1991

J. agric. Engng Res. (1991) 49, 197-208

The Fertilizer Value of Agricultural Manure: Simple Rapid Methods of Assessment SERGIO PICCININI; GIUSEPPE BORTONE Centro Ricerche Produzioni Animali (Research Centre for Animal Production), Via Crispi, 3-42100 Reggio Emilia, Italy (Received

2 April

1990; accepted

in revised

form

25 November

1990)

This paper presents the results of a series of analytical tests performed on pig and dairy cattle manure in order to establish the extent of the correlation between: dry matter (TS) and specific gravity (SG); TS and total Kjeldhal nitrogen (TKN) and total phosphorus (Pt); SG and TKN and Pt. In addition, two N-meters for field use were also used to estimate the ammonium (NH,-H) content. All the variables (TS, SG, TKN, Pt, NH,-N) show a high index of correlation for both the pig and dairy cattle slurry and the linear relations applied proved adequate in all cases. Though the precision of the equations is not very high, the estimate for TKN and Pt content, obtained from the relationship between the SG and these elements is nevertheless acceptable for practical farm use of animal manure. 1. Introduction

Agricultural waste is extremely variable in its composition and it is therefore difficult to estimate its fertilizer value accurately before spreading. Chemical analyses using standard laboratory methods are accurate but take time are costly and not practical for most farms. Fast, simple field tests that can be used by farmers are needed to give adequate estimates of the fertilzer elements (in particular nitrogen) and this would be a considerable advantage in deciding the rate of application of manure to meet crop needs. There would be both economic advantages (savings on fertilizers) as well as environmental advantages (a reduction in the risk of leaching and run off of the nutrients). Researchld has shown that there is a statistically significant correlation between the TS and fertilizer elements in agricultural manure, particular pig and cattle. Research ld has also shown a significant correlation between TS and SG of manure. Simple density measurements would help to establish the content of dry matter and from this the content of fertilizer elements. A correlation between SG and fertilizer elements has also been found. This paper presents the results of a series of analytical tests performed on pig and dairy cattle manure in order to establish, under farming conditions in Italy, the correlation between TS and SG (for this, both a normal hydrometer and one calibrated in dry matter were used); TS and TKN and Pt; and SG and TKN and Pt. In addition two N-meters suitable for field use were also tested in order to measure the NH4-N content of manure. The aim was to check the precision, accuracy and reliability of such instruments and compare with standard laboratory analyses. 2. Materials

and methods

Analyses were carried out on pig and dairy cattle manure. For pig slurry, 222 samples were tested with a dry matter content in the range O-4-12%. The samples were taken from 21 farms including all the different kinds of pig units. 197

198

S. PICCININI;

G.

BORTONE

Different types of slurry were analysed; raw, after screening, after storage, after subjecting to aerobic treatment and after anaerobic digestion in mesophilic completely mixed reactors. For dairy cattle slurry (slurry from cattle where bedding is used to a limited extent only or not at all), 108 samples were tested with a dry matter content in the range 2-18%. Both raw slurry and slurry after anaerobic digestion in plug-flow type mesophilic reactors were used. The samples were taken from houses with cubicles and with mechanical manure removal. The liquid manure samples were collected from terminal collection pits and/or directly from the scraper lane. The digested slurry samples were taken from the end pit of two plug-flow type anaerobic digestors installed on two of the farms under survey. The following chemical parameters were determined on the pig and cattle slurry samples: total solids (TS) total Kjeldhal nitrogen (TKN), ammonium nitrogen (NH,-N), total phosphorus (Pt), and specific gravity (SG). The samples were homogenized using special equipment before undergoing analysis. The analytical laboratory methods were carried out according to methods recommended by APHA’ and FA07. Specific gravity was determined using a normal hydrometer and a hydrometer calibrated in % dry matter which is known as a Slurry Meter.ls5 During analysis particular attention was paid to the previous homogenization and mixing of the sample, and the choice of a cylinder only slightly higher than the instrument itself for containing the slurry. Readings were taken within a few seconds of immersion in the slurry in order to avoid errors due to the sedimentation process which is especially fast in pig slurry. Ammonium nitrogen was also determined using two simple field instruments:*” The Agros nitrogen meter* was developed in 1983 by Agros, a Swedish company. It consists of a stainless steel container with an airtight lid in which the slurry and the reagent (calcium hypochlorite) are placed using special measuring cups; the ammonium reacts with the chlorine to give gaseous molecular nitrogen and a pressure gauge, already calibrated in kg of N/m3 of slurry, registers the pressure that has developed. The reaction that takes place is as follows. 3Ca02C12 + 4NH:

+ 40H-+

CaCl, + 2N2 + lOH*O

The quantity of nitrogen may be deduced from the relationship PV = nRT, where P = pressure, V = volume, n = number of moles of Nz, R = gas constant, T= absolute temperature. Since R, V and T are constant the pressure is proportional to the molecular nitrogen formed. The Quantofix n-volumeterg was developed by the Rimu-Luftungstecmnik Company in the Federal Republic of Germany. The basic principle is the same as for the previous instrument. The differences are the construction materials (plastic instead of stainless steel), the method of measuring the pressure (a column of water), the reagent used, sodium hypochlorite (NaOCl, 160 g active Cl/l), mixed using a ratio of 5 : 1 with a 30% NaOH solution. The reaction equation is as follows: 3NaOCl+

2NH:

+ 20H-+

3NaCI + 5Hz0 + N,

Ammonium nitrogen of a fixed volume of slurry is oxidized by the reagent to gaseous molecular nitrogen (N2) which displaces an equal volume of water from the base container in the metric cylinder, whose scale is directly calibrated in kg NH4-N/m3 of slurry.

ASSESSING

AGRICULTURAL

199

MANURE

3. Results

3.1. Correlation

and discussion

between TS and SG

The results from the examinations made of 102 samples of pig slurry and 93 samples of cattle slurry are shown plotted in Figs 1 and 2. Regression equations have been fitted and the 95% confidence limits are shown; the significance of these will be further explained. The two regression equations proved highly significant (p < O-01) and show correlation coefficients (r) of 0.94 and 0.88 respectively. Both regressions, however, show a precision no better than 22%. Precision in this case refers to the repeatability of the measurement calculated on the basis of the ratio between the standard error of the estimated variable (total solids) and its average value. This means that when the equation of the straight line regression is used to estimate the total solids content of the slurry, it is based on only one determination of specific gravity, and the data calculated is liable to a minimum error of 22% (see in Figs 1 and 2 the two external lines representing the 95% confidence limits). In order to reduce such an error, it is necessary to carry out further determinations of specific gravity on the same slurry sample, so that the average value obtained is liable to less error, as shown in the Figs I and 2 by the two internal lines representing the 95% confidence limits.

1.00

1.01

Fig. 1. Pig slurry: regression

1.02 Specific

1.03

1.04

between total soli&

Regression

1.05

1.06

gravity

(TS%)

and specec gravity

equation:

TS% = -221.6 + 221.0 SG (p < 0.01) r = O-94

St. err. of the estimation = O-75 SG TS% Average. 1.018 348 St. Dev. 9.2 x 1O-3 2.22 Min. 1.006 0.36 Max.

Number of points

l-053

102

12-07

102

(SG)

200

S.

PICCININI:

G.

BORTONE

16 . 0

2 VI- 12-

.

-

.x i

I

-

,’

/

I,”

38-

0

1.00

’ ’

I.’ ’ 1.02

Fig. 2. Cattle slurry: regression

.

’ ’

’ ’ 1.04 Specific

’

’ ’ ’ 1.06

’

’

’ ’ 1.08

’

1.10

gravity

between total solids (TS%)

Regression

’

and specific grauiv

(SC)

equation :

TS% = -215.6 + 216.1 SG (p < 0.01) r = 0.88

St. err. of the estimation = 2.11 SG TS% Average. 1.040 9.46 St. Dev. 0.019 4.32 Min 1.009 1.98 Max l-092 18.07 Number of points 93 93

The hydrometer (Slurry Meter), already calibrated in % total solids, was also used to experiment on 49 samples of pig slurry and 63 samples of cattle slurry. The results obtained confirm the good correlation that exists between SG and %TS. In fact the %TS data registered by the Slurry Meter show good correlation with the %TS data determined using standard laboratory methods, as can be deduced from the straight-line regression: (a) Pig slurry TS%(lab.)

= -0.749 + 0.991 TS% (Slurry Meter) r = 0.93

p < 0.01

TS% (lab.) = -0.306 + 1.107 TS% (Slurry Meter) r = 0.93

p < 0.01

(b) Cattle slurry

The two regression equations show excellent agreement between the two methods of measuring the TS (standard method and Slurry Meter) showing the high accuracy level of the calibration of the Slurry Meter.

