Seasonal Impacts of Organic Fertilizers, Cover Crop Residues, and Composts on Soil Health Indicators in Sandy Soils: A Case Study with Organic Celery

Abstract

While integrated practices are used in organic vegetable production for soil fertility management, their impacts on short- and long-term soil health across diverse cropping systems and environments need to be better understood, especially in sandy soils. In this two-year study (2022–2023 and 2023–2024) conducted on certified organic land, a suite of soil physical, chemical, and biological properties at the end of each organic celery (Apium graveolens L. var. dulce) production season were analyzed, with one set of field experiments assessing the influence of preplant organic fertilizers and the other set examining the effects of composts and sunn hemp (Crotalaria juncea L.) as a rotational cover crop before celery planting. Compared to feather meal-based organic fertilizer, the poultry litter-based organic fertilizer enhanced soil K and Mg base saturation, promoted micronutrient availability, and increased the overall soil fertility score. Sunn hemp cover cropping impacted soil N dynamics, and both yard waste compost and vermicompost increased the overall soil health score by over 4.0% compared to the no compost control, with yard waste compost resulting in the highest level of soil active C (10.8% higher than the control). Seasonal variations were observed in many soil parameters measured, along with marked interactions among nutrient management practices and production seasons. This study highlights the complexity of soil health assessments and improvement for sandy soils with low water and nutrient retention, and the importance of long-term, systematic studies under organic crop production.

1. Introduction

Organic cropping systems place an emphasis on integrating cultural, biological, and mechanical practices to foster resource cycling, promote ecological balance, and conserve biodiversity, with maintaining and promoting soil health as a foundational principle [1,2,3]. Soil can be considered a dynamic living system, and soil quality and health indicators that span the physical, chemical, and biological realms can be used as a basis for setting sustainable management goals and developing monitoring tools to help enhance soil ecological functions toward improving plant health and resilience [4,5,6].

Maintaining soil fertility and ensuring nutrient availability, especially nitrogen (N) availability, are among the major challenges in organic crop production in the southeastern US and beyond [7], largely driven by the need to synchronize nutrient supply, primarily through soil microbiological processes, with crop demand at critical stages for optimizing crop growth and development. Organic fertilizers can be used to provide N and other macro- and micronutrients to vegetable crops; however, the mineralization rates of organic fertilizers are input-specific, and often unpredictable due to interactions with site- and season-specific factors such as soil moisture and temperature [8]. While recent efforts have vastly improved our understanding of N mineralization from organic fertilizers [9,10], continued development of new organic fertilizer products and interactions among fertilizer ingredients, soil texture and organic matter (OM), and abiotic factors as well as soil biological activity drive a continued need for research in this area to enhance nutrient use efficiency in organic systems. On the other hand, it should be pointed out that the mere reliance on input substitution is not well aligned with the ecological foundation of organic production. Despite the short-term or seasonal contribution of organic fertilizers to soil fertility and crop nutrition, holistic applications of soil-building practices are essential in organic systems to help maintain the soil nutrient pool and buffering capacity through increased soil OM content and pools of soil organic carbon (SOC) as well as enhanced diversity and activity of soil microbiota [11,12].

Integrated nutrient management employs a suite of practices besides direct fertilization, such as crop rotations, incorporating green manure and cover crop residues, and adding composts, to build soil OM and provide essential nutrients. It helps to exploit the potential synergies between organic fertilizers and other amendments while improving the synchrony of N supply with crop demand [13]. Cover crop residues have been shown to promote N availability and broader aspects of soil health, including reduction of bulk density (BD), weed and disease pressure, and soil erosion, while enhancing SOC and soil microbial populations [14,15,16]. Still, more work is warranted to elucidate their interactions with organic fertilizers, other amendments, and environmental conditions, particularly in sandy soils, where building SOC and promoting many of the physical and biological aspects of soil health poses unique challenges. Similarly, the promising benefits of composts as contributors of SOC and soil microbes, and for improving N use efficiency and plant disease suppression are becoming increasingly recognized [17,18], though more research is needed to ascertain the role of composts derived from various feedstocks, as well as short- and long-term impacts on soil health indicators.

To date, research-based information is scarce with respect to short-term, seasonal responses of soil health parameters to integrated management practices in organic vegetable production systems under sandy soil conditions. Building upon organic celery (Apium graveolens L. var. dulce) field trials designed to further develop integrated nutrient management strategies for Florida sandy soils, this soil assessment-focused study was aimed at determining the seasonal influence of granular organic fertilizers, leguminous cover cropping, and different compost products on a range of physical, chemical, and biological indicators of soil health at the end of each organic celery production season.

2. Materials and Methods

2.1. Experimental Site

Soil samples were collected from two sets of field experiments conducted on certified organic land at the University of Florida Plant Science Research and Education Unit (PSREU) in Citra, FL (about 20 m above sea level), aiming to compare the impacts of preplant organic fertilizer practices, and the integration of cover cropping, composts, and organic fertilizer application on soil health at the end of a celery production season. At all experimental sites for both studies and over two years, the soil type is Gainesville loamy sand, characterized as hyperthermic coated typic quartzipsamments and composed of approximately 93.5% sand, 6.4% clay, and 0.1% silt with an average soil pH of 6.4–6.6 and OM contents ranging from 0.56–0.91% in 2022 and 2023, respectively.

2.2. Organic Fertilization Trial Setup

Field preparation for the organic fertilization trials was initiated on 2 November 2022 and 6 November 2023 for the 2022 and 2023 trials, respectively. Land was initially tilled prior to the formation of false beds formed on 1.5 m centers (1.5 m between centers of adjacent beds). Each of the two contrasting granular organic fertilizers was applied to false beds and incorporated into the soil with a rotary tiller to a depth of 15 cm. On the same day, raised beds (89 cm wide and 18 cm high) were formed and covered with 0.3 mm (thickness) black plastic mulch (Intergro, Inc., Clearwater, FL, USA) with two drip tapes (30.5 cm emitter spacing, 3.4 L h⁻¹ m⁻¹ flow rate; Jain Irrigation Inc., Haines City, FL, USA) offset by 20 cm on each bed. Organic celery ‘Tall Utah’ (Seed Savers Exchange, Decorah, IA, USA) was transplanted in double rows (30 cm between- and within-row spacing) on 4 November 2022 and 8 November 2023, 12 weeks after seeding in 128-cell transplant trays in greenhouse conditions (23 °C/16 °C for day/night temperature setting). Each experimental unit contained 30 plants, and irrigation was scheduled according to approximations of reference crop evapotranspiration from the Florida Automated Weather Network (FAWN: https://fawn.ifas.ufl.edu/tools/et/) weather station in Citra, FL, USA. Throughout the production season, crop irrigation requirements ranged from 10,666 L ha⁻¹ to 14,221 L ha⁻¹, supplied in 30–60 min of daily drip irrigation over 2–3 cycles per day. Celery was harvested on 17 February 2023 and 27 February 2024 in the 2022 and 2023 trials, respectively.

The experiment followed a randomized complete block design with four replications, comparing Nature Safe 10N-0.9P-6.6K (Darling Ingredients, Irving, TX, USA) to Everlizer 3N-1.3P-2.5K (Organic Growing Solutions, Mayo, FL, USA), two types of commonly used organic fertilizers for vegetable production in Florida. Nature Safe is a granular organic fertilizer containing feather meal, meat and bone meal, and blood meal as well as supplemental mined sulfate of potash, while Everlizer is a heat processed poultry litter product. In this experiment, the two organic fertilizers were applied to the false beds at 280 kg N ha⁻¹. Potassium (K) supply was balanced by adding additional 0N-0P-41.5K sulfate of potash (Big K; JH Biotech, Inc., Ventura, CA, USA) as a preplant fertilizer to plots receiving Nature Safe, so both organic fertilizer treatments supplied 233 kg K ha⁻¹. Supplemental phosphorus (P) was not applied, other than the P supplied from the two granular organic fertilizers, due to very high levels of soil P (77–130 mg kg⁻¹) at the experimental sites according to pre-season baseline soil testing.

2.3. Integrated Nutrient Management Trials

A sunn hemp (Crotalaria juncea L.) (Hancock Seed Co., Dade City, FL, USA) cover crop was seeded at a rate of 45 kg ha⁻¹ on 31 August 2022 and 31 August 2023. The seeds were inoculated using Guard-N seed inoculant (Verdesian Life Sciences, Cary, NC, USA) at a rate of 7.5 g kg⁻¹ seed before broadcasting in plots (9 m × 9 m) arranged in a randomized complete block design with four replications, in comparison with the weedy fallow control. Sunn hemp was selected in this study as it is a warm-season leguminous cover crop with fast growth and robust biomass accumulation, commonly grown in the southeastern United States. The cover crop was terminated using a flail chopper (model 5700; Hiniker Company, Mankato, MN, USA) on 8 November 2022 and 7 November 2023, and the sunn hemp and weedy fallow plots were rototilled to a depth of 15 cm on 15 November 2022 and 14 November 2023.

