- Email: [email protected]
Estimation of maturity of compost from food wastes and agro-residues by multiple regression analysis
Bioresource Technology 97 (2006) 1979–1985
Estimation of maturity of compost from food wastes and agro-residues by multiple regression analysis Miyuki Chikae, Ryuzoh Ikeda, Kagan Kerman, Yasutaka Morita, Eiichi Tamiya
*
School of Materials Science, Japan Advanced Institute of Science and Technology, 1-1, Asahidai, Nomi, Ishikawa 923-1292, Japan Received 30 August 2005; received in revised form 28 September 2005; accepted 30 September 2005 Available online 9 November 2005
Abstract The composting process of food wastes and tree cuttings was examined on four composting types composed from two kinds of systems and added mixture of microorganisms. The time courses of 32 parameters in each composting type were observed. The efficient composting system was found to be the static aerated reactor system in comparison with the turning pile one. Using the multiple regression analysis of all the data (159 samples) obtained from this study, some parameters were selected to predict the germination index (GI) value, which was adopted as a marker of compost maturity. For example, using the regression model generated from pH, NHþ 4 concentration, acid phosphatase activity, and esterase activity of water extracts of the compost, GI value was expressed by the multi-linear regression equation (p < 0.0001). High correlations between the measured GI value and the predicted one were made in each type of compost. As a result of these observations, the compost maturity might be predicted by only sensing of the water extract at the composting site without any requirements for a large-size equipment and skill, and this prediction system could contribute to the production of a stable compost in wide-spread use for the recycling market. 2005 Elsevier Ltd. All rights reserved. Keywords: Composting; Food waste; Tree cutting; Maturity; Multiple regression analysis
1. Introduction The present trend about waste management is recycling and the recovery of waste as new materials or as energy. The waste organic materials produced by the city life, such as domestic refuse, residues from food processing and wastes from food industry are accumulating to become a significant amount. The composting of these by-products is more encouraged to avoid the loss of energy. The compost from domestic wastes possesses a high nutritional value, with high concentration of nitrogen, potassium and phosphorus, while the contamination by heavy metals and other toxic substances is negligible (Hogland et al., 2003). However, for the development of a market of recycle and recovery of waste to compost, not only must the qual-
*
Corresponding author. Tel.: +81 761 51 1660; fax: +81 761 51 1665. E-mail address: [email protected] (E. Tamiya).
0960-8524/$ - see front matter 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.biortech.2005.09.026
ity of products be stable, but also the composting system must be efficient. The evaluation of the maturity of domestic wastes compost has been widely recognized as one of the most important factors concerning the compost process and the application of these by-products. The land-applied immature compost gives rise to a serious N-deficiency in crops, and the rapid decomposition of immature compost causes a decrease of the O2-concentration around the root system (Mathur et al., 1993a). Additionally, land-applied immature compost inhibits the plant growth by the production of phytotoxic substances, fundamentally ammonia, ethylene oxide, and organic acids (Mathur et al., 1993a). Many parameters to evaluate the maturity of compost from food wastes or city refuse, such as the change of physicochemical properties (Mathur et al., 1993a), the enzymatic activity (Ranalli et al., 2001), germination tests (Zucconi et al., 1981), calorimetric and spectroscopic methods (de Oliveira et al., 2002; Ouatmane et al., 2000; Provenzano et al., 2001)
1980
M. Chikae et al. / Bioresource Technology 97 (2006) 1979–1985
have been reported, and reviewed (Jime´nez and Garcia, 1989). To the best of our knowledge, no parameters for using at the composting site in a simple way have been reported. Food wastes of school lunch, which are similar to domestic refuse, and on which source separation can be done easily, were incinerated in Kaga City, a town on the Honshu Island of Japan with a population of 67,600. On the other hand, tree cuttings from parks and roadsides are also occupying a significant amount of land in Kaga City. The composting of these materials requires a reasonable recycling management. The aims of this study were: (1) to search an efficient composting system of food wastes and tree cuttings from four comparing types; and (2) to establish a detection way of the compost maturity, which could be used easily at the composting site. Then, the time courses of physicochemical and biological parameters of the products from four composting types were observed, and the relationships between germination index (GI) values and other parameters in each sample were determined by using stepwise multiple regression analysis.
the compost reactor, placed outdoors, was a cylinder with 1.5 m diameter, 1.5 m in height and a hole in the center for aeration. The sidewalls were made of stainless steel mesh and the top was covered with a sheet. Temperature measuring and sampling were carried out at three different locations, such as, at the surface, 10 cm depth from surface, and the center of the reactor once per week. Composting experiments were performed at the Mizushimabussan Co. Ltd. in Kaga City from October, 2003 till March, 2004. The mean atmospheric temperatures of day during experiment were between 1.6 and 11.7 C. The total weight and the main components of the raw materials in four composts are shown in Table 1. 2.2. Physicochemical analysis Water content was determined by oven-drying at 105 C for 5 h. Total carbon content and total nitrogen content were measured by auto elemental analyzer Vario EL III (Elementar, German), and C/N was calculated from them. Water extracts of the compost samples were prepared by shaking the fresh sample with distilled water at 1:10 w/v (dry weight basis) using a horizontal shaker for 30 min at room temperature by filtration. The fresh extracts were subjected to measuring pH, electrical conductivity (EC), and the remaining extracts were kept at 20 C until use. The thawed extracts were used for determination of concentrations of NHþ 4 by the Indophenol Blue method (Koshino et al., 1988a), and NO 3 by the salicylic acid–sulfuric acid method derived from the phenol–sulfuric acid one (Koshino et al., 1988b).
2. Methods 2.1. Composting and compost sampling The organic fractions of the food waste of school lunch, separated into three groups of bread/rice, vegetablesÕ waste, and leftovers of lunch, were collected from a junior high school in Kaga City. These obtained food wastes were mixed with rice hulls and similar size of chips of tree cuttings, as bulk materials, and a mixture of natural microorganisms, commercialized as Uron C (A; Tsurumi Soda, Japan) or Uchishiro (B; Seiwa, Japan), was added for accelerating the composting process. The mixture was stirred mechanically for 3 h in 65 C (Daido, Japan), and the water content was adjusted to approximately 65% by control of volume of bulk materials. After then, the treated mixture was composted by two systems, the turning pile system, and the static aerated system using the compost reactor (Mizushimabussann, Japan). On the turning pile system, the heaps in 5 m · 5 m yard were turned manually two times per week. After each turning, temperatures and samples were taken from three different locations of the compost pile once per week. On the static aerated system,
2.3. Enzyme activity assay 2.3.1. Agar plate assay Protease, amylase, and cellulase activity of water extracts were determined by agar plate assay by the method of Cowan and Daniel (1982) with a slight modification. All chemicals were purchased from Wako Pure Chemicals Inc. (Tokyo, Japan). (1) Protease: 2 g of agar and 0.1 g of NaN3 were dissolved completely in 0.1 M phosphate buffer (pH 7.0) by heating, and then 1 ml of 1 M CaCl2 and 10 ml of 1% casein in 5 mM NaOH were added to the mixture. After cooling to approximately 50 C,
Table 1 The total weight and the components of raw materials on four types of compost Compost name
Pile A Pile B Reactor A Reactor B
Composting system
The The The The
turning pile system A turning pile system B static reactor system A static reactor system B
Total weight (kg)
6615 4540 400 400
Percentage of the component Chip of tree cutting
Rice hull
Bread and/or rice
Vegetable waste
Catering waste
13 13 13 14
13 13 13 14
19 20 19 18
26 25 25 37
29 29 30 28
The mixture of natural microorganisms, named Uron C (A) or Uchishiro (B), was added.
