Landbrugsministeriet
Statens Planteavlsforsøg
Ammonia volatilization from cattle and pig slurry during storage and after application in the field
Ph.D. Dissertation
Sven Gjedde Sommer
The Royal Veterinary and Agricultural University Section of Soil, W ater and Plant Nutrition
Copenhagen
The Danish Institute of Plant and Soil Science Department of Plant Nutrition and Physiology Research Centre for Agriculture
Tidsskrift for Planteavls Specialserie
Beretning nr. S 2209 -1992
Contents
CONTENTS ... II LIST O F F IG U R E S ... IV LIST OF TABLES ... V 2. FOR OR D ...VI 3. S A M M E N F A T N IN G ... IX 4. S U M M A R Y ...XI 5. N O M E N C L A T U R E ... XIII
6. IN T R O D U C T IO N ...1
6.1 O bjective o f the s t u d y ... 1
6.2. Processes o f volatilization used in the development o f a model ...2
7. T EC H N IQ U ES FOR M EASURING AM M ONIA V O LA TILIZA TIO N L O S S E S ... 5
7.1 . Indirect measurements ... 5
7.2 . D irect measurements ... 7
7 .2 .1 . Enclosures ... 7
7 .2 .2 . Wind tu n n e l s ... 8
7 .2 .3 . M icrom eteorological m e th o d s ... 9
8. FACTORS IN FLU EN C IN G AM M ON IA L O S S E S ...13
8.1. C oncentration o f T A N ...13
8 .1 .1 . Concentration gradient in stored s l u r r y ... 13
8 .1 .2 . The am m onia flux related to the content o f T A N in surface applied s lu r r y ... 13
8.2. Acidity in the slurry ...14
8 .2 .1 . Changes in pH in the stored slurry ... 14
8 .2 .2 . Ammonia loss related to pH o f applied s l u r r y ...15
8.3. Structure material in the s l u r r y ...17
8 .3 .1 . E ffect o f surface crusting on am m onia loss from
stored s l u r r y ...
8 .3 .2 . E ffect o f dry matter content on loss o f am m onia
from surface applied s l u r r y ...
8.4. Climate
8 .4 .1 . T he effect o f wind speed and tem perature on
am m onia loss from stored s l u r r y ...
8 .4 .2 . T he effect o f tem perature, air humidity and wind speed on am m onia loss from surface applied
slurry ...
8.5. M ethods for reducing am m onia l o s s e s ...
8 .5 .1 . Slurry s to r a g e s ...
8 .5 .2 . M ethods for reducing losses following
application ...
8.6. Loss o f am m onia from slurry applied to a recently
lim ed s o i l ...
8.7. A m m onia volatilization during application o f slurry ...
9. M ODELLING A M M O N IA V O LA TILIZA TIO N FRO M STO RED S L U R R Y ...
10. CON CLUSIO N . . 11. OUTLOOK 12. LITERATU RE . .
]
1 2
2
2:
2
‘2<
2f
27 27 29 30 32 33
Appendix I - XII; Papers published or submitted for publication.
List of figures
En liste o ver fig u rer m ed angivelse a f side
Fig. 1 Cross section o f a wind tunnel unit ... 8 Fig. 2 A m m onia loss from surface applied slurry determ ined
with wind tunnels...9 Fig. 3 Relationship between N H 3 losses measured by a new
and a conventional m icrom eteorological m ethod... 12 F ig .4 D escription o f proton and hydroxylic ion producing processes during
volatilization o f am m onia and carbondioxide from a stored s l u r r y ... 16 Fig. 5 Mean daily am m onia loss rates from stored cattle
and pig slurry ... 17 F ig .6 Am m onia loss from surface applied cattle slurry
related to slurry dry m atter content ... 19 Fig. 7 Predicted and measured am monia loss from stored
pig slurry ... 20 F ig .8 A m m onia loss from surface applied cattle slurry
during a sum m er and a w inter period... 21 Fig. 9 Am m onia loss from surface applied slurry related
to air tem perature... 23 Fig. 10 Loss o f am m onia from pig slurry injected, surface applied
and cultivated o r surface applied ... 25
List of tables
En liste over tabeller m ed angivelse a f side
T able 1. Initial bulk pH after stirring o f stored slurry and surface pH
o f slurry during the experim ent... 15 T able 2. V olatilization o f am m onia from surface applied pig slurry to a
new ly cultivated and an uncultivated soil... 26
2. Forord
D enne afhandling er udarbejdet med henblik på erhvervelse a f P h .D .-g rad en ved D en K ongelige V eterinær- og Landbohøjskole. Herm ed afsluttes 3 års studier financieret af F orskerakadem iet og Statens Planteavlsforsøg. D et eksperim entelle arbejde er udført p å Askov Forsøgsstation med kontakter til sektion for Kulturteknik og P lanteem æ ring, Landbohøjskolen.
Studiets hovedfag har væ ret Planteem æ ring og fysiologi under vejledning af professor N iels Erik Nielsen med bifag i analytisk fysisk kemi under vejledning a f lektor Søren Storgaard Jørgensen og lektor L eif Skibsted. Vejledere på projektet har væ ret afdelingsforstander Bent Tolstrup Christensen, professor N iels E rik Nielsen og lektor Jan K ofoed Schjørring.
Jeg vil gerne rette en tak til Bent T. Christensen, fordi jeg fik m uligheden for at gennem føre dette studium på Askov Forsøgsstation, og for en grundig, positiv og ikke m indst konstruktiv kritik af de artikler, je g har forfattet i løbet a f de 3 år. Niels E rik Nielsen takkes for ideer og inspiration til projektet. En tak til Jan K. Schjørring for et godt sam arbejde gennem årene, sam t for interessante og behagelige sam taler og diskussioner. Endelig vil je g takke medarbejderne p å Askov Forsøgsstation, fordi de tog godt imod mig og gav en stor hånd til gennem førelse a f eks
perim enterne. Især skal Kirsten Vang takkes for sin store og dygtige indsats i forbindelse med opbygning a f forsøgsudstyr og gennem førelse a f forsøgene.
Til slut en stor tak til min kone Birgitte og mine børn Niels og T ine, fordi I tog med til Jylland, og ovenikøbet accepterede, at je g det følgende år drog til K øbenhavn næsten ugentligt.
I løbet a f studiet er der publiceret følgende artikler, hvori resultaterne fra studiet beskrives i detaljer. I denne afhandling gives en sammenfatning a f artiklerne, hvortil der henvises ved num rene I-XII.
A rtikler publiceret internationalt:
I. Schjørring, J.K , Sommer, S. G. and Ferm, M. 1992. A sim ple passive sam pler for m easuring am m onia emission in the field. W ater A ir Soil P ollut. 62, 13-24.
II. Somm er, S.G . and Jensen, E.S. 1991. F oliar absorption o f atmospheric am m onia by
ryegrass in the field. J. Environ. Qual. 20, 153-156.
III. Som m er, S . G ., Kjellerup, V .K and Kristjansen, O. 1992. D eterm ination o f total am m onium nitrogen in pig and cattle slurry: Sam ple preparation and analysis. Acta. Agric.
Scand. In press.
IV . Som m er, S .G . and Olesen, J .E . 1991. Effects o f dry m atter content and tem perature on am m onia loss from surface-applied cattle slurry. J. E nviron. Q ual. 20, 679-683.
