Nitrogen Balance of Effluent Irrigated Silage Cropping Systems in Southern Australia

The nitrogen (N) balance in a double-cropped, effluent spray irrigation system was examined for several years in southern Australia. The amounts of N added by irrigation, removed in the crop, and lost by ammonia (NH3) volatilisation, denitrification, and leaching were measured. Results from the project provide pig producers with the knowledge necessary to evaluate the efficiency of such systems for managing N, and enable sustainable effluent reuse practices to be developed. Oats were grown through the winter (May to November) without irrigation, and irrigated maize was grown during the summer/autumn (December to April). Approximately 18 mm of effluent was applied every 3 days. The effluent was alkaline (pH 8.3) and the average ammoniacal-N (NH4 + NH3) concentration was 430 mg N/l (range: 320 to 679 mg N/l). Mineral N in the 0- to 1.7-m layer tended to increase during the irrigation season and decrease during the winter/spring. About 2000 kg N/ha was found in the profile to a depth of 2 m in October 2000. N removed in the aboveground biomass (oats + maize) was 590 and 570 kg N/ha/year, equivalent to ≈25% of the applied N. Average NH3 volatilisation during the daytime (6:00 to 19:00) was 2.74 kg N/ha, while volatilisation at night (19:00 to 6:00) was 0.4 kg N/ha, giving a total of 3.1 kg N/ha/day. This represents ≈12% of the N loading, assuming that these rates apply throughout the season. The balance of the N accumulated in the soil profile during the irrigation season, as 15N-labelled N studies confirmed. The high recovery of the N-labelled N, and the comparable distribution of 15N and Br in the soil profile, implied that there was little loss of N by denitrification, even though the soil was wet enough for leaching of both tracers.


INTRODUCTION
Commercial pig production requires sound environmental management, particularly in the treatment and reuse of effluent high in nitrogen (N) and salt. When effluent is used for irrigation of crops and pastures, environmental sustainability is judged mainly on its ability to control organic matter, N, other nutrients, salt, and bacteria. For N, sustainable management will only be achieved when the amount exported in the harvested crop and lost to the atmosphere by volatilisation and biological denitrification matches the N-loading (effluent and mineralisation) of the system. If this is the case, the accumulation of large amounts of NO 3 -N within the soil profile and its subsequent leaching to the groundwater will be prevented. However, such practices would contribute to other off-site effects of ammonia (NH 3 ) emissions and global climate change effects from emissions of nitrous oxide. Piggery effluent is saline (2 to 6 dS/m), so the effluent-irrigated areas must be leached to avoid excessive salt build-up in the root zone. In this paper we report results of a project designed to measure and model losses of N by NH 3 volatilisation, denitrification, and leaching from spray-irrigated effluent. The results from the project will assist regulatory authorities in assessing effluent reuse schemes and pig producers in evaluating the cost/benefit of using centre-pivot spray irrigation.

