2001 Cambridge University Press. DOI: 10.1017\S0021859601001186 Printed in the United Kingdom. 235. Emissions of ammonia, nitrous oxide and methane.
Journal of Agricultural Science, Cambridge (2001), 137, 235–250. # 2001 Cambridge University Press DOI : 10.1017\S0021859601001186 Printed in the United Kingdom
235
Emissions of ammonia, nitrous oxide and methane from different types of dairy manure during storage as affected by dietary protein content D. R. K U= L L I N G", H. M E N Z I "†, T. F. K R O= B E R #, A. N E F T E L", F. S U T T E R#‡, P. L I S C H E R$ M. K R E U Z E R#* " Institute of Environmental Protection and Agriculture, Federal Research Station of Agroecology and Agriculture (FAL), Liebefeld, CH-3003 Berne, Switzerland # Institute of Animal Sciences, Animal Nutrition, Swiss Federal Institute of Technology Zurich (ETH), ETH Zentrum\LFW, CH-8092 Zurich, Switzerland $ Constat, Ahornweg 46, CH-3095 Spiegel, Switzerland (Revised MS received 2 February 2001)
SUMMARY In a storage experiment with dairy cow manure, the effects of dietary protein content and manure type on ammonia, nitrous oxide and methane volatilization as well as overall nitrogen (N) loss from manure were investigated. Early-lactating cows received rations with 175, 150 and 125 g crude protein\kg dry matter. Each ration was tested in four manure storage systems reflecting typical farm conditions. These either contained total excreta with high amounts of straw (deep litter manure) or no straw (slurry) or, proportionately, 0n9 of urine and 0n1 of faeces (urine-rich slurry) complemented by the residuals with a low amount of straw (farmyard manure). Manure samples were stored for 7 weeks under controlled conditions and trace gas emission was repeatedly measured. Reduction of N intake decreased daily N excretion and urine N proportion and, on average, led to 0n7-fold lower storage ammonia emission rates on average. Total storage N loss was simultaneously reduced with the extent depending on urine N proportion of the respective manures. A lower dietary protein content furthermore reduced nitrous oxide emission rates in most manure types but increased methane emission from urine-rich slurry ; however, global warming potential (based on trace gas output) of all manures was similar with low and high dietary protein content. In deep litter manure, characterized by the highest C : N ratio, emission rates of total N, ammonia and methane were lowest, whereas nitrous oxide values were intermediate. Substantial emission of nitrous oxide occurred with farmyard manure which also had the highest methane values and, consequently, by far the highest global warming potential. C : N ratio of manure was shown to be suitable to predict total N loss from manure during storage in all manure types whereas urine N proportion and manure pH were only of use with liquid manures. INTRODUCTION Recently trace gas emissions from manure during storage were identified as a problem (e.g. Sommer & Møller 2000). The current inevitable N deposition in wide parts of Europe exceeds the critical nutrient loads of sensitive ecosystems (Eugster et al. 1998) * To whom all correspondence should be addressed. Email : michael.kreuzer!inw.agrl.ethz.ch † Current address : Swiss College of Agriculture, CH3052, Zollikofen, Switzerland. ‡ Current address : Swiss Centre for Agricultural Extension (LBL), CH-8315 Lindau, Switzerland.
threatening endangered species (Ellenberg 1990). Reduced nitrogenous trace gases such as ammonia (NH ) make up a greater part of these emissions than $ oxidized nitrogenous gases. In Switzerland, agricultural processes proportionately contribute to 0n9 of total NH , with cattle being responsible for 0n6 of $ total NH (Menzi et al. 1997 ; Stadelmann et al. 1998). $ Manure storage also significantly contributes to the atmospheric pool of greenhouse gases such as nitrous oxide (N O) and methane (CH ). The global warming # % potential of N O and CH , respectively, is approxi# % mately 310-fold and 23-fold of that of carbon dioxide (IPCC 1996). In Switzerland, N O and CH released # % from manure were estimated to account for 0n1–0n15
236
. . $
and 0n08 of total anthropogenic emission of N O # (Schmid et al. 2000) and CH (Minonzio et al. 1998), % respectively. Reviewing different approaches for reducing NH $ emissions from livestock buildings, Phillips et al. (1999) identified the best options as (i) dietary manipulation and (ii) increasing the C : N ratio by generous use of bedding. In livestock systems, NH $ emissions are closely related to nitrogen (N) excretion of livestock. In particular the use of forage-based rations in dairy cows often results in a high N excretion due to high protein and low energy contents of the diet. Reduction in dietary protein, when without major effect on performance, increases N utilization and simultaneously reduces NH emissions. Apart $ from the reduced total N excretion, this is mainly due to a more than proportional reduction of urine N excretion (Kirchgessner et al. 1991 a ; Van Vuuren et al. 1993 ; Bussink & Oenema 1998). In only a few studies on cows fed with rations of different dietary protein contents have N losses during storage of excreta been quantified (Smits et al. 1995, 1997 ; Paul et al. 1998 ; James et al. 1999), mostly achieved by measurements of NH in exhaust air of cattle units. $ Beside the dietary influence, trace gas emission from dairy manure is also influenced by the manure storage system, which is related to the housing system. While in most stalls with slatted floors only complete slurry is produced, at least part of the excreta is retained as solid manure by adding straw or other waterabsorbing substrates in stalls with solid floors. In traditional tethered housing systems with solid floors using low straw addition, most faeces and a small proportion of urine are retained in a solid manure, which is stacked into anaerobic piles outside the building. The remaining excreta are stored as a urinerich slurry. In modern loose housing systems with solid floors using large straw addition, all excreta are retained in an aerobic, compost-like deep litter manure. For some of these systems of dairy manure treatment, quantitative data are available (Isermann 1990 ; Bussink & Oenema 1998 ; Freibauer & Kaltschmitt 2000 ; Jungbluth et al. 2000) but these systems were not directly compared at similar conditions and interrelationships between manure storage system and feeding schedules were not studied. Furthermore it still remains unclear whether or not feeding or manure-related strategies to reduce NH $ emission are counterbalanced by increased production of the greenhouse gases N O and CH . # % The objective of the present study was to simultaneously evaluate the impact and the interactions of dietary and manure storage attempts to reduce NH emission under controlled experimental con$ ditions. At generally sufficient nutrient supply, diet crude protein contents (g\kg dry matter) were reduced from typically high values, appropriate for summer rations, to medium and extra low contents. From the
ET AL.