M. Chikae et al. / Bioresource Technology 97 (2006) 1979–1985
15 ml of the mixture was poured into plastic petri dishes (9 cm diameter), and was allowed to solidify. (2) a-Amylase: 0.8 g of soluble starch and 2 g of agar were dissolved completely in 100 ml of 0.1 M acetate buffer (pH 5.0) by heating. Test plates were prepared in the same manner. (3) Cellulase: 0.8 g of carboxymethylcellulose (sodium salt) and 2 g of agar were dissolved completely in 0.1 M acetate buffer (pH 5.0) by heating. Test plates were also prepared similarly. Fourteen sample wells with 3 mm diameter were made on a test plate using cork-borer. An aliquot (10 ll) of the water extract of the compost sample was applied to each well, and incubated at 55 C for 18 h. After incubation, proteolysis was observed as a clear halo around the sample well. Amylolysis and cellulolysis were observed as a halo visualized by staining agar plate. Visualization of amylolysis was carried out by adding 5 ml of 0.2% KI and 0.02% I2 solution. Visualization of cellulolysis was carried out by exposing the agar to 5 ml of 0.1 M Na2CO3 for 3 min, and then to 5 ml of 0.1% Congo Red for 20 min, and followed by the addition of 1 M NaCl with moderate shaking. The diameter of a halo was measured by a caliper. 2.3.2. API ZYMTM assay Nineteen enzymes were assayed using the kit of API ZYMTM (BioMerieux, France). The assayed enzymes were as follows: alkaline phosphatase, acid phosphatase, phosphohydrolase, esterase, esterase–lipase, lipase, leucine arylamidase, valine arylamidase, cystine arylamidase, trypsin, chymotrypsin, a-galactosidase, b-galactosidase, b-glucuronidase, a-glucosidase, b-glucosidase, N-acetyl-b-glucosaminidase, a-mannosidase, a-fucosidase. 2.4. Komatsuna seed germination test A fresh water extract (10 ml) was dropped into a plastic petri dish with a filter paper. Thirty Komatsuna (Campestris brassica) seeds, a kind of Chinese cabbage, were distributed on the filter paper, and incubated at 26 C in the dark for 48 h under cover. The numbers of germinating seeds were counted and the lengths of the root radical were measured. As a control, 10 ml of distilled water was replaced with the extract at every treatment. The GI was calculated according to the following formula by Zucconi et al. (1981). Seed germination root length of treatment 100 GI ð%Þ ¼ Seed germination root length of control 2.5. Statistics The GI value and other parameters, water content, tem perature, pH, EC, NHþ 4 or NO3 concentration, C or N content, C/N ratio, and 22 kinds of enzyme activities, were
1981
correlated by stepwise multiple regression using StatView (version 5.0: SAS Institute Inc., USA). The correlation coefficients between each parameter were calculated using the same software. 3. Results and discussion 3.1. Physicochemical properties The values for the physicochemical parameters of each compost measured in selected times are shown in Tables 2a and 2b. The maximum temperature was approximately 65–72 C in each composting system. The periods until decrease of temperature to 30 C were about 85 and 81 days in the turning pile systems A and B, also 42 and 35 days in the static aerated reactor systems A and B, respectively. However, it should be noted that the static aerated reactor systems were more influenced from low atmospheric temperature because of its smaller volume. Water contents also reduced with the composting procedure. On the piles A and B, water was added on a turning time basis, when water content dropped below 45%. The reactors A and B required no manual addition of water, but observed high water contents caused by the accumulated snow. Total N contents of final products tested composts ranged from 1.7 to 2.6, and these were corresponding with the composts from animal manure (Huang et al., 2001). In general, the total N contents of composts by turning pile system was increased with the composting procedure, while those by static aerated reactor systems were slightly increased. It was reported that the volatilization loss of ammonia was increased above pH 8 (Ekinci et al., 2000). In the static aerated reactor system, the volatilization loss of ammonia seemed to be more increased. The pH values were ranged from 5.5 to 8, and then decreased to around 7.5. Only a little change was observed in EC. This change could be correlated with NHþ 4 concentration (r = 0.618). The NO concentration increased with 3 the reduction of NHþ . However, the reduction of NO 3 4 observed at later points indicated to be the loss caused by rain or snow. 3.2. Enzyme activities The enzyme activities of water extracts from compost sample are shown in Tables 3a and 3b. In the reactors A and B, there was a wide variety of activities according to the location, (the change of value appeared on the surface initially, extended to the center, and the effect of different location began to disappear on later composting stages) then, only the value of the center of the reactor are shown here. In the API ZYMTM assay, data of eleven kinds of enzymes showing stronger activity are shown in the Tables 3a and 3b. The enzyme activities of reactors A and B were relatively higher than those of piles A and B. The changing pattern of activities of enzymes, such as a-amylase, protease, alkaline
1982
M. Chikae et al. / Bioresource Technology 97 (2006) 1979–1985
Table 2a Physicochemical properties of the products on the turning pile system at selected times Parameter
Pile A: composting time (days)
Pile B: composting time (days)
0
35
70
95
130
0
32
68
95
124
Solid Temperature (C) Water content (%) Total carbon (%) Total nitrogen (%) C/N ratio
66.3 61.3 42.3 1.58 26.7
65.4 53.8 42.6 1.75 24.1
66.6 49.7 39.2 2.23 17.9
28.1 49.0 41.1 2.27 18.1
24.3 51.7 38.7 2.6 14.9
65.1 62.2 43.3 1.38 31.5
71.1 55.5 42.3 1.46 29.0
49.3 48.1 41.1 1.67 24.7
44.3 49.3 40.4 2.07 19.6
24.6 41.1 39.0 1.70 22.3
Water extract pH EC (mc/cm) NHþ 4 (mg/L) NO 3 (mg/L)
5.6 3.2 13.3 ND
4.7 5.2 108.6 6.7
8.6 3.7 93.4 ND
7.6 4.3 40.5 94.6
7.7 3.1 2.6 148.6
5.4 2.7 15.3 ND
5.7 4.3 78.3 ND
8.5 3.1 66.6 ND
7.8 3.4 16.1 62.2
7.4 3.3 2.87 ND
ND: not detected.