After the cover crop termination and initial tillage to incorporate biomass, false beds were formed in sunn hemp and weedy fallow plots for the application of composts and granular organic fertilizer. The experiment was arranged in a split plot design with four replications, with cover crop as the whole plot factor and compost as the subplot factor. Each compost subplot consisted of one raised bed (15.2 m long, 89 cm wide, 18 cm high) within a cover crop whole plot. The compost treatments included (1) yard waste compost applied at 22.4 t ha⁻¹ (Yard; Veransa Group, Seffner, FL, USA), (2) vermicompost applied at 11.2 t ha⁻¹ (Vermi; Black Star Organic Products, LLC., Bell, FL, USA), (3) mixed compost applied at 22.4 t ha⁻¹ and made from a 1:1 gravimetric mix of the yard waste compost and vermicompost (Mixed), and (4) no compost control (Control). These composts were selected based on their unique feedstocks, where vermicompost is cow manure-based and yard waste compost is plant-based, as well as their availability in the region. Nature Safe 10N-0.9P-6.6K was applied at 98 kg N ha⁻¹ to all plots for preplant fertilization. Organic celery ‘Tall Utah’ was transplanted the following day (16 November 2022 and 15 November 2023) into similar black plastic-mulched beds and spaced as previously described, where each experimental unit contained 30 celery plants. Throughout the production season, celery was fertilized with Aqua Power 5N-0.4P-0.8K liquid fish fertilizer (JH Biotech, Inc.) and supplemental Big K sulfate of potash 0N-0P-41.5K to meet total rates (preplant application plus in-season fertigation) of 280 kg N ha⁻¹ and 150 kg K ha⁻¹. Celery was harvested on 8 March 2023 and 5 March 2024 in the 2022 and 2023 trials, respectively.

2.4. Soil Sampling and Laboratory Analyses

After the final organic celery harvests in both the organic fertilization and integrated nutrient management trials, soil was sampled for the analysis of a diverse array of physical, chemical, and biological soil health indicators. In the organic fertilization studies, soil was sampled at 106 and 112 days after transplanting (DAT) on 18 February 2023 (2022 trial) and 28 February 2024 (2023 trial), respectively. In the integrated nutrient management studies, soil was sampled at 113 DAT, on 9 March 2023 (2022 trial) and 7 March 2024 (2023 trial). Soil was sampled from underneath the plastic mulch in eight randomly selected locations within each experimental unit, using a soil probe with a 1.9 cm internal diameter. Soil cores were collected to a depth of 30 cm, composited into polyethylene bags, stored on ice during transportation to the University of Florida campus, and frozen at −20 °C with minimal disturbance before overnight shipping to Waters Agricultural Laboratories, Inc. (Camilla, GA, USA) for further analysis of a suite of physical, chemical, and biological soil health indicators. Upon arrival at the service laboratory, all soil samples were dried at 40 °C and either passed through a 2 mm sieve or hand-rolled to remove large chunks, in the case of aggregate stability, prior to analysis.

Aggregate stability is a soil physical property and was determined through the volumetric aggregate stability test (VAST), which uses a visual assessment of soil aggregates remaining after wet slaking, read on a mat with concentric lines corresponding to degrees of aggregate stability and reported on a percent basis [19]. Soil chemical properties included contents of soil nitrate N (NO₃-N), ammonium N (NH₄-N), P, K, magnesium (Mg), calcium (Ca), sulfur (S), aluminum (Al), boron (B), zinc (Zn), manganese (Mn), iron (Fe), and copper (Cu), cation exchange capacity (CEC), and base saturation (BS). Soil NO₃-N (mg kg⁻¹) was determined via flow injection analysis and cadmium reduction following extraction with KCl as described by Devi and Townshend [20], and soil NH₄-N (mg kg⁻¹) was similarly determined from KCl soil extractions via flow injection analysis and the Berthelot reaction as described by Alves et al. [21]. Following soil extraction using the Mehlich-III method [22], soil P, K, Mg, Ca, S, B, Zn, Mn, Fe, and Cu contents were simultaneously determined using inductively coupled plasma–optical emission spectroscopy (ICP-OES) and expressed in mg kg⁻¹ dried soil. CEC and BS were measured using concentrations of the exchangeable cations Ca, Mg, and K, where CEC (meq 100 g⁻¹) was calculated as the sum of the cations and BS (%) was determined as the proportion of each base cation relative to the CEC. Using soil nutrient status from Mehlich-III extractions and ICP-OES analysis, the proportions of P to Al and Fe (P:[Al + Fe]) and P to Ca and Mg (P:[Ca + Mg]) were further calculated to help evaluate P saturation, where values under 5% may indicate potential for P fixation. Soil pH was also determined following a 1:1 (v:v) soil to water suspension ratio [23].

Soil biological parameters were also assessed as a part of the suite of soil health assessments. Soil OM content (%) was determined by loss on ignition [24]. Solvita labile amino-N (SLAN) was analyzed to measure labile fractions of soil N by alkali hydrolysis, which are associated with soil microbial activity and mineralizable pools of soil N, and was expressed in mg kg⁻¹ [25,26]. Solvita CO₂ burst (CB), a measurement of soil respiration (mg CO₂-C kg⁻¹ day⁻¹), was assessed by re-wetting air-dried (40 °C) soil to 50% available pore space and quantifying CO₂ release using thin-layer gel chromatography [26]. Active carbon (C) was quantified through potassium permanganate oxidation and reported in mg permanganate oxidizable C (POXC) kg⁻¹, serving as an estimation of the pools of soil OM that are readily decomposable and freely available to soil microorganisms, and often considered as “biologically active” [27,28]. Potentially mineralizable N (PMN) (kg ha⁻¹) was estimated by the service laboratory based on SLAN and CB, where PMN represents potential biological contributions to plant-available N. Soil health scores (SHS) were also calculated by the service laboratory, based on a model incorporating weighted CB, active C, OM, aggregate stability, and SLAN. The SHS score ranges from 0 to 50, where 0–9 represents the lowest soil health. Soil fertility scores (SFS) were then determined based on extractable P and K, potentially mineralizable N, and soil health scores. The SFS ranges from 0 to 100, where scores above 60 are considered good.

2.5. Statistical Analysis

The suite of soil health parameters were analyzed following a linear mixed model using the GLIMMIX procedure in SAS (Version 9.4; SAS Institute, Cary, NC, USA). Year was included as a fixed effect in the model, and data for the organic fertilization trials (year as a subplot factor) and the integrated nutrient management studies (year as a sub-subplot factor) were subsequently analyzed as split plot and split–split plot experiments, respectively. Natural logarithmic or square root transformations were applied to the data as necessary to meet the model assumptions including linearity, normality, and homogeneity of variance after examining residual plots and model fit statistics. Multiple comparisons were conducted using Fisher’s least significant difference test at p ≤ 0.05.

Principal component analysis (PCA) was also conducted using JMP Version 17 (SAS Institute) to reduce dimensionality among nine major soil health parameters (active C, NH₄-N, aggregate stability, CB, CEC, NO₃-N, OM, pH, and SLAN), where data were pooled over two seasons in the organic fertilizer trials. Variables were standardized (with the mean centered and scaled to unit variance) and loading plots for each fertilizer source were used to visualize patterns and highlight variable correlations within each of the two contrasting organic fertilizer sources. For the integrated nutrient management trials, canonical discriminant analysis was conducted in JMP Version 17 with backward stepwise selection (p ≤ 0.10) to examine multivariate differentiation among treatments. A biplot was presented where rays represented significant variables, and separate canonical plots with ellipses representing 95% confidence intervals were generated to visualize group dispersion, comparing each compost treatment within the weedy fallow and sunn hemp scenarios, respectively, and comparing cover crop treatments within each compost scenario. All soil physical, chemical, and biological properties measured in the study were evaluated in the discriminant analysis, except for soil health and fertility scores, as they are composite metrics that may introduce redundancy and multicollinearity.

3. Results

3.1. Soil Health Assessment in the Organic Fertilization Trials

In the organic fertilization trials, soil OM, Solvita CO₂ burst, potentially mineralizable N, Solvita labile amino-N, active C, aggregate stability, pH, CEC, NH₄-N, and overall soil health score were not significantly affected by the preplant organic fertilizer source or its interaction with year at the end of the celery production season (

Table 1. Effects of preplant organic fertilizer source and year on soil physical, chemical, and biological properties measured at the end of the organic celery production season in the organic fertilization trials.

