Top Navigation Bar

Article Request Page ASABE Journal Article

Efficiency of Drainage Practices for Improving Water Quality in Lithuania

A. Povilaitis, A. Rudzianskaite, S. Miseviciene, V. Gasiunas, O. Miseckaite, I. Živatkauskiene


Published in Transactions of the ASABE 61(1): 179-196 (doi: 10.13031/trans.12271). Copyright 2018 American Society of Agricultural and Biological Engineers.


Submitted for review in February 2017 as manuscript number NRES 12271; approved for publication as part of the “Advances in Drainage: Select Works from the 10th International Drainage Symposium” collection by the Natural Resources & Environmental Systems Community of ASABE in June 2017. Presented at the 2016 ASABE Annual Meeting as Paper No. 162513379.

The authors are Arvydas Povilaitis, Professor, Aurelija Rudzianskaite, Researcher, Stefanija Miseviciene, Researcher, Valerijus Gasiunas, Researcher, Otilija Miseckaite, Lecturer, and Ina Živatkauskiene, Doctoral Student, Institute of Water Resources Engineering, Aleksandras Stulginskis University, Kaunas, Lithuania. Corresponding author: Arvydas Povilaitis, Aleksandras Stulginskis University, Universiteto 10, LT-53361 Kaunas-Akademija, Lithuania; phone: +370-37-752337; e-mail: arvydas.povilaitis@asu.lt.

Abstract. Artificial drainage is a common agricultural practice in Lithuania. In this country, the total drained land area occupies 47% of the total land area and 87% of the agricultural land area. Therefore, this article presents recent research findings on agricultural drainage in Lithuania related to the practices designed to reduce nutrient, i.e., nitrogen (N) and phosphorus (P), losses from the soil via tile drainage and transport in open drains. Temporal changes in tile drainage flow over the last four decades are also discussed in this article. The results from experiments with controlled drainage practices in Lithuania showed promise. Compared to conventional drainage, controlled drainage reduced inorganic N by 42% to 77% and reduced total P by 34% to 72%. The reduced loads were the result of reduced drainage outflow. Moreover, research on the effects of additives in drainage trench backfills showed that woodchips, chopped straw, and lime additives mixed in the drainage trench backfill led to reductions in NO3-N concentrations of 78%, 69%, and 52%, respectively, in the drainage water. The addition of lime to drainage trench backfill reduced PO4-P concentrations in the drainage water by 39%, while woodchips and chopped straw increased the concentrations by 11% and 22%, respectively. It was determined that NO3-N in the drainage water was removed most effectively by woodchips and that PO4-P was removed most effectively by the addition of lime. The experiments with reactive filter materials used as in-ditch measures to remove phosphorus showed that the filter materials can be ranked as follows based on their P removal efficiencies: Polonite > slag > Filtralite-P > dolomite chips. Polonite had an advantage over the other tested materials due to its higher porosity, low sensitivity to clogging, and greater permeability. Laboratory-scale experiments using denitrification bioreactors filled with three types of woodchips (deciduous, coniferous, and mixed) showed no significant differences in NO3-N removal efficiency among the three materials. However, the tests showed that woodchip media are capable of achieving higher NO3-N removal rates due to higher flow rates. Therefore, better optimization and proper evaluation of the effects of hydraulic retention time are needed to improve the design of denitrifying woodchip bioreactors.

Keywords.Agricultural drainage, Controlled drainage, Denitrifying bioreactors, Drainage trench backfills, In-ditch filters.

Lithuania historically is an agricultural country. It covers a total area of 65,300 km2 and is situated along the southeastern shore of the Baltic Sea (fig. 1). The climate is temperate, and the spatial variations in the long-term annual averages (derived from observations at 27 meteorological stations over the last 60 years) of the main climatic characteristics are as follows: solar radiation = 2298 to 3185 MJ m-2, air temperature = 6.1°C to 8.0°C, precipitation = 560 to 910 mm, and evapotranspiration = 500 to 580 mm (Galvonaite et al., 2013). There are 30,000 streams longer than 0.25 km. The average density of the stream network is 0.99 km km-2. The runoff coefficient (defined as the ratio between average annual runoff and precipitation) is 0.32. The annual specific discharge (i.e., the rate of discharge per unit area) within the country varies from 5.0 to 12.0 L s-1 km-2 (Galvonaite et al., 2013). The atmospheric supply of water, the topography characterized primarily by plains, and the low permeability of the prevailing glacial tills result in more than 33,700 km2 of soils (51.6% of the total area) suffering from excess moisture (Aleknavicius, 1989).

To address the excess moisture, drainage of agricultural land in Lithuania started at the end of the 19th century. Throughout the history of land drainage, the overall drained area (ditches plus tile drains) has totaled 30,214 km2, or 87% of the agricultural land area. With a tile-drained area of 26,205 km2 (86.7% of the overall drained area), Lithuania remains one of the most extensively drained countries in the world (Povilaitis, 2015).

The subsurface drainage was largely achieved by burying tile pipes in the ground. The drain spacing, depending primarily on soil texture and longitudinal gradient, ranged between 16 and 30 m and had a depth of 0.9 to 1.2 m. Plastic (high-density polyethylene) pipes have only been used since 1986. Depending on local soil and climate conditions, the average annual specific subsurface drainage discharge in different parts of the country ranges from 4.0 to 6.0 L s-1 km-2 (Povilaitis et al., 2015).

The implementation of agricultural drainage has increased agricultural intensi?cation and productivity, but these gains have not been without environmental impact. In the 1990s, it was understood that the accelerated and enhanced lowering of the water table by subsurface drains and the expansion of arable areas, as well as regulation (deepening, widening, and straightening) of streams and destruction of riparian zones, had caused damage to rivers and topsoil (Pauliukevicius and Juodis, 1987; Zalakevicius, 2001). Intensive land drainage has also raised issues regarding its impact on local hydrology and water quality (Povilaitis, 2015).

Figure 1. Map of Lithuania showing locations of the study sites. Symbols in the legend are as follows: 1 = rivers and streams, 2 = lakes, 3 = tile drainage water monitoring site, 4 = meteorological station, and 5 to 8 = percentage distribution of drained agricultural area (5 = <65%, 6 = 65% to 75%, 7 = 75% to 85%, and 8 = >85%).

It is well known that tile drainage systems serve as transport pathways of contaminants directly from agricultural land to streams. When entering drains, water leaches nutrients (mostly inorganic forms of N and P), and the increased nutrient inflow into surface water bodies leads to their eutrophication, with subsequent degradation of the biological, ecological, social, and economic value of the environment (Jaynes et al., 2001; Galloway et al., 2004; Tiemeyer et al., 2006). In Lithuania, 80% of total nitrogen and 53% of total phosphorus that enters streams originates from agricultural areas (Povilaitis et al., 2014). Numerous investigations conducted in Lithuania have shown that the typical NO3-N concentrations in tile drainage water can vary from 3 to 20 mg L-1, while the total P concentrations can vary from 0.1 to 0.15 mg L-1, with annual losses ranging from 5 to 40 kg ha-1 for total N and from 0.10 to 0.40 kg ha-1 for total P (Povilaitis et al., 2015). As a result, Lithuania is referred as one of the “hot spots” of Baltic Sea pollution (HelCom, 2013). Consequently, by the year 2021, Lithuania is obligated to reduce the amounts of total N and total P entering the sea by 11,800 and 880 metric tons (26% and 66%), respectively. This is a considerable challenge that requires new solutions. Therefore, the main purpose of this article is to present the latest research findings on agricultural drainage in Lithuania with specific reference to practices designed to reduce nutrient losses from the soil via tile drainage and transport in open drains.

Worldwide, a variety of management practices and treatment systems have been investigated for better nutrient control in agricultural areas. These include in-field nutrient management, crop rotation, alternate land use, controlled drainage, and wetlands (Hoover et al., 2016). For example, by mixing lime, woodchips, or chopped straw into drainage trench backfill, the soil can be quickly drained, and the added materials can serve as pollutant treatments (Aškinis and Miseviciene, 1998). Moreover, reactive filter materials can be used for the removal of P from natural and waste waters (Douglas et al., 2004). Schärer (2003) and Smith et al. (2005) determined that Fe and Al oxides tended to adsorb P very rapidly. Renman (2008) reported that reactive filter materials, such as Polonite, lightweight expanded clay aggregate (LECA), and metallurgic slag, could be effective because of their high P sorption capacity. Phosphorus is removed by sorption to filter materials, which can then be recycled in agriculture as fertilizers and soil amendments (Hylander et al., 2006). These materials have also been tested as in-ditch measures in Lithuania.

As a new technology, woodchip denitrification bioreactors for tile drainage are being investigated for practical edge-of-field NO3 removal. This technology is based on routing tile drainage water through the bioreactors, where nitrate is used by bacteria to oxidize carbon while reducing NO3 to nitrogen gas. The first attempts to apply such biotechnologies in tile drainage systems were performed in the U.S. (Blowes et al., 1994; Cooke et al., 2001; van Driel et al., 2006). Subsequently, later studies (Jaynes et al., 2008; Cameron and Schipper, 2010; Christianson et al., 2009, 2011) suggested that application of “nature-driven” measures can substantially reduce NO3-N in drainage water. However, organic material is needed as a carbon source to promote the growth of heterotrophic denitrifying bacteria in bioreactors. The type of carbon fill is one of the most important considerations for denitrification systems. In general, woody media are the preferred type of fill material due to their longevity, availability, and high practicality in different locations (Robertson et al., 2005; Schipper et al., 2010). However, the efficiency of different types of woody media in the reduction of N in drainage water has not been widely investigated. Therefore, laboratory-scale experiments were conducted to test the efficiency of three types of local woodchips (deciduous, coniferous, and mixed) in the removal of NO3-N in tile drainage systems in Lithuania (Povilaitis, 2016). The NO3-N removal efficiencies and NO3-N removal rates observed in the experiments are presented here.

