Nitrogen Balance in a Poorly Draining Intensively Cultivated Soil

The increasing use of nitrogen in fertilizers has not only increased the productivity of agricultural crops but also the concern for the effects of high N inputs on the natural environment. Economic, agronomic and environmental concerns have led to greater efforts towards more effective utilization of nitrogenous fertilizers, which can be achieved through a better quantitative understanding of the fate of nitrogen when applied to the soil. So Nitrogen balance was investigated in a poorly draining intensively cultivated soil. Nitrogen inputs and outputs, such as additions by precipitation (Natm), irrigation water (Nirr), commercial fertilizers (Nfert), mineralization (Nmin), denitrification (Ndenitr), volatilization (Ngas), capillary rise (Ncap), uptake (Nupt), residual (Nres) and leaching (Nleac) were studied and quantified during a period of 72 weeks for two growing crops (maize and wheat). Nitrogen excess or deficit at time ‘t’ (Nt) was calculated from the following balance equation: Nt = Nres + Natm + Nirr + Nmin + Nfert + Ncap − Nupt − Ndenitr – Ngas − Nleac. Existing models were used after proper adjustments to calculate these fluxes. Only a single model, based on soil temperature at 60 cm and NO3-N concentration in ground water was developed for the estimation of denitrification at the ground water boundary. The calculated values satisfactorily matched the respective experimental data. The results shown that the greater part of the applied nitrogen fertilizer was lost to ground water during the winter causing considerable environmental concern. For sustainable land use, inputs and outputs of nutrients must be balanced in order to avoid negative impacts on the environment, especially in intensively exploited agricultural ecosystems.


Introduction
Much research has been carried out over the years into the fate and efficiency of nitrate-nitrogen (NO3 --N) in agricultural ecosystems.The increasing use of nitrogen in fertilizers has not only increased the productivity of agricultural crops but also the concern for the effects of high N inputs on the natural environment (Janzen et al., 2003;Eickout et al., 2006;Ross et al., 2008).This concern has focused on the concentration of NO3 − -N in surface water and its leaching into groundwater, its release into the atmosphere as ammonia and its subsequent deposition onto the soil surface, leading to soil acidification and eutrophism (Merrington et al., 2002).Nitrous oxide is also produced, which contributes as a greenhouse gas to global warming and to destruction of the ozone layer (Crutzen, 1981;Franzluebbers, 2007;Herrero et al., 2010).These losses are connected to the nitrogen cycle in such a way that an attempt to reduce one source of N-pollutant may induce an increase in another.Consequently, an integrated study of such pollutants is required.Such work will lead to the development of local, national and international strategies for minimizing N losses from agricultural ecosystems.The EEC directive 91/676/EEC requires member states to minimize direct and indirect water pollution by NO3 --N from agricultural sources and to implement measures to avoid further pollution.Thus, the farmer can more accurately determine the amount of fertilizer to use on a crop to manage yield yet avoid over fertilization, while political decision makers can identify agricultural best practices Different models of nitrogen balance describe the nitrogen flows (inputs and outputs) in the soil-water-plantatmosphere system (Vereecken et al., 2016).These models vary in their conceptualization, complexity and the concerns, which they address.Studies on N balance have been reported in several long-term experiments (Davis et al., 2003;Ross et al., 2008;Pieri et al., 2011).Accurate measurement of N balance is, however, difficult because of the complexity of measurement of some parameters and increased time, labor, and cost constraints.Differences in soils, climatic conditions, crop species, and management practices can influence N balance in various agro-ecosystems due to variations in N inputs, outputs, and retention in the soil (Meisinger and Randall, 1991;Ross et al., 2008;Pieri et al., 2011).
Recent research on dealing with water pollution from NO3 − -N has focused on the factors and processes, which regulate the concentration of N in the root zone.Economic, agronomic and environmental concerns have led to greater efforts towards more effective utilization of nitrogenous was applied by sprinkler on 25/6, 13/7, 29/7, 18/8 with 180, 150, 160, 150 m 3 H2O ha -1 , respectively (Table 1).At the end of the cultivation period, the aerial parts of the crop were removed from each plot and the plant mass divided into leaves/stems and seeds.These were dried to constant weight at 65 °C, weighed and the dry weight per ha calculated.The Kjeldahl method was used to determine the total N content.

