On-Farm Conversion of Cannabis Sativa Waste Biomass into an Organic Fertilizer by Microbial Digestion

Abstract

Nearly 90% of the total biomass of hemp plants grown for cannabidiol (CBD) is left to rot in piles, burned, or hauled to landfills post-harvest even though these residues are nutrient-rich. Composting is an option for waste management and nutrient recovery for hemp producers. However, materials, labor costs, time, and space needed to properly construct compost piles pose significant logistic challenges to smallholder farmers who want to manage their waste on-farm. Additionally, the high lignin content of cannabis hurd is often poorly degraded if the compost fails to meet national organic standards for thermophilic conditions. A potentially more efficient and cost-effective alternative is to convert leftover cannabis waste into bio-fertilizers through a semi-anaerobic process by adding microbial inoculants that specialize in lignin degradation. Imio Technologies, Inc. (Imio) developed cannabis-specific microbial inoculants designed to optimize the conversion of hemp waste into a bio-available form. Applied as a liquid culture, Imio microbes colonize wasted stalks and produce organic acids and oxidative enzymes that decompose lignin and release plant nutrients. This study aimed to investigate the ability of Imio microbial inoculants to degrade lignin-rich material into a fertilizer substitute. At the conclusion of a 3-week digestion period, biomass was transformed into a more bioavailable form with a nutrient profile similar to finished composts. Two-week cured fertilizer generated through this process promoted greater root:shoot ratios, which suggests plant growth promotion, compared to a synthetically fertilized control group. The digested hemp fertilizer also showed a value-added property of disease suppression against the soilborne pathogen Rhizoctonia solani.

Introduction

Over 16,000 acres of cannabis (Cannabis sativa L.) was harvested in the United States (US) as floral hemp [also referred to as hemp cannabidiol (CBD) or medical hemp] in 2021. With an average yield of 550 Kg (1200 lbs) of biomass per acre, that production likely generated over 7.9 M Kg (∼8600 tons) of hemp waste biomass (USDA National Hemp Report 2021). This substantial waste stream is a result of the crop being cultivated solely for its floral components that contain concentrated levels of cannabinoids such as CBD. Therefore, harvesting medical hemp is focused on the removal of floral tissue and largely disregards the remaining 90% of the plant as waste (Goldstein 2021). Leftover stalks and leaves are often left to rot in piles, burned, or hauled to landfills post-harvest although they are nutrient-rich (Heard, Watson, and Kostiuk 2007). This practice is not only damaging to the environment but expensive to farmers considering the high cost of fertilizers and labor needed to produce hemp CBD (de Bertoldi, Vallini, and Pera 1983; Jain, Bhatia, and Pathak 2014; Jelliffe, Lopez, and Ghimire 2020; Yang et al. 2008) and overlooks the potential of crop residues to restore fertility and improve soil health (Singh and Rengel 2007). As a relatively new crop in the US, states are just beginning to standardize cultivation guidelines specific to hemp CBD. However, the issue of sustainable waste management of the leftover stalks and subsequent biomass remains largely unaddressed.

The high value of hemp CBD products reduces economic incentive to recycle unused biomass leading many producers to rely on landfilling or incineration as a primary means for waste management. Stalks and leaves account for the majority of discarded biomass even though they store up to 80% of the macronutrients the plant consumes (Heard, Watson, and Kostiuk 2007). These nutrients are primarily housed in two types of tissues in the stalk: long outer fibers called bast and the inner dense wood called hurd. Bast fibers make up 20–40% of the stalk and are characterized by high cellulose (57–77%) and low lignin content (5–9%). Hurd makes up the other 60–80% and contains less cellulose (40–48%) and high amounts of lignin (21–24%) (Stevulova et al. 2014). With state-level policies in New England and other regions of the US now mandating the diversion of organic materials away from landfills, hemp producers need new ways to efficiently repurpose their wastes (O’Brien et al. 2019).

