Water Dynamics Combined with a Supply of NPK Solutes and Urea in a 3-Layer Soil Profile under Drip Irrigation

The intensive and inappropriate use of water, fertilizers and phytosanitary products is sources of water and soil pollution. It is thus necessary to improve the management of irrigation water in order to optimize its use and productivity, especially in regions where water resources are becoming increasingly scarce. The water flow and non-reactive solutes’ transport simulation under drip irrigation were carried out in a 3-layered soil profile distributed from top to bottom i.e., sandy, sandy-silty, silty-sandy-clay. The aim of this study was thus, to provide a good practice of water management associated with solutes’ application, in order to retain as much solute as possible in the root zone, which will increase the residence time of the solutes. Three treatments of water flux corresponding to 100% ET c , 75% ET c , 50% ET c , combined with 100 mmol /L/ m 2 of NPK and 246 mmol/L/m 2 of urea applicable in two doses, were carried out over a period of 110 days corresponding to the duration of the cropping cycle for the intermediate variety of maize. The 100% ET c and 75% ET c treatments cause more loss of water and solutes, because of the sandy texture of the soil. However, a 50% ET c water flux would reduce more water loss through drainage, and solutes’ loss due to leaching beyond the root zone, which would increase the residence time of solutes in the soil profile. Application tests of the NPK


Introduction
The search for a continuous increase in agricultural productivity, the standardization of technology, and the intensification of the production led in the 1980s to negative environmental impacts on agro-ecosystems, such as erosion, reduction of biodiversity, water and soil pollution. Intensive and inappropriate use of water, fertilizers and phytosanitary products are sources of water and soil pollution [1] [2] [3]. This water and soil pollution by pesticides has become a major health problem [4]. It is thus necessary to improve the management of irrigation water in order to optimize its use and productivity, especially in regions where water resources are becoming increasingly scarce [5]. Irrigated agriculture, the main objective of which is to improve irrigation water management and increase productivity, will be used. Irrigation techniques can indeed meet the challenge of reasonable management of water resources and fertilizers. Among the existing irrigation methods, drip irrigation systems offer enormous potential as compared to other systems. Indeed, they significantly reduce evaporation, apply water and fertilizers directly to the root zone and greatly reduce loss. Due to these advantages, drip irrigation has become the most accepted method of irrigation/fertigation in order to improve the efficient use of water and nitrogen, as well as minimize nitrate leaching [6]. However, a potential problem associated with drip irrigation is the deep percolation and leaching of nutrients beyond the root zone [7] [8] [9] [10] [11], which could be a source of soil and groundwater pollution. In Burkina Faso, where the lack of water continues to increase each year, the adoption of drip irrigation as an alternative by producers is slowly taking place, despite the enormous benefits that this irrigation system could bring to producers. In order to better support producers, studies therefore need to be carried out, especially on how to manage water and fertilizers for the proper planning of irrigated agriculture with drip irrigation. The main objective of this study was thus, to propose good practices for the application of water associated with fertilizers (NPK; urea) in order to retain as many fertilizers as possible in the root zone as long as possible, for the production of corn under drip irrigation. Doing so, could maximize the residence time of solutes, which could also increase the solute's uptake by the plant. More specifically, our objective was to: 1) study the dynamics of NPK (14-23-14) and urea (46-0-0) in the soil; 2) predict the risks of leaching according to the dose and the time of application; 3) provide a good management of the soil and a good use of the crop. Proper design of drip systems requires knowledge of the wetting front distribution around the dripper. Although several studies have been carried out to investigate the dy-namics of water in soil under drip irrigation (e.g. [12] [13] [14]), it is still difficult to design management strategies for the optimization of the quantity of irrigation water, its frequency, and the location of drippers, in order to achieve the highest water use efficiency [15]. Simulation models have been valuable research tools for studies taking into account the interactive and complex processes of water flow and solutes' transport in the soil, and also the effects of management practices on yields as well as the environment [16] [17]. These simulations can be used to assess the efficiency of irrigation systems over several seasons, and advise producers if this requires improvement in several aspects of the functioning of the irrigation systems [18]. Numerical simulation as a tool to optimize the management of irrigation practices is a fast and an inexpensive approach; it has been used a lot over the past ten years. Among the numerical models used, the Hydrus model, thanks to its flexibility in taking into account different boundary conditions, and the uptake of water and nutrients by the roots of the plant, and an easy-to-access user interface, was much successfully used in several studies (e.g. [7] [10] [15] [19]- [24]) to simulate the movement of water under drip irrigation. In this study, we used Hydrus 1D to simulate water flow and non-reactive solutes' transport in one dimension (vertical axis), in a 3-layer soil, and we did not take into account water and solutes' root uptake.