Materials and Methods

Because the results presented in this article are rather diverse (i.e., the analysis of temporal changes in drainage flow, the efficiency of in-ditch filters, laboratory-scale denitrifying bioreactor tests, and the effects of controlled drainage and trench backfill additives on nutrient concentrations in drainage water), the corresponding research experiments were not performed based on a unified methodology. Therefore, the main principles and routines along with the sampling methods and descriptions of the study sites are provided in detail below.

Drainage water quality, including ammonium-nitrogen (NH4-N), nitrate-nitrogen (NO3-N), total nitrogen (TN), total phosphorus (TP), and orthophosphate-phosphorus (PO4-P), as well as soil agrochemical properties (NH4-N, NO3-N, and plant-available P2O5) were analyzed according to approved standards in Lithuania. All concentrations in drainage water were determined via the spectrometric method. The plant-available P2O5 content was determined via the acidic ammonium lactate (A-L) extraction method (known as the A-L method from Egnér et al., 1960), and the NH4-N and NO3-N contents in the soil were determined via the spectrometric method.

Temporal Changes in Drainage Outflow

The study site where temporal changes in drainage outflow during the period 1977 to 2016 were investigated has a tile-drained area of 0.44 ha and is located in the middle part of Lithuania (54° 52' 46 N, 23° 51' 30 E). The drain depth ranges from 0.8 to 1.40 m, and the drain spacing varies from 12 to 18 m. The average surface slope of the site is 0.008 m m-1 (Sakalauskas, 1995). The prevailing soil is sod podzolic (Hypogleyic Luvisol, according to the FAO) light loam overlying medium loam. The topsoil thickness varies from 0.20 to 0.25 m. The permeability rate ranges from 1.0 to 2.0 m d-1 in the arable soil layer and from 0.01 to 0.004 m d-1 in the lower layers. The main tile drain collector of the system enters a separate well equipped with a V-notch (Thompson) weir. The notch angle is 30°. The water level in the well is measured continuously. Drainage outflow is determined based on the known head-discharge relationship for the V-notch weir.

In addition to drainage outflow, meteorological variables (average annual precipitation and average air temperature) during the period 1977 to 2016 were analyzed. The data were obtained from the Kaunas Meteorological Station, which is the nearest station to the investigated site (a distance of 0.5 km). The study period was divided into four seasons: winter (December to February), spring (March to May), summer (June to August), and autumn (September to November).

Effects of Controlled Drainage

Experiments with controlled drainage were carried out at the Lipliunai site (55° 19' N, 23° 50' E), which is situated in the Middle Lithuanian Lowland, in the catchment of the Graisupis stream (with an area of 16.6 km2). The study area was drained in 1960, and tiles were installed at a depth of 0.90 to 1.10 m, with a drain spacing of 20 to 24 m. Tile drains with a diameter of 40 mm were placed in a regular herringbone pattern. In the 10.3 ha drained area, reconstruction was carried out in 1999. After installing a drainage manhole at the junction of two collectors, two separate systems were arranged: a 4.9 ha free conventional drainage (CD) system and a 5.4 ha controlled drainage (CWD) system. In the CWD area, a water level control device with a riser pipe and a hand-operated rigid flap door was installed at the outlet of the drainage collector in the manhole. The groundwater table rose to a maximum of 68 cm above the tiles. The area impacted by the groundwater table rise represented approximately 52% of the CWD treatment plot. The average surface slope in the area is 0.5% and increases to 1.0% to 1.3% in ascents (Ramoška and Morkunas, 2006; Ramoska et al., 2011).

The soil at the study site is a non-acidic Endocalcari-Endohypogleyic Cambisol (CMg-n-w-can) (FAO, 2015). In the CD area, the soil is sandy loam and loamy sand in the 0-30 cm layer and sandy loam in the 31-70 cm layer. In the CWD area, the soil is loamy sand in the upper layer (0-30 cm) and sandy loam and loamy sand in the subsoil (31-70 cm).

Soil agrochemical properties (NH4-N, NO3-N, and plant-available P2O5) were measured in the topsoil (0-30 cm) and in the subsoil (31-70 cm) during the drainage flow period. For each plot, three subsamples were pooled into one sample and analyzed for the topsoil, while four subsamples were pooled into one sample and analyzed for the subsoil. Only mineral fertilizers were applied on the study fields. The cropping sequences and fertilization rates are presented in table 1.

Effects of Additives in Trench Backfill

The research on the effects of additives in the drainage trench backfills was carried out during 1990-1993 on pig slurry irrigated field at the Juodkiškis site (55° 17' N, 24° 1' E). The research area is situated in the catchment of the Obelis stream (with an area of 673.8 km2) in the Middle Lithuanian Lowland. The slurry applied to the field was supplied by a pig-breeding farm. Liquid pig manure is stored on the farm in reservoirs, where it stratifies into solid and liquid layers. During the warm season, the content of the liquid layer was used to irrigate perennial grasses (i.e., the slurry was spread on the soil surface using a high-pressure volume sprinkler system). The ratio of water to solids in the slurry was 1.25:1.0. The soil at the study site is a non-acidic Endocalcari Endohypogleyic Cambisol (CMg-n-w-can) sandy loam (FAO, 2015). In the 1.22 ha area, separate tile drains were installed, and the trench backfill was mixed with various components to improve the purification of the pig slurry. The trench soil was mixed with 3.6% woodchips in the first treatment, 0.3% chopped straw in the second treatment, and 1.2% lime (active matter) in the third treatment (% based on trench volume weight). The trenches in the 1.62 ha control treatment were filled with the spoil. The installed tile drains were made of 50 mm diameter tubes, while the collectors were 75 mm diameter non-perforated, high-density polyethylene pipe that did not contribute to the drainage water. In all treatments, the drain spacing was 15 m, and the drainage depth was 1.4 m. The experimental treatments were installed in three replications (fig. 2).

Table 1. Cropping sequences and fertilization for the controlled drainage experiment.
CropFertilization DateFertilizer TypeFertilizer Amount
(kg ha-1)
Nutrient Load (kg ha-1)
NitrogenPhosphorus
Winter wheat19 Sept. 2014Blended fertilizers (N5P15K25)25012.537.5
Winter wheat3 and 10 Apr. 2015Ammonium nitrate (N34.4)250 and 20086 and 68.8-
Winter wheat2 May 2015Blended fertilizers (N20P20K20)40.80.8
Winter wheat17 July 2015Urea (N46.4)52.3-
Winter wheat22 Sept. 2015Blended fertilizers (N4P16K34)2501040
Winter wheat5 Apr. 2016Ammonium nitrate (N34.4)20068.8-
Winter wheat12 May 2016Ammonium nitrate (N34.4)20068.8-
Winter rape19 Aug. 2016Blended fertilizers (N4P16K34)2501040
(a)(b)
Figure 2. (a) Layout of the Juodkiškis study site and (b) cross-sections of the trenches used to investigate the treatment effects of additives in drainage trench backfill: 1 = trench soil mixed with woodchips, 2 = trench soil mixed with chopped straw, 3 = trench soil mixed with lime, 4 = trench filled with spoil (control treatment), 5 = polyethylene film, 6 = irrigation net, and 7 = fiberglass drain tube.

The trench in each treatment was filled using a drainage trencher, and each linear meter contained 2.3 kg of chopped straw, 0.04 m3 of technological woodchips, or 11 kg of lime. These materials were blended with the excavated trench soil in the correct proportions by the drainage trencher transporter. To prevent one drainage system from impacting another, polyethylene film screens were installed between the systems to the drainage depth. The irrigation schedule was determined by the amount of nutrients in the slurry. During the study period, the slurry contained 0.401 kg TN m-3 and 0.045 kg TP m-3 on average. The irrigation rate of 600 m3 ha-1 of slurry was distributed equally over the area in three applications: 200 m3 ha-1 was applied in spring, at the start of perennial grass growth, and after mowing. The average TN and TP loads applied on the research field were 240.6 and 27 kg, respectively. Drainage outflow samples were taken at the drainage outlet on slurry application days and three and five days later.

Efficiency of In-Ditch Filter Materials

The site for investigation of in-ditch N and P filter materials was in the Middle Lithuanian Lowland (Nevežis River basin; 55° 33' 49? N, 24° 2' 57? E). The proportion of agricultural land in this district is approximately 57.3%, of which 92.4% is arable land and only 6% is used as grassland and pastures. In this drained agricultural land, most of the streams are modified, i.e., converted into ditches. One such stream, the Aukupis ditch (with a catchment area of 3.6 km2 and specific runoff of 7.9 L s-1 km2), was chosen for filter well installation. The experiment aimed to achieve contaminated stream water purification in conditions that were close to reality. The Aukupis ditch collects tile drainage and surface water from slurry manured fields and from a pig-breeding enterprise. The average P2O5 concentration in the plow layer of the soil (0-20 cm) ranges from medium to very high (120 to 207 mg kg-1).

Three filter wells were installed in the side slope of the ditch (fig. 3). Each structure consisted of a 1.0 m diameter concrete well in which filter media were tested. Each well was filled with 1.0 m3 of filter material. Ditch water was supplied to the filter bottom and was filtered during upward flow through the filter material, which was permanently flooded.

Figure 3. Schematic diagram of a filter well.

Two phases of research were carried out. The first phase lasted from December 2012 to May 2014. Three filter materials were tested: Polonite, Filtralite-P, and granulated blast furnace slag. Because the tested filter materials showed little nitrogen removal efficiency, three other materials were selected for a second phase of research: bark, charcoal, and dolomite chips. The second phase lasted from July 2016 to December 2016.