Effect	OM i	CB	SLAN	AC	PMN	pH	CEC	NH4-N	AS	SHS	SFS
Fertilizer (F)
Everlizer ii	0.86	37.5	16.5	462.7	37.9	6.9	6.1	3.6	2.3	13.2	60 a
Nature Safe	0.82	31.7	13.5	425.5	34.3	6.6	5.6	3.2	1.8	12.4	52 b
Year (Y)
2022	0.85	39.7	16.9	470.5	39.8	6.8	6.2 a	3.4	2.7	13.2	58
2023	0.84	29.9	13.1	418.5	32.7	6.7	5.5 b	3.4	1.4	12.4	54
	p value
F	0.541	0.202	0.314	0.143	0.246	0.071	0.077	0.692	0.548	0.096	0.005
Y	0.708	0.085	0.165	0.090	0.081	0.426	0.051	0.962	0.208	0.157	0.071
F × Y	0.412	0.772	0.261	0.519	0.800	0.215	0.098	0.348	0.656	0.781	0.472

ⁱ OM = organic matter content (%); CB = Solvita CO₂-C burst (mg CO₂-C/kg/day); SLAN = Solvita labile amino-N (mg/kg); AC = active carbon (mg/kg); PMN = potentially mineralizable N (kg N/ha); CEC = cation exchange capacity (meq/100g); NH₄-N = ammonium nitrogen (mg/kg); AS = aggregate stability (%); SHS = soil health score; SFS = soil fertility score. ⁱⁱ Preplant organic fertilizer treatments included Everlizer 3N-1.3P-2.5K, a heat processed poultry litter product, and Nature Safe 10N-0.9P-6.6K, containing feather meal, meat and bone meal, blood meal, and sulfate of potash. Within a parameter, means followed by the same letter are not significantly different at p ≤ 0.05, according to Fisher’s least significant difference test.

). In contrast, the organic fertilization × year interaction showed a significant impact on soil NO₃-N (

Table 2. Effects of the interaction between preplant organic fertilizer source and year on soil nutrients and P:[Al + Fe] measured at the end of the organic celery production season in the organic fertilization trials.

Preplant Fertilizer	P iii (mg kg−1)		K (mg kg−1)		Mg (mg kg−1)		Mn (mg kg−1)		B (mg kg−1)		NO3-N (mg kg−1)		P:[Al + Fe] (%)
	2022	2023	2022	2023	2022	2023	2022	2023	2022	2023	2022	2023	2022	2023
Everlizer i	291.0 Aa	170.0 Ba	76.3 Aa	42.3 Ba	61.4 Aa	53.0 Aa	10.5 Aa	6.1 Ba	0.3 Aa	0.2 Ba	6.3 Aa	2.1 Ba	38.7 Aa	19.8 Ba
Nature Safe	210.6 Ab	159.4 Ba	33.2 Ab	31.1 Ab	26.5 Bb	37.0 Ab	5.8 Ab	5.0 Aa	0.2 Ab	0.2 Aa	4.8 Aa	2.8 Ba	27.7 Ab	18.2 Ba
	p value
F ii	0.023		<0.001		<0.001		0.017		0.010		0.600		0.007
Y	0.007		0.005		0.345		0.016		<0.001		<0.001		<0.001
F × Y	0.036		0.015		0.026		0.039		<0.001		0.041		0.017

ⁱ Preplant organic fertilizer treatments included Everlizer 3N-1.3P-2.5K, a heat processed poultry litter product, and Nature Safe 10N-0.9P-6.6K, containing feather meal, meat and bone meal, blood meal, and sulfate of potash. ⁱⁱ F = preplant organic fertilizer source; Y = year (2022 and 2023 trials). ⁱⁱⁱ P = phosphorus; K = potassium; Mg = magnesium; Mn = manganese; B = boron; NO₃-N = nitrate nitrogen; P:[Al + Fe] = proportion of phosphorus to aluminum and iron. Within a parameter, means in a column followed by the same lowercase letter and means in a row followed by the same uppercase letter are not significantly different at p ≤ 0.05, according to Fisher’s least significant difference test.

). While a significantly lower level of soil NO₃-N was observed in the 2023 trial compared to the 2022 experiment under both preplant fertilizer sources, the percent reduction between the two seasons was greater in Everlizer (by 66.7%) vs. Nature Safe (by 41.7%) (Table 2). Soil P, K, Mg, Mn, and B were also significantly affected by the interaction effect (Table 2). In the 2022 trial, soil P, Mn, and B contents were significantly higher under Everlizer relative to Nature Safe, whereas similar levels were observed in the 2023 trial. In both organic fertilizer treatments, soil P showed a significantly higher level in the 2022 trial compared to the 2023 experiment, but with a greater magnitude in Everlizer (by 71.2%) vs. Nature Safe (by 32.1%) (Table 2). Soil K and Mg contents were significantly higher under Everlizer fertilization compared to Nature Safe in both years, but to a greater degree in the 2022 trial. Soil K, Mn, and B contents in the 2022 trial were significantly higher than in the 2023 trial under Everlizer but reached a similar level between the two trials under Nature Safe (Table 2). While soil Mg showed similar levels between the two trials under Everlizer, it was significantly higher in the 2023 trial compared to the 2022 trial under Nature Safe (Table 2). Soil Zn and Cu contents were significantly greater under Everlizer vs. Nature Safe fertilization regardless of production season, and soil S, Zn, Cu, and Fe contents were significantly higher in the 2022 trial. In contrast, organic fertilizer source and year did not exhibit any impact on soil Ca (

Table 3. Effects of preplant organic fertilizer source and year on soil nutrients, base saturation, and P:[Ca + Mg] measured at the end of the organic celery production season in the organic fertilization trials.

Effect	Ca i	S	Zn	Cu	Fe	BS-Ca	BS-K	BS-Mg	P:[Ca + Mg]
Fertilizer (F)
Everlizer ii	847.4	11.4	4.1 a	2.7 a	110.0	71.4	2.4 a	8.1 a	23.7
Nature Safe	725.1	11.5	1.5 b	0.8 b	112.2	67.3	1.5 b	4.9 b	23.3
Year (Y)
2022	840.8	13.3 a	3.8 a	2.8 a	120.1 a	72.0	2.4	6.1	26.2 a
2023	730.7	9.8 b	1.6 b	0.7 b	102.8 b	66.7	1.5	6.8	21.1 b
	p value
F	0.159	0.938	0.002	<0.001	0.213	0.268	0.014	<0.001	0.828
Y	0.270	0.018	0.003	<0.001	<0.001	0.164	0.156	0.208	0.024
F × Y	0.382	0.079	0.498	0.889	0.700	0.977	0.206	0.158	0.630

ⁱ Ca = calcium (mg/kg); S = sulfur (mg/kg); Zn = zinc (mg/kg); Cu = copper (mg/kg); Fe = iron (mg/kg); K = potassium; Mg = magnesium; BS-Ca = calcium base saturation (%); BS-K = potassium base saturation (%); BS-Mg = magnesium base saturation (%); P:[Ca + Mg] = proportion of phosphorus to calcium and magnesium (%). ⁱⁱ Preplant organic fertilizer treatments included Everlizer 3N-1.3P-2.5K, a heat processed poultry litter product, and Nature Safe 10N-0.9P-6.6K, containing feather meal, meat and bone meal, blood meal, and sulfate of potash. Within a parameter, means followed by the same letter are not significantly different at p ≤ 0.05, according to Fisher’s least significant difference test.

).

Year showed a marginally significant impact on CEC, with a higher level in the 2022 trial (6.2 meq 100 g⁻¹) relative to the 2023 trial (5.5 meq 100 g⁻¹) (Table 1). In both seasons, soil K and Mg base saturation values (i.e., proportions of exchangeable K and Mg on the CEC) were significantly greater in the Everlizer treatment compared to Nature Safe (Table 3). The P:[Ca + Mg] was significantly affected by year, higher in the 2022 trial (26.2%) than in the 2023 trial (21.1%) (Table 3). P:[Al + Fe] was impacted by the significant organic fertilizer source × year interaction, where the increase by Everlizer relative to Nature Safe was only observed in the 2022 trial (Table 2). Under both fertilizer treatments, P:[Al + Fe] was lower in the 2023 trial than the former season, but with a greater reduction in Everlizer (Table 2). The soil fertility score and soil health scores generated by the service laboratory showed that Everlizer increased the soil fertility score in comparison to Nature Safe (Table 1).

