Assessing the Suitability of Digestate and Compost as Organic Fertilizers: A Comparison of Different Biological Stability Indices for Sustainable Development in Agriculture

Abstract

Nowadays, biowaste valorization is a key point in the circular economy. Digestate and compost from organic waste treatment can be used as nutrient-rich fertilizers. In Europe, the use of biowaste-derived fertilizers is promoted by the European Fertilizer Regulation (EU) 2019/1009, which requires verification of their biological stability through regulated indices; however, it is not clear whether the proposed indices and threshold values indicate the same level of stability and what correlations there are between them. This study compared four biological stability indices, namely Oxygen Uptake Rate (OUR), Self-Heating (SH), Residual Biogas Potential (RBP), and Dynamic Respirometric Index (DRI), which were tested on 50 samples of compost and digestate. Overall, the results revealed that most of the compost and digestate samples were quite far from European standards. On the contrary, the RBP test seemed to be less stringent than the other indices, since a much larger number of samples was closer to or in compliance with the established threshold. Data analysis using Pearson’s coefficients showed a strong linear correlation between the indices. Nevertheless, the linear regression predictive model based on experimental data demonstrated that the indices could not represent the same level of stability, providing poor consistency and variability in the predicted values and established threshold. In particular, the DRI test appeared to be more severe than the other aerobic indices. This work could provide valuable support in improving evaluation criteria and promoting a sustainable use of compost and digestate as organic fertilizers from a circular economy perspective.

1. Introduction

Nowadays, biowaste management is a crucial point in the political strategies of sustainable development and circular economy [1,2]. Annually, biowaste production in the European Union ranges from 118 to 138 million tons; among these, around 88 million tons come from municipal waste and 30–50 million tons from industrial activities, such as food processing facilities. Biowaste constitutes 30–40% of municipal solid waste (MSW), although it can vary from 18% to 60% [3]. In this context, digestate and compost can be produced from the treatment of organic waste through the anaerobic digestion and composting processes, respectively, representing valuable organic fertilizers rich in nutrients [4,5,6,7].

The use of digestate and compost from biowaste is regulated by the European Fertilizer Regulation (EU) 2019/1009 [8], which aligns with the broader goals of the EU Action Plan for the Circular Economy and the Bioeconomy Strategy, emphasizing the importance of sustainable use of resources by minimizing waste and environmental impact [9]. According to the above-mentioned regulation, it is necessary to verify the biological stability of compost and digestate before their application in agriculture by appropriate indices, which should certify their suitability in terms of human and environmental safety [10,11], while providing information to plant operators about process performance. Indeed, biological stability is defined as the degree of decomposition of biodegradable organic matter and reflects the material’s potential for further microbial activity. Stable compost or digestate minimizes odor generation, phytotoxicity, pathogen regrowth, and greenhouse gas emissions during its storage and application [9,12,13,14]. In addition, it is worth noting that biological stability indices are mentioned in end-of-waste (EoW) [3] and for EU Ecolabel guidelines [15], which identify the process or set of processes that cause a waste to cease being a waste and a multi-criteria label certifying excellent environmental performances for a product throughout its life-cycle, respectively.

Recent analyses of compost and digestate quality standards have also highlighted that European countries adopted heterogeneous stability methods and threshold values, leading to non-equivalent interpretations of biological stability [3,8,16,17,18]; this variability further complicates the harmonization of criteria under the EU Fertilizer Regulation. Prasad and Foster (2023) showed that OUR, RBP, and self-heating tests often lead to divergent classifications of stability, emphasizing the need for clearer and more consistent evaluation frameworks across Europe [17]. Anyway, it is not clear whether the proposed indices and threshold values indicate the same level of stability and what correlations there are between them. Therefore, this study compared four biological stability indices, namely Oxygen Uptake Rate (OUR), Self-Heating (SH), Residual Biogas Potential (RBP), and Dynamic Respirometric Index (DRI), which were tested on 50 samples of compost and digestate collected from different organic waste treatment plants in Italy, according to the EU Fertilizer Regulation [12,19]. Digestate and compost are classified by the EU Fertilizer Regulation as Component Material Categories (CMCs): CMC3 refers to compost, while CMC5 refers to digestate (other than fresh crop digestate). Each category must satisfy at least one stability criterion, based on standardized methods. For compost (CMC3), either the OUR or SH test is required; on the other hand, either the OUR or RBP test may be applied for digestate (CMC5) [8].

