Feasibility of near infrared spectroscopy for estimating suppressiveness of carnation (Dianthus cariophyllus L.) fusarium wilt in different plant growth media

Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 280 (2022) 121528 Available online 18 June 2022 1386-1425/© 2022 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). Feasibility of near infrared spectroscopy for estimating suppressiveness of carnation (Dianthus cariophyllus L.) fusarium wilt in different plant growth media Víctor M. Fern´ andez-Caban´ as a , * , Celia Borrero a , Daniel Cozzolino b , Manuel Avil´ es a a Dpto. Agronomía, Universidad de Sevilla, ETSIA, Ctra. Utrera km.1, 41013 Seville, Spain b Centre for Nutrition and Food Sciences, Queensland Alliance for Agriculture and Food Innovation (QAAFI), The University of Queensland, Brisbane, QLD, Australia HIGHLIGHTS GRAPHICAL ABSTRACT •Fusarium is an important disease in carnation crops worldwide. •Suppressive media can be an alternative to phytosanitary products. •NIR technology can predict carnation Fusarium wilt suppressiveness in compost. ARTICLE INFO Keywords: NIRS Compost Severity Biological control Fusarium oxysporum f. sp. dianthi, soilless ABSTRACT Fusarium wilt is one of the most widespread diseases in carnation crops in a large number of countries. Plant protection products commonly used to remedy the disease have been considered ineffective and environmentally unsafe for commercial use. As an alternative, the use of suppressive growth media has been proposed. The aim of this study was to evaluate the use of a rapid method such as near infrared spectroscopy (NIRS) to monitor and evaluate suppressive media potential. The NIR spectra were collected from 6 plant growth media used in a series of trials to evaluate suppressiveness to carnation Fusarium wilt namely grape marc compost, cork compost, olive oil husk +cotton gin trash composted and mixed with rice husk, spent mushroom composted and mixed with peat, coir fiber and light peat. The NIR calibration models showed promising results for estimating pH, β-glucosidase activity, disease severity (AUDPC and RLSBX) in the growth media evaluated, with coefficients of determination of 0.99, 0.98, 0.98 and 0.90; SECV of 0.09, 11.63, 0.05 and 0.10; and RPD values of 13.86, 6.62, 7.19 and 3.24; respectively. NIR spectroscopy could become a useful non-destructive and fast analytical tool for the identification of Fusarium wilt suppressive composts, avoiding the use of reagents. * Corresponding author. E-mail addresses: [email protected] (V.M. Fern´ andez-Caban´ as), [email protected] (C. Borrero), [email protected] (D. Cozzolino), [email protected] (M. Avil´ es). Contents lists available at ScienceDirect Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy journal homepage: www.journals.elsevier.com/spectrochimica-acta-part-a- molecular-and-biomolecular-spectroscopy https://doi.org/10.1016/j.saa.2022.121528 Received 23 March 2022; Received in revised form 9 June 2022; Accepted 16 June 2022 Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 280 (2022) 121528 2 1. Introduction One of the diseases that causes the greatest losses in crops worldwide is Fusarium wilt. In addition, the control of this disease is very complicated in production systems [1]. The carnation (Dianthus cariophyllus L.) is susceptible to Fusarium oxysporum Schlecht. f. sp. dianthi (Prill. & Del.) Snyd. & Hans (Fod). Fusarium wilt is prevalent in many carnation production areas, and its control is critical for crop systems in SW Spain, one of the leading carnation cultivation areas in the European Union [2]. In order to limit the propagation of the disease, soil fumigation based on phytosanitary products is applied, although many of them are currently banned or subjected to regulatory limitations due to potential human health and environmental concerns. As an alternative to the use of chemicals for soil disinfection, soilless cultivation of carnations in growth media has been recently proposed as one approach to control this disease [3]. Compost is a material commonly used as a substrate in soilless crops, presenting numerous advantages from the environmental and phytosanitary approaches [4]. Composting of organic waste from