Principal component analyses were performed individually for each granular organic fertilizer, projecting nine (active C, NH₄-N, aggregate stability, CB, CEC, NO₃-N, OM, pH, and SLAN) of the major physical, chemical, and biological soil health indicators onto two-component loading plots (Figure 1). The two PCAs demonstrated a similar level of variation, where principal component (PC) 1 and PC2 explained 46.3% and 22.0% of the total observed variation in soils under Nature Safe fertilization for a total of 68.3% (Figure 1A), and 38.1% and 28.1% of the total observed variation in soils under Everlizer fertilization for a total of 66.2% (Figure 1B). Under Nature Safe fertilization, soil active C, aggregate stability, CEC, OM, and CB appeared to drive overall variability, and negative loading for pH and aggregate stability along PC1 suggests these parameters correlated negatively with CB, CEC, OM, and active C. Under Everlizer, soil OM, CEC, active C, and CB appeared to similarly drive the variability, whereas aggregate stability contributed more to PC2. While SLAN, NO₃-N, and NH₄-N were highly positively correlated under Nature Safe, SLAN and NH₄-N were highly negatively correlated under Everlizer and NO₃-N was not well correlated with either SLAN or NH₄-N. Aggregate stability and pH exhibited a positive correlation under Nature Safe fertilization, but they appeared to have a negative correlation under Everlizer. CB, CEC, OM, and active C were consistently positively correlated under both organic fertilizers, while stronger associations were observed under Nature Safe fertilization vs. Everlizer.

3.2. Soil Health Assessment in the Integrated Nutrient Management Trials

In the integrated nutrient management trials, PMN and soil S and B contents were not affected by cover cropping, compost application, or year (

Table 4. Effects of cover cropping, compost application, and year on soil physical, chemical, and biological properties measured at the end of the organic celery production season in the integrated nutrient management trials.

Effect	OM i	CB	SLAN	AC	PMN	pH	CEC	NO3-N	NH4-N	AS	SHS
Cover cropping (Cv)
Sunn hemp	0.91	31.3	16.0	443.6	34.3	6.6	6.1	- ii	6.4 a	2.7	12.9
Weedy fallow	0.92	31.4	14.9	433.9	33.9	6.5	6.0	-	4.8 b	2.4	12.8
Compost (Cp)
Control iii	0.91	- ii	14.9	420.6 b	35.3	6.5	-	-	6.0	-	12.4 b
Vermi	0.91	-	18.5	436.5 ab	33.5	6.5	-	-	5.4	-	12.9 a
Yard	0.91	-	15.2	466.2 a	32.5	6.6	-	-	6.0	-	12.9 a
Mixed	0.93	-	13.4	432.8 b	35.1	6.6	-	-	5.0	-	13.1 a
Year (Y)
2022	0.87 b	- ii	15.2	396.5 b	33.9	6.8 a	-	-	7.2 a	-	12.6 b
2023	0.96 a	-	15.7	485.1 a	34.4	6.2 b	-	-	4.3 b	-	13.1 a
	p value
Cv	0.313	0.952	0.407	0.382	0.726	0.058	0.421	0.495	0.009	0.336	0.445
Cp	0.650	0.322	0.071	0.052	0.219	0.102	0.083	0.251	0.646	0.038	0.006
Y	<0.001	0.487	0.688	<0.001	0.625	<0.001	<0.001	0.839	<0.001	0.325	<0.001
Cv × Cp	0.226	0.097	0.275	0.321	0.321	0.900	0.194	0.048	0.608	0.204	0.492
Cv × Y	0.642	0.897	0.819	0.065	0.596	0.555	0.072	0.004	0.189	0.057	0.445
Cp × Y	0.157	0.005	0.753	0.380	0.140	0.321	<0.001	0.886	0.297	0.036	0.387
Cv × Cp × Y	0.554	0.176	0.359	0.867	0.302	0.736	0.871	0.717	0.798	0.204	0.604

ⁱ OM = organic matter content (%); CB = Solvita CO₂-C burst (mg CO₂-C/kg/day); SLAN = Solvita labile amino-N (mg/kg); AC = active carbon (mg/kg); PMN = potentially mineralizable N (kg/ha); CEC = cation exchange capacity (meq/100g); NO₃-N = nitrate nitrogen (mg/kg); NH₄-N = ammonium nitrogen (mg/kg); AS = aggregate stability (%); SHS = soil health score. ⁱⁱ dashes indicate multiple comparisons were not performed on main effects due to a significant two-way interaction. ⁱⁱⁱ Control = no compost control; Vermi = vermicompost (11.2 t ha⁻¹); Yard = yard waste compost (22.4 t ha⁻¹); Mixed = a 1:1 gravimetric mix of yard waste compost and vermicompost (22.4 t ha⁻¹). Within a parameter, means followed by the same letter are not significantly different at p ≤ 0.05, according to Fisher’s least significant difference test.

and

Table 5. Effects of cover cropping, compost application, and year on soil nutrients, base saturation, and P:[Ca + Mg] measured at the end of the organic celery production season in the integrated nutrient management trials.

Effect	Ca i	Mg	S	B	Zn	Cu	Mn	Fe	BS-Ca	BS-Mg	P:[Ca + Mg]
Cover cropping (Cv)
Sunn hemp	806.5	- ii	10.8	0.2	1.3	0.8	5.9	112.3	65.9	-	25.8
Weedy fallow	785.6	-	10.8	0.2	1.4	0.8	6.3	113.6	64.4	-	25.5
Compost (Cp)
Control iii	- ii	-	10.6	0.2	1.3	0.8	6.1	115.3	65.3 a	-	26.9 a
Vermi	-	-	10.7	0.2	1.3	0.7	6.0	112.8	62.4 b	-	28.2 a
Yard	-	-	10.4	0.2	1.4	0.8	6.4	111.2	67.2 a	-	23.2 b
Mixed	-	-	11.6	0.2	1.4	0.9	6.0	112.4	65.7 a	-	24.4 b
Year (Y)
2022	- ii	37.7	10.5	0.2	1.5 a	1.3 a	5.1 b	130.3 a	71.2 a	5.5	24.2 b
2023	-	40.1	11.2	0.2	1.2 b	0.4 b	7.4 a	97.8 b	59.6 b	5.2	27.1 a
	p value
Cv	0.462	0.906	0.984	0.528	0.139	0.679	0.090	0.492	0.268	0.833	0.708
Cp	0.019	0.680	0.381	0.887	0.743	0.616	0.655	0.140	0.010	0.429	<0.001
Y	0.007	0.190	0.166	0.263	<0.001	<0.001	<0.001	<0.001	<0.001	0.237	<0.001
Cv × Cp	0.418	0.031	0.454	0.752	0.198	0.568	0.412	0.129	0.960	0.006	0.836
Cv × Y	0.847	0.302	0.608	0.928	0.782	0.953	0.500	0.795	0.401	0.639	0.583
Cp × Y	0.011	0.213	0.312	0.917	0.074	0.658	0.881	0.877	0.262	0.297	0.559
Cv × Cp × Y	0.655	0.542	0.206	0.879	0.527	0.734	0.940	0.446	0.168	0.131	0.739

ⁱ Ca = calcium (mg/kg); Mg = magnesium (mg/kg); S = sulfur (mg/kg); B = boron (mg/kg); Zn = zinc (mg/kg); Cu = copper (mg/kg); Mn = manganese (mg/kg); Fe = iron (mg/kg); BS-Ca = calcium base saturation (%); BS-Mg = magnesium base saturation (%); P:[Ca + Mg] = proportion of phosphorus to calcium and magnesium (%). ⁱⁱ dashes indicate multiple comparisons were not performed on main effects due to a significant two-way interaction. ⁱⁱⁱ Control = no compost control; Vermi = vermicompost (11.2 t ha⁻¹); Yard = yard waste compost (22.4 t ha⁻¹); Mixed = a 1:1 gravimetric mix of yard waste compost and vermicompost (22.4 t ha⁻¹). Within a parameter, means followed by the same letter are not significantly different at p ≤ 0.05, according to Fisher’s least significant difference test.

). Soil NH₄-N content was significantly affected by cover cropping, where sunn hemp increased soil NH₄-N contents by 33.3% compared to the weedy fallow control (Table 4). Soil NH₄-N was also affected by year, with a higher level in the 2022 trial relative to the 2023 trial (Table 4).

The BS-Ca was significantly affected by compost application and year, where Vermi reduced the proportion of exchangeable Ca on the CEC compared to Control and other compost treatments, and the BS-Ca was greater in the 2022 trial (71.2%) compared to the 2023 trial (59.6%) (Table 5). P:[Ca + Mg] was also affected by compost application and year, where significantly lower values were found in Yard and Mixed compared to Control and Vermi, and P:[Ca + Mg] was higher in the second season (Table 5). Additionally, compost application and year impacted the overall soil health score, with higher levels in all compost treatments compared to Control, and greater levels in the 2023 trial (13.1) relative to the 2022 trial (12.6) (Table 4).

Soil pH was significantly affected by year, where pH after the end-of-season soil sampling was 6.8 in the 2022 trial and 6.2 in the 2023 trial (Table 4). Soil OM and active C contents also varied significantly with year, with higher levels in the 2023 trial (0.96% and 485.1 mg/kg) relatively to the 2022 trial (0.87% and 396.5 mg/kg) (Table 4). Moreover, soil Zn, Cu, Mn, and Fe contents were significantly impacted by year, with higher levels of Cu, Zn, and Fe in the 2022 trial but a greater level of Mn in the 2023 trial (Table 5).