Table 2b Physicochemical properties of the products on the static reactor system at selected times Parameter
Reactor A: composting time (days)
Reactor B: composting time (days)
0
28
42
70
136
0
28
42
95
132
Solid Temperature (C)a Water content (%) Total carbon (%) Total nitrogen (%) C/N ratio
62.8 61.3 42.3 1.59 26.7
46.3 37.9 43.2 1.66 26.5
42.2 35.4 41.6 1.85 22.8
34.9 40.9 42.4 1.99 21.3
13.2 68.4 42.8 2.00 21.5
65.0 62.2 43.4 1.38 32.4
35.1 47.8 42.1 1.70 25.1
7.9 39.2 41.5 1.95 21.4
18.9 67.4 42.5 1.80 23.8
15.1 68.5 39.8 1.77 22.7
Water extract pH EC (mc/cm) NHþ 4 (mg/L) NO 3 (mg/L)
5.6 3.1 12.7 8.7
6.3 2.8 45.8 8.8
7.4 3.2 46.1 ND
6.8 3.0 30.2 53.1
7.2 3.3 34.0 10.3
5.4 2.7 15.0 ND
8.1 3.5 69.0 6.4
8.1 3.6 86.9 28.4
7.8 4.7 6.4 ND
7.8 4.7 5.0 ND
ND: not detected. a The temperature of the center of reactor.
Table 3a Enzyme activity of water extracts of the piles A and B at selected times Enzymes
Pile A: composting time (days)
Pile B: composting time (days)
0
28
42
70
130
0
28
42
95
124
Agar plate assay a-Amylase Protease Cellulase
1 ND 5
4 2 5
4 3 2
ND 4 5
ND ND 1
1 ND 5
5 3 5
4 5 4
ND ND 1
ND ND ND
API ZYMTM assay Alkaline phosphatase Acid phosphatase Phosphohydrolase Esterase Esterase–lipase Leucine arylamidase a-Galactosidase b-Galactosidase b-Glucosidase N-Acetyl-b-glucosaminidase
2 5 3 1 1 2 1 2 1 4
ND 1 1 ND ND ND 1 1 1 ND
2 1 ND ND ND ND 2 ND 1 3
4 2 2 1 2 5 ND 4 ND 2
3 2 1 2 2 1 ND ND 1 1
ND 5 3 ND ND ND 1 3 1 1
1 4 1 ND ND 1 2 2 1 1
5 2 1 2 2 4 1 4 3 4
2 1 1 1 1 2 ND 1 1 ND
2 1 1 1 1 2 ND ND 1 1
ND: not detected.
and acid phosphatase, leucine arylamidase, and b-galactosidase were similar to the changing pattern of the tempera-
ture. It was suggested that the enzymatic activity during composting could reflect the dynamics of the composting
M. Chikae et al. / Bioresource Technology 97 (2006) 1979–1985
1983
Table 3b Enzyme activity of water extracts of reactors A and B at selected times Reactor Aa: composting time (days)
Enzymes
Reactor Ba: composting time (days)
0
28
42
70
136
0
28
42
95
132
Agar plate assay a-Amylase Protease Cellulase
ND ND 5
2 ND 3
8 5 6
6 4 4
ND ND ND
ND ND 5
8 7 4
5 8 6
ND ND ND
ND ND ND
API ZYMTM assay Alkaline phosphatase Acid phosphatase Phosphohydrolase Esterase Esterase–lipase Leucine arylamidase a-Galactosidase b-Galactosidase b-Glucosidase N-Acetyl-b-glucosaminidase
2 5 3 1 1 2 1 2 1 4
5 3 2 2 2 5 5 5 5 5
4 4 3 2 3 5 5 5 5 5
5 4 3 1 1 5 3 4 5 4
2 2 1 1 1 1 ND ND ND 1
ND 5 3 ND ND ND 1 3 1 1
5 2 2 2 4 3 4 4 5 5
5 3 3 3 1 3 5 5 5 5
1 3 2 1 ND 1 ND ND 1 1
1 1 1 1 1 1 ND ND ND 1
ND: not detected. a The data of the samples from the center of the reactor were shown.
process in terms of the decomposition of organic materials, and probably provided information about the maturity of the product. Vuorinen (1999, 2000) reported that the acid and alkaline phosphatase were widely observed with composting process, and b-D-glucosidase was a key enzyme to cellulosic plant materials in the compost from animal manure. Additionally, extracellular enzyme activities during manure composting had a positive correlation with the population of three microbial groups, total aerobic heterotrophs, actinomycetes and fungi (Tiquia, 2002). In general, the present data are in good agreement with these reports, and both agar plate and API ZYMTM assays were simple and easy ways to estimate the enzyme activity of a crude solution. 3.3. Germination test Germination index (GI) is one of the most sensitive parameters for evaluating the toxicity and the degree of compost maturity, and the GI of 50% has been used as
an indicator of phytotoxin-free composts (Zucconi et al., 1981). Since sample storage influenced the parameterÕs value for maturity evaluation (Wu and Ma, 2001), freshly collected and prepared water extracts were applied. The changes of the GI values in this study are shown in Fig. 1. The piles A and B required 70 and 60 days of composting for maturation, while reactors A and B required 56 and 28 days, respectively. Judging from GI values, the efficiency of composting system in this study could be shown as follows: the static aerated reactor system B > the static aerated reactor system A > the turning pile system B > the turning pile system A. In this study, the changing pattern of GI values was similar among all tested composts, it increased to the level of 60–80%, and decreased during approximately two weeks, then increased again. These values were different from the changing pattern of the animal manure compost, which showed only an increasing pattern or a slight reduction at the first step of composting (Mathur et al., 1993b; Huang et al., 2001). This slight reduction may be attributed to the 120
120
b
100
100
80
80
GI (%)
GI (%)
a
60
60
40
40
20
20 0
0 0
50
100
150
Composting time (days)
200
0
50
100
150
200
Composting time (days)
Fig. 1. The change of the GI value in each compost type: (a) the turning pile system A (d) and B (s), (b) the static reactor system A (d) and B (s). Each value indicates the mean ± SE (n = 3).
1984
M. Chikae et al. / Bioresource Technology 97 (2006) 1979–1985
release of high concentrations of NH3 and low molecular weight organic acids (Wong, 1985). However, in this study, there was no relationship between GI values and NH3 concentrations (r = 0.019), therefore, further examination to clarify this reduction event is undertaken in our laboratory. 3.4. Multiple regression analysis The GI values, which were adopted as the markers of compost maturity, and other parameters, water content,
Table 4 Selected parameters and coefficients estimated by multi-linear stepwise regression analysis Factor
Regression coefficient
Standard error
S.R.C.a
F-value
Intercept pH EC NHþ 4 concentration NO 3 concentration Temperature Alkaline phosphatase Acid phosphatase Phosphohydrolase Esterase b-glucosidase
43.028 15.430 4.982 0.291 0.183 0.244 3.181 5.504 2.962 9.675 3.530
13.312 1.341 2.019 0.039 0.052 0.094 1.178 1.026 1.429 1.954 1.168
43.028 0.703 0.132 0.489 0.133 0.140 0.183 0.307 0.110 0.267 0.150
10.447 132.478 6.088 55.463 12.203 6.725 7.284 28.761 4.297 24.526 9.133
a
Standard regression coefficient.