Specifically, OUR measures the rate of oxygen consumption by microorganisms in the compost or digestate samples, indicating their metabolic activity and biological stability [20]. The SH or Dewar test evaluates the propensity of compost or digestate samples to undergo spontaneous heating due to microbial activity, providing insights into their biological activity and maturity [10]. RBP assesses the remaining biogas production potential of compost or digestate samples, indicating the extent of organic matter degradation and biological stability achieved during the composting or anaerobic digestion process [21,22]. DRI is a comprehensive index that integrates various parameters such as oxygen consumption rate, temperature, and moisture content to assess the overall biological stability and maturity of compost and digestate samples [23]. It should be noted that DRI is only used at the national level in Italy (according to the Legislative Decree No. 75 of 29 April 2010) [23,24,25]; however, it was considered as a further parameter for comparison with European standards.

To our knowledge, no studies have compared the biological stability indices based on the European Fertilizer Regulation (EU) 2019/1009. In the literature, many authors have dealt with this topic, testing several physical, chemical, and biological indices; however, they preceded and/or do not consider all the indicators and thresholds established by the legislation as the basis for their work. The DRI was evaluated as an indicator for organic waste and compost in [23,25,26]. Biological and thermoanalytical indices were applied to assess the stability of 15 compost end-products in [27]. A massive estimate of the indices suitable for measuring the organic substance content in solid waste has been carried out by [28]. Ref. [29] compared several stability indices for mechanical/biological waste treatment and composting plants. According to the authors, each of the above studies found good correlations between the tested indices.

This work could provide valuable support for the European governments in improving evaluation criteria and promoting compost and digestate use as organic fertilizers in agriculture from a circular economy perspective.

2. Materials and Methods

2.1. Experiment Description and Sample Characterization

OUR, SH, RBP, and DRI stability indices were tested on 50 samples (25 of digestate and 25 of compost), which were collected from different Italian organic waste treatment plants. The waste treatment plants considered are medium-to-large, located in the same region (Tuscany, central Italy), and serve urban areas with similar social, environmental, and economic characteristics. This choice was made to ensure organic waste was as comparable as possible. Compost and digestate samples were sampled at the end of the treatment chain, prior to any potential agricultural applications.

The chemical and physical parameters of compost and digestate samples were analyzed according to standard procedures provided by the American Public Health Association (APHA).

In the following sections, the procedure and instrumentation used for the calculation of the biological stability indices are described. The results obtained on digestate and compost samples were compared with standard limits; in particular, the European Fertilizer Regulation (EU) 2019/1009 admits a maximum value of 25 mmolO₂/kg_VS/h and 0.25 Nl/g_VS for OUR and RBP, respectively, while SH must reach at least a level of III, defined on the basis of the Rottegrad classification (

Table 1. Dewar self-heating increments, rating, and description of stability classification based on the European system.

Temperature Rise Above Ambient in C	Official Class of Stability	Descriptors of Class or Group	Major Group
0–10°	V	Very stable, well-aged compost	Finished compost
10–20°	IV	Moderately stable; curing compost
20–30°	III	Material still decomposing; active compost	Active compost
30–40°	II	Immature, young, or very active compost
40–50° (or more)	I	Fresh, raw compost, just-mixed ingredients	Fresh compost

). On the other hand, a maximum value of 1000 mgO₂/kg_VS/h is allowed by the Italian Regulation for DRI.

2.1.1. Self-Heating (SH)

The SH test serves as an indirect assessment of the aerobic biological activity for organic materials that are smaller than 10 millimeters in size. This technique relies on measuring the peak temperature attained by the biomass contained within a vessel under standardized conditions over several days, in accordance with [30].

The experimental apparatus comprises an adiabatic vessel with a volume of 2 liters, two thermocouples, and a data acquisition system (refer to Figure 1a). Approximately 1.0–1.5 kg of organic material was introduced into the vessel following sieving (utilizing a 10-millimeter sieve), and the ideal moisture content (approximately 35% w/w) was established in accordance with the Technical Regulation UNI/EN 16087-2:2011 [31]. Two T-type thermocouples consistently monitored the temperatures of the room and biomass. One thermocouple was situated in the operational area, whereas the other was positioned at the midpoint of the vessel. All data from the thermocouples were collected and analyzed using a cRIO 9030 data acquisition system (National Instruments, Austin, TX, USA). During the experimental period, the temperature rose as a result of the exothermic nature of the degradation process. The experimental tests were halted when the biomass temperature consistently declined for two consecutive days after reaching its peak value. Following this protocol, the duration of the experiment could range from 5 to 10 days. The decomposition rate was assessed by calculating the difference between the highest biomass temperature and the surrounding temperature. The biological stability level was evaluated based on the Rottegrad classification (Table 1). SH was measured in duplicate for each sample [30].

