, and the form in which P is applied, all influence the effectiveness of P fertilisation and P balance. The investigation highlights the importance of maintaining long-term field experiments and archived soil and crop samples on a world-wide basis for understanding nutrient cycling and fertility dynamics.     

Validation of Submodels for CFD Simulation of n‐Hexane Pool Flames including Interferometry

Computational fluid dynamics simulation is used to predict transient and time‐averaged flame temperatures and species concentrations of an n‐hexane pool flame. Employing a combination of an assumed probability density function approach with laminar flamelets using detailed kinetic data and large eddy simulation with Smagorinsky submodel is shown to be a promising way in modeling pool and tank fires. The measured species concentration and flame temperature profiles from gas chromatography, thermocouple measurements and holographic interferometry are used to validate the submodels for CFD simulation of pool flames.