V. Som m er, S .G ., Olesen, J .E . and Christensen, B .T. 1991. E ffects o f tem perature, wind speed and a ir hum idity on am m onia volatilization from surface applied cattle slurry. J.
Agric. Sei. Cam b. 117, 91-100.
V I. Som m er, S .G ., Christensen, B .T ., Nielsen, N .E. and Schjørring, J.K . 1992. A m m onia volatilization during storage o f cattle and pig slurry: Effect o f surface cover. J. A gric. Sei.
Subm itted.
V II. Som m er, S . G ., Jensen, E .S. and Schjørring, J.K . 1992. L eaf absorption o f gaseous am m onia after application o f pig slurry on sand between rows o f w inter wheat. 395-402.
In A ir P o llu tio n R e se a rc h R e p o rt 39; F ield M e a su re m e n ts a n d In te r p r e ta tio n o f Species r e la te d to P h o to o x id an ts a n d A cid D eposition. Eds. A ngeletti, G ., B eilke, S.
and S lanina, S. U dgivet a f EEC. Belgien.
Artikler publiceret i T idsskrift for Planteavl op G røn Viden. Landbruget:
V ill. Sommer, S.G . 1989. Udspredning a f gylle: Fordam pning a f am m oniak og fordeling a f udbragt gylle. (Spreading o f slurry: Volatilization o f am m onia and distribution o f applied slurry). T idskr. Planteavl. 93, 323-329. (Summ ary and legends in English).
IX. Christensen, B .T . and Sommer, S.G . 1989. Fordam pning a f am m oniak fra udbragt gød
ning. M etode og am m oniaktab fra urea og urea-am m onium nitrat. (V olatilization o f
am m onia from fertilizers and manure. M ethodology and loss o f am m onia from urea and
urea-am m onium -nitrate). Tidsskr. Planteavl. 93, 177-190. (Sum m ary and legends in
English).
X. Som m er, S.G . and Christensen, B .T. 1989. Fordam pning a f am m oniak fra svinegylle udbragt på jordoverfladen. (Volatilization o f am m onia from surface-applied pig slurry).
T idsskr. Planteavl. 93, 307-321. (Summ ary and legends in E nglish).
X I. Somm er, S.G . 1990. A m m oniakfordam pning fra svinegylle på nykalket jord. (Am m onia volatilization from pig slurry applied to recently lim ed soil). G røn viden, Landbrug. 51,
1-4.
X II. Som m er, S.G . and Christensen, B. T. 1990. A m m oniakfordam pning fra fast husdyrgødning
sam t ubehandlet, afgasset og filtreret gylle efter overfladeudbringning, nedfældning,
nedharvning og vanding. (Am m onia volatilization from solid m anure and raw, ferm ented
and separated slurry after surface application, injection, incorporation into the soil and
irrigation) T idskr. Planteavl. 94, 407-418. (Sum m ary and legends in English).
3. Sammenfatning
F ordam pning a f am m oniak reducerer husdyrgødnings indhold a f plantetilgæ ngelig kvælstof.
Ammoniakken tabes i stalden, fra lager, under udbringning, fra gødning der ligger på jorden og fra græssende dyr. Den fordam pede am m oniak vil blive afsat som am m oniak eller am m onium , og kan derved m edføre uønskede æ ndringer a f kvælstofbegrænsede plantesam fund. I dette studie er faktorer a f betydning for am m oniaktabet fra gyllebeholdere, under udbringning og fra udbragt gødning, blevet undersøgt.
Forskellige m etoder til bestem m else a f am m oniaktabet fra handelsgødning og fra husdyrgødning er benyttet eller afprøvet. Størrelsen af am m oniaktabet kan som regel ikke bestem m es som forskellen i tilført og genfundet kvæ lstof eller udtrykt som m erudbytter a f tørstof eller indhøstet kvælstof. Am m oniaktab m ålt i et kam m er kan ikke direkte relateres til tab i m arken. K am m ertek
nikken kan derfor mest hensigtsmæssig anvendes under kontrollerede forhold i laboratoriet, i forsøg hvor am m oniaktabet ved forskellige behandlinger ønskes sammenlignet. M ålinger a f am m oniaktabet fra gylle med vindtunneler har i m arkforsøg givet sam m e væ rdier som m ålinger med m eteorologi
ske m etoder, når vindhastighed og nedbør i vindtunnelerne kontinuert ju steres til sam m e niveauer som udenfor. V indtunnelerne er m eget anvendelige til undersøgelser i m arken, hvor effekten af forskellige faktorer ønskes bestem t, og til at tilvejebringe data for m odellering a f am m oniaktabspro
cessen. Til direkte m ålinger a f am m oniaktabet fra udbragt gødning i m arkskala-forsøg, anses meteorologiske m assebalance-m etoder for at væ re mest velegnede. En ny m etode baseret p å passive flux m ålere har i en afprøvning vist sig at give korrekte tabsm ålinger. M etoden viste sig enkel at anvende i et forsøg, hvor am m oniaktabet fra seks forsøgsfelter i en m ark blev bestem t samtidig.
F orsøgene er im idlertid areal krævende, og m ålingerne kan ikke gennem føres, hvis der i vindretningen er en am m oniakkilde inden for en afstand a f 50-100 m.
V indtunneler blev benyttet til måling a f am m oniakfordam pningen fra gylle lagret i pilot-skala gyllebeholdere (0,90 x 2,89 m), og fra gylle nedfældet direkte, nedharvet i jo rd en eller udbragt på jorden (0,5 x 2 ,0 m). Am m oniaktabet ved udspredning a f gylle blev undersøgt i et enkelt forsøg og blev målt som differencen i am m onium indholdet før og efter udbringning.
F ra gyllelagre med ugentlig om røring var am m oniaktabet 3-5 g N H 3-N n r 2 dag '1. Om som m eren
(16,9 °C) var am m oniaktabet 50% større end i en vinter-forårs periode (7,3°C ), som følge a f de
højere tem peraturer. Tilstedeværelse a f et flydelag begrænsede am m oniaktabet med 80% i forhold til tabet fra gylle om rørt ugentligt. Et forsøg fra en vinterperiode viste, at et lag halm på 15 cm kunne erstatte flydelaget. Ekstra overdækning i form a f leca sten, rapsolie, sphagnum , flydende plastfolie og et trælåg begrænsede am m oniaktabet væsentligt. D er er blevet udviklet en model til beregning a f am m oniaktabet fra lagret gylle, og de beregnede tab er i overenstem m else med tabene målt med vindtunneler.
U nder udbringning a f gylle, dvs. fra gyllen forlader gyllesprederen til det ram m er jo rd en , er am m oniaktabet m indre end 4% a f det udbragte am m onium i et forsøg med klapspredere.
A m m oniakfordam pningen fra gylle udbragt på jordoverfladen påvirkes a f klim aet, gyllens sam m ensætning, beskaffenhed a fjo rd e n s overflade og tiden fra gyllen udspredes til den nedbringes i jo rd en . I de første 6 tim er efter udbringning steg am m oniaktabet eksponentielt med tem peraturen, hvorefter tabet steg lineæ rt med tem peraturen. A m m oniaktabet øgedes med stigende vindhastighed indtil 2-3 m s 1. Ved vindhastigheder derover ændredes tabsraten ikke. N edbør eller vanding reducerede am m oniaktabet fra overfladeudbragt husdyrgødning.