resulting excreta, four different types of manure were produced which represented three typical manure storage systems and differed in their urine N contents and C : N ratios. Total N losses as well as gaseous emissions of NH , N O and CH from these manures $ # % during storage were studied in detail. MATERIALS AND METHODS Animal experiment and excreta collection Manure samples for subsequent storage investigations were obtained from 12 Brown Swiss cows in lactation for 7n5p2n2 weeks and weighing 637p34 kg (meanp..) on average in the excreta collection phase. These cows were selected from a more extended experiment (Kro$ ber et al. 2000). Their output was 30n9p2n4 kg milk\day containing 41n5p4n7 g fat\kg and 33n5p2n5 g protein\kg. Four cows were each allocated to three dietary treatments based on live weight and performance. The complete rations contained forage and pelleted concentrate in a ratio of 1 : 1 in terms of dry matter (DM) and were consumed at an amount of 20n6p1n3 kg DM\day. The rations were characterized by a stepwise decrease in crude protein (CP) content from 175 to 150 and 125 g\kg DM. Table 1 presents diet composition ; analysed CP contents closely matched predicted values. The reduction in CP content from 175 g\kg (typical for rations high in rumen-degradable protein) to 150 g\kg was almost entirely achieved by the removal of urea thus avoiding any direct effect of dietary protein source. In contrast, concentrate formulation was varied by replacing proteinaceous feeds mostly by starchy feeds without changing energy content for further reduction to the lowest CP content which remained just beyond the threshold value of about 120 g\kg DM considered to be necessary for optimum rumen fermentation (Roffler & Satter 1975). All rations were designed and allocated in a way which ensured a sufficient supply of net energy, protein and amino acids (Jarrige 1989 ; Rulquin & Ve! rite! 1996). For this reason, all rations were supplemented with 0n75 g rumen-protected methionine (Mepron2 M85, Degussa, Hanau, Germany) per kg. Diet CP content was without significant effect on performance of the cows (data not shown). All animals were tethered in individual stalls complete with slatted floors designed for balance measurements (Kro$ ber et al. 2000). After 2 weeks of adaptation to the experimental rations, faeces and urine were quantitatively collected in week 3. During the collection period urine was diverted from faeces using urinals, which were fixed onto the clipped skin, and a tank collected the urine at the lower end of the urinal. This procedure ensured that urine was practically devoid of faecal particles and so had no tendency to rapid ammonia formation and emission
237
Dairy manure type and dietary protein effects on emissions Table 1. Composition of the complete rations (g\kg dry matter (DM )) Dietary crude protein content
175
150
125
Ingredients Forage 500 500 500 Grass silage (Harvestore2) 250 250 250 Maize silage 200 200 200 Meadow hay 50 50 50 Concentrate 500 500 500 Barley 73 75 45 Cassava – – 165 Sugar beet molasses 20 20 12 Sugar beet pulp 98 100 60 Citrus pulp 54 55 33 Wheat bran 48 50 30 Soybean hulls 85 88 82 Soybean meal 32 32 20 Maize gluten meal 30 30 22 Maize germ 35 35 21 Urea 10 – – Mineral premix* 15 15 10 Methionine† 1n5 1n5 1n5 Analysed composition of the complete ration as consumed (meansp..) DM (related to wet weight) 658p28 654p22 646p21 Organic matter 918p8 915p7 914p7 Neutral detergent fibre 410p9 410p3 383p10 Crude protein 175p4 147p3 124p4 * Supplied per kg concentrate : 5n1 g Ca, 1n5 g P, 1n2 g Mg, 1n2 g Na, 0n3 mg Se, 30 000 IU vitamin A, 6000 IU vitamin D , $ 45 mg vitamin E. † Supplied twice daily in a rumen-protected form together with 0n5 kg crumbled concentrate.
which was confirmed prior to the experiment. For the collection of the faeces sliding trays were used. Urine and faeces were sampled daily and stored separately at 4 mC during the collection period and at k18 mC afterwards. Excreta collected from the four cows per dietary treatment were pooled into two equal portions to provide defined manure for two subsequent series of the manure storage experiment. Manure was remixed before storage based on the proportions of faeces and urine actually excreted. The animal experiment was conducted in accordance with the Swiss guidelines for animal welfare.
Manure storage experiment In order to simulate three typical complete manure storage systems on farms, two liquid manure types (slurry, SL ; urine-rich slurry, USL) and two solid manure types (deep litter manure, DLM ; farmyard manure, FYM) were prepared. SL is typically produced in stalls with slatted floors (tethered or loose housing systems). DLM corresponded to typical deep litter loose housing systems with on average 11n8 (10 to 15) kg straw utilized per cow per day. SL and DLM both contained all excreta. FYM and USL are
typically produced separately in traditional tethered housing systems with solid floors, where the faeces are largely retained by, on average, 1n75 (1 to 2n5) kg straw per cow per day (Menzi & Besson 1991) in FYM which is removed twice daily and stacked outside in anaerobic piles and most of the urine is collected in an extra pit. This is still the most common housing system in Switzerland and is also found in other European countries. As assumed in the Swiss tables of standard values for manure (Walther et al. 1994), USL samples contained 0n9 and 0n1 of the total urine and faeces, respectively, with the remainder being incorporated in FYM samples. The three CP contents were subjected to each of the four manure types resulting in 12 treatments with six replicates per treatment, three in each of two subsequently performed series (total n l 72). In the first series, long straw (length of 16n0p0n8 cm) was used whereas, for practical reasons, chopped straw (5n1p0n2 cm) was applied in the second series. All other conditions inclusive of the initial C : N ratio were kept constant between series. Each manure sample was filled into an open polyethylene bucket of a volume of 10 litres and stored for 7 weeks (49–51 days) at 20 mC and 70 % ambient humidity in a 40 m$ sized laboratory ventilated at 0n33 m$\s. The SL and
238
. . $
USL samples used in the storage experiment consisted of 5 kg excreta and 2n5 kg deionized water in order to simulate a dilution through flushing water as is typical farm practice in Switzerland (Walther et al. 1994). Excreta and water were thoroughly mixed with an electric mixer. FYM samples contained 5 kg excreta and 0n125 kg straw whereas DLM samples were composed of only 2n5 kg excreta (in order to get similar sample volumes in all treatments) and 0n438 kg straw, each corresponding to the commonly applied amounts of straw additions mentioned above. Excreta and straw were thoroughly mixed by hand in a bulk sample. The buckets containing FYM and DLM samples were placed into a heated water bath and subjected to a temperature schedule over the storage period which was derived from the change in temperature over time determined in a preliminary investigation on the self-heating process of a 2i2i0n5 m manure stack containing insulated 10litre buckets filled with FYM. Accordingly starting temperature was set to 41 mC and then linearly declined to 25 mC within the 7 weeks of storage. This procedure was found to be superior to a simple insulation of the buckets with Styropor balls which was not sufficient to maintain the typical self-heating temperature. The temperature of the individual manure samples was recorded every 0n5 h with a thermo logger (Delta-T Logger DLZe ; Delta-T Devices Ltd., Cambridge, UK). Individual FYM and DLM samples were compacted after being filled into the buckets with a weight of 2 kg\cm# in order to better reflect practice conditions as volume affects NH emission rate $ (Schulze Lammers et al. 1997).
Trace gas collection and analysis The chamber technique was employed for measurement of the trace gas emission from the manure samples. The basic principle as developed by Hill (1971) is to measure trace gas emission from a small surface by the mass balance of this gas within the chamber air. This was found to be the most reliable way to compare emission rates of treatment groups located closely beside each other (Ludwig 1994). However, in contrast to micro-meteorological methods, this technique interferes with the emissions process itself which has to be considered (Mosier 1989). In cases of gases with high emission rates Mosier (1989) suggested the use of dynamic chambers (i.e. chambers with a defined flux of ambient or cleansed air) to determine input\output mass balance. For trace gases with low emission rates, closed chambers without air exchange are preferred where the air is gradually enriched with the trace gas. From the resulting gradients, emission rates are calculated. In the current experiment NH was measured in a $ dynamic chamber and, for N O and CH , a closed # %
ET AL.