Measurements of soil physical and chemical properties
Soil samples were taken from each horizon in each experimental plot.The samples were transferred to the laboratory, air-dried, and ground to pass a 2 mm sieve.This soil fraction was used to determine: particle size distribution by hydrometric method; organic matter content using a modified Walkley-Black method; calcium carbonate equivalence by CO2 release on reaction with HCl; total soil N by Kjedahl, pH in 1:1 soil:water suspension; Bulk density using appropriate metal cylinders; field capacity and water content at wilting point by use of Richard's apparatus.All the above methods are described in detail in Soil Survey Laboratory Methods (2014).
For the requirements of the water balance model (WATBAL) (Berghuijs-van Dijk, 1985;Yates, 1996) a) field capacity and wilting point were calculated for two soil layers: 0-38 and 38-109 cm, b) Saturated water hydraulic conductivity in the field for three layers, 0-38, 38-72 and 72-109 cm, and for 38-109 cm, were determined using the crust infiltration method (Bouma and Benning, 1972).The measured soil physical and chemical properties are given in Table 2.
Composite soil samples were taken before sowing from each plot (from three sites per plot) for 0-15, 15-30, 30-60, 60-90 cm three times during the each cultivation period before sowing, in the middle of the growing season and after harvest (Table 1).In order to show the variation of NO3 --N and NH4 + -N with time, these N forms were determined by extraction with 2 M KCl.NH4 + -N was determined using the indophenol blue method, and NO3 --N by cadmium reduction method (Keeney and Nelson, 1982).For monitoring NO3 --N and NH4 + -N in the soil solution, at the experimental site, vertical pipes with ceramic cups at their lower end, connected with a tube to a pump, were installed at a depth of 80 cm in each plot.The soil solution was sampled by applying suction of 250 mm Hg every two weeks and the concentration of NO3 --N and NH4 + -N determined.
The depth of the water table was assessed using piezometers installed in each plot.Meteorological data were sourced from the national meteorological station at Aliartos, located 2 km from the experimental field.

ET o and ET a calculations
Potential evapotranspiration (ETo) was calculated using the Penman-Monteith equation (Allen et al., 1998), while actual evapotranspiration was calculated from the equation: ETa = Ea + TRM, where TRM is the true transpiration rate (mm d -1 ) and Ea the true evaporation rate (mm d -1 ).Ea was calculated using the equation: where LAI is the leaf area index (cm 3 cm -3 ), CMC the water content of the surface horizon, Eo the open pan evaporation rate, SMO the saturated water content, SMAD the water content of the respective air-dried soil (Driessen, 1988).The LAI at different growth stages of crops was determined using a leaf area analyzer (LI-COR 2000).
fertilizers, which can be achieved through a better quantitative understanding of the fate of nitrogen when applied to the soil.
The aim of the present study was the quantitative assessment of nitrogen inputs and outputs for a poorly draining intensively cultivated soil situated on an alluvial plain formed on lake deposits, with the intent of sustainable utilisation of nitrogenous fertilizers and correspondingly reducing environmental pollution.

Location and soil classification of the experimental site
The soil studied is located in the drained Kopaida basin, Veotia prefecture, Greece (38°32' N 23°06 E, 110 m asl), in an alluvial plain of lake deposits, intensively cultivated with maize, wheat and cotton.The soil has a perched water table, is very poorly draining and is classified as a Humic Fluvaquent, fine mixed calcaric (Soil Survey Staff, 1975).The water table level is influenced by local use of irrigation (lateral flow) and usually fluctuates at a depth of 85-145 cm.

