Modeling the impact of nitrate fertilizers on groundwater quality in the southern part of the Nile Delta, Egypt

The Nile Delta is a vast leaky aquifer that directly connects with the Nile River and is bounded by its two branches (Mabrouk et al. 2013). Consequently, the groundwater quality in the Nile Delta is strongly affected by the changes in the Nile River flows (Dawoud 2004). The horizontal expansion of lands under reclamation for increasing the agricultural area in Nile Delta, with extensive groundwater extraction, has resulted in the increasing pollution and deterioration of groundwater resources (Abdel-Shafy & Aly 2002). The use of fertilizers is now becoming a common practice in agriculture in Egypt, while both consumption and application rates have been increased significantly (FAO 2005). Excessive use of nitrogen fertilizers is likely to be responsible for increasing nitrate concentrations in the shallow groundwater system. Thus, appropriate water and nutrient management is required to minimize groundwater pollution and to maximize nutrient use efficiency (Shekofteh et al. 2013). Groundwater contamination under agricultural catchments occurs when nitrogen rich fertilizer application as urea exceeds the plant demand and the denitrification capacity of the soil, leading to nitrogen leaching to the groundwater system, usually in the form of nitrate, which is highly mobile with little sorption (Birkinshaw & Ewen 2000).

The Nile Delta and Valley aquifers have been studied by RIGW/IWACO (1999) and results showed that Egypt faces imminent threats in addition to the urgent need for a comprehensive management plan for drought mitigation, based on limiting extraction rates. The annual extraction rates that were reported as follows; 3.02 × 10⁹ m³ yr⁻¹, 3.5 × 10⁹ m³ yr⁻¹, and 4.6 × 10⁹ m³ yr⁻¹ in years 2000, 2003, and 2010 respectively (RIGW 2010). The highest ion concentrations that have been detected in a groundwater aquifer in the Nile Delta were ⁠, ⁠, K⁺, and (Al-Agha et al. 2015). In addition, the vulnerability of water resources to pollution was mainly related to the excess use of fertilizers in agricultural activities. Therefore, it is important to carry out the groundwater vulnerability evaluation for the management of groundwater resources (Bai et al. 2013, 2016). On the other hand, the impact of local hydrogeological conditions and human activity on water resources in the south east of the Nile Delta has been studied by El-Fakharany & Mansour (2009), it was found that groundwater quality issues are related to these activities.

Agriculture is the main activity in the study area, and the land under cultivation is in increase that equals 326,260 feddans, which is estimated at about 70% of the total Governorate area (Mohamed 2004). Therefore, the problem of increasing nitrogen fertilizer application is becoming worse as more nitrate is leached to the groundwater table, causing the deterioration of the Nile Delta aquifer. The objective of this study is to investigate the effect of using urea fertilizer in agricultural lands in El Menoufia Governorate on the groundwater contamination with nitrate for the central southern part of the Nile Delta. The groundwater modeling system (GMS) package, including MODFLOW and MT3D, was used in the model simulations for groundwater flow and nitrate transportation respectively. The model calibration for the groundwater heads and observed nitrate concentrations at 40 m depth in year 2014 was performed. Then, the current fertigation strategy which applied with the flood irrigation method was assessed for better groundwater management in the aquifer.

El-Menoufia Governorate is located in the central southern part of the Nile Delta of Egypt. It extends between latitude 30 ° 20′–30° 40′ N and longitude 30 ° 50′–31 ° 15′E. The governorate, whose total area is about 2,543.82 km², is bounded on the east by Damietta branch and on the west by Rosetta branch. The governorate extends to El-Gharbiya in the north and Cairo in the south. The governorate consists of nine administrative districts, all of them are cultivated land and densely populated, covered by an extensive irrigation network consisting of canals and drains. The average ground elevation slope is 10 cm/km, and gradually decreases from 12 m to 10 m above mean sea level (MSL) from south to north.

The Quaternary and Pliocene deposits that constitute the main water-bearing formations of the Nile Delta in the study area comprise the aquifer system, which is intercalated by semi-pervious clay and silty layers. The Nile silt aquitard, which belongs to the Holocene age, caps the main aquifer that belongs to the Pleistocene, which consists of sand, gravel and rock fragments representing the groundwater aquifer itself. The formations are underlain by an impermeable base of Pliocene clay deposits that act as an aquiclude (Schlumberger 1984).

In El-Menoufia, the agricultural practices are the main activity of the people. Crops are cultivated in a mixed pattern following two or three crop rotations per year. The main crops consists of wheat and clover in the winter, starting from October to November, and maize and cotton in the summer starting from May to June (Bader 2004). Irrigation is mainly by the traditional flooding method, 0.25 m submergence of agricultural land by irrigation water, which occurs frequently, three and two times per month in the summer and winter, respectively (Kruseman & Vullings 2007).

The nitrogen fertilizer is the major chemical used in the study area, where urea is the most common type, which is supplied to farmers through the Agriculture Directorate, El-Menoufia, Ministry of Agricultural and Land Reclamation). The application rate of nitrogen fertilizer is recorded by Agricultural cooperative societies, and from field data as follows: 270 Kg/ha and 187 Kg/ha in summer and winter respectively, and the duration of application is from July to August and from December to January in summer and winter respectively (Mohamed 2004). The time and method of urea application depends on the crop type (see for instance FAO 2005).

The GMS was chosen to build the required simulation models used in this study. MODFLOW is a modular, three dimensional, finite difference groundwater flow model which was coupled with the MT3D model for nitrate transport simulation. GMS integrates these two models by coupling, conceptualizing, running the simulation codes, and visualizing the simulation results.