A m m oniaktabet øgedes med stigende tørstofindhold i gyllen samt ved stigende pH og alkalinitet.
Det akkum ulerede am m oniaktab efter 6 tim er steg lineært med tørstofindholdet i gyllen; men derpå var forløbet kurvet. Hvis tabsraterne blev korrigeret for effekten a f tem peratur og pH , var tabet sigm oidalt relateret til tørstofindholdet. Dette resultat antyder, at æ ndringer i tørstofindholdet ved niveauer under 4% og over 12% er a f mindre betydning for størrelsen a f am moniaktabet.
E fter nedfældning eller om hyggelig nedharvning var am m oniaktabet ringe. H vis jorden blev harvet (til 10 cm dybde) før overfladeudbringning a f gylle, blev am m oniaktabet begræ nset med ca.
50% i forhold til gylle udbragt på ubehandlet jo rd . U dlægning a f gylle mellem ræ kker a f afgrøder
ved slangeslæbning kan begrænse am m oniaktabet fra gyllen, idet forflygtigelsen a f am m oniak fra
gyllen mindskes, og am m oniak optages a f planterne.
4. Summary
Volatilization losses o f am m onia reduces the fertilizer value o f anim al m anure for plant production.
Ammonia losses related to livestock farm ing occur from anim al houses, from m anure storages, during spreading o f m anure, from surface applied m anure and from grazing anim als. Deposition o f aerial am m onia and am m onium may cause undesired changes o f oligotrofic ecosystem s. The objective o f this study was to increase the know ledge o f the factors, w hich influences the am m onia loss potential from slurry during storage, spreading and from slurry applied in the field.
Techniques for m easuring am m onia losses from slurry w ere tested and are discussed. It is concluded, that changes in the nitrogen content o f slurry and crop responses are not adequate for determination o f am m onia loss from slurry stored o r applied to the field. T he use o f small enclosures should be used only for com paring relative am m onia loss rates between treatm ents, as losses cannot be related directly to losses in the field. T he technique is best for m easurem ents in the laboratory w here the environm ent can be controlled. A wind tunnel system gave results sim ilar to measurements in the open, if wind speed and precipitation in the tunnels w ere adjusted to the environment. T h e tunnels can be used under field conditions in studies o f the effect o f various treatments in replicate experim ents. T he systems have been shown to provide reliable data for modelling the process o f am m onia volatilization. T he dem and for big hom ogeneous areas with a uniform source strength restricts the applicability o f the eddy correlation techniques, the gradient technique and the Z IN S T mass balance technique. A test o f a new passive flux sam pler used in a mass balance m ethod gave accurate determ inations o f the am m onia loss from a circular plot.
Wind tunnel system s w ere used for determ ining am m onia volatilization losses from stored slurry and from slurry applied in the field. A m m onia loss during spreading was determ ined by the change in nitrogen content o f the slurry.
From slurry storages which w ere stirred weekly, average am m onia losses w ere 3-5 g N H 3-N n r2
d '1. The losses during sum m er (16.9°C) w ere 50% higher than during w inter-spring (7.3°C) due to
higher tem peratures. S urface crustings reduced am m onia losses by 80% com pared to the losses
from slurry stirred w eekly. O ne experim ent showed that a layer o f chopped straw (15 cm) could
replace a surface crusting layer. Am m onia losses from slurry covered by rape oil, leca pebbles,
sphagnum peat, floating foil o r a wooden lid w ere less than 60% o f the loss from stirred slurry.
A model for predicting am m onia loss from stored slurry w as developed. T he model predictions agreed with am m onia loss values from uncovered slurry determ ined w ith wind tunnels.
D uring application o f slurry with conventional spreaders am m onia losses were less than 4% o f the am m onium content.
A m m onia loss from surface applied slurry is influenced by clim ate, slurry composition, soil conditions and tim e from application until incorporation o f the slurry. Ammonia losses w ere exponentially related to air tem perature during the first 6 hours. Subsequent loss rates were linearly related to tem perature. Am m onia losses increased with wind speed up to 2-3 m s'1. At higher w ind speeds, loss rates w ere not influenced by change in wind speed. R ain o r irrigation reduced am m onia losses from surface applied slurry.
A m m onia losses increased with increasing slurry dry m atter content and pH . The accum ulated am m onia loss after six hours was linear related to slurry dry m atter content, but in follow ing periods the relation was curved. A fter adjusting the results for tem perature and pH effects the relationship was sigm oidal. Changes in slurry dry m atter content, therefore, had little effect on am m onia loss, when the dry matter content was higher than 12% o r low er than 4%.
L ittle am m onia was lost from slurry injected directly into the soil, and from slurry incorporated
into the soil. Cultivation before application o f slurry onto the soil reduced ammonia losses
com pared to an unharrowed soil by 50% . Ammonia loss from slurry applied to a crop m ay be
reduced by application on the soil between rows o f the crop. T he reduction is caused by a reduced
transfer o f am m onia from the slurry to the atm osphere and absorption o f ammonia by the plant
leaves.
5. Nomenclature
Sym bol Definition Units
K„
K
nNH„
n h
3
sTA N
N H 3I
n h4+
J
D K(v)
R.
R h
H enrys constant
The equilibrium constant o f NH4+ and N H 3
Vertical flux o f am m onia
Concentration o f gaseous am m onia in equilibrium with the concentration o f NH3, in the slurry solution
Am m onia concentration in the im m ediate atm osphere
Concentration o f total am moniacal nitrogen [ N H / ] + [NH3]
Concentration o f am m onia in solution Concentration o f am m onium in solution Diffusive transport o f T A N in the slurry
Diffusion coefficient o f TA N Transfer coefficent for am monia volatilization
Resistance o f the turbulent layer above the slurry
Resistance o f the lam inar layer between the surface o f the slurry and the turbulent layer
The height o f the internal boundary layer
atm kg m ol'1 mol I'1
atm osphere
atm osphere
mol l'1
mol I'1 mol I'1 mol cm 2 s'1
cm 2 s '1
mol N H 3-N m s '1
s n r 1
s n r 1
cm
Symbol Definition Units
z„ Roughness length
z0 Roughness length
k von Karmans constant
v. Roughness speed
v(z) M ean wind speed at different heights above the slurry
T A ir tem perature
z D istance from the surface o f the slurry, zero is the surface o f the slurry,
negative is dow nward and positive is upwards.
X The distance downwind from the edge o f the slurry treated area or the slurry tank
cm
cm k = 0 .4 m s'1 m s'1
°K cm
cm
6. Introduction
In Denmark about 60% o f the nitrogen excreted from housed cattle and pigs is collected and transported in form o f slurry, equivalent to a yearly production o f 180.000 ton N (Kofoed and Hansen, 1990). M o re than half the nitrogen is am m onium -N which is readily available to plants.
Ammonia volatilization, therefore, reduces the value o f slurry for plant production.
In Denmark and in E urope, em ission o f am m onia from anim al m anures represents the most important source o f atm ospheric am m onia (Buijsman et a l. 1987). T he deposition o f am m onia may be detrimental to nitrogen lim ited ecosystems (Roelofs, 1986). In D enm ark the use o f anim al manure, therefore, has been regulated by law.
