Nitrogen is the main crop nutrition element that has some impact on its productivity (Oenema et al., 2009). Continuous fertilisation with N fertiliser ensures large and constant harvests. Nitrogen excess in agricultural production has a negative impact on the environment (leached nitrates, evaporating ammonia and N oxide) (Ahlgren et al., 2008). The abuse of N fertiliser is becoming a global problem and affects the environment in a number of ways: (1) human health problems caused by nitrates (Powlson et al., 2008); (2) livestock health problems (Lundberg et al., 2008); (3) surface water eutrophication (Smith & Schindler, 2009); (4) formation of nitrous acid, which is linked to acid rain (Menz & Seip, 2004); (5) depletion of the ozone layer in the atmosphere because of atmospheric nitrate oxides, which strengthen ultraviolet radiation (Savci, 2012; Rosenstock et al., 2013); and (6) causes the global warming effect (Martinez-Blanco et al., 2014; Payen et al., 2015).

In terms of non-point source pollution, N constitutes 52 to 61% (Klaushal et al., 2011). Agricultural pollution constitutes on average 30 to 35% of N leaching (Klaushal et al., 2011). Due to particular agricultural operations in Lithuania, 36,000 tonnes of N per year reach the Baltic Sea waterborne load (Larsson & Granstedt, 2010). Previous research has shown that leaching of elements and compounds from the soil is essentially determined by the hydrothermal regime, the amount of these element in the soil and their fertilisation (Beeston et al., 2010; Komonweeraket et al., 2011). A higher rate of N_min leaching occurs in humid years (Sogbedji et al., 2001). Views on fertilisation systems constituting human influence on water quality are rather mixed in the scientific literature. While some authors claim that mineral fertilisation systems are more environmentally dangerous (Kokkora & Hann, 2007; Ulen et al., 2012), others claim the opposite, that organic systems are more dangerous (Hossain et al., 2007). Nonetheless, it is known that unproductive fertiliser losses and water pollution increase considerably through the irrational and unbalanced use of both mineral and organic fertiliser (Hongxu et al., 2011; Hyytiäinen et al., 2011). It was determined in Finland that the yield of crops increases substantially, while N leaching falls with balanced fertilisation with N carried out several times during the vegetation of crops (Salvioni et al., 2011). As tests have shown, NO₃- N leaching from the soil can be reduced significantly by changing scattered fertilisation into local (Nakamura et al., 2004).

Current knowledge suggests that the scale of leaching of chemical elements and compounds should be inversely proportional to the agroecosystem’s total biomass synthesis. Cabrera et al. (2007) determined that a bigger biomass synthesis and dryer year limit N leaching by drainage and ground water. Securing a rich harvest (crop rotation with cover crops and continuously occupied soil, plants of high biological potential, balance and rational fertilisation and application of organic production) reduces the potential water pollution risk (Thorup-Kristensen et al., 2012; Buckley & Carney, 2013).

The quality of water which filters though the soil and in the surface runoff depends on the changes that occur in the soil when a variety of crops is grown. The fact that around 67% of the leached N is from the soil has been known for a very long time (Ulen & Johansson, 2009). Research data on N migration and its compounds’ dynamics in the soil in the agroecosystems affected by human activity is not substantial.

Based on statistical data, cereals crop areas have increased since 1991 in Lithuania and constitute 34.3 to 41.8% of the total crop area. Perennial and annual grass areas reduced from 46 to 41%. For many years, large areas of perennial grasses, which are used for a longer time without inclusion in crop rotations, are used to grow cereals successively. The farmer’s decision to grow more of one or another kind of crop, as well as the general farming culture, is affected by economic interest. Such a one-sided approach can worsen the indicators of the agro landscape and become a potential non-point pollution source, even when other agrarian measures are used correctly and rationally.

On the basis of the data obtained, we have proposed the hypothesis that non-point source pollution of the agro-landscape is inversely proportional to the synthesis of total biomass of the ecosystem, and that all of the available means ensuring high crop productivity actually reduce the potential risk of water pollution. The novelty of this paper is the evaluation of the effect of crop rotations under different fertilisation levels on the environment. The aim of the research was to evaluate the relationship between crop rotations of different intensity and biological potential with the migration of N.

Data show that perennial grasses and sugar beets produced the biggest yield (3.7-10.9 and 10.3-10.4 t/ha of dry material respectively). In all years, the higher fertilisation (843 N kg/ha) cereal crop rotation yield was determined to be larger than in the lower fertilisation (540 N kg/ha) cereal crop rotation (Table 2).

