Determining water and nitrogen balances for beneficial management practices using lysimeters at Wagna test site (Austria)

Determining water and nitrogen balances for beneficial management practices using lysimeters at Wagna test site (Austria)

Science of the Total Environment 499 (2014) 448–462 Contents lists available at ScienceDirect Science of the Total Environment journal homepage: www...

3MB Sizes 0 Downloads 90 Views

Science of the Total Environment 499 (2014) 448–462

Contents lists available at ScienceDirect

Science of the Total Environment journal homepage:

Determining water and nitrogen balances for beneficial management practices using lysimeters at Wagna test site (Austria) Gernot Klammler ⁎, Johann Fank 1 JOANNEUM RESEARCH Graz, RESOURCES, Institute for Water, Energy and Sustainability, Elisabethstraße 18/II, A-8010 Graz, Austria

H I G H L I G H T S • The presented lysimeter design did not show any oasis or fringe effects. • N-leaching from investigated BMPs did not endanger groundwater quality. • Comparison between calculated ET0 and ET0was determined with a lysimeter.

a r t i c l e

i n f o

Article history: Received 9 January 2014 Received in revised form 29 April 2014 Accepted 2 June 2014 Available online 27 June 2014 Keywords: Low nitrogen input farming Organic farming Nitrogen/nitrate leaching Evapotranspiration Grass-reference evapotranspiration weighable, monolithic lysimeter

a b s t r a c t The shallow Murtal aquifer south of Graz, Austria, provides easily withdrawable groundwater, which is supplied as drinking water without any chemical treatment. The aquifer is also used intensively by agriculture. Common agricultural management practices are the main source for diffuse nitrogen leaching and high groundwater nitrate concentrations. To safeguard the coexisting use of these two important resources, lysimeters are operated at the agricultural test site Wagna, Austria, and the influence of two beneficial management practices – low nitrogen input and organic farming – on nitrogen leaching towards groundwater is investigated. The technical lysimeter design as presented here consists of: (1) high-resolution weighing cells, (2) a suction controlled lower boundary condition for sucking off seepage water, thus emulating undisturbed field conditions, (3) comparative soil temperature, water content and matrix potential measurements inside and outside the lysimeter at different depths, (4) an installation of the lysimeters directly into test plots and (5) a removable upper lysimeter ring enabling machinery soil tillage. Our results indicate that oasis effects or fringe effects of the lysimeter cylinder on unsaturated water flow did not occur. Another lysimeter cultivated with lawn is operated for observing grass-reference evapotranspiration, which resulted in good agreement with calculated grassreference evapotranspiration according to the FAO-Penman–Monteith method. We conclude that lysimeters installed at Wagna test site did not show any fringe effects and, thus, are appropriate tools for measuring water balance elements and nitrogen leaching of arable and grass land at point scale. Furthermore, our results for the period of 2005 to 2011 show that beneficial management practices reduced nitrate leaching and, hence, may allow for a sustainable coexistence of drinking water supply and agriculture in the Murtal aquifer. © 2014 Elsevier B.V. All rights reserved.

1. Introduction Since agriculture is the largest non-point source of groundwater pollution (Sutton et al., 2011), the need for water quality and nutrient management as well as the need for a better understanding of water drainage and chemical leaching through the vadose zone is continuing to grow. The Murtal aquifer between Graz and Bad Radkersburg, Austria (Fig. 1), is an important resource for regional and supraregional drinking water supply, but the region also provides excellent agricultural ⁎ Corresponding author. Tel.: +43 316 876 6000. E-mail addresses: [email protected] (G. Klammler), [email protected] (J. Fank). 1 Tel.: +43 316 876 6000. 0048-9697/© 2014 Elsevier B.V. All rights reserved.

conditions. This dual use implicates conflicts due to high groundwater nitrate concentrations (cNO3) caused by diffuse nitrogen (N) pollution, which have to be harmonized for a sustainable coexistence. Therefore, monitoring and measuring techniques, that can determine drainage fluxes from undisturbed soil profiles, are critical for the determination of nutrient budgets and the evaluation of land-use practices on water quality (Masarik et al., 2004). Accurate crop evapotranspiration (ET) data are required to improve agricultural water resources management. Lysimeters are still considered to be the standard method to determine ET from measurements. If the lysimeters are weighable, the current ET can be deduced from their weight change (Young et al., 1996). Furthermore, Meissner et al. (2007) show that also a precise measurement of dew, fog, and rime is possible using a high-resolution weighing system. Thus, a large

G. Klammler, J. Fank / Science of the Total Environment 499 (2014) 448–462


Fig. 1. Location and overview of the test site Wagna, Austria.

weighable lysimeter is the best method for obtaining reliable ET data and also seepage water (SW) quantity and quality. The evaluation of lysimeter data allows a much more reliable calculation of the solute load carried towards the groundwater than any other method (Klocke et al., 1993), e.g., hydrochemical investigations in the groundwater, isotopic analysis, and water and solute balancing. Due to these characteristics, lysimeters are also excellent tools to derive or calibrate water and solute transport models (Wriedt, 2004) for unsaturated zone simulation. Lysimeters in Europe are predominantly used for agricultural research (approx. 63%; Lanthaler and Fank, 2005), but have also become important concerning climate change research. For example, the SoilCan project (soil can make a difference in climate policy) is designed as a long term large scale experiment to study the effects of climate change on soil systems. Based on 126 lysimeters, SoilCan focuses on water and matter fluxes in soil (Pütz et al., 2011; Zacharias et al., 2011), which also reflects the ability of lysimeters as an appropriate tool for water resources research. In 2004 two monolithic, weighable lysimeters were implemented at the agricultural test site Wagna in southern Austria (Fank and Von Unold, 2007) for investigating the influence of beneficial management practices (BMPs) on groundwater quantity and quality. BMPs are management practices that reduce or eliminate environmental risks in general. At Wagna test site the BMPs on low nitrogen input farming and organic farming are cultivated at the test plots and lysimeter observations focus on diffuse nitrogen pollution of groundwater due to agricultural fertilization. Shortcomings assigned to lysimeter measurements (e.g., oasis effects, preferential flow paths at the walls of the lysimeter cylinders or the influence of the lower boundary condition on the outflow rates) are prevented by a special design of lysimeters installed. In the present work, based on this new type of field-lysimeters – high-resolution, weighable, monolithic lysimeters directly installed into arable land – we determine exactly measured water balance parameters and best possible information about the influence of different farming systems on shallow groundwater. Furthermore, but not investigated here, the gathered data can be used to develop, calibrate and validate models for water and solute transport in the unsaturated zone in order to transfer point-data from the lysimeter to field and regional scales. The unsaturated zone models can also be coupled with groundwater flow and transport models to predict effects of different farming systems on saturated groundwater quality and quantity at regional scales. The objective of this paper is to present the lysimeter based measuring equipment at the agricultural test site Wagna and to

summarize the results of a seven year period of water and nitrogen balance determination at the interface between atmosphere and groundwater. We will demonstrate that BMPs can significantly reduce N leaching into the groundwater, which may be of fundamental importance for avoiding conflicts between agriculture and groundwater protection in the highly sensitive Murtal aquifer, in Austria.