Ammonia volatilization losses related to livestock farm ing occur from anim al houses, m anure storages, applied m anure and from grazing anim als. As stated by Jarvis and Pain (1990), there are few dependable m easurem ents o f am m onia loss from stored manure. From applied slurry it has been shown that am m onia volatilization is related to the general w eather conditions, slurry composition, soil characteristics and method o f application (Brunke et a l. 1988; Pain et a l. 1989;
Döhler 1991; H orlacher and M arschner 1990; Thom pson e t a l. 1990). Experim ents under different environmental regim es, and differences in the origin and chem ical com position o f the slurries employed have given a qualitative understanding o f the factors influencing am m onia loss. H ow ever, the limited num ber o f experim ents have not allowed a general quantitative relationships to be established betw een am m onia loss rates and the various variables influencing am m onia losses.
6.1 O bjective o f th e stu d y
- To increase the understanding o f the factors influencing am m onia loss from stored and surface applied animal slurry.
- To evaluate different techniques for determ ining am m onia losses from confined areas.
- To obtain quantitative relationship between the rate o f am m onia loss (F) and the intensity o f
various factors affecting am m onium concentrations, eg. pH and dry m atter content o f the slurry.
6 .2 . Processes o f volatilization used in the developm ent o f a m odel
T he main source o f am m onium in slurry is urine which contains m ore than 60% of the excreted nitrogen from pigs and cattle (Koefoed and H ansen, 1990). In urine abo u t 70% of the nitrogen is in urea (M uck and Richards, 1983; Jarvis e t al. 1989). A m m onium and carbonate are produced during the hydrolyses o f urea by the enzym e urease:
C 0 (N H 2)2 + 2H 20 = 2 N H 4+ + C 0 3-
U rease is released from living and disintegrating microbial cells and is found in a wide ran g e o f environm ents including soil and the floors o f animal houses. D ue to high temperatures and the moist substrate, this reaction will be rapid in slurry produced in anim al houses. Only little o f the excreted urea therefore is left in stored slurry, and the conversion o f urea is not included in the model for describing the am m onia loss from slurry storages.
Am m onia volatilization from a liquid surface such as recently stirred slurry can be considered to be the transfer o f am m onia from the liquid surface o f the slurry to the im m ediate atmosphere. T he rate o f am m onia loss is given by:
F = K (v )(N H 3 g - N H 3 J (1)
K(v) depends mainly on wind speed, surface roughness and tem perature (Rachpal-Singh and N ye, 1986a).
T he concentration o f atm ospheric am m onia in equilibrium with a solution will be proportional to the concentration o f N H 3I in solution. The relationship between the different ammoniacal species in solution is as follows:
(NH 3)g
It kh
(N H 4+) + H 20 —
( N Hj)i+ H 30
T he concentration o f N H 3 at equilibrium can be calculated by the equations:
[TA N ] = [N H 3J + [N H 4+] (2)
[ N H / ] = [NH3J[H 30 +]/K
n(3)
[N H 3I1 = [TA N ]/(1 + [H30 +]/K n) (4)
T he atmospheric am m onia concentrations in equilibrium with the concentrations o f TA N in the surface o f the slurry can be calculated by the H en ry ’s Law equation:
NH3,g= K „ [N H 3]] (5)
Combining 1, 4 and 5 gives the equation:
F = K (v )(K „([T A N ]/(l + [H30 +]/K n)) - N H 3,J (6)
Volatilization o f am m onia from the surface o f the slurry reduces the concentrations o f TA N in the surface layers. F o r the volatilization process to continue, TA N m ust be transported to the surface layer. Depletion o f T A N will cause a decrease in concentrations tow ard the top layer, and TAN will diffuse to the surface. This transport can be described by F ic k ’s law:
J = -D d [T A N ]/d z (7)
It is assumed the slurry is not stirred, and that turbulent transfer caused by wind o r tem perature gradients in the slurry may be neglected due to the viscosity and dry m atter content o f most slurries.
T he transfer coefficient K(v) used in equation (1) may be calculateded by the equations (Molen et al. 1990):
2
K (v )= 1 /(R .+ R b) ( 8 )
R, = ln (/ z0 ‘) k '1 v .'1 (9)
v . = v(z) k (ln(z z,,1) )'1 ( 10)
/(ln(//z0 )-l)=k2 X ( 11 )
Rb is the resistance o f the lam inar layer dependent on slurry surface roughness and v.. The relation between Rb, surface roughness and v. is not know n, and has to be estimated from am m onia loss m easurem ents.
The tem perature dependent equilibrium constant (KN) and H enry L aw constant (K,J can b e cal
culated by the equations (Beutier and Renon, 1978):
lnKN= -177.9 5 2 9 2 -1 8 4 3 .22/T + 31.43351n(T )-0.0544943T (12)
lnK H = 160.559-8621.06/T -25.67671n(T )+ 0.035388T (13)
7. Techniques for measuring ammonia volatilization losses
Ammonia losses may be determ ined indirectly from the crop response to nitrogen in applied manure, or as the difference in am monium content added and the recovery o f inorganic nitrogen in the soil. The d irect m ethods o f determ ining am m onia losses from slurry are based on changes in atmospheric am m onia concentration in enclosures placed over the treated surface, o r on m icrometeorological techniques in which the volatilization is determ ined in the free air above the surface.
7 .1 . Indirect m e a su re m e n ts
Ammonia volatilization from slurry added to the soil and determ ined by indirect techniques will be influenced by leaching, denitrification, mineral fixation o f am m onia and m ineralization/im m obili
zation processes. Estim ating am m onia volatilization from crop responses (dry m atter yield o r plant uptake o f nitrogen) to the field applied m anure do not discrim inate between these processes.
Am m onia loss estim ated from cropping m ethods will depend on the crop response to the added nitrogen (the crop/nitrogen response curve). I f the yield is at the maximum o f the crop response curve, there will be no significant increase in crop production due to the applied nitrogen, and the am m onia volatilization is overestim ated. Christensen (1986) found that the am m onia loss determined with indirect methods tends to be twice as high as losses found by direct methods.
U sing the nitrogen uptake, additional errors arise due to variations in uptake o f nitrogen induced by climate or infection o f the crop, and to the exchange o f nitrogen between the atm osphere and the crop (Farquhar e t a l. 1980). Introducing l5N will not reduce the errors as the assim ilated 1SN will take part in the exchange processes (Schjørring et a l. 1989). T herefore recovery o f nitrogen in the harvested crop only broadly reflect differences in the gaseous loss measured with wind tunnels (Larsen e t a l. 1992; Thompson et a l. 1987).
Nömmik (1966) studied am m onia volatilization from nitrogen fertilizers by the N -recovery technique. To reduce leaching, the experim ents w ere carried out in periods with no precipitation, and in dry soils to avoid denitrification. F urtherm ore, the soil had a low am m onium fixing capacity
2 *
and the experim ental periods w ere kept short to reduce effects o f fertilizer nitrogen on net- m ineralization.
The precaution needed for this technique, however, lim its its use for estimating am m onia volatilization from surface applied slurry, as the soil will be m oistened and inorganic nitrogen, therefore, exposed to denitrification (Thom pson and Pain, 1988; M olen e t al. 1988). M olen e t a l.
