Modelling P Dynamics in soil - Decomposition and Sorption: Concepts and User Manual

Summary in the original language of the document

Weather-driven simulation modelling has become an important component of studies of soil nu-trients, both for crop growth and for losses by leaching to the environment. In order to model phosphorus (P) dynamics in soil, the mobilisation and immobilisation processes of inorganic and organic P species is important for quantifying P leaching from the unsaturated zone. The mobilization and immobilisation processes of P are represented in this project by sorption and desorption of P and by decomposition of organic matter with mineralization and immobilisation of P.
The code/model presented here, referred to as the P-Model, works with another model, called Daisy. The Daisy code delivers input data of water content, water flux, temperature, inorganic nitrogen (N), crop uptake of N, and crop root exudates, to the P-Model. The P-Model must be considered as a prototype of a P module in Daisy, as the user-friendliness is limited and not all aspects are described in details. For instance crops are described as very simple sink terms, in spite of the fact that they are very important in the P dynamic of soils. Soil ploughing and harrowing are not included in the P-Model and therefore some limitations of the Daisy set up must be considered.
The concept of the turnover of soil organic matter is based on the organic matter module in Daisy. The simulation of the organic matter balance and the nitrogen dynamics is strongly inter-connected in Daisy as the organic matter model is considered an integral part of the overall ni-trogen balance model. It is assumed that the activity of soil microbial biomass is related to the substrate availability; and that it is reasonable to relate the N mineralization and immobilization with decomposition of soil organic matter (SOM). Daisy does not include P in the organic matter module. However, like N it seems reasonable to relate P mineralization and immobilization to the decomposition of SOM.
In the P-Model organic matter is defined with respect to C, N, and P and is represented by different pools in the model, representing easily degradable and recalcitrant pools, microbial pools, and a dissolved pool. The dissolved organic matter (DOM) pool is produced by microbial degradation of organic matter and by physical/chemical desorption from the most bio available pool of soil organic matter. Hence, DOM may also be immobilized by sorption to the soil organic matter pool. The P-model also includes addition of organic and inorganic fertilizers, crops, and solute transport by convection.
One of the main issues during parameterisation of the DOM dynamics has been to identify and particularly to quantify the sources of DOM in soil. It is assumed that DOM is produced by the soil microbial biomass (SMB) pool and the SOM pool. DOM production is not directly linked to the input of organic material as fresh litter, dead roots, and organic fertilizer, because the contri-bution of DOM from all these different sources probably is very different and not known at this stage. However, theses sources is very bioavailable to the microbial biomass promoting fast turnover and growing microbial biomass, which contribute to the production of DOM. Hence, the DOM dynamic is closely linked to the dynamic of the AOM pools. The biological decomposition of DOM and associated P in the P-Model has been parameterised from literature (e.g. C/P ratios, decay rates and diffusion coefficients of DOM).
The P-Model considers sorption and desorption of DOM in soil, by assuming that dissolved organic P (DOP) and dissolved organic N (DON) follows dissolved organic C (DOC) sorp-tion/desorption in soil. When DOM is sorbed it is assumed to be associated to a part of the SOM pool. Sorption/desorption of DOC takes advantages of estimated relationships linking soil properties to sorption properties across a range of soil types. The Initial Mass approach is used to estimate a DOC concentration specific for the soil, where no net sorption/desorption of DOM occurs. The sorption/desorption dynamics of DOM and associated P in the P-Model have been parameterised from sorption data and field experiments. An empirical, kinetic term couples the ‘equilibrium’ concentration to the actual DOM concentration in the soil, incorporating considera-tions as diffusion and sorption time into the description of the sorption/desorption process. De-pending on whether the DOM concentration is above or below the ‘equilibrium’ concentration then the DOC concentration in the soil is reduced or increased, respectively. Reducing the sorp-tion/desorption rate coefficients the exchange between part of the SOM pool and the DOM pool is reduced. The best fit of simulations to measured DOC and DOP concentrations in batch ex-periments was found by using desorption and sorption rate coefficients of 0.001 hour-1.
Additionally, the P-Model considers sorption of inorganic P, which is described by a three-step mechanism: 1) A fast sorption mechanism, and, 2) a relative slow absorption mechanism, plus 3) a very slow fixation process. Thus, its is hypothesised that inorganic phosphate first binds to easily available sorption sites with high affinity, and then the less available sites which is limited by diffusion and further migration into the soil particle to sorption sites less available. Langmuir describes all three processes at ‘equilibrium’. As for DOM, kinetic terms couples the ‘equilibrium’ concentrations of the different inorganic P pools to the actual P concentration in the soil, incorporating considerations as diffusion and sorption time into the description of the sorption/desorption processes. The sorption of inorganic P has been parameterised from sorption data and literature (e.g. sorption affinity constant and sorption capacities). Due to the concept of instantaneous process of adsorption, which implies an adsorption rate coefficient of 1 h-1 in this model, the other rates have been fitted based on this assumption. The process of absorption has been fitted to obtain equilibrium within 3-4 days. The fixation process has been adjusted not obtain equilibrium during 14 days. The slow desorption rate has not been included during these test. So, whether the slow process is considered reversible or irreversible is still up to discussion in the concept. The distribution of the sorption capacities between the three P pools was shown to have limited effect of the sorption dynamic The distribution of the total sorption capacity between the three sorbed phases has been parameterised based on the assumption that the quickly sorbed fraction has 1/3 of total sorption capacity. By fitting the P-Model to data of sorption experiments on two different soils, it was found that 1/4 of the total sorption capacity is allocated to the absorbed phase for both top and subsoils. The rest of the sorption capacity is allocated to the slow, fixed pool. The sorption affinities are very dominating for the sorption dynamic. The sorption affinities were initially assumed to be in the range of 10-120 mM-1 as of-ten found by fitting Langmuir constants during sorption experiments. However, the best fit to measured sorption data to two subsoils were made by using rather high values for the sorption affinity constants (400-200 mM-1) for the adsorbed, absorbed and fixed processes. The P sorp-tion of the two topsoils behaves very different which could be due to a pH effect. It was not possible to parameterise the affinity constants for the three sorption processes similar for the two top soils, maybe because effects of initial soil pH are not taking into account in the parameteri-sation.
Only part of the P-Model has been validated on field data. The calculation of DOC and DON mobilisation / immobilisation is validated at three fields with different soil treatments located at the Burrehøjvej field at Research Center Foulum in the central part of Jutland. The simulations are compared with measured DOC, DON, NO3 and NH4 concentrations sampled from suctions cups installed in the three fields. The simulations show that in the topsoil the DOM is mainly produced by biological SOM turnover during summertime and chemical/physical release from SOM at all time. The yearly fluctuation of DOM concentration is related to the microbial activity. High fluctuations of DOM which are related to microbial activity are diminished by the chemical/physical sorption/desorption processes which attempts to keep the DOC concentration steady at a certain level. For the soil treated with 9 year of grass clover, the higher DOCand DON fluctuations in the topsoil are slightly diminished in the simulations in relation to the measured DOC and DON concentration. For the two other soil treatments the DOC and DON concentration show less fluctuations and are simulated better.