Biodegradation Activity of Eight Organic Substrates: A Correlation Study of Different Test Methods

The biological activity of eight organic substrates was assessed using different techniques under aerobic and anaerobic environments at different scales. The used substrates included simulated fresh to actual stabilized composts. Experiments included dynamic respiration (with 30 and 5.5 L reactors), static respiration (with 1 L reactors) and biochemical methane potential (BMP) tests (with 1.1 L flasks). The indices evaluated were: the 24-h dynamic respiration index (DRI24) and the cumulative dynamic respiration indices at 4 and 7 days (DCRI4 and DCRI7, respectively) measured with two different methods, the maximum dynamic CO2 generation rate (D-CO2_24) at 24 h and a dynamic cumulative CO2 generation after 7 days (D-CO2_7), the maximum static O2 consumption rate in 12 and 24 h (SRI12 and SRI24), the static cumulative O2 consumptions after 4 and 7 days (SCRI4 and SCRI7) and the static CO2 generation after 7 days (S-CO2) and the BMP after 30 days. The DRI24 ranged from 25 to 3000 mg O2 kg−1 VS h−1 in one lab and from 150 to 3500 mg O2 kg−1 VS h−1 in the other. A positive statistically significant correlation was achieved between the two types of dynamic indices. In addition, the CH4 production after 30 days showed a strong positive correlation with both DRI24 indices and DCRI4 and DCRI7, as measured in both labs. The static respiration indices did not correlate well with the dynamic respiration ones. The practical implications of the use of the biodegradation activity indices were also analysed and discussed.


Introduction
A wide research on respiration activity indices has been conducted in order to characterize the biodegradation activity and stability of organic substrates [1][2][3][4][5][6][7][8][9][10]. Stability usually refers to the resistance of organic matter to biodegradation [11] and/or the extent to which readily biodegradable organic matter has decomposed [12]. A high content of an easily degradable fraction means a low biological stability and vice versa [13]. Stability is commonly assessed via measurements of oxygen consumption and carbon dioxide generation [4,12,14,15], methane generation [15][16][17] as well as via temperature evolution through the self-heating test [18][19][20]. All above techniques can be adopted to measure the biodegradation activity of fresh to stable substrates.
In the current study aerobic microbial respiration activity (in which O 2 consumption and CO 2 generation are the dominant indices) and anaerobic activity (in which the CH 4 generation is the key index) were measured. These biological indices are essential for evaluating biomass degradation activity for compliance or process control purposes, since these methods, although in a smaller scale can reproduce the actual waste degradation process [21] and can designate treatment technologies.
In particular, with regard to the aerobic respiration activity, different aeration modes have been tried; static with intermittent air flow and dynamic with continuous air flow. In dynamic respiration methods, further differences exist regarding the selected airflow rate. Airflow rates between 20 and 30 ml min -1 have been reported by [9,21] in order oxygen content at the exhaust gases be above 10 % (v/v). Similarly, [4] continuously adjusted the airflow rate in order to constantly maintain an oxygen content of 140 ml L -1 air (14 % v/v) at the reactor's outlet stream.
How representatively can all these reactors measure the actual respiration activity of the material? Can they all assess the same microbial activity regardless of sample size and boundary conditions?
Although larger reactors are usually better in terms of sample representation, they are usually more complex during their operation, difficult and costly to be constructed. Moreover, they require large sample sizes and replication can be sometimes very limited to non-existent due to cost constraints. On the other hand, smaller reactors are easier to construct and operate and allow the experimenter to run simultaneously many replicated experiments. The Di.Pro.Ve-CosTech Ò system is more automated and allows the measurement of several variables at short time intervals; in addition, air flow can be controlled either through the oxygen content or through the temperature readings. However, due to economical limitations in several waste treatment facilities, plant operators are seeking less cost-benefit techniques to assess biodegradation activity and compost stability. Stability indices are used for compliance tests in Germany at a national level [26], in Italy at regional/provincial level [27] and in England and Wales [28], for both MBT materials and composts. Moreover, DRI has recently received official recognition at EU level for purposes of standardization [29] and validation [30].
The objective of this work refers to (1) assess the biodegradation activity of 8 organic substrates of different biodegradabilities in two different laboratories, using different microbial activity assessment techniques per laboratory and (2) investigate correlations among those indices. One laboratory is located at the Mediterranea University of Reggio Calabria (Italy), and will be herein referred to as either UNIRC or ITA, whilst the other is located at the Democritus University of Thrace in Xanthi (Greece) and will be herein referred to as DUTH or GRE.