(1988) established a mass balance for nitrogen applied to the soil surface in urine, using m icrom eterorological m easurem ents and the nitrogen recovery technique. Between 10-30 % o f the nitrogen added was not accounted for. The authors suggested that this w as due to denitrification, as all nitrogen added was recovered when nitrification inhibitors w ere added.
T he nitrogen recovery approach was used in a study o f am m onia absorption by leaves after application o f slurry on the soil between the rows o f w inter w heat (V II). 15N enriched am m onium was added to the slurry. Im m obilization, mineralization and denitrification processes w ere inhibited by mixing m ercury chloride into the slurry. T o im prove the calculation o f the nitrogen balance, the slurry w as applied to nitrogen-free sand. F rom this system, am m onia loss was higher than the losses determ ined with direct methods in the field. The higher losses w ere ascribed to th e low CEC o f the sand and the absence o f nitrification which reduces am m onia losses.
Recovery o f TA N in stored slurry cannot be used for determ ination o f the ammonia loss, because am m onia is form ed by net-m ineralization o f nitrogen during the experim ents (Bode, 1991). T h e loss o f am m onia is small com pared to the total-N content in the slurry. D ifferences in total-N, therefore, may give erroneous estim ates o f am m onia loss due to sampling variability. Inhomogenity o f total-N is great in stored slurry, as the organic nitrogen precipitates o r floats to the surface.
A m m onia loss during application (i.e. from the slurry leaves the spreader and until it hits the
ground) was m easured by determ ining the difference in content o f am m onium (V ill). T h e slurry
in the slurry tanker was stirred to elim inate variations in TA N content o f the samples related to
inhom ogenity. In this experim ent the experim ental period was short, and microbial processes,
therefore, o f m inor im portance.
7 .2 . D irect m easurem ents
7 .2.1. Enclosures
E nclosures used in the field have been made o f cylinders with a lid on the top, and an inlet and outlet at the top for air to pass over the experim ental plot (Ferguson e t a l. 1988; Kissel et al.
1977). Ammonia loss was measured by collecting the gaseous am m onia in acid. T he air may be drawn through the enclosure continuously, the enclosures being perm anently closed (Sherlock and Goh, 1984). In the system o f Kissel e t al. (1977) the enclosures w ere periodically closed w ith the lid, and air was only drawn through the enclosure during closure.
Am m onia losses measured with enclosures increase with the air flow up to 15 volum e changes/m inutes (Kissel e t a l., 1977; Ferm 1986). A t higher flow rates, am m onia loss w as not significantly influenced by change in air flow rate (Kissel et al. 1977). H o ff et al. (1981) proposed that change in aerodynam ics caused 41 % o f lost am m onia from slurry applied on a plastic foil not being accounted for by the determ ination with enclosures only closed during m easurem ents. T he direct m easurem ents w ere com pared to the loss determ ined by the N -recovery technique, but the difference could not be ascribed to denitrification o r im m obilization. W hen determ ined with an enclosure, am monia loss from urea spread upon soil covered by a 3 cm layer o f chopped w heat straw, was much lower than determ ined by a m icrom eterological method (Ferguson et al. 1988).
The straw seemed to have directed the air flow over the surface o f the w heat straw , by that increasing the diffusion resistance. If windspeed is high and the roughness o f the surface low , the accum ulated loss from a treated plot determined with an enclosure will be sim ilar to direct m easure
ments in the open. This was shown by determ ining am m onia loss from urea broadcasted to a pasture (grass height 1 cm) with a perm anent closed enclosure (air change 17 volum e m in u te 1), com pared to sim ultaneous m easurem ents with an aerodynam ic technique. D uring the experim ents, the wind speed was high and there was no rain (Black e t al. 1985).
The most appropriate use o f the enclosure technique is determ inations o f the am m onia loss potential from animal m anure or fertilizers applied to an uncovered soil in the laboratory (D öhler, 1991; W hitehead and Raistrick, 1990). T he air exchange should be 17 volum es minute"1 o r m ore.
The enclosures change the air tem perature with less than 2°C at this flow rate (Sherlock and Goh,
1984).
A
Figure 1. Cross section o f a w ind tunnel unit. D im ensions are given in cm (XI). A, m otor; B, air
sam pling p oints; C, fa n ; D , Steel-net; E, steel duct; F, vane anem om eter head; G, temperature
sensor; H, steel fu n n e l; I, tunnel o f polycarbonate; J, m etal fra m e ; K, spade-edge.
7.2.2. W ind tu n n e ls
A system o f portable wind tunnels (Fig. 1) has been designed to allow m easurem ents o f am m onia volatilization in the field without inducing marked changes in the m icroclim ate (Lockyer, 1984;
IX ). The loss o f am m onia from the plot is calculated by the product o f air flow and the difference in am m onia concentration o f air entering and leaving the tunnels.
W ind speed within the tunnels is controllable, and a test showed that air tem peratures inside the tunnel generally are less than 1 °C low er than am bient air tem peratures (2 m height) and soil tem peratures outside tend to be 1.9°C higher than inside (V; Ryden and Lockyer, 1985). Com pared to the m icrom eteorological mass balance method, sim ilar am m onia losses w ere determ ined with wind tunnels, provided the wind speeds in the tunnels w ere adjusted to am bient wind speeds 0.25 m above the ground (Ryden and Lockyer, 1985). W ind tunnels may, therefore, be used for estim ating the effect o f wind speed on am m onia volatilization from surface applied cattle slurry (V;
Thom pson e t a l. 1990). D ue to turbulence, the wind tunnels could not be used for estim ating the reduction in am m onia volatilization when slurry was applied on the soil in a tall crop (X).
The coefficient o f variation in am m onia loss between wind tunnels with sim ilar treatm ent w as less
than 25% . T here were no significant differences in am m onia losses determ ined in tw o trials with
sim ilar treatm ents and clim ate (Fig. 2). This gives an opportunity for com paring experim ents
epeated during different periods, exam ining the effect o f clim ate o r com paring the effect o f treatments, slurry com position etc. by m ore measurem ents during the sam e period.
In the studies IV -V I and IX -X II, wind tunnel systems w ere used to study the effect o f clim ate and slurry com position factors on am m onia loss from surface applied anim al m anure and stored slurry. Ammonia loss was measured from the soil plots o f 50 x 200 cm and from pilot slurry tanks o f 90
X289 cm.
7.2.3. M icrom eteorological m ethods
The microm eteorological methods may be divided into three categories, eddy correlation methods, gradient diffusion m ethods and atm ospheric mass balance methods (Denm ead, 1983).
<X 2
Days from s t a rt of e x p e r i m e n t
Fig. 2. Cumulative am m onia loss as percentage o f TAN (Ammonium + am m onia) in surface applied
slurry determined with three wind tunnels. Both experim ents were carried out in D ecem ber 1986
during periods with near sim ilar clim atic conditions (V).
With the Eddy correlation technique am m onia loss is determ ined by sim ultaneously m easuring the
vertical wind speed and am m onia concentration during short tim e intervals. The gradient diffusion
technique requires m easurem ent o f mean gas concentrations at different heights above the surface
and knowledge o f the appropriate eddy diffusion coefficient. T he eddy diffusion coefficient is
calculated either from wind speed and tem perature (Aerodynam ic m ethod) o r temperature and air hum idity (energy balance method), at d ifferent heights above the surface. T hese techniques can only be used in situations w here the air has transversed a field with hom ogeneous ammonia so u rce strength and uniform surroundings, as horizontal concentration gradients must be negligible. T h e experim ental area has to be large (several ha). Therefore the m ethods are not widely used for determ ination o f am m onia loss from areas treated with am m onium fertilizer or manure.