Due to the presence of the lower Pliocene clay layer, which represents the impermeable boundary of the aquifer, only the upper 200 m portion was modelled, as shown in the hydrogeological section in Figure 1(a). The upper layer belongs to the Holocene and consists of silt and clay with variable thickness from 5 m to 20 m (Fadlelmawla & Dawoud 2006). The conceptual model approach was used to construct the MODFLOW simulation, then layers parameters were defined as coverage. Subsequently, the conceptual model is converted to the grid model with elevation data interpolation to MODFLOW and MT3D.

The aquifer is dominated by unconsolidated coarse sand and gravel with occasional clay lenses. Therefore, the aquifer is subdivided in the vertical direction in to four layers: the upper is a silty clay cap with a thickness varying between 5 m to 20 m; the second is fine sand with a depth ranging from 10 m to 25 m; the third is a coarse sand layer with an average thickness of 30 m; and the fourth layer is sand and gravel, which is assumed to be deeply extended to the impermeable layer. For the boundary condition, the two branches of the River Nile were assigned as constant heads. The irrigation network was assigned using river and drain packages, where the infiltration of the groundwater is controlled by the bed conductance. The southern boundary of the study area is parallel to the head gradient of the groundwater; therefore, it was set as a flux boundary with a variable inflow rate.

The same grid configuration used in MODFLOW was also used in the nitrate transport modeling (MT3D). The calibration of the MT3D model parameters was accomplished by matching the measured nitrate concentrations at a level 40 m below the ground surface with those calculated at the end of the calibration period from 1990 to 2014. The longitudinal dispersivity was assigned a constant value of 120 m, and the ratios of transverse and vertical to longitudinal dispersivity adopted were 0.05 and 0.005 respectively. The molecular diffusion, D₀, for nitrate was adopted after (Daniel & Shackelford 1988), equal 1.6416 × 10⁻¹⁰ m²/d.

According to Lowrance (1992), more than 95% of the leached amount of nitrogen from the root zone is assumed to be transformed to the form at the water table, with low concentrations of nitrite and ammonium. Goderya et al. (1996) reported that approximately 30% to 50% of this transformed amount of leaches to the groundwater. From the applied dose of urea and the estimated plant uptake, which equals 142 and 146 kg N/ha in summer and winter, the calculated loadings to groundwater were 128 and 41 kg N/ha in summer and winter respectively, which equals about 35% of the annual fertilizer application.

A preliminary sensitivity analysis, as an inherent part of model calibration, was carried out to identify the most sensitive parameters that affect model results. Then, MODFLOW steady state calibration for the hydraulic conductivities, recharge rates, and the conductance was performed based on the groundwater spatial distribution map after RIGW (1990). The correlation–based measures including the correlation coefficient (r), coefficient of efficiency (E), and index of agreement (d), besides the error–based measures, including the root mean squared error, the mean absolute error, and the mean relative error, have been used to evaluate the goodness–of–fit of the model.

In our study, the nitrate concentrations in recharge water were estimated from dividing the leaching mass by the recharge volume during the fertilizer application period. Nitrate losses by denitrification in the unsaturated zone are about 6 to 9 kg/ha annually (Lowrance 1992), which is considered relatively small compared to the microbiological denitrification that affects nitrate movement in the saturated layer under the anaerobic conditions. MT3D calibration began with the average initial concentration in the Nile Delta aquifer, which in 1990 was 0.4 mg/l. Besides, the denitrification was represented as a first order reaction with a rate constant λ that was firstly estimated using the half life time of ⁠, which was assumed to be equal to 2.5 years (Korom 1992). The field measured concentrations at 40 m depth from the ground surface have been obtained from eight observation wells (as shown in Table 1) after RIGW records in 2014.

It can be seen from Figure 8 that all points are within the ±5% error bounds, and the R² indicates a good agreement between the MODFLOW results and the groundwater levels presented in the hydrogeological map (Figure 3). From the primary sensitivity analysis, it was concluded that any change in the recharge rate either from the two branches of the Nile or the excessive irrigation water will significantly affect the model results rather than hydraulic conductivity or the bed conductance of irrigation canals and drains.

The groundwater in the Nile Delta aquifer is in steady state equilibrium, flowing towards the northwest. The most significant recharge of the aquifer is the Damietta Branch, then the main irrigation canals and excessive irrigation water from the agriculture fields. The vulnerability of groundwater in aquifer depends on the existence or absence of the silty-clay layer and, where applicable, on its thickness. In addition, the change of discharge rates and operation strategies for production wells affect the nitrate distributions and concentrations in the aquifer. The excessive practices of the unofficial extraction from the groundwater aquifer by the hand pumps at lower elevations increase the susceptibility of the aquifer to the higher degree of pollution and deterioration. In order to maintain the acceptable range of nitrate concentrations, it is recommended that the existing wells operate according to regulated pumping rates and discharge. The adjustment of the applied urea dose according to plant requirements for higher crop productions for the minimum leaching amount to the water table is highly recommended. Finally, chemical analyses of groundwater samples should be taken periodically, including the analysis of ⁠, K₂SO₄, (NH₄)₃PO₄, and total dissolved solids (TDS) in particular for agricultural lands.

The first author would like to thank the Egyptian Ministry of Higher Education (MOHE) for providing him with the financial support (PhD scholarship) for this research, also the Egypt-Japan University of Science and Technology (E-JUST) and JICA for offering the facilities and tools needed to conduct this, as well as the Tokyo Institute of Technology.