2.1.2. Oxygen Uptake Rate (OUR) Test

Aerobic stability tests for OUR determination were carried out by using Oxitop-IDS instrumentation, developed by WTW GmbH (Figure 1b). Oxitop instrumentation allows us to calculate the OUR by determining the oxygen demand for the biological oxidation (BOD-Biological Oxygen Demand) of the organic material. BOD calculation is based on pressure measurements (differential measurement), using a piezo-resistive electronic pressure sensor (Hagman and Box, n.a.), as described by the standardized methodology UNI EN 16087-1: 2020 [32].

The samples were sifted with a 10 mm sieve to evaluate the amount of residual material; samples showing a residual material amount greater than 20% were considered unsuitable for testing. Suitable samples were instead tested by filling a glass reactor (vessel) based on the TS and VS content; specifically, the following volumes were added:

• 180 mL of demineralized water;
• 10 mL of a complete nutrient solution;
• 10 mL of buffer solution;
• 5 mL of ATU solution.

Open vessels were stirred (250 rpm) and kept in contact with the air from 4 to 8 h. At the end of this time, vessels were closed and sealed through the OxiTop head and the special hooks. Tests were then carried out in duplicate for a total duration of 7 days [33]. The calculation of oxygen consumption, evaluated in mmol O₂ per kg of organic material (%TVS/TS), was determined by the decrease in pressure at the reactor’s headspace, according to the following formula (Equation (1)):

$$\mathrm{O}\mathrm{C}={\displaystyle \frac{\mathsf{\Delta }\mathrm{P}}{\mathrm{R}(273.15+\mathrm{T})}}·{\displaystyle \frac{{\mathrm{V}}_{\mathrm{g}\mathrm{a}\mathrm{s}}·1000}{\mathrm{W}·\mathrm{D}\mathrm{M}·\mathrm{O}\mathrm{M}}}$$

(1)

where:

• OC is the oxygen consumption [mmolO₂ per kg of organic matter].
• ΔP is the change in pressure in the reactor’s headspace [kPa].
• R is the gas constant [8.314 L ∙ kPa K⁻¹ mol⁻¹].
• T is the test temperature [°C].
• W is the initial mass of the sample [kg].
• DM is the dry matter content [% by weight].
• OM is the organic matter content [%TVS/TS].
• Vgas is the volume of the gas phase in the reactor [ml].

2.1.3. Dynamic Respirometric Index (DRI) Test

DRI analysis was carried out using stainless-steel adiabatic reactors with a capacity of 30 L each, according to the UNI/TS 11184: 2016 [34] standard. The experimental apparatus consists of two continuous-flow aerobic respirometers filled with approximately 20 kg of sample (Figure 1c). Before starting the analysis, sample parameters such as humidity, density, and pH were standardized according to the guidelines of reference [24].

To establish ideal conditions, the sample is subjected to continuous blowing of dry air from a compressor. Airflow was measured using two flowmeters (Aalborg Instruments, Orangeburg, NY, USA), which were positioned at the inlet and outlet of the system. During the analysis, the inlet airflow was regulated by the flowmeter to maintain an oxygen concentration in the exhaust air above 14%.

Special probes were installed at the outlet of the system to measure oxygen (Zirconium oxide sensor, TEC-ZRC, Tecnosens SpA, Brescia, Italy) and CO₂ concentrations (GasCard NG Gascheck 10%, Edinburgh Sensors, Livingston, UK) in the exhaust airflow. The reactors are externally coated with a water jacket heated by a thermostat (FA90, FALC instruments, Treviglio, Italy) to ensure that the adiabatic conditions of the system are maintained. The temperature of the incoming flow, the outgoing flow, the environment, and the biomass subjected to the process were monitored through the installation of thermocouples. The sensor signal installed in respirometric reactors was managed and acquired by a National Instrument acquisition system and developed by software (LabView environment, 2024 Q3 version).

The DRI value was measured in accordance with [23], representing the average hourly index and measured during the 24 h when biomass breathing is highest.