MATERIALS AND METHODS
The site (lat. 36° 1' S., long. 146° 22'E.) had grown one maize crop (97/98) that was irrigated with 680 mm of treated effluent before the work reported in this paper began. Prior to this time the site had been maintained as unimproved pasture. The soil is classified as a Red Dermosol [1]. Clay content increased with depth from 290 mg/kg soil in the surface 0.1 m to a maximum of 690 mg/kg soil in the layer at 1.15-to 1.8-m depth. The pH (1:5 soil:0.01 M CaCl 2 ) of the surface 0.1 m was 4.75 and increased to about 7.0 at 1.15 to 1.8 m. Additional properties of the surface 0.1 m were (g/kg soil): total C, 16; total N, 1.4; cation exchange capacity (cmol (+) /kg of soil), 9.7.
Rainfall, solar radiation, wind run and direction at 2 m, and temperature and relative humidity were recorded every 15 min with an automatic weather station.
Eight intensive measurement sites were established within the area irrigated by the centre pivot (600-m radius). At each site we installed soil solution samplers (0.25-, 0.45-, 1.0-, and 1.4-m depth), neutron probe access tubes, and Br and 15 N subplots (3.4 m 2 ). Crop growth and changes in mineral N in the soil were monitored near each of these sites. Oats were grown during the winter/spring (May to November) and maize during the summer months (December to April). Details of sowing and harvesting dates, irrigation period, and amounts over the period are presented in Table 1. All of the aboveground biomass was harvested for silage. Irrigation with treated piggery effluent started in mid-December after emergence of the summer crop ( Table 1). The irrigator was set so that it took approximately 72 h to complete an irrigation cycle and approximately 18 mm of irrigant was applied in each cycle. Effluent characteristics are given in Table 2. Samples of the crops (oats and maize) were collected from near the eight measurement sites at approximately 3-week intervals throughout the growing season. Sampling consisted of cutting all the aboveground biomass from 1 m of three adjacent rows. The area sampled was 0.9 m 2 for the oats and 2.25 m 2 for the maize. The plant material was divided into green leaf, senescent leaf, stem, and head fractions, dried at 60°C, and weighed to determine dry weight. The dried material was ground to pass a 0.42-mm sieve and retained for total N analysis. Subsamples of the plant material were ground to pass a 0.42-mm sieve and total N content was determined on a Leco CHN analyser (Model CHN-1000). Biomass and total N data are expressed on an oven-dry weight basis.
At the end of each crop, soil cores (≈40 mm i.d.) were taken near each of the eight sampling sites. The cores were sectioned into 0.05-m intervals to 0.2-m depth, 0.1-m intervals from 0.2to 0.5-m depth, and 0.2-m intervals from 0.5-to 1.9-m depth. After thorough mixing of the soil samples, NH 4 + and NO 3 -(including NO 2 -) were extracted by shaking the equivalent of 10 g of dry soil with 100 cm 3 of 2 M KCl for 1 h [2]. The suspensions were filtered and the filtrate stored frozen until analysed. Ammonium was measured by reacting with hypochlorite liberated from dichlor-isocyanurate under alkaline conditions. The absorbance was measured at 660 nm using a dual channel segmented flow analyser [3]. NO 3 was determined by quantitative reduction to nitrite by cadmium metal in a packed-bed cadmium column using a segmented flow analyser [3]. All results were expressed on a oven-dry weight (105°C) basis.
Bromide (Br) and 15 N-labelled (NH 4 ) 2 SO 4 (10 atom %) were applied uniformly to three microplots in the two maize crops (1999 and 2000) and to the 2000 oat crop to allow estimates of losses by denitrification and leaching. The unconfined microplots were 3.4 m 2 (1.5 × 2.25 m) in the maize and 1 m 2 in the oats. The 15 N and Br solution was applied as a liquid to the soil surface of the maize plots using a hand sprayer at the rate of 800 cm 3 /m 2 that added 5 gN/m 2 (10 atom %) and 12.7 g Br/m 2 . Spray marker was applied with the solution to assist uniform application onto the soil surface. For the oats, the tracer was applied by injecting 2 cm 3 of solution on a 5-× 5-cm grid [4]. At harvest, only 0.75 m of the two centre rows of maize was harvested, whereas 0.5 m of the centre two rows was harvested in the oats. Two soil cores (≈40 mm i.d.) were taken from the centre area of each microplot and sectioned as for the mineral N samples. NH 3 volatilisation was measured using micrometeorological techniques during two intensive field campaigns (January 11 to January 29, 1999; February 1 to February 21, 2000). Atmospheric measurements were made at two sites positioned 65 m from the centre of the pivot irrigation boom. In 1999, the instruments and air intakes were positioned at 1.9, 2.9, and 3.9 m aboveground and were lowered and covered during the times that the irrigation boom passed over the sites.
For the second field campaign in February 2000, the sampling strategy was revised to better estimate NH 3 lost to the atmosphere, because water droplets containing NH 3 evaporate when they are ejected from the boom sprayers. Standard 1-D methods were used to measure the NH 3 emission for the effluent irrigated area, whereas a mass balance approach was employed when the irrigation boom was within an arc of 40° of the sampling masts. This required the measurement of profiles of NH 3 concentrations both upwind and downwind of the irrigation boom. The instruments and air intakes were positioned at 0.5, 1.0, 1.8, 3.58, 4.56, 6.06, and 9.21 m aboveground. The measurement stations were shut down, sampling towers lowered, and instruments covered during the times that the irrigation boom passed over the sites. Ultrasonic anemometers measured turbulent wind velocities, and rapid fluctuations in CO 2 and H 2 O concentrations were measured using open-path infrared analysers. Fluxes of water vapour and carbon dioxide were then computed for half-hourly periods using the eddy covariance technique. Profiles of wind speed, atmospheric temperature, and relative humidity were measured over 30-min averaging periods, as well as net radiation, soil temperature, and atmospheric pressure. NH 3 and water vapour mixing ratios averaged over 30 min were measured by Fourier Transformed Infrared Spectroscopic (FTIR) instrumentation [5]. The measurements were combined to calculate NH 3 volatilisation fluxes using micrometeorological theory.

Climatic Conditions During the Measurement Period
In evaluating the results, it is important to know how climatic conditions during the experiments relate to long-term averages. More than 100 years of rainfall data are available from the SILO Point Patched Dataset[6] for Corowa, which was within 5 km of the site [7]. Although rainfall at the measurement sites can be different from that measured in Corowa, especially during the summer, there was a general linear relationship between the two (slope 0.97; r 2 = 0.7). Because of the good agreement, we used the historical data to indicate the general relationship between rainfall during the measurement period and the long-term average (Fig. 1). With few exceptions, monthly rainfall was below average between March 1998 and July 1999; after that, there were many months both above and below average by more than 20 mm, with a slight above average trend (11 compared with 9). The highest monthly rainfall was measured in August 1999, which was 70 mm above average and would have a significant effect on leaching. Rainfall above average by more than 20 mm was recorded for August, September, and November 2000.