chamber was used (Fig. 1). Both chambers were constructed from polyethylene buckets. The buckets containing the manure sample and the chamber buckets each had a height of 26 cm, an inner diameter of 24 cm on average and a volume of 11n5 litres with the manure samples using about 8 litres and covering 0n05 m#. NH emission rate was measured by repeatedly $ covering the manure storage buckets with the dynamic chamber. During the first 2 weeks of each series, this was performed on 4 subsequent days and later on 3 days per week. The dynamic chamber was constructed following Ludwig (1994), but improved by prolonging and narrowing the inlet collar for ambient air and turning the ventilation inside the chamber from almost vertical to nearly horizontal. This improved NH $ recovery rate within the chamber from proportionately 0n70 to 0n92. Furthermore, in the recovery trials, the variation in NH emission rate was reduced from $ p25 % to p1n2 % by the introduction of two 10 volt ventilators inside the chamber (Fig. 1). As this procedure also increased NH emission rates by some $ 40 %, these measurements therefore reflected potential losses rather than passive emissions during undisturbed storage. Thus they are suitable particularly for relative comparison of different treatments and increase in NH emission. The completely turbulent $ air stream of 91n6 l\min was passed over the manure surface ; 91 l\min were immediately evacuated by a vacuum pump (TF 71\8 W Korr 51031 ; ASF Thomas, Memmingen, Germany) through a flow meter (G\T 1000, Nr. 1024\MKC23BAL10000, Brooks Instruments B.V., Veenendaal, The Netherlands) and directly removed from the laboratory. The remaining air was pumped through an external converter module and a chemi-luminescent analyser (model 17 & 42C ; Thermo Environmental Instruments Inc., Franklin, MA, USA), which analysed NH concentration. Ammonia emission rate per unit $ of manure surface was calculated from the NH $ concentration gradient to ambient air and the air flow rate after covering buckets for 30 min when NH $ concentration reached a steady state. In order also to be able to quantify total passive N losses reliably, the N balance technique was applied which compares the N amount present at the beginning and the end of storage (Pollet et al. 1998). In addition to the N balance carried out in each sample, an NH balance $ of the storage laboratory was established by continuously recording the NH input and output concen$ trations of the 0n33 m$\s ambient air evacuated. On the basis of both balances, the relative contribution of NH -N to total N loss was estimated as an average $ over all treatment groups. From the N balance data and the cumulated trace gas emissions over 7 weeks, the emissions were calculated which occurred from the daily amount of excreta per cow within the manure storage period.
239
Dairy manure type and dietary protein effects on emissions Dynamic Chamber
Closed Chambers Gas accumulation phase
Gas sample evacuation phase
In
Turbulence
Out to flowmeter and vacuum pump
Manure
Turbulence
Turbulence
Manure
Manure
Out to vacuum pump and gas sampling bag
Out to chemiluminescence detector
Fig. 1. Construction of the dynamic chamber and the closed chamber used to determine trace gas emission rates.
Data on N O and CH emission were only obtained # % in the second experimental series. The closed chamber used for these measurements also was equipped with two 10 volt ventilators. The chamber volume was variable from 11n5 to 15n5 l which allowed the repeated extraction of air samples without interrupting the accumulation. The time intervals between the four successive samples taken varied from 5 to 30 min depending on the anticipated emission rate. When the accumulation was linear, the measurement was considered to represent the actual emission rate of the manure sample. One half of each air sample was injected into a gas chromatograph (CH \Non-CH % % Carbohydrogen Analyser 55C, Thermo Environmental Instruments Inc., Franklin, MA, USA) to determine CH concentration, the other half was % analysed by a tunable diode laser spectroscope (TDLAS 004, Aerodyne, USA) for N O concen# tration. Global warming potential was calculated as CO -equivalents in grams from the gas emissions # measured using a factor of 23 for CH and a factor of % 310 for N O (IPCC 1996). For the present calculation, # emissions of CO from storage were not considered # since in this case CO just represents a step in cycling # following biological fixation without creating additional amounts of CO . # Diet and manure analysis Diets were analysed for proximate contents according to standard methods (Naumann & Bassler 1997). At the start and the end of storage, homogeneous manure samples were taken. These samples, as well as samples
of urine and faeces used to mix manure samples, were analysed for pH (model 632 equipped with the electrode 6n0202n000 containing 3 KCl electrolyte ; Metrohm, Herisau, Switzerland), N (Kjeldahl technique) and total ammoniacal N (TAN ; MgO distillation ; Amberger et al. 1982) within 30 min of mixing. Carbon contents were determined by an automatic analyser (CHN-600 ; LECO Corporation, St. Joseph, MI, USA) in dried (6 h at 80 mC, then 2 h at 105 mC) manure samples. Manure pH was furthermore measured weekly at a depth of 10 cm with a pH meter. The average pH at 10 cm depth was found to reflect closely pH of the mixed manure at the end of storage. Statistical evaluation Data were analysed by ANOVA applying the statistic package Sj (version 4, 1997 ; MathSoft Inc., Data Analysis Products Division, Seattle, WA, USA). For analysis of variance data from all storage weeks and all replicates were used considering dietary CP content, manure type and storage experimental series as effects as well as the interaction of dietary CP content and manure type. Data of both series were combined as interactions, with diet and manure treatments being minor according to a preliminary statistical analysis, although average values differed between series in some variables. As ANOVA supposes homogeneous variances most analyses have been carried out with logarithmically transformed data. However, mean values and ... are given in the untransformed scale. As there are incremental numerical increases in the effect of dietary CP the response to dietary CP or to the interaction of CP and
. . $
240
ET AL.
Table 2. Effect of dietary crude protein (CP) content and type of dairy manure on manure compositon and storage loss* Analysis of variance Diet CP content (g\kg)
Manure†
P
USL
FYM
DLM
Mean
Factor
..
Total
Linear
Quadratic
175 6n01 7n34 150 4n63 4n89 125 3n33 3n19 Mean 4n66 5n14 Urine N‡ 175 2n96 5n64 (g\kg excreta) 150 1n61 3n26 125 0n78 1n67 Mean 1n78 3n52 C : N ratio‡ 175 10 4 150 13 7 125 14 8 Mean 12 6 pH 175 8n29 8n75 150 7n62 8n27 125 7n04 7n41 Mean 7n65 8n14 Manure stored for 7 weeks Total N 175 4n59 3n74 (g\kg excreta) 150 3n78 3n10 125 2n98 2n45 Mean 3n78 3n10 C : N ratio 175 9 7 150 10 8 125 13 10 Mean 11 9 pH 175 6n95 7n40 150 6n89 7n48 125 6n28 7n45 Mean 6n71 7n44 Loss during 7 weeks of manure storage Total N 175 23n7 49n0 (% of initial) 150 18n0 36n7 125 10n7 23n5 Mean 17n5 36n4 Wet weight 175 27n1 40n1 (% of initial) 150 31n3 38n1 125 30n3 35n0 Mean 29n6 37n7 Dry weight 175 26n7 20n0 (% of initial) 150 29n4 23n6 125 16n9 17n0 Mean 24n3 20n2
4n99 4n44 3n61 4n35 0n54 0n29 0n13 9n32 17 16 18 17 7n02 6n38 6n09 6n50
5n11 4n40 4n37 4n63 2n93 1n62 0n78 1n78 25 26 25 25 8n30 8n38 8n22 8n30
5n86 4n59 3n63
CP Manure CPiM
0n075 0n086 0n149
0n001 0n001 0n001
0n001
ns
0n001
ns
8n09 7n66 7n19
CP Manure CPiM
0n088 0n101 0n176
0n001 0n001 0n05
0n001
ns
0n01
ns
4n30 3n61 3n28 3n73 11 10 12 11 7n90 7n83 8n01 7n91
4n51 4n36 3n70 4n19 18 19 24 20 7n83 8n55 8n83 8n40
4n29 3n71 3n10
CP Manure CPiM
0n097 0n112 0n194
0n001 0n001 ns
0n001
ns
0n05
ns
CP Manure CPiM
0n9 1n0 1n7
0n001 0n001 ns
0n001
ns
ns
ns
CP Manure CPiM
0n069 0n079 0n137
ns 0n001 0n001
ns
ns
0n001
ns
11n4 18n2 9n2 12n9 47n5 49n2 48n7 48n5 25n5 30n8 32n1 29n5
10n6 0n3 15n6 8n8 71n7 71n6 69n9 71n1 40n8 43n0 27n5 37n1
23n6 18n3 14n8
CP Manure CPiM
1n21 1n39 2n41
0n001 0n001 0n001
0n001
ns
ns
0n001
46n6 47n6 46n0
CP Manure CPiM
1n57 1n82 3n15
ns 0n001 ns
ns
ns
ns
ns
28n3 31n7 23n4
CP Manure CPiM
2n32 2n67 4n63
ns 0n05 ns
ns
ns
ns
ns
Item Fresh manure Total N (g\kg excreta)
SL
3n02 1n70 0n84 14 15 16
11 12 15 7n52 7n69 7n64
* Values were logarithmized before analysis of variance (except for pH). Means and ... values (.. l 55) refer to nontransformed data. † SL l slurry ; USL l urine-rich slurry ; FYM l farmyard manure ; DLM l deep litter manure. ‡ Only one mixed sample per treatment analysed.