2.1.4. Residual Biogas Potential (RBP) Test

The RBP tests can be assessed through anaerobic assays, and the production of biomethane potential (BMP) can be evaluated through the composition of the generated biogas. The experiment was conducted in triplicate, and the production of biogas was measured after twenty-eight days (RBP 28 and BMP 28) for each reactor. The experiments were performed using stainless-steel batch reactors of 1 L (Figure 1d) [12]. These reactors were kept at a stable temperature of 37 + 0.1 °C within a water bath regulated by a thermostat (FA90, FALC Instruments, Treviglio, Italy). Reactors were loaded with the organic samples and combined with an inoculum (retrieved from an anaerobic digestion plant that processes OFMSW and cattle manure), maintaining a VS mass ratio of 1.5:1.

The reactors were secured with a ball valve cap to enable biogas sampling. Previously, each reactor was purged with inert gas (N₂) to guarantee anaerobic conditions [35]. Biogas yield was assessed by measuring the pressure in the reactor’s headspace with a membrane pressure gauge (Model HD2304.0, Delta Ohm S.r.l., Selvazzano Dentro, Italy).

To assess the methane concentration in BMP28, a gas chromatograph (INFI-CON, Bad Ragaz, Switzerland) was utilized [31]. The experiment adhered to the protocols outlined in UNI/TS 11703:2018 [36].

2.2. Predictive Model and Analytical Relationships Between the Analyzed Indices

Based on experimental data obtained from biological stability indices of compost and digestate samples, Pearson’s correlation coefficient (r) was used to evaluate the strengths and directions of the linear relationship of any pair of indices. A value of “+1” means a perfect positive correlation, “−1” means a perfect negative correlation, and 0 means no linear correlation [37,38].

The correlation coefficient (hereinafter referred to as X and Y) was calculated as follows [39]:

• Standardization of the variables X (“zX”) and Y (“zY”), which are re-expressed to have means equal to 0 and standard deviation (s.d.) equal to 1 (Equations (2) and (3)).

zX_i = [X_i − mean (X)]/s.d. (X)

(2)

zY_i = [Y_i − mean (Y)]/s.d. (Y)

(3)

• Calculation of the correlation coefficient as the mean product of the paired standardized scores, considering sample size n (Equation (4)).

r_x,y = sum of [zX_i · zX_i]/(n − 1)

(4)

In addition, to compare the stability levels obtained by different indices, predictive models were developed for three datasets (compost, digestate, and all combined samples) using a second-order polynomial regression.

Based on experimental data, two group models were developed:

• Relationship between RBP, OUR, and SH, representing a comparison between the biological stability indices permitted by European Fertilizer Regulation (EU) 2019/1009.

The considered key variables were OUR (mmolO₂/kg_VS/h), SH (°C), and RBP (Nl/g_VS). A second-order polynomial regression was implemented according to the following form (Equation (5)):

RBP = β₀ + β₁ · OUR + β₂ · SH + β₃ · OUR² + β₄ · SH² + β₅ · (OUR · SH)

(5)

2.

DRI as a function of OUR and SH, representing a comparison between biological stability indices conducted under aerobic conditions.

DRI (mgO₂/kg_VS/h) was the dependent variable, while OUR (mmolO₂/kg_VS/h) and SH (°C) were the predictors, fitting the quadratic model as follows (Equation (6)):

DRI = β₀ + β₁ · OUR + β₂ · SH + β₃ · OUR² + β₄ · SH² + β₅ · (OUR · SH)

(6)

Models were fitted using the least squares method implemented in Python’s (3.12 version) scikit-learn library. The fit quality was assessed by the coefficient of determination (R²) and Root Mean Square Error (RMSE).

It should be noted that this is a preliminary attempt to correlate and develop predictive models of biological stability indices, as methodologies for model validation are lacking. This would require a larger amount of data and additional analyses, which were considered out of scope due to sample limitations. The objective of this study was to preliminarily verify potential discrepancies between the indices, which could be further investigated in larger future studies.

3. Results and Discussion

3.1. General Insights from the Analysis of Biological STABILITY Indices

Biological stability indices revealed significant variations in both compost and digestate samples, confirming strong heterogeneity in the collected samples. Figure 2 and Figure 3 show the results obtained from the application of biological stability indices on compost and digestate samples, highlighting a wide range of values in all the indices analyzed.

Indeed, collected samples exhibited heterogeneous characteristics.

Table 2. Minimum (Min), maximum (Max), and average (Av.) values of the main physical and chemical parameters measured for digestate and compost.