2. Material and methods 2.1. Description of the test site The agricultural test site Wagna with a total area of 4.4 ha is located within the Mur Valley between Graz and Bad Radkersburg, Austria, and consists of 32 test plots with approx. 1000 m2 each (Fig. 1) and is situated on a gravel terrace of Würm glaciation (aquifer thickness approx. 8 m). Clayey–sandy Cambisols are predominant (soil depths are very heterogeneous ranging between 15 and 230 cm) and the content of clay and sand is about 15% and 52%, respectively. The humus content ranges between 1.3% and 2.2%. All in all, the location is composed of very light soils with a low water storage capacity and its characteristics are representative for most parts of the Mur Valley between Graz and Bad Radkersburg. Since 1987 different cultivation strategies have been researched concerning N fertilizer input, nitrate (NO3) leaching and crop yields. Primarily, comparisons were made between maize-mono-cropping and crop rotations (maize-maize-winter grain-winter rape) with the objective of yield increase. In 1998, due to high NO3 concentrations in the groundwater, the research question and cultivation strategy changed to reduction of N fertilizer, no fertilization in autumn, hardy grass catch crops (no legumes) and cultivation of oil pumpkin instead of winter rape (due to increasing importance of oil pumpkin in the Mur Valley). Since 2005 the difference between low N input farming and organic farming has been researched (crop rotations only). The two crop rotations are cultivated in four variations (i.e., crop rotation starts with different crop in the same year) and four repetitions. In this context, two of the test plots were equipped with high-precision lysimeters and soil hydrologic measuring profiles (SCIENCELYS; Fig. 1). An additional grass-reference lysimeter (HYDROLYS; Fig. 1) in combination with a weather station (weather station 1; Fig. 1) was also installed at the southeast limit of the test site. Since 2003, further weather data has been acquired at a weather station of the national


G. Klammler, J. Fank / Science of the Total Environment 499 (2014) 448–462

meteorological service ZAMG (weather station 2; Fig. 1), which is also situated at the test site. 2.2. Measuring equipment 2.2.1. SCIENCE-lysimeter In 2004 two weighable monolithic SCIENCE-lysimeters (UMS, 2013a) were installed at the test site, one of them cultivated by low N input farming (SCI 1), the other one by organic farming (SCI 2). To avoid oasis effects the lysimeters are directly implemented into the test plots and are cultivated using agricultural standard machinery (e.g., tractor, plow). A number of sensors and sampling devices allow describing the soil water and solute transport situation stepwise down the lysimeter. The measurements inside the lysimeter monolith are supplemented by sensor readings in a soil hydrologic measurement profile in the undisturbed soil outside the lysimeter to investigate if there is an influence of the lysimeter construction on soil hydrologic properties inside the lysimeter. The bottom of the SCIENCELYS is designed as a potential boundary condition where a certain water potential can be applied at a suction cup rake to suck of SW (controlled by a tensiometer installed in undisturbed soil outside the lysimeter). Design and technical equipment. Fig. 2 shows the design and technical equipment of the SCIENCE-lysimeter described as follows: • Dimensions: 2 m depth, 1 m2 surface • Cultivation ring: This is actually not a regular component of the SCIENCELYS by UMS, but at Wagna test site it was installed for machine operated tillage of the lysimeter. In other words, the upper 30 cm of the lysimeter cylinder can be removed to cultivate the top soil of the lysimeter with standard machinery (weighing cells have to be unloaded first). • Soil water sampler: For analyzing soil water quality, suction cups are installed in 35, 60 and 90 cm depth. • Soil moisture probes: TDR probes (Time Domain Reflectometry; TRIME-IT) are installed in 35, 60, 90 and 180 cm depth (temporarily also in 10 and 20 cm depth). • Tensiometer: For measuring the hydraulic head, tensiometer with a measuring range between +100 hPa and −850 hPa are installed in 90 and 180 cm depth. • Matrix sensors: For measuring the hydraulic head up to −2000 hPa, matrix sensors are installed in 35, 60 and 90 cm depth (temporarily also in 10 and 20 cm depth).

• Soil temperature probes: Installed in 35, 60, 90 and 180 cm depth (not illustrated in Fig. 2). Temporarily soil temperature was also measured in 10 and 20 cm depth. • Silicon carbide porous suction cup rake: For a clearly defined lower boundary condition, the SW is sucked off by a suction cup rake (surface 3600 cm2) in 180 cm depth. Tension is applied by a vacuum pump which is controlled by a tensiometer in 180 cm depth of an undisturbed soil outside the lysimeter. This guaranties the same flow rates in the lysimeter as in the undisturbed field. • Precision weighing system: For measuring mass changes of the lysimeter (e.g., due to ET, precipitation (P), SW outflow) the lysimeter cylinder is situated on three high-precision weighing cells (resolution of 35 g or 0.035 mm water equivalent). • Tipping bucket: For measuring the quantity of SW, a tipping bucket with 0.1 mm resolution is installed (not illustrated in Fig. 2, where SW is quantified by a weighing gauge); water is sampled for chemical analysis. • The following measurements are taken outside the lysimeter to detect possible influences of the lysimeter cylinder on the SW and N dynamics: o Soil temperature probes in 35, 60, 90 and 180 cm depth (not illustrated in Fig. 2) o Soil moisture probes in 35, 60, 90 and 180 cm depth (not illustrated in Fig. 2) o Matrix sensors in 35, 60 and 90 cm depth o Tensiometer in 90 and 180 cm depth The measuring interval of the probes is 10 s; average values are recorded every 10 min. Soil profiles. In the course of installing the lysimeters in 2004 detailed soil samples were taken and analyzed by the Institute for Land and Water Management Research of the Austrian Federal Agency for Water Management. Thus, the following parameters are available for the location of the two SCIENCE-lysimeters: humus content, pH in soil, CaCO3-content, density of substrate, dry bulk density, moist bulk density, pore volume, hygroscopicity, particle size distribution, saturated hydraulic conductivity and relations between water content, hydraulic head and unsaturated hydraulic conductivity (steady state and transient method). Table 1 gives an overview of the soil horizons and the soil textures of the two SCIENCE-lysimeters. Cultivation. Since the installation of the SCIENCE-lysimeter in 2004 the research objective of the test site Wagna has changed to

Fig. 2. Design of a SCIENCE-lysimeter from UMS Munich (UMS, 2013a).

G. Klammler, J. Fank / Science of the Total Environment 499 (2014) 448–462


Table 1 Soil horizons and soil textures for the locations of SCIENCELYS 1 and SCIENCELYS 2. SCIENCELYS 1 (low N input farming) Depth [cm] 0–30 30–50 50–60 60–200

a b c

SCIENCELYS 2 (organic farming)


Soil texturea,b (clay/silt/sand/gravel)a, c

Depth [cm]


Soil texturea, b (clay/silt/sand/gravel)a, c

Ap R B B C

Ls3 Ls4 Sl4 Ss

0–20 20–30 30–80 80–100 100–110 110–120 120–200

Ap R Ap R B B B B C

Sl4 Sl4 Ls3 Sl3/Sl4 Su2/Sl2 Su3 Ss

(20/33/45/2) (20/27/53/0) (14/24/62/0) (0/2/31/67)

(11/21/37/31) (12/25/41/22) (20/32/47/1) (12/22/66/0) (5/23/72/0) (7/35/58/0) (1/1/31/67)

According to AG Boden (2005). Abbreviations valid for grain size distribution b2 mm. Classification: 0 b 0.002 b 0.063 b 2 b 63 mm.

investigating the following two beneficial management practices to reduce the environmental risks on groundwater (especially N leaching) due to agricultural production: Low N input farming

(e.g., in 2008) depending on the amount of dry matter and the C/Nratio of the slurry. Typical catch crops for low N input farming are ryegrass or forage rye; for organic farming clover and/or lucerne are generally cultivated between two main crops.