Polonite is a calcium silicate-based material with a grain size of 2 to 5.6 mm. It is used for absorption and recycling of phosphorus in small treatment plants, but it can also be used for larger-scale treatments, such as in filters for farm land, municipal/large treatment plants, animal farms, and industrial water treatment. Filtralite-P is a fine-grained filter material, with a large surface area and a high capacity for phosphorus removal. This product is formed by heating and firing natural plastic marine clay in a rotary kiln at temperatures up to 1150°C. The grain size is 0.5 to 4 mm. Slag is a byproduct of steelmaking. It is largely limestone or dolomite that has absorbed phosphate from iron ore during smelting. The product mainly contains silicon, calcium, and magne-sium. The grain size is 0 to 4 mm. The deciduous tree bark was derived mostly from oak. The thickness of the bark varied from 5 to 10 mm with pieces 40 mm × 20 mm in size. Charcoal is a porous black solid consisting of an amorphous form of carbon obtained by heating wood in the absence of air. The particle size varies from 10 × 20 × 20 mm to 10 × 40 × 20 mm. The dolomite chips had a grain size of 2.0 to 5.0 mm.

Lab-Scale Denitrifying Bioreactors

Three rectangular (130 cm length, 35 cm width, and 70 cm depth) denitrification bioreactors (0.32 m3 volume each) were constructed at the Drainage Laboratory of Aleksandras Stulginskis University in Kaunas, Lithuania. These steel containers were connected to two plastic water tanks (1.0 m3 volume each) by flexible hoses. The inside surface of the bioreactors was painted with exterior house paint and coated with non-toxic silicone sealant. Each bioreactor was filled with different types of woodchips made from local raw materials: deciduous, coniferous, and mixed trees woodchips. Alder tree scraps dominated the deciduous woodchips, spruce tree chips dominated the coniferous woodchips, and pine tree chips dominated the mixed woodchips. The prevailing (at 50% of the cumulative distribution) particle diameter of the woodchips varied from 1.1 to 3.0 cm. All three bioreactors were filled with woodchips to a depth of 70 cm, and a saturation level of 65 cm was maintained. A polyethylene liner was folded over the top of each bioreactor to seal it from the soil, and a mound with a 5 cm thickness was formed and sown with grass.

The woodchip porosity was determined using the standard porosity determination procedure described by Christianson et al. (2010). The analysis revealed that the conifer woodchip porosity was 54%, the deciduous woodchip porosity was 53%, and the mixed woodchip porosity was 53%. A schematic representation of the three laboratory-scale woodchip bioreactors is shown in figure 4.

The NO3-N removal efficiency tests of the three woodchip bioreactors started on June 5, 2015, and the results presented in this article cover the period until October 5, 2016. The measurements were performed at various irregular time intervals by applying the same sampling procedures. Due to some technical problems, fewer water samples were taken from the bioreactors filled with coniferous and mixed woodchips than from the bioreactor with deciduous woodchips (N = 124, 124, and 155, respectively).

The bioreactors were fed nitrate (via the addition of NaNO3 to the water tanks) during the study period at concentrations ranging from 1.0 to 9.10 mg N L-1. These concentrations are typical (82% of the cumulative frequency) of the range of NO3-N values observed in drainage water under field conditions. The water from the tanks was supplied to each bioreactor by gravity. The flow velocity was determined by the difference in hydraulic head (3.5 m maximum) between the water levels in the tanks and in the bioreactors. The inflow and outflow rates were adjusted manually with valves. The valves were also used to achieve steady-state flow conditions at various time intervals and to achieve different water retention times.

The NO3-N concentrations in the bioreactors were determined using a MaxiDirect Photometer MD600 (Lovibond, Amesbury, U.K.) with powder reagents. The dissolved oxygen (DO) contents and water temperatures at the inlet, outlet, and inside the sampling wells were measured with a portable multimeter (18.28, Eijkelkamp, Giesbeek, The Netherlands), and pH values were measured with a portable meter (WTW pH340i, Trescal, Paris, France).

Figure 4. Schematic of the three laboratory-scale woodchip bioreactors: 1 = water tank, 2 = water mixing chamber, and 3 = bioreactors (A = with deciduous woodchips, B = with coniferous woodchips, and C = with mixed woodchips).

The NO3-N removal rate in the bioreactors was calculated using the following equation:

(1)

where

E = NO3-N removal rate (g N m-3 h-1)

Cin(NO3-N) and Cout(NO3-N) = NO3-N inflow and outflow concentrations (g m-3), respectively, during an event (defined as the period between two water sample collections when the water flow rate through the bioreactor is larger than zero)

HRT = hydraulic retention time (h), which was determined as follows:

(2)

where

? = porosity of woodchips

V = bioreactor volume (m3)

Q = flow rate (m3 h-1).

Statistical Analysis

To determine the reliability of differences between the investigated experiments, Student’s t-test was used. Additionally, the rank-based Mann-Kendall trend analysis and Sen’s slope estimator were applied to meteorological and drainage flow time series data to determine if significant trends (p < 0.05) were present among the variables. Moreover, the Shapiro-Wilk test was conducted to assess the normality of the NO3-N removal efficiency data from different bioreactors. Because the assumption of normality was not met, a non-parametric Kruskal-Wallis test was performed to determine whether the three groups of measurement data with different woodchip fillers were significantly different. Statistical analyses were performed by applying Statistica (ver. 7.0), PAST (ver. 2.0), and MAKESENS (ver. 1.0) software packages.

Results and Discussion

The analysis revealed that the average annual precipitation during 1977 to 2016 was 642 mm, or 1.4% higher than the climate normal (CN, defined as the three-decade 1981-2010 average). At the study site, the driest year occurred in 1992 (69% lower than the CN), and the wettest year occurred in 2010 (33.8% higher than the CN). The annual air temperatures for the 1977-2016 period (with an average of 7.0°C) showed increasingly variable patterns (fig. 5). The highest air temperature was observed in 2015 (8.7°C, or 1.7°C higher than the CN), and the lowest air temperature was observed in 1990 (5.0°C, or 0.3°C lower than the CN).

The dynamics of the annual drainage outflow during the study period exhibited a statistically insignificant (p > 0.05) decreasing trend (fig. 5). However, the analysis showed that the traditional seasonal outflow pattern remained (fig. 6), i.e., the outflow was highest in spring and lowest in summer.

When comparing the changes in seasonal drainage outflow during the last four decades, it becomes obvious that the outflow increased during winter months (figs. 6 and 7). It is likely that higher winter outflows are related to higher winter air temperatures (a statistically significant increasing pattern at p < 0.100), resulting in earlier and faster snowmelt, less snow accumulation, and less water storage capacity in the snowpack.

According to the Mann-Kendall test, a statistically significant increasing trend in drainage outflow was detected in winter, and a statistically significant decreasing trend was detected in spring. The Sen’s estimator for the true slope of the linear trend confirmed the Mann-Kendall test results: the changes in winter drainage outflow have a positive slope, while the changes in spring drainage outflow have a negative slope (table 2).

Figure 5. Variations in annual precipitation, drainage outflow, and temperature.
Figure 6. Seasonal drainage outflow by decade for 1977 to 2016.
Figure 7. Variations in precipitation, drainage outflow, and air temperature in winter.
Table 2. Trend statistics for seasonal drainage outflow in 1977-2016.
SeasonMann-Kendall Trend[a]Sen’s Slope Estimate[a]
Test ZSSmin95Smax95
Winter2.06[b]0.41700.835
Spring-1.99[b]-0.500-1.0830
Summer-0.730.000-0.0430
Autumn0.300.000-0.5710.297

    [a]S is slope, and Smin95 and Smax95 are the lower and upper limits, respectively, of the 95% confidence interval of S.

    [b] Level of significance is less or equal to 0.05.

Effects of Controlled Drainage on N and P Losses from the Soil

In both treatments (CD and CWD), the inorganic nitrogen (NO3-N + NH4-N) and plant-available P2O5 contents differed between soil layers, i.e., the nutrient contents were higher in the 0-30 cm soil layer than in the 31-70 cm layer (table 3). In the CD area, the amount of plant-available P2O5 in the soil was up to 1.8 times (0-30 cm soil layer) and 3.5 times (31-70 cm) greater than in the CWD area. This dif-ference (tact = 3.65 to 5.83 > t0.05 = 1.73) was statistically significant. Similarly, the amount of inorganic N fluctuated. The largest fluctuations in inorganic N content were found in the topsoil layer, and the largest fluctuations in plant-available P2O5 were found in the subsoil in the CWD area.

Table 3. Inorganic N and available P2O5 in the soil (mg kg-1).[a]
Soil Thickness
(cm)
CDCWDtact
Inorganic N
0-301.25 - 17.5
8.02
0.76 - 17.18
6.30
0.65
31-700.69 - 10.47
5.40
0.89 - 9.40
4.44
0.63
tact1.210.88
Available P2O5
0-30115 - 158
139
84 - 132
103
5.83
31-7080 - 154
125
40 - 167
76
3.65
tact1.642.14

    [a] Values are (minimum - maximum) / mean; t0.05 = 1.73.

Figure 8. Percentages of nitrogen compounds forming the inorganic N content in the soil.
Figure 9. Inorganic N and TP concentrations in tile drainage water.

Higher inorganic N and plant-available P2O5 in the upper layers of the soil were determined by the organic matter content, fertilizer, and soil texture. In the upper soil layer, the soil humus content was average to high (1.79% to 4.32%) (Guzys and Miseviciene, 2015). Additionally, the top layer of soil was more affected by the mineral fertilizer. The CD area was dominated by sandy loam, while the CWD area was dominated by loamy sand. Higher proportions of colloids and silt particles (<0.001 mm) in the soil are associated with higher soil cation sorption capacities. As the soil organic matter increases, the sorption capacity increases as well (Vaišvila, 1999; King et al., 2015). Ulén and Mattsson (2003) stated that, when aerobic conditions improve, phosphorous quickly transforms into forms available to plants.