100
temperature, pH, EC, NHþ 4 and NO3 concentrations, C and N contents, C/N ratio, and 22 kinds of enzyme activities, were correlated by using a multi-linear stepwise regression analysis method (159 samples). The selected parameters and those properties are shown in Table 4. In this model, a highly significant (p < 0.0001) coefficient of multiple determination (R2) as 0.896 was obtained. For a more simple model, four parameters having higher þ F-value, pH, NHþ 4 concentration (½NH4 ), acid phosphatase activity (AP), and esterase activity (E), were further selected. The regression model generated from them was expressed by the following equation:
GI ¼ 60:239 þ 19:057pH 0:238½NHþ 4 5:381AP 5:117E A highly significant (p < 0.0001) coefficient of multiple determination (R2) of 0.791 for this model indicated that GI value could be estimated by these parameters. Using this equation, the predicted GI values were calculated in each compost type. High correlations between the measured GI values and the predicted ones in each compost type were obtained (Fig. 2), and PearsonÕs correlation coefficients (r) of piles A and B, reactors A and B were found as 0.96, 0.93, 0.84, and 0.89, respectively. The slopes of regression lines were close to 1 in both of the turning systems, but they were found to be lower, 0.51 or 0.6, in both of the reactor systems. For the prediction of GI values, it was sug100
Pile A
Pile B 80
Predicted GI (%)
Predicted GI (%)
80 60 40 20 0
60 40 20
0
y = 1.0813x + 1.2732 r = 0.96
y = 0.9595x + 5.4255 r = 0.93
-20
-20 0
20
40
60
80
100
0
20
40
100
Reactor B
Reactor A
80
Predicted GI (%)
80
Predicted GI (%)
80
100
100
60 40 20 0 -20
60
GI (%)
GI (%)
20
40
60
GI (%)
80
100
40 20 0
y = 0.5162x + 13.365 r = 0.84
0
60
120
-20
y = 0.6007x + 12.476 r = 0.89
0
20
40
60
80
100
120
GI (%)
Fig. 2. The correlations between the measured GI values and predicted ones for each compost type. The equation of regression line and PearsonÕs correlation coefficient (r) were shown.
M. Chikae et al. / Bioresource Technology 97 (2006) 1979–1985
gested that a correction according to the composting system might be needed as performed here. The selected parameters in this study, pH, NHþ 4 concentration, acid phosphatase activity, and esterase activity had a low correlation with GI value, their correlation coefficients were 0.369, 0.019, 0.047, and 0.321, respectively. Thus, these parameters could not estimate a GI value only by themselves, but they could be utilized to predict it by a combination of them together. The multiple regression analysis was found to be a useful method to combine them. Additionally, the accuracy of the predicted value would be increased by a re-correlation according to the composting process or the source materials. As a conclusion, the compost maturity could be predicted by only sensing of the water extract at the composting site without a large-sized equipment and skill, and this prediction system may contribute to the production of stable composts in the recycling market. Acknowledgements Supports for this work from Kaga City, Ishikawa Agricultural Research Center, Resource Ecology Recycle Common Facility Cooperatives, and Mizushimabussan Co. Ltd. are gratefully acknowledged. The authors would like to thank Mr. S. Kitamura, Mr. S. Singo, and Mr. N. Kimoto for their experimental discussion and assistance throughout the investigation. References Cowan, D.A., Daniel, R.M., 1982. A modification for increasing the sensitivity of the casein–agar plate assay: a simple semiquantitative assay for thermophilic and mesophilic proteases. J. Biochem. Biophys. Meth. 6, 31–37. de Oliveira, S.C., Provenzano, M.R., Silva, M.R.S., Senesi, N., 2002. Maturity degree of composts from municipal solid wastes evaluated by differential scanning calorimetry. Environ. Technol. 23, 1099–1105. Ekinci, K., Keener, H.M., Elwell, D.L., 2000. Composting short paper fiber with broiler litter and additives. Compost Sci. Util. 8 (2), 160–172.
1985
Hogland, W., Bramryd, T., Marques, M., Nimmermark, S., 2003. Physical, chemical and biological processes for optimizing decentralized composting. Compost Sci. Util. 11 (4), 330–336. Huang, G., Wu, Q., Li, F., Wong, J.W.C., 2001. Nitrogen transformations during pig manure composting. J. Environ. Sci. 13 (4), 401–405. Jime´nez, E.I., Garcia, V.P., 1989. Evaluation of city refuse compost maturity: a review. Biol. Waste. 27, 115–142. Koshino, M., Imai, J., Sanpei, M., Yamazoe, F., Yoshida, N., Fujii, K., Miwa, E., 1988a. Analysis of Compost (Japanese), second ed., Yokendo, Tokyo, pp. 42–45. Koshino, M., Imai, J., Sanpei, M., Yamazoe, F., Yoshida, N., Fujii, K., Miwa, E., 1988b. Analysis of Compost (Japanese), second ed., Yokendo, Tokyo, pp. 52–55. Mathur, S.P., Owen, G., Dinel, H., Schnitzer, M., 1993a. Determination of compost biomaturity. I. Literature review. Biol. Agric. Hortic. 10, 65–85. Mathur, S.P., Dinel, H., Owen, G., Schnitzer, M., Dugan, J., 1993b. Determination of compost biomaturity. II. Optical density of water extracts of composts as a reflection of their maturity. Biol. Agric. Hortic. 10, 87–108. Ouatmane, A., Provenzano, M.R., Hafidi, M., Senesi, N., 2000. Compost maturity assessment using calorimetry, spectroscopy and chemical analysis. Compost Sci. Util. 8 (2), 124–134. Provenzano, M.R., de Oliveira, S.C., Silva, M.R.S., Senesi, N., 2001. Assessment of maturity degree of composts from domestic solid wastes by fluorescence and fourier transform infrared spectroscopies. J. Agric. Food Chem. 49, 5874–5879. Ranalli, G., Bottura, G., Taddei, P., Garavani, M., marchetti, P., Sorlini, C., 2001. Composting of solid and sludge residues from agricultural and food industries. Bioindicators of monitoring and compost maturity. J. Environ. Sci. Health A36 (4), 415–436. Tiquia, S.M., 2002. Evolution of extracellular enzyme activities during manure composting. J. Appl. Microbiol. 92, 764–775. Vuorinen, A.H., 1999. Phosphatases in horse and chicken manure composts. Compost Sci. Util. 7 (2), 47–54. Vuorinen, A.H., 2000. Effect of bulking agent on acid and alkaline phosphomonoesterase and b-D-glucosidase activities during manure composting. Biores. Technol. 75, 133–138. Wong, M.H., 1985. Phytotoxicity of refuse compost during the process of maturation. Environ. Pollut. 40, 127–144. Wu, L., Ma, L.Q., 2001. Effects of sample storage on biosolids compost stability and maturity evaluation. J. Environ. Qual. 30, 222–228. Zucconi, F., Forte, M., Monaco, A., de Bertoldi, M., 1981. Biological evaluation of compost maturity. BioCycle 22 (2), 27–29.