different sources, including agriculture, tree bark, animal manure, food industries, sawdust, domestic and municipal solid waste, as well as sewage sludge, has turned out to be an interesting option for the sustainable management of potential pollutants [4]. Currently, environmental policies, laws and regulations focus on the transformation of organic waste into valuable resources through composting processes, all within the framework of the circular economy [5]. Nevertheless, stimulation of biological properties of growth media plays an important role in the application of compost, as both microbial biomass and enzyme activities can be incremented, inducing suppressiveness against some diseases [6,7]. Previous studies have shown the possibility of estimating and evaluating the suppression of Fusarium wilt in carnations from analytical parameters such as pH or beta-glucosidase activity of growth media [8]. To determine the quality of the compost, a large number of physical, chemical and biological methods have been used in different works that, in general, are time consuming and arduous [9]. These difficulties in finding a precise and simple methodology for the analysis of this material, so heterogeneous in its composition, has hindered the effective evaluation of compost quality in commercial production systems [10]. However, the development of faster analytical techniques would allow a more efficient evaluation of the usefulness of the growth media for the control of the disease. Among the rapid analytical techniques, those based in spectroscopic tend to stand out since, in addition, they are usually non-destructive. Some spectroscopic measures such as cross-polarized magic-angle spinning 13 C nuclear magnetic resonance ( 13 C-CPMAS-NMR) have been related to suppressiveness of compost amendments to peats [11] and of peats [12]. This technique has been also combined with diffuse reflectance Fourier-transform infrared (DR-FTIR) to estimate suppressiveness of Pythium damping-off of cucumber [13]. Among the spectroscopic techniques, Near Infrared Spectroscopy (NIRS) is one of the most outstanding for the rapid analysis of a wide variety of products and parameters. This technology has advanced spectacularly in recent years and is now used routinely for qualitative and quantitative analysis in many fields, including industries such as agri-food, pharmaceuticals and chemicals, but also in other sectors such as human health and environmental studies [14]. NIRS technology is used to analyse organic substances rapidly and cost-effectively. In composts and other plant growth media it has been used to predict quality parameters such as dry matter, total C, organic C, total N, C:N, S, K, ammonia, potentially mineralizable N, heavy metals, ash, pH, nitrogen dry matter, fibre fractions and thermophilic population [15–20]. It is therefore considered as a technique of great interest for quality control in the industrial compost process from agro-industrial waste materials, with excellent results for estimating total organic matter and total organic carbon, satisfactory results for EC, Fe and Mn, and moderate estimates for total nitrogen, C/N ratio and contents in P, K, Cu and Zn [21]. However, not many studies have been found in the literature on the estimation of compost suppressiveness by NIR spectroscopy and, for the few found, the results were not satisfactory. Michel et al. [22] reported moderate success for estimating fluorescein diacetate hydrolysis (FDAHR), specific enzyme activity and suppression of pathogens at an inoculation level of 5 ‰ related to rating or fresh weight; with some low coefficients of determination for the last two parameters (R 2 =0.49 and 0.47, respectively). More recently, Herrmann et al. [23] found that NIR calibrations to estimate FDA-hydrolysis and the germination tests with cress and salad only were able to discriminate between high and low values, while unsuccessful predictions with R 2 values below 0.50 were obtained for germination tests with ryegrass and bean and the disease suppression test using R. solani. Thus, considering the existing indications on the potential of NIR spectroscopy for the estimation of parameters related to the suppressiveness of the growth media, the following study was proposed whose main objective was to determine the value of using the NIRS technique to predict carnation Fusarium wilt suppressiveness related parameters in compost. 