As different types of crops were grown in the research field, in order to determine which crop rotation field was the most productive, each of the yields of dry material amount was converted to GJ/ha and cumulative curves were drawn. After eight years of research, the biggest productivity was found in crop rotation of cereal with higher fertilisation, which accumulated 1483 GJ/ha of total energy harvest (Fig. 2).

Less fertilised cereal crop rotation productivity was 35% lower, and it was the lowest in the research range (966 GJ/ha). Perennial grass crop rotation productivity was 11% lower than the higher fertilised cereal crop rotation and constituted 1318 GJ/ha. With respect to humus and N, the soil of the site was not very homogeneous, as the variation coefficients of these indicators were 17 and 14, respectively. At the end of the research the highest humus reserves were found in the field, where perennial grasses crop rotation was applied (4.11%) (Fig. 3). A slightly lower soil humus is obtained by applying higher fertilisation cereal crop rotation (3.7%), while the lowest is obtained using a lower fertilisation cereal crop rotation (2.94%). The data analysis revealed that the effect of different crop rotations on humus accumulation was significant: between TI and TII t_fact.=|–5.07|>t_theor.95%=2.11, between TI and TIII t_fact.=|–8.02|>t_theor.95%=2.11, and between TII and TIII t_fact.=|–2.28|>t_theor.95%=2.11. An increasing tendency in N_total accumulation in the soil was determined when using different fertilisation rates in the crop rotations (–0.25%, in TIII, –0.23% in TII, and –0.19% in TI). However, the differences between the treatments were not significant, between TI and TII t_fact.=|–0.96|<t_theor.95%=2.11; between TI and TIII – t_fact.=|–1.53|<t_theor.95%=2.11; and between TII and TIII - t_fact.=|–0.53|<t_theor.95%=2.11.

The most stable N_min concentration and accumulated amount were in TII, where variation coefficients of this element were 22% and 18%, respectively. With TI and TIII, N_min concentration and accumulated amount in the soil varied more (from –41 to 35% and from 51 to 57%, respectively; Table 3). The average perennial data for N_min concentration and accumulated amount in the soil are also rather uneven. The data show that in TI N_min concentration and its accumulation in the soil was lower than in TII, however, this difference was not statistically significant (t_fact.=|–0.94|<t_theor.95%=2.14 and t_fact.=|–0.79|<t_theor.95%=2.20, respectively). Statistically significant differences between N_min concentration and its accumulation in the soil were found in TI and TIII and in TII and TIII (t_fact.=2.98>t_theor.95%=2.14 and t_fact.=2.23>t_theor.95%=2.14; and t_fact.=5.27>t_theor.95%=2.14 and t_fact.=3.52>t_theor.95%=2.14, respectively).

N_min concentration and its accumulated amount in the soil in higher and lower fertilisation cereal crop rotations were not statistically significant, even though the amount of this element in a lower fertilisation cereal crop rotation was determined to be lower, 13 and 9% up to 6.6 mg/kg and 69 kg/ha, respectively. The lowest N_min concentration and its accumulated amount in the soil had the most productive perennial grasses crop rotation, and these indicators reached 3.3 mg/kg and 42 kg/ha, respectively. Correlation analysis results showed that N_min amount was closely related to fertilisation. Also, its amount in the soil had a tendency to fall when the yield was high (Fig. 4).

A weak reverse linear was determined between N_min amount in the soil and crop productivity. Increasing field crop productivity reduces the amount of N_min in the soil, whereas increasing field crop fertilisation by N fertiliser increases the amount of N_min in the soil. Field crop productivity determined the fluctuation of N_min quantity by around 15%, and fertilisation by around 55%.

Precipitation dispersion was uneven during the research period, and it had effect on the formation of drainage (Table 4). It can be seen from the results that four of the years were dry (2006, 2008, 2011, and 2013), when precipitation was 80, 97, 97, and 91%, respectively, of the climate normals; 2007, 2009, 2010, and 2012 were damp years, because precipitation was 113, 125, 122, and 123%, respectively, of the climate normals. During almost the entire research period, the average air temperature was higher than the climate normals, except in 2010, when it corresponded to it.

Drainage quantity depended on the precipitation level and changed according to the direct linear pattern, as a higher level of precipitation increased the drainage.

According to the research data, on average during 8 years in a more productive, higher fertilisation cereal (TII: 55 ± 23 mm) and grass crop rotation there was a falling tendency in drainage determined, while in the least productive and lower fertilisation cereal (TI: 71 ± 29 mm) crop rotation there was an increasing tendency observed.