In the soil of the CD area, inorganic N consisted of 87% NO3-N and 13% NH4-N. In the CWD area, inorganic N consisted of 85% NO3-N and 15% NH4-N (fig. 8). In both treatments, the highest NO3-N percentage occurred in winter when the average air temperature was positive (five-day average air temperature was 2.6°C in January 2015). Additionally, a large amount of NO3-N was measured in autumn (September 2016), when the post-harvest period created favorable conditions for mineralization of plant residue (five-day average air temperature was 15.8°C).

Higher NO3-N and TP concentrations in winter were related to meteorological conditions. The literature (Tripolskaja et al., 2002; Jansons et al., 2009) suggests that the abundance and activity of some groups of microorganisms change in the autumn-winter period in association with positive temperatures. During the cold period, ammonifying nitric bacteria are more active. Due to the activity of these microorganisms in the soil, the content of inorganic nitrogen changes, and the soil receives a constant supply of nitrate. Wesström and Messing (2007) stated that peak P loads occurred in late winter, and P concentrations were positively correlated with soil temperature.

During the study period, the inorganic N concentration (55% of all measurements) was greater in CWD water (fig. 9). In both drainage systems, up to 99.9% of the inor-ganic N content was NO3-N. The largest difference (1.8 times) in inorganic N concentration between CD and CWD occurred in March 2016.

The NO3-N concentration (55% of all measurements) and the NH4-N concentration (72% of all measurements) were higher in the CWD treatment. However, these differences were insignificant (NO3-N: tact = 1.07 < t0.05 = 1.68; NH4-N: tact = 0.74 < t0.05 = 1.68). In both drainage systems, the inorganic N concentration increased when winter wheat, grown in the study area, was fertilized (April to May 2015) with a higher rate (156 kg N ha-1) of mineral nitrogen fertilizer.

Significant differences (tact = 1.80 > t0.05 = 1.69) in TP concentrations between the two areas were confirmed, with higher concentrations (0.001 to 0.045 mg L-1) for the CWD treatment (fig. 9). Compared with the TP concentrations of the CD treatment, the TP concentrations of the CWD treatment were 3.2 times higher during the spring flood (in March 2016). Additionally, these concentrations were 2.3 and 1.9 times higher in the autumn of 2014 and 2016, respectively. On average, 90% and 88% of the TP was PO4-P in CD and CWD, respectively.

The inorganic N and TP concentrations increased when the water table rose closer to the topsoil layer, which contained more nutrients, on February 7, 2016, and March 1, 2016. Higher total phosphorus concentrations also occurred in autumn. The application of fertilizer is likely responsible for these increases in concentration. King et al. (2015) reported that large amounts of particulate P were measured in tile discharge resulting from the first precipitation event after fertilizer application.

The reason why lower inorganic N and TP concentrations were commonly found in the CD system relative to the CWD system might be related to the fact that the former area contained predominantly lighter textured soils (sandy loam), which are more susceptible to leaching of nutrients that are not taken up by plants. The water in the CD system flowed to the drains from deeper layers, which contain fewer nitrogen compounds. In a waterlogged soil, a variety of biochemical processes can begin. In the pores of the waterlogged soil, the oxygen content decreases. A reduction in the oxygen content can create anaerobic conditions, leading to denitrification, methanogenesis, and the reduction of manganese, iron, and sulfate, resulting in a lowering of the soil redox potential (Beumer et al., 2007). Changes in the pH and redox reactions change the bio-availability of nutrients in the soil (Violante et al., 2010; Lynch et al., 2014). The NH4-N and PO4-P levels increase when the soil conditions become anaerobic (Beumer et al., 2007; Loeb et al., 2007).

At the study site, CWD affected both the volume of drainage outflow as well as the inorganic N and TP loads. The monthly drainage outflow was 55% to 80% lower in the CWD area as compared to the CD area. Consequently, inorganic N and TP loads in the CWD area were lower by 42% to 77% and by 34% to 72%, respectively, compared to the loads in the CD area. The reduced loads are related to the difference in drainage outflow. Similar results were obtained by other researchers. Šaulys et al. (2011) reported that N and P concentrations in tile drainage water were not directly affected by outflow control conditions. However, the control of outflow resulted in lower nutrient loads. The volume of controlled drainage outflow in that study was 21% to 24% less than that of conventional drainage outflow. The results of other studies (Evans et al., 1992; Wesström and Messing, 2007) also showed that controlled drainage reduced TN losses at the field edge by an average of 45% and reduced TP losses by 35% when drain outflow was reduced by 30%.

Effects of Additives in Trench Backfills

All additives mixed in the drainage trench backfills reduced NO3-N concentrations in the tile drainage water: woodchips reduced the concentrations by 78% (tact = |-7.12| > t0.05 = 2.04), chopped straw reduced the concentrations by 69% (tact = |-5.61| > t0.05 = 2.07), and lime reduced the concentrations by 52% (tact = |-4.30| > t0.05 = 2.07). The best NO3-N purification effect (2.63 mg L-1) was obtained when the drainage trench backfill soil was mixed with woodchips (fig. 10).

Figure 10. NO3-N and PO4-P concentrations in tile drainage water using different trench backfills (AVG = average; SD = standard deviation).

It has been reported that woodchips effectively reduce NO3-N in tile drainage water (Woli et al., 2010; Christianson et al., 2011). Studies have shown that, using woodchips, NO3-N concentrations were reduced by 60% from 22.1 to 8.8 mg L-1 (Jaynes et al., 2008) and by 94% from 9.4 to 0.58 mg L-1 (Chichlowski, 2014). Our research confirms this finding; woodchips achieved the greatest reduction of all the additives, reducing NO3-N concentrations by 78%. A higher purification effect is obtained at higher inflow NO3-N concentrations and at higher temperatures (Hoover et al., 2016). According to Porter et al. (2015), the composition of the denitrifying microbial community, which is responsible for the reduction of pollutants in tile drainage water, can be influenced by seasonal humidity changes, the depth of the backfill, and the temperature. Therefore, it is necessary to look for ways to reduce the pollution from agricultural land at low temperatures, at high flow rates, and with aging backfill (Addy et al., 2016). As slurry filters through the soil, its inorganic compounds react with mineral compounds in the soil. Consequently, the contaminants that appear in the soil aggregate become surrounded by microorganisms; this is the start of the purification process for biochemical contaminants. Improvements in the infiltration capacity of the trench fill improve the water and air regime, which in turn improves conditions for the development of microorganisms, which are directly involved in the mineralization of slurry organic matter. Lime application increases the number of ammonifying and nitrifying microorganisms, which increases the soil nutrient reserves and ensures good conditions for plant growth (Bambara and Ndakidemi, 2010; Moreira and Kumar Fageria, 2010; Jokubauskaite et al., 2015; Narendrula-Kotha and Nkongolo, 2017). In our research, the NO3-N concentrations in drainage water decreased by 52% with the application of lime compared to the control treatment. Bastiene et al. (2012) also found that average concentrations of NO3-N in drainage water were 15% higher for the control treatment compared to the treatment with lime in a silty clay loam soil.

The incorporation of chopped straw into the drainage trench backfill significantly reduced the concentration of NO3-N in tile drainage water. Compared with the control treatment, the concentration decreased by 69%. After adding straw to the drainage trench soil, the degradation process begins. Soil nitrogen is consumed by the microorganisms involved in this process, and very little nitrogen is leached into the tile drainage, unless more precipitation occurs (Romanovskaja and Tripolskaja, 2003). Nitrogen leaching decreased by 47% when 3 t ha-1 of dry material was added to the soil (Powlson et al., 1985). However, degradation of the straw increases the amount of mobile phosphorus in the soil (Jenkyn et al., 2001). Therefore, our research showed an increase in PO4-P concentrations of 22% (0.078 mg L-1) in tile drainage water compared to the control treatment.

According to Andersson et al. (2016), a very effective tool for reducing phosphorus concentrations in drainage water is mixing lime into the drainage trench backfill, which lowered the phosphorus leaching to 49% versus 51% without the addition of lime. The addition of lime to the sandy soil also reduced phosphorus leaching compared to using only natural soil. When clayey soil as drainage trench backfill was mixed with 0.6% CaO, the TP and PO4-P concentrations decreased by 50% and 64.4%, respectively, relative to the control treatment (Šaulys and Bastiene, 2008; Bastiene et al., 2012). As stated by Bergström et al. (2007), the soil pH decreases during the slurry nitrification process. This results in a greater solubility of phosphorus due to the amount of ammonium ions, which is high in manure and slurry. According to Murphy and Stevens (2010), soil acidity neutralization with lime and gypsum to a pH of 6.5 results in decreasing molybdate-reactive P solubility by 14% to 56% and decreasing organic P by 10% to 53%.

The research data for the sandy loam soil at the Juodkiškis study site show that lime mixed into the trench backfill reduced the PO4-P concentrations in drainage water by 39% (tact = |-2.73| > t0.05 = 2.08). The use of other additives in the trench backfill did not have any larger impacts on the PO4-P concentrations because, after statistical analysis of the data, the differences between the treatments were found to be non-significant. However, woodchips and chopped straw increased the PO4-P concentrations in tile drainage water by 11% (tact = 0.27 < t0.05 = 2.10) and 22% (tact = 0.52 < t0.05 = 2.11), respectively, compared with the control treatment. Chichlowski (2014) found similar results; woodchips increased the phosphorus concentrations in tile drainage water from 0.02 to 1.29 mg L-1.

At the Juodkiškis study site, all additives in the trench backfill increased the drainage outflow: by 44.2% with chips, by 21.4% with straw, and by 68.7% with lime. According to Bastiene et al. (2012), soil filtration properties were particularly improved by various drainage trench backfills installed in clay soils. It was found that drainage outflow can be 5 to 24 times higher compared with the control treatment in drainage trenches with various backfills. The same study determined that soil porosity was 15% to 20% higher than in the soil in which lime was not used.