Process
The studied scenario consisted of an application of urea (46-0-0) and NPK (14-23-14) for the production of maize (intermediate variety 95 -110 days) through a simulated drip irrigation with the Hydrus 1D software. In order to assess deep drainage, and solutes' leaching according to the day of application, the water flux and the solutes' concentration, neither the crop nor the water and solutes' uptake by roots were considered in this simulation. The two solutes applied were considered as tracers, therefore having no interaction with the environment. Also, they were provided in liquid form with the irrigation water. The water supplies were first made at 100% ET c . This, in order to observe possible loss of water and solutes if the irrigation were scheduled daily, and then, to proceed with an irrigation planning when the water is considered as a limiting factor, or to reduce loss. The quantities of solutes supplied were in accordance with those recommended for maize production in Burkina Faso, according to J. Sanou (2004). The scenarios were simulated according to the initial and boundary conditions in order to get closer to the reality

Plot of Land Meshing
We considered a plot of land with an area of 500 m 2 (25 m × 20 m) irrigated by surface drip irrigation system, with a flow rate of 2 l/h (the most frequently used flow under in our tropics). The crop used was maize (intermediate variety) with a production cycle of 110 days. We used the recommendations of the Ministry of Agriculture in terms of inter-crop and inter-row spacing of 40 cm × 80 cm. For the area to be irrigated, there will be 31 lines of pipes and 50 drippers per line, i.e., a total of 1550 drippers (pockets) are given in Figure 1.

Fertilizers Dosage
For semi-intensive agriculture, the quantities of urea (46-0-0) and NPK (14-23-14) to be used for maize production are respectively 150 kg/ha and 200 kg/ha (J. Sanou, 2004). Table 1 gives the quantities of fertilizer and their date of application. For a small village irrigation (500 m 2 ), we determined the concentration of urea and NPK to apply according to the area: • NPK concentration From sowing date until to the 15th day after sowing, the NPK dose to use is 200 kg/ha, i.e., 0.02 kg/m 2 or 20 g/m 2 . For NPK (14-23-14) or 14% of N 2 , we will have 2.8 g/m 2 . The concentration being the number of moles per volume quantity, we considered a liter (1 L) of solution. The molar concentration would therefore be 0.1 mol/L/m 2 or 100 mmol/L/m 2 . This concentration value was considered as input data for the NPK simulation.
• Urea concentration In total, we used 150 kg/ha of urea (46-0-0) in two applications: On the 30 th day after sowing We used 100 kg/ha of urea (46-0-0) or 46 kg/ha or 4.6 g/m 2 of urea. The concentration being the number of moles per quantity of volume, we considered a liter (1 L) of solution. The molar concentration would therefore be 0.164 mol/L/m 2 or 164 mmol/L/m 2 .
On the 45 th day after sowing The quantity envisioned is 50 kg/ha of urea (46-0-0) or 23 kg/ha of urea, which

Field of Study
The study area Figure 2 was a one meter deep soil profile, made up of 3 layers and initially dry. These three layers are distributed from top to bottom in a sandy layer, sandy-silt, and sandy-clay silt. The initial water contents in these different layers were equal to the residual water contents. In this study domain, we placed observation nodes N1, N2, N3, N4, and N5 respectively on the soil surface of 0 cm, 20 cm, 50 cm, 75 cm and 100 cm in depth. The soil properties used for the simulation were determined experimentally thanks to a complete characterization of the physical, morphological and hydrodynamic properties in [26]. We presented the physical parameters of the 3 different layers necessary for the simulation in Table 2. The parameters alpha α and n were derived from particle size distribution and bulk density in [26] with the ROSETTA pedotransfer function. And the other parameters s θ and s K were determined experimentally by [26]. At the beginning of the simulation, the soil did not contain any solute, the irrigation water also did not. The ground surface was subject to an atmospheric boundary condition with meteorological values that were those of Burkina Faso, obtained from the National Meteorological Agency. The surface of the domain was also subjected to variable fluxes of water and solutes in order to take into account the inputs of water and fertilizer. The bottom boundary of the domain was subject to free drainage, and the side walls, to zero flux of water and solutes. The various meteorological data are shown in Table 3. Water supplies were made according to the plant's water   (1).
In which, ET c is the crop water evapotranspiration (mm/day), ET 0 the potential evapotranspiration (mm/day); and K c the dimensionless maize crop coefficient depending on plant growth. In Table 4 we summarize the different inputs of water and solutes during the simulation period.