241
Dairy manure type and dietary protein effects on emissions manure (predominantly in their logarithmically transformed form) was assessed by establishing linear and non-linear contrasts. Microsoft Excel (version 5n0, 1993 ; Microsoft Corporation, Redmond, WA, USA) was used for linear regression analysis. RESULTS Total nitrogen loss during storage as affected by dietary protein content and manure type Cows were excreting on average 68 kg manure per day. They lost far lower amounts of N (P 0n001) when dietary CP content was reduced from 175 g\kg (191 g\day) to 125 g\kg (58 g\day). Accordingly, dietary protein supply substantially affected the N contents of the excreta used for the storage experiment in all manure types (Table 2). As most of the variation came from urine N content and its proportion of total manure N, differences between dietary treatments were highest with urine-rich slurry (USL), mostly consisting of urine, and lowest with farmyard manure (FYM), mainly derived from faeces, with intermediate differences in complete slurry (SL) and deep litter manure (DLM). Manure types differed considerably in C : N ratio depending on faeces proportion and straw addition. Within manure types, the C : N ratio was inversely related to manure N content since mainly N varied. Furthermore pH differed within and between manure types with the values linearly declining when dietary CP content decreased. After 7 weeks of storage, manure N content was reduced but still differed (P 0n001) between dietary treatments and manure types. Manure pH also declined during storage but dietary treatments had no longer significant effects at the end of storage. The C : N ratio showed similar differences between dietary treatments as observed initially but was increased in USL and decreased in FYM and DLM. Both dietary CP content and manure type had a significant effect (P 0n001) on storage N loss ( % of initial) and interactions were also significant (Table
N loss (g\kg excreta) l 0n063j0n50 iinitial urine N content (g\kg excreta) 0n36k15n2i1 (initial C : N)
(2)
N loss (g\kg excreta) l
k1n70k28n8i1 (final C : N).
(3)
The regressions were highly significant in all three cases and R# was 0n72, 0n86 and 0n58 for equations 1, 2 and 3, respectively. The correlation between N loss, given as reduction in manure N content relative to initial N content was not significant. Furthermore, using only the solid types of manure, the regression of N loss on initial urine N content was not significant either because of a relatively lower extent of N loss (DLM) or a low urine N proportion (FYM). Initial urine N content, initial C : N ratio and N loss were significantly correlated with initial pH with correlation coefficients of 0n93, 0n89 and 0n80, respectively.
(b) 4 3 2 1
(1)
N loss (g\kg excreta) l
N loss (g/kg excreta)
N loss (g/kg excreta)
(a)
2). In contrast, feeding had no clear effects on weight loss during storage which significantly differed between manure types. Overall, but not in all manures and CP contents, relative N loss was lower with decreasing dietary CP content. The group differences were high with SL and USL, with the loss found with 125 g CP\kg to be only about half of that occurring with 175 g CP\kg, and weak with FYM. DLM showed an unexpectedly dramatic decrease with 150 but not with 125 g CP\kg relative to 175 g CP\kg dietary DM. N losses during 7 weeks of storage accounted for about half of total urine N amount with SL and USL. In the two solid manures no clear relationship to urine N proportion was obvious. In all treatment groups, total N loss during storage clearly and linearly increased with initial urine N content and inversely to the carbon : nitrogen ratio of the manures (Fig. 2). The relationships are described by the following equations :
5 4 3 2 1 0
0 0
2
4
Initial urinary N (g/kg excreta)
6
0
0·1
0·2
0·3
Initial (C:N)–1 ratio
Fig. 2. N-loss over all manure types within 7 weeks of storage depending on their initial urine N-content and on C : N ratio. 4, slurry ; , urine-rich slurry ; >, farmyard manure; $, deep litter manure. Bars represent standard errors.
. . $
242
ET AL.
Table 3. Effect of dietary crude protein (CP) content and type of dairy manure on trace gas emissions from manure* Analysis of variance
Item
Diet CP content (g\kg)
Average rates of emission NH (µg\m#\s) 175 $ 150 125 Mean N O† (ng\m#\s) 175 # 150 125 Mean N O† 175 # (% of N loss) 150 125 Mean CH † (µg\m#\s) 175 % 150 125 Mean Emission from daily excreted Total N (g) 175 150 125 Mean N O† (mg) 175 # 150 125 Mean CH † (g) 175 % 150 125 Mean GWP§ (eq) 175 150 125 Mean
Manure SL
USL
FYM
P DLM
135 246 181 87 71 129 110 67 42 67 28 31 83 147 106 62 9n1 14n4 787n9 11n4 6n0 17n4 748n7 13n5 1n7 4n4 132n7 3n0 5n6 12n1 556n4 9n3 0n03 0n02 24n83 0n54 0n02 0n04 6n85 2n78 0n02 0n02 7n89 0n04 0n02 0n03 13n19 1n12 3n01 1n89 8n20 0n10 3n18 4n26 14n98 0n98 3n45 9n71 8n38 0n37 3n21 5n29 10n52 0n48 amounts per cow during 7 weeks 90n9 104n7 38n4 58n3 80n7 2n8 24n8 31n2 46n1 58n0 72n2 29n1 18 1157 46 14 1249 70 5 251 11 12 886 42 8n3 16n6 0n6 9n7 34n4 6n0 10n9 27n7 2n4 9n6 26n2 3n0 196 740 28 227 1178 160 252 715 59 225 878 82
Mean
Factor
SE
Total Linear Quadratic
163 94 42
CP Manure CPiM
8n2 9n5 16n4
0n001 0n001 ns
0n001
ns
ns
ns
205n7 196n4 35n4
CP Manure CPiM
36n38 42n01 72n76
0n001 0n001 ns
–
–
–
–
6n36 2n42 1n99
CP Manure CPiM
0n094 0n001 0n108 0n001 0n188 0n001
–
–
–
–
3n30 5n85 5n48
CP Manure CPiM
1n034 0n01 1n193 0n001 2n067 0n01
–
–
–
–
4n67 5n40 9n35
0n05 0n001 0n05
0n05
ns
0n05
ns
CP 58n3 Manure 67n4 CPiM 116n7
0n01 0n001 ns
–
–
–
–
CP Manure CPiM
0n001 0n001 0n05
–
–
–
–
0n01 0n001 ns
–
–
–
–
of storage‡ 78n0 CP 47n3 Manure 34n0 CPiM 407 444 89 8n5 16n7 13n7 321 522 342
2n17 2n50 4n35
CP 53n2 Manure 61n5 CPiM 106n4
* Values were logarithmized before analysis of variance (except for pH). Means and ... values (nitrous oxide and methanerelated variables : .. l 19 ; for other data : .. l 55) refer to non-transformed data. † Nitrous oxide and methane measurements only from the second series are presented in this table ; calculation of contrasts not possible. ‡ Values of urine-rich slurry are combined with those of farmyard manure as the two components of one system. § GWP (eq) l Global warming potential of CH and N O emissions expressed as equivalents of g CO .