The degradation of lignin represents the biggest technical challenge in repurposing hemp stalks because it is highly resistant to decomposition and requires the activity of many enzymes for mineralization (Kumar and Chandra 2020; Lopez et al. 2006). These enzymes, primarily oxidases, are produced by specialized bacteria and fungi that have evolved to breakdown and extract nutrients from lignocellulosic plant residues. The capacity of microorganisms to break down lignocellulose and mineralize the stored nutrients depends on their ability to produce the necessary enzymes required for the complete degradation of the substrate. Generally, the more complex the substrate, the more complex and extensive are the enzymes required for decomposition (Ryckeboer et al. 2003; Tuomela et al. 2000). In the case for lignin, degradation proceeds relatively slowly with many microorganisms and enzymes required to mineralize the nutrients stored in these complexes (Datta et al. 2017).

Composting is an aerobic, biologically driven process that can serve as an effective tool for reclaiming nutrients locked up in lignocellulosic plant wastes and is an option for waste management of residual cannabis biomass. However, materials, cost of labor, time required, and space needed to properly construct compost piles pose significant logistic challenges to small-scale producers who want to manage their waste on-farm (Onwosi et al. 2017; Viaene et al. 2016). Further complicating this issue is the especially high lignin content of cannabis hurd which is often poorly degraded if the compost fails to meet national organic standards for thermophilic composting (Lopez et al. 2006; Tuomela et al. 2000). Many hemp farmers are restricted in their ability to produce quality composts because factors that dictate compost efficiency (e.g., oxygen, temperature, moisture, particle size, and chemical composition) require special attention and machinery. Even for farmers willing to invest time, labor, and resources in composting—the consistency of the microbial succession and output of the process is variable (Neher et al. 2013). A potentially more efficient and cost-effective alternative to managing cannabis waste is by converting the leftover organic material into bio-fertilizers using microbial inoculants on-farm.

Bokashi composting is a cost and space-efficient waste management strategy used to generate nutrient-rich fertilizers through organic waste fermentation using microbial inoculants (Murillo-Amador et al. 2015; Quiroz and Céspedes 2019). Typically, this microbial inoculant is Effective Microorganisms^TM (EM) which consists of lactic acid bacteria, Saccharomyces spp. (yeast), photosynthetic bacteria, and actinobacteria (Higa 1991; Higa and Parr 1994; Shin et al. 2017). Often grown on wheat bran before applied to waste, the added microbes have specialized biochemical properties that facilitate the rapid conversion of organic material into a bioavailable form (Christel 2017; Quiroz and Céspedes 2019; Xu 2001). The final product, a nutrient-dense material, rich with organic matter and teeming with beneficial microbes, is produced in a fraction of the time it takes for traditional composting (Olle 2020, 2021). Although this system effectively recycles mixed agricultural waste products such as wheat straw, manures, sawdust, and various plant residues, waste piles consisting solely of highly lignified hemp stalks can be difficult to break down in bokashi composts. A solution to this pitfall could be a microbial inoculant optimized for the degradation of lignin.

Microbial inoculants can optimize composts and improve decomposition of highly recalcitrant materials such as lignified plant residues (Greff et al. 2022). Introducing lignin-degrading microbes directly to lignocellulosic plant wastes can shorten the recycling process and improve the output consistency of the resulting fertilizer. In collaboration with local Vermont startup company Imio Technologies, Inc. (Imio), cannabis-specific microbial inoculants were designed to optimize the conversion of hemp waste into a bio-available form. The inoculants consist of proprietary formulations of lactic acid bacteria and white-rot fungi that work best under semi-anaerobic conditions and require little maintenance once added to waste. Applied as liquid or solid (bran) culture, the Imio microbes colonize wasted stalks and produce organic acids and oxidative enzymes that degrade lignin and release plant nutrients.