Water Flow and Solutes' Transport Modeling
The Hydrus 1D software [27] was used for the one-dimensional numerical simulation of water, and solutes' transport in the soil column. The water flow in a variably saturated medium is described by the Richards equation: where z is the vertical coordinate (cm); and h the soil water pressure head (cm); t the time (day); K the hydraulic conductivity function (cm•day −1 ); θ the volume-tric water content (cm 3 •cm −3 ). The hydraulic conductivity function is given by the van-Genuchten (1980) Mualem (1976) relationnships in Equation (3) [28].
In which e S is the effective saturation, r θ and s θ respectively denote the residual and saturated water contents (cm 3 •cm −3 ). K s is the saturated hydraulic conductivity (cm•day −1 ), α (cm −1 ) and n (−) are empirical shape parameters, l is a pore connectivity/tortuosity parameter (−).
The solutes' transport is represented by a convective-dispersive transport equation. Convection accounts for the macroscopic transport of solutes, which accompany the fluid. Diffusion accounts for both molecular diffusion and hydrodynamic dispersion. No nitrogen transformation was considered in the simulation process. Thus, the adsorption, precipitation/dissolution, volatilization of nitrogen were neglected. The equation governing the transport of non-reactive solutes in a homogeneous porous medium is written as: where c is the concentration of the species in the liquid phase (M•L −3 ); D the hydrodynamic dispersion coefficient of the chemical species (L 2 •T −1 ), and q the water flow (L•T −1 ).

Results and Discussion
We presented the results of the simulation for a 110 days duration (duration of the crop cycle) below. The first results obtained were in conformity with the technical route recommended for maize production in Burkina Faso using of fertilizers. Figure 3 and Figure 4 respectively gave the potential pressure head and the water content, versus time at the various observation nodes. These two Figure 3 and Figure 4 show how the water is distributed through the soil, from the surface (node N1: 0 cm) to the bottom of the domain (node N5: −100 cm). Figure 4 and Figure 5 respectively show the concentration of NPK and urea versus time at the different observation nodes. Through Figure 5 and Figure 6, we saw how the solutes infiltrate and then, spread throughout the soil from the application date. By analyzing Figure 3, we noticed that, with a water flux of 0.266 cm/day applied to the surface (node N1) during the first 21 days, the water reached the layers N2, N3, N4, and N5 respectively 1 day, 7 day, 12 day, 17 day after application. Beyond the 40 th day of irrigation, the pressure potential in nodes N4 and N5 was greater than that of nodes N1, N2, N3, which means that the quantity of water leaving the deep layers was greater than that coming from the upper layers. Beyond the 65 th day, the curves of the various observation  nodes were almost parallel, until the end day of simulation. We therefore, saw the water flux reach an equilibrium state in the entire soil profile. We also noticed that, the water flux leaving the soil profile at node N5 was greater than the flux of water infiltrating the surface at node N1, which means a loss of water from the 40 th day until the end day of the simulation (110 th day). Figure 4 showing the water content versus time at the different observation nodes, looks like Figure 3, which confirmed the previous results. The observation of nodes N1, N3 and N5 in Figure 3, clearly shows that the water stock above 50 cm was greater than that in the first 50 cm of the ground.  The NPK (concentration 1) was injected on the 15 th day. As the water seeps into the soil, it carries the solutes with it through the entire soil profile. The concentration of solute available on the surface of the soil gradually decreased until it was completely leached on the 30 th day. The solute, respectively reaches a depth of 20 cm (node N2) and 50 cm (node N3) on the 17 th day and 22 nd day after sowing, which means respectively, 2 days and 7 days after application of the NPK. The NPK continues its distribution and reaches the bottom of the domain (node N5: 100 cm) on the 47 th day. As the water continues to seep into the soil, all of the solute concentration applied to the soil surface will be leached, and thus, returning the soil profile to its initial state of zero concentration. What occurred after the 75 th day, with the solute concentration in all layers being zero, the soil was completely leached. The solute concentration in the different layers of the soil remained low, a peak of around 6 mmol/cm 3 was reached at a depth of 20 cm on the 25 th day. The concentration beyond 20 cm of depth remained lower than the observed peak. In Figure 6, two supplying of urea (concentration 2) were made on the 30 th and 45 th day after sowing, which is observable at node N1: 0 cm. Because of the dispersion, this concentration of solute progressively reaches respectively nodes N3: −50 cm and N4: −75 cm on the 32 nd and 40 th day. The solute concentration not being zero at node N5 at the end of the simulation, we deduced that the solute remains in the soil at the end of the simulation. These analyzes show that, a large part of the water and solutes used, was lost through drainage and leaching beyond the root zone, especially in the early stages of the crop development. As for the volume of water and the quantity of solute stored in the soil, the quantity (NPK in particular) would be unusable by the crops, because they was stored beyond a certain depth (−60 cm) of the non-colonized soil by the roots at the first stage of the crop development. We performed different simulations in order to limit the volume of water, and a certain high concentration of solutes in the first half of the soil, where the root density is higher for the first phase of corn growth between 1 and 21 days. As the root volume is still growing until the maturity of the maize during which time, urea was applied in two doses, the urea could still be absorbed by the roots even beyond 60 cm depth. It was therefore necessary to optimize the application of NPK, as it can be applied from seedling to the 15 th day after sowing, a period included in the initial phase of growth, for which root development is not important.. In order to reduce the water loss by drainage as well as NPK loss, we carried out water supplies at 100% ET c , 75% ET c , and at 50% ET c with new treatments for NPK's inputs and thus found a favorable day in which there would be less loss and longer residence time, as compared to other days of application.