	TS (%)			VS (%)			pH			C/N
	Min	Max	Av.	Min	Max	Av.	Min	Max	Av.	Min	Max	Av.
Digestate	1.44	88.9	35.5	47.0	91.0	66.0	4.5	8.7	7.3	6.0	21.5	11.4
Compost	9.6	92.7	70.8	29.0	87.1	49.3	4.0	9.0	7.4	12.0	15.9	13.7

provides a comprehensive overview of the main physical and chemical parameters for digestate and compost, namely total solids (TS), volatile solids (VS), pH, and carbon-to-nitrogen (C/N) ratio, showing a wide variability in the measured values.

3.1.1. Biological Stability Indices on Compost Samples

As shown in Figure 2a, the OUR test found a maximum value of 275 mmolO₂/kg_VS/h (11 times higher than the regulatory limit), and the average value of the analyzed samples was quite far from the threshold value of 25 mmolO₂/kg_VS/h.

Regarding the SH test (Figure 2b), some samples obtained the level indicating the highest degree of stability according to the Rottegrad scale (level V, 0–10 °C, Table 1); however, most of the samples did not comply with the minimum regulatory standard (level III, 20–30°), since the average value of the samples was on the lowest grade of the Rottegrad scale, which indicates a compost that is still fresh (level I, 40–50 °C).

Concerning the RBP test (Figure 2c), a maximum value more than three times higher than the standard limit (0.25 Nl/g_VS) was found; however, the average value obtained (0.28 Nl/g_VS) is slightly higher than the threshold established by the European Fertilizer Regulation (EU) 2019/1009, indicating that most of the samples have complied with the regulatory threshold.

Finally, the DRI test (Figure 2d) reported values up to six times higher than the standard limit (1000 mgO₂/kg_VS/h); also, the average value of the samples (2831.5 mgO₂/kg_VS/h) is well above the maximum established level.

In this case, the results obtained showed that the OUR, SH, and DRI indices were more in agreement with each other, showing that most of the compost samples were quite far from the stability requirements of the European Fertilizer Regulation (EU) 2019/1009. On the contrary, the RBP test seemed to be less stringent than the other indices, since the average value of the samples was found to be not far from the established threshold.

3.1.2. Biological Stability Indices on Digestate Samples

Figure 3a shows the results of the OUR test on digestate samples, marking maximum values about five times higher than the standard limit (25 mmolO₂/kg_VS/h), and the average value (57 mmolO₂/kg_VS/h) was much higher than the required threshold.

The SH test (Figure 3b) showed that a few digestate samples reached Level I (40–50 °C, Table 1) of the Rottegrad classification, indicating the lowest degree of biological stability. However, most samples fell within Level III (20–30 °C), which represents the minimum stability level required by the European Fertilizer Regulation. This is consistent with the average temperature rise (24 °C), suggesting that digestate generally exhibits moderate biological activity and acceptable stability under aerobic conditions.

Regarding the RBP test (Figure 3c), the maximum value detected was only slightly above the regulatory limit (approximately 1.4 times higher). More importantly, the average RBP value (0.17 Nl/g_VS) was well below the threshold set by the European Fertilizer Regulation (0.25 Nl/g_VS), indicating that most digestate samples displayed limited residual anaerobic biodegradability and complied with the normative requirement.

Conversely, the DRI test (Figure 3d) showed maximum values up to 4.4 times higher than the limit of 1000 mgO₂/kg_VS/h. Although the mean DRI value (1252 mgO₂/kg_VS/h) was closer to the threshold than the maximum value, it still exceeded the regulatory limit, highlighting residual aerobic biodegradation potential in several samples. Compared with compost, digestate exhibited lower maximum values across all indices and average values generally closer to their respective thresholds. SH and RBP confirmed good compliance with regulatory criteria, whereas OUR and DRI frequently exceeded the allowed limits, reinforcing the inconsistency between indices and the difficulty of deriving a uniform stability assessment across different methodologies.

3.2. Pearson’s Correlation Coefficient r

Figure 4 shows a Pearson correlation heatmap. The Pearson coefficient (r) was calculated to identify potential linear relationships between the biological stability indices (i.e., OUR, SH, RBP, DRI) and the main chemical-physical parameters (i.e., TS, VS, pH, C/N ratio) for compost (Figure 4a) and digestate (Figure 4b) samples.