M ass balance techniques have been used for measuring am m onia volatilization from sm aller land areas treated with sewage sludge, pig o r cattle slurry o r grazed areas (Beauchamp et a l. 1978;
D enm ead e t a l. 1982; Ryden and M cNeill, 1984; Pain et a l. 1989; Bless et al. 1991). T he am m onia volatilization is calculated from the difference in the horizontal flux o f gaseous am m onia through hypothetical vertical planes w indward and leeward o f a treated experimental area.
T he profile o f the am m onia flux is often calculated from determ ination o f ammonia by gas traps and o f wind speed by cup anem om eters at five o r m ore heights. T he m ethod is most simply applied when circular plots are used with the leeward sampling position at its centre. By that the distance between the leeward and w indward sampling position (ie the fetch) is constant despite changing wind direction (Denm ead, 1983). T he method is usefull for agricultural studies because it only needs small experim ental areas and due to the sim plicity o f the m ethod.
F o r circular plots located in large and uniform areas, the am m onia flux profiles hav e a theoretically predictable shape determ ined by surface roughness, plot geom etry and atm osphere stability (Wilson e t a l. 1982; M clnnes e t a l. 1985). W ilson e t a l. (1982) predicted a h eight (ZIN ST), at which the ratio o f the horizontal to the vertical flux (u X c/F ) has almost the sam e value in all atm ospheric stability regim es. The ratio u X c/F may be calculated knowing su rface roughness, plot geom etry and atm ospheric stability, and the vertical flux o f ammonia can be calculated from determ inations o f the horizontal flux u Xc. D eterm ination o f ammonia volatilization from circular plots 20 o r 50 m in radius by the ZIN ST method, agreed satisfactorily with estim ates based on a mass balance which em ployed am m onia fluxes determ ined in five or more heights (W ilson et a l. 1983; Pain e t a l. 1989). The labour needed for the m easurem ents of the horizontal flux o f am m onia can be reduced by using a rotating passive flux sam pler (Leuning et a l. 1985;
Sherlock e t a l. 1989). T he ZIN ST method has been used m easuring am m onia losses from a circ u lar
experimental area w ith a radius o f 3.5 m (fetch), and determ ination o f am m onia fluxes 12.5 cm above the surface (G ordon et a l., 1988).
Knowing the m ean air tem perature and wind speed at a reference height z and surface temperature, u x c / F can be calculated as the function o f the fetch (M clnnes e t a l. 1985). T he fetch and the reference height may, therefore, be varied, and the vertical flux o f am m onia can be determined at a reference height w here the coefficient o f variation o f the m easurem ents are small.
A test showed (M cinnes e t a l., 1985) this method to agree better with the mass balance technique w here the flux was m easured at 5 heights, than the method o f W ilson e t a l. (1982), because the concentration o f am m onia may be low at the height ZIN ST. T he techniques do not account for the horizontal flux in the plant canopy, and the methods cannot be used in studies o f am m onia loss from slurry applied to a tall crop.
A new mass balance technique was developed (I). This method does not m ake extensive dem ands to the surrounding area, clim atic variables have not to be measured and no electricity is needed.
T he ammonia flux is determ ined with passive flux sam plers, each m ade o f two glass tubes connected in series. T he inner surface o f the tubes is coated with oxalic acid which absorbs the am monia in the a ir flow ing through the tubes. The passive flux sam plers are mounted at four heights on four m asts. The masts are placed at right angles to each other on the periphery o f a circular experimental plot. At each mast, the flux o f am m onia into and out o f the experim ental plot is measured. The vertical flux o f am m onia from the plot is calculated as the difference in the flux o f ammonia into and out o f the plot, measured at the four masts. The vertical flux per square m eter is calculated, by dividing the horizontal flux by the diam eter (fetch, m) o f the experim ental plot.
T he vertical flux from the circular plot determ ined with passive absorbers was in good agreem ent with the determination based on gas traps and wind speed m easurem ents (Fig. 3).
T he passive flux sam plers w ere used in a study o f am m onia loss from slurry applied to plots o f
15 X 15 m in a w inter w heat field (Som m er and Pedersen, unpublished). T he slurry was applied
onto the crop with splash plates, injected into the soil o r applied with trail hoses on the soil between
the rows of the crop. Six experim ents w ere carried out during the same period. The method
perform ed very w ell on these relatively small plots placed w ith a distance o f 40 m between plots,
Fig. 3. Relationship between vertical flu x densities o f NH3 fag N H 3- N m'2 s'1) m easured by a
m icrometeorological m ethod with a new type o f passive flu x sa m p ler and a conventional
m icrom eteorological m ethod (reference m ethod) (I).
and about 30 m to the nearest building or hedge. It is therefore suggested, that this method provides a research tool able to com pare the effects o f differences in slurry composition and different application techniques. The experim ents may be carried out in th e sam e field and in th e sam e period, excluding the effect o f differences in soil conditions and clim ate.
In the experim ent three o f the flux sam plers w ere mounted above the canopy and one flux sam pler 5 cm below the top o f the canopy (Som m er and Pedersen, unpublished). Within the crop canopy, about 10% o f the total horizontal flux from the plot was m easured, assum ing that the am m onia flux measured 5 cm below the top o f the wheat represents the flux gradient within the cro p (see D enm ead e t a l. 1982). Consequently the horizontal flux profile within the canopy must be d eterm i
ned, if am m onia loss from slurry applied to a tall crop ( > 2 0 cm) is exam ined by a mass balance
technique.
8. Factors influencing ammonia losses
Ammonia loss during storage and from surface applied slurry will partly be influenced by sim ilar factors, i.e., clim ate, dry matter content and pH o f the slurry. D ue to the different environm ents, the factors may act different. The process o f am m onia volatilization from stored slurry is not significantly influenced by the material o f the storage tank, whereas the loss from surface applied slurry is influenced by the soil. T he am m onia loss may, therefore, be related to the sam e factors, but the influence o f the factors must be dealt with separately.
8 .1 . C o n c en tra tio n o f TA N
8 .1 .1 . Concentration gradient in stored slurry
W ind tunnel m easurem ents o f the am m onia loss from a weekly stirred slurry showed that losses increased two to three fold after stirring (VI). M ixing o f the slurry increased the TA N concentrations in the surface layers, causing the transfer o f am m onia to the atm osphere to increase.
One to two days after stirring the loss rates declined, as concentrations o f TA N in the surface layers declined. T h e reduction o f TA N in the surface is supposed to be caused by the slow diffusive transport of TA N from the layers below, which cannot replace am m onia lost by volatilization, and a curvated decline in T A N concentrations towards the surface o f the slurry is created. This gradient has been measured in the pilot slurry storages (Olesen and Som m er, not published), and is depicted by a model describing the diffusion o f TA N during am m onia volatilization (M uck and Steenhuis,
1982).