100% ET c and NPK Applied Either on the 1 st , 5 th , 10 th or 15 th Day after Sowing
When the NPK was applied the 1 st day after sowing Figure 7, we can see through the various observation nodes that the solute concentration was watched out from the surface (node N1) of the soil on the 25 th day after application, and the maximum concentration was 60 mmol/cm 3 . The solution continues its distribution and reached the N2 node 2 days after application, with a maximum (6 mmol/cm 3 ) reached on the 20 th day. The solute left the 20 cm depth on the 39 th day. After 9 days, the solute then reached a depth of 50 cm (node N3) for which, peak concentration was 5 mmol/cm 3 , and beyond the 60 th day, it leached over the depth of 50 cm. The N4 node was reached on the 15 th day with a peak concentration of 3.75 mmol/cm 3    disappeared from the soil surface on the 28 th day. The maximum concentration observed at this level was 56 mmol/cm 3 . The solute front reached N2 on the 7 th day with a maximum concentration of 7.75 mmol/cm 3 on the 15 th day. The solute was leached to the depth of 20 cm on the 40 th day. The water continuing its infiltration carried the solute with it to node N3 on day 12, and leaves the depth on the 60 th day. On the 20 th day, the solute reached node N4 and leaves this depth on the 75 th day. Figure 9 (NPK applied on the 10 th day after sowing) on the soil surface, the solute was leached before the 30 th day. The solute continues to progress until it reached node N2 on the 12 th day with a maximum concentration of 6 mmol/cm 3 ; the solute leaves this region of the soil on the 43 rd day. Node N3 was reached on the 17 th day, and the solute was leached from this horizon on the 65 th day. 10 days after reaching node N3, the solute then reaches node N4 in order to flow past N4 by the 90 th day.
The application of NPK on the 15 th day after sowing Figure 10 shows a disappearance of the solute from the soil surface (N1) on the 32 nd day. The solute front continues to distribute and reached N2 on the 17 th day and then vanished on the 46 th day. Nodes N3 and N4 were reached after the 22 nd and 30 th day  respectively. The solute front leaves the N3 and N4 horizon on the 67 th and the 95 th day, respectively.
Note that for 100% ET c , regardless of the period of application of NPK to the soil surface, the residence time of NPK on the soil surface did not exceed 25 days, which was the longest stay in the surface when the solute was applied on the 1 st day after sowing. For each date of application, the solution reached nodes N2 and N3 respectively, 2 days and 7 days after application. On the other hand, the concentration in the various observation nodes decreases as one approaches the 15 th day in order to carry out the solute's intake. All solute in the soil was lost on the 110 th day after sowing, regardless of when the solute was supplied.