A high degree of correlation was found between the biological stability indices, as the r-coefficient ranged from 0.82 to 0.99, indicating a strong linear relationship between the analyzed variables. Indeed, according to [39], values between 0.70 and 1.0 (−0.70 and −1.0) indicate a strong positive (negative) linear relationship.

Compost samples obtained very high correlations (Figure 4a), as the r-coefficient ranged from 0.90 (coefficient r between SH/OUR) to 0.99 (coefficient r between SH/RBP and OUR/RBP). Digestate samples achieved lower correlation coefficients than compost samples (Figure 4b); however, the value of the coefficient r in digestate samples still fell into a high range, varying from 0.82 (coefficient r between SH/OUR, RBP/OUR, and RBP/DRI) to 0.93 (coefficient r between SH/DRI).

In the case of digestate samples, it should be emphasized that the lowest values of the r-coefficient were obtained from the correlations between the RBP index with OUR and DRI indices, although they still have strong correlations (>0.70). As is well known, digestate is the solid and liquid residue of the anaerobic digestion of organic matter [40]. The RBP test assesses the potential production of residual methane (i.e., the amount of biogas that can potentially still be produced from residual matter) and is conducted under anaerobic conditions [12,21]. On the contrary, the OUR and DRI indices evaluate the decomposition of organic matter in the presence of oxygen and are conducted under aerobic conditions [20,41,42]. Therefore, the lower correlation between these indices could indicate that aerobic indices are less suitable in assessing the degree of stability and biodegradability on anaerobic matrices [42].

Finally, regarding the chemical-physical parameters, weak positive/negative (values between 0 and 0.3/−0.3) or moderate positive/negative (values between 0.3/−0.3 and 0.7/−0.7) linear relationships were generally found. Specifically, the few strong correlations were obtained in compost samples for the following pairs of variables: pH/OUR (−0.88), pH/TS (0.87), VS/OUR (0.9), VS/RBP (0.85), and TS/SH (0.7). On the other hand, no strong correlations related to chemical-physical parameters were observed in digestate samples.

Despite extremely high Pearson correlation coefficients (up to 0.99), the linear regression model may suffer from multicollinearity. This is because the independent variables may be highly correlated with each other, making it difficult to estimate the individual effect of each variable on the dependent variable and compromising the reliability of the estimated coefficients. This phenomenon can be analyzed using diagnostics such as the Variance Inflation Factor (VIF), which is missing, thus representing a limitation for this study. As previously mentioned, the developed model is a preliminary attempt to predict stability indices; therefore, future studies will have to take this aspect into consideration for comprehensive analysis.

3.3. Second-Order Polynomial Regression Predictive Model

3.3.1. Overview of Predictive Model Equations

To predict and compare the stability levels obtained by different indices, second-order polynomial regression models were developed, considering two group models. The first group model evaluated the relationship between RBP, OUR, and SH, which are the biological stability indices regulated at the European level, using RBP as the target index. The second group evaluated the relationship between DRI, OUR, and SH, representing a comparison between biological stability indices conducted under aerobic conditions and considering DRI as the target index, which is an index permitted by the Italian legislation. In addition to compost and digestate, predictive models were developed considering all sample datasets, since the standard does not distinguish stability levels according to the type of matrix [23,25,26,28]. Model results for DRI and RBP, predicted as functions of OUR and SH, are listed in

Table 3. Results of second-order polynomial regression models, summarizing sample type, target index, model equation, R-squared (R²), and the Root Mean Squared Error (RMSE).

Dataset	Target Index	Equation	R2	RMSE
Compost	RBP	RBP = −0.0852 + 0.0018·OUR + 0.0104·SH + −0.0000·OUR2 + −0.0002·SH2 + 0.0000·OUR·SH	0.96	0.038
Compost	DRI	DRI = −223.3858 + 21.4709·OUR + 45.2233·SH + −0.0817·OUR2 + −0.7418·SH2 + 0.4477·OUR·SH	0.97	257.85
Digestate	RBP	RBP = 0.0598 + 0.0026·OUR + −0.0003·SH + 0.0000·OUR2 + 0.0002·SH2 + −0.0002·OUR·SH	0.60	0.052
Digestate	DRI	DRI = −173.9605 + 28.1243·OUR + 28.3748·SH + 0.0282·OUR2 + 0.9664·SH2 + −0.6159·OUR·SH	0.95	266.32
All sample	RBP	RBP = 0.0295 + 0.0005·OUR + 0.0046·SH + −0.0000·OUR2 + −0.0001·SH2 + 0.0000·OUR·SH	0.84	0.065
All sample	DRI	DRI = −57.9241 + 19.7859·OUR + 32.0990·SH + −0.0182·OUR2 + 0.1918·SH2 + 0.0046·OUR·SH	0.96	299.58

(the positive or negative sign of the coefficients defines the direction or position of the curve). The table summarizes all six second-order polynomial regression models fitted on the analyzed dataset. R² indicates the proportion of variance explained, and RMSE (Root Mean Squared Error) quantifies the average prediction error.