Estimation of maturity of compost from food wastes and agro-residues by multiple regression analysis Miyuki Chikae, Ryuzoh Ikeda, Kagan Kerman, Yasutaka Morita, Eiichi Tamiya
*
School of Materials Science, Japan Advanced Institute of Science and Technology, 1-1, Asahidai, Nomi, Ishikawa 923-1292, Japan Received 30 August 2005; received in revised form 28 September 2005; accepted 30 September 2005 Available online 9 November 2005
Abstract The composting process of food wastes and tree cuttings was examined on four composting types composed from two kinds of systems and added mixture of microorganisms. The time courses of 32 parameters in each composting type were observed. The efficient composting system was found to be the static aerated reactor system in comparison with the turning pile one. Using the multiple regression analysis of all the data (159 samples) obtained from this study, some parameters were selected to predict the germination index (GI) value, which was adopted as a marker of compost maturity. For example, using the regression model generated from pH, NHþ 4 concentration, acid phosphatase activity, and esterase activity of water extracts of the compost, GI value was expressed by the multi-linear regression equation (p < 0.0001). High correlations between the measured GI value and the predicted one were made in each type of compost. As a result of these observations, the compost maturity might be predicted by only sensing of the water extract at the composting site without any requirements for a large-size equipment and skill, and this prediction system could contribute to the production of a stable compost in wide-spread use for the recycling market. 2005 Elsevier Ltd. All rights reserved. Keywords: Composting; Food waste; Tree cutting; Maturity; Multiple regression analysis
1. Introduction The present trend about waste management is recycling and the recovery of waste as new materials or as energy. The waste organic materials produced by the city life, such as domestic refuse, residues from food processing and wastes from food industry are accumulating to become a significant amount. The composting of these by-products is more encouraged to avoid the loss of energy. The compost from domestic wastes possesses a high nutritional value, with high concentration of nitrogen, potassium and phosphorus, while the contamination by heavy metals and other toxic substances is negligible (Hogland et al., 2003). However, for the development of a market of recycle and recovery of waste to compost, not only must the qual-
*
Corresponding author. Tel.: +81 761 51 1660; fax: +81 761 51 1665. E-mail address: [email protected] (E. Tamiya).
0960-8524/$ - see front matter 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.biortech.2005.09.026
ity of products be stable, but also the composting system must be efficient. The evaluation of the maturity of domestic wastes compost has been widely recognized as one of the most important factors concerning the compost process and the application of these by-products. The land-applied immature compost gives rise to a serious N-deficiency in crops, and the rapid decomposition of immature compost causes a decrease of the O2-concentration around the root system (Mathur et al., 1993a). Additionally, land-applied immature compost inhibits the plant growth by the production of phytotoxic substances, fundamentally ammonia, ethylene oxide, and organic acids (Mathur et al., 1993a). Many parameters to evaluate the maturity of compost from food wastes or city refuse, such as the change of physicochemical properties (Mathur et al., 1993a), the enzymatic activity (Ranalli et al., 2001), germination tests (Zucconi et al., 1981), calorimetric and spectroscopic methods (de Oliveira et al., 2002; Ouatmane et al., 2000; Provenzano et al., 2001)
1980
M. Chikae et al. / Bioresource Technology 97 (2006) 1979–1985
have been reported, and reviewed (Jime´nez and Garcia, 1989). To the best of our knowledge, no parameters for using at the composting site in a simple way have been reported. Food wastes of school lunch, which are similar to domestic refuse, and on which source separation can be done easily, were incinerated in Kaga City, a town on the Honshu Island of Japan with a population of 67,600. On the other hand, tree cuttings from parks and roadsides are also occupying a significant amount of land in Kaga City. The composting of these materials requires a reasonable recycling management. The aims of this study were: (1) to search an efficient composting system of food wastes and tree cuttings from four comparing types; and (2) to establish a detection way of the compost maturity, which could be used easily at the composting site. Then, the time courses of physicochemical and biological parameters of the products from four composting types were observed, and the relationships between germination index (GI) values and other parameters in each sample were determined by using stepwise multiple regression analysis.
the compost reactor, placed outdoors, was a cylinder with 1.5 m diameter, 1.5 m in height and a hole in the center for aeration. The sidewalls were made of stainless steel mesh and the top was covered with a sheet. Temperature measuring and sampling were carried out at three different locations, such as, at the surface, 10 cm depth from surface, and the center of the reactor once per week. Composting experiments were performed at the Mizushimabussan Co. Ltd. in Kaga City from October, 2003 till March, 2004. The mean atmospheric temperatures of day during experiment were between 1.6 and 11.7 C. The total weight and the main components of the raw materials in four composts are shown in Table 1. 2.2. Physicochemical analysis Water content was determined by oven-drying at 105 C for 5 h. Total carbon content and total nitrogen content were measured by auto elemental analyzer Vario EL III (Elementar, German), and C/N was calculated from them. Water extracts of the compost samples were prepared by shaking the fresh sample with distilled water at 1:10 w/v (dry weight basis) using a horizontal shaker for 30 min at room temperature by filtration. The fresh extracts were subjected to measuring pH, electrical conductivity (EC), and the remaining extracts were kept at 20 C until use. The thawed extracts were used for determination of concentrations of NHþ 4 by the Indophenol Blue method (Koshino et al., 1988a), and NO 3 by the salicylic acid–sulfuric acid method derived from the phenol–sulfuric acid one (Koshino et al., 1988b).
2. Methods 2.1. Composting and compost sampling The organic fractions of the food waste of school lunch, separated into three groups of bread/rice, vegetablesÕ waste, and leftovers of lunch, were collected from a junior high school in Kaga City. These obtained food wastes were mixed with rice hulls and similar size of chips of tree cuttings, as bulk materials, and a mixture of natural microorganisms, commercialized as Uron C (A; Tsurumi Soda, Japan) or Uchishiro (B; Seiwa, Japan), was added for accelerating the composting process. The mixture was stirred mechanically for 3 h in 65 C (Daido, Japan), and the water content was adjusted to approximately 65% by control of volume of bulk materials. After then, the treated mixture was composted by two systems, the turning pile system, and the static aerated system using the compost reactor (Mizushimabussann, Japan). On the turning pile system, the heaps in 5 m · 5 m yard were turned manually two times per week. After each turning, temperatures and samples were taken from three different locations of the compost pile once per week. On the static aerated system,
2.3. Enzyme activity assay 2.3.1. Agar plate assay Protease, amylase, and cellulase activity of water extracts were determined by agar plate assay by the method of Cowan and Daniel (1982) with a slight modification. All chemicals were purchased from Wako Pure Chemicals Inc. (Tokyo, Japan). (1) Protease: 2 g of agar and 0.1 g of NaN3 were dissolved completely in 0.1 M phosphate buffer (pH 7.0) by heating, and then 1 ml of 1 M CaCl2 and 10 ml of 1% casein in 5 mM NaOH were added to the mixture. After cooling to approximately 50 C,
Table 1 The total weight and the components of raw materials on four types of compost Compost name
Pile A Pile B Reactor A Reactor B
Composting system
The The The The
turning pile system A turning pile system B static reactor system A static reactor system B
Total weight (kg)
6615 4540 400 400
Percentage of the component Chip of tree cutting
Rice hull
Bread and/or rice
Vegetable waste
Catering waste
13 13 13 14
13 13 13 14
19 20 19 18
26 25 25 37
29 29 30 28
The mixture of natural microorganisms, named Uron C (A) or Uchishiro (B), was added.