Water Flux 75% ET c and NPK Applied Either on the 1 st , 5 th , 10 th or 15 th Day after Sowing
For a treatment of 75% ET c as water supply, we considered the same treatments as above for the solutes' supply. By observing the pressure head curve in the soil at different observation nodes, we noticed that the quantity of water stored in the soil and lost at node N5 was less as compared to the quantity stored and lost for a treatment of 100% ET c . 75% ET c therefore reduced water loss through drainage.
To save space, we did not present the figure giving the pressure head potential. Supply of solute on the 1 st day of simulation Figure 11: the concentration reached a peak of 55 mmol/cm 3 , the solute made approximately 30 days on the soil surface (Node N1). On the 4 th day, the solute has reached node N2 with a peak concentration of 5 mmol/cm 3 , the solute has left this horizon on the 46 th day after application. At 11 days of the application, the solute has reached node N3 with a peak of 4 mmol/cm 3 , and left this depth on the 72 nd day. Node N4 was reached on the 20 th day and cancels itself out on the 106 th day. All the concentration has left the profile after 110 days of application. Figure 12 shows us a solute supply made on the 5 th day of simulation. A further peak of 52mmol/cm 3 was observed and the solute remained on the surface until the 32 nd day. The N2 horizon was reached on the 7 th day with an approximate peak of 5mmol/cm 3 , and was washed out of this horizon on the 48 th day. On the 15h day the solute reached node N3 with a peak of 4mmol/cm 3 and then vanished on day 75. Node N4 was reached on the 24 th day with a concentration of 2mmol/cm 3 and was canceled on the 110 th day.
For an application dose made on day 10 Figure 13, the solute front disappears  com pletely from the surface on day 34. The N2 layer was reached on the 12 th day with the same peak of 5 mmol/cm 3 , this layer then loses the solute on the 50 th day. The solute front reached the N3 level on the 19 th day with a peak concentration of 4 mmol/cm 3 . The solute has left this horizon on day 76, reaching the N4 layer on day 28 and canceling out on day 110. Applying a solute dose on day 15 after sowing Figure 14 shows a solute front that persisted at the surface until day 36. The front continues its distribution to reach level N2 on the 18 th day and remained there until the 52 nd day. The N3 layer was then reached on the 25 th day, the solute continues its progression to leave the N3 horizon on the 80 th day. The N4 layer was reached by the 32 nd day and the solute finally drains from the horizon on the 110 th day.
With a flux corresponding to 75% ET c , the residence time of the solute on the soil surface decreased as one approaches the date recommended for the application of the solute (15 th day after sowing).The solute remained on the soil surface 30, 28, 25, 22 days respectively for a solute applied on the 1 st , 5 th , 10 th and 15 th day after sowing. It then took an average of 3 days to reach the N2 horizon and 11 days to reach the N3 layer. The concentration of solutes in the different layers decreased with the decrease in the inflow of water. With a lower water flux (compared to that of 100% ET c ), the solute front distributed more slowly and therefore took longer to reach the other layers (N2, N3, N4 and N5) and would reduce the leaching loss. The solute's stay in the soil was even longer before reaching the bottom limit of the domain.