8 .1 .2 . The am m onia f lu x related to the content o f TAN in surface applied slurry
In the experiments IV , V , X and X I, the am ount o f slurry applied was generally 3 1 m'2. The
accumulated loss o f am m onia during the experim ent was expressed as percentage o f T A N added
in the slurry, and the loss during a period was expressed as percentage o f TA N rem aining in the
slurry at the beginning o f the period. This transform ation enables com parison o f accum ulated
am monia losses from slurries with different contents o f TA N , because o f the linear relation between
am m onia loss and concentration o f TA N (Equation (6); Brunke et a l. 1988). The transform ation w ill give an opportunity to relate the determ ined am m onia losses to factors such as clim ate, differences in com positions o f slurry and soil conditions (IV; V; B eaucham p et al. (1978)).
In most experim ents am m onia loss rates w ere high the first 24 hours after surface application, thereafter the losses decreased (Fig. 2). This pattern is due to th e reduction in the content o f applied TA N caused by am m onia volatilization, but also to changes in slurry acidity and infiltration o f slurry into the soil (Lauer et al. 1976; Beauchamp et al. 1982). In m ost experiments m o re than half the total am m onia loss during a 6-day period occured within the first day (IV,V, IX -X II), a pattern dem onstrated in several other studies (Thompson et al. 1987; Pain et al. 1989).
8.2 . Acidity in th e slurry
8 .2 .1 . Changes in p H in the stored slurry
Initially it was assum ed that pH in the surface o f newly stored slurry w ould be modified, m ainly due to transport and loss o f the acidic am monium ion and the basic bicarbonate (Fig. 4). S im ilar processes have been described by a model o f am m onia volatilization from urea applied on the soil (Rachpal-Singh and N ye, 1986ab). T he model predicted, and an experim ent showed, that in this system pH declined 1-1.5 units at the soil surface. T he decrease in pH in the soil surface w as due to a faster diffusive transport to the surface o f TA N com pared to carbonate species, and thereby a g reater volatilization o f am m onia. Loss o f am m onia decreases pH , w hile the loss of carbondioxide increases pH.
In stored slurry the system differs from the one described by R achpal-Singh and Nye (1986ab).
In recent produced slurry pH will rise as urea is hydrolysed to am m onia and carbonate ions. I f the buffer system only consisted o f TA N and carbondioxide pH would be 9 .3 after hyrolyzes o f the urea. In slurry organic acids modify pH to between 7 and 8, and T A N and carbondioxide are mainly am monium and bicarbonate. In a soil with recently hydrolysed urea, pH is about 9 , and T A N and carbondioxide are mainly am m onia and carbonate.
These processes have to be exam ined. In this study, therefore, pH in the surface o f stored pig
slurry was determ ined in a wind tunnel experim ent, carried out to pro v id e data for the developm ent
o f a model. Im m ediately after stirring, bulk pH in one m depth was 6 .9 -7 .0 and four days la ter pH
Table 1. Initial bulk p H after stirring o f the slurry and surface p H o f slurry during the experim ent
(Sommer and Olesen, not published). Immediately after the slurry was stirred p H was m easured in
samples from 0 .7 5 m depth. A fter that p H was determ ined at the slurry surface with a fla t glass
membrane electrode.
Wind speed m s'1
D ays from stirring o f the slurry
0 3 8 10
pH
8.7 6.9 7.7 7.3 7.3
5.3 7 .0 8.0 7.8 7.9
2.2 6.9 7.7 7.4 7 .7
in the surface w as 0 .4 to 1.0 units higher (Table 1). The high surface pH was probably due to a greater loss of carbondioxide than o f am m onia after stirring. Surface pH did not decrease from day 4 to day 10, indicating that am m onia volatilization did not cause an increase in the acidity o f the surface layers. T his may be due to a greater loss o f carbondioxide com pared to am m onia, o r to aerobic digestion o f fatty acids in the slurry surface, a process which also will increase pH (Georgacakis ct a l. 1982; Husted e t al. 1991)).
8 .2 .2 . Ammonia loss related to p H o f applied slurry
Accumulated am m onia loss during 6 day periods increased with the initial bulk pH in the surface applied pig slurries (X). Interaction o f dry matter content o f the slurry and air tem perature weakened the relationship, and the am m onia flux could not be related to initial differences in ammonia/am monium ratios induced by proton activity. This may be due to buffering o f pH by volatile organic acids, the bicarbonate system and am monium (Husted e t a l. 1991). Husted e t al.
(1991) showed that a slurry with a high bulk pH may have a low total alkalinity, and losses o f
small amounts o f am m onia vill reduce pH , while a slurry with a low er bulk pH may have a high
S l ur r y storage
N H 3 --- 14- ----
NH3
■f C02 J
- W -
co 2
i
1 N H 4 - N H 3 + H+
CO^H2O - H 2C 0 3^ H C 0 5^ K 1-- CO3“ H+
NH4 = NH3
hC C V H 2 O ~ H 7 CO 3 ^ HC05 CO 3 ^ l-T
Fl g. 4. D escription o f proton and hydroxylic ion producing processes during volatilization o f
ammonia and carbondioxide fr o m a slurry tank.
total alkalinity, and loss o f am m onia does not change the acidity significantly. This could explain why the accum ulated am m onia loss during 6 days in som e experim ents was higher from slurries with a low bulk pH than a high bulk pH (X). Anaerobic digestion o f slurry increased p H , but am m onia loss values w ere sim ilar from surface applied anaerobic digested and raw slurry (XII;
Pain e t a l. 1990). This may be due to an unchanged alkalinity o f the slurry.
Im m ediately follow ing application o f a pig slurry, pH in the surface increased to 8.4 from an initial bulk pH o f 7 .6 in the stored slurry (V). The increase was due to a greater loss o f carbon
dioxide than o f am m onia, because w ater solubility o f carbondioxide is lower than that o f
am m onium . W ithin 3-4 days, pH declined to 7.4 as am m onia volatilization produced protons in the
surface. The initial high pH could explain the high am m onia loss rate during the first day after
application (Fig. 2). Subsequently the acidification o f the slurry has influenced the reduction in
am m onia loss rate.
Lid PVC-Foil Peat Leca Rapooil Straw Surface Weekly crust stirring I December 1989-June 1990 [|] September-December 1990
I I July-Septcmber 1990 !$$ February-June 1991
Fig. 5. Mean d aily am m onia loss rales fro m stored cattle (Dec. 1989-Sep. 1990) and p ig slurry
(Sep. 1990-Jun. 1991) with eight different surface coverings. * No fo rm a tio n o f surface crust.
Ammonia losses were m easured with wind tunnels on slurry tanks (0.90 m x 2 .8 9 m). M ean air
temperatures during the experiments were: Dec. 1989-Jun. 1990, 7°C; J u l.- Sep. 1 9 9 0 ,17°C; Sep.-
Dec. 1990, 7°C; Feb.-Jun. 1991, 6 °C .