M. Chikae et al. / Bioresource Technology 97 (2006) 1979–1985
15 ml of the mixture was poured into plastic petri dishes (9 cm diameter), and was allowed to solidify. (2) a-Amylase: 0.8 g of soluble starch and 2 g of agar were dissolved completely in 100 ml of 0.1 M acetate buffer (pH 5.0) by heating. Test plates were prepared in the same manner. (3) Cellulase: 0.8 g of carboxymethylcellulose (sodium salt) and 2 g of agar were dissolved completely in 0.1 M acetate buffer (pH 5.0) by heating. Test plates were also prepared similarly. Fourteen sample wells with 3 mm diameter were made on a test plate using cork-borer. An aliquot (10 ll) of the water extract of the compost sample was applied to each well, and incubated at 55 C for 18 h. After incubation, proteolysis was observed as a clear halo around the sample well. Amylolysis and cellulolysis were observed as a halo visualized by staining agar plate. Visualization of amylolysis was carried out by adding 5 ml of 0.2% KI and 0.02% I2 solution. Visualization of cellulolysis was carried out by exposing the agar to 5 ml of 0.1 M Na2CO3 for 3 min, and then to 5 ml of 0.1% Congo Red for 20 min, and followed by the addition of 1 M NaCl with moderate shaking. The diameter of a halo was measured by a caliper. 2.3.2. API ZYMTM assay Nineteen enzymes were assayed using the kit of API ZYMTM (BioMerieux, France). The assayed enzymes were as follows: alkaline phosphatase, acid phosphatase, phosphohydrolase, esterase, esterase–lipase, lipase, leucine arylamidase, valine arylamidase, cystine arylamidase, trypsin, chymotrypsin, a-galactosidase, b-galactosidase, b-glucuronidase, a-glucosidase, b-glucosidase, N-acetyl-b-glucosaminidase, a-mannosidase, a-fucosidase. 2.4. Komatsuna seed germination test A fresh water extract (10 ml) was dropped into a plastic petri dish with a filter paper. Thirty Komatsuna (Campestris brassica) seeds, a kind of Chinese cabbage, were distributed on the filter paper, and incubated at 26 C in the dark for 48 h under cover. The numbers of germinating seeds were counted and the lengths of the root radical were measured. As a control, 10 ml of distilled water was replaced with the extract at every treatment. The GI was calculated according to the following formula by Zucconi et al. (1981). Seed germination root length of treatment 100 GI ð%Þ ¼ Seed germination root length of control 2.5. Statistics The GI value and other parameters, water content, tem perature, pH, EC, NHþ 4 or NO3 concentration, C or N content, C/N ratio, and 22 kinds of enzyme activities, were
1981
correlated by stepwise multiple regression using StatView (version 5.0: SAS Institute Inc., USA). The correlation coefficients between each parameter were calculated using the same software. 3. Results and discussion 3.1. Physicochemical properties The values for the physicochemical parameters of each compost measured in selected times are shown in Tables 2a and 2b. The maximum temperature was approximately 65–72 C in each composting system. The periods until decrease of temperature to 30 C were about 85 and 81 days in the turning pile systems A and B, also 42 and 35 days in the static aerated reactor systems A and B, respectively. However, it should be noted that the static aerated reactor systems were more influenced from low atmospheric temperature because of its smaller volume. Water contents also reduced with the composting procedure. On the piles A and B, water was added on a turning time basis, when water content dropped below 45%. The reactors A and B required no manual addition of water, but observed high water contents caused by the accumulated snow. Total N contents of final products tested composts ranged from 1.7 to 2.6, and these were corresponding with the composts from animal manure (Huang et al., 2001). In general, the total N contents of composts by turning pile system was increased with the composting procedure, while those by static aerated reactor systems were slightly increased. It was reported that the volatilization loss of ammonia was increased above pH 8 (Ekinci et al., 2000). In the static aerated reactor system, the volatilization loss of ammonia seemed to be more increased. The pH values were ranged from 5.5 to 8, and then decreased to around 7.5. Only a little change was observed in EC. This change could be correlated with NHþ 4 concentration (r = 0.618). The NO concentration increased with 3 the reduction of NHþ . However, the reduction of NO 3 4 observed at later points indicated to be the loss caused by rain or snow. 3.2. Enzyme activities The enzyme activities of water extracts from compost sample are shown in Tables 3a and 3b. In the reactors A and B, there was a wide variety of activities according to the location, (the change of value appeared on the surface initially, extended to the center, and the effect of different location began to disappear on later composting stages) then, only the value of the center of the reactor are shown here. In the API ZYMTM assay, data of eleven kinds of enzymes showing stronger activity are shown in the Tables 3a and 3b. The enzyme activities of reactors A and B were relatively higher than those of piles A and B. The changing pattern of activities of enzymes, such as a-amylase, protease, alkaline
1982
M. Chikae et al. / Bioresource Technology 97 (2006) 1979–1985
Table 2a Physicochemical properties of the products on the turning pile system at selected times Parameter
Pile A: composting time (days)
Pile B: composting time (days)
0
35
70
95
130
0
32
68
95
124
Solid Temperature (C) Water content (%) Total carbon (%) Total nitrogen (%) C/N ratio
66.3 61.3 42.3 1.58 26.7
65.4 53.8 42.6 1.75 24.1
66.6 49.7 39.2 2.23 17.9
28.1 49.0 41.1 2.27 18.1
24.3 51.7 38.7 2.6 14.9
65.1 62.2 43.3 1.38 31.5
71.1 55.5 42.3 1.46 29.0
49.3 48.1 41.1 1.67 24.7
44.3 49.3 40.4 2.07 19.6
24.6 41.1 39.0 1.70 22.3
Water extract pH EC (mc/cm) NHþ 4 (mg/L) NO 3 (mg/L)
5.6 3.2 13.3 ND
4.7 5.2 108.6 6.7
8.6 3.7 93.4 ND
7.6 4.3 40.5 94.6
7.7 3.1 2.6 148.6
5.4 2.7 15.3 ND
5.7 4.3 78.3 ND
8.5 3.1 66.6 ND
7.8 3.4 16.1 62.2
7.4 3.3 2.87 ND
ND: not detected.