Crop Growth and N Accumulation
The biomass of the oats increased rapidly during the spring and at harvest: 9.13 ± 1.83, 12.6 ± 1.8, and 12.9 ± 1.8 t/ha was removed from the site in 1998, 1999, and 2000, respectively. The oat crops removed 200, 260, and 270 kg N/ha; whereas the maize crops removed 390 and 310 kg N/ha (Fig. 2). The lower N removal by the 1999/2000 crop may be due to increased soil salinity that reduced the growth of the maize. Over the measurement period (May 98 to December 2000), 72 t/ha of biomass containing 1430 kg N/ha was exported from the site in silage. These data indicate that for a double-cropping cycle (oat and irrigated maize) 570 to 590 kg N/ha was exported, equivalent to 23 to 26% of the N applied in the effluent.

NH 3 Volatilisation
Volatilisation of NH 3 was highest during and soon after irrigation, consistent with high soil pH values that increase the partial pressure of NH 3 -N at the soil surface. NH 3 fluxes were variable and were lower at night when there was less energy available for evaporation. Average volatilisation rates during daytime (6:00 to 19:00) ranged from 2.7 to 1.2 kg N/ha/day, while volatilisation at night (19:00 to 6:00) varied from 0.1 to 0.4 kg N/ha. The daily emissions (excluding NH 3 emission from the spray droplets) were 3.1 kg N/ha/day in 1998/99 and 1.3 kg N/ha/day in 1999/00. Volatilisation from the boom was estimated to be 1.9 kg N/ha/ day in February 2000. If we assume that the emission from the boom was the same for both years, the total emission of NH 3 was estimated to be 462 and 298 kg N/ha for the 98/99 and 99/00 irrigation seasons, respectively, which is equivalent to about 12 to 18% of the N loading.

Soil Mineral N
Ammonium was the predominant N species in the effluent ( Table 2). Once applied to the soil, the ammonium was rapidly nitrified to NO 3 as evidenced by the low ammonium but high NO 3 -N values in the soil. There was an increase over time in the NO 3 -N concentration in the surface 1 m between May 1998 and May 1999, which is attributed to the below-average rainfall during this time. After that time, the concentrations decreased in the surface but increased at depth (>1 m), suggesting that NO 3 was leaching beyond the root zone of the crop (Fig. 3). This tended to occur during the winter months when rainfall significantly exceeded potential evapotranspiration.
The soil contained about 2250 ± 630 kg of mineral-N/ha to a depth of 1.7 m after the oats had been sown in May 1998 (Fig.  3). The storage increased to a maximum of about 4740 kg N/ha by May 1999, shortly after the maize crop had been harvested. The net increase in mineral N storage between May 1998 and May 1999 mainly occurred in the 0.5-to 1.1-m layer, with little change in the surface 0.5 m and in the 1.1-to 1.7-m layers. In November 1998 it was not physically possible to sample all sites to 1.7 m because the soil was dry. Consequently, data below 1.0 m are biased towards the wetter sites. No effluent was applied during the winter period, but there was net mineralisation of the soil organic matter. Between May 1999 and October 2000, mineral N stored to 1.7 m decreased by ≈2995 kg N/ha; the net decrease was 1950 kg N/ha in the surface 0.5-m layer and 837 kg N/ha in the 0.5-to 1.1-m soil layer. The decrease in the amount of mineral N stored in the profile coincided with the period of above-average winter/spring rainfall (Fig .1). As shown above, uptake and export in the oats and maize accounted for 840 kg N/ha, and 2248 kg N/ha (Table 1) was added in effluent (1999/2000). These data suggest that either leaching or denitrification was responsible for the large decline in mineral N.

Mineral N Losses through Leaching and Denitrification
There is evidence of some leaching of mineral N beyond the root zone of the crops, as evidenced by the downward movement of NO 3 -N during the winter/spring period of 1999 (Fig. 3). Because maize uses more water than rainfall plus irrigation, there is a trend of net accumulation of NO 3 -N during the irrigation season (summer/autumn) in the upper soil layers with limited downward movement. During the winter/spring (non-irrigation period) concentration decreased in the surface soil (Fig. 3) as a result of leaching, crop uptake, and denitrification. However, N balance calculations suggest that the amount of N applied in the effluent and from the breakdown of organic material is far greater than the amount removed in the crop or lost from the system by volatilisation.
The trend of mineral N accumulation during the irrigation season is confirmed from the distribution of 15 N-labelled material in the soil profile (Fig. 4). These data show that most of the 15 N was recovered in the upper part of the soil profile when applied to the 1998/99 maize crop (Fig. 4a, b). In contrast, the 15 Nlabelled material moved to a greater depth during the wet winter of 2000 (Fig. 4c,d), compared with that applied during the irrigation season. Between 94 and 108% of the 15 N applied in January and February 1999 was recovered in the soil and plant. The high recovery of the 15 N during the irrigation season confirms   Table 1). that there was little loss of mineral N during this period. Although data on this is not shown, the Br had a similar distribution to 15 N in the soil profile. The shape of the 15 N distribution with depth ( Fig. 4c) indicates that the tracer applied in December 1999 had been leached to about 0.8 m by October 2000. The high recovery of the 15 N-labelled N, and the comparable distribution of 15 N and Br in the soil profile, implied that there was little loss of N by denitrification, even though the soil was wet enough for leaching of both tracers. It is concluded that leaching was the dominant loss mechanism on these sites.