The overall goal of this project was to identify microbial formulations that best degrade cannabis waste and provide farmers a sustainable and cost-effective strategy for handling their leftover biomass. This three-part study aimed to evaluate: (1) the ability of Imio microbial inoculants to degrade lignin-rich hemp stalks; (2) if the generated fertilizer exhibits phytotoxicity or suppresses plant disease; and (3) how the fertilizer affects the growth of hemp CBD plants. Each aim was captured as a question and addressed by separate experiments. Performance of Imio microbial inoculants were compared to the Effective Microorganisms^TM (EM) commercial inoculant and water to serve as a negative control.

Materials and Methods

Microbial Inoculants

Microbial inoculants were evaluated for their ability to degrade hemp waste. Four formulations of proprietary microbial blends (MI1-4) were compared to commercial and negative microbial controls of Effective Microorganisms^TM (EM) (TeraGanix Inc., Alta, TX) and water, respectively. Proprietary inoculants were a mixture of 7–10 different species of ligninolytic bacteria and fungi varying in concentration and species combination. All microbes were added at equal ratios to make solutions with OD600 values of 1 or 2. Bacteria from the genus Lactobacillus represent a portion of the proprietary formulations, as well as a variety of bacteria including Bacillus subtilis and Rhodopseudomonas palustris, supplemented with yeast (Saccharomyces spp.) and various white-rot fungi including Irpex lacteus and Phanerochaete chrysosporium. Top performing inoculant (MI4) was optimized by adjusting the ratio of Lactobacillus (MI4.1) and using an alternate strain of lignin degrading bacteria with twice the concentration of microbes (MI4.2) which were tested in sequential experiments. EM contains primarily lactic acid bacteria, photosynthetic bacteria, yeast, and fermenting fungi (Higa 1991; Higa and Parr 1994).

Microbial blends were evaluated both as a liquid and solid. Solid (bran) formulations of inoculants were created using the protocol from Christel (2017) with the following modifications. 85 g of molasses and 85 g of EM inoculant or MI liquid culture was dissolved in 3.8 L of deionized water. This mixture was then added to 5.4 kg of wheat bran and homogenized. The inoculated bran was placed inside double-layered garbage bags (0.95 MIL) to exclude light and oxygen. Bran fermented in the dark at room temperature for 2 weeks prior to use.

Question 1: What is the Best Microbial Inoculum to Decompose Hemp?

Experimental Design

Three successive experiments were conducted with a complete factorial combination of treatments and experimental units completely randomized. Experiment 1 compared four liquid inoculant blends (1, 2, 3, 4) with both a commercial inoculant (EM) and water included as controls (6 inoculants x 15 replications = 90 experimental units). Experiment 2 compared optimized variations of two top-performing blends (4.1, 4.2) and a water control only (3 inoculants x 15 replications = 45 experimental units). Experiment 3 compared one top-performing blend (4.1) as either a liquid or as a solid (bran) formulation with EM bran as a reference control (3 inoculants x 15 replications = 45 experimental units). Each treatment was sampled destructively at three times during fermentation and replicated five times per experiment. Containers were sampled to estimate mass loss, microbial enzyme activity, and available nutrients. At the end of each 21-day fermentation period, containers were emptied, and material was consolidated by treatment, mixed with steam pasteurized soil or coconut coir at a 1:1 v/v ratio and cured in open 19-liter buckets for 1, 2 or 4 wk in preparation for plant bioassays and growth trials (questions 2 and 3) (Table 1).

Table 1. Summary of treatments and controls in each experimental run. All treatment combinations were replicated 15 times.

Experiment	Inoculum	Water control	EM control
1	1–4	x	x
2	4.1, 4.2	x
3	4.1		x

Decomposition

Decomposition was measured as mass loss through time. No-see-um polyester netting (The Rain Shed, Inc., Corvallis OR, USA) sachets (5 cm x 5 cm) were filled with 3.5 ± 0.5 g dried shredded hemp and placed into 1 L autoclaved (121°C, 25 min) polycarbonate plastic containers. Shredded hemp (40 g) was then packed into each container and moistened with 1 mL g⁻¹ of respective inoculant. Following inoculation, containers were closed tightly to create anaerobic conditions and incubated at room temperature for 21 days. Five containers from each treatment were opened and sampled destructively every 7 days. Mass loss was determined as the difference in weight of hemp before and after decomposition after subtracting the weight of the empty sachet. Additional 5 g subsamples were removed from the bulk hemp to determine gravimetric moisture and enzyme activity. Gravimetric moisture was determined by drying digested hemp at 90°C for 48 h and computing as the percentage of weight of water to weight of dry hemp.