Table 2b Physicochemical properties of the products on the static reactor system at selected times Parameter
Reactor A: composting time (days)
Reactor B: composting time (days)
0
28
42
70
136
0
28
42
95
132
Solid Temperature (C)a Water content (%) Total carbon (%) Total nitrogen (%) C/N ratio
62.8 61.3 42.3 1.59 26.7
46.3 37.9 43.2 1.66 26.5
42.2 35.4 41.6 1.85 22.8
34.9 40.9 42.4 1.99 21.3
13.2 68.4 42.8 2.00 21.5
65.0 62.2 43.4 1.38 32.4
35.1 47.8 42.1 1.70 25.1
7.9 39.2 41.5 1.95 21.4
18.9 67.4 42.5 1.80 23.8
15.1 68.5 39.8 1.77 22.7
Water extract pH EC (mc/cm) NHþ 4 (mg/L) NO 3 (mg/L)
5.6 3.1 12.7 8.7
6.3 2.8 45.8 8.8
7.4 3.2 46.1 ND
6.8 3.0 30.2 53.1
7.2 3.3 34.0 10.3
5.4 2.7 15.0 ND
8.1 3.5 69.0 6.4
8.1 3.6 86.9 28.4
7.8 4.7 6.4 ND
7.8 4.7 5.0 ND
ND: not detected. a The temperature of the center of reactor.
Table 3a Enzyme activity of water extracts of the piles A and B at selected times Enzymes
Pile A: composting time (days)
Pile B: composting time (days)
0
28
42
70
130
0
28
42
95
124
Agar plate assay a-Amylase Protease Cellulase
1 ND 5
4 2 5
4 3 2
ND 4 5
ND ND 1
1 ND 5
5 3 5
4 5 4
ND ND 1
ND ND ND
API ZYMTM assay Alkaline phosphatase Acid phosphatase Phosphohydrolase Esterase Esterase–lipase Leucine arylamidase a-Galactosidase b-Galactosidase b-Glucosidase N-Acetyl-b-glucosaminidase
2 5 3 1 1 2 1 2 1 4
ND 1 1 ND ND ND 1 1 1 ND
2 1 ND ND ND ND 2 ND 1 3
4 2 2 1 2 5 ND 4 ND 2
3 2 1 2 2 1 ND ND 1 1
ND 5 3 ND ND ND 1 3 1 1
1 4 1 ND ND 1 2 2 1 1
5 2 1 2 2 4 1 4 3 4
2 1 1 1 1 2 ND 1 1 ND
2 1 1 1 1 2 ND ND 1 1
ND: not detected.
and acid phosphatase, leucine arylamidase, and b-galactosidase were similar to the changing pattern of the tempera-
ture. It was suggested that the enzymatic activity during composting could reflect the dynamics of the composting
M. Chikae et al. / Bioresource Technology 97 (2006) 1979–1985
1983
Table 3b Enzyme activity of water extracts of reactors A and B at selected times Reactor Aa: composting time (days)
Enzymes
Reactor Ba: composting time (days)
0
28
42
70
136
0
28
42
95
132
Agar plate assay a-Amylase Protease Cellulase
ND ND 5
2 ND 3
8 5 6
6 4 4
ND ND ND
ND ND 5
8 7 4
5 8 6
ND ND ND
ND ND ND
API ZYMTM assay Alkaline phosphatase Acid phosphatase Phosphohydrolase Esterase Esterase–lipase Leucine arylamidase a-Galactosidase b-Galactosidase b-Glucosidase N-Acetyl-b-glucosaminidase
2 5 3 1 1 2 1 2 1 4
5 3 2 2 2 5 5 5 5 5
4 4 3 2 3 5 5 5 5 5
5 4 3 1 1 5 3 4 5 4
2 2 1 1 1 1 ND ND ND 1
ND 5 3 ND ND ND 1 3 1 1
5 2 2 2 4 3 4 4 5 5
5 3 3 3 1 3 5 5 5 5
1 3 2 1 ND 1 ND ND 1 1
1 1 1 1 1 1 ND ND ND 1
ND: not detected. a The data of the samples from the center of the reactor were shown.
process in terms of the decomposition of organic materials, and probably provided information about the maturity of the product. Vuorinen (1999, 2000) reported that the acid and alkaline phosphatase were widely observed with composting process, and b-D-glucosidase was a key enzyme to cellulosic plant materials in the compost from animal manure. Additionally, extracellular enzyme activities during manure composting had a positive correlation with the population of three microbial groups, total aerobic heterotrophs, actinomycetes and fungi (Tiquia, 2002). In general, the present data are in good agreement with these reports, and both agar plate and API ZYMTM assays were simple and easy ways to estimate the enzyme activity of a crude solution. 3.3. Germination test Germination index (GI) is one of the most sensitive parameters for evaluating the toxicity and the degree of compost maturity, and the GI of 50% has been used as
an indicator of phytotoxin-free composts (Zucconi et al., 1981). Since sample storage influenced the parameterÕs value for maturity evaluation (Wu and Ma, 2001), freshly collected and prepared water extracts were applied. The changes of the GI values in this study are shown in Fig. 1. The piles A and B required 70 and 60 days of composting for maturation, while reactors A and B required 56 and 28 days, respectively. Judging from GI values, the efficiency of composting system in this study could be shown as follows: the static aerated reactor system B > the static aerated reactor system A > the turning pile system B > the turning pile system A. In this study, the changing pattern of GI values was similar among all tested composts, it increased to the level of 60–80%, and decreased during approximately two weeks, then increased again. These values were different from the changing pattern of the animal manure compost, which showed only an increasing pattern or a slight reduction at the first step of composting (Mathur et al., 1993b; Huang et al., 2001). This slight reduction may be attributed to the 120
120
b
100
100
80
80
GI (%)
GI (%)
a
60
60
40
40
20
20 0
0 0
50
100
150
Composting time (days)
200
0
50
100
150
200
Composting time (days)
Fig. 1. The change of the GI value in each compost type: (a) the turning pile system A (d) and B (s), (b) the static reactor system A (d) and B (s). Each value indicates the mean ± SE (n = 3).
1984
M. Chikae et al. / Bioresource Technology 97 (2006) 1979–1985
release of high concentrations of NH3 and low molecular weight organic acids (Wong, 1985). However, in this study, there was no relationship between GI values and NH3 concentrations (r = 0.019), therefore, further examination to clarify this reduction event is undertaken in our laboratory. 3.4. Multiple regression analysis The GI values, which were adopted as the markers of compost maturity, and other parameters, water content,
Table 4 Selected parameters and coefficients estimated by multi-linear stepwise regression analysis Factor
Regression coefficient
Standard error
S.R.C.a
F-value
Intercept pH EC NHþ 4 concentration NO 3 concentration Temperature Alkaline phosphatase Acid phosphatase Phosphohydrolase Esterase b-glucosidase
43.028 15.430 4.982 0.291 0.183 0.244 3.181 5.504 2.962 9.675 3.530
13.312 1.341 2.019 0.039 0.052 0.094 1.178 1.026 1.429 1.954 1.168
43.028 0.703 0.132 0.489 0.133 0.140 0.183 0.307 0.110 0.267 0.150
10.447 132.478 6.088 55.463 12.203 6.725 7.284 28.761 4.297 24.526 9.133
a
Standard regression coefficient.