The cover cropping × compost application interaction exhibited a significant effect on soil NO₃-N (

Table 6. Effects of the interaction between compost and year and the interaction between compost and cover cropping on soil physical, chemical, and biological properties measured at the end of the organic celery production season in the integrated nutrient management trials.

Effect	Solvita CO2 Burst (mg CO2-C/kg/day)		CEC iii (meq/100 g)		Aggregate Stability (%)		Ca (mg/kg)		Mg (mg/kg)		Base Saturation -Mg (%)		NO3-N (mg/kg)
	2022	2023	2022	2023	2022	2023	2022	2023	SH i	WF	SH	WF	SH	WF
Control ii	33.1 Aab	33.4 Aa	5.4 Bb	6.7 Aa	2.5 Aab	2.1 Aab	754.5 Ab	803.0 Aa	39.2 Aab	35.7 Aa	5.3 Aab	4.7 Ab	2.0 Aab	1.7 Aa
Vermi	28.5 Ab	31.7 Aa	5.8 Aa	6.0 Ab	2.0 Ab	3.3 Aa	788.9 Aab	738.4 Aa	33.2 Bb	44.1 Aa	4.6 Bb	6.2 Aa	2.5 Aa	2.0 Aa
Yard	26.9 Bb	34.0 Aa	5.9 Ba	6.5 Aa	2.9 Aab	1.6 Bb	898.0 Aa	788.2 Ba	41.2 Aa	38.2 Aa	5.6 Aa	5.2 Ab	1.7 Ab	2.3 Aa
Mixed	35.8 Aa	28.5 Ba	5.9 Aa	6.2 Ab	4.0 Aa	3.2 Aa	850.4 Aab	758.6 Ba	43.3 Aa	37.8 Aa	6.0 Aa	5.1 Ab	2.0 Aab	1.8 Aa

ⁱ SH = sunn hemp; WF = weedy fallow. ⁱⁱ Control = no compost control; Vermi = vermicompost (11.2 t ha⁻¹); Yard = yard waste compost (22.4 t ha⁻¹); Mixed = a 1:1 gravimetric mix of yard waste compost and vermicompost (22.4 t ha⁻¹). ⁱⁱⁱ CEC = cation exchange capacity; Ca = calcium; Mg = magnesium; NO₃-N = nitrate nitrogen. Within a parameter, means in a column followed by the same lowercase letter and means in a row followed by the same uppercase letter are not significantly different at p ≤ 0.05, according to Fisher’s least significant difference test.

). In the weedy fallow plots, soil NO₃-N was similar across compost treatments, but in the sunn hemp plots, Yard significantly reduced soil NO₃-N by 32.0% compared to Vermi. Soil NO₃-N was also affected by the significant cover cropping × year interaction (Table 4), where sunn hemp increased soil NO₃-N (2.3 mg kg⁻¹) compared to the weedy fallow (1.7 mg kg⁻¹) in the second production season but not in the first season (1.9 mg kg⁻¹ vs. 2.2 mg kg⁻¹ under sunn hemp and weedy fallow, respectively). While soil NO₃-N content was significantly higher in the 2023 trial than in the 2022 trial under sunn hemp, the opposite was observed under weedy fallow.

The cover cropping × compost interaction also significantly impacted soil Mg content and BS-Mg (Table 5). Under weedy fallow management, no significant differences were observed among compost treatments. In contrast, Vermi resulted in significantly lower soil Mg content compared to Yard and Mixed when the sunn hemp cover crop was planted (Table 6). Vermi also led to lower soil Mg content in the sunn hemp plots, compared to the weedy fallow control. As for soil BS-Mg, it was higher in Vermi compared to Control and other compost treatments under weedy fallow; however, a lower level was observed in Vermi relative to Yard and Mixed under sunn hemp, while similar levels were found between Vermi and Control. Moreover, the difference between sunn hemp and weedy fallow was only apparent under Vermi, with a lower BS-Mg in sunn hemp vs. weedy fallow (Table 6).

The significant compost × year interaction was detected in the measurements of soil aggregate stability, CB, CEC, and Ca content (Table 6). In the 2022 trial, aggregate stability in Mixed was significantly greater than Vermi, but similar to Yard and Control. In the 2023 trial, Mixed and Vermi increased aggregate stability relative to Yard, but did not differ significantly from Control. Within compost treatments, soil aggregate stability was similar between years, except under Yard, where aggregate stability was lower in the 2023 trial compared to the 2022 trial. Vermi and Yard resulted in lower CB compared to Mixed in the 2022 trial, whereas similar levels among all compost treatments were found in the 2023 trial. Moreover, Yard led to higher CB in the second season, while Mixed showed higher CB in the first season (Table 6). All three compost treatments led to significantly higher CEC values compared to Control in the 2022 trial, but in the 2023 trial, Vermi and Mixed reduced CEC compared to Control (Table 6). The CEC values in Vermi and Mixed were similar between the two seasons, but they were significantly higher in the second season for Yard (by 10.2%) and Control (by 24.1%). In the 2022 trial, Yard demonstrated higher soil Ca content than Control, while no differences among compost treatments were found in the 2023 trial (Table 6).

Soil P and K contents as well as soil K base saturation and P:[Al + Fe] were significantly affected by the cover cropping × compost application × year three-way interaction (

Table 7. Effects of the three-way interaction among cover cropping, compost application, and year on soil P and K contents, K base saturation, P:[Al + Fe], and soil fertility score at the end of the organic celery production season in the integrated nutrient management trials.

	P iii (mg kg−1)				K (mg kg−1)				Base Saturation-K (%)				P:[Al + Fe] (%)				Soil Fertility Score
	2022		2023		2022		2023		2022		2023		2022		2023		2022		2023
Compost	SH i	WF	SH	WF	SH	WF	SH	WF	SH	WF	SH	WF	SH	WF	SH	WF	SH	WF	SH	WF
Control ii	207.7 Aa	200.2 Aa	246.2 Aa	225.3 Aa	46.9 Aa	43.2 Ab	42.1 Aa	41.0 Ab	2.2 Aa	2.1 Ab	1.6 Aa	1.6 Ab	25.7 Aa	23.7 Aa	25.8 Aa	24.8 Aa	56 Aa	55 Aa	55 Aa	56 Aab
Vermi	203.5 Aa	207.2 Aa	262.6 Aa	208.7 Ba	51.8 Aa	54.0 Aa	33.1 Ab	31.9 Ac	2.4 Aa	2.8 Aa	1.4 Aa	1.4 Ab	26.0 Aa	24.4 Aa	28.6 Aa	23.2 Ba	57 Aa	56 Aa	54 Aa	51 Ab
Yard	208.2 Aa	187.8 Aa	192.9 Bb	230.8 Aa	51.0 Aa	35.6 Bb	34.2 Bb	58.0 Aa	2.2 Aa	1.4 Bc	1.3 Ba	2.4 Aa	27.2 Aa	23.5 Aa	21.0 Bb	24.8 Aa	56 Aa	52 Bb	54 Ba	60 Aa
Mixed	216.8 Aa	215.7 Aa	210.4 Ab	224.9 Aa	48.9 Aa	52.6 Aa	41.3 Aab	33.5 Abc	2.2 Aa	2.3 Ab	1.6 Aa	1.3 Ab	27.5 Aa	25.1 Aa	21.1 Ab	24.4 Aa	58 Aa	58 Aa	54 Aa	53 Ab
	p value
Cv	0.654				0.804				0.634				0.394				0.964
Cp	0.118				0.760				0.765				0.396				0.595
Y	<0.001				<0.001				<0.001				0.048				0.013
Cv × Cp	0.105				0.503				0.416				0.127				0.653
Cv × Y	0.749				0.059				0.011				0.034				0.150
Cp × Y	0.113				<0.001				<0.001				0.042				0.001
Cv × Cp × Y	<0.001				<0.001				<0.001				0.018				0.018

ⁱ SH = sunn hemp; WF = weedy fallow. ⁱⁱ Control = no compost control; Vermi = vermicompost (11.2 t ha⁻¹); Yard = yard waste compost (22.4 t ha⁻¹); Mixed = a 1:1 gravimetric mix of yard waste compost and vermicompost (22.4 t ha⁻¹). ⁱⁱⁱ P = phosphorus; K = potassium; P:[Al + Fe] = proportion of phosphorus to aluminum and iron. Within a parameter, means in a column followed by the same lowercase letter and means in a row followed by the same uppercase letter are not significantly different at p ≤ 0.05, according to Fisher’s least significant difference test.