A significant impact on drainage efficiency was determined (Šaulys and Bastiene, 2006) when the trench backfill was mixed with lime (0.40% of soil mass). It was found that such backfill can flush drainage water 55 times faster than the control treatment. Other literature (Velykis et al., 2003) suggests that straw significantly increased water permeability in clay soil, while in sandy loam soil, straw increased water permeability by 1.2 times compared with the control treatment.

To compare the efficiency of the additives, the NO3-N and PO4-P losses were calculated. Even though the drainage outflow was lowest (54.9 mm) in the control treatment, the high outflow concentration of NO3-N (11.83 mg L-1) resulted in the highest (5.7 kg ha-1) N losses. The loss of NO3-N was reduced the most (66.7% (tact = |-4.36| > t0.05 = 2.01) by mixing the trench backfill with woodchips. This resulted in the lowest outflow concentration (2.63 mg L-1), even though the drainage outflow was 44.2% higher compared with the control treatment. After mixing the trench backfill with chopped straw, NO3-N losses decreased by 64.9% (tact = |-4.36| > t0.05 = 2.01) compared with the control treatment. This was influenced by the lowest drainage outflow compared to all the treatments (66.7 mm), along with low concentration of NO3-N (3.69 mg L-1). The mixing of lime in the trench soil significantly increased the drainage outflow (92.7 mm); however, due to the low NO3-N concentration (5.62 mg L-1), the loss of NO3-N was only 1.8% (tact = 0.27 < t0.05 = 1.99) lower than in the control treatment (5.6 and 5.7 kg ha-1, respectively). Statistically, the difference was not significant. Literature data confirm the results obtained in this study. According to Šaulys et al. (2011), NO3-N losses in tile drainage water did not increase by improving trench permeability with lime.

Statistical analysis of PO4-P losses did not show any significant differences between the control treatment and treatments in which the trench backfill soil was mixed with woodchips (tact = 1.26 < t0.05 = 2.03), chopped straw (tact = |1.68| < t0.05 = 2.02), or lime (tact = |-0.85| < t0.05 = 1.99). However, some extenuating effects can be observed. The results indicate that the trench backfill with lime reduced PO4-P loss by 19.6%, compared with the control treatment. This was a result of 39.1% lower PO4-P concentration compared to the control treatment. Other research also revealed that addition of lime to drainage trench backfills reduced TP leaching by 2.5 times (Šaulys et al., 2011). Meanwhile, the trench back-fills with woodchips and chopped straw increased PO4-P losses by 47.8% and 13.0%, respectively, compared with the control treatment. The loss of PO4-P was mainly affected by the high drainage outflow in the treatment with woodchips (79.2 mm) and by the PO4-P concentration (0.078 mg L-1) in the treatment with chopped straw (fig. 11).

Figure 11. NO3-N and PO4-P losses in tile drainage water using different trench backfills (AVG = average; SD = standard deviation).

The results from the Juodkiškis site showed that the pig slurry purification efficiency was not always sufficient when mixing the trench backfill soil with various materials. In all treatments, the average NO3-N concentrations in drainage water were determined to be low, except in the control (11.83 mg L-1). However, the peak values of NO3-N were observed in the lime and chopped straw treatments (13.13 and 12.52 mg L-1, respectively). The peak PO4-P concentrations in drainage water were observed in the treatments with woodchips and chopped straw (0.417 and 0.411 mg L-1, respectively). This study has shown that NO3-N concentrations were most reduced by a soil mixture with woodchips and that the PO4-P concentrations were most reduced by a soil mixture with lime.

This study also found that trench backfills with woodchips and chopped straw reduced NO3-N losses the most, by 66.7% and 64.9%, respectively. Trench backfills with lime did not show any greater effect on NO3-N loss. The loss of PO4-P was reduced by 19.6% using trench backfill with lime, while the backfills with chips and straw increased the PO4-P losses by 47.8% and 13%, respectively.

In-Ditch Nitrogen and Phosphorus Filters

The TN and TP retention capacity of different filtration materials in the filter wells was evaluated by periodic (once per week) water sampling. The water discharge from the filter wells was measured at the same time. Water samples from the ditch were also collected and analyzed. Because the ditch collects tile drainage and surface water from manure-fertilized plots and from a pig-breeding farm located upstream, it is constantly polluted with nutrients. Therefore, the average nutrient concentrations in the ditch water during the study period were 33.1 ±13.6 mg TN L-1 and 0.44 ±0.29 mg TP L-1. The filters were sometimes frozen in winter.

To assess the TP retention efficiency of different filter materials, analysis of the retention dynamics was performed. The cumulative retention of TP by the 1.0 m3 filters in the first phase of research is presented in figure 12. In summer, the retention capacity decreased, and P leaching began from the filters. When the weather became colder, the capacity of the filters to retain phosphorus increased. However, at the end of the monitoring period (May 2014), phosphorus again began to be released from the filters.

The discharge and concentration of the inlet water were unstable. To assess the TP retention efficiency of different materials, the TP retention with respect to the load, excluding periods when TP was leached from the filters, was calculated. The calculated dependencies are presented in figure 13.

In the first stage of the research, all three tested materials showed poor TN retention efficiency. The cumulative retention values for TN and TP during the second stage of research are presented in figure 14. As in the first stage, there were periods in which N and P leaching occurred from the filters, resulting in higher concentrations at the filter outlet than at the inlet.

The TN retention capacity with respect to the load was determined only for bark, and the TP retention capacity with respect to the load was determined only for dolomite chips. The relationships are shown in figure 15. For other filter materials, the retention capacity was insignificant.

Similar trends associated with N and P leaching from the filters were observed in the first and second stages of the research. In both cases, leaching occurred during the months of July and August. This process can be related to the fact that high N and P concentrations in ditch water stimulate intense development of algae. The algae clog the filters; therefore, the filters must be flushed. After flushing, the filter permeability increases; however, the N and P leaching continues for some time. After an assessment of leaching, the overall balance of P retention showed that leaching was approximately 10% to 15% of the retained amount.

Figure 12. Cumulative retention of TP in filters of 1.0 m3 volume in the first phase of research.

The differing porosity of the filter materials influenced the permeability of the filters. The filters with Polonite, Fil-tralite-P, and slag filtered 1776, 1709, and 1190 m3 of water, respectively. Because the P retention capacity of filters depends on the P load, according to the relationships presented in figures 9 and 11, the TP retention was calculated based on an average load of 1.5 g d-1 and a 1.0 m3 filter volume. The results revealed that Polonite, slag, Filtralite-P, and dolomite chips retained 0.7, 0.56, 0.36, and 0.28 g TP d-1, respectively. Except for bark, the other tested materials showed poor TN retention. With an average load of 120 g d-1 in the 1 m3 bark filter, the TN retention was 8.5 g d-1, i.e., the retention efficiency was 7.1%.

(a)
(b)
(c)
Figure 13. TP retention in the filters (volume of 1.0 m3) with respect to TP load in the first stage of research: (a) Polonite, (b) Filtralite-P, and (c) slag. The influent P concentration was: 0.1 < Co < 0.8 mg L-1.

Nitrate-Nitrogen Removal in Lab-Scale Denitrification Bioreactors

Many factors affect NO3-N removal in bioreactors. The results obtained in this study showed that biological activity facilitated reduction of oxygen and further nitrate transfor-mation in the bioreactors. Figure 16 shows the dynamics of water temperature, DO concentration, and NO3-N in the inflows and outflows of the bioreactors. The results are summarized in table 4.

(a)
(b)
Figure 14. Cumulative retention of (a) TN and (b) TP in filters (volume of 1.0 m3) in the second phase of research.

The water temperature measured at the inlet and outlet ranged from 13.9°C to 26.1°C. This range of temperatures is comparable to tile drainage temperatures that may occur in the summer in Lithuania. Typically, the water temperature during the growing season ranges between 7.0°C and 16.0°C, although temperatures above 20°C occasionally occur (Sakalauskas, 1986). However, most of the flow that occurs from November to April (the critical period for NO3-N leaching) has a relatively low temperature. Therefore, the effect of low temperatures in conjunction with practices to enhance nitrate removal should be investigated in further research. During the tests, the temperature was higher (20°C to 26°C) at the initial (June 5 to September 3, 2015) and later (May 2 to October 5, 2016) stages and was lower (14°C to 17°C) during the middle (September 4, 2015, to May 1, 2016) stage of the tests. The Kruskal-Wallis tests revealed statistically significant differences (p < 0.05) in the inflow water temperatures between the time periods. Moreover, the inflow NO3-N concentrations during the same time periods were also statistically different. However, the patterns of change in the inflow water temperature and NO3-N concentration were similar for all three bioreactors. Considering the entire test period, the patterns of change in the inflow temperature and NO3-N concentration among the three bioreactors did not show any statistically significant differences. The absence of differences in the changes of these two input variables suggests that they were about the same for all three bioreactors. This allowed us to assume that the inflow water temperatures and the inflow NO3-N concentrations during the study period affected NO3-N removal equally in all three bioreactors.

(a)
(b)
Figure 15. Retention of (a) TN for bark and (b) TP for dolomite chips in filters (volume of 1.0 m3) with respect to TN and TP loads in the second phase of research. The nutrient concentrations in the influent were: TN = 14 < Co < 50 mg L-1, and TP = 0.1 < Co < 0.8 mg L-1.

The pH values in the inflow ranged from 5.2 to 8.6, and those of the outflow ranged from 5.0 to 8.3. The DO concentrations at the inlets and outlets ranged from 3.0 to 9.0 mg L-1 and from 0.0 to 3.35 mg L-1, respectively, while the values inside the sampling wells ranged from 0.0 to 4.35 mg L-1 (average of 0.52 mg L-1). The tests also revealed that low oxygen conditions (<4.0 mg L-1) developed in the lower layers of bioreactors (approx. 20 cm above the bottom) within the first 2 or 3 h after the water inflow started and that the minimal DO values (0.00 to 0.50 mg L-1) were reached during the first ten days after startup. The ratio of inflow DO concentrations to outflow DO concentrations ranged from 2.5 to 122. These values imply that the DO in the bioreactors was consumed by heterotrophic bacteria and that carbon was made available to the bacterial community.