Substrate Description and Preparation
Eight different organic substrates of different origin and composition were investigated in this study (Table 1). Samples were selected in such a manner so that to cover a wide range of organic matter contents and expected degradabilities.
HC_1 was a mature home compost obtained from household organic waste, produced by kitchen and garden waste mixed with a limited quantity of wood combustion residues, constituted by coal and ash. Sampling for that substrate took place after one year of composting in a custom made plastic composter, by opening the top cover and grabbing randomly samples from several heights of the composter. HC_2 was a 21-day aerobically stabilized home compost produced in Italy from a mixture of kitchen waste (60 % wb) and wood chips (40 % wb). OFMSWC was a compost prepared from the organic fraction of commingled municipal solid waste (OFMSW) and was obtained from a full scale mechanically and biologically treatment (MBT) facility at Reggio Calabria after 28 days of aeration. Approximately 10 kg of material was randomly sampled by the plant managers.
Substrates SIM_FW1 and SIM_FW2 were raw food wastes that were artificially prepared in both labs on-site. This was done in order to achieve an expected high degradability and to check the reproducibility of measurements between the labs with those simulated wastes. SIM_FW1 was prepared by mixing 25 % of cooked pasta, 25 % of white bread, 25 % of chopped apples and 25 % of grilled minced beef meat (on a wet weight basis, wb). SIM_FW2 was prepared by mixing 50 % of uncooked pasta, 30 % of uncooked frozen fries and 20 % of beef dog food (all in wb). Those ingredients were obtained from the same commercial chain that is located in Italy and Greece. The two aforementioned substrates (SIM_FW1 and SIM_FW2) were left for a week in a sealed plastic bag after preparation, and before the initiation of the experimental runs, so that to better simulate actual food wastes. FORW was a sample of leaves obtained from a forest floor during spring, by collecting all the available leaves from a 2 m 2 randomly selected area. FAMN was a fresh (1 week old) animal manure, collected from a cow breeding facility, that consisted of cow manure and straw. AMNC was obtained from the same facility, after passing through a solid-liquid separator and after stored in piles for 1 week.
For substrates HC_1, OFMSWC, FORW, FAMN and AMNC almost 4-10 kg (wet weight) were collected, while sampling was followed by a sequential quartering process in order to ensure randomness and homogenization. In the case of simulated substrates (SIM_FW1 and SIM_FW2) and HC_2, the whole amounts were used for the analysis.
HC_1, HC_2, OFMSWC and FORW were collected from Reggio Calabria (Italy), and then divided and individually sealed in two different plastic bags, containing approximately the same substrate quantity (about 4-10 kg). One of the two bags was placed in a non-insulated paper box, that was also thoroughly sealed and was shipped by airmail to Xanthi (DUTH, Greece). The other bag was kept in UNIRC at room temperature until the other part reached Greece, usually after 4-5 days. As soon as samples arrived in Greece, characterization started at the same time in both labs. FAMN and AMNC were collected from a cow breeding plant near Xanthi (Greece) and were shipped to Italy following the same abovementioned procedure. Substrates SIM_FW1 and SIM_FW2 were prepared concurrently in both laboratories, using the same raw materials and proportions, right before the initiation of the analyses.