2. Materials and methods 2.1. Plant growth media Four different composted residues sourced from the agricultural industry waste were evaluated and characterised for carnation Fusarium wilt suppression as reported in our previous studies [8]. For this work, the following composted materials were used to obtain the NIR predictive models (a) Cork (Quercus suber L.) residues from industry (CC) industry, (b) grape marc (GMC) proceeding from the alcohol distillation industry, (c) olive oil husk blended with cotton gin trash, 1:1 v/v, composted and mixed with rice husk (1:1 v/v) (OC +R) and (d) spent mushroom compost combined with amended light peat (1:1 v/v) (SM + P). In order to be able to evaluate the suppressiveness of the growth media, two other materials commonly used in Spain as substrates were also selected as controls: fertilized Sphagnum light peat (Klasmann, Valimex, Palleter, Spain) and coir fibre (Cocopeat, Projar, Valencia, Spain). Amendment of peat samples was carried out with the addition of 4 g L −1 CaCO 3, and 0.33 g L −1 of K 2 O (50% high solubility granulated; Compo Agricultura S.L., Barcelona, Spain) and 4.15 g L −1 of P 2 O 5 (18% granulated; Fertiberia, Madrid, Spain). The different plant growth media were incubated at a water tension of 1000 Pa (adjusted by weight) for 14 days at a temperature of 25 ◦C, to standardize the initial conditions. A total of 36 growth media were evaluated in this study. Assessment of disease severity Disease suppressive properties of plant growth media were measured by a Fusarium wilt bioassay described elsewhere [8], using both the susceptible carnation cultivar ‘Medea’ and a monosporic isolate of Fusarium oxysporum f. sp. dianthi (Fod). Disease incidence was supervised every other day starting 55 days after planting, using scores for both wilted leaves and stems based on a symptom severity scale described by Baayen and van der Plas [24]. Disease severity was calculated as the ratio of the maximum possible disease severity. The standardized area under the disease progress curve (AUDPC) per pot was computed by disease severity integrated between symptoms onset and bioassay final time, dividing by the total duration (days) of the epidemic in each bioassay, in order to compare the different bioassays, which had variable epidemic durations. Mean values calculated per pot (containing four plants) of the relative length of the stem with brown xylem (RLSBX) were recorded. 2.2. Plant growth media characteristics At the beginning of the bioassays, both the pH and the microbial V.M. Fern´ andez-Caban´ as et al. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 280 (2022) 121528 3 activity in the incubated growth media were determined [8]. The pH was measured in triplicate in each of the plant growth media in an aqueous extract (2:1; vol/vol), as described by Gabri¨ els et al. [25]. Microbial activity was determined using the β-glucosidase activity assay, according to Bandick and Dick [26]. Six samples were analysed for each plant growth medium and bioassay. Physical-chemical composition and disease suppressive properties of growth media obtained in a previous work [8] were used as reference data to develop the NIR calibrations. Mean, standard deviation, maximum and minimum values for the parameters evaluated for the 6 samples of each growth media are displayed in Table 1. 2.3. NIR analysis and calibration development The NIR spectra of two subsamples were obtained from each organic plant growth media at the beginning of the bioassays. Reflectance spectra of compost samples were scanned on a Foss NIRSystems 6500 SY-II monochromator, from 400 to 2498 nm, every 2 nm (spectral bandpass 10 nm ±1 nm). Instrument had attached a transport module, and used a rectangular “natural product sample cell” cup (ref: IH-0331). Software WINISI ver. 1.5 (Infrasoft International, Port Matilda, PA) was used to manipulate and process the spectra, and to obtain calibration equations. A total of 14 combinations of pre-treatments were applied to obtain