Water Flux 50% ET c and NPK Applied Either on the 1 st , 5 th , 10 th or 15 th Day after Sowing
For a 50% ET c treatment, the curve giving the pressure head potential in the soil, showed an equilibrium-state of the water in the soil from the 70 th day until the end of the simulation. The pressure head potential being constant in all observation nodes. Referring to the water mass balance, it can be seen that there was no more water loss in the soil profile after the 66 th day but rather water storage in the soil profile. For a supply of solutes on the 1 st day of the simulation Figure 15 the solute spent approximately 49 days on the soil surface (node N1) with a peak concentration of 37.5 mmol/cm 3 before disappearing. It then arrived at node N2 on day 9 to stay there for 86 days before leaving horizon N2; the maximum concentration for N2 being 2.5 mmol/cm 3 . Node N3 was reached on the 22 nd day, at this level, the solute has left horizon N3 on the 110 th day of the simulation. On the 35 th day, the solute reached node N4 and leaved this horizon on the 110 th day. The bottom of the domain (node N5) was reached on the 59 th day, at this layer, the concentration of the solute did not cancel out even after the 110 th day.
For an NPK treatment on the 5 th day after sowing Figure 16 the solute has spent 50 days on the soil surface (node N1) before dispersing inside the soil. Continuing its progression through the soil, the solute arrived at node N2 on day 12 and canceled out on day 96. The distribution of the solute being done with the water infiltration, node N3 was reached on the 25 th day; the solute has left the N3 horizon on the 110 th day. 10 days after reaching node N3, the solute has arrived at node N4 and was no longer washed out of this horizon, even after the 110 th day of simulation. Likewise, after reaching node N5 on the 59 th day, the solute remained at the bottom of the domain after the 110 th day, but at low concentration. Figure 17 shows us a supply of solutes made on the 10 th day, we observed through node N1 that the solute has made 52 days at the surface of the soil with   a peak concentration of 36 mmol/cm 3 . Horizon N2 was reached on the 18 th day, the solute left this horizon on the 98 th day of simulation. Continuing its progression inside the ground, the solute arrived at node N3 on the 28 th day and was washed away from this horizon on the 110 th day. The two nodes N4 and N5 were reached by the solute respectively on the 38 th and 62 nd day. At these last two nodes, the solute was no longer leached even after the last day of simulation. But concentrations remained low at these horizons. Applying a dose of solute on day 15 Figure 18, the solute remained on the soil surface until the 54 th day. The solute front has continued its distribution to reach With a 50% ET c treatment, the time that the solute did on the soil surface decreases as one approaches the recommended day for the application of the solute (15 th day after sowing). The same observation was made for the flows corresponding to 100% ET c , and 75% ET c . The solute remained on the surface of the soil 49, 46, 43, 40 days respectively for a solute applied on the 1 st , 5 th , 10 th and 15 th day after sowing. It then took an average of 8 days to reach the N2 horizon and 19 days to reach the N3 layer. The concentration of solutes in the different layers decreased with the decrease in the inflow of water. With a lower water flow (compared to that of 75% ET c , or 100% ET c ), the solute front distributed more slowly and therefore took longer to reach the other horizons (N2, N3, N4 and N5) and reduced leaching loss. With 50% ET c , the residence time of the solute on the surface was significantly longer. No solute was lost at the bottom edge of the domain for the various treatments on the last day of simulations.
The first two layers of the soil being sandy, this explains the fairly rapid leaching towards the last layer for high water flux (100% ET c , and 75% ET c ).
Only the 50% ET c treatment made it possible to gain water and solutes (no deep drainage or leaching beyond the limit of the bottom of the domain). An addition of solute on the 1 st or 5 th day after sowing regardless of the water flux provided (100% ET c , 75% ET c , or 50% ET c ) will not have a great impact on the absorption of solute by the crop because the roots are not yet developed at this stage. We analyzed the solute inputs on the 10 th and 15 th day after sowing. The closer we got to the 15 th day, the more we have got a loss of solutes, probably due to the gradual storage of water, as it adds up. A favorable day for solute supply would therefore be between the 10 th and 14 th day after sowing. To reduce the solute infiltration velocity into the soil layers, one could inject solute without first supplying water the day before or the day after application. By simulating such a process, we realized that this treatment in addition to reducing the solute infiltration rate in the soil, also increases the concentration of the solute in the different horizons of the soil. This would increase the availability of the solute for the roots of the crop.

Conclusion
We simulated water and non-reactive solutes' transport into a 3-layered soil profile, with the Hydrus 1D software. For the simulation, we evaluated the doses of solutes that should be provided for the production of the intermediate variety of maize, the crop cycle of which was estimated at 110 days over an area of 500 m 2 .
Different simulations were performed at 100% ET c , 75% ET c , and 50% ET c in terms of water supplies combined at the doses of 100 mmol/L/m 2 NPK and 246 mmol/L/m 2 of urea for the solute inputs, referring to the technical itinerary for maize production in Burkina Faso. For the three simulated treatments, there was more deep drainage for water flux at 100% ET c or 75% ET c , leading at the same time to loss of solutes beyond a certain depth of the soil not colonized by roots.
The water flux at 50% ET c further reduced the loss of water by drainage and of solutes by leaching, this flux hardly caused any loss beyond the bottom of the domain. It has also improved the residence time of solute in the soil. The closer one gets to the 15 th day after sowing, the greater the risk of loss, because the soil has already absorbed water. This loss can be reduced by spacing the water supply and the solute supply, or by amending soil with organic matter in order to increase its retention capacity, which could improve the residence time of the solute.