Initial Characterization
The preliminary characterization of the substrates was performed simultaneously at both labs using similar standard techniques. Measurements of moisture (M), volatile solids (VS), pH, water holding capacity (WHC) were performed in both labs, whilst the elemental analysis (total carbon and total nitrogen contents) was performed in DUTH only. Moisture content was measured through weight difference at 75°C till constant weight, whereas volatile solids were measured in a muffle furnace through the loss on ignition of dried and ground material at 550°C for 2 h. Volatile solids, moisture and pH were measured according to [31,32]. Total carbon and total nitrogen were measured using an elemental analyzer (CE Instruments, CHNS-O Model EA-1110), according to [31,33].
Moisture, VS and pH measurements were done in 2 or up to 5 replications, whereas total C and N were done in six replications. The replications aided to check the statistical relationship between the initial parameters performed in UNIRC and DUTH using ANOVA principles.
The WHC was measured according to method TS 11184 [34] by placing 1000 g of sample in a previously weighed wet cotton bag, which was immersed in a water container for 24 h. The bag was removed from the water and allowed to drain for 6 h and then re-weighed. Table 2 shows which properties and respiration indices (discussed in next section) were measured in each laboratory.
In both labs, prior to the initiation of the dynamic respiration tests, all samples were optimized for moisture content by adding, if necessary, water to reach the 75 % of the corresponding WHC [34]. In addition, bulk density was adjusted, where needed, by adding a styrofoam material to maintain a value for the mixture below 450-500 kg/m 3 .

Dynamic Respiration Activity Tests
Dynamic Respiration Index at UNIRC (Italy) The dynamic respiration activity in UNIRC was measured using a 30 L adiabatic respirometric reactor (Costech International, Cernusco sulNaviglio, Italy) according to TS 11184 method [30,34]. The characteristic of the Costech Ò reactor is that it is adiabatic and that air flow entering the reactor is continuously adjusted so that to keep the O 2 concentration in the outlet stream between 14 and 17 % (v/v), as suggested by [4][5][6]25]. Advantages of the Costech Ò system include: (1) the large volume of the reactor that can help minimizing sampling errors, especially for heterogeneous materials characterized by a relatively large size of the particles, (2) the automation of oxygen and temperature measurements, and (3) the ability for continuous automatic adjustment of the air flow. The main disadvantages are related to the high investment cost and the maintenance of the apparatus. Moreover, there is a need of controlling the temperature of the inlet air, especially in case of low biological activity, since the results of the test are influenced by eventual significant changes in the temperature of the air supplied to the reactor. In terms of applicability, the main issue is the availability of the substrate, each time measured, because the operational quantity needed is-at least-3.0 kg.
The hourly oxygen uptake rate (OUR h ) was calculated according to [30]: where Q (L h -1 ) is the airflow rate, DO 2 (mL L -1 ) is the difference between the inlet and outlet O 2 concentrations of the reactor, V g (L mol -1 ) is the volume of 1 mol of gas at the inlet air temperature, 31.98 (g mol -1 ) is the molecular weight of O 2 and VS (kg) is the initial weight of volatile solids placed in the reactor.
Oxygen content measurements were performed every one hour. The DRI 24 was then calculated as the moving average of the 24 OUR h values, taken over the 24 h period that had the most intense biological activity. The amount of each substrate placed in the reactor ranged from 3.7 to 8.2 kg (wet weight) depending on bulk density (and Cumulative O 2 consumption after 7 days (SCRI 7 )

7
Respiratory quotient (RQ) 4 7 availability of materials), whilst the volume occupied in the reactor was always maintained similar for all runs (approximately 15 L). The temperature in CosTech Ò reactors was measured via a probe placed in the centre of the biomass. One run was performed for each substrate (n = 1), which is considered sufficient and representative due to the relatively large amount ([3.5 kg) of the sample placed in the reactor [8].