). In the first season, there were no differences in soil P content in response to cover cropping or compost application. In the second season, Yard and Mixed resulted in lower extractable soil P contents compared to Control and Vermi within the sunn hemp treatment; however, no difference was observed under weedy fallow. In the first season, soil K content did not vary among compost treatments within the sunn hemp treatment, while Vermi and Mixed led to higher soil K contents in comparison with Control and Yard under weedy fallow. In the second season, Vermi and Yard resulted in lower soil K contents than Control under sunn hemp, while Yard significantly increased soil K content compared to Control and other compost treatments under weedy fallow (Table 7).

While there was a lack of compost effects in the sunn hemp treatment in either season on soil K base saturation, compost differences were observed under weedy fallow in both seasons (Table 7). In the 2022 trial, Vermi showed significantly higher BS-K and Yard led to significantly lower BS-K, in comparison with Control and Mixed. In contrast, Yard resulted in significantly higher BS-K than all other treatments in the 2023 trial. In the 2022 trial, no differences in P:[Al + Fe] were observed among treatments; however, in the 2023 trial, Yard and Mixed exhibited a significantly lower P:[Al + Fe] compared to Control and Vermi under sunn hemp, with no difference observed under weedy fallow (Table 7). With respect to the overall soil fertility score, compost treatments did not differ significantly under sunn hemp in either trial; however, significant compost effects were detected in the weedy fallow plots (Table 7). In the 2022 trial, Yard resulted in a lower soil fertility score than Control and other compost treatments. In contrast, Yard led to a higher soil fertility score than Vermi and Mixed but was similar to Control in the 2023 trial.

The canonical discriminant analysis was also performed to help identify multivariate differentiation among compost and cover crop treatments regarding their impacts on soil health parameters, with canonical plots paneled to compare compost treatments within the sunn hemp and weedy fallow scenarios, respectively (Figure 2A–C), and to compare sunn hemp vs. weedy fallow within each compost treatment (Figure 2D–G). Overall, the first and second canonical functions explained 41.5% and 24.4% of the total variation, respectively (Figure 2). Soil parameters positively correlated with the first canonical function included soil P, aggregate stability, Zn, and active C, while Al, BS-Ca, CEC, and Ca were negatively correlated with the first canonical function. Meanwhile, Ca, active C, and aggregate stability were positively correlated with the second canonical function, and BS-Ca, CEC, Al, Zn, and P were negatively correlated with the second canonical function (Figure 2B) (all associations in order of strongest to weakest correlations).

Under sunn hemp, Yard was separated from Vermi along the first canonical function, whereas Mixed overlapped with the two individual compost treatments. While the 95% confidence ellipse for Control partially overlapped with all compost treatments, there was partial separation between Control and Yard and between Control and Vermi along the second canonical function (Figure 2A). Soil P, Al, and aggregate stability were the key soil indicators differentiating Yard and Vermi along the first canonical function, and Ca, BS-Ca, and CEC were key attributes contributing to the partial separation of Yard and Vermi from Control along the second canonical function (Figure 2B). Under weedy fallow, there was less separation observed among compost treatments. The 95% confidence ellipses of all compost treatments largely overlapped, though Control appeared to partially separate from the other compost treatments along the second canonical function (Figure 2C), influenced by the negative loading of BS-Ca, CEC, and Zn (Figure 2B). Among all compost treatments, Vermi resulted in the largest magnitude of separation between sunn hemp and weedy fallow cover crop treatments, though all comparisons exhibited partial overlap in 95% confidence intervals (Figure 2D–G). Under Vermi, sunn hemp was partially differentiated from weedy fallow along the first and second canonical functions, driven by P, Al, and aggregate stability along the first canonical function, and Ca, BS-Ca, and CEC along the second canonical function.

4. Discussion

4.1. Soil Organic Matter, Aggregate Stability, and Active Carbon

As cohesive groups of soil particles integral to soil structure, soil aggregates impact the distribution and arrangement of soil particles and therefore water infiltration and permeability, and are bound by increased hyphal biomass, soil bacteria and polysaccharides, and other organic material [29,30,31]. Considering the coarse texture of sandy soils and the low baseline soil OM contents at the experimental site, it is not surprising to see the lack of a nutrient management program impact on soil OM at the end of each organic celery production season in this study. The significant effect of year on soil OM in the integrated nutrient management trial may be partly explained by the higher baseline OM content in the 2023 trial compared to the 2022 trial. Yilmaz and Sönmez [32] found aggregate stability increased from 7.8% to 46.5% after 90 days of monthly additions of organic fertilizers and vermicompost products to a clay loam soil, with the greatest benefits observed when amended with arbuscular mycorrhizal fungi. While the two granular organic fertilizers did not impact aggregate stability differentially in the organic fertilizer trials, and composts did not significantly improve aggregate stability in the integrated nutrient management trials, the present study in sandy soils with low levels of OM and soil aggregation highlights a need for longer-term studies under sandy soil conditions and across different environmental conditions and diverse organic cropping systems to better understand limitations and the potential for fertilizer selection and integrated nutrient management approaches to improve soil structure.

The volumetric aggregate stability test (VAST) was used to assess aggregate stability in our study, which focuses on macroaggregates greater than 2 mm. While studies in clay soils with higher OM contents have shown the majority of soil aggregates are greater than 1 mm [33], research indicates that aggregate fractions greater than 1 mm are minimal under Florida’s sandy soil conditions [34]. Aggregate stability in the range of 20–50% using VAST has been reported for soils containing 20–46% sand [35], but research-based information from sandy soils is rather limited. Our findings point out the challenges of macroaggregation in sandy soil conditions and highlight the need for more detailed categorization of soil aggregate fractions. While aggregate stability showed some positive response to compost application, particularly Mixed in this study, aggregate stability was still exceedingly low.

Soil active C reflects the labile fraction of the SOC pool. Permanganate oxidizable C (POXC) is considered responsive to changes in management practices and has been linked to practices that increase C sequestration and build soil OM in the long term [36]. Management practices that promote microbiological diversity and develop SOC pools are generally thought to increase active C, promote soil aggregation, and build OM content [37]. The addition of cow manure compost in a five-year study showed increases in SOC (by 27–87%) and concurrent decreases in bulk density (by 3–6%), parameters inherently related to soil aggregation, in an alfisol (1.3% OM) under a wheat–maize rotation [38]. Cover crop residues have been shown to improve soil aggregation and alter the distribution of organic C among soil aggregates and across soil profiles [39]. In a six-year study in a ferrasol (15% sand, 21% silt, 64% clay), the impact of incorporating various leguminous and non-leguminous cover crop residues (with tillage), including white mustard (Sinapis alba L.), lacy phacelia (Phacelia tanacetifolia Benth.), lopsided oat (Avena strigosa Schreb.), and Egyptian clover (Trifolium alexandrinum L.), was shown to increase SOC at 15–30 cm soil depths after 11 crop rotations between cover crops, maize, and soybean, whereas improvements in soil aggregation were limited [40].

Labile fractions of SOC and soil microbial biomass have exhibited positive responses to the addition of cover crop residues in the medium term (six to eight years) [41], and annual cover crop rotations in loamy sand soils (about 8% clay content) enhanced POXC more than compost addition in an organic vegetable production system, despite the greater relative contribution of compost toward stocks of SOC [42]. However, these multi-year studies did not examine potential synergy between cover crop residues and compost application on SOC and active C. In an 18-month compost incubation study examining biowaste, yard waste, and manure-based composts, all of the composts consistently increased proportions of alkyl and aromatic C in soil OM with little influence on labile fractions of O/N-alkyl C, supporting the idea that composts are primarily contributing stabilized C recalcitrant to microbial decomposition [43]. In the present six-month study (average of a 68-day cover crop followed by a 16-week celery crop in each season) that was repeated across two production seasons, the short-term impacts of cover crop residues on active C were not detected. Since the field trial in our study did not involve multiple cover crop genotypes or celery cultivars, future research can include diverse cover crop and vegetable crop treatments to better examine plant–soil interactions that impact soil health management and improvement.

Interestingly, Yard increased active C compared to Mixed and Control, suggesting that certain composts may induce more immediate changes to pools of labile SOC and stimulate microbial activities. Compost compositional characteristics and feedstock-specific interactions likely drive these differences, where the higher C:N ratio of the yard waste compost (32:1 and 23:1 in 2022 and 2023 trials, respectively) could have promoted a more persistent supply of SOC compared to the vermicompost (12:1 C:N in both 2022 and 2023 trials) and the mixed compost (22:1 and 17.5:1 C:N). Finished composts typically have a C:N ratio of about 18:1 [44], and even with similar C:N ratios, composts of varying feedstocks contain distinct levels of humic acid and humic acid:fulvic acid ratios, impacting their recalcitrance. On the other hand, cover crop residues, particularly legumes with relatively low C:N ratios, are expected to primarily contribute labile C to pools of SOC [45]. The observed differences in active C may reflect not only differences in compost composition and C:N ratios, but also nutrient release kinetics and associated shifts in soil microbial activity. Nevertheless, more research is needed to link compost characteristics to broader changes in soil health and crop productivity. Examining the influence of repeated applications of composts and the incorporation of cover crop residues on specific aspects of soil health including POXC in sandy soils is also warranted, considering unique challenges posed in the southeastern US.