• N-fertilizer input according to a guideline for appropriate fertilization (BMLFUW, 2006). This guideline provides recommendations concerning fertilization rates, which assume that the applied fertilizer amounts do not exceed the plant's uptake. Since the guideline does not provide recommendations for oil pumpkin, the N fertilization rate (exclusive of volatilization losses) is set to 50–60 kg/ha/a based on expert opinion. • Application of mineral and farm fertilizer as well as pesticides • Conventional tillage (e.g., plow)

2.2.2. HYDRO-lysimeter (grass-reference lysimeter) In 2006, a weighable monolithic HYDRO-lysimeter (UMS, 2013b) was installed for measuring highly resolved and precise water balance parameters (lower boundary condition tension-controlled). The HYDROLYS at Wagna test site is cultivated with a permanent grass cover of 12 cm height. This surface corresponds to the reference surface for calculating the grass-reference ET according to Allen et al. (1998) and allows a comparison between measured and calculated ET. Design and technical equipment. The design of the HYDROLYS is similar to the SCIENCELYS (Fig. 2), however, there is less measuring equipment installed:

Organic farming • • • •

Cultivation according to EC (2008) and EC (2007) No application of pesticides, mineral or farm fertilizer N-input exclusively from atmospheric N fixation of legumes Shallow tillage (e.g., ripper)

Table 2 summarizes the cultivated main crops and the applied N fertilization rates for low N input farming and organic farming at the two SCIENCE-lysimeters from 2005 to 2011. According to the Integrated Administration and Control System IACS of the EU, maize (grain; 53%), oil pumpkin (13%) and winter grain (i.e., winter wheat, winter barley and triticale; 9%) are the main crops in the Mur Valley between Graz and Bad Radkersburg and, consequently, also at the Wagna test site. The average N rate applied for low N input farming during our test period was 113 kg/ha/a; for organic farming the N input by legumes was not quantified. The N fertilization for low N input farming is usually split into two applications: the first one with farm fertilizer (slurry), the second one with mineral fertilizer (74% NH4NO3 and 26% CaCO3). Since farm fertilizer may vary in its N content, the applied N rates may differ from the recommended fertilization rate according to BMLFUW (2006). The input of organic nitrogen Norg can increase significantly

Dimensions: 1 m depth, 1 m2 surface No cultivation ring Tensiometer installed in 90 cm depth inside and outside the lysimeter Precision weighing system (resolution of 10 g or 0.01 mm water equivalent) • Silicon carbide porous suction cup rake in 90 cm depth for SW extraction (applied tension is controlled by a tensiometer in 90 cm depth outside the lysimeter) • Bidirectional vacuum pump for transferring water back into the lysimeter over the suction cup rake in case of capillary rise • Weighable leachate tank (resolution of 1 g or 0.001 mm water equivalent) • • • • Soil profile. Table 3 shows the soil horizons and the soil texture of the HYDROLYS. Further soil parameters analyzed for the HYDROLYS monolith comply with the parameters for the SCIENCELYS given in Section

Table 2 Cultivated main crops and nitrogen fertilization rates applied (separated into mineralized nitrogen Nmin and organic nitrogen Norg) at SCIENCELYS 1 and SCIENCELYS 2 from 2005 to 2011. Nmin in this table describes the N rate applied at the lysimeter and does not consider N losses of farm fertilizers due to volatilization of NH3. Year

2005 2006 2007 2008 2009 2010 2011 a

SCIENCELYS 1 (low N input farming)

SCIENCELYS 2 (organic farming)


Nmin [kg/ha/a]

Norg [kg/ha/a]


Nmin [kg/ha/a]

Norg [kg/ha/a]a

Oil pumpkin Maize Maize Winter Barley Oil pumpkin Maize Triticale Annual average:

56 118 121 94 52 112 146 100

7 12 9 38 0 9 17 13

Oil pumpkin Clover Maize Triticale Oil pumpkin Maize Triticale

0 0 0 0 0 0 0 0

– – – – – – – –

Norg-input by legumes not quantified.


G. Klammler, J. Fank / Science of the Total Environment 499 (2014) 448–462

2.4. Climatic water balance

Table 3 Soil horizons and soil textures for the locations of the HYDROLYS.

The determination of the real evapotranspiration ETr is based on the solution of the climatic water balance. Eq. (1) shows the elements of the climatic water balance used for lysimeters and neglecting irrigation, surface runoff, horizontal subsurface flow and capillary rise.

HYDROLYS (grass-reference) Depth [cm]


Soil texturea, b (clay/silt/sand/gravel)a, c

0–20 20–50 50–100

A C1 C2

Ls4 Sl4 Ss

a b c

(11/14/32/43) (7/9/32/52) (0/1/26/73)

P−ETr −SW−ΔS ¼ 0

According to AG Boden (2005). Abbreviations valid for grain size distribution b2 mm. Classification: 0 b 0.002 b 0.063 b 2 b 63 mm.


where Cultivation. The permanent grass cover between 2006 and 2011 was cultivated without any fertilization; in the vegetation period the grass was cut once a week and grass residuals remained on the lysimeter. 2.2.3. Meteorological observation At Wagna test site several weather elements are measured by different operators. Weather station 1 is situated right next to the HYDROLYS (see Fig. 1) and measures among others all weather elements required for calculating the grass-reference ET according to Allen et al. (1998). Data of weather station 2 (see Fig. 1) is used to verify and correct data of weather station 1 in case of measuring errors. Precipitation is measured at the location of weather station 2 using three different devices: a tipping bucket, a precipitation gauge and a totalisator. At the same location also soil temperature in 5, 10 and 20 cm depth under permanent grass cover is observed. 2.2.4. Groundwater observation Two groundwater observation boreholes are located near weather station 2 (see Fig. 1). Groundwater table, electric conductivity and groundwater temperature are measured automatically every 10 min. During the period of record (including our test period) the groundwater table is on average about 4.5 m below ground surface and ranges between 3.5 and 5.3 m depth. Since the top of the gravel layer is located at 0.6 and 1.2 m depth at SCI 1 and SCI 2, respectively, no influence of the groundwater table due to capillary rise on water content and nitrate concentrations for locations above the bottom lysimeter outlet in 1.8 m depth is expected. 2.3. Chemical analysis Seepage water and soil water in 35, 60, and 90 cm depth (only SCI 1 and SCI 2) from the lysimeters are sampled in a weekly interval (if there is seepage/soil water flow) and parameters Cl, NO3, SO4, pH and electric conductivity are monitored; for the period 2005 to 2009 also Na, K, Mg, and Ca have been analyzed (hydro-chemical parameters analyzed by ion-exchange chromatography). Nitrogen contents are also analyzed for crop yields, crop residuals and hardy catch crops (Kjeldahl) and for slurry which is applied at the test site.


precipitation real evapotranspiration seepage water soil water storage change.

Because of the rather plane surface and the fact that the lysimeter edge is 5 cm above ground, surface runoff from the lysimeter can generally be ignored for the determination of ETr. However, in case of heavy rain events small runoff may occur through the gap between the inner and the outer lysimeter cylinder. To quantify the error of neglecting it in the water balance, a runoff sampling system was installed in the course of maintenance at SCIENCELYS 2 in March 2007. 2.4.1. Calculation of grass-reference evapotranspiration ET0 The FAO-Penman–Monteith method presented in Eq. (2) is recommended after Allen et al. (1998) as the sole method for determining ET0. This is because it closely approximates grass ET0 at the location evaluated, it is physically based, and it explicitly incorporates both physiological and aerodynamic parameters. This method provides consistent ET0 values for all seasons, regions and climates worldwide requiring the climatic element solar radiation, temperature, air humidity and wind speed.

ET0 ðΔt Þ ¼

Cn u ðe −e Þ T þ 273 2 s a Δ þ γð1 þ C d u2 Þ

0:408ΔðRn −GÞ þ γ


where ET0 Δt Rn G T u2 es ea Δ γ Cn Cd

grass-reference evapotranspiration [mm Δt−1] time step (e.g. day, h, min) net radiation at the crop surface [MJ m−2 Δt−1] soil heat flux density [MJ m−2 Δt−1] mean daily air temperature at 2 m height [°C] wind speed at 2 m height [m s−1] saturation water pressure [kPa] actual water pressure [kPa] slope of vapor pressure curve [kPa °C−1] psychrometric constant [kPa °C−1] coefficient concerning calculation time step [K mm s3 mg−1 Δt−1] coefficient concerning surface resistance [s m−1].