Enzyme Activity

Oxidase, hydrolase, and amino-peptidase activity were quantified as indicators of microbial functional activity and measured as nmol h⁻¹ g⁻¹ of dry hemp. Oxidases (peroxidase and phenol oxidase) serve as an indicator for lignin degradation, a primary component in hemp stalks. Hydrolases such as ß-1,4-glucosidase (BG) and ß-1,4-N-acetylglucosaminidase (NAG) indicate cellulose and chitin hydrolysis, respectively, while L-leucine aminopeptidase (LUC) and phosphatase (PHOS) activity are used to quantify the degradation of proteins and phosphate, respectively. Enzyme activity of the inoculated shredded hemp was measured at days 7, 14 or 16, and 21 during the digestion period using the protocol from Neher, Fang, and Weicht (2017) in which 0.3 g of shredded hemp was mixed in 200 mL of autoclaved nanopure (18.2 megohm) water.

Nutrient Analysis on Fermented Hemp Tissue

At the conclusion of the fermentation period, digested hemp samples from each treatment in experiment 3 were analyzed for nutrient content. Soil used for curing as well as 2-week cured material were also analyzed. Samples were extracted using a Modified Morgan method (Lunt, Swanson, and Jacobson 1950; Morgan 1941). Elements included were phosphorus (P), potassium (K), calcium (Ca), magnesium (Mg), sulfur (S), iron (Fe), manganese (Mn), boron (B), copper (Cu), zinc (Zn), sodium (Na), molybdenum (Mo) and aluminum (Al). Ammonium and nitrate were measured using KCl extraction methods (Keeney and Nelson 1982). Cation exchange capacity (CEC) was calculated from cation ratios of Ca, Mg, and K while soil pH was determined using 0.01 M CaCl₂. Percent organic matter (% OM) was measured by loss on ignition and converted to a Walkley-Black equivalence (Walkley and Black 1934). Internal standard reference material was included in each batch of 20 samples and output was corrected for calibration and reagent blanks for quality assurance.

Question 2: Does the Inoculum Exhibit Phytotoxicity or Suppress Plant Disease?

Phytotoxicity was evaluated using two indicator plant species, garden cress (Lepidium sativum) and red clover (Trifolium pratense). Disease suppression against fungal pathogen Rhizoctonia solani was quantified using a plate competition approach (Neher, Fang, and Weicht 2017).

Experimental Design

Two successive plant bioassays were conducted with a complete factorial combination of treatments, and experimental units completely randomized. Bioassay 1 compared germination of garden cress seedlings in mineral soil amended with hemp waste from four different microbial treatments (MI1, MI2, MI3, MI4), at two cure times (0, 4 wk). Two microbial controls (EM, water) and an unamended soil control were included as references. Uncured and cured material were replicated two and four times, respectively, to give a total of 40 experimental units in an unbalanced design to accommodate limited materials [(6 inoculants uncured x 2 reps) + (6 inoculants cured x 4 reps) + (uninoculated soil control x 4 reps)].

Bioassay 2 compared recycled waste from two top-performing blends (MI4.1, MI4.2) in either of two media (mineral soil, coconut coir) at three curing times (0, 1, 2 wk), using both garden cress and clover. One microbial control (water) and uninoculated soil or coir media were included as references for soil and coir treatments, respectively. Treatments and controls were replicated four times. Uncured (0 week) material was evaluated only in mineral soil, not coir. In summary, there were 80 experimental units with soil [(2 plants x 3 inoculants x 3 cure times x 4 reps) + (2 plants x 1 uninoculated soil control x 4 reps)] and