Dynamic Respiration Index at DUTH (Greece)
The dynamic respiration activity in DUTH was measured in 5.5 L dynamic reactors made from plexi-glass, as described in detail by [24]. The air flowrate was adjusted with a volumetric flowmeter (Aalborg, USA). Reactors were placed in a heated room at a constant temperature of 35 ± 2°C, but no temperature measurements within the reactor were performed. A wet sample of 500 g was placed in each reactor for each substrate. The unit air flow (UAF) rate for all eight substrates was maintained around 12 L air kg -1 VS h -1 ; thus, the actual air flow rate was adjusted accordingly from 15 to 60 mL min -1 . In any case, the UAF rate never exceeded 140 mL air dry kg -1 min -1 according to [35] suggestions. Measurements of the instantaneous O 2 and CO 2 contents at the reactors' outlet were performed manually, using a portable gas analyzer (GA-2000P, Geotechnical Instruments Ltd., U.K). The frequency of the measurement varied, from three to eight times per day, depending on the biological activity of each substrate. The most intense biological activity, characterized by high O 2 consumption and high CO 2 generation, was observed during the first 4 days from the initiation of the experiments. In this initial period, more frequent measurements were made to better formulate the respiration activity profile. During the last days of the experiments, the measurement frequency was decreased, since O 2 consumption and CO 2 generation profiles tended to stabilize. The same sampling frequency had been performed and justified in [24].
The DRI 24 was calculated similarly to above [4,9]. The maximum CO 2 generation rate (D-CO 2 _24) was calculated as the maximum of the averages of the instantaneous carbon dioxide generation rates recorded during the same 24-h period of higher activity. DRI measurements were done in triplicates for six substrates (OFMSWC, SIM_FW1, FORW, FAMN, AMNC and SIM_FW2) and in duplicates for substrates HC_1 and HC_2. All replicates were run concurrently.

Other Dynamic Respiration Activity Indices
Other dynamic respiration indices that were calculated on the basis of the same measurements done when calculating DRI 24 , were: cumulative dynamic O 2 consumption after 4 and 7 days (DCRI 4 and DCRI 7 , respectively); calculated as the integral below the OUR curve from day 0 to days 4 and 7, respectively. In DUTH, the maximum dynamic CO 2 generation rate (D-CO 2 _24) at 24 h and a dynamic cumulative CO 2 generation after 7 days (D-CO 2 _7) were also calculated.

Biologic Methane Potential (BMP)
BMP tests were performed in UNIRC in duplicates under mesophilic conditions. The amount of substrate used was between 1.3 and 22.8 wet g for each vessel. Tests were performed using a WTW OxiTop Ò AN6 (Germany) apparatus that consisted of three sets of 6 glass bottles (1.1 L volume) placed in thermostatic cabinet at 35 ± 0.5°C and equipped with a manometric head that recorded pressure increases due to biogas production. Each bottle was continuously mixed by a magnetic stirrer. Biogas produced passed through a NaOH solution (3 M), for CO 2 adsorption and methane production was then measured by an eudiometer (water displacement method). Anaerobic inocula were collected at three different periods from the second stage of an anaerobic digester that was working in mesophilic conditions and was fed with agrowastes (cattle manure, chicken manure, agriculture residuals and other industrial waste coming from the transformation of agriculture products such as olive and citrus). Immediately after sampling, inocula were sieved (\1 mm) to remove large fibrous materials (e.g. straw) and were kept under endogenous anaerobic conditions at 35°C (mesophilic inocula) for a minimum of 7-10 days to reduce blank biogas production. The BMP of the substrate was calculated (in NmL CH 4 g -1 VS) by subtracting the methane production of the inoculum (i.e. blank biogas production) from the gross methane production. The characteristics of the inocula ranged from 21.0 to 59.0 g TS L -1 with a VS content from 55 to 73 % (% TS). The substrate to inoculum ratio (in g VS/g VS) was always kept at around 0.4 to 0.5. The methane yield of the inocula ranged from 3 to 20 % of the methane yield of the main substrates, except in the case of HC_1 for which the methane yield was comparable to that of the inoculum.

Static Respiration Index (SRI)
Static respiration indices were measured only in DUTH using a 1 L static solid phase respirometer equipped with WTW OxiTop-C Ò manometric heads, following the methodology described in [10,23]. 50-70 g of wet substrate were placed in each static respirometer and moisture content was adjusted to an optimal moisture content range between 50 and 60 % ww. A 50 mL plastic beaker, used as an alkaline trap (containing a 2N KOH solution), was placed inside the respirometer in order to quantify the amount of C-CO 2 generated by the substrate. Respirometers were firmly closed and momentarily opened daily, so that the maximum pressure drop inside the vessels within two consecutive aerations would never exceed 170 hPa for any of the substrates. Based on the ideal gas law, this maximum pressure drop corresponded to a residual O 2 content in the vessel equal to 5 % v/v, which still sustains aerobic conditions. These maximum pressure drops were observed after 24 h during the first couple of days. All measurements were performed at 35°C and lasted 7 days in an incubator at the absence of light. Experiments were performed with 4-6 replicates per substrate.
The indices calculated for the static solid phase test in this study were: the maximum O 2 consumption rate recorded over 12 and 24 h periods of highest activity (SRI 12 , SRI 24 , in mg O 2 kg -1 VS h -1 ), the cumulative O 2 consumptions after 4 and 7 days (SCRI 4 , SCRI 7 , in g O 2 kg -1 VS h -1 ) and the CO 2 generation after 7 days (S-CO 2 , in g C-CO 2 kg -1 VS) as explained in [10].