100
temperature, pH, EC, NHþ 4 and NO3 concentrations, C and N contents, C/N ratio, and 22 kinds of enzyme activities, were correlated by using a multi-linear stepwise regression analysis method (159 samples). The selected parameters and those properties are shown in Table 4. In this model, a highly significant (p < 0.0001) coefficient of multiple determination (R2) as 0.896 was obtained. For a more simple model, four parameters having higher þ F-value, pH, NHþ 4 concentration (½NH4 ), acid phosphatase activity (AP), and esterase activity (E), were further selected. The regression model generated from them was expressed by the following equation:
GI ¼ 60:239 þ 19:057pH 0:238½NHþ 4 5:381AP 5:117E A highly significant (p < 0.0001) coefficient of multiple determination (R2) of 0.791 for this model indicated that GI value could be estimated by these parameters. Using this equation, the predicted GI values were calculated in each compost type. High correlations between the measured GI values and the predicted ones in each compost type were obtained (Fig. 2), and PearsonÕs correlation coefficients (r) of piles A and B, reactors A and B were found as 0.96, 0.93, 0.84, and 0.89, respectively. The slopes of regression lines were close to 1 in both of the turning systems, but they were found to be lower, 0.51 or 0.6, in both of the reactor systems. For the prediction of GI values, it was sug100
Pile A
Pile B 80
Predicted GI (%)
Predicted GI (%)
80 60 40 20 0
60 40 20
0
y = 1.0813x + 1.2732 r = 0.96
y = 0.9595x + 5.4255 r = 0.93
-20
-20 0
20
40
60
80
100
0
20
40
100
Reactor B
Reactor A
80
Predicted GI (%)
80
Predicted GI (%)
80
100
100
60 40 20 0 -20
60
GI (%)
GI (%)
20
40
60
GI (%)
80
100
40 20 0
y = 0.5162x + 13.365 r = 0.84
0
60
120
-20
y = 0.6007x + 12.476 r = 0.89
0
20
40
60
80
100
120
GI (%)
Fig. 2. The correlations between the measured GI values and predicted ones for each compost type. The equation of regression line and PearsonÕs correlation coefficient (r) were shown.
M. Chikae et al. / Bioresource Technology 97 (2006) 1979–1985
gested that a correction according to the composting system might be needed as performed here. The selected parameters in this study, pH, NHþ 4 concentration, acid phosphatase activity, and esterase activity had a low correlation with GI value, their correlation coefficients were 0.369, 0.019, 0.047, and 0.321, respectively. Thus, these parameters could not estimate a GI value only by themselves, but they could be utilized to predict it by a combination of them together. The multiple regression analysis was found to be a useful method to combine them. Additionally, the accuracy of the predicted value would be increased by a re-correlation according to the composting process or the source materials. As a conclusion, the compost maturity could be predicted by only sensing of the water extract at the composting site without a large-sized equipment and skill, and this prediction system may contribute to the production of stable composts in the recycling market. Acknowledgements Supports for this work from Kaga City, Ishikawa Agricultural Research Center, Resource Ecology Recycle Common Facility Cooperatives, and Mizushimabussan Co. Ltd. are gratefully acknowledged. The authors would like to thank Mr. S. Kitamura, Mr. S. Singo, and Mr. N. Kimoto for their experimental discussion and assistance throughout the investigation. References Cowan, D.A., Daniel, R.M., 1982. A modification for increasing the sensitivity of the casein–agar plate assay: a simple semiquantitative assay for thermophilic and mesophilic proteases. J. Biochem. Biophys. Meth. 6, 31–37. de Oliveira, S.C., Provenzano, M.R., Silva, M.R.S., Senesi, N., 2002. Maturity degree of composts from municipal solid wastes evaluated by differential scanning calorimetry. Environ. Technol. 23, 1099–1105. Ekinci, K., Keener, H.M., Elwell, D.L., 2000. Composting short paper fiber with broiler litter and additives. Compost Sci. Util. 8 (2), 160–172.
1985
Hogland, W., Bramryd, T., Marques, M., Nimmermark, S., 2003. Physical, chemical and biological processes for optimizing decentralized composting. Compost Sci. Util. 11 (4), 330–336. Huang, G., Wu, Q., Li, F., Wong, J.W.C., 2001. Nitrogen transformations during pig manure composting. J. Environ. Sci. 13 (4), 401–405. Jime´nez, E.I., Garcia, V.P., 1989. Evaluation of city refuse compost maturity: a review. Biol. Waste. 27, 115–142. Koshino, M., Imai, J., Sanpei, M., Yamazoe, F., Yoshida, N., Fujii, K., Miwa, E., 1988a. Analysis of Compost (Japanese), second ed., Yokendo, Tokyo, pp. 42–45. Koshino, M., Imai, J., Sanpei, M., Yamazoe, F., Yoshida, N., Fujii, K., Miwa, E., 1988b. Analysis of Compost (Japanese), second ed., Yokendo, Tokyo, pp. 52–55. Mathur, S.P., Owen, G., Dinel, H., Schnitzer, M., 1993a. Determination of compost biomaturity. I. Literature review. Biol. Agric. Hortic. 10, 65–85. Mathur, S.P., Dinel, H., Owen, G., Schnitzer, M., Dugan, J., 1993b. Determination of compost biomaturity. II. Optical density of water extracts of composts as a reflection of their maturity. Biol. Agric. Hortic. 10, 87–108. Ouatmane, A., Provenzano, M.R., Hafidi, M., Senesi, N., 2000. Compost maturity assessment using calorimetry, spectroscopy and chemical analysis. Compost Sci. Util. 8 (2), 124–134. Provenzano, M.R., de Oliveira, S.C., Silva, M.R.S., Senesi, N., 2001. Assessment of maturity degree of composts from domestic solid wastes by fluorescence and fourier transform infrared spectroscopies. J. Agric. Food Chem. 49, 5874–5879. Ranalli, G., Bottura, G., Taddei, P., Garavani, M., marchetti, P., Sorlini, C., 2001. Composting of solid and sludge residues from agricultural and food industries. Bioindicators of monitoring and compost maturity. J. Environ. Sci. Health A36 (4), 415–436. Tiquia, S.M., 2002. Evolution of extracellular enzyme activities during manure composting. J. Appl. Microbiol. 92, 764–775. Vuorinen, A.H., 1999. Phosphatases in horse and chicken manure composts. Compost Sci. Util. 7 (2), 47–54. Vuorinen, A.H., 2000. Effect of bulking agent on acid and alkaline phosphomonoesterase and b-D-glucosidase activities during manure composting. Biores. Technol. 75, 133–138. Wong, M.H., 1985. Phytotoxicity of refuse compost during the process of maturation. Environ. Pollut. 40, 127–144. Wu, L., Ma, L.Q., 2001. Effects of sample storage on biosolids compost stability and maturity evaluation. J. Environ. Qual. 30, 222–228. Zucconi, F., Forte, M., Monaco, A., de Bertoldi, M., 1981. Biological evaluation of compost maturity. BioCycle 22 (2), 27–29.