4.2. Soil Nutrient Composition and Chemical Indicators of Soil Health

A wide array of plant- and animal-based inputs, as well as commercial formulations and blended products, can be used as organic fertilizer sources to supply plant-available N and other macro- and micronutrients. In the organic fertilization trials, soil NO₃-N content at the end of the organic celery season was lower in the 2023 vs. 2022 trial, and the difference between the two seasons was greater with Everlizer despite similar levels of soil NO₃-N between the two organic fertilizer treatments. Throughout the 2023 trial, precipitation exceeded 38 cm at the experimental site, compared to only 15 cm of total precipitation in the 2022 trial, possibly leading to a greater risk of nitrate leaching. The N mineralization pattern of the two types of organic fertilizers as related to crop demand needs to be examined to better understand the soil N dynamics. Interestingly, in the integrated nutrient management trials, sunn hemp led to a higher level of soil NO₃-N relative to the weedy fallow in the 2023 trial, although no difference was detected in the first season. Differences in soil NO₃-N content between the two production seasons were likely driven by differences in cover crop biomass accumulation and distribution, where the higher sunn hemp biomass and greater abundance of lignified stem tissue in the second season may have contributed more persistently to pools of soil N. This seasonal variation was not detected in regard to soil NH₄-N content, which was consistently higher under sunn hemp vs. weedy fallow across two seasons. It also needs to be pointed out that soil NH₄-N content was higher (4.3–7.2 mg kg⁻¹) than the minimal remaining soil NO₃-N content (1.7–2.3 mg kg⁻¹) at the end of the celery production season in the integrated nutrient management trials. This may be associated with seasonal precipitation events, crop uptake, and termination of fertigation for about 14 days prior to celery harvest which tended to diminish remaining pools of soil NO₃-N [15]. The other intriguing finding is that the impact of compost on soil NO₃-N content was dependent on cover cropping. No differences in soil NO₃-N content were observed among compost treatments under weedy fallow management, while under the sunn hemp summer cover crop, higher soil NO₃-N contents were detected in Vermi vs. Yard. While the sunn hemp residues had an overall C:N ratio ranging from 18.2:1 to 27.3:1 in this study, the more persistent, lignified stem tissue (24.6:1 and 39.3:1 C:N ratios in 2022 and 2023 trials, respectively) may have promoted N immobilization, particularly when combined with the yard waste compost containing higher C:N ratios.

Leguminous cover crop residues can undoubtedly contribute N to agricultural production systems; however, rapid cover crop decomposition, asynchrony of N mineralization with crop uptake, and cover crop incorporation outside of cash cropping beds tends to reduce the feasibility of relying on cover crop N contributions to support vegetable crop growth and development in organic systems [46]. In previous studies monitoring mineral N release from sunn hemp residues in sandy soils, N mineralization occurred early in the production season, and late season differences between sunn hemp-amended soils and the weedy fallow control were not apparent [15]. As soil health assessments were conducted after the final harvest of organic celery in both seasons in the present study, limited impacts of sunn hemp residues on mineral N were expected. Stable and mature composts, including both the yard waste compost and the vermicompost used in this study, have been shown to mineralize less than 10% of total N over the entire production season [47], and composts with higher C:N ratios (>25:1) have been demonstrated to increase the rate of N immobilization [48]. Hence, more systematic research is needed to better understand N mineralization dynamics in sandy soils under integrated nutrient management practices across multiple locations and diverse environmental conditions, and to link these dynamics to broader implications for soil health.

Managing soil P availability can be challenging in both organic and conventional production systems, due to the adsorption of P to Fe and Al hydroxides in acidic soils, and to Ca in calcareous soils, impacting P availability to plants and limiting the eventual uptake of orthophosphates [49]. Cover crops have been shown to reduce P losses in soils by minimizing soil erosion and P runoff, as well as recycling legacy pools of existing soil P [50]. Composts can supply organic and mineral sources of P, and may stimulate microbial activity driving the mineralization of organic P. In the integrated nutrient management trials, the three-way interaction among year, cover cropping, and compost application exhibited pronounced impacts on Mehlich-III extractable P. While no cover crop or compost effects were found in the 2022 trial, the 2023 trial demonstrated that under Yard, sunn hemp led to lower Mehlich-III extractable P than weedy fallow, and the opposite was observed under Vermi. The interaction effects of cover crop, compost, and seasonal environmental conditions may be largely attributed to the different C:N ratios and nutrient contents of compost products, together with the discrepancy in sunn hemp dry biomass accumulation between the two seasons (4640 kg/ha in 2022 and 6630 kg/ha in 2023).

The ratio of P:[Al + Fe] could be used as an indicator of P saturation [51], and according to Waters Agricultural Laboratories, levels below 5% may indicate P fixation. Similar to soil P, P:[Al + Fe] was affected by the three-way interaction among year, cover cropping, and compost application, while the lack of significant cover crop or compost effects on soil Al and Fe content suggests that the Mehlich-III extractable P played a major role in driving differences in the ratio of P:[Al + Fe]. In general, P:[Al + Fe] and P:[Ca + Mg] ranged from 18.2 to 38.7% in the present study, indicating a limited concern associated with P fixation, and in line with the high level of Mehlich-III extractable P detected. Soil Mg content was lower under Vermi compared to Yard, in accordance with the higher Mg content in the yard waste compost (1400–1600 mg kg⁻¹ compost based on fresh weight) vs. the vermicompost (800 mg kg⁻¹ based on fresh weight). It is also important to note that Vermi was applied at a lower rate than Yard. Additionally, soil Mg content was reduced under sunn hemp vs. weedy fallow in response to Vermi application, where more rapid cover crop decomposition stimulated by a N-rich compost may have impacted microbial Mg immobilization, and bivalent metals including Mg may have resulted in the formation of complexes with organic ligands such as humic and fulvic acids from composts [52]. Further, Yard and Mixed consistently reduced P:[Ca + Mg], possibly associated with the temporary increase in P immobilization by Yard with higher C:N ratios [53] as well as the higher Ca content in the yard waste compost (1500–2100 mg kg⁻¹ compost based on fresh weight) compared to Vermi (800–1000 mg kg⁻¹ compost based on fresh weight). In the 2022 trial, Yard application increased soil Ca content compared to Control, likely due to the higher Ca content of the yard waste compost used in the first season, i.e., a 42.1% increase relative to the second season. The interaction of cover crop residues and compost application on soil phosphatase activities and microbial P [54] also deserves further examination.

In the organic fertilization trials, significantly higher Mehlich-III extractable P was observed under Everlizer vs. Nature Safe in the 2022 trial. Given that no additional P was applied due to the very high baseline soil P content (77–130 mg kg⁻¹), this result driven by the higher content of P in Everlizer is expected. Likewise, the P:[Al + Fe] was increased under Everlizer compared to Nature Safe in the first season, likely associated with the increased level of P since soil Al or Fe contents did not show differential responses to the two organic fertilizers. Interestingly, soil P content and P:[Al + Fe] in the second season were comparable between the two organic fertilizers, which may be partly ascribed to heavier precipitation in the second season leading to greater P losses through leaching and runoff. Overall, the soil P content well exceeds the 45 mg kg⁻¹ threshold for high soil P [55], which may indicate the need for planting cover crops or other catch crops following the celery crop to minimize nutrient losses in the off-season. Organic fertilizer selection also needs to be carefully considered toward nutrient stewardship, as applying high-P organic fertilizers to meet crop N demands may not be suitable for sandy soils already rich in P.

In the organic fertilizer trials, Everlizer significantly increased soil K content compared to Nature Safe in both seasons, despite the addition of sulfate of potash to balance K supply in Nature Safe. The K⁺ in the soil solution can be readily leached in sandy soils with low clay content, particularly in soils with high concentrations of Ca²⁺ [56], while the higher content of K⁺ in the heat processed poultry litter-based Everlizer may be retained for longer, potentially reducing soil K losses. Meanwhile, according to fertilizer testing results, Everlizer also contained more K than indicated in the guaranteed analysis (3% K₂O), with approximately 5% K₂O. Soil Mg, Mn, B, Zn, and Cu contents were also generally higher under Everlizer compared to Nature Safe, and accordingly increased the BS-K and BS-Mg. These differences in soil minerals highlight how poultry litter-based organic fertilizers may differ from organic fertilizers containing feather meal and bloodmeal in terms of supplying essential macro- and micronutrients, suggesting the importance of characterizing nutrient composition and availability of organic fertilizer products derived from different nutrient sources.