Substrates' Characterization
The initial properties of the substrates are included in Table 3, where it is shown that there was a very good agreement between the initial properties (moisture content, volatile solids and pH) of all substrates from both labs; this ensured that respirometric experiments on all materials started with identical initial properties in both countries. WHCs for half of the studied substrates were almost the same (HC_1, FORW, AMNC, SIM_FW2), while for the rest seemed to slightly differ. The as-received moisture contents of the eight substrates ranged from 18 % (ww) for the OFMSWC, to 83 % (ww) for the FAMN. Volatile solids (i.e. organic matter) ranged from a low 23 % (dw) for the OFMSWC, to 96-98 % (dw) for the simulated food waste substrates (SIM_FW1 and SIM_FW2). C/N ratios, as measured through elemental analysis, ranged from 8.2 to 40, for FAMN and HC_2, respectively. pH values ranged from 5.1 (for SIM_FW2, which had the higher organic matter content) to 9.5 for HC_1 (composted kitchen waste and wood ashes with a VS at 42 % dw). Most of the substrates had a pH between 6 and 8, while manures (FAMN and AMNC) were alkaline. Table 4 presents the dynamic respiration indices and the BMP results. The profiles of the oxygen uptake rates of the eight substrates, as measured by both laboratories, are included in Fig. 1. Note that the mean values are shown only in the case of the Greek tests (for which there was replication),while the errors bars in the graphs show the standard errors. Apparently, a very good replication (coefficient of variation was always less than 5 %) existed for the Greek dynamic respiration tests.

Oxygen Uptake Profiles in Dynamic Methods
According to Fig. 1, there were quite similar OUR profiles between the ITA and GRE dynamic tests for substrates HC_1, HC_2, SIM_FW1, FAMN and SIM_FW2. The largest deviation among the OURs recorded between ww wet weight basis, dw dry weight basis, WHC water holding capacity the two labs was observed for OFMSWC; this can be also confirmed by looking at the corresponding dynamic indices of Table 4. For this substrate, the Greek index was unexpectedly high, whilst the Italian DRI 24 was close to values suggested in the past by [36], who had used similar materials and similar equipment. This contradiction with regard to OFMSWC is hard to explain and could be attributed to an inherent variability and unpredictability of biological experiments, or to a heterogeneity of the sample which was derived from MSW. It might be also attributed to the fact that the temperature of OFMSWC in CosTech Ò remained similar to that of the ambient air (around 20°C) throughout the run, whilst artificial mesophilic temperatures (&35°C) were always maintained in the Greek method. For FORW, a higher early peak was observed in the Greek method, although a rather stable and slightly fluctuating OUR was observed in the Italian method with a peak at the end of the run. This periodic fluctuating OUR profile in the case of the CosTech Ò runs was also obvious for substrates HC_1, FORW, FAMN, AMNC, SIM_FW2 and might be attributed to diurnal temperature variations during the experiment, since the air that was introduced into the reactor was at ambient room temperature. A similar disagreement seemed to exist in the case of FORW, in which the GRE method resulted in higher OURs compared to the ITA method. In AMNC too, the Greek method resulted in higher OURs compared to the Italian method; yet, this substrate was a rather stable one with some of the lowest OURs recorded in both methods. According to Fig. 1, the OUR profiles in the Greek system were much smoother than the CosTech Ò OUR profiles, despite the fact that data recording was less infrequent in the former case.
In the CosTech Ò system, only 2 substrates (SIM_FW1 and SIM_FW2) reached temperatures up to 45°C after 1.5 days which were then reduced and maintained to between 30 and 40°C till the end of the process. This agrees with the fact that SIM_FW1 and SIM_FW2 had two of the highest respiration activities. On the other hand, FAMN, which also had a high respiration activity in Italy, never reached such mesophilic temperatures. This was caused due to a malfunction of the automatic airflow adjusting system after day 8. However, the DRI 24 values were still calculated during the first 7 days, in which the maximum biological activity was anyway observed.   According to Fig. 2, only SIM_FW1 and SIM_FW2 reached temperatures up to 45°C, yet none reached thermophilic temperatures. Both these substrates had two of the highest respiration activities, as measured in both Italy and Greece. Yet, FAMN, which also had a high respiration activity as measured in Italy, had, surprisingly, one of the lowest temperatures among all substrates, which is difficult to explain. The internal temperatures of all other substrates (i.e. except SIM_FW1 and SIM_FW2) remained at ambient values between 20 and 25°C.