Even though drainage was 29% higher in TI and 25% higher in TIII than in TII, the differences between these treatments were not significant: between TI and TII, t_fact.=1.20<t_theor.95%=2.14; between TI and TIII, t_fact.=0.12<t_theor.95%=2.14; and between TII and TIII, t_fact.=–0.98<t_theor.95%=2.14.

Drainage essentially depended on precipitation (r=0.68; t_fact.=4.31>t_theor.95%=2.15). More intensively grown, larger biological potential crops, which use greater amounts of water, reduce not only soil humidity, but also drainage (r=0.43); t_fact.=2.23>t_theor.95%=2.15) (Fig. 5). In TIII, N_min and N_total concentrations were found to be 12.1 and 12.8 mg/L, respectively, in drainage water (Table 5).

Using a lower fertilised cereal crop rotation (TI) meant that N_min and N_total concentrations had a tendency to increase by 13 and 24% to 13.7 and 15.9 mg/L, respectively. An increasing tendency was also determined for N_min and N_total concentrations in drainage water in TII (15.6 and 18.5 mg/L). According to correlation and regression data analysis (Table 6), N concentration in water was determined by a number of factors.

Thus, as the field crops‘ productivity increased, N concentration in the drainage water changes according to the parabolic pattern and it was at its lowest when the field crops‘ productivity was 150-170 GJ/ha. When a higher productivity is achieved by fertilising field crops, both N_min and N_total concentrations in the drainage water increase. The relationship between soil N_min and fertilisation by N_min fertiliser with N concentration in drainage water can be expressed by this function: y = a – b / x (Fig. 6); as soil N_min as well as fertilisation with N fertiliser increases, the concentration of both N forms in drainage water increases.

The relationship of annual and cumulative N balance with N concentration in drainage water is of a direct linear nature and, as N balance increases, the concentration of both N forms in drainage water increases. Direct linear consistency means that N_min concentration in drainage water increases as the soil N and humus increase. Due to the impact of field occupation coefficient N concentration in drainage water reduces.

Research data of leaching of N compounds by drainage showed (Table 5), that both N_min and N_total leaching had a downward trend when using grass crop rotation (7.0 and 8.0 kg/ha respectively). 


An increasing tendency was found for N leaching in lower fertilised cereal crop rotation conditions (12 and 13%) and in higher fertilised cereals crop rotation conditions, 10.0 kg N_min/ha and 11.5 kg N_total/ha, respectively. Correlation and regression data analysis has shown that both N_min and N_total leaching is related to the productivity of agroecosystem and factors affecting it (Fig. 6). Crop fertilisation with mineral fertiliser had a very big effect on N_min leaching (r=0.78), while it had only average effect on N_total leaching (r=0.65). The yield affected N_min leaching on average (r=0.57), while N_total was weakly affected (r=0.46); however, it was statistically significant.

Nitrogen leaching through the drainage relationship between the fertilisation of an agroecosystem with mineral fertiliser and its productivity is parabolic. Fertilisation with N fertiliser until 66-71 kg/ha a.m. does not increase N leaching losses. Increasing the harvest by fertilisation leads to the increase in N leaching. The lowest N_min leaching losses were found when productivity reached 213 GJ/ha, whereas the lowest N_total leaching losses were when productivity reached 212 GJ/ha. The relationship between N_min and N_total leaching by drainage with the balance of agroecosystem can be expressed by the formula y = a – b / x. As N balance changes from –300 to –20 kg N/ha, N_min and N_total leaching barely increases. It increases more significantly when the N balance becomes zero and positive. The relationship of N_min leaching (kg/ha) with soil N % (x₁) and humus % (x₂) is as follows:

During the 2006-2013 period, we were analysing several research questions, such as how N_min affects the yield, how fertilisation affects N_min in the soil, the effect of different crop rotations on the annual N balance, and the differences in mineral and N_total concentrations in drainage water with fertilisation, as well as how they depend on the soil.