Modified Partial Least Squares (MPLS) calibration equations [27,28]. Scatter correction Standard Normal Variate combined with Detrending (SDT) [29] and six different derivative math treatments were used as mathematical spectral pre-treatments. WINISI derivative math treatments are defined by a four-digit nomenclature (a, b, c, d), where a is the derivative order; b is the derivative gap; c is the smoothing segment and d is the second smoothing segment [30]. Derivatives used were (1,5,5,1), (1,10,5,1), (1,10,10,1), (2,5,5,1), (2,10,5,1) and (2,1010,1), and scatter corrections were none and SDT. All calibrations developed were obtained using the full spectral range (400–2498 nm). During calibration, T and H outlier detection limits were set to 2.5 and 3, respectively. T outliers are referred to samples with significant differences between their reference and predicted values, while H outliers are samples with extreme spectral distances (H >3) to the centre of the calibration set [30]. NIR calibration models were evaluated by examining the statistical values obtained for calibration procedures, selecting the best combination of pre-treatments for each analytical parameter. Cross-validation was used to establish the optimum number of factors for the regression models and to avoid overfitting. Because of the reduced size of the calibration set, the data set was split into seven sub-sets for crossvalidation, each group then being predicted with a calibration model obtained with the other samples. In the last stage, validation errors were combined to calculate a standard error of cross-validation (SECV). Main calibration statistics used to evaluate the models were determination coefficient for calibration (R 2 ), standard error of calibration (SEC), determination coefficient for cross validation (r 2 ), SECV and the RPD, computed as the ratio of the standard deviation of the original data (SD) to the SECV [31]. Selection of best calibration models was based on their highest r 2 value and the lowest SECV for each of the studied constituents. Analytical parameters used for the NIR calibrations were AUDPC, RLSBX, pH and microbial activity. No external validation was possible due to the reduced number of samples and the setup of the experiment. However, the SECV is the single best estimate of the standard error of prediction (SEP), and similar to the average SEP obtained from 10 randomly-chosen prediction sets [30]. 3. Results and discussion The growth media samples studied in this work showed adequate variability for the physical-chemical parameters and severity indices analysed (Table 1), which is quite convenient for the development of NIR calibrations. Both pH and β-glucosidase activity values were higher for the four of the compost types analysed, while peat and coir fibre samples had the minimum values. Severity indices were lower for composts and higher for peat and coir fibre as a consequence of their correlation with tomato and carnation Fusarium wilt severity as reported in previous studies [8,32]. The average visible (VIS) and NIR spectra of the studied samples (Fig. 1) provided an insight into the variability of the growth media used in the experiment. The bands observed in the VIS wavelength range might be associated with the variation in colour of the samples. The most prevalent were observed around wavelengths associated with blue (427–476 nm), green (497–570 nm) and red (618–780 nm) colours where differences among samples were notable. In the NIR region, bands located at 1100–1390 nm can be associated to the second overtone of symmetric and antisymmetric C-H stretch vibration (-CH, -CH 2 and –CH 3 ), attributed to carbohydrates, lipids and proteins [33]. Absorption bands at 1390–1660 nm are due to O-H vibration and H-bridges in water molecules, which overlaps with stretch and deformation of C-H of carbohydrates and lipids [33]. In this zone, characteristic bands were also observed at 1448 nm associated with N-H primary aromatic amine (N-H asymmetric) mainly related with nitrogen containing compounds (e.g. Table 1 . Physical-chemical composition (pH and β-glucosidase activity) and disease suppressive properties (AUDPC and RLSBX) of the six organic plant growth media and