In the integrated nutrient management trials, soil K content exhibited inconsistent responses to cover crop residues and compost application across the two seasons. Longer-term studies are necessary to better understand seasonal and legacy impacts of cover cropping and compost application on soil K⁺ retention. Yard improved soil Zn content, in accordance with the higher level of Zn (54 mg kg⁻¹) in the yard waste compost than that of the vermicompost (45 mg kg⁻¹), demonstrating the role that composts can play in supplying micronutrients in organic production systems. The compost impact on soil CEC also varied by year, where all composts increased CEC compared to the no compost control in the 2022 trial, but in the 2023 trial, Vermi and Mixed reduced the overall CEC. The addition of OM generally improves CEC, although changes in the formation of stable organic complexes and the adsorption of humic substances to soil mineral surfaces may explain reductions in CEC observed in the second season. Notably, the substantially higher precipitation during the 2023 trial may have contributed to greater leaching of exchangeable cations in the compost-amended soils. This is a particularly important consideration in the treatments containing Vermi, which has a finer texture and lower base cation contents. Specifically, Vermi reduced BS-Ca compared to all other compost treatments regardless of cover cropping, and it also lowered BS-Mg relative to Yard and Mixed in sunn hemp-amended soils. The findings aligned with higher levels of Ca (by 118%) and Mg (by 88%) in the yard waste compost relative to the vermicompost. As cover crops and composts are increasingly used by organic growers in Florida to overcome water and nutrient management challenges in sandy soils, more in-depth research is needed to elucidate the specific roles of integrated soil-building practices in enhancing overall CEC and soil nutrient retention and availability in both the short and long term.

4.3. Soil CO₂ Burst and Solvita Labile Amino-N

Biological indicators including CB and SLAN can provide more integrative insights into the health and sustainability of cropping systems and are responsive to changes in management to project longer-term improvements to soil aggregation and OM [57,58]. In a turfgrass study conducted on a sandy loam soil, CB and SLAN were found to be well-correlated and could be associated with soil N mineralization potential [26], though others have argued that single soil biological indicators do not comprehensively represent soil biochemical health attributes [59]. In the present study, there was a lack of clear correlations between CB and SLAN, and CB and SLAN were not affected by the organic fertilizer treatment or sunn hemp cover cropping at the end of the organic celery production season. In the integrated nutrient management trials, no consistent compost impact on CB was found over the two seasons except that in the first season, CB was enhanced in Mixed compared to Yard or Vermi. CB relies on re-wetting dried soils to capture immediate flushes in CO₂, which might be more predominantly linked to the impacts of labile C. Compost application also tended to affect SLAN, where Vermi led to the highest level of SLAN among other compost treatments at the end of the organic celery production season. The relatively higher N content in the vermicompost (900 mg kg⁻¹ compost based on fresh weight) compared to the yard waste compost (600 mg kg⁻¹ compost based on fresh weight) might have contributed to the increased level of labile amino-N, a pool of potentially available N in the longer term, despite both compost treatments showing negligible amounts of mineral N at soil sampling after celery harvest. Interestingly, the mixed compost, with an equal quantity of Vermi, did not exhibit similar SLAN, highlighting how the intrinsic composition and C:N balance of composts may impact soil N dynamics and exert broader impacts on soil health. While existing research measuring soil CB and SLAN are quite limited, particularly in integrated and organic systems, the impacts of organic fertilizers and composts have been shown to improve soil microbial activities [60]. Moreover, cumulative improvements in soil respiration over a three-year period in response to annual amendments with 15–45 ton ha⁻¹ of a food waste and yard waste-based compost have been reported [61]. Cover crops have been shown to impact CB and SLAN, while inconsistent results were found between years [62]. Hence, more comprehensive cover crop studies in combination with other soil amendments are warranted to further understand relationships between cover crop species, repeated cover cropping cycles, and integrated soil and nutrient management practices as well as environmental conditions on soil biological parameters.

4.4. Overall Soil Health and Fertility Scoring, and Multivariate Approaches for Soil Health Assessment

Soil health and fertility scores have been developed by commercial service labs to offer a benchmark tool that serves as a “whole picture” view of soil parameters. In this study, they helped to visualize how organic fertilizers can differ in their immediate supply of nutrients while also differentially impacting soil health in the longer term. Everlizer, the heat processed poultry litter, exhibited potential to improve soil health through continued application with its higher C:N (12:1) resulting in continued additions of SOC, compared to Nature Safe (4:1 for C:N). Immediate soil fertility was also enhanced under Everlizer, as shown by the enhanced CEC, and increases in soil NO₃-N and various macro- and micronutrients. In the integrated nutrient management trials, all composts similarly enhanced soil health as reflected by the soil health score, while the compost impact on soil fertility score was dependent on the production season and sunn hemp cover cropping. Overall, soil health scores across organic fertilizers and cover crop and compost treatments were generally considered low to medium, according to the service lab, further illustrating the challenges of maintaining and improving soil health in sandy soils containing low intrinsic levels of OM.

Principal component analysis with loading plots for each organic fertilizer provides insights into the key soil health parameters driving differences between Everlizer and Nature Safe. Labile amino-N and mineral N are positively associated under Nature Safe fertilization, but negatively associated under Everlizer, suggesting that the intensity and quantity of soil nutrient pools under different organic fertilizer products may be distinct, exhibiting differential influence on soil N availability and transformation during the production season. A stronger positive relation between active C, CB, and OM is found under Nature Safe vs. Everlizer, and the latter seems to demonstrate a stronger negative relation between OM and SLAN as well as aggregate stability and SLAN.

The canonical discriminant analysis compared compost treatments under sunn hemp and weedy fallow scenarios and compared cover crop treatments under each compost treatment. The composts showed clearer separation under sunn hemp compared to the weedy fallow, suggesting that integrated nutrient management approaches may synergistically affect aspects of soil health and nutrient status. The no compost control was influenced by negative loading of CEC in both cover crop scenarios. Likely, the compost products promoted an improved CEC, with a greater magnitude of improvement in conjunction with sunn hemp cover crop residues. Similarly, the increased separation of Yard and Vermi from Control can be attributed to the positive loading of Ca and negative loading of BS-Ca, respectively, suggesting that impacts of composts and cover crop residues on CEC and base cation distribution are intertwined. Separation between Yard and Vermi is also observed in the sunn hemp scenario, driven by positive loading of P and aggregate stability, and negative loading of Al. It is intriguing that the separation diminished in the weedy fallow scenario, implying that the differential compost impacts on soil nutrients and physical indicators of soil health might be dependent on the use of cover cropping systems. Overall, the canonical discriminant analysis helped demonstrate the potential benefits of integrating cover crop residues and composts, particularly as it relates to promoting the overall improvement of soil physical, chemical, and biological parameters. Future research targeting soil and rhizosphere microbial community abundance and diversity will help to further contextualize the impacts of management practices on soil health improvement and dynamics. Nonetheless, impacts on microbial community structure appear to be cover crop-dependent, and few studies have examined the synergy of different sustainable management practices on microbial community dynamics and the persistence of compositional changes, particularly in the sandy soils of the southeastern US [63].

5. Conclusions

This study examined the impacts of different organic nutrient management practices on soil physical, chemical, and biological parameters following organic celery production seasons, over two years, under subtropical sandy soil conditions with low OM. In the organic fertilizer trials, soil CEC and Zn, Cu, and Fe levels were significantly higher in the season with less precipitation. Soil Zn, Cu, BS-K, BS-Mg, and the overall soil fertility score were consistently higher with the poultry litter-based organic fertilizer compared to the feather meal-based product, and the former also tended to increase the overall soil health score. In the integrated nutrient management trials, soil OM, pH, active C, NH₄-N, Zn, Cu, Mn, Fe, BS-Ca, and the overall soil health score varied greatly by production season. Soil active C was consistently higher with yard waste compost application across both seasons, and all compost treatments improved the overall soil health score compared to the no compost control, exhibiting short-term benefits to soil health. On the other hand, compost impacts on CO₂ burst, CEC, and aggregate stability were season-dependent, substantiating challenges in characterizing compost effects as influenced by environmental conditions, with respect to short term soil health enhancement, especially for sandy soils in warm and humid environments. Sunn hemp cover crop residues affected soil N availability, while synchronizing N releases with vegetable crop uptake is needed for reducing N losses. The principal component analysis further stressed major differences between the two organic fertilizers regarding their seasonal impacts on soil health, and the canonical discriminant analysis helped illustrate synergistic effects of composts and cover crop residues on soil nutrient pools and other physical and biological indicators of soil health. Long term studies of both monoculture and polyculture cover crop rotations on soil N status, SOC, microbial biomass, and systematic soil health attributes are warranted, particularly within the context of integrated organic production systems across diverse environments and soil types. While organic fertilizers play an important role in supporting organic vegetable crop nutrition and productivity, integrated practices like cover cropping, crop rotation, and compost are critical to building pools of SOC, improving soil structure, and stimulating soil microbiological diversity and activity toward the goal of sustainable agroecological systems in organic farming.