NO3-N removal occurred in all bioreactors over the course of the experiment, with average inflow and outflow concentrations of 4.93 and 1.53 mg L-1, respectively. However, the total cumulative volume of water that passed through the bio-reactor filled with coniferous woodchips was 22% to 24% larger than the volumes that passed through the other two bioreactors. This caused differences in the calculated hydraulic retention time (HRT), which ranged from 0.48 to 25.3 h, from 0.63 to 15.3 h, and from 0.70 to 10.6 h for the bioreactors with deciduous, coniferous, and mixed woodchip fillers, respectively. The resulting NO3-N removal efficiency (determined as the difference between the inlet and outlet NO3-N concentrations divided by the inlet concentration) during an event ranged from 8.3% to 90.1% in the deciduous bioreactor, from 16.7% to 88.2% in the coniferous bioreactor, and from 7.7% to 88.6% in the mixed bioreactor. The nitrogen removal rates during events ranged from 0.058 to 6.80 g N m-3 h-1 in the bioreactor with deciduous filler, from 0.14 to 7.25 g N m-3 h-1 in the bioreactor with coniferous filler, and from 0.064 to 6.41 g N m-3 h-1 in the bioreactor with mixed filler.

The observed removal efficiencies are similar to the values obtained in other studies. However, the average NO3-N removal rates determined in this study are in the higher range of removal rates reported in the literature. This was most likely due to the relatively high inlet water temperatures maintained in the tests. Chun et al. (2009) reported NO3-N concentration reductions of 10% to 40% for retention times of less than 5 h and 100% removal at retention times of 20 h. Hoover et al. (2016) found that NO3-N removal efficiency in bioreactors filled with hardwood chips changed from 18% to 78%, and the average load reduction ranged from 0.87 g N m-3 h-1 at 1.7 h HRT to 0.32 g N m-3 h-1 at 7.3 h HRT. Lepine at al. (2016) showed that NO3-N removal efficiency changed from 45% to 71% with corresponding removal rates of 1.60 and 1.33 g N m-3 h-1. Husk et al. (2017) reported that after three years of denitrification with woodchip bioreactors in cold climate conditions (mean annual temperature of 5.6°C), there was a significant reduction (99%) in NO3-N concentration in the effluent water, with an average removal rate of 0.30 g N m-3 h-1 at 4 h HRT. The published results of other studies revealed NO3-N removal rates ranging from 0.02 to 0.33 g N m-3 h-1 (Christianson et al., 2012), from 0.0 to 1.23 g N m-3 h-1 (Youssef et al., 2016), and from 0.10 to 0.92 g N m-3 h-1 (Schipper et al., 2010) for a variety of hydraulic, water temperature, and inflow concentration conditions.

According to Christianson et al. (2012), the type of carbon fill is one of the most important considerations in denitrification systems because the properties of the medium affect factors that influence nitrogen removal. Therefore, there is considerable discussion in the literature about the use and effects of the physical, chemical, and other properties of various fill materials (Schipper and Vojvodic-Vukovic, 1998; Greenan et al., 2006; Jaynes et al., 2008; Cameron and Schipper, 2010; Warneke et al., 2011). However, woody media are preferred due to their cost and longevity. This experiment showed that there was no significant difference (p > 0.050) in NO3-N removal efficiency among all three bioreactors. However, NO3-N removal rates in the bioreactor with coniferous filler were higher compared to the other two bioreactors. According to the Kruskal-Wallis test, there was no statistically significant difference (p > 0.050) between the removal rates in the bioreactors with deciduous and mixed woodchip fillers. However, the test revealed significant dif-

Figure 16. Dynamics of water temperature, dissolved oxygen, and nitrate concentrations in bioreactors with (a) deciduous woodchips filler, (b) coniferous woodchips filler, and (c) mixed woodchips filler.

ferences in the removal rates between the bioreactors filled with coniferous and deciduous woodchips (p = 0.003) and between the bioreactors filled with coniferous and mixed woodchips (p = 0.020). Comparison of the HRTs among the bioreactors revealed similar patterns. The HRTs in the bioreactor with coniferous woodchips were significantly shorter (p = 0.002 and p = 0.030) than the HRTs in the other two bioreactors. Therefore, the difference in HRTs was the main reason for the higher NO3-N removal rates in the bioreactor with coniferous filler.

Table 4. Average parameter values for each bioreactor filled with different type of woodchips.
ParameterWoodchip Material
DeciduousConiferousMixed
Inflow
Temperature (°C)19.119.119.1
pH8.08.08.0
DO (mg L-1)5.75.65.7
NO3-N (mg L-1)5.04.94.9
Outflow
Temperature (°C)18.918.718.9
pH7.16.87.0
DO (mg L-1)0.850.700.80
NO3-N (mg L-1)1.601.501.50
HRT (h)3.12.73.0
NO3-N removal efficiency (%)65.768.366.7
N removal rate (g m-3 h-1)1.491.821.50

Generally, higher NO3-N removal efficiency in longer HRT treatments indicates that the denitrification process is N-limited, whereas higher NO3-N removal rates for shorter HRTs are governed by reaction kinetics (Woli et al., 2010; Christianson et al., 2012; Youssef et al., 2016; Lepine et al., 2016; Hoover et al., 2016). According to Schipper et al. (2010), variations in NO3-N removal rates are predominantly attributed to operating temperatures and NO3-N inputs. Most often, removal rates increase with increased temperature in bioreactors where NO3-N is not fully depleted. Because the temperature change patterns among all bioreactors in this study did not show any significant differences, it is likely that higher NO3-N removal rates are limited by factors other than inflow water temperature. Under these circumstances, the effect of NO3-N input load (i.e., a transition to first-order kinetics) is the most probable reason for the difference in NO3-N removal rates.

The HRT directly affects the NO3-N removal rate; lower HRT indicates a higher flow rate of water passing through the bioreactor. When the flow rate increases, the NO3-N input load also increases. Therefore, in this experiment, when the three types of woodchips were compared, the nitrogen removal efficiency did not show any significant differences among the three carbon media. However, the higher NO3-N removal rates in the bioreactor with coniferous filler were caused by higher flow rates with subsequently shorter HRTs. Consequently, more water was treated, contributing to higher NO3-N removal rates. The results from this short-term study imply that woodchip media are capable of achieving higher nitrogen removal rates through higher flow rates. This confirms the findings of Christianson et al. (2010), Cameron and Schipper (2012), Hoover et al. (2016), and others and indicates the importance of optimization and proper evaluation of the effects of HRT when designing denitrifying woodchip bioreactors. The HRT should not be too long; otherwise, the nitrogen removal rate will decline.

Conclusions

In Lithuania, artificial drainage is a common agricultural practice. It is used in 87% of the agricultural land area. Over the past four decades (1977-2016), a statistically significant increase in drainage flow has been detected in winter (December to February), and a statistically significant decrease in drainage flow has been detected in spring (March to May). Changes in the other seasons and in the annual drainage flow were not significant. The increased winter flow has forced researchers to pursue new solutions to the increased nutrient runoff problem.

The results from experiments with controlled drainage systems showed much promise. Controlled drainage reduced inorganic N and TP loads while reducing drainage outflow. Monthly drainage flow was 55% to 80% lower with controlled drainage compared to conventional drainage. Consequently, lower loads of inorganic N (by 42% to 77%) and TP (by 34% to 72%) were observed in the controlled drainage system. The reduced loads are related to the differences in drainage outflow.

Woodchips, chopped straw, and lime additives mixed in drainage trench backfill led to reductions of NO3-N concentrations in drainage water by 78%, 69%, and 52%, respectively. The addition of lime to drainage trench backfill reduced PO4-P concentrations in the drainage water by 39%, while the addition of woodchips and chopped straw increased PO4-P concentrations by 11% and 22%, respectively. The NO3-N and PO4-P in drainage water were removed to the greatest extent by the addition of woodchips and lime, respectively.

This study also revealed that reactive filter materials can be used as in-ditch measures for phosphorus removal. Based on their P removal efficiencies, the filter materials can be ranked as follows: Polonite > slag > Filtralite-P > dolomite chips. Based on the average inflow load of 1.5 g P d-1 per 1.0 m3 of filter volume, Polonite was able to retain 0.7 g d-1, slag was able to retain 0.56 g d-1, Filtralite-P was able to retain 0.36 g d-1, and dolomite chips were able to retain 0.28 g d-1 of total phosphorus. The advantage of Polonite over the other tested materials was due to its higher porosity, lower sensitivity to clogging, and greater permeability. For nitrogen removal, bark can also be used as a filter material. However, its removal efficiency was shown to be relatively low. With an average inflow load of 120 g N d-1 per 1 m3 of bark filter, a retention of 8.5 g d-1 of total nitrogen was achieved (i.e., a removal efficiency of 7.1%).

The laboratory-scale experiment using denitrification bioreactors filled with three types of woodchips (deciduous, coniferous, and mixed) showed no significant differences in NO3-N removal efficiency among the three materials (average of 65.7%, 68.3%, and 66.7%, respectively). However, the tests showed that woodchip media are capable of achieving higher NO3-N removal rates due to higher flow rates. Higher NO3-N removal rates for shorter HRTs are governed by NO3-N removal reaction kinetics. This highlights the importance of better optimization and proper evaluation of the effects of HRT when designing denitrifying woodchip bioreactors. This experiment is the first attempt to test the applicability of denitrifying bioreactors for NO3-N removal in tile drainage systems in Lithuania. More investigations both at the laboratory scale and under field conditions are needed to quantify the long-term performance of these bioreactors in order to optimize the NO3-N removal rates for various flow conditions.