Table 4 Coefficients Cn, Cd and soil head flux density G for different calculation time steps. Time step

1 daya 1 houra 1 hourb 10 minc a b c


900 37 37 6.25



Daytime (Rn N 0)

Nighttime (Rn b =0)

Daytime (Rn N 0)

Nighttime (Rn b =0)

0.34 0.34 0.24 0.24

0.34 0.34 0.96 0.96

0 0.1 Rn 0.1 Rn 0.1 Rn

0 0.5 Rn 0.5 Rn 0.5 Rn

According to Allen et al. (1998). According to Allen et al. (2006) & ASCE-EWRI (2005). Modified after Allen et al. (2006) & ASCE-EWRI (2005).

G. Klammler, J. Fank / Science of the Total Environment 499 (2014) 448–462


Fig. 3. Average monthly totals of precipitation (P), seepage water (SW) and evapotranspiration (ETr) of SCIENCELYS 1 (low N input farming) for the period 2005–2011.

Cn and Cd are coefficients that differ with calculation time step and with time of day. In areas where substantial changes in wind speed, dew point or cloudiness occur during the day, calculation of ET0 using time steps shorter than 1 day are recommended. For this purpose, Allen et al. (1998, 2006) provide coefficients for determining ET0 for hourly time steps presented in Table 4. However, implementing an adequate Cn coefficient in Eq. (2) allows also a determination of ET0 for 10 min time steps, which correspond to the measuring interval of weather station 1 at Wagna test site.

2.4.2. Direct determination of ET The common determination of ET using lysimeters is based on solving the climatic water balance given in Eq. (1) for daily time steps, where quantity of SW is measured at the bottom lysimeter outlet and ΔS is determined by measuring the lysimeter mass change. P is measured by a separate device (e.g. P gauge or tipping bucket). Thus, the measured amount of P may deviate from the amount actually falling on the lysimeter surface, because P is not measured exactly at the same location where the lysimeter is positioned, P samplers are mostly situated at a measuring height of 1 m or higher above ground, they have an approx. 15 times smaller surface area than a lysimeter (compared to a lysimeter surface of 1 m2) and they are not covered with vegetation. In some cases this can lead to physically impossible result of ET b0. For example, at SCIENCELYS1 on January 13th 2005 SW = 0.7 mm and ΔS = +0.65 mm are measured. The P gauge (located a few meters next to the lysimeter) measured P = 1.1 mm. According to Eq. (1), ET would result in ET = 1.1 − 0.7 − 0.65 = − 0.25 mm for this day. In such cases ET is defaulted to zero.

2.5. Nitrogen balance Eq. (3) describes the N balance consisting of the N inputs, N outputs and the N change in the lysimeter soil: F þ DP þ BF−L−CY−CR−V−DN−ΔS ¼ 0



mineral and farm fertilizer input atmospheric deposition biological nitrogen fixation due to legumes nitrogen leaching nitrogen removal by crop yield nitrogen removal by crop residuals (only if crop residuals are removed from the field/lysimeter) volatilization denitrification nitrogen storage change in the soil.

In fact, the N balance cannot be solved entirely, because DP and V were not measured at the Wagna test site. Furthermore, BF (e.g., significant for SCIENCELYS 2) and ΔS were not quantified either. Thus, only F, L, CY and CR (indicated in bold letters in Eq. (3)) were analyzed here for the lysimeters installed in Wagna (DN insignificant; Leis, 2009), where L is calculated by means of the amount (SW) and nitrate concentration (cNO3) of the seepage water according to Eq. (4): h i h i −1  −2 SW l m =442:7: L½kg=ha ¼ cNO3 mg l

Fig. 4. Average monthly totals of precipitation (P), seepage water (SW) and evapotranspiration (ETr) of SCIENCELYS 2 (organic farming) for the period 2005–2011.



G. Klammler, J. Fank / Science of the Total Environment 499 (2014) 448–462

Table 5 Annual totals of precipitation (P), evapotranspiration (ETr), seepage water (SW) and soil water storage change (ΔS) for SCIENCELYS 1 and 2 from 2005 to 2011. [mm]

















2005 2006 2007 2008 2009 2010 2011 Mean

Pumpkin Maize Maize W. Barley Pumpkin Maize Triticale

879 833 892 893 1360 1014 730 943

563 532 553 694 742 573 725 626

319 320 296 192 609 429 63 318

−3 −19 43 7 9 11 −58 −1

2005 2006 2007 2008 2009 2010 2011 Mean

Pumpkin Clover Maize Triticale Pumpkin Maize Triticale

879 833 892 893 1360 1014 730 943

505 612 571 741 692 577 748 635

374 276 217 151 662 436 61 311

−1 −55 104 1 6 1 −80 −3

– – 5 8 21 7 6 9


Runoff measured since March 2007; not considered for determining ETr.

2.6. Tracer test Particular emphasis should be paid on a double-tracer test in 2005 for verifying the influence of the lysimeter cylinder wall on unsaturated water flow. For this purpose, deuterium (D2O) was applied on April 12th 2005 in the center of the SCIENCE-lysimeters and sodium bromide (NaBr) at the outer segment of the lysimeters (each tracer covers 50% of the lysimeter surface). The applied amount of D2O is 10 ml with a deuterium content of 99.8 at.% and 50 g of NaBr. For calculating the recovery rate of bromide in the SW it was also required to analyze the bromide uptake by crops. 3. Results 3.1. Water and nitrogen balance Results for water balance are presented according to the standard determination method of ETr as given in Section 2.4.2. Since not all elements of the N balance can be analyzed, the focus within this paper is set on N leaching, but also the N uptake by plants and the applied N fertilization rates are considered. 3.1.1. SCIENCE-lysimeter (SCI 1 and SCI 2) Figs. 3 and 4 show the average monthly SW and ET rates compared to P for SCI 1 and SCI 2, respectively (based on measurements from 2005 to 2011). There is no period without SW at Wagna test site, even if ET exceeds P rates as it is the case in some months. In general, lowest SW rates occur in July, where ET is highest. Furthermore, low SW rates also seem to be representative for April with approx. 12 mm/month. In Table 5 annual totals of P, ET, SW and storage change are given. It can also be seen that annual average SW rates are 318 mm/a at SCI 1 and 311 mm/a at SCI 2. In contrast, P may range between 61 mm/a in 2011 and 662 mm/a in 2009 (at SCI 2) mainly as a consequence of annual P rates. Average ETr is calculated (according to Eq. (1)) to be 9 mm/a higher at SCI 2 (635 mm/a) than at SCI 1 (626 mm/a). Independent of

different soil conditions and cultivation strategies between SCI 1 and SCI 2, crop specific mean ETr is 561 mm/a for maize, 626 mm/a for oil pumpkin and 738 mm/a for triticale. Table 5 also shows a measured runoff for SCI 2 of 9 mm/a on average, which is approx. 1% of P. Monthly totals for P, SW and ET are given in Figs. 5 and 6 for SCI 1 and SCI 2, respectively. It is obvious that contrary to the long term annual averages presented in Figs. 3 and 4, SW may not occur in certain periods. Especially in the second half of 2011 which shows a long absence of SW. High P in winter 2008/09 as well as high P between May and September 2009 goes along with high SW rates in 2009 in general. The highest monthly P occurred in August 2010 with approx. 240 mm/month, which resulted in highest monthly SW rates of approx. 150 mm in the subsequent month at both SCI 1 and SCI 2. Table 6 shows annual sums of N leaching (in form of NO3), cNO3 in SW, N content in crop yield and applied N fertilization rates, which are measured for SCI 1 and SCI 2 at Wagna test site. Due to a potential influence of the previous cultivation strategy using higher N fertilization rates, the years 2005 and 2006 are not contained in the results. At the low N input farming system (SCI 1) the average N fertilization rate between 2007 and 2011 was 118 kg/ha/a. The organic farming system neither applied mineral nor farm fertilizer, but instead used cultivation of legumes to fix atmospheric N. The average N content in crop yield was between 104 kg/ha/a at SCI 2 and 116 kg/ha/a at SCI 1. N leaching and NO3-concentrations in SW were generally low at both low N input and organic farming. An exception is the year 2009 where the organic farming system (SCI 2) leached 133 kg/ha/a N towards groundwater. These high leaching rates in 2009 are also visible in Fig. 7, where the total monthly as well as the cumulative monthly N leaching rates are presented. It can also be seen that until 2009 the low N input farming system resulted in higher N leaching than organic farming; after 2009 N leaching of SCI 1 and SCI 2 was almost equal. 3.1.2. HYDRO-lysimeter (grass-reference lysimeter) The combination of the HYDROLYS with a weather station at the same location allows comparing measured ET0 with calculated ET0

Fig. 5. Monthly totals of precipitation (P), seepage water (SW) and evapotranspiration (ETr) of SCIENCELYS 1 (low N input farming) for the period 2005–2011.