%
#
Emission of individual trace gases during storage as affected by dietary protein content and manure type Within 1 week of storage at 20 mC, 900 g\kg of urine N was found to have been converted into ammonium N. Ammonia volatilization was responsible for the major proportion of total gaseous N loss during the 7 weeks of storage. The reduction of dietary CP content
#
from 175 to 125 g\kg significantly (P 0n001) reduced NH emission rate in every manure type (linear after $ logarithmic transformation of data ; Table 3). In detail, the proportionate reduction was 0n68, 0n73, 0n85 and 0n64 for SL, USL, FYM and DLM, respectively. As expected from the steadily decreasing ammonium concentration in manure and the constantly growing surface crust, generally an exponential decrease with
243
Dairy manure type and dietary protein effects on emissions Urine-rich slurry
Slurry 3 NH3 (log g/m2/s)
NH3 (log g/m2/s)
3
2
1
0
2
1
0 0
7
14
21
28
35
42
49
0
7
Storage time (days)
21
28
35
42
49
42
49
Storage time (days)
Deep litter manure
Farmyard manure 3
3 NH3 (log g/m2/s)
NH3 (log g/m2/s)
14
2
1
2
1
0
0 0
7
14
21
28
35
42
49
Storage time (days)
0
7
14
21
28
35
Storage time (days)
Fig. 3. Effect of dietary crude protein content ($, 175 g\kg ; , 150 g\kg; >, 125 g\kg) and manure type on the development of ammonia emission rate (logarithmized) during storage.
time was found in the change in NH emission rate $ (Fig. 3). The slope of regression was greatest for DLM and small for SL and USL indicating a high persistence in the emission rates during the storage period for these manure types. All manures showed clear differences in the extent of emission between 175 and 125 g CP\kg throughout the storage period, whereas differences between 175 and 150 g CP\kg were less clear, particularly with FYM. In detail, NH $ emission (y ; log µg NH \m#\s) significantly $ (P 0n001 ; P 0n01 for 125 g CP\kg) varied with time (x ; days) in all treatments. The regression equations, in the order of 175, 150 and 125 g CP\kg dietary DM, were : y l 1n671k0n007 x, y l 1n431k0n017 x and y l 0n890k0n011 x for SL ; y l 2n144k0n010 x, y l 1n820k0n012 x and y l 1n263k0n005 x for USL ; y l 1n746k0n027 x, y l 1n727k0n028 x and y l 0n975k0n019 x for FYM ; y l 2n495k0n061 x, y l 1n904k0n048 x and y l 1n014k0n032 x for DLM. As the NH emission $ measurement represented a potential rather than an effective N loss, the extent of NH emission was 4n0, $ 2n9, 2n5 and 20n1 times higher than total N loss
calculated from the variation in N balance measured in SL, USL, FYM and DLM, respectively. With SL, USL and, mostly, DLM, N losses as N O # were negligible at clearly less than 1 % of total initial N (Table 3). However, in FYM N O made up a # proportion of 0n07 to 0n25 of total N with the highest values found with 175 g CP\kg. On average N O # emission rates were only 0n01, 0n02 and 0n02-fold of that of FYM with SL, USL and DLM, respectively. In all manure types N O emission rate was lower with # 125 than with 175 g CP\kg whereas this was not always the case with the reduction from 175 to 150 g CP\kg (Table 3, Fig. 4). Overall the proportionate reduction feeding the low-protein ration accounted for 0n81, 0n69, 0n83 and 0n74 of the values found with 175 g CP\kg with SL, USL, FYM and DLM, respectively. The values noted with 150 g CP\kg were either intermediate (SL) or similar to those found with 175 g CP\kg (other manure types). As expected substantial N O emission was only found after about # 3 weeks of storage had passed (Fig. 4) which is the time when NH emission rate declined. Measurements $ therefore concentrated on the period between week 4
. . $
244
ET AL.
Urine-rich slurry
Slurry 100
40
85 N2O (ng/m2/s)
N2O (ng/m2/s)
30 20 10
70 55 40 25 10
0 0
7
14
21
28 35
42 49
–5
–10
0
7
Storage time (days)
1600
30
400
35 42 49
Deep litter manure 40
N2O (ng/m2/s)
N2O (ng/m2/s)
Farmyard manure
800
21 28
Storage time (days)
2000
1200
14
20 10 0 0
0 0
7
14
21 28 35 42 49
Storage time (days)
7
14
21 28 35 42
49
–10 Storage time (days)
Fig. 4. Effect of dietary crude protein content ($, 175 g\kg; , 150 g\kg; >, 125 g\kg) and manure type on the development of nitrous oxide emission rate during storage. Bars represent standard errors.
and 7 of manure storage. After 6 weeks of storage N O emission from FYM and DLM reached a # maximum. By that time, DLM in some cases simultaneously acted as a sink for NH and as a $ source for N O whereas SL generally acted as a sink # for N O before it was covered by a solid crust after # about 3 weeks. The diet effects largely differed between manure types in the case of CH emission from manure (Table % 3). With SL, FYM and DLM there was no clear response to feeding in CH emission whereas with % USL the rate significantly increased with decreasing dietary CP content. The extent of CH emission was % high in FYM and low in DLM (even considering that only half of the excreta amount was used in this treatment) with intermediate values found in SL and USL. Similar to N O, CH emissions from SL and # % USL did not reach noticeable amounts before 3 weeks of storage had passed whereas with DLM and FYM emission started immediately and a maximum (DLM at 1 week, FYM at 5 to 6 weeks of storage) was reached before the end of storage (Fig. 5). The latter was only partially the case with SL and USL where in particular the low-protein treatment (125 g CP\kg)
seemed to have caused a delayed and still substantial CH emission after 7 weeks of storage. No other clear % dietary effects on the increase of CH emission rate % were obvious. DLM even acted as a CH sink in half % of the cases after 4 weeks of storage. Absolute losses and global warming potential as affected by diet and manure type In order to be able to compare manure storage systems approximately on the same basis, FYM and USL were combined for the calculation with the daily excreta per cow (Table 3) as they are two components of one system. The amount of N and N O volatilized # from daily excreted amounts per cow during the 7 weeks of storage were proportionately reduced with decreasing dietary CP content from 175 to 125 g CP\kg by about 0n75 for the SL and the FYM\USL system, whereas there was no systematic variation in the DLM system. Total N loss per cow was lowest in DLM while N O emissions per cow were lowest in SL. # Both were highest in FYM together with USL. The global warming potential (GWP) of trace gases emitted from stored manures and expressed as CO #
245
Dairy manure type and dietary protein effects on emissions Slurry
Urine-rich slurry
25
35 30 CH4 (g/m2/s)
CH4 (g/m2/s)
20 15 10 5
25 20 15 10 5 0
0 0
7
14
21
28
35
0
42 49
7
Storage time (days)
21
28 35
42 49
Storage time (days)
Farmyard manure
5
50
Deep litter manure
4 CH4 (g/m2/s)
40 CH4 (g/m2/s)
14
30 20
3 2 1
10 0 0
0 0
7
14
21
28
35
42 49
Storage time (days)
7
14
21 28
35
42
49
–1 Storage time (days)
Fig. 5. Effect of feed crude protein content ($, 175 g\kg; , 150 g\kg; >, 125 g\kg) and manure type on the development of methane emission rate during storage. Bars represent standard errors.
