Simulation modelling has been found to be a useful tool in evaluating soil fertility technologies. In less developed countries resources are not always available to adapt such work to other environmental conditions. The overall goal of this study has been to develop a simulation model that will quantify and explain the dynamics of nitrogen in the agro-forestry systems of sub-humid Zimbabwe and calculate the uptake of N by maize after a legume improved fallow. Specific objectives were formulated as:
(i) to identify and understand the processes which cause changes in mineral N over time in legume-based agroforestry systems, (ii) to develop a dynamic model incorporating the processes and forcing functions, (iii) to simulate and analyse the model against available data and make better estimates of N leaching, (iv) to predict N uptake by maize crop following a two year improved fallow and (v) to derive a plot N balance and make recommendations for N management.
Several processes have been identified as influencing the dynamics of N in the soil and their incorporation into the model has been done in consideration of the available knowledge. Organic matter turnover is modelled by separating into two pools namely: added organic matter from the legume residues and native organic matter. The first pool decomposes using first order kinetics while the latter has a more or less constant rate of mineralization. Soil moisture content influences both rates. Microbial biomass has been assumed to be relatively constant over the season.
Mineralization of nitrogen is assumed to produce nitrate which thus excludes the processes of ammonification and volatilization of ammonia from the model. The C/N ratio of the residues is such that there is net mineralization of nitrogen and since microbial biomass is kept constant, there is no immobilization of N. Simulation of denitrification is through a rate constant. Potential denitrification rate is controlled by oxygen status of the soil which is indirectly expressed via a moisture correction factor.
Rainfall is the main driving variable. Other processes of water balance include infiltration, water uptake by maize roots and vertical movement of water in the profile. The vertical movement of water is modelled using the Richards equation. Uptake depends on root depth and distribution and crop demand. Nitrogen uptake is modelled using Michaelis-Menten kinetics, which is based on the concentration of N in the soil and is modified by root distribution. Tillage is included in the model as a management strategy and has been modelled as a parameter by lowering the bulk density of the soil. Soil temperature is incorporated as a parameter, which can be changed to assess its influence on the various process rates. This approach assumes that temperature is almost optimal and does not limit rates of change. A number of other parameters are used to characterize the system.
Simulation starts at the onset of the growing season. The rate at which OM mineralizes is generally high owing to suitable temperature and moisture conditions. About 150kg N ha-1 is added into the system through legume residues. The C/N ratio was shown to be an important determinant of the N mineralized. Most of the nitrogen is made available when the crop is still developing its root system. Total uptake for the season is therefore relatively low, about 60 kg ha-1, half of which is removed from the system through harvest of grain. Uptake of N by the subsequent crop becomes an important issue which has to be addressed.
The results indicate that leaching is another major form of N loss from the system and amounts to 9 kg ha-1 per season. The amount of N which mineralizes at the onset of rains apparently cannot be taken up by the crop and is left susceptible to leaching. The results show that denitrification accounts for a considerable loss of N in the magnitude of 2 kg ha-1 season-1. The N balance was positive for the duration of the season from incorporation of residues. Further research would need to address possible ways of complimenting N release with crop uptake. The overall performance of the model is satisfactory although some adjustments and proper validation may be necessary to improve its performance. For this, independent experimental data is required.