G. Klammler, J. Fank / Science of the Total Environment 499 (2014) 448–462


Fig. 6. Monthly totals of precipitation (P), seepage water (SW) and evapotranspiration (ETr) of SCIENCELYS 2 (organic farming) for the period 2005–2011.

according to Allen et al. (1998). Since the calculated ET0 does not consider water stress, a comparison to lysimeter measurements is only valid for conditions without water stress. The comparison in Fig. 8a is presented for the period from June 2006 to April 2008 where water stress was generally absent, except for a few short periods, which are neglected in this evaluation. Especially for the first 14 months it can be seen that both the ET0 calculation using daily as well as 10 min weather data agree very well with ET0 measured at the HYDROLYS. Over the entire period of 23 months the deviation between measured and calculated ET0 is less than 1% for the daily interval method and less than approx. 3% for the 10-minute method. Short-term changes of ET0 during the day cannot be addressed using daily weather data. Fig. 8b illustrates that 10 min weather data is able to reflect the influence of changing weather conditions during the day. This effect is shown for July 10th 2007 where a thunderstorm at about noon reduced measured ET0 and calculated ET0 rates from the 10-minute method, while the calculation method using a daily interval only yields one value per day. The water balance regime of the grass-reference lysimeter (Fig. 9) is similar to the water balance of the SCIENCELYS with agricultural management: highest ETr in June and July, lowest SW in April and July. Table 7 shows that annual average ETr of the cultivated permanent grass cover was 649 mm/a. Annual average SW was 311 mm/a ranging between a minimum of 116 mm/a in 2011 and a maximum of 593 mm/a in 2009. Fig. 10 shows the monthly sums of P, SW and ETr. In comparison to SCI 1 and SCI 2, it can be seen that also in 2007 and 2010 periods without SW flow occurred. The highest monthly SW flow of approx. 120 mm/month was measured in June 2009. It is also obvious that ETr was occasionally rather low as in the summer of 2010 and 2011. Fig. 11 shows the total monthly and the cumulative monthly N leaching rates for HYDROLYS. The average annual N leaching rate was 4 kg/ha/a; the mean NO3-concentration in the SW was 6 mg/l (also given in Table 7). The N leaching rate was generally very low. For example, there was almost no N leaching until July 2008 and

between July 2008 and April 2010 the monthly N leaching rate was 0.7 kg/ha/month. The highest N leaching of approx. 3 kg/ha/month was observed in August 2010 (due to high P). 3.2. Soil temperature (SCI 1 and SCI 2) In Figs. 12 and 13 the soil temperatures in 35, 60, 90 and 180 cm depth are given for SCI 1 and SCI 2, respectively. The temperature is measured inside the lysimeter cylinder as well as outside in the field at the same depths (horizontal distance approx. 3 m). During our investigation period, it can be seen that the temperatures inside and outside complied well in 35, 60 and 90 cm depth for both SCI 1 and SCI 2. The 180 cm depth temperatures inside and outside were similar during spring and summer; however, in fall and winter temperatures inside the lysimeter were up to 3 °C lower than the reference temperature outside in the field. 3.3. Water content (SCI 1 and SCI 2) Figs. 14 and 15 illustrate the measured water contents inside the lysimeter compared to the water content in the field. Due to measuring errors and sensor breakdowns the time series presented are partially discontinuous. With the exception of SCI 2 in summer 2005, it is obvious that water contents inside possessed the same overall behavior as outside the lysimeter. However, differences of more than 10 vol.% occurred in 90 cm depth at SCI 2. Maximum water contents of 40 vol.% could be observed inside SCI 1 in 60 cm depth. 3.4. Tracer test Results in Fig. 16 of the double-tracer test (see description in Section 2.6) in 2005 show that the flow velocity at the outer segment of the lysimeter (near the lysimeter wall) is similar to the velocity in the center for both SCI 1 and SCI 2. While the highest deuterium and bromide-concentrations at the bottom lysimeter outlet of SCI 1 were

Table 6 Annual totals of applied nitrogen fertilizer (F), nitrogen offtake by crop yield (CY), nitrogen leaching (L) and nitrate concentrations in seepage water (cNO3) for SCIENCELYS 1 and 2 from 2007 to 2011. SCIENCELYS 1 (low N input farming)

Fa [kg/ha/a]

CYb [kg/ha/a]

L [kg/ha/a]

cNO3 [mg/l]

SCIENCELYS 2 (organic farming)

Fc [kg/ha/a]

CYb [kg/ha/a]

L [kg/ha/a]

cNO3 [mg/l]

2007 2008 2009 2010 2011 Mean

121 120 52 116 150 112

93 (9) 132 (330) 57 (2) 142 (9) 156 (400) 116

18 20 25 8 4 15

27 45 18 8 29 20

2007 2008 2009 2010 2011 Mean

– – – – – –

143 (10) 125 (400) 93 (3) 99 (9) 62 (400) 104

5 3 133 10 4 31

10 10 89 10 33 45

a b c

Maize W. Barley Pumpkin Maize Triticale

Maize Triticale Pumpkin Maize Triticale

Farm fertilizer reduced by N losses due to volatilization during application (according to BMLFUW, 2006). N in yield corresponds to number of cultivated crops on the lysimeter surface of 1 m2 (given in brackets). N input from legumes not quantified.


G. Klammler, J. Fank / Science of the Total Environment 499 (2014) 448–462

Fig. 7. Monthly totals and cumulative sums of nitrate nitrogen leaching for SCIENCELYS 1 and 2 from 2007 to 2011.