equivalents per cow per day differed considerably between manure storage systems. GWP was by far lowest in the DLM system and highest in the FYM\USL system. DISCUSSION
experiment and farm practice. From control measurements varying the volume to surface ratio, it seems that there was a certain over-estimation of NH and $ N O emissions, but not of CH emission. # %
Manure storage types
Influence of dietary protein content and manure composition on trace gas emissions
In the present study, an attempt was made to quantify the effects of dietary CP reduction on trace gas emission from manure under various storage conditions. The use of the same defined excreta for all storage forms allows a comparison of the relative effects of dietary treatments. Direct comparison of the manure types is basically only possible when data are related to the same unit as has been done when relating data to total amount of manure excreted per cow per day. The model storage methods were designed to reflect those common in farm practice as closely as possible. Nevertheless there were some constraints which limit the applicability of the absolute extents of emission. For instance, the ratio of volume to surface may differ between this laboratory
Protein reduction in the nutrition of dairy cows is known to decrease N excretion substantially, particularly urinary N (Kirchgessner et al. 1991 a ; Bussink & Oenema 1998) containing mostly urea which is rapidly converted to ammonia and so highly susceptible to volatilization. Accordingly, the use of rations with lower CP contents is considered to be an efficient means of reducing gaseous N and NH loss $ from manure. In the present study, gaseous N loss from slurry during 7 weeks of storage was reduced by 15 g\kg total N per 1 g CP less per kg diet. This amount was in the range of 10 to 20 g less gaseous N\kg total N found by Smits et al. (1995), Paul et al. (1998) and James et al. (1999). The present study, for the first time, compared these effects in different
246
. . $
manure types and also included measurement of N O # and CH . The reduction of the dietary CP content % from 175 to 125 g\kg decreased NH and total N $ emissions considerably both in complete slurry and the FYM\USL system, whereas the effects remained unclear in deep litter manure although at a low overall extent of NH and total N emission. This $ suggests that the full advantage of low-protein rations is only apparent in systems with a relatively high emission rate. Except for deep litter manure, the extent of the effect of the variation in dietary CP content clearly depended on the urine N proportion of the respective manure, thus being most effective in urine-rich slurry followed by complete slurry and farmyard manure (mostly faeces). Kirchmann & Witter (1989) reported N immobilization in aerobic (deep litter manure) but not in anaerobic (farmyard manure) solid manure and a decrease of net N mineralization with increasing C : N ratio in manure. Presumably these factors are the reasons for the generally low emission rates found with deep litter manure. NH emission measurements indicated 20 $ times higher potential N losses in deep litter manure than had been effectively measured according to the N balance. However, even the potential N losses found with deep litter manure remained below the effective losses occurring in all other manure types. As with NH emissions, N O emissions in most $ # treatments were positively related to dietary CP content. This was partially expected because both NH and N O formation depend on N availability. $ # However, a higher dietary CP content also slightly enhances the digestibility of the diet, particularly of fibre (Kro$ ber et al. 1999), and so was to some extent negatively related to manure fibre content and to crust formation which enhances N O formation. As # NH emissions already decreased when N O emissions $ # were still rising in the two slurries, the proportion of N O-N of total N losses could be expected to increase # with a further extension of storage time. N O # emissions only played a significant role for total N loss in the FYM\USL system where these losses almost exclusively came from the farmyard manure component. N O emissions were only slightly in# creased in the deep litter manure system compared to the slurry system, possibly because, due to the even more aerobic conditions, the higher N immobilization never supported a high N emission. In contrast to gaseous N losses, methane emissions were either not reduced or even slightly increased by dietary protein reduction. A direct effect of dietary CP content on methanogenesis can be generally excluded from data obtained in respiratory chambers with cows (Kirchgessner et al. 1991 b). One obvious reason for this effect would be the above-mentioned slight reduction in digestibility of fibre as the main substrate for methanogenesis with low CP contents. However, the diet effect on methane emission rate was
ET AL.
only clearly occurring in urine-rich slurry which contained almost no faeces and fibre at all. As methane formation can be inhibited by NH (Hansen $ et al. 1998) the considerable reduction in ammonium content with dietary CP reduction in urine-rich slurry may actually explain this phenomenon. This is further supported by the development of NH and CH $ % emissions with storage time with earlier CH increase % in deep litter manure showing a particularly rapid decline in NH emission. A non-linear effect of dietary $ CP content on CH emission in the solid manure % types could have resulted from the change in diet components between the 175\150 g CP\kg groups on one hand and the ration containing 125 g CP\kg on the other hand where a higher proportion of fermentable carbohydrates was available. However, in these manures, the straw addition should have provided sufficient available carbon and microbial NH binding capacity to overrule any greater influence $ of diet composition on CH formation. % The effects of diet and particularly of manure storage form were found to be of importance for the global warming potential per unit of livestock. From the present data, deep litter manure may be superior to slurry and particularly the FYM\USL manure system. The results for the trace gas emissions are, furthermore, of interest for the estimation of feeding effects in composting systems which in part resemble the farmyard manure and the deep litter manure in the current study. Although composting allows a strong reduction in both fresh weight and dry matter of substrates and production of a cleaner product of neutral pH within a short time, it is often reported to cause a highly variable N balance of between 77 % loss and 14 % gain (Meyer & Sticher 1983 ; Schuchardt 1990 ; Martins & Dewes 1992 ; Eghball et al. 1997). Favourable conditions for a low N loss are C : N ratios above 25, temperatures below 50 mC, dry matter contents at around 250 g\kg, oxygen contents between 10 and 15 % and air contents of above 50 % (Hu$ mbelin et al. 1980 ; Kirchmann 1985 ; Schuchardt 1990 ; Maeda & Matsuda 1997 ; Glenn 1998). This may explain the particularly low emission rates found with deep litter manure which, among the manures investigated, fits these conditions best. Furthermore, the high proportion of litter increasing C : N ratio (Dewes 1996 ; Hu$ ther et al. 1997) and the compacting of the samples restricting the extent of NH diffusion $ (Schulze Lammers et al. 1997) may specifically have assisted in reducing NH emissions from deep litter $ manure, although estimated NH volatilization from $ dairy housing systems was assumed to be similar with and without use of litter from comparison of the NH $ concentrations in air in the survey of Seedorf & Hartung (1999). However, straw addition to manure may increase CH and N O emissions, unless aeration % # is limiting CH release (Hu$ ther et al. 1997). In the % present study a similar antagonism between NH on $
Dairy manure type and dietary protein effects on emissions one hand and N O and CH emission on the other # % hand was observed in deep litter manure between the groups receiving 175 or 150 g CP\kg but not with 125 g CP\kg where all three trace gases were simultaneously minimized.