The study showed that a higher fertilised cereal crop rotation had the highest cumulative productivity (1483 GJ/ha), while cereals with less fertilisation and grass crop rotations had a productivity which tended to be lower by 11 to 35%, respectively, suggesting that the crop yield depends on fertilisation rate. Similar results were found by Křen et al. (2014), who analysed spring barley yields applying different N fertilisation rates. The yield increase was determined to be higher using N rate from 0 kg N/ha to 45 kg N/ha, rather than the rate from 45 kg N/ha to 90 kg N/ha. Grover et al. (2009) claimed that crop rotation is a prerequisite for a large crop yield, especially when legume crops are included in the rotation. However, the cultivation of corn monoculture produced the same yield as the crop rotation (Grover et al., 2009). Babuliková (2014) contributed to the findings using wheat in the crop rotation. It was found that in order to obtain the highest wheat yield the best N rate was 96-168 kg N/ha. There was a small difference by fertilising with a rate of 150-200 kg N/ha; however, by increasing the N rate to 240 kg N/ha, the grain yield and N effectiveness dropped. In addition to this, a long-term research revealed that crop rotation and different fertilisation had only a slight effect on yield and soil fertility parameters (Cuvardic et al., 2004). On the other hand, according to Coulter et al. (2011), a high yield in the “extended crop rotations for corn and soybean” can be obtained, even by reducing the fertilisation rates. Additionally, Øgaard (2014) found that excessive fertilisation with N leads to a fall in crop yields, an increase in N losses, and the formation of high N surplus.

In general, the results suggest that higher fertilisation increases mineral N concentration and accumulated amount in the soil. This is evident from our research, which showed that the highest mineral N concentration and accumulated amount in the soil was found in the higher fertilisation cereal crop rotation (respectively 7.6 mg/kg and 76 kg/ha), while the lowest amounts and concentrations were found in the grass crop rotation (3.3 mg/kg and 42 kg/ha), with values that were 57 and 45%, respectively, lower than for the higher fertilised cereal crop rotation. Similar results were found by Cuvardic et al. (2004), who observed an increasing trend in soil N (from 0.32 to 0.36%). Additionally, our research showed an increase in humus (4.11%) in perennial grasses crop rotation (Fig. 3). Other researchers found that using N-fixing legumes and red clover increase mineral N concentrations in the soil and might even accumulate the amount of 200 kg/ha of biological N for further crop rotation (Lapinskas, 1998; Riedell et al., 2009). In addition, Röing et al. (2005) claimed that 50% more N mineralises in crop rotations with ley rather than in crop rotations where only cereals are grown. According to Sonneveld & Bouma (2003), on old pastures, reducing N application levels can lower the probability of exceeding the environmental threshold for nitrate by up 20%, whilst hardly affecting N uptake. Köhler et al. (2006), in sand soils, showed that nitrate leaching depended on N fertilisation and different crop rotations as well as mineralisation from soil organic matter. Catch crops were the most efficient way to reduce the NO₃ ^– concentrations in the groundwater recharge of sandy soils. These results suggest that the N_min concentrations in the soil depend not only on the fertilisation rate, but also on the type of crop rotation.

We found that the application of all crop rotations generally resulted in a negative N balance. However, the use of different cereals crop rotations showed a lower N balance deficit (–31 and –14 kg N/ha), while the grass crop rotation resulted in a negative N balance, which increased by 3.2 and 7.1 times and reached -100 kg N/ha. According to Kutra & Aksomaitiene (2003), the most negative N balance (-426 kg N/ha) was found in the perennial grass fields; the most positive balance (+ 174 kg N/ha) was in fields with sugar beets rotation.

The results from a number of countries show that the main environmental problem related to agriculture is N leaching. According to Arregui & Quemada (2006), the amount of leached N highly depends on the amount of it found in drainage and soil before planting crops. When N balance is positive, there is a high possibility that N concentrations in drainage water will be above the maximum allowable values. When N balance is negative, its concentrations are determined to be low. Crop rotation where perennial grasses are grown is regarded to be the best in terms of the impact to the environment, as N_total concentrations in drainage water were determined to be the lowest (12.8 mg/L) and not to exceed the required environmental regulations (VZ, 2011). This was also observed in our research, as the N_min and N_total leaching had a tendency to fall in grass crop rotation (7.0 and 8.0 kg/ha, respectively), while in cereal crop rotations with lower and higher fertilisation there was an increasing tendency in N leaching (by 13 to 42%) observed. Additionally, Kutra & Aksomaitiene (2003) showed that in fields where cereals were dominating, N surplus was + 74 kg N/ha. Weber & Kubiniok (2013) also found that in the soils sown with perennial grasses, chemical elements’ leaching losses were lower than in cereals. Thus, due to the fact that grass crop rotations reduce N concentrations in the soil, they also reduce the risk of it leaching into the drainage water. However, there are more factors that affect N leaching, as fertilization rate not higher than plant need and minimized soil tillage systems (Kutra et al., 2006).