As highlighted, regression models showed that OUR and SH are strong predictors for both RBP and DRI in the compost dataset, which exhibited the highest R² (>0.96) and lowest RMSE. Digestate and all datasets displayed greater variability, indicating more complex dynamics. The model equation for RBP in the digestate dataset exhibited a very low R² (0.6), indicating poor predictive ability. However, in the other cases, the R² value was still high or acceptable, since it was between 0.84 and 0.96.

3.3.2. Relationship Between RBP, SH, and OUR (Compost Dataset)

Figure 5 shows the 3d (Figure 5a) and 2d (Figure 5b) contour plots of the second-order polynomial regression model on the relationships between RBP, SH, and OUR, considering the compost dataset. The compost model achieved an excellent fit (R² ≈ 0.96, Table 3), suggesting a well-defined quadratic response surface. RBP increased with OUR and SH up to intermediate levels, followed by a decline likely due to microbial oxidation and stabilization (Figure 5a). According to the prediction model, the values of OUR and SH corresponding to the threshold limit of RBP (0.25 Nl/g_VS) ranged from ≈13.1 to 90.1 mmolO₂/kg_VS/h and ≈5.7 to 40.5 °C, respectively (Figure 5b). Levels predicted by the model showed great variability in the expected results, reaching biological stability values outside the thresholds established by the European Fertilizer Regulation (EU) 2019/1009.

3.3.3. Relationship Between RBP, SH, and OUR (Digestate Dataset)

Figure 6 shows the 3d (Figure 6a) and 2d (Figure 6b) contour plots of the second-order polynomial regression model on the relationships between RBP, SH, and OUR, considering the digestate dataset. The digestate model showed lower predictive strength (R² ≈ 0.60, Table 3), reflecting higher variability in the dataset. While OUR had a positive influence on RBP, the interaction with SH introduced nonlinearity and possible inhibitory effects (Figure 6a). According to the prediction model, the values of OUR and SH corresponding to the threshold limit of RBP (0.25 Nl/g_VS) ranged from ≈12.6 to 112.1 mmolO₂/kg_VS/h and ≈3.0 to 52.7 °C, respectively (Figure 6b). Levels predicted by the model showed great variability in the expected results, reaching biological stability values outside the thresholds established by the European Fertilizer Regulation (EU) 2019/1009.

The relatively poor performance of the RBP digestate model suggested that aerobic and respirometric indices may have limited applicability to anaerobic matrices and can be attributed to both physicochemical and microbiological factors. Specifically, digestate has already undergone anaerobic stabilization, which substantially reduces its residual biogas potential. However, a low RBP value does not necessarily imply equally low aerobic biodegradation potential, as residual organic matter may still sustain aerobic microbial activity. Moreover, digestate is characterized by a heterogeneous composition, often including fractions with different degrees of recalcitrance and stabilization history. This heterogeneity, combined with the intrinsic differences between anaerobic and aerobic degradation pathways and microbial consortia, can decouple RBP from aerobic stability indices such as OUR and DRI, resulting in weaker and less predictable relationships for digestate compared to compost.

3.3.4. Relationship Between RBP, SH, and OUR (All Sample Dataset)

Figure 7 shows the 3d (Figure 7a) and 2d (Figure 7b) contour plots of the second-order polynomial regression model on the relationships between RBP, SH, and OUR, considering the all sample dataset. The global model demonstrated a strong correlation between RBP, OUR, and SH, explaining approximately 80% of the data variance. RBP increased moderately with SH and OUR, but excessive oxygen uptake appeared to limit RBP (Figure 7a). According to the prediction model, the values of OUR and SH corresponding to the threshold limit of RBP (0.25 Nl/g_VS) ranged from ≈12.6 to 114.6 mmol O₂/kg_VS/h and ≈3.0 to 45.2 °C, respectively (Figure 7b). Levels predicted by the model showed great variability in the expected results, reaching biological stability values outside the thresholds established by the European Fertilizer Regulation (EU) 2019/1009.