ASSESSING

AGRICULTURAL

201

MANURE

It needs to be pointed out, nevertheless, that the analysis of the two regressions has reconfirmed the low precision level and repeatability of the measurements (>29% for pig slurries and >21% for cattle slurries) when only one determination is made on a sample. 3.2. Correlation

between TS and TKN and Pt

The results for TS versus TKN with regression lines fitted and 95% confidence limits from the analysis of 218 samples of pig slurry and 76 samples of cattle slurry are shown in Figs 3 and 4 respectively. The regression is highly significant for pig slurry (p < O-01) with a correlation coefficient r = 0.90. Though the regression is equally significant for cattle slurry (p < 0.01) the correlation coefficient is decidedly lower (r = 0456). Both regressions show a low level of precision no better than 27% for pig slurry and 32% for cattle slurry. Results for TS versus Pt with regression lines fitted and 95% confidence limits determined from the analysis of 168 pig slurry samples and 67 cattle slurry samples are shown in Figs 5 and 6 respectively. The two regression equations emerge as being highly significant (p
‘*C

Total

solids,glkg

Fig. 3. Pig slurry: regression between total Kjeldhal nitrogen (TKN) and total solids (TS) Regression equation : TKN, g/kg = 1.095 + 0.060 TS, g/kg (p < 0.01) r = O-90 St. err. of the estimation = 0.93

g/kg 3844 32.35 3.57 12264

TKN, g/kg 3.408 2.141 0.609 10.85

218

218

TS, Average. St. dev. Min Max Number of points

S. PICCININI:

202 12

,

2 10 0-l

-

z [JI e

-

a-

,

,

,

I,

I,

I

I,,

I

I,,

BORTONE

I,

.

,..”

-

l

.

. ..

c c

I

G.

_

.’

..’ 0

..

IIll.IIIIL~lIIII’f” 0

80

40

Total

Fig. 4. Cattle slurry: regression Regression

solids,

160

200

nitrogen

(TKN)

120 g/kg

between total Kjeldhal

and total solids (TS)

equation :

TKN, g/kg = 1.554+ 0.0216TS, g/kg (p < 0.01) r = O-66

St. err. of the estimation= 1.21 TKN TS, g/kg g/kg Average. 101.79 3.750 49.01 1.610 St. dev. Min 19.08 1.46 180.67 10.18 Max Number of points 76 76

Both show a low precision level, no better than 40% for pig slurry and 23% for cattle shIy. In Fig. 6 a smaller scatter around the straight-line regression for samples with a total solids content of less than 8% (anaerobically digested cattle slurry) can be seen. One possible explanation for this is the lower content of cellulose materials (straw) in these last samples as compared to the raw cattle slurry (the samples with a total solids content higher than 8%). For management reasons the loading of anaerobic digestors is, in fact, done using slurry which does not contain coarse fragments. The correlation between TS and the nutrients was also examined within a group of slurry samples from pig and dairy cattle, homogeneous both in terms of age and the type of treatment. Comparison of the various straight-line regressions did not indicate significant differences (parallel tests). The direct ratio between SG and TKN and Pt was also looked into on the basis of previous work2C and for the interest created from a practical point of view. A significant correlation was found both for the pig and cattle slurry (p ~0.01) even though the corresponding correlation coefficients were shown to be less than those found for the regression equations between TS and TKN and Pt. The precision of the estimates

ASSESSING

AGRICULTURAL

203

MANURE 6

,,,,,,,,,,,,,, .

Total

Fig. 5. Pig slurry:

regression

solids.

1

g/kg

between total phosphorus

(Pt) and total solids (‘IS)

Regression equation : Pt, g/kg = 0.032 + 0.0312TS, g/kg @
is also no better. (a) Pig slurry

The equations of the straight-line

TKN (g/kg) = -67.40 + 69.19 SG n = 54

are as follows:

p < 0.01

r = 0.75 standard error estimate = 25%

Pt (g/kg) = -46*61+ n = 48

regressions calculated

46.88 SG

(1)

p co-01

r = 0.75 standard error estimate = 46%

(2)

(b) Cattle slurry TKN (g/kg) = -44.61+ n = 62

46.22 SG

r = 0.56 standard error estimate = 38%

Pt (g/kg) = -14.05 + 14-40 SG n=54

p < 0.01

(3)

p < 0.01

r = 0.74 standard error estimate = 27%

(4)

204

S. PICCININI; 2.4

,

0

I,..

, ,

.,I

0

40

Fig. 6. Cattle slurry: regression

,

,

,

I

I I 80 Total

I

,

,

,

I

I I

,

,

,

I I,, 120 solids,g/kg

between total phosphorus

, ,

I I 160

G.