Effect of Temperature and Air Flow on Respiration Activity
As shown in Fig. 1, maximum deviations for OUR profiles are observed for OFMSWC and AMNC, which can also be confirmed through their dynamic indices (Table 4). This can be attributed to the fact that in Italian reactor, these two substrates did not exceed 25°C. Adversely, Greek dynamic reactors were always in a temperature range between 34 and 37°C. The lack of high temperatures in the other 5 substrates might be charged to the low respiration activity of those substrates (Table 4). Moreover, difference in reactor scale is important since it might affect the temperature profile of the material.
When designing solid waste treatment plants, it is also necessary to know the amount of air required to maintain optimum conditions for the degradation of organic substrates. For this reason, Fig. 3 was drawn in order to study the effect of the cumulative (total) air flow that passed through the CosTech Ò system after 7 days, on the respiration activity. In Fig. 3, only data from the Italian reactor are shown, since the flowrate in that system was continuously adjusted throughout the process. On the contrary, in the Greek dynamic method, a constant UAF rate was applied to all reactors throughout the experiments.
According to Fig. 3, it appears that as the total flow of air increases, the respiration activity increases too; yet it seems that a constant DRI 24 value tends to be reached at high cumulative flows. This trend was also shown by [37] and can be explained by the fact that beyond a certain cumulative aeration threshold value, the respiration activity is maximized and stabilizes and no more excess air is necessary. Moreover, an excessive aeration could decrease microbial activity due to the reduction of the temperature of the biomass at sub-optimal levels. Table 5 includes the static respiration indices and the ratios of the dynamic respiration (measured in the Greek lab) over the static respiration indices. As shown in Table 5, the dynamic respiration indices were in general much higher than the corresponding static indices (both in terms of rate and total oxygen consumption). No consistency was also observed between the dynamic and the static test indices in the recent work of Aspray et al. [20]. After categorizing the 8 substrates in fresh (SIM_FW1, SIM_FW2, FAMN) and stabilized composts (HC_1, HC_2, OFMSWC, FORW, AMNC), it was observed that the ratio of DRI 24 /SRI 24 (measured in Greece) for the fresh substrates ranged from 3.3 to 11.4 as opposed to the stable substrates for which that ratio ranged from only 0.9 to 3.3. Similarly, the ratio of DCRI 7 /SCRI 7 ranged from 18 to 40, for the fresh substrates and from 1.2 to 13, for the stable substrates. The above indicates that the static methods underestimate actual oxygen consumption, as had been also suggested by [4]. This variability in the ratios of DRI to SRI contradicts the findings of [4], who showed that the dynamic respiration rate indices (DRI 24 ) of severalmunicipal solid waste derived organic substrates (fresh, composted and stable) were, on average, 2 times higher than the corresponding static respiration indices (SRI 24 ). It is noted, however, that  [4] had measured oxygen consumption in their static methods with an oxygen probe at 20°C. Comparison with dynamic and static test were also conducted in the work of Godley et al. [28], who had also reported that dynamic respiration rates were at least double than the static ones. Similar results had been found by Adani et al. [6] in which all dynamic respiration rates were far higher compared to the static ones, based on a sample size of 18 organic substrates.