already observed 7 months after the tracer injection, the highest concentrations at SCI 2 occurred only after 10 to 12 months for bromide and deuterium, respectively. Since the amounts of P and ET play a significant role for the residence time it is more meaningful to express the residence time in the form of SW amount required to percolate a certain percentage of the tracer (Fig. 17). At SCI 1 10% of the bromide tracer could already be detected after 90 mm SW; 90% recovery rate was reached after 465 mm. SCI 2 has a higher water storage capacity and, therefore, the band width of residence time was 165 to 560 mm SW between 10% and 90% recovery rates. Under predominant SW rates in 2005 and 2006 the median residence times were 8 months for SCI 1 and of 9 months for SCI 2, which were reached after 230 mm and 310 mm of SW, respectively. Table 8 shows the bromide mass and water balance for the period of the tracer test. It can be seen that the percentages of bromide uptake by vegetation correspond well with the ET as well as percentages of bromide leached go along with the SW rate. 4. Discussion 4.1. Water and nitrogen balance Results presented in this paper show that long term average monthly totals of the climatic water balance of low N input farming (SCI 1) and organic farming (SCI 2) are similar (see Figs. 3 and 4). However, differences of SW and ET between SCI 1 and SCI 2 may occur in certain periods for different reasons. For example, in 2006 ET was approx. 80 mm higher for SCI 2 than for SCI 1 due to different vegetation (maize vs. clover). Even when the same crops were cultivated at SCI 1

and SCI 2 like in 2007 and 2010, ET rates for maize varied between SCI 1 (maximum ET in June) and SCI 2 (maximum ET in July). The higher ET rates in 2008 and 2011 at SCI 2 can be attributed to undersown crops, which prevent a period of bare soil after harvesting the main crop. It can also be seen that after the harvesting of triticale in June 2011 the ET rate decreased continuously until the catch crop in August 2011 started to transpire again. As described in Section 2.5, not all elements for solving the entire nitrogen balance were measured. However, comparing the annual average applied N input due to fertilization of 112 kg/ha/a at SCI 1 (volatilization assumed according to BMLFUW, 2006) with the total N output of 131 kg/ha/a through N leaching and N offtake by crop yields, results in a N deficit of 19 kg/ha/a. Since no legumes were cultivated at SCI 1 this difference can only be attributed to N storage change in the soil and/or atmospheric N deposition. As a consequence, in 2012 an instrument for measuring the total integrated N input was installed at Wagna test site for investigating not only information concerning wet and dry deposition, but also the N uptake directly through stomata of plants (Russow and Weigel, 2000; still under development). N contents in the crop yields may vary between same crops from year to year and may also be different between the two cultivation practices (e.g., maize in 2007 and 2010; Table 6). Reasons for this are that on the one hand crop yields are varying and on the other hand also N contents in the same crops may be different. For example, in 2010 maize yields for SCI 1 and SCI 2 were almost the same, but the N content in maize at SCI 1 was higher leading to higher N removal. Maize at SCI 1 and SCI 2 in 2007 had approx. same N contents per kilogram, but in this year maize yields measured at the lysimeter were significantly higher for SCI 2.

Fig. 8. Grass-reference evapotranspiration (ET0) measured at HYDROLYS compared to ET0 according to the FAO-Penman–Monteith method (Allen et al., 1998) for daily and 10 min time steps: (a) Cumulative ET0 from June 2006 to April 2008 and (b) short-term ET0 for a thunderstorm event on July 10th 2006.

G. Klammler, J. Fank / Science of the Total Environment 499 (2014) 448–462


Fig. 9. Average monthly totals of precipitation (P), seepage water (SW) and evapotranspiration (ETr) of HYDROLYS for the period 2007–2011.

N-input by legumes at SCI 2 was not quantified, but comparing the N leaching rates and N contents in crop yield between low N input and organic farming may suggest that N inputs by legumes are considerable. According to Gisi (1997) lucerne and clover can bring 200–250 kg/ha/a N into the soil; Scheffer and Schachtschabel (2002) assume that the N-fixation potential of lucerne is 225 kg/ha/a. Especially the N leaching rates at SCI 2 in 2009 were rather high, which may be caused by a long precedent period of legumes (undersown in April 2008 to triticale) before the cultivation of oil pumpkin in April 2009. In addition, high groundwater recharge rates due to high P in 2009 further potentiated the N leaching. To reduce this risk of high N leaching rates from oil pumpkin, the sequence of the organic crop rotation was modified and maize was cultivated in 2012 instead of oil pumpkin after triticale. However, both the low N input farming (SCI 1) and organic farming (SCI 2) resulted in overall averages of cNO3 of 20 and 45 mg/l, respectively, between 2007 and 2011. Neglecting the outlier of N leaching in 2009 for the organically management strategy, the average cNO3 in the seepage water would decrease to 11 mg/l. Thus, agriculturally used locations operating beneficial management strategies (even if groundwater cNO3 is not diluted by surface waters) are generally recharged by infiltrating precipitation with a lower cNO3 than the groundwater cNO3 threshold value of 50 mg/l according to the EU Groundwater Directive (EC, 2006). In comparison to the presented results of BMPs between 2005 and 2011, previous suction plate measurements from 1992 to 1996 at the same location showed a cNO3 of 103 mg/l on average for maize monocropping applying 150 kg/ha N-fertilizer and a cNO3 of 136 mg/l for a crop rotation (winter wheat, winter rape, maize, maize, winter barley) applying 143 kg/ha N-fertilizer. For the interpretation of these cNO3 values it has to be considered that the cultivation strategy in that period was without any hardy catch crop for both cultivation strategies (except for the crop rotation in 1993 where legumes have been cultivated from July to October) and that winter wheat, winter rape and winter barley have been also fertilized in autumn before seeding. According to the comparison of measured and calculated ET0, the FAO-Penman–Monteith method (Allen et al., 1998) proves to be valid for the location of Wagna test site, Austria. While the daily interval method is satisfactory for longer investigation periods (e.g., several years), the 10-minute calculation approach is recommended for shorter term investigations, where changing weather conditions during the day influence ET0. In the context of measuring ET0 using a lysimeter, it has to be considered that a changing vegetation (e.g., due to a natural transfer of other grass sorts) will not correspond to the grass-reference evapotranspiration according to Allen et al. (1998) anymore. However, the long term average water balance for a permanent grass cover at HYDROLYS is similar to the SCIENCELYS with low SW rates in April and July and a maximum in September. ET rates are slightly higher than for the SCIENCELYS because of the permanent vegetation without periods of bare soil. N leaching at HYDROLYS where neither N fertilizer

is applied nor legumes are cultivated is very low compared to SCI 1 and SCI 2. At this point the use of external P measurement for solving the climatic water balance is also discussed briefly. As already mentioned in Section 2.4.2 common measurement of rainfall is strongly influenced by P gauge errors. Most significant errors result from wind loss, wetting loss, evaporation loss, and due to in- and out-splashing of water (WMO, 2008). Thus, measured P is generally underestimated. Measuring errors can be reduced by a larger area of the measuring gauge's surface and positioning the collecting vessel at ground level. Modern weighable lysimeters commonly have a surface of 1 m2, are integrated into their typical surroundings of vegetation cover (to avoid oasis effects) and allow scaling the mass change of monolithic soil columns in high measuring accuracy (0.01 mm water equivalent) and high temporal resolution. Thus, also P can be quantified by measuring the positive mass changes of the lysimeter. However, this method implicates external effects (background noise, influence of vegetation and wind) which manifest in the mass time series. While most of the background noise of the weighing is rather well known and can be filtered out of the mass time series, the influence of wind, which blows through the vegetation and affects measured lysimeter mass, cannot be corrected easily. There is no clear relation between wind speeds and the measured outliers of lysimeter mass. Moreover, the influence of wind does not seem to be identical for different lysimeters. Overall, it can be said, that the potential of solving the climatic water balance for calculating ETr using P directly measured at the lysimeter is rather high and would result in a more accurate determination of ETr. However, methods to compensate for external effects on lysimeter weighing (e.g., random noise due to the influence of wind) have to be investigated prior to a global application of lysimeters as P gauges.