BORTONE

,

I ,200

(Pt) and total solids (TS)

Regression equation : Pt, g/kg = 0.3698 + 0.007 TS, g/kg @ < 0.01) r = 0.82 St. err. of the estimation = O-24 TS, Pt, g/kg dkg Average. 99.38 1.050 St. dev. 50.84 0.420 Min 19.80 0.270 Max 180.67 2.040 Number of points 67 67

3.3. Field application

of results

The significant linear regressions found could be used to estimate the Pt and TKN in pig and cattle slurry. Specific gravity of the slurry is measured using a hydrometer, the percentage of dry matter is calculated using the respective equation. Then the nitrogen and phosphorus concentrations are determined using the regression equations between dry matter and nutrients. This method, however, is inadvisable in as much as it is subject to a high level of error on account of the combination of poor precision levels or the repeatability determined in all the linear regressions examined. It is far more advisable to use the linear regression found between specific gravity and fertilizer elements. This avoids a series of calculations and thereby reduces the margin of error when estimating TKN and Pt even if the relative correlation coefficients and the relative precisions both for pig and cattle slurry are not better than the direct correlations between TS and SG and between TS and TKN and Pt. Although the margin of error remains rather high [of Eqns (11, (21, (3) and (4)l I‘t would be more acceptable within the limits of the use of agricultural manure.

ASSESSING

AGRICULTURAL

205

MANURE

Laboratory

Fig. 7. Pig slurry:

comparison

Regression

ammonium

nitrogen,

mg/l

of Agros meter and standard

laboratory

method

equations:

NH,-N(Ag) = 83.94 + 0.922NH,-N r = 0.96 St. err. of the estimation = 242.46

Average. St. dev. Min Max Number of points

NH,-N (Ag), mgll 1959 854 200 4800 64

(Lab)

@ < 0.01)

NH,-N (Lab), w/l 2033 888 196 4200 64

Hydrometers with a scale already calibrated with the value of fertilizer elements would give an instant reading without the necessity to consult conversion tables. 3.4. Agros-nitrogen

3.4.1. Agros-nitrogen

merer

and Quantujix N-uolumeter

meter

Pig slurry samples (63) and cattle slurry samples (13) were analysed using this instrument. Figs 7 and 8 show the results and straight-line regressions with 95% confidence limits for NH4-N determinations with the Agros meter versus NH.,-N determined using standard methods for pig slurry and cattle slurry respectively. The regression equations are highly significant (p < O-01) and show a correlation coefficient of 0.96 and 0.94 respectively. For pig slurry, statistical tests were carried out (Student’s t test) on the value of the intercept which did not prove to be significantly different from 0, and on the slope which did not differ significantly from one. This indicates the high accuracy of the instrument excluding the existence of systematic errors. The precision of the estimations made with the Agros meter proved high (212%).

206

S. PICCININI;

0

1000

2000

Laboratory

Fig. 8. Cattle slurry: comparison Regression

3000 ammonium

4000

5000

G.

BORTONE

6000

nitrogen,mgll

of Agros meter and standardlaboratory method

equation :

NH,-N(Ag) = 292.58+ 1.11NH,-N (Lab) (p < 0.01) r = 0.94

St. err. of the estimation= 462.24 NH,-N (Ag), NH,-N (Lab), mg/l mg/l Average. 2469 1963 1074 St. dev. 1269 Min 1300 961 Max 5700 5005 Number of points 13 13

For cattle liquid manure, statistical tests were carried out (Student’s t test) on the value of the intercept which did not differ significantly from 0, and on the slope value which did prove to be significantly different from one (p ~0.01). When used with the cattle liquid manure, the instrument is therefore considered to be less accurate and affected by a systematic error of about 20%. As such, the application of a correction factor represented by the straight-line regression equation is considered appropriate. The level of precision (320%) proved to be less than for the pig slurry. The lower accuracy and precision found for the cattle liquid manure are probably attributable to the high straw content which hinders reliable sampling and analysis. 3.4.2. Comparison between the Agros nitrogen meter and Quantojk N-uolumeter

A comparison was also made between the Agros meter and the Quantofix which has been available on the market only more recently. A total of 16 pig slurry samples were used from which the N&-N content was determined using both the two field instruments and standard laboratory techniques. Table 1 shows the results of the N&-N calculations carried out on the 16 pig slurry samples using the three methods.