Experimental design and treatments
The experimental plot was divided into three rows of four plots, each of 12 m 2 , with 0.5 m between plots and 1 m between rows.The experiment followed the complete randomized block design, with four levels of N and three replicates.
Two different cultivation periods with different crops were studied for 72 consecutive weeks (504 days).The experiment started in March and finished in November the following year.Table 1 presents the cultivation periods and the weeks with the corresponding treatments.In the first cultivation period, March to October, maize was grown 'Pioneer', while in the following, November to July, wheat was grown 'Mexicali-81'.
For maize cultivation, the following N treatments were applied: 0, 95, 193 and 287 kg N ha -1 , while those for wheat were 0, 40, 70, 80 and 128 kg N ha -1 .Ammonium phosphate (20-10-10) was used for both crops.The fertilizer was incorporated into the upper 15 cm of the soil during sowing.For maize, irrigation 141 Table 1.Treatments through the study periods Soil temperature measurements and calculations Soil temperature was measured using a thermocouple (Fe-Constantan) installed in the field at depths of 5, 15, 30 and 60 cm.Soil temperature at different depths was calculated on a daily basis using the following equation: T(z, t) = Tm + Aoe (z-dt) sin[ω (t − t0) − z/d], where T(z, t) is the temperature at depth z at time t (Scott, 2000).The equation describes a curve for temperature behaving as a sine wave with angular frequency of ω = 2π/ρ (where ρ is the period, 24 h for daily wave, 365 days for yearly), amplitude Ao around the temperature Tm which characterizes the soil, and d is the damping depth.The soil temperature was expressed on a weekly basis for use in WATBAL model.

N min determination
Mineralization of organic N (Nmin) was determined by soil incubation experiments at 7, 20, 30 and 35 °C for 28 weeks following the procedure described by Stanford and Smith (1972).Residual nitrogen in the soil after incubation was removed by successive washing with 10 ml of 0.01 M CaCl2.Then 25 ml of nutrient medium was added containing 0.002 M CaSO42H2O, 0.002 M MgSO4, 0.005 M Ca(H2PO4)2H2O and 0.0025 M K2SO4.There were two replicates for each sample.The samples were placed in an incubator and after two weeks incubation, the tubes were removed, and soil samples were extracted with successive applications of 10 ml of 0.01 M CaCl2 to a total of 100 ml in which NO3 --N and NH4 + -N were determined.This process was repeated after 5, 7, 8 and 9 weeks, and the cumulative amounts of NO3 --N and NH4 + -N (Ncum) mineralized between weeks were calculated for the laboratory estimate.Ncum was also determined at field conditions using 1800 g of soil (< 2 mm) uniformly mixed with an equal quantity of quartz sand (1 mm) and placed in plastic cylinders of 40 cm height and 12 cm diameter, each with a mesh at the base to prevent loss of material and a funnel for collection of soil leachate.The columns were placed in the field such that its upper surface was at the same level as the field soil surface.Soil temperature and moisture of the columns were measured at a depth of 5 and 30 cm using thermocouples and soil moisture meters.New columns were prepared for the second period.For field determination of Ncum, the same procedure was followed as for the laboratory determination.
Nmin was calculated using a first-order reaction equation proposed by Stanford and Smith (1972) where Ncum is the cumulative N-mineralization for incubation time t, Nmin is the potential mineralization, and k the constant for mineralization rate.Nmin and k were derived from non-linear regression, since this gives more accurate results (Talpaz et al., 1981) than with the techniques originally proposed by Stanford and Smith (1972).Losses of N (Ngas) in the form of gaseous ammonia (NH3) from the N applied in fertilizer were measured in a separate identical experiment in the same field following the procedure proposed by Kissel et al. (1977).NH3 gas lost from the soil surface was forced to pass through a chemical trap containing 175 ml of 2% H3BO3 for 10 min.The amount of NH3 trapped was determined by titration with HCl (Keeney and Nelson, 1982).
In order to determine the amount of N deposited from the atmosphere (Natm), rain collectors were located in the field.After each rainfall event, the volume of collected water and its NO3 --N content were determined.In addition, the NO3 --N concentration in irrigation water (Nirr) was determined before irrigating.

WATBAL model requirements
The soil profile was divided into two layers (0-38, 38-109 cm) in accordance with the requirements for WATBAL model for determining the water balance on a weekly basis, adjusted for the level of the water table, the measured actual evaporation, and the calculated evapotranspiration.According to the model, the soil moisture in the first zone was uniformly distributed, while that in the second increased linearly with depth until reaching saturation at the level of the water table.The physical properties of the soil taken into account in applying the model were:  The depth of the first (38 cm) and the second (71 cm) layers  Soil water content at wilting point of the two layers,  Soil water content at field capacity of the two layers,  Soil moisture content at saturation of the two layers,  The balance equation for the first layer can be written as: and for the second layer as: dM2/dt = fc(M2) + ft(M1) + fd(M1), where PR is rainfall, ETa is actual evapotranspiration, fc(M2) is water from the second layer by capillary transfer, ft(M1) the transfer of water from the first to second layer, fd(M1) the drainage of water from the first layer.