Lysimetric research revealed that nitrate concentrations in lysimeter water were highly dependent on fertilisation rates. Moreover, according to a 30-year research work (Adomaitis et al., 2008), by fertilising crop rotation plants (winter wheat, sugar beet, spring barley, annual grasses and perennial grasses), using the fertilisation rates 112 and 224 kg N/ha, the nitrate concentrations observed in lysimeter water at a depth of 0-80 cm were 112.1 and 187.2 mg/L, respectively. Tripolskaja & Verbylienė (2014) determined that by fertilising with N fertiliser (on average 120 kg/ha), compared to unfertilised soil, the average nitrate concentration in lysimeter water was increased by 37.5 to 54.3%. However drainage, due to more dense and more exuberant crops, which reduce precipitation filtration, fell during the year on average by 13.7-16.2%. This is also supported by our research results, which showed that larger biological potential crops reduced drainage by 23% compared to lower biological potential crops. Tiemeyer et al. (2006) determined that N concentrations in drainage water were increasing by increasing flow rates, high loss rates always occurred at high flow rates, and higher loss rates are to be expected in wetter years. This was also confirmed by our research: when the precipitation level was higher, the drainage was also higher. Drainage essentially depended on meteorological conditions. Previous research showed that agroecosystem´s hidrological regime is determined not only by natural, but also by human farming activity factors (the crops grown, the farming level). More intensively grown, higher biological potential crops use up more water and thus reduce not only the soil moisture but also drainage (Aksomaitienė et al., 2002). The problematic nature of the research is that the majority of results were obtained from small drainage sites or even by lysimetric research methods. To determine similar patterns by performing tests in large drainage systems is also difficult, not only because of their large amortisation effect (Cornelese et al., 2001), but also because of the research background promiscuity (soil cover variegation, ground water stratification depth, etc.); however, under some circumstances, patterns remain (Sileika & Guzys, 2003).

Both N_min and N_total leaching by drainage, due to the crop productivity and their fertilisation with N fertiliser, change according to y = a – bx / cx ². The smallest N leaching losses were obtained by fertilising 66-71 kg N/ha and crop productivity of 212 GJ/ha. Furthermore, N_min and N_total concentration in drainage water depended on crop rotation, its productivity and the factors determining productivity. Both N_min and N_total concentration in drainage water had a tendency to fall in grass crop rotation (12.1 and 12.8 mg/L respectively). The use of cereal crop rotations with higher and lower fertilisation increased N_min concentration by 13 and 29%, while total N by 24 and 44%. According to Adomaitis et al. (2008), a larger amount of nitrates is leached when annual crops are grown, compared with perennial grasses. Regular grass sown in arable land and not fertilised with N reduced nitrate concentration to the level of perennial pasture in just one year. The risk of exceeding the maximum allowable limit of nitrates in crop rotation in arable land was significant and did not depend on the level of fertilisation with N (Mašauskas et al., 2006).

Moreover, Beaudoin et al. (2005), determined that nitrate concentrations change subject to the year, crops and soil type; in a deep loamy soil, concentrations fluctuated between 31 mg/L and 92 mg/L in shallow sand. Correa et al. (2005) and Tonitto et al. (2006) observed that the biggest leaching losses occurred in the bare, fallow soil. The lowest concentrations were determined in crop rotation with sugar beet/wheat (38 mg/L), whereas the highest concentrations were found in the pea/wheat crop rotation (66 mg/L) (Beaudoin et al., 2005). Kutra et al. (2006), found that the majority of N was leached from row crops (22.4 kg/ha), slightly less from cereal fields (16.6-18.9 kg/ha) and least of all from grasslands (10.5 kg/ha). On the contrary, Bohm et al. (2009) showed that legume crop cultivation can considerably increase N_min concentration in the soil solution and leaching as well.

The data imply that the amount of N_min in the soil is a very labile indicator. Loosening during the mineralisation period and spreading during fertilisation ensure that N_min is quickly used up by crops. The longer the crop vegetation period and the greater their biological potential, the stronger this preventative effect (Kutra et al., 2006; Askegaard et al., 2011; Doltra et al., 2011).

Finally, as the data collected by many researchers show, the advanced agricultural technologies and efficient application (not the rates) of mineral fertilisers are among the essential preconditions for reduced non-point source water pollution with N (Goulding et al., 2000; Sileika & Guzys, 2003).

The obtained results showed that the application of perennial grasses crop rotation in agricultural fields was the best environmental tool, reducing N migration to drainage. By fertilising with other treatments using higher N rates, mineral N accummulation in the soil increased, stimulating a higher yield and reducing N leaching by drainage. By fertilising with higher N rates than it is neccesary for plants, both mineral and total N leaching to drainage had a tendency to increase.