ASSESSING

AGRICULTURAL

207

MANURE

NH,-N

values

Table 1 determinated by three slurry samples

methods

N&N

Observations made Average Standard deviation Min. Max.

on 16 pig

(w/l)

Standard method

Agros meter

16 1414 719 773 3360

16 1347 715 600 3200

Quantotix N-volumeter 16 1372 876 400 4ooa

The equations of the two straight-line regressions of the NH,-N calculated using the standard laboratory method versus the NH,-N calculated using the two field instruments were as follows: NH4-N NH4-N

(standard, mg/l) = 78.47 + 0.992 NH4-N (Agros, mg/l) n= 16 r = 0.99 p
(standard, mg/l) n= 16

= 313.33 + O-802 NH,-N r = 0.98

(Quantofix,

mg/l)

p < o-01

The Quantofix N-volumeter proved less accurate when compared to the Agros, showing a systematic error of about 5%. Comparison of precision levels (Fisher test) did not show any significant differences; both instruments gave a result of ll-12% precision. The two instruments tested also proved to be simple to use as well as reliable from a maintenance and hardwearing point of view. The trend of the NH,-N/TKN ratio in pig slurry was also studied because both the ammonium nitrogen data and the total nitrogen data are important in the utilization of agricultural manure. A total of 206 samples were examined and the following results were found: average value NH,-N/TKN = 0.65 standard deviation = O-12 It would, therefore, be possible to arrive at the TKN value by dividing the NH4-N determined using the Agros meter by the coefficient 0.65. This value, however, is subject to a high level of error attributable to the combination of the errors of the numerator and the denominator. The most acceptable approximation would be to obtain the TKN data directly from the TKN straight-line regression equation determined in the laboratory VU-SW the NHrN determined by the Agros meter. The equation of this straight-line regression is as follows: TKN (kg/m3) = 0.347 + l-439NH4-N (Agros, kg/m3) p < o-01 r = O-81, standard error estimate = 29% n = 63 4. Conclusion

AI1 the variables studied (TS, SG, TKN, NH,-N, Pt) show a high linear correlation-for both the pig and dairy cattle slurry. Poor precision was found, however, in straight-line regressions between TS and SG, between TS and fertilizer elements and between SG and fertilizer elements. The estimate for nitrogen and phosphorus content, obtained from

208

S. PICCININI;

G.

BORTONE

the existing ratio between these elements and specific gravity was nevertheless acceptable within the limits of agronomic use of animal manure. It should be remembered that the SG data will prove representative of the total volume of slurry contained in the storage tank only if the slurry sample, or samples, are taken after careful mixing and homogenization of the tank contents. The two field instruments tested, the Agros nitrogen meter and the Quantofix N-volumeter were considered useful for determining the ammonium-N content on account of their good agreement with standard laboratory analyses. A previous mixing of the slurry in the storage tanks is nevertheless also required for their use. References

’ Tunney, H. Dry matter, specific gravity, and nutrient relationshipsof cattle and pig slurry. In: Engineering Problems with Effluents from Livestock. (Hawkins, J. C., ed.) Proceedings. C.E.C. Seminar, Cambridge,England. Luxembourg, PublicationsC.E.C. 1979 ’ Cbeschei, G. M.; Westeman, P. W. Rapid methodsfor determining fertiliser value of livestock manures. American Society Agricultural Engineers, Paper N”84-4082, presented at ASAE summermeeting, University of. Tennessee,Knoxville, June 1984 ’ Tunney, H. Manure nutrient composition: rapid methodsof assessment.In: Processingand Use of Organic Sludge and Liquid Agricultural Wastes. (I’Hermite, P., ed.) Proceedingsof 4th. International SymposiumRome, Italy, 8-11 Oct. 1985 ’ Tunney, H. Field testsfor estimatingfertiliser value of agricultural wastes.In: Agricultural Waste Management and Environmental Protection. (Walte E.; Szabolcs, I., eds). Proceedings4th International SymposiumCIEC, Braunschweig,F.R.G., 11-14 May 1987 5 Tunney, H.; Bertrand, M. Rapid field testsfor estimatingdry matter and fertiliser value of animal slurries. In: Agricultural Engineering. (Dodd, V. A.; Grace, P.M., eds) Proceedingsof the 11th International Congresson Agricultural Engineering, Dublin, 4-8 September 1989 ’ American Public Health Association (APHA). Standard methodsfor examination of water and wastewater1985 ’ Food and Agriculture Organisatioo (FAO). Recommendedanalytical methods for first priority componentsof liquid manure. Intermediate report of Subnetwork 4, Budapest, 1980 ’ KjeUerup, V. Agros Nitrogen Meter for estimation of ammonium nitrogen in slurry and liquid manure. In: Efficient land useof sludgeand manure. (Dam Kofoed, A., ed.) Barking: Elsevier Applied SciencePublishers,1985 9 Khss, J.; Weber, W. Method for rapid determination of ammonianitrogen in animal slurriesand sewagesludge.In: Agricultural Waste Management and Environmental Protection. (Welte, E.; Szabolcs, I., eds) Proceedings4th International Symposium CIEC, Braunschweig, F.R.G., 11-14 May 1987.