Correlations
Potential correlations among the initial physicochemical properties measured (OM content, pH, total carbon and total nitrogen) and the main biodegradation indices were calculated. The only positive correlation existed between total nitrogen content and the DRI 24 and CH 4 production after 30 days (in both cases Spearman's q = 0.8), as these were measured by the Italian lab. With regard to the aerobic index, this correlation can be explained by the nitrification processes that consume oxygen. On the other hand, neither total carbon nor organic matter contents correlated significantly with any of the main biodegradability indices, as also revealed in Fig. 4. According to Fig. 4, a clear outlier exists (OFMSWC), whilst it is observed that the substrates with VS contents above 85 % (db) achieved a very wide range of respiration activities. As opposed to OFMSWC, FORW had a rather high VS content (around 80 % dw) which would be expected to result to a relatively high respiration activity. However, this was not true, and the dynamic respiration indices (from both labs) were relatively low for that substrate (i.e. \500 mg O 2 kg -1 VS h -1 ). This can be attributed to the likely low content of readily biodegradable substrate present in the leaves. Their high volatile solids content is most probably attributable to the high content of lignin, which is recalcitrant to fast biodegradation in aerobic conditions. From the above it appears that no clear relationship exists between the VS content of the materials and the respiration activity. Similar  observations have been found out by Barrena et al. [8], in which the organic matter content did not correlate sufficiently with any of the dynamic or static respiration indices. On the contrary, Aspray et al. [20] found that the VS content correlated positively with two different dynamic methods, while no correlation was evident with the static test indices. Positive correlation between organic matter content and both O 2 and CO 2 static indices have been reported in [10,23]. Apparently, specific constituents of the organic matter, such as sugar content (i.e. cellulose, hemicelluloses, lignin, starch), as well as the level of microbial inoculation, determine the biodegradation extents (and thus the respiration activity) regardless of total VS content. Figure 5 graphically depicts the most important correlations among the biodegradation indices measured in this work. Table 6 includes all Spearman's rank order correlation coefficients among the studied indices (only the statistically significant values at p \ 0.05 are shown). Note that the Spearman coefficients calculate any type of correlation and not necessarily a linear one (as in the case of the Pearson coefficients).
According to Table 6, the DRI 24 indices from both labs correlated between them adequately (q = 0.74) with a marginally statistically significant correlation at p \ 0.05. According to Fig. 5a, OFMSWC appears to be an extreme value. Similarly to DRI 24 , the cumulative oxygen consumptions (DCRI 4 , DCRI 7 ) also correlated significantly in both labs (coefficients ranged from 0.83 to 0.95), verifying the visible positive correlation trend illustrated in Fig. 5b. According to Table 6, the cumulative dynamic respiration indices (DCRI 4 , DCRI 7 ) also correlated well with the corresponding DRI 24 in both labs. This was also shown in [24] and practically indicates that the higher the maximum O 2 consumption rate, the higher the total oxygen consumption too.  Table 6 also reveals a very strong correlation between the methane generation after 30 days and the dynamic respiration indices of both labs. This is also evident in Fig. 5c, d that both illustrate that positive correlation between both Greek and Italian dynamic respiration indices (DRI 24 ) and the 30-d methane yield. This is an important finding, because it indicates that the degradation potential of an organic substrate could be similar under both aerobic and anaerobic environments, as has been shown by other researchers too [8,15,38]. However the values of the dynamic respiration indices measured, respectively, in the Italian laboratory, for OFMSWC, and in the Greek laboratory, for SIMFW_2, worsened significantly this correlation.
According to Fig. 5e, f, the correlation among the static and dynamic respiration indices (DRI 24 and SRI 24 ) was poor as was discussed in ''Static and Dynamic Respiration Indices'' section too.
The correlation between the total CO 2 generation and the total oxygen consumption under dynamic conditions measured in the Greek lab was strong (correlation coefficients above 0.98), as shown in Table 6. This indicates that the moles of generated CO 2 were steadily proportional to the moles of O 2 consumed for all substrates. This was also noticed for the peak rates of CO 2 generation (D-CO 2 _24) and the peak rates of O 2 consumption (DRI 24 ) indicating that CO 2 generation rate and OUR followed a parallel profile (in the dynamic tests in Greece). Moreover, respiratory quotients (RQ: moles of CO 2 generated over the moles of O 2 consumed) calculated for the dynamic experiments conducted in Greece, ranged between 0.7 and 1.0 (Table 4), which are reasonable values and comparable to [24].
Statistically significant correlation was also observed between the static CO 2 generation and the static O 2 consumption values obtained in the Greek lab (q = 0.91). Therefore, CO 2 generation could be also used as a biodegradation activity index, although a caution is required, since static indices are affected by temperature and CO 2 evolution is pH-dependent due to its solubility in aqueous solutions [12]. In the case of static method RQs ranged from 1.3 to 2.3, with an average value of 1.7 (Table 5). High RQ values-using the same static method-had been also reported by Komilis et al. [10] indicating variations in the oxidation stage and the degradability of the carbon contained in each substrate.