3.3.5. Relationship Between DRI, SH, and OUR

The following figures display the 3d and 2d contour plots of the second-order polynomial regression model on the relationships between DRI, SH, and OUR, considering the compost dataset (Figure 8), the digestate dataset (Figure 9), and the all sample dataset (Figure 10).

Levels predicted by the model showed great variability in the expected results; however, it is worth noting that the DRI seemed to be a more severe index than OUR and SH. Predicted values of OUR and SH corresponding to the threshold limit of DRI (1000 mgO₂/kg_VS/h) are summarized in

Table 4. Predicted values of OUR and SH corresponding to the threshold limit of DRI (1000 mgO₂/kg_VS/h), based on the second-order polynomial regression model.

Dataset	OUR Min (mmol O2/kgVS/h)	OUR Max (mmol O2/kgVS/h)	SH Min (°C)	SH Max (°C)
Compost	12.6	42.7	3.0	17.5
Digestate	13.1	35.3	5.7	20.2
All sample	12.6	44.1	3.0	19.8

, based on the second-order polynomial regression model. In this case, compared to the models discussed in the previous sections, the maximum values obtained for the OUR are much lower and closer to the threshold limit (25 mmol O₂/kg_VS/h); on the other hand, the maximum SH values are always below 30 °C, thus indicating a level of biological stability that complies with the established standards of the European Fertilizer Regulation (EU) 2019/1009.

4. Conclusions

This study compared four biological stability indices—OUR, SH, RBP, and DRI—which were tested on 50 samples of compost and digestate collected from different organic waste treatment plants in Italy [6,7]. In addition, Pearson’s coefficients and second-order polynomial regression models were developed to measure the strength and direction of the linear relationship between any pair of indices and compare the stability levels obtained by different indices.

Overall, the results revealed that most of the compost and digestate samples were quite far from European standards, according to OUR, SH, and DRI indices. On the contrary, the RBP test seemed to be less stringent than the other indices, since a much larger number of samples was closer to or in compliance with the established threshold [43]. Data analysis using Pearson’s coefficients showed a strong linear correlation between the indices. Nevertheless, the linear regression predictive model based on experimental data demonstrated that the indices could not represent the same level of stability, providing poor consistency and variability in the predicted values and established threshold. Although it belongs to the Italian normative, it should be noted that the DRI test appeared to be more severe than the other aerobic indices of the OUR and SH [44].

The results obtained suggest revising the biological stability indices and threshold values, depending on the type of matrix. In particular, the observed discrepancies suggest that a single set of interchangeable indices and thresholds may not adequately reflect the biological behavior of different matrices [14,45]. From a policy and regulatory perspective, several practical pathways can be envisaged to improve coherence and reduce the risk of inconsistent classifications under the EU Fertiliser Regulation. These include: (i) the adoption of matrix-specific threshold values that explicitly distinguish between compost and digestate; (ii) tiered or stepwise testing strategies, in which one index is used as a screening tool and additional tests are applied when values are close to regulatory limits; and (iii) combined or decision-tree approaches integrating aerobic and anaerobic indices, thereby accounting for different degradation pathways.

Such approaches could enhance regulatory robustness without necessarily increasing testing complexity or costs, while better aligning stability assessment with the actual biological characteristics of the treated materials. For clarity, the main outcomes of this study can be summarized as follows:

Scientific findings

• Biological stability indices permitted under EU Regulation 2019/1009 are strongly correlated but not equivalent in terms of stability classification.
• DRI emerged as the most stringent aerobic index (probably due to methodological procedure, ensuring complete and continuous aeration of samples); whereas RBP appeared to be the least restrictive, particularly for digestate samples [13].
• Regression analysis demonstrated that different indices and thresholds may lead to inconsistent regulatory outcomes, despite similar trends in experimental data.

Regulatory and operational implications

• The current flexibility in index selection under EU Regulation 2019/1009 may result in non-uniform stability assessments across facilities and Member States.
• Matrix-specific criteria and multi-level assessment strategies could improve harmonization and reduce misclassification risks.
• A regulatory framework integrating complementary indices may better support circular economy objectives and the safe agricultural use of compost and digestate [46].

Overall, this study suggests a more differentiated and robust regulatory approach to the assessment of biological stability, contributing to the harmonization of assessment criteria and the promotion of fertilizers derived from organic waste from a circular economy perspective [47,48].