8 .3 . Structure m aterial in the slurry
8.3.1. E ffect o f su rfa c e crusting on am m onia loss fr o m stored slurry
Ammonia loss from a slurry with a surface crust was less than 20% o f the losses from a weekly
stirred slurry (V I). T he surface crust created a stagnant air layer above the liquid slurry and
increased the surface roughness. Thereby the surface resistance increased and the transfer
coefficient (K(v) decreased, (Equation 8). Ammonia loss rate from a cattle slurry with no surface
crust was sim ilar to the loss rates from a stirred slurry (Fig. 5). The great loss o f am m onia from
unstirred slurry w ith little dry m atter content may be due to a low viscosity and an increase in
natural convection o f the slurry. Furtherm ore the transfer coefficient is higher above a rough surface with surface crust than above a liquid surface. In D utch wind tunnel experiments sim ilar am m onia losses from stored cattle slurry w ere determ ined (Bode 1991). In the present study (VI) am m onia losses from stored pig slurry w ere low er than in the D utch, probably due to the developm ent o f a surface crust, w hich did not appear in the study o f B ode (1991).
8 .3 .2 . E ffect o f dry m atter content on loss o f am m onia fro m su rface applied slurry
A m m onia loss rates from surface applied cattle slurry adjusted to differen t dry matter contents w ere related to the content o f dry m atter (Fig. 6). The am m onia loss from 0 to 6 h was linearly related to dry m atter content, w hile the relationship during follow ing periods w as nonlinear. A lo w er dry m atter content in the slurry may cause an increased infiltration into the soil. Sorption to soil colloids reduces the concentration o f am monium in the solution, and the transfer of am m onia from the liquid in the soil to the atm osphere is reduced. In addition the transport o f gaseous am m onia is reduced in soil.
W hen the effect o f pH and tem perature was eliminated from the am m onia loss data, the loss o f am m onia during the periods 0-6 h, 6-12 h, 12-24 h and 24-144 h tended to be sigmoidally related to dry m atter content. This shows that at low ( < 4% ) and high contents ( > 12%) o f dry m atter, small changes in dry m atter have a limited influence on am m onia loss. Probably little liquid rem ains on the soil at dry matter content low er than 4% and little liquid infiltrates the soil a t dry m atter content higher than 12%.
H igher loss values w ere determ ined from cattle slurry than from adjusted slurry with sim ilar dry m atter content (X II). Visually it appeared that the cattle slurry had a higher viscosity than the adjusted slurry, the high viscosity may have contributed to a sm aller infiltration compared to the adjusted slurry. Soluble carbohydrates, proteins and fatty acids co ntribute to the viscosity o f the slurry, but may not be related to the dry matter content. T herefore viscosity may be an additional factor in the am m onia loss process.
In w inter experim ents, high losses w ere determ ined from the fibrous fraction (20% DM ) b u t not
from the liquid fraction (1 % D M ) o f mechanical separated pig slurry. T h is indicates that dry m atter
content and thus infiltration interact with tem perature (X II).
S l u r r y d r y m a t t e r c o n t e n t , %
Fig. 6. Accum ulated am m onia loss in percent o f TAN in applied cattle slurry related to slurry dry
m atter content f o r fo u r time periods (IV). The am m onia loss was m easured with wind tunnels (0.5
X
2 m). Slurry dry m atter content was adjusted by mixing the fib ro u s and liqued fra c tio n o f
m echanical separated cattle slurry. M ean a ir temperatures were fr o m 0 .5 to 19.6°C.
3
W in d s p e e d [ m s “ l ]
Fig. 7. Predicted (solid line) and m easured (points) ammonia loss fr o m stored p ig slurry at different
wind speeds. The slurry was stirred initially. Am m onia loss was determ ined with wind tunnels on
p ilo t slurry tanks (0,9 x 2 ,8 9 m ), m ean air temperature was 3°C. The lam inar resistance was
assum ed to be Rb= 2 0 s m '1. (Olesen og Sommer, unpublished).
8.4. Climate
8 .4 .1 . The effect o f w ind speed a n d a ir tem peralure on am m onia loss fr o m stored slurry
By increasing wind speeds the boundary resistance at the surface o f stored slurry will be reduced (Equation (9)-(12)), and the transfer o f am m onia from the slurry to the atm osphere increases. At high wind speeds transfer o f am m onia through the boundary layers to the atm osphere will be faster than the rate o f diffusive transport o f TA N to the surface layers, and the slow diffusive transport o f TA N will lim it the rate o f am m onia volatilization. T he flux o f am m onia from stored pig slurry, therefore, showed a curvated relationship with increasing wind speed in the periods 0-4.5 h, 4.5- 12.5 h and 21.5-27.5 h after an initial stirring (Fig. 7).
From stored cattle slurry, am m onia loss at mean air tem peratures o f 17°C was 50% higher than at 7°C. During the day the slurry was stirred, am m onia loss was related to air tem perature (VI).
On the day o f stirring, TA N concentrations in the surface only varied a little during the experimental period, and there was little interaction from differences in TA N concentrations and development o f surface crust.
Days from s t ar t of e x p e r i m e n t
Fig. 8. Accumulated am m onia loss in percent o f added TAN in surface applied cattle slurry ( 3 1 m '2)
during a summer and a w inter period (V). Am m onia loss was m easured with wind tunnels (0,5
X2 m).
8 .4 .2 . The effect o f a ir tem perature, a ir hum idity an d wind speed on am m onia loss fro m s u t f ace applied slurry
A t low tem peratures, the rate o f am m onia loss was generally reduced (Fig. 8). The accum ulated loss during 6 days was high, how ever, due to an appreciable loss rate from day 2 to day 6. D uring the experim ent the soil was saturated with w ater and partially frozen, factors assumed to prom ote am m onia losses (V). Sustained am m onia losses from slurry during periods with low tem peratures have also been observed in a study o f T hom pson et a l. (1987). At 19°C initial loss rates w ere high but after 12 h alm ost no further loss occurred (Fig. 8). This was ascribed to surface crusting and rapid infiltration into the relatively dry soil.
T he results o f 20 wind tunnels-experim ents with near identical cattle slurries w ere related to clim atic conditions (Fig. 9). During the initial 6 h, the accum ulated am m onia loss was exponentially related to tem perature. D uring the three succeeding periods o f 6-12 h, 12-24 h and 24-144 h, the am m onia volatilization rate was low and only slightly related to tem perature. In these periods the am m onia loss pattern shifted to a linear relationship to tem perature. T h e reason for the increase in loss rate with tem perature is that equilibrium -constants change with tem perature. At a given con
centration o f TA N , an increase in tem perature will cause a higher am m onia/am m onium ratio in the slurry, and the solubility o f am m onia declines with increasing tem perature. Initially pH is high (8.5) and changes in tem perature will influence am m onia loss significant, as small changes in pH produce significant changes in the am m onia/am m onium ratio (V). T herefore the am m onium loss rate is exponential related to tem perature during the first six hours. In the following periods pH declined to 7.5, at this level the am m onia/am m onium ratio is less influenced by changes in pH . A fter a period factors like infiltration and soil properties interact with temperature.
F our equations describing the relationship between am m onia loss and temperature w ere determ i
ned for the four periods (Fig. 9). F o r all experim ents 46 % o f the m easured loss over 6 days was accounted with the models (V), indicating that other factors like soil properties influenced the am m onia losses.
T he am m onia loss rate after 6 h increased when tunnel wind speed increased up to 1.75 m s '.
N o consistent increase in volatilization was found when the wind speed increased from 1.75 to 2.8
m s '1 (V). In a study o f Thom pson e t a l. (1990), the accum ulated am m onia loss from cattle slurry
was measured with sim ilar wind tunnels. The loss o f am m onia increased with wind speed up to
A i r t e m p e r a t u r , °C