Manure attributes as indicators to predict N loss from manure Since dietary CP primarily affected the urinary but, less so, the faecal N excretion, N loss particularly from the slurries depended on initial urine N content, but not on total N content. As also reported by Whitehead & Raistrick (1993) the majority of initial urine N was hydrolysed within 1 week of manure storage at 20 mC to ammoniacal N, which is potentially volatilizable. Therefore, similar to Maeda & Matsuda (1997) who related N losses to initial manure ammoniacal N and Smits et al. (1995) who related NH emissions to urinary urea, a close relationship to $ urine N proportion was found in the current study. However, urine N content of manure was only satisfactorily describing N loss when urine made up a major proportion and when straw addition simultaneously was low. Similarly, pH turned out not to be suitable as an indicator for the straw containing manures. In these manures the change of pH over storage time is not primarily caused by NH emissions, $ but is due to their CO loss (Sommer & Husted 1995). # Under conditions of the current study, the initial C : N ratio of the manure appears to be a good indicator to predict N loss for a storage period of 7 weeks over all types of manure inclusive of the solid forms. Although adding a large amount of straw to excreta (DLM) initially multiplied the cumulative N loss, these losses were minimized from week 4 of storage onwards, as compared to any other treatment. Thus, the larger amount of available carbon must have enhanced both the initial ammonification process and the later incorporation of N into microbial biomass. The regression equation of N loss on C : N ratio suggests that a net N gain instead of N loss can be expected when the C : N ratio exceeds 43. Indeed, a net N gain occasionally occurred even at a lower C : N ratio in farmyard manure (two out of nine samples) and in deep litter manure (three out of nine samples). The observed net N gain occurred at an ambient NH $ concentration of about 300 ppb. The present results are consistent with the findings of other studies, where NH emissions of manures decreased by a factor of $ three when the C : N ratio was increased from 14 to 25 (Schuchardt 1990 ; Beck et al. 1997 ; Maeda & Matsuda 1997). Furthermore the current results are consistent with those of five out of 12 cases described by Maeda & Matsuda (1997), where no NH emissions $ occurred at all when the C : N ratio of the manure was above 40. In another study on farmyard manure
247
storage (Amon et al. 2000) frequent turning of the heaps reduced greenhouse gas emissions by a factor of seven but multiplied NH emissions during storage by $ a factor of six. However, the C : N ratio of manure in that study was rather low (14). Consequences for farm practice Animal waste is the major source of NH loss from $ agriculture. In the present experiment, a proportionate reduction of dietary CP content by less than 0n3-fold reduced NH emissions during manure storage on $ average by more than 0n7-fold. By adding straw (deep litter system) instead of water (slurry) to the excreta, total N loss was even lower by more than 0n9-fold. Accordingly, dietary CP reduction and changing manure type in order to reduce NH emissions, both $ without consequences for animal productivity (Kro$ ber et al. 2000), were at least partially additive with a maximum achievable reduction during manure storage of more than 0n85-fold. To minimize trace gas emissions from manure storage in practice, dairy cows therefore are ideally kept in deep litter housing systems applying demand-oriented individual feeding schedules. However, focus has to be extended from manure storage to whole farm N utilization rate (Paul et al. 1998) as a reduction of N loss in one component may be compensated by other components of the system (Bussink & Oenema 1998), particularly by postponing the losses from storage to spreading of manure with a higher content of residual N (Saether 1997). The impact of NH reduction strategies on $ whole farm N efficiency depends on the component of the system to which the strategy is applied. Factors reducing the input of dietary CP have a seven-fold greater effect on whole farm N efficiency than those which reduce N loss just at the manure storage level (Kohn et al. 1997). Finally, it is important that a lower dietary CP input is not counterbalanced by a higher N input via mineral fertilizers. CONCLUSIONS The current study showed that N loss from manure as NH and N O can be considerably reduced either by $ # avoiding a dietary CP surplus or a manure carbon deficiency. In regions using mainly forage-based rations, for instance in mountainous areas which are not suitable for arable crops, a reduction of NH $ emissions by half or more would, however, require a shift to concentrate-based rations. The feed input into these farms would endanger the regional nutrient balance and would counteract the important role of cattle in producing food of high nutritive value from forage which cannot be used for human consumption. The emission reduction potential using rations consisting mostly of forage still remains to be explored.
248
. . $
Storage system comparison in the current study favoured manure types with high or without straw addition. The loose housing system with deep litter resulted in the lowest emissions of environmentally hazardous trace gases, a housing system which is currently also promoted for animal welfare reasons. However, as the persistent emission after 7 weeks of storage indicates, the result of storage system comparisons is strongly influenced by the overall duration of storage. Finally, for rapid assessment of the susceptibility of manures to emissions of total N during storage, the C : N ratio appears to be the most appropriate attribute to be used over all manure types
ET AL.
whereas, in manure forms rich in urine, urine N proportion and pH of manure could also be useful for assessment. We are grateful to Dr H. Leuenberger, A. Felder, A. Blatter, M. Fahrni and M. Schmid for their help in organizing and performing the experiment at the ETH research station Chamau and at the Institute of Environmental Protection and Agriculture. We also owe thanks to H. Shariat-Madari, W. Stauffer, J. Aeberhard and O. Fankhauser for assistance in the laboratory.
REFERENCES A, B., A, T. & B, J. (2001). Emissions of NH , N O and CH from milking cows housed in a $ # % farmyard manure tying stall. Nutrient Cycling in Agroecosystems, Special Issue. A, A., V, K. & G, R. (1982). Stickstofffraktionen verschiedener Gu$ llen und deren Wirkung im Pflanzenversuch. Zeitschrift fuW r PflanzenernaW hrung und Bodenkunde 145, 283–294. B, J., K$ , M., H, A., C, K. & J, T. (1997). Ammonia emissions from composting animal wastes in reactors and windrows. In Ammonia and Odour Emissions from Animal Production Facilities, International Conference Proceedings (Eds J. A. M. Voermans & G. J. Monteny), pp. 381–388. Vinkeloord, The Netherlands : Dutch Society of Agricultural Engineering. B, D. W. & O, O. (1998). Ammonia volatilization from dairy farming systems in temperate areas : a review. Nutrient Cycling in Agroecosystems 51, 19–33. D, T. (1996). Effect of pH, temperature, amount of litter and storage density on ammonia emissions from stable manure. Journal of Agricultural Science, Cambridge 127, 501–509. E, B., P, J. F., G, J. E. & D, J. W. (1997). Nutrient, carbon, and mass loss during composting of beef cattle feedlot manure. Journal of Environmental Quality 26, 189–193. E, H. (1990). O= kologische Vera$ nderungen in Biozo$ nosen durch Stickstoffeintrag. In Ammoniak in der Umwelt. KreislaW ufe, Wirkungen, Minderung, Conference Proceedings (Eds J. Hartung, M. Paduch, S. Schirz, H. Do$ hler & H. Van den Weghe), pp. 44n1–44n24. Darmstadt, Germany : KTBL. E, W., P, S., W, H., L, A., L, M., R, M., G$ , P., G, M. & S, J. (1998). Spatial variation in annual nitrogen deposition in a rural region in Switzerland. Environmental Pollution 102, 327–335. F, A. & K, M. (E) (2000). Emission Rates and Emission Factors of Greenhouse Gas Fluxes in Arable and Animal Agriculture. European Summary Report of the EU Concerted Action ‘ Biogenic Emissions of Greenhouse Gases Caused by Arable and Animal Agriculture ’ (FAIR3-CT96-1877). Stuttgart, Germany : Institut fu$ r Energiewirtschaft und rationelle Energieanwendung, Universita$ t Stuttgart.