4.2. Soil temperature The behavior of soil temperature inside the lysimeter was generally the same as in the surrounding field, except for the temperature sensor Table 7 Annual totals of precipitation (P), evapotranspiration (ETr), seepage water (SW) and soil water storage change (ΔS) for HYDROLYS from 2007 to 2011. HYDROLYS [mm] Year

P [mm]

ETr [mm]

SW [mm]

ΔS [mm]

N-leach [kg/ha]

cNO3 [mg/l]

2007 2008 2009 2010 2011 Mean

892 893 1360 1014 730 978

662 665 760 611 545 649

235 214 593 400 116 311

−6 15 7 3 69 18

1 4 10 6 1 4

2 8 7 6 4 6


G. Klammler, J. Fank / Science of the Total Environment 499 (2014) 448–462

Fig. 10. Monthly totals of precipitation (P), seepage water (SW) and evapotranspiration (ETr) of HYDROLYS for the period 2007–2011.

in 180 cm depth, which showed lower temperatures in fall and winter in both SCI 1 and SCI 2. These differences may derive from inflow of cold air from the surface through the gap at the lysimeter surface (see Fig. 1c). This affects especially the lower part of the lysimeter, where the entire bottom side stays in contact with the air. Furthermore, the lower 10 cm of the lysimeter column is filled with filter gravel which has a different thermal conductivity than native soil between 190 and 200 cm. Moreover, a massive concrete block where the lysimeter is put on (see Fig. 1a) disturbs the original temperature condition at that depth. However, since N turnover rates decrease with lower temperatures and since N turnover processes are not expected at that depth at all, the temporary presence of a lower lysimeter temperature at the bottom is judged to be insignificant for the investigations at Wagna test site. 4.3. Soil water content Soil texture conditions at Wagna test site are very heterogeneous. Even for small horizontal distances between the TDR probes inside and outside the lysimeter, differences of the absolute levels of water contents are obvious. This can be related to variable water storage capacities at the same measuring depths. At SCI2 and 90 cm depth, for example, the water content inside the lysimeter is approx. 10 vol.% higher than outside. However, except for SCI 2 in summer 2005, the relative fluctuations inside and outside were similar for equal measuring depths. The behavior of water content fluctuation is driven by P events and root extraction. The latter causes the typically low water contents in summer (e.g., in 60 cm at SCI 1 and in 60 and 90 cm at SCI 2). While inside of SCI 2 in 90 cm depth the typical water content minimum in

summer 2005 was present, the water content outside remained at the same level during the entire year. This fact can be attributed to the cultivation of oil pumpkin in combination with the location of the TDR probes in the field. Due to rather high row distances of 140 cm, the TDR-probes at SCI 2 were located between two pumpkin rows, while the sensor locations at SCI 1 were directly beneath a row. Furthermore, there was no typical drop of water content in summer 2009 because of the high P rates in that year.

4.4. Tracer test The application of two different tracers (deuterium in the center and bromide at the outer segment of the lysimeter) allows detecting fringe effects of the lysimeter wall on unsaturated flow velocity. In other words, it can be investigated, if during the installation of the lysimeter (lysimeter cylinder is pressed into the soil) the soil conditions near the lysimeter wall were disturbed (e.g., forming preferential flow paths). Comparing the breakthrough curves of deuterium and bromide (Fig. 16) it can be seen, that there is no significant difference of residence time between the center and the outer segment of the lysimeter at Wagna test site, which is an indicator that no fringe effects are present. In comparison to the mean annual SW rates which are similar between SCI 1 and SCI 2, the tracer test confirmed different residence times of water in the soil. This difference results from the transient behavior of water flow and solute transport depending on differences in the thickness of the fine soil, in the different occurrence of macropores and in small differences of unsaturated hydraulic conductivity in soil layers.

Fig. 11. Monthly totals and cumulative sum of nitrate nitrogen leaching for HYDROLYS from 2007 to 2011.

G. Klammler, J. Fank / Science of the Total Environment 499 (2014) 448–462


Fig. 12. Comparison of soil temperatures inside and outside SCIENCELYS 1 for 35, 60, 90 and 180 cm depth.

4.5. Lysimeter design Lysimeters installed at Wagna test site were designed to minimize known lysimeter shortcomings. Oasis effects were not recognizable neither at bare soil nor under growing vegetation due to installing the lysimeter directly into the test plots (lysimeter is surrounded

with same vegetation). Comparing SW behavior of the monolithic SCIENCELYS with backfilled lysimeters (also installed at Wagna test site with same soil and vegetation than SCIENCELYS; not presented in this paper) shows the importance of using undisturbed lysimeters for water balance investigations. Soil tillage does have a significant influence on N mineralization processes in the upper soil. Therefore, the

Fig. 13. Comparison of soil temperatures inside and outside SCIENCELYS 2 for 35, 60, 90 and 180 cm depth.


G. Klammler, J. Fank / Science of the Total Environment 499 (2014) 448–462

Fig. 14. Comparison of water contents inside and outside SCIENCELYS 1 for 35, 60, 90 and 180 cm depth.

lysimeter design allows soil tillage with a plow (or other standard machinery) even on the lysimeter. However, the operation of the lysimeter is more time consuming (upper lysimeter ring has to be removed before tillage), but considering that N turnover is initiated by soil tillage, investigations on N leaching are more realistic. The lysimeter outlet at the bottom is realized with suction cups, which are tension-controlled

(measured in the field) and suck off water so that tensions in 180 cm depth inside the lysimeter comply with tensions outside in the field. In comparison to gravity lysimeters where SW outflow is only possible if saturation is reached, the extraction by suction cups guaranties that outflow from the lysimeter bottom occurs at the same time and magnitude as in the undisturbed field. Moreover, chemical processes

Fig. 15. Comparison of water contents inside and outside SCIENCELYS 2 for 35, 60, 90 and 180 cm depth.

G. Klammler, J. Fank / Science of the Total Environment 499 (2014) 448–462


Fig. 16. Bromide and deuterium breakthrough curves for SCIENCELYS 1 and 2.

(e.g., denitrification) caused by saturated conditions at the lysimeter bottom are avoided with this bottom boundary design. If it is intended to investigate ET at rather small time steps, a determination of the soil water storage change by weighing the lysimeter mass is required. At Wagna test site the lysimeters are situated on three high precision load cells, where the number of load cells has proved to be sufficient. However, the gap between the lysimeter cylinder and the outer cone at the surface must not be blocked by stones, because the friction introduced affects the recorded lysimeter weight and, consequently, ET results. Lysimeters may also be covered by snow in winter. If the snow cover exceeds a height of a few centimeters the weighing is also affected by bridging of the snow over the gap. At Wagna test site this disturbance was avoided by cutting the snow cover right over the gap. Runoff after heavy rainfall events through the lysimeter gap was measured to be only approx. 1% of P. According to this finding a runoff sampling system does not necessarily have to be part of the lysimeter design and runoff can be neglected for precipitation regimes like in Wagna. 5. Conclusions Lysimeter measurements show that the BMPs of (1) low N input farming and (2) organic farming, as cultivated at Wagna test site, Austria, produce cNO3 in seepage water, which were generally below the groundwater nitrate threshold of 50 mg/l established by the EU Groundwater Directive. In other words, the investigated BMPs will on the one hand not increase existing low groundwater cNO3 beyond this threshold and, furthermore, they will decrease groundwater cNO3, which are currently higher than this threshold. Based on this, the N fertilization rates for low N input farming according BMLFUW (2006) for maize (grain), winter barley and triticale as well as 60 kg/ha/a for oil pumpkin (valid for predominant soil conditions at Wagna test site) are recommended for a sustainable groundwater protecting agricultural

land-use. N inputs by legumes at organically managed fields are not easy to control. Rhizobia are able to fix rather high amounts of N which can lead to high N leaching under certain conditions (e.g., oil pumpkin after a long period of legumes). Under consideration of this aspect both cultivation strategies operated at Wagna test site are generally suitable to decrease the recent diffuse N pollution for protecting groundwater quality. The cultivation of grassland without applying any N fertilizer (HYDROLYS) results in insignificant low N leaching rates. N turnover processes are strongly dominated by soil temperature. It is found that the lysimeter design only disturbs the native soil temperature characteristics near the lysimeter bottom in winter (i.e., in depths and periods where usually no N turnover occurs). In general, soil temperature is less sensitive to soil texture than water contents. The absolute level of measured soil water contents at the same depths in Wagna may deviate significantly even within a short horizontal distance. However, the fluctuation of the water content time series for same depths is similar. For the purpose of our study, monolithic lysimeters were deployed to minimize soil disturbance during installation (e.g., through backfilling). This may cause fringe effects (e.g., preferential flow paths) at the lysimeter wall, but a double-tracer test confirmed sufficiently similar flow velocities near the center and near the lysimeter wall. One drawback of the presented lysimeter design is that snow can bridge over the gap next to the lysimeter wall, which may affect the lysimeter weight irregularly and has to be removed manually. It can finally be stated that the design of the presented lysimeters is applicable for investigations on the climatic water balance under different vegetations as well as for determining leaching rates of various nutrients. Results from the Wagna test site provide excellent data for development and calibration of vadose zone models, which can further be applied at aquifer scale for designing measures to reduce N leaching from non-point source pollution.