of the global population (n =36). pH β-glucosidase activity AUDPC RLSBX Growth media Mean ± SD Range Mean ± SD Range Mean ± SD Range Mean ± SD Range OC +R 7.92 ±0.022 7.91–7.95 252.20 ±6.818 244.11–259.25 0.07 ±0.034 0.03–0.09 0.31 ±0.100 0.19–0.41 SM +P 7.14 ±0.024 7.11–7.17 59.10 ±13.409 44.13–74.13 0.10 ±0.041 0.06–0.15 0.40 ±0.078 0.30–0.47 CC 7.76 ±0.017 7.74–7.78 91.77 ±15.907 77.16–111.58 0.12 ±0.074 0.05–0.21 0.52 ±0.152 0.36–0.69 GMC 7.47 ±0.003 7.47–7.47 84.00 ±4.822 80.12–90.15 0.01 ±0.005 0.00–0.01 0.10 ±0.034 0.07–0.14 Coir fibre 5.66 ±0.011 5.65–5.67 32.56 ±3.383 28.38–35.75 0.88 ±0.028 0.85–0.90 0.93 ±0.103 0.80–1.00 Peat 5.29 ±0.289 5.08–5.66 30.25 ±4.883 24.15–34.67 0.80 ±0.056 0.73–0.84 1.00 ±0.000 1.00–1.00 All samples 6.87 ±1.044 5.08–7.95 91.65 ±77.020 24.15–259.25 0.33 ±0.371 0.00–0.90 0.54 ±0.340 0.07–1.00 Fig. 1. Average vis-NIR spectra of the six growth media. V.M. Fern´ andez-Caban´ as et al. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 280 (2022) 121528 4 amines, proteins) [19,31]. At 1660–1870 nm it is reported a first overtone of symmetric and antisymmetric C-H stretch vibration of –CH 2 and –CH 3 groups associated to lipids and proteins [33] and O-H combinations at 1792 nm [19,31]. Absorptions in the region 1870–2015 nm are attributed to the first overtone of carbohydrates [33], while bands at 1930 nm are associated to O-H stretching and H-O-H combinations mainly related with O-H in cellulose and water [19,31]. At 2015–2230 nm it can be appreciated absorption bands characteristic associated to C- H, N-H and C=O, corresponding to a combination of C-H stretching and bending of –CH 3 and –CH 2 groups for carbohydrates, lipids and proteins. Finally, bands at 2360–2500 nm are associated to C-H combination bands for lipids and proteins [33]. The development of robust NIR prediction models is usually associated with large data sets. It has been reported by several authors that the minimum number of samples necessary for stable calibration is between 50 and 100. Nevertheless, more samples are recommended when sources of variance are high [34]. The literature on prediction models in compost to estimate chemical or physical composition shows calibration sets developed using 70 to 300 samples [15,19]. In the present study, although the number of samples for calibration was low, prediction models with excellent calibration statistics were obtained, as it can be appreciated in Table 2. In relation to the spectral pre-treatments that provided the best calibration results, it stands out that, except for the prediction of AUDPC, it was not necessary to use the SDT correction. Moreover, no derivative treatment was used for the development of calibration for RLSBX, which seems to indicate that the characteristic texture of the samples could be related to their suppressive properties. Satisfactory prediction models were obtained for the prediction of both pH and microbial activity, with calibration equations explaining more than 90% of the variation, and RPD values of 13.86 and 6.62; respectively. Previous studies reported lower precision for pH estimation with coefficients of determination of 0.89–0.90 and RPD values of 2.54–2.98 [17,21]. For the prediction of β-glucosidase activity in Mediterranean soils, r 2 value of 0.93 and RPD of 3.66 was reported by other authors [35]. For other enzymatic parameters related to microbial activity, Chakraborty et al. [36] found r 2 and RPD values of 0.91 and 3.2, respectively for the estimation of FDA-HR, while Michel et al. [22] reported r 2 =0.75 and RPD =2.2; and Herrmann et al. [23] r 2 =0.53 and RPD =1.5, for the same parameter. In relation to the estimation of severity, calibration equations explained 98% and 90% of the variations in AUDPC and RLSBX, respectively, with cross validation errors of 0.06 and 0.11, and RPD values of 7.19 and 3.24; respectively. Few references can be found in the scientific literature on the estimation of parameters related to the suppressiveness of diseases in growth media using NIR spectroscopy. Disease suppression (damping-off rate) for the pathosystem Rhizoctonia solani / basil was not able to be predicted by NIR, reporting a coefficient of determination of 0.31 [23]. In another