N leac calculation
The amount of NO3 − -N leached from the root zone (Nl eac) was calculated using the equation Nl eac = [(DP/MC × CNO3)] × R, where DP is the amount of water removed from the root zone calculated from the WATBAL model, CNO3 and MC are the NO3 − -N concentration and soil moisture, respectively, in the root zone, and R is a loss factor.R was calculated from NO3 − -N leaching of undisturbed soil.For this purpose, five columns 40 cm in length and 20 cm in diameter with undisturbed soil were taken to the laboratory.Leaching of the columns was carried out using 4 L of water per column and the NO3 − -N in leachate was measured.The NO3 − -N loss in leachate was then calculated as Nl eac = [(DP/MC × CNO3)].The amount of NO3 − -N leached was less than that calculated from the equation due to the presence of immobile water (Vanclooster et al., 1992;Manzoni and Porporato 2009).The difference was calculated as a percentage and used as an estimate of R.

Statistical analysis
Statistical analysis was carried out using STATISTICA (StatSoft 1995).

Soil temperature
The soil temperature followed a sine wave curve at all depths.The daily temperature difference at 5 cm was from 3.5-13.0°C, while that for 60 cm was less than 0.5 °C.The soil temperature varied throughout the year, with an average value of 16.5 °C depth overall.Although soil temperature is related to soil water, physical soil properties, climate and vegetation cover, the curve satisfactorily described the soil temperature for the studied depths.

Soil moisture
The soil moisture in the root zone (0-38 cm) ranged from 0.198 cm 3 cm −3 (approximately the wilting point of this soil) to 143 0.46 cm 3 cm −3 (Fig. 1(b)) depending on evapotranspiration.The soil moisture during the first cultivation period was relatively high since the water table was between 85 and 145 cm (due to irrigation method-lateral flow) (Fig. 2).During the next cultivation period, partially during the summer, the soil remained dry due to the unusually dry winter, resulting in the water table dropping below 170 cm depth.The WATBAL model predicted soil moisture relatively well for the root zone, except in the summer of the second cultivation period, for which soil moisture was underestimated by 30% (Fig. 1(b)).This difference is likely to be the result of the mulch zone, which was created in the top 5 cm of soil from the wheat crop when evapotranspiration was very high.

N addition by rain and irrigation (N atm + N irr )
The total contribution of Natm + Nirr was calculated as 10 kg N ha −1 year −1 .

N Removal by crop uptake (N upt )
The dry weight of crop plants and the total N uptake by crops are shown in Table 3.The result show that N uptake by the two crops was significantly greater than that applied in nitrogenous fertilizer, reflecting the significant contribution to crop nutrition, particularly by Nmin.