Use of Biodegradation Activity Indices for Compliance Purposes
According to the findings of this research, the use of both aerobic dynamic respiration and anaerobic methods is an important tool to evaluate the biological activity of fresh  All Spearman's rank-order correlation coefficients shown are significant at p \ 0.05 n/s: non-significant, [G] Greek method, [I] Italian method and composted biodegradable wastes/residues. With regard to fresh substrates, this is important knowledge for the selection of the appropriate treatment technologies. For practical purposes, and by accounting for the good correlation of the aerobic and anaerobic indices, the use of the former indices would be preferred (shorter duration than the anaerobic ones, no need of seed). Moreover, the contemporary use of aerobic and anaerobic methods allows a cross check and the detection of errors or of measurement problems (e.g. inhibition) that are frequent with this type of complicated biological measurements. This is very important when for practical reasons, as in the Di.Pro.Ve-CosTech Ò system, replicates are difficult to perform. Therefore, the biological stability index presented in this paper, and especially the DRI 24 and BMP, could be adopted for process control at the level of single plant. In this case, when the indices are used only for internal purposes, the method and equipment to be adopted for the measurements could be chosen based on the cost of the equipment and on the knowledge of the plant's technical personnel.
However, the use of stability indices for compliance purposes (e.g. evaluation of compost quality, acceptance of waste in landfill) is more complicated. In fact, compliance threshold values must be referred to a specific method as in norms in Germany and Italy [26,27] and should be set considering the uncertainty connected to this type of measurements that is higher when replicates are not provided. In this case, also, the adoption of a double compliance (both to aerobic and anaerobic threshold values) could help to ensure the reliability of the evaluation and the detection of errors/measurement problems as done in Germany and Austria. However, practical problems linked to the long duration of the tests (at least 7 days for DRI and 30 days for BMP) should be solved prior to a generalized adoption of compliance tests based on biological activity methods.
A reasonable possibility would be to measure the stability indices (both dynamic aerobic and anaerobic) with a frequency related to the amount of biomass treated in the plant; the evaluation of the compliance should then be done on a statistical basis (e.g. 95 % of the samples analysed yearly should comply with the threshold values). If the frequency of samples with respiration indices above the threshold imposed by the norms exceeds a pre-set value, some form of penalty should be provided.
In the case of using indices for compliance purposes, the adoption of standard and unified methods among countries, at least at a European Level, would be greatly beneficial.

Conclusions
Results of the work indicate that although an agreement existed among the different methods for some of the substrates, differences did exist for others so that there is no unique ideal method to propose. In addition, the OUR profile was much smoother in the case of the Greek dynamic reactor compared to the CosTech Ò system, despite the smaller recording interval in the former case. A statistically significant correlation was calculated between the dynamic respiration indices in both labs, with regard to both rates and cumulative values. However, a weak agreement was observed for one of the substrates. A strong correlation was found between the methane yield after 30 days and the 24-h based dynamic respiration rate and the cumulative dynamic respiration indices, at 4 and 7 days. The ratio of the dynamic to static respiration rates ranged from 1 to 11 and was higher in fresh materials compared to the more stable ones. The dynamic respiratory quotients were very close to 1.0, whilst the static ones had an average of 1.7. Finally, it appears that the volatile solids content should not be used as a predictor of biodegradation activity.