study, Michel et al. [22] evaluated predictive models to estimate the suppression of Pythium ultimum to the disease rating and to the fresh weight of cucumber plants, obtaining RDP values of 1.4 and coefficients of determination of 0.47 and 0.71, respectively, lower again than those obtained in this study. The results of this study also emphasized the improvements achieved in the estimation of suppressiveness in growth media for the disease evaluated with a direct and rapid method based on NIR spectra alone. This method avoided the use of chemical and biochemical determinations, indicating that NIR spectroscopy could be used to predict Fusarium wilt suppressiveness of organic plant growth media. The good prediction ability obtained with the NIR calibrations developed for the estimation of pH, β-glucosidase activity, AUDPC and RLSBX might be due to the inter correlation between both pH and β-glucosidase activity with severity. It is important to note that in a previous study we have demonstrated that this correlation (using the same set of samples as the one used in this study) explained more than 67% of the variation in RLSBX [8]. The NIR calibration models developed for the parameters measured explained more than 90% of variation in RLSBX. This amount of variation was higher than the obtained using parameters such as pH and β-glucosidase activity. These results suggested that the NIR provides with more information about the severity of this disease than the measurement of single physico-chemical parameters. In addition, this technique is cheaper, faster and more precise than pH and β-glucosidase activity measurements for the identification of Fusarium wilt suppressive plant growth media and their mixes as demonstrated by the calibration statistics obtained. 4. Conclusions Soilless cultures using peat or coir fibre are susceptible to Fusarium wilt. These plant growth media are mostly imported from outside Spain, where peat is sourced from valuable ecosystems. On the other hand, durable management of wastes generated by agriculture is a complex problem. Composting is one of the most desirable techniques for reducing waste volume. Therefore, the use of suppressive compost for plant growth media can solve these problems. NIR spectroscopy could become a useful non-destructive and fast analytical tool for the identification of Fusarium wilt suppressive composts, avoiding the use of reagents. However, further research must be done to validate the calibration models with more samples and types of compost. CRediT authorship contribution statement Víctor M. Fern´ andez-Caban´ as: Conceptualization, Formal analysis, Writing – original draft. Celia Borrero: Investigation, Writing – review & editing. Daniel Cozzolino: Writing – review & editing, Visualization. Manuel Avil´ es: Conceptualization, Methodology, Resources, Writing – review & editing. Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgements This research was supported by grants AGL2002-04313-C03-03 from the Ministerio de Ciencia y Tecnología and AGL2005-08137-C03-02 Table 2 NIRS calibration statistics for the prediction of standardized area under disease progress curve (AUDPC) and relative length of stem with brown xylem (RLSBX) for carnation plants and of pH and β-glucosidase activity in their six organic plant growth media (n =36). Component Pre-treatment Mean SD SEC R 2 SECV r 2 RPD pH None +2,10,5,1 6.94 1.030 0.04 1.00 0.09 0.99 13.86 β-glucosidase activity None +1,5,5,1 91.65 77.020 9.62 0.98 11.63 0.98 6.62 AUDPC SDT +2,10,10,1 0.31 0.366 0.03 0.99 0.05 0.98 7.19 RLSBX None +0,0,1,1 0.54 0.340 0.08 0.94 0.10 0.90 3.24 R 2 : determination coefficient for calibration, SDT: Standard normal variate +Detrending, SEC: standard error of calibration, r 2 : determination coefficient for cross validation, SECV: standard error of cross validation. V.M. Fern´ andez-Caban´ as et al. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 280 (2022) 121528 5 from the Ministerio de Educaci´ on y Ciencia (Proyectos I+D del Plan Nacional de Investigaci´ on Científica, Desarrollo e Innovaci´ on Tecnol´ ogica) of Spain. We thank S. Castillo and F. Arestoy for excellent technical assistance and Recomsa for supplying composted spent mushroom samples. The authors also thank the Agricultural Research Service of the University of Seville for technical support and access to its facilities. The authors have no conflict of interest to declare. References [1] K. Toyota, M. Kimura, Autecology of Fusarium oxysporum f. sp. raphani in soil, in: Fusarium Paul E. Nelson Meml. Symp., 2001, p. 138. [2] R. Casta˜ no, B. Scherm, M. Avil´ es, Genetic Diversity of Fusarium oxysporum f. sp. dianthi in Southern Spain, J. Mycol. 2014 (2014) 1–14, https://doi.org/10.1155/ 2014/582672. [3] M. Pizano, Floriculture and the environment: growing flowers with out methyl bromide UNEP, Paris, Fr. (2001). [4] M. Avil´ es, C. Borrero Vega, M. Trillas, Review on Compost as an Inducer of Disease Suppression in Plants Grown in Soilless Culture, Dyn. Soil, Dyn. Plant. 5 (2011) 1–11. [5] A. Vergnoux, M. Guiliano, Y. Le Dr´ eau, J. Kister, N. Dupuy, P. Doumenq, Monitoring of the evolution of an industrial compost and prediction of some compost properties by NIR spectroscopy, Sci. Total Environ. 407 (2009) 2390–2403, https://doi.org/10.1016/j.scitotenv.2008.12.033. [6] K. Debosz, S.O. Petersen, L.K. Kure, P. Ambus, Evaluating effects of sewage sludge and household compost on soil physical, chemical and microbiological properties, Appl. Soil Ecol. 19 (2002) 237–248, https://doi.org/10.1016/S0929-1393(01) 00191-3. [7] K.L. Bailey, G. Lazarovits, Suppressing soil-borne diseases with residue management and organic amendments, Soil Tillage Res. 72 (2003) 169–180, https://doi.org/10.1016/S0167-1987(03)00086-2. [8] C. Borrero, I. Trillas, M. Avil´ es, Carnation Fusarium wilt suppression in four composts, Eur. J. Plant Pathol. 123 (2009) 425–433, https://doi.org/10.1007/ s10658-008-9380-4. [9] C. Borrero, S. Castillo, E. Casanova, G. Segarra, M.I. Trillas, R. Casta˜ no, M. Avil´ es, Capacity of composts made from agriculture industry residues to suppress different plant diseases, Acta Hortic. 1013 (2013) 459–464. 10.17660/ actahortic.2013.1013.56. [10] R. Albrecht, R. Joffre, R. Gros, J. Le Petit, G. Terrom, C. P´ erissol, Efficiency of nearinfrared reflectance spectroscopy to assess and predict the stage of transformation of organic matter in the composting process, Bioresour. Technol. 99 (2008) 448–455, https://doi.org/10.1016/j.biortech.2006.12.019. [11] C. Pane, R. Spaccini, A. Piccolo, F. Scala, G. Bonanomi, Compost amendments enhance peat suppressiveness to Pythium ultimum, Rhizoctonia solani and Sclerotinia minor, Biol. Control. 56 (2011) 115–124, https://doi.org/10.1016/j. biocontrol.2010.10.002. [12] M.J. Boehm, T. Wu, A.G. Stone, B. Kraakman, D.A. Iannotti, G.E. Wilson, L. V. Madden, H.A.J. Hoitink, Cross-polarized magic-angle spinning 13C nuclear magnetic resonance spectroscopic characterization of soil organic matter relative to culturable bacterial species composition and sustained biological control of pythium root rot, Appl. Environ. Microbiol. 63 (1997) 162–168, https://doi.org/ 10.1128/aem.63.1.162-168.1997. [13] A.G. Stone, S.J. Traina, H.A.J. Hoitink, Particulate Organic Matter Composition and Pythium Damping-Off of Cucumber, Soil Sci. Soc. Am. J. 65 (2001) 761–770, https://doi.org/10.2136/sssaj2001.653761x. [14] E.W. Ciurczak, B. Igne, J. Workman Jr, D.A. Burns, Handbook of near-infrared analysis, Fourth, CRC Press, 2021. [15] J.F. Collard, R. Agneessens, D. Stilmant, P. Dardenne, Use of near infrared spectroscopy for compost characterization, in: Near-Infrared Spectrosc. Proc. 11th Int. Conf., 2004, pp. 769–773. [16] H.J. Ko, H.L. Choi, H.S. Park, H.W. Lee, Prediction of heavy metal content in compost using near-infrared reflectance spectroscopy, Asian-Australasian, J. Anim. Sci. 17 (12) (2004) 1736–1740. [17] D.F. Malley, C. McClure, P.D. Martin, K. Buckley, W.P. McCaughey, Compositional analysis of cattle manure during composting using a field-portable near-infrared spectrometer, Commun. Soil Sci. Plant Anal. 36 (2005) 455–475, https://doi.org/ 10.1081/CSS-200043187. [18] M.J. O’Mahony, S.M. Ward, J. Lynch, Near infrared sensors (NIR) for peat moisture determination, in: Acta Hortic., International Society for Horticultural Science (ISHS), Leuven, Belgium, 1998: pp. 185–192. 