References

Addy, K., Gold, A. J., Christianson, L. E., David, M. B., Schipper, L. A., & Ratigan, N. A. (2016). Denitrifying bioreactors for nitrate removal: A meta-analysis. J. Environ. Qual., 45(3), 873-881. https://doi.org/10.2134/jeq2015.07.0399

Aleknavicius, P. (1989). Land cadastre [in Lithuanian]. Vilnius, Lithuania: Mokslas.

Andersson, H., Bergström, L., Djodjic, F., Ulen, B., & Kirchmann, H. (2016). Lime placement on subsoil as a strategy to reduce phosphorus leaching from agricultural soils. Soil Use Mgmt., 32(3), 381-389. https://doi.org/10.1111/sum.12290

Aškinis, S., & Miseviciene, S. (1998). An investigation of the soils’ sewage cleaning efficiency with the addition of different materials in the trenches of drains [in Lithuanian]. Water Mgmt. Eng., 5(27), 111-118.

Bambara, S., & Ndakidemi, P. A. (2010). The potential roles of lime and molybdenum on the growth, nitrogen fixation, and assimilation of metabolites in nodulated legume: A special reference to Phaseolus vulgaris L. African J. Biotech., 8(17), 2482-2489.

Bastiene, N., Šaulys, V., & Gurklys, V. (2012). An assessment of lime filter drainage systems. In M. S. Javaid (Ed.), Drainage Systems (pp. 181-210). Rijeka, Croatia: InTech. https://doi.org/10.5772/34941

Bergström, L., Djodjic, F., Kirchmann, H., Nilsson, I., & Ulén, B. (2007). Phosphorus from farmland to water: Status, flows, and preventive measures in a Nordic perspective. Uppsala, Sweden: Swedish University of Agricultural Sciences.

Beumer, V., van Wirdum, G., Beltman, B., Griffioen, J., & Verhoeven, J. T. (2007). Biogeochemical consequences of winter flooding in brook valleys. Biogeochem., 86(1), 105-121. https://doi.org/10.1007/s10533-007-9150-y

Blowes, D. W., Robertson, W. D., Ptacek, C. J., & Merkley, C. (1994). Removal of agricultural nitrate from tile-drainage effluent water using in-line bioreactors. J. Contam. Hydrol., 15(3), 207-221. http://dx.doi.org/10.1016/0169-7722(94)90025-6

Cameron, S. G., & Schipper, L. A. (2010). Nitrate removal and hydraulic performance of organic carbon for use in denitrification beds. Ecol. Eng., 36(11), 1588-1595. http://dx.doi.org/10.1016/j.ecoleng.2010.03.010

Cameron, S. G., & Schipper, L. A. (2012). Hydraulic properties, hydraulic efficiency, and nitrate removal of organic carbon media for use in denitrification beds. Ecol. Eng., 41, 1-7. http://dx.doi.org/10.1016/j.ecoleng.2011.11.004

Chichlowski, E. (2014). Nitrate removal from subsurface drainage by denitrifying bioreactor. MS thesis. West Lafayette, IN: Purdue University. Retrieved from http://docs.lib.purdue.edu/open_access_theses/311

Christianson, L. E., Bhandari, A., & Helmers, M. J. (2009). Emerging technology: Denitrification bioreactors for nitrate reduction in agricultural waters. J. Soil Water Cons., 64(5), 139-141. https://doi.org/10.2489/jswc.64.5.139A

Christianson, L. E., Bhandari, A., & Helmers, M. J. (2011). Pilot-scale evaluation of denitrification drainage bioreactors: Reactor geometry and performance. J. Environ. Eng., 137(4), 213-220. https://doi.org/10.1061/(ASCE)EE.1943-7870.0000316

Christianson, L., Bhandari, A., & Helmers, M. J. (2012). Practice-oriented review of woodchip bioreactors for subsurface agricultural drainage. Appl. Eng. Agric., 28(6), 861-874. https://doi.org/10.13031/2013.42479

Christianson, L., Castelló, A. L., Christianson, L., Helmers, M., & Bhandari, A. (2010). Hydraulic property determination of denitrifying bioreactor fill media. Appl. Eng. Agric., 26(5), 849-854. https://doi.org/10.13031/2013.34946

Chun, J. A., Cooke, R. A., Eheart, J. W., & Kang, M. S. (2009). Estimation of flow and transport parameters for woodchip-based bioreactors: I. Laboratory-scale bioreactor. Biosyst. Eng., 104(3), 384-395. http://dx.doi.org/10.1016/j.biosystemseng.2009.06.021

Cooke, R. A., Doheny, A. M., & Hirschi, M. C. (2001). Bio-reactors for edge-of-field treatment of tile outflow. ASAE Paper No. 012018. St. Joseph, MI: ASAE. https://doi.org/10.13031/2013.7373

Douglas, G. B., Robb, M. S., Coad, D. N., & Ford, P. W. (2004). A review of solid-phase adsorbents for the removal of phosphorous from natural and wastewaters. In E. Valsami-Jones (Ed.), Phosphorus in environmental technology: Principles and applications (pp. 291-320). London, UK: IWA Publishing.

Egnér, H., Riehm, H., & Domingo, W. R. (1960). Untersuchungen über die chemische Bodenanalyse als Grundlage für die Beurteilung de Nährstoff zustandes der Böden: II. Chemische Extraktionsmethoden zur Phosphor- und Kaliumbestimmung [in German]. Ann. Royal Agric. College Sweden, 26, 199-215.

Evans, R. O., Parsons, J. W., Stone, K., & Wells, W. B. (1992). Water table management on a watershed scale. J. Soil Water Cons., 47(1), 58-64.

FAO. (2015). World reference base for soil resources 2014: International soil classification system for naming soils and creating legends for soil maps (update 2015). World Soil Resources Reports No. 106. Rome, Italy: United Nations FAO. Retrieved from http://www.fao.org/3/a-i3794e.pdf

Galloway, J. N., Dentener, F. J., Capone, D. G., Boyer, E. W., Howarth, R. W., Seitzinger, S. P., ... Vöosmarty, C. J. (2004). Nitrogen cycles: Past, present, and future. Biogeochem., 70(2), 153-226. https://doi.org/10.1007/s10533-004-0370-0

Galvonaite, A., Valiukas, D., & Kilpys, J. (2013). Climate atlas of Lithuania [in Lithuanian]. Vilnius, Lithuania: Lithuanian Hydrometeorological Service.

Greenan, C. M., Moorman, T. B., Kaspar, T. C., Parkin, T. B., & Jaynes, D. B. (2006). Comparing carbon substrates for denitrification of subsurface drainage water. J. Environ. Qual., 35(3), 824-829. https://doi.org/10.2134/jeq2005.0247

Guzys, S., & Miseviciene, S. (2015). Nitrogen migration in crop rotations differing in fertilisation. Spanish J. Agric. Res., 13(2), e0303. https://doi.org/10.5424/sjar/2015132-6672

HelCom. (2013). Approaches and methods for eutrophication target setting in the Baltic Sea region. Baltic Sea Environment Proceedings No. 133. Helsinki, Finland: Helsinki Commission. Retrieved from http://www.helcom.fi/Documents/Ministerial2013/Associated%20documents/Background/Eutorophication%20targets_BSEP133.pdf

Hoover, N. L., Bhandari, A., Soupir, M. L., & Moorman, T. B. (2016). Woodchip denitrification bioreactors: Impact of temperature and hydraulic retention time on nitrate removal. J. Environ. Qual., 45(3), 803-812. https://doi.org/10.2134/jeq2015.03.0161

Husk, B. R., Anderson, B. C., Whalen, J. K., & Sanchez, J. S. (2017). Reducing nitrogen contamination from agricultural subsurface drainage with denitrification bioreactors and controlled drainage. Biosyst. Eng., 153, 52-62. http://dx.doi.org/10.1016/j.biosystemseng.2016.10.021

Hylander, L. D., Kietlinska, A., Renman, G., & Siman, G. (2006). Phosphorus retention in filter materials for wastewater treatment and its subsequent suitability for plant production. Bioresour. Tech., 97(7), 914-921. http://dx.doi.org/10.1016/j.biortech.2005.04.026

Jansons, V., Abramenko, K., & Berzina, L. (2009). Risk assessment of the agricultural pollution with nitrates in Latvia. LLU Raksti, 22(317). Retrieved from http://llufb.llu.lv/proceedings/n22/1/LLU-raksti-nr22-1-11.pdf

Jaynes, D. B., Colvin, T. S., Karlen, D. L., Cambardella, C. A., & Meek, D. W. (2001). Nitrate loss in subsurface drainage as affected by nitrogen fertilizer rate. J. Environ. Qual., 30(4), 1305-1314. https://doi.org/10.2134/jeq2001.3041305x

Jaynes, D. B., Kaspar, T. C., Moorman, T. B., & Parkin, T. B. (2008). In situ bioreactors and deep drain-pipe installation to reduce nitrate losses in artificially drained fields. J. Environ. Qual., 37(2), 429-436. https://doi.org/10.2134/jeq2007.0279

Jenkyn, J. F., Christian, D. G., Bacon, E. T. G., Gutteridge, R. J., & Todd, A. D. (2001). Effects of incorporating different amounts of straw on growth, diseases and yield of consecutive crops of winter wheat grown on contrasting soil types. J. Agric. Sci., 136(1), 1-14. https://doi.org/10.1017/S0021859600008558

Jokubauskaite, I., Slepetiene, A., & Karcauskiene, D. (2015). Influence of different fertilization on the dissolved organic carbon, nitrogen, and phosphorus accumulation in acid and limed soils. Eurasian J. Soil Sci., 4(2), 137-143. https://doi.org/10.18393/ejss.91434

King, K. W., Williams, M. R., Macrae, M. L., Fausey, N. R., Frankenberger, J., Smith, D. R., ... Brown, L. C. (2015). Phosphorus transport in agricultural subsurface drainage: A review. J. Environ. Qual., 44(2), 467-485. https://doi.org/10.2134/jeq2014.04.0163

