Modelling soil nutrient dynamics under alternative farm management practices in the Northern Highlands of Ethiopia

Agricultural production in the Northern Highlands of Ethiopia is low, stagnant or unsustainable. The objectives of this study were to explore long-term dynamics of soil organic carbon (C), nitrogen (N) and phosphorus (P) and the consequences for crop-available N and P to support the design of sustainable farm management practices for higher yields and improved livelihoods in the Northern Highlands of Ethiopia. Simplified soil N and P dynamics modules are described. C dynamics have been linked to the dynamics of organic N via the C:N ratio. The model has been calibrated on the basis of empirical data from the study area. The N and OC modules have been validated on the basis of an empirical data set for fields continuously cultivated for 7–53 years in smallholder farms in the Highlands of Ethiopia. The model has been applied for exploration of long-term dynamics of soil N, OC and P and crop available N and P under alternative farm management regimes. The simulation results indicate that, in terms of soil OC, the control management results in 44, 42 and 38% depletion, respectively, in Cambisols, Luvisols and Leptosols; current management practice (Alt1) results in 16% reduction in Cambisols, 32% in Luvisols, but a 22% increase in Leptosols; Alt2 (returning all non-economic organic material to the soil) results in 27% reduction in Luvisols, whereas in Cambisols and Leptosols it increases by 1 and 57%, respectively, after 50 years of cultivation. The rates of changes in soil N are similar to those in OC under current management, Alt1 and Alt2. In terms of total soil P, the control management and Alt2 result in 46 and 43% depletion in Cambisols and 53 and 52% in Luvisols over the 50 years. On the other hand, Alt1 results in build-up of total soil P in Cambisols (69% higher after 50 years), but still to depletion (8%) in Luvisols. All other management regimes are not ‘sustainable’ in terms of soil N, OC and P, and lead to ‘soil mining’. Finally, the model has been used to estimate the required organic amendments and inorganic P inputs to maintain the current status of soil OC and P, as a benchmark of management practices. To maintain the current status of soil OC, the required composted organic amendments were 5.3, 15.0 and 2.1 Mg ha−1 annually for Cambisols, Luvisols and Leptosols, respectively. To maintain the current soil P-levels, required inorganic P-doses (in addition to organic P contributions from composted organic fertilizer from 5.3 (in Cambisols) and 15.0 (in Luvisols) Mg ha−1 year−1) were 8 kg ha−1 year−1 in Cambisols and 23 kg in Luvisols. The model is relatively easy to parameterize for specific situations and reproduces the most important aspects of soil nutrient dynamics. The modelling approach developed in this study can support the design of appropriate soil N, OC and P management practices that eventually should lead to higher yields and improved livelihoods.