10.17660/ActaHortic.1998.421.19. [19] H.S.S. Sharma, M. Kilpatrick, L. Burns, Determination of phase II mushroom (Agaricus bisporus) compost quality parameters by near infrared spectroscopy, J. Near Infrared Spectrosc. 8 (2000) 11–19, https://doi.org/10.1255/jnirs.259. [20] G.D. Smith, A. Jervis, M. Lennartsson, W.F. Bourne, Laboratory methods of estimating potentially mineralizable nitrogen in organic potting mixes. II. Development of near infrared reflectance spectroscopy method, Commun. Soil Sci. Plant Anal. 32 (2001) 2769–2781, https://doi.org/10.1081/CSS-120000960. [21] L. Galvez-Sola, R. Moral, M.D. Perez-Murcia, A. Perez-Espinosa, M.A. Bustamante, E. Martinez-Sabater, C. Paredes, The potential of near infrared reflectance spectroscopy (NIRS) for the estimation of agroindustrial compost quality, Sci. Total Environ. 408 (2010) 1414–1421, https://doi.org/10.1016/j. scitotenv.2009.11.043. [22] K. Michel, C. Bruns, T. Terhoeven-Urselmans, B. Kleikamp, B. Ludwig, Determination of chemical and biological properties of composts using near infrared spectroscopy, J. Near Infrared Spectrosc. 14 (2006) 251–259, https://doi. org/10.1255/jnirs.642. [23] S. Herrmann, J. Mayer, K. Michel, B. Ludwig, Predictive capacity of visible-near infrared spectroscopy for quality parameter assessment of compost, J. Near Infrared Spectrosc. 17 (2009) 289–301, https://doi.org/10.1255/jnirs.850. [24] R.P. Baayen, C.H. van der Plas, Localization ability, latent period and wilting rate in eleven carnation cultivars with partial resistance to Fusarium wilt, Euphytica 59 (1992) 165–174, https://doi.org/10.1007/BF00041269. [25] R. Gabri¨ els, W. Van Keirsbulck, O. Verdonck, Reference method for physical and chemical characterization of growing media: An international comparative study, in: II Symp. Hortic. Substrates Their Anal. XXIII IHC 294, 1990, pp. 147–160. [26] A.K. Bandick, R.P. Dick, Field management effects on soil enzyme activities, Soil Biol. Biochem. 31 (1999) 1471–1479, https://doi.org/10.1016/S0038-0717(99) 00051-6. [27] H. Martens, T. Naes, Multivariate calibration, John Wiley & Sons, Chichester, U.K., 1992. [28] J.S. Shenk, M.O. Westerhaus, New Standardization and Calibration Procedures for Nirs Analytical Systems, Crop Sci. 31 (6) (1991) 1694–1696, https://doi.org/ 10.2135/cropsci1991.0011183X003100060064x. [29] R.J. Barnes, M.S. Dhanoa, S.J. Lister, Standard Normal Variate Transformation and De-Trending of Near-Infrared Diffuse Reflectance Spectra, Appl. Spectrosc. 43 (1989) 772–777, https://doi.org/10.1366/0003702894202201. [30] J.S. Shenk, M.O. Westerhaus, Routine operation, calibration, development and network system management manual, NIRSystems Inc., Silver Spring, MD, USA, 1995. [31] P. Williams, Implementation of Near-infrared Technology, in: P.C. Williams, K. Norris (Eds.), Near-Infrared Technol. Agric. Food Ind., second ed., American Assoc. of Cereal Chemists Inc., St. Paul, Minnesota, USA, 2001, p. 143. [32] C. Borrero, M.I. Trillas, J. Ordov´ as, J.C. Tello, M. Avil´ es, Predictive factors for the suppression of fusarium wilt of tomato in plant growth media, Phytopathology 94 (2004) 1094–1101, https://doi.org/10.1094/PHYTO.2004.94.10.1094. [33] G. Lequeue, X. Draye, V. Baeten, Determination by near infrared microscopy of the nitrogen and carbon content of tomato (Solanum lycopersicum L.) leaf powder, Nat. Publ. Gr. (2016) 1–9, https://doi.org/10.1038/srep33183. [34] P.C. Williams, K. Norris, Variables affecting near-infrared spectroscopic analysis, in: P.C. Williams, K. Norris (Eds.), Near-Infrared Technol. Agric. Food Ind., second ed., American Association of Cereal Chemists, Inc. St. Paul, MN, USA, 2001, pp. 171–198. [35] R. Zornoza, C. Guerrero, J. Mataix-Solera, K.M. Scow, V. Arcenegui, J. Mataix- Beneyto, Near infrared spectroscopy for determination of various physical, chemical and biochemical properties in Mediterranean soils, Soil Biol. Biochem. 40 (2008) 1923–1930, https://doi.org/10.1016/j.soilbio.2008.04.003. [36] S. Chakraborty, B.S. Das, M. Nasim Ali, B. Li, M.C. Sarathjith, K. Majumdar, D. P. Ray, Rapid estimation of compost enzymatic activity by spectral analysis method combined with machine learning, Waste Manag. 34 (2014) 623–631, https://doi. org/10.1016/j.wasman.2013.12.010. V.M. Fern´ andez-Caban´ as et al.