Fig. 17. Bromide concentration in the seepage water and bromide recovery rate in the leachate of SCIENCELYS 1 and 2.


G. Klammler, J. Fank / Science of the Total Environment 499 (2014) 448–462

Table 8 Bromide mass balance and water balance from April 2005 to September 2007 for SCIENCELYS 1 and 2. Bromide mass balance


Water balance













38.3 100 38.3 100

– – – –

23.9 62 24.1 63

– – – –

13.2 34 13.9 36

– – – –

1.2 4 0.3 1

= = = =

0 [g] 0 [%] 0 [g] 0 [%]

1860 100 1811 100

– – – –

1161 62 1155 64

– – – –

679 37 584 32

– – – –

20 1 72 4

= = = =

0 [mm] 0 [%] 0 [mm] 0 [%]

I—injection, L—leachate, P—precipitation, SW—seepage water, V—vegetation, ΔM—not recovered, ET—evapotranspiration, ΔS—storage change.

Future work will also focus on economic aspects of low N input and organic farming, where profits of these cultivation strategies are compared to common (higher N input) cultivation strategies. Acknowledgment The work was carried out as part of the GENESIS project on groundwater systems (http:/ financed by the European Commission 7FP contract 226536. References Allen RG, Pereira LS, Raes D, Smith M. Crop evapotranspiration (guidelines for computing crop water requirements). FAO irrigation and drainage paper no. 56; 1998 [300 pp.]. Allen RG, Pruitt WO, Wright JL, Howell TA, Ventura F, Snyder R, et al. A recommendation on standardized surface resisdence for hourly calculation of reference ET0 by the FAO56 Penman–Monteith method. Agric Water Manag 2006;81:1–22. AG Boden (Ad-hoc-Arbeitsgruppe Boden). Bodenkundliche Kartieranleitung. Bundesanstalt für Geowissenschcbeaften und Rohstoffe in Zusammenarbeit mit den Staatlichen Geologischen Diensten5. Aufl; 2005 [Hannover, 438 pp.]. ASCE-EWRI. The ASCE standardized reference evapotranspiration equation. Technical committee report to the Environmental andWater Resources Institute of the American Society of Civil Engineers from the task committee on standardization of reference evapotranspiration. 1801 Alexander Bell Drive, Reston, VA 20191–4400: ASCE-EWRI; 2005. [173 pp.]. BMLFUW (Bundesministerium für Land- und Forstwirtschaft, Umwelt und Wasserwirtschaft). Richtlinien für die Sachgerechte Düngung. Anleitung zur Interpretation von Bodenuntersuchungsergebnissen in der Landwirtschaft6. Aufl. Wien: Bundesministerium für Land- und Forstwirtschaft, Umwelt und Wasserwirtschaft; 2006 [79 S.]. EC (European Council). Council regulation (EC) no 834/2007 of 28 June 2007 on organic production and labelling of organic products and repealing regulation (EEC) no 2092/ 91. 20081010:EN:PDF, 2007. [valid on July 27, 2013]. EC (European Council). Commission regulation (EC) no 889/2008 of 5 September 2008 laying down detailed rules for the implementation of council regulation (EC) no 834/2007 on organic production and labelling of organic products with regard to organic production, labeling and control., 2008. [valid on July 27, 2013]. EC (European Council). Directive 2006/118/EC of the European Parliament and of the council of 12 December 2006 on the protection of groundwater against pollution and deterioration. 2006:372:0019:0019:EN:PDF, 2006. [valid on June 19, 2013].

Fank J, von Unold G. High-precision weighable field lysimeter — a tool to measure water and solute balance parameters. Int Water Irrig 2007;27(3):28–32. [Tel Aviv]. Gisi U. Bodenökologie. 2. AuflStuttgart - New York: Thieme; 1997. p. 194–7. Klocke NL, Todd RW, Hergert GW, Watts DG, Parkhurst AM. Design, installation, and performance of percolation lysimeters for water-quality sampling. Trans Am Soc Agric Eng 1993;36:429–35. Lanthaler C, Fank J. Lysimeter stations and soil hydrology measuring sites in Europe — results of a 2004 survey. Proceedings of the 11th Gumpensteiner Lysimetertagung, Lysimetrie im Netzwerk der Dynamik von Ökosystemen. Lehr- und Forschungszentrum für Landwirtschaft Raumberg-Gumpenstein, Irdning; 2005. p. 19–24. Leis A. Chemical, isotopic, and microbiological evidence for nitrification below the plant root zone from intensive fertilized agricultural area in Austria — insights from lysimeter studies and soil cores. IAEA-TECDOC-1618. Application of isotopes to the assessment of pollutant behaviour in the unsaturated zone for groundwater protection. Final report of a coordinated research project 2004–2005Vienna: International Atomic Energy Agency; 2009. p. 15–30. Masarik KC, Norman JM, Brye KR, Baker JM. Improvements to measuring water flux in the vadose zone. J Environ Qual 2004;33:1152–8. Meissner R, Seeger J, Rupp H, Seyfarth M, Borg H. Measurement of dew, fog, and rime with a high-precision gravitation lysimeter. J Plant Nutr Soil Sci 2007;170:335–44. Pütz Th, Kiese R, Zacharias S, Bogena H, Priesack E, Wollschläger U, et al. TERENO-SOILCan Ein Lysimeter Netzwerk in Deutschland. Proceedings of the 14th Gumpensteiner Lysimetertagung. Lehr- und Forschungszentrum für Landwirtschaft RaumbergGumpenstein, Irdning; 2011. p. 5–10. Russow R, Weigel A. Atmogener N-Eintrag in Boden und Pflanze am Standort Bad Lauchstädt: Ergebnisse aus 15N-gestützten Direktmessungen (ITNI-System) im Vergleich zur indirekten Quantifizierung aus N-Bilanzen des statischen Dauerdüngungsversuches. Arch. Acker- Pfl. Boden 2000;45:399–416. Scheffer F, Schachtschabel P. Lehrbuch der Bodenkunde. 15. AuflHeidelberg: Spektrum Akad. Verlag; 2002 [593 S.]. Sutton MA, Howard C, Erisman JW, Billen G, Bleeker A, Grennfelt P, et al. The European nitrogen assessment. Cambridge University Press; 2011 [612 pp.]. UMS (Umwelt-Monitoring-Systeme). Science-lysimeter., 2013. [valid on June 6, 2013]. UMS (Umwelt-Monitoring-Systeme). Hydro-lysimeter., 2013. [valid on June 6, 2013]. WMO (World Meteorological Organization). Guide to meteorological instruments and methods of observation. WMO-no. 8; 2008 [140 pp.]. Wriedt G. Modelling of nitrogen transport and turnover during soil and groundwater passage in a small lowland catchment of Northern Germany [Ph.D. thesis] Germany: University of Potsdam; 2004. Young MH, Wierenga PJ, Mancino CF. Large weighing lysimeters for water use and deep percolation studies. Soil Sci Soc Am J 1996;161:491–501. Zacharias S, Bogena HR, Samaniego L, Mauder M, Fuß R, Pütz T, et al. A network of terrestrial environmental observatories in Germany. Vadose Zone J 2011;10:955–73.