ارزیابی و مقایسه دو مدل کودآبیاری جویچه‌‌ای در تخمین غلظت نیترات رواناب

Surface fertigation in furrow irrigation has potential to decrease water and fertilizer losses. Fertigation is as a technique having more advantages than conventional fertilizer application methods such as flexibility, potential for improving fertilizer distribution uniformity and high application efficiency. Due to time restriction and experimental field implementation cost, fertigation simulation models have been developed to achieve the desired management conditions. Research has been performed in the last decades to out for simulate surface water and solute flow under surface fertigation. These researches reported that simulation could be effectively applied to improve surface fertigation design and management, reducing water and fertilizer losses. Playan and Faci (1997) stated that a short duration of the fertilizer injection resulted in low fertilizer distribution uniformity in border fertigation. However, Sabill?n and Merkley (2004) advocated relatively short injection times and relatively high injection rates in furrow fertigation. Abbasi et al. (2003) conducted a blocked-end furrow fertigation experiment and simulated overland water flow and solute (bromide) transport. These authors reported that high solute uniformity was obtained when the solute was applied during the entire irrigation event or during the second half of the irrigation event. Burguete et al. (2009) presented a fertigation model, named SURCOS, for furrow (and level furrow) irrigation. Model simulations succeeded in predicting fertilizer concentration in irrigation water at different times and distances along the furrow for different fertilizer application strategies. The objective of this study was to simulate nitrate concentration in runoff in conventional, variable alternate and fixed alternate furrow irrigation regimes by SURCOS fertigation model and Abbasi et al. model and comparison of these two models. For this purpose, experimental field data were used in the two fertigation events during maize growing season.
A field experiment was conducted in the experimental station of College of Agriculture and Natural Resources, University of Tehran, Karaj, Iran. This zone has a semi-Mediterranean climate with an average annual rainfall of 265 mm and an average annual temperature of 16 ?C. Three furrow irrigation methods; conventional, variable alternate, and fixed alternate furrow irrigation (CFI, AFI and FFI, respectively) were investigated in this work. Inflow and outflow (runoff) rates were measured by WSC flume installed at the inlet and outlet of the experimental furrows. The experimental furrows were marked for measuring the water advance times. The average basic infiltration rate (f_0) was determined by the inflow-outflow method (Elliot and Walker 1987). The two-point method (Elliott and Walker 1982) was also used to determine the Kostiakov-Lewis parameters, coefficients a and k. The Manning’s n was assumed to be 0.04. Two different numerical models (Abbasi and SURCOS) were used and assessed for the simulation of the surface fertigation which both are capable of simulating overland water flow and solute transport. The governing equations for water flow were solved in the form of a zero-inertia model of the Saint-Venant’s equations. Also, Solute transport was modeled with the advection-dispersion equation.
Both models were run for the first and second fertigation events regarding the input data. The runoff nitrate concentrations were predicted and compared with the measured data. The predicted nitrate concentrations were almost constant whereas the measured values were relatively fluctuating. These variations could be related to spatial variability in soil infiltration and roughness properties and small changes in water and fertilizer rates in the field. Having high solubility and mobility, nitrate could easily move with water. There were good agreements between the measured and predicted nitrate concentrations. The R2, RMSE and NRMSE values for simulation of nitrate concentration in runoff were obtained 0.989, 13.5 mg/L, 4.3 % for the Abbasi et al. model and 0.913, 17 mg/L, 5.45 % for the SURCOS model, respectively. The results indicated that both models were able to successfully simulate nitrate concentration in runoff.
Although the water and fertilizer movements in furrow irrigation are bi-dimensional, both 1D surface fertigation models could successfully simulate the fertigation process. However, the Abbasi model had relatively better performance than the SURCOS model due to the less error in predicting nitrate concentration while the SURCOS model could better distinguished differences of the irrigation treatments. The results indicated that the simulation models can be used to get better design and management of the fertigation in alternate and conventional furrow irrigation for alleviating the water crisis and environmental risks