G, J. (1998). Dairy farm thrives with manure composting. Biocycle 39, 40. H, K. H., A, I. & A, B. K. (1998). Anaerobic digestion of swine manure : inhibition by ammonia. Water Research 32, 5–12. H, A. C. (1971). Vegetation : A sink for atmospheric pollutants. Journal of Air Pollution Control Assessment 21, 341–346. H$ , M., M, M. & S, H. (1980). StickstoffFixierung bei der Kompostierung von Rindermist. Schweizerische Landwirtschaftliche Forschung 19, 85–90. H$ , L., S, F. & W, T. (1997). Emissions of ammonia and greenhouse gases during storage and composting of animal manures. In Ammonia and Odour Emissions from Animal Production Facilities, International Conference Proceedings (Eds J. A. M. Voermans & G. J. Monteny), pp. 327–334. Vinkeloord, The Netherlands : Dutch Society of Agricultural Engineering. IPCC (I P C C) 1996 Climate Change 1995 – The Science of Climate Change (Eds J. T. Houghton, G. J. Jenkins & J. J. Ephraums). Cambridge : Cambridge University Press. I, K. (1990). Ammoniakemissionen der Landwirtschaft als Bestandteil ihrer Stickstoffbilanz und Lo$ sungsansa$ tze zur hinreichenden Minderung. In Ammoniak in der Umwelt. KreislaW ufe, Wirkungen, Minderung. Conference Proceedings (Eds J. Hartung, M. Paduch, S. Schirz, H. Do$ hler & H. Van den Weghe), pp. 1n1–1n75. Darmstadt and Du$ sseldorf, Germany : KTBL and VDI. J, T., M, D., E, E., D, E. J. & P M, H. (1999). Effects of dietary nitrogen manipulation on ammonia volatilization from manure from Holstein heifers. Journal of Dairy Science 82, 2430–2439. J, R. (Ed) (1989). Ruminant Nutrition. Recommended Allowances and Feed Tables. Paris : INRA. J, T., H, E. & B, G. (2001). Greenhouse gas emissions from animal husbandry. Nutrient Cycling in Agroecosystemss, Special Issue. K, M., W, W. & K, M. (1991a). Stickstoffemission laktierender Milchku$ he u$ ber die Gu$ lle in Abha$ ngigkeit von der Leistungsintensita$ t. Agribiological Research 44, 1–13. K, M., W, W., M$ , H. L. &
Dairy manure type and dietary protein effects on emissions K, M. (1991b). Release of methane and of carbon dioxide by dairy cattle. Agribiological Research 44, 91–102. K, H. (1985). Losses, plant uptake and utilization of manure N during a production cycle. Acta Agriculturae Scandinavica Supplement 24, 77 pp. K, H. & W, E. (1989). Ammonia volatilization during aerobic and anaerobic manure decomposition. Plant and Soil 115, 35–41. K, R. A., D, Z., F, J. D. & B, R. C. (1997). A sensitivity analysis of nitrogen losses from dairy farms. Journal of Environmental Management 50, 417–428. K$ , T. F., S, H., F, R. & D, W. (1999). Einflu$ sse unterschiedlicher Rohproteingehalte in der Ration auf Grundfutteraufnahme, Verdaulichkeit, NAusscheidungen und Leistung von Milchku$ hen u$ ber den Zeitraum einer Laktation. ZuW chtungskunde 71, 182–195. K$ , T. F., K, M., S, M., L, W. & S, F. (2000). Effects of rumen-protected methionine in a low-protein ration on metabolic traits and performance of early lactating cows as opposed to rations with elevated crude protein content. Journal of Animal Physiology and Animal Nutrition 84, 148–164. L, J. (1994). Untersuchungen zum Austausch von NO und NO zwischen AtmosphaW re und BiosphaW re. Doctoral # Thesis. Bayreuth, Germany : Universita$ t Bayreuth. M, T. & M, J. (1997). Ammonia emissions from composting livestock manure. In Ammonia and Odour Emissions from Animal Production Facilities, International Conference Proceedings (Eds J. A. M. Voermans & G. J. Monteny), pp. 145–153. Vinkeloord, The Netherlands : Dutch Society of Agricultural Engineering. M, O. & D, T. (1992). Loss of nitrogenous compounds during composting of animal wastes. Bioresource Technology 42, 103–111. M, H. & B, J. M. (1991). Gu$ llenanfall von Milchku$ hen auf Praxisbetrieben. Landwirtschaft Schweiz 4, 173–177. M, H., F, R. & K, R. (1997). AmmoniakEmissionen in der Schweiz : Ausmass und technische Beurteilung des Reduktionspotentials. Schriftenreihe der FAL 26. Zurich\Berne : Federal Research Station for Agroecology and Agriculture. M, M. & S, H. (1983). Die Bedeutung des Strohgehaltes fu$ r die Erhaltung des Stickstoffs wa$ hrend der Kompostierung von Rindermist. Zeitschrift fuW r PflanzenernaW hrung und Bodenkunde 146, 199–206. M, G., G, A. & F, J. (1998). MethanEmissionen der schweizerischen Landwirtschaft. Berne : Bundesamt fu$ r Umwelt, Wald und Landschaft. M, A. R. (1989). Chamber and isotope techniques. In Exchange of Trace Gases between Terrestrial Ecosystems and the Atmosphere (Eds M. O. Andreae & D. S. Schimel), pp. 175–187. New York : John Wiley & Sons. N, K. & B, R. (1997). Methodenbuch Vol. III. Die chemische Untersuchung von Futtermitteln. Darmstadt, Germany : VdLUFA-Verlag. P, J. W., D, N. E., K, T. & F, L. J. (1998). Protein content in dairy cattle diets affects ammonia losses and fertilizer nitrogen value. Journal of Environmental Quality 27, 528–534. P, V. R., C, D. A., S, R. W., C, T. R., W, A. G., D, T. G. M. & S, D. L. (1999). An assessment of ways to abate ammonia emissions from UK livestock buildings and waste stores.
249
Part 1 : ranking exercise. Bioresource Technology 70, 143–155. P, I., C, J. & V L, H. (1998). Determination of the ammonia emission from cubicle houses for dairy cows based on a mass balance. Journal of Agricultural Engineering Research 71, 239–248. R, R. E. & S, L. D. (1975). Relationship between ruminal ammonia and nonprotein nitrogen utilization by ruminants. II. Application of published evidence to the development of a theoretical model for predicting nonprotein nitrogen utilization. Journal of Dairy Science 58, 1889–1898. R, H. & V! ! , R. (1996). Amino acid nutrition of dairy cows : productive effects and animal requirements. In Recent Developments in Ruminant Nutrition (Eds P. C. Garnsworthy & D. J. A. Cole), pp. 71–93. Nottingham : Nottingham University Press. S, T. (1997). Liquid compost reactor with ammonia emissions control. In Ammonia and Odour Emissions from Animal Production Facilities, International Conference Proceedings (Eds J. A. M. Voermans & G. J. Monteny), pp. 467–473. Vinkeloord, The Netherlands : Dutch Society of Agricultural Engineering. S, M., N, A. & F, J. (2000). Lachgasemissionen aus der Schweizer Landwirtschaft. Zurich\Berne : Federal Research Station for Agroecology and Agriculture, and Ta$ nikon : Swiss Federal Research Station for Agricultural Economics and Engineering, Publication Series No. 33. S, F. (1990). Ammoniakverluste bei der Kompostierung tierischer Exkremente. In Ammoniak in der Umwelt. KreislaW ufe, Wirkungen, Minderung, Conference Proceedings (Eds J. Hartung, M. Paduch, S. Schirz, H. Do$ hler & H. Van den Weghe), pp. 37n1–37n14. Darmstadt, Germany : KTBL. S L, P., R$ , G. & B, P. (1997). Amount and limitation of ammonia emissions from stored solid manure. In Ammonia and Odour Emissions from Animal Production Facilities, International Conference Proceedings (Eds J. A. M. Voermans & G. J. Monteny), pp. 43–48. Vinkeloord, The Netherlands : Dutch Society of Agricultural Engineering. S, J. & H, J. (1999). Survey of ammonia concentrations in livestock buildings. Journal of Agricultural Science, Cambridge 133, 433–437. S, M. C. J., V, H., E, A. & K, A. (1995). Effect of protein nutrition on ammonia emissions from a cubicle house for dairy cattle. Livestock Production Science 44, 147–156. S, M. C. J., V, H., M, G. J. & V, A. M. (1997). Gaseous nitrogen emissions from grasslands. In Effect of Protein Nutrition on Ammonia Emission from Cow Houses (Eds S. C. Jarvis & B. F. Pain), pp. 101–107. Wallingford, UK : CAB International. S, S. G. & H, S. (1995). A simple model of pH in slurry. Journal of Agricultural Science, Cambridge 124, 447–453. S, S. G. & M, H. B. (2000). Emission of greenhouse gases during composting of deep litter from pig production–effect of straw content. Journal of Agricultural Science, Cambridge 134, 327–335. S, F., M, H., P, S. & Z, A. (1998). Ammonia Emissions in Switzerland. Present Situation, Development, Technical and Economic Assess-
250
. . $
ment of Abatement Measures, Recommendations. Zurich\ Berne : Federal Research Station for Agroecology and Agriculture, and Ta$ nikon : Swiss Federal Research Station for Agricultural Economics and Engineering. V V, A. M., K, C. J., V, H. & V, H. (1993). Effects of partial replacement of ryegrass by low protein feeds on rumen fermentation and nitrogen loss by dairy cows. Journal of Dairy Science 76, 2982–2993.
ET AL.
W, U., M, H., R, J.-P., F, R., J, B., K, W., M, A., S, A. F. & V, P. A. (1994). Grundlagen fu$ r die Du$ ngung im Acker- und Futterbau. Agrarforschung 1, Supplement, 40 pp. W, D. C. & R, N. (1993). Nitrogen in the excreta of dairy cattle : changes during short-term storage. Journal of Agricultural Science, Cambridge 121, 73–81.