N loss due to denitrification (N den )
The losses of N as Nden in this particular soil were expected to be high due to (a) the relatively high water table (Fig. 2), (b) the high pH ( > 7.7) (Table 2) and (c) the high temperature in the denitrification zone (7.5-21°C) (Fig. 3).The soil layers below the root zone (>38 cm) remained at moisture contents above their field capacity.The WATBAL model showed that capillary water contributed significantly to the evapotranspiration demands of root zone soil from starting to the end of studied period (Fig. 2).
The concentration of NO3 − -N in soil solution sampled extracting by ceramic cups was lower than that of groundwater.The difference was related to the temperature at 60 cm.The NO3 − -N concentration in the soil solution was approximately 25 times lower than that of the groundwater during the summer, when the average temperature at 60 cm was 19.8 °C.In contrast, during winter when the average temperature at 60   cm was 6.5°C, the NO3 − -N concentration in the soil solution was 1.2 fold lower than that of groundwater.The groundwater level varied from 85-135 cm between winter and summer.The soil moisture at sampled depths remained relatively stable (46.8-47.9%)due to the irrigation method.
Therefore, soil oxygenation did not differ significantly between summer and winter.Consequently, the only factor contributing to loss of N during the summer months was the increase in reduction by micro-organisms due to increasing temperature.Another possible factor which could have contributed to N loss during this period is immobilisation of N by micro-organisms, which also increases due to increasing temperature.The presence of roots and their decomposition at the sampled depth (80 cm) was insignificant.Therefore the immobilisation of N at this depth is considered negligible.
During the summer, when the soil temperature at 60 cm was 18-21 °C, the activity of denitrifying micro-organisms was very high and the NO3 --N concentration in the soil solution of the denitrifying zone was less than 1 mg L -1 .However, when the temperature fell below 7.1 °C, the NO3 --N concentration in the soil solution of this zone was > 9 mg L -1 .The average NO3 --N concentration of groundwater for the study period was 11.76 ± 3.2 mg L -1 (n = 48).
The NO3 --N concentration in the soil solution in the denitrifying zone, regardless of N treatment, followed an exponential curve dependent on the soil temperature at 60 cm (Fig. 3).The equations shown in Fig. 3 are of first-order reaction kinetics, similar to those describing the decomposition of organic matter (Y = A e (−kt) ).The constant 'k' (average k = 0.1546, with standard deviation 2×10 −3 ) and the value of 'A' (average A = 23.57, with standard deviation 0.103) were independent of N treatment.The value of A is approximately twice the average NO3 --N concentration in groundwater (i.e.2×11.76 = 23.5).Therefore, the equation for estimation of denitrification can be written as: Css = 2Cgw e (−kt) , where Css and Cgw are the NO3 --N concentration in the soil solution (at 80 cm) and groundwater, respectively.The NO3 --N concentration in the soil solution of the denitrifiying zone was calculated per week using the above equation, the temperature at 60 cm and the NO3 --N concentration in groundwater.The N transported to the atmosphere was calculated cumulatively as the product of the differences (Css − Cgw) and the quantity of rising capillary water calculated by the WATBAL during the execution of the experiment.
The greatest losses of N recorded during the summer months can be attributed to the high soil temperatures and high quantities of capillary water moving to the rhizosphere during this period (Fig. 4).The total loss of N due to denitrification during the first cultivation period 18 kg ha -1 , while for the second it was only 7.1 kg ha -1 The total N lost to the atmosphere for the 37 week period of capillary rise calculated by WATBAL was 25.3 kg ha -1 .

Loss of N as NH 3 (N gas )
The amount of N lost from ammoniacal fertilizers is the result of a combination of soil properties (pH, particle size distribution, cation exchange capacity, equivalent CaCO3; fertilizer type, quantity and means of application; and climatic conditions (Sommer et al., 2004)).For the studied soil, the CaCO3 (52.5%) and pH (7.7) were at levels, which favour the loss of N as NH3 gas.
The highest amount of loss as NH3 was recorded when ammonium fertilizer was applied to the soil surface (Fig. 5).This is in accordance with the findings of Sommer et al. (2004).Relatively low losses of N occurred at times of the day when soil Fig. 3. Relationship between NO 3 --N concentration in soil water extracted with ceramic cups and soil temperature Fig. 4. N losses (N gas ) from denitrification (N den ), soil temperature at 60 cm depth, amount of capillary water raise to the root zone and NO 3 --N concentration of ground water temperature falls and relative humidity rises, i.e. late in the afternoon until early the next morning.A significant loss of N as NH3 was recorded between 11 am and 4 pm, when soil temperature was high and relative humidity low, conditions which favour evaporation.Generally, conditions which favour the loss of water also favour the loss of N as NH3 (Sommer et al., 2004).Consequently, two processes are mainly responsible for N loss as NH3: i) conditions which favour the evaporation of water increase the escape of NH3 gas from the soil surface; ii) the upward movement of water due to surface drying facilitates NH3 transport.During the second experiment (May), rainfall of 18 mm during the second night after fertilizer application was followed by an almost totally cloudless sunny day.This resulted in greater losses of water through evaporation and of NH3 from the experimental plots to which fertilizer had been applied (Fig. 5).Loss of N as NH3 was significantly high during the first two days after fertilizer application (Fig. 5).After four days, the loss of NH3 declined massively and measurements ended.The incorporation of the lowest N treatment (88 kg N ha -1 ) is the upper 15 cm of the soil led to the lowest gaseous loss.A significant difference between NH3 losses in April and May existed, with those in May being much higher.The lower soil moisture in May (0.30 cm 3 cm −3 ) compared with April (0.40 cm 3 cm −3 ) on the one hand, and on the other, the higher average soil temperature in May (20 °C ) compared with April (14 °C) (Figs.1(a), 1(b)) resulted in greater losses of N as NH3.