Lepine, C., Christianson, L., Sharrer, K., & Summerfelt, S. (2016). Optimizing hydraulic retention times in denitrifying woodchip bioreactors treating recirculating aquaculture system wastewater. J. Environ. Qual., 45(3), 813-821. https://doi.org/10.2134/jeq2015.05.0242

Loeb, R., van Daalen, E., Lamers, L. P., & Roelofs, J. G. M. (2007). How soil characteristics and water quality influence the biogeochemical response to flooding in riverine wetlands. Biogeochem., 85(3), 289-302. https://doi.org/10.1007/s10533-007-9135-x

Lynch, S. F. L., Batty, L. C., & Byrne, P. (2014). Environmental risk of metal mining contaminated river bank sediment at redox-transitional zones. Minerals, 4(1), 52-73. https://doi.org/10.3390/min4010052

Moreira, A., & Kumar Fageria, N. (2010). Liming influence on soil chemical properties, nutritional status, and yield of alfalfa grown in acid soil. Revista Brasileira de Ciencia do Solo, 34(3), 1231-1239. http://dx.doi.org/10.1590/S0100-06832010000400022

Murphy, P. N. C., & Stevens, R. J. (2010). Lime and gypsum as source measures to decrease phosphorus loss from soils to water. Water Air Soil Pollut., 212(1-4), 101-111. https://doi.org/10.1007/s11270-010-0325-0

Narendrula-Kotha, R., & Nkongolo, K. K. (2017). Microbial response to soil liming of damaged ecosystems revealed by pyrosequencing and phospholipid fatty acid analyses. PLoS ONE, 12(1), e0168497. https://doi.org/10.1371/journal.pone.0168497

Pauliukevicius, G., & Juodis, J. (1987). Estimation of land in danger of deflation [in Russian]. In Ecological optimisation of the agrolandscape (pp. 71-74). Moscow, Russia: Nauka.

Porter, M. D., Andrus, J. M., Bartolerio, N. A., Rodriguez, L. F., Zhang, Y., Zilles, J. L., & Kent, A. D. (2015). Seasonal patterns in microbial community composition in denitrifying bioreactors treating subsurface agricultural drainage. Microb. Ecol., 70(3), 710-723. https://doi.org/10.1007/s00248-015-0605-8

Povilaitis, A. (2015). Hydrological effect of artificial drainage in lowland river catchments in Lithuania. Environ. Eng. Mgmt. J., 14(9), 2243-2253.

Povilaitis, A. (2016). Management practices and environmental effects of agricultural drainage in Lithuania. ASABE Paper No. 162513379. St. Joseph, MI: ASABE. https://doi.org/10.13031/IDS.20162513379

Povilaitis, A., Lamsodis, R., Bastiene, N., Rudzianskaite, A., Miseviciene, S., Miseckaite, O., ... & Balevicius, G. (2015). Agricultural drainage in Lithuania: A review of practices and environmental effects. Acta Agric. Scandinavica B, 65(S1), 14-29. https://doi.org/10.1080/09064710.2014.971050

Povilaitis, A., Šileika, A., Deelstra, J., Gaigalis, K., & Baigys, G. (2014). Nitrogen losses from small agricultural catchments in Lithuania. Agric. Ecosyst. Environ., 198, 54-64. http://dx.doi.org/10.1016/j.agee.2014.02.002

Powlson, D. S., Jenkinson, D. S., Pruden, G., & Johnston, A. E. (1985). The effect of straw incorporation on the uptake of nitrogen by winter wheat. J. Sci. Food Agric., 36(1), 26-30. https://doi.org/10.1002/jsfa.2740360105

Ramoška, E., & Morkunas, V. (2006). Influence of drainage runoff regulation on soil water regime [in Lithuanian]. Water Mgmt. Eng., 30(50), 5-14.

Ramoška, E., Bastiene, N., & Šaulys, V. (2011). Evaluation of controlled drainage efficiency in Lithuania. Irrig. Drain., 60(2), 196-206. https://doi.org/10.1002/ird.548

Renman, A. (2008). On-site wastewater treatment: Polonite and other filter materials for removal of metals, nitrogen, and phosphorus. PhD diss. Stockholm, Sweden: Royal Institute of Technology.

Robertson, W. D., Ford, G. I., & Lombardo, P. S. (2005). Wood-based filter for nitrate removal in septic systems. Trans. ASAE, 48(1), 121-128. https://doi.org/10.13031/2013.17954

Romanovskaja, D., & Tripolskaja, L. (2003). The agroecological evaluation of various organic fertilizer usage on sandy loam soil [in Lithuanian]. Zemdirbyste, 84(4), 3-22.

Sakalauskas, A. (1986). Hydrometric yearbook: Drainage outflow [in Lithuanian]. Kaunas, Lithuania: RVUP.

Sakalauskas, A. (1995). Hydrometric yearbook: Drainage and surface outflow [in Lithuanian]. Kaunas, Lithuania: LŽUA.

Šaulys, V., & Bastiene, N. (2006). The effect of lime admixture to trench backfill on the functioning of tile drainage in heavy soils. Irrig. Drain., 55(4), 373-382. https://doi.org/10.1002/ird.262

Šaulys, V., & Bastiene, N. (2008). The impact of lime on water quality when draining clay soils [in Lithuanian]. Ecol., 54(1), 22-28. https://doi.org/10.2478/V10055-008-0005-1

Šaulys, V., Bastiene, N., Gurklys, V., & Kincius, L. (2011). Assessment and application priorities of environmental measures in renovation of drainage systems [in Lithuanian]. Water Mgmt. Eng., 39(59), 52-60.

Schärer, M. (2003). The influence of processes controlling phosphorus availability on phosphorus losses in grassland soils. PhD diss. Zurich, Switzerland: Federal Institute of Technology.

Schipper, L. A., Robertson, W. D., Gold, A. J., Jaynes, D. B., & Cameron, S. C. (2010). Denitrifying bioreactors: An approach for reducing nitrate loads to receiving waters. Ecol. Eng., 36(11), 1532-1543. http://dx.doi.org/10.1016/j.ecoleng.2010.04.008

Schipper, L., & Vojvodic-Vukovic, M. (1998). Nitrate removal from groundwater using a denitrification wall amended with sawdust: Field trial. J. Environ. Qual., 27(3), 664-668. https://doi.org/10.2134/jeq1998.00472425002700030025x

Smith, D. R., Haggard, B. E., Warnemuende, E. A., & Huang, C. (2005). Sediment phosphorus dynamics for three tile-fed drainage ditches in northeast Indiana. Agric. Water Mgmt., 71(1), 19-32. http://dx.doi.org/10.1016/j.agwat.2004.07.006

Tiemeyer, B., Kahle, P., & Lennartz, B. (2006). Nutrient losses from artificially drained catchments in northeastern Germany at different scales. Agric. Water Mgmt., 85(1-2), 47-57. http://dx.doi.org/10.1016/j.agwat.2006.03.016

Tripolskaja, L., Bagdanaviciene, Z., & Romanovskaja, D. (2002). Changes in mineral nitrogen and microbiological activity of soil during decomposition of organic fertilizers in the autumn-winter season [in Lithuanian]. Agric. Sci., 2, 3-12.

Ulén, B., & Mattsson, L. (2003). Transport of phosphorus forms and of nitrate through a clay soil under grass and cereal production. Nutr. Cycling Agroecosyst., 65(2), 129-140. https://doi.org/10.1023/a:1022133031584

Vaišvila, Z. (1999). The soil and its chemical properties [in Lithuanian]. In Z. Vaisvila (Ed.), Agrochemija. Kaunas, Lithuania: Lutute.

van Driel, P. W., Robertson, W. D., & Merkley, C. (2006). Denitrification of agricultural drainage using wood-based reactors. Trans. ASABE, 49(2), 565-573. https://doi.org/10.13031/2013.20391

Velykis, A., Satkus, E., & Ramoška, E. (2003). Straw usage for heavy soil hydro physical properties and moisture regime. Zemdirbyste-Agric., 79(3), 17-29.

Violante, A., Cozzolino, V., Perelomov, L., Caporale, A. G., & Pigna, M. (2010). Mobility and bioavailability of heavy metals and metalloids in soil environments. J. Soil Sci. Plant Nutrition, 10(3), 268-292. https://doi.org/10.4067/S0718-95162010000100005

Warneke, S., Schipper, L. A., Matiasek, M. G., Scow, K. M., Cameron, S., Bruesewitz, D. A., & McDonald, I. R. (2011). Nitrate removal, communities of denitrifiers, and adverse effects in different carbon substrates for use in denitrification beds. Water Res., 45(17), 5463-5475. http://dx.doi.org/10.1016/j.watres.2011.08.007

Wesström, I., & Messing, I. (2007). Effects of controlled drainage on N and P losses and N dynamics in a loamy sand with spring crops. Agric. Water Mgmt., 87(3), 229-240. http://dx.doi.org/10.1016/j.agwat.2006.07.005

Woli, K. P., David, M. B., Cooke, R. A., McIsaac, G. F., & Mitchell, C. A. (2010). Nitrogen balance in and export from agricultural fields associated with controlled drainage systems and denitrifying bioreactors. Ecol. Eng., 36(11), 1558-1566. http://dx.doi.org/10.1016/j.ecoleng.2010.04.024

Youssef, M. A., Chescheir, G. M., Skaggs, R. W., Appelboom T., W., Poole C., A., & Liu, Y. (2016). Performance of in-line denitrifying woodchip bioreactors for nitrogen reduction from subsurface drained pasture receiving swine lagoon effluent. ASABE Paper No. 162489231. St. Joseph, MI: ASABE. https://doi.org/10.13031/IDS.20162489231

Zalakevicius, M. (2001). Lithuanian rivers: On the way from degradation to restoration. In H. J. Nijland & M. J. Cals (Eds.), River restoration in Europe: Practical approaches (pp. 125-129). European Center for River Restoration.