Potential mineralization of organic N (N min )
Nmin at time t was calculated for the soil layer 0-38 cm under both laboratory and field conditions.Although the process of mineralization of organic N can take place at a greater depth, the quantity was not considered significant.The cumulative rate of mineralisation (Ncum) decreased over time, under both laboratory and field conditions (Fig. 6).Entering the laboratory data into the equation for the first-order reaction Ncum = Nmin (1 − e (−kt) ) Stanford and Smith (1972) showed that the equation was successful.Although the field temperatures measured at the time of Ncum determination during the first cultivation period were higher (18.7-26.4°C) than the lowest temperature used for laboratory soil incubation (7 °C), Ncum was actually lower (80.7 vs. 97.8mg kg -1 soil).These differences are clearly the result of more favourable soil moisture conditions in the laboratory setting.During the incubation experiment in the field, soil moisture was 28.5-36.9%,while in the laboratory it remained stable at field capacity (44.5%).
Nmin as calculated using the equation of Stanford and Smith was 88.4 and 106 mg kg −1 soil for the field and laboratory (7 °C) conditions, respectively.Nmin for the rhizosphere was calculated as 257 kg ha -1 , resulting in 10% lower yields in plots which did not receive N fertilizer compared with that which received the highest (287.4 kg ha -1 ).During the subsequent cultivation period (wheat), Nmin in the field decreased significantly to 128.6 kg ha -1 due to low temperatures and unusual weather conditions during spring and summer.Nmin during the winter period was about 3-fold lower than during the spring (Fig. 1(c)).Nmin was predicted well during the spring and summer, and for both cultivation periods using the equation of Stanford and Smith, adjusted accordingly by calculated soil temperature and moisture (Figs.1(a), 1(b), and 6).Correct prediction of soil temperature and moisture are critical for the correct prediction of Nmin.Overestimation of Nmin (by 34%) at the end of autumn and during winter can be attributed to immobilization of N (De Bruin et al. 1989) and possible dependence of k not only on temperature but also on the composition of organic matter (El-Harris et al., 1983).At the end of the first cultivation period fresh columns were installed in the experiment field to investigate changes in mineralisation after incorporation of plant residues (mostly roots and some leaves) and N mineralisation was measured for 28 weeks.As can be seen from Table 5, Nmin for columns installed after the first cultivation period compared with those installed at its beginning was higher.
Residual soil N Fig. 8 (1 st sampling) shows that at the beginning of the first cultivation period (March), the amount of residual NO3 --N was low.During this period, the greatest values of residual NO3 - Fig. 5. NH 3 losses (N gas ) under different modes, rates and time of applied NH 4 Fig. 6.Cumulative N mineralized (N min ) under laboratory and field conditions -N were found for the 38-72 cm layer, immediately below the root zone.The low values for NO3 --N and its distribution in the soil profile during spring indicate its leaching during the winter.Approximately in the middle of the cultivation period, high amounts of NO3 --N were found in the upper 15 cm of the soil profile due to the preceding fertilizer application, while below 38 cm the concentration declined abruptly (Fig. 8, 2 nd sampling).This pattern was found for all treatments.Leaching of NO3 − -N during this period was minimal due to the lack of rain and of irrigation.Notably, there was a significant increase in NO3 --N in the surface horizon of the experimental plots which did not receive fertilizer.The increase in NO3 --N content in this horizon, despite removal of NO3 --N by the crop, should be the result of mineralisation of N during this period (Figs.1(c) and 7).At the end of the first cultivation period (Fig. 8, 3 rd sampling), the NO3 --N had practically all been consumed in plots which had received no fertilizer or 94.6 kg N ha -1 , while NO3 --N representing 9.2% and 14.37% of applied N remained in soil in plots which received 193 and 287.4 kg N ha -1 , respectively.The NO3 --N content of soil below the root zone was quite low (Fig. 8), due to the high level of denitrification recorded during the warm summer.
At the beginning of the second cultivation period (Fig. 9, 4 th sampling), the distribution of NO3 --N in the soil profile shows the movement of NO3 --N, through leaching to deeper layers due to moderate rainfall (86.3 mm) during this period.As a result of two major rainfall events approximately in the middle of the second cultivation period, a great quantity of NO3 --N was leached to groundwater (Fig. 9, 5 th sampling).At the end of the cultivation period, there was practically no rainfall, and leaching was therefore not possible (Fig. 9, 6 th sampling).

N Input -Output
At the beginning of the experiment, the residual N (Nres) in the rhizosphere (0-38 cm), considered to be the initial N (Ninit) in the calculations, did not differ significantly between experimental plots (18 ± 1.4 kg N ha -1 ).At the end of the first cultivation period, after maize harvest, Nres was 29, 28, 27, 57, 103 kg N ha -1 for 0, 95, 193 and 287 kg N -1 , respectively, indicating increasing Nres with increasing Nfert (r = 0.92, p < 0.001).At the end of the second cultivation period, Nres did not differ significantly by treatment, indicating that a significant proportion of N had been leached to deeper soil layers during the winter.The outputs Nupt and Nres, which represent the "apparent losses", increased with increasing Nfert during the first (r = 0.88, p < 0.001) and second (r = 0.85, p < 0.001) cultivation period but also for the whole study period (r = 0.92, p < 0.001), indicating the increase of these N outputs with increasing Nfert.The apparent losses during the second cultivation period did not differ significantly from those of the first.However, Nleac was significantly greater than during the first and contributed to the apparent losses by more than 55%, showing that these losses of N were mainly due to leaching.The losses of N in gaseous form (Ngas + Nden) were less than those of the first cultivation period.The statistically significant correlation between apparent losses and Nleac not only for the first (r = 0.88, p < 0.001) and second (r = 0.85, p < 0.001) cultivation period but also for the whole study period (r = 0.92, p < 0.001) supports the conclusion that leaching is the primary route of N loss in this agricultural ecosystem.

Conclusions
The total N balance (Nt) for each cultivation period reveals a significant deficit, which can be attributed to forms of N which were not possible to be included in this study.Losses of these forms of N were relatively high, especially during the first cultivation period, leading to the conclusion that a significant amount of N was immobilized or taken up by weeds.
The potential N for this soil showed that high values of Nleac were the result of the application of high quantities of nitrogenous fertilizer.
The losses of N from the root zone (< 38 cm) in gaseous form and from denitrification were relatively low compared with losses by leaching.The reduction of losses of N by leaching requires accurate estimation of demands for nitrogen by plants under intensively cultivated agroecosystems.
Nmin was found to be a significant input into this particular ecosystem, and its values were greater than those for Nupt for both crops.This input of N also defines the response of crops to the application of N fertilizer.
Nitrogen which is taken up and removed by the crops increased with increasing N fertilization.
For sustainable land use, inputs and outputs of nutrients must be balanced in order to avoid negative impacts on the environment, especially in intensively exploited agricultural ecosystems.
Fig. 1(a) shows the temperature recorded at different times and the changes at a depth of 15 cm and the values predicted by the equation: T(z, t) = Tm + Aoe (z -dt) sin[ω (t − t0) − /d].This equation can be used to satisfactorily predict soil temperature of this soil.

Fig. 1 .
Fig. 1.Predicted and measured values for: (a) soil temperature at 15 cm, (b) soil moisture at rooting depth (0-38 cm) and (c) N min mineralized at rooting depth (0-38 cm), for 72 weeks' continuous study Fig. 2. Fluctuation of ground water table and amount of capillary water raise to the root zone (mm week -1 ) for the study period of 72 weeks the same column followed by the same letter are not significantly different according to Duncan multiple range test at p<0.05

Fig. 7 .
Fig. 7. Predicted and measured values of N mineralized (N min ) under field conditions 147

Fig. 8 .
Fig. 8. NO 3 --N distribution with soil depth under different rates of N fertilizer applications at different growth stages of plant, during the first study period (maize) Fig. 9. NO 3 --N distribution with soil depth under different rates of N fertilizer applications at different growth stages of plant, during the second study period (wheat)

Table 2 .
Selected physicochemical properties of the studied soil

Table 3 .
Dry matter of shoots, total yields and total N uptake by maize and wheat under different N treatments

Table 4 .
Total ammonia volatilization losses (NH3) under different rates and modes of applied Ν

Table 5 .
Total N mineralized (Nmin)under field conditions measured in soil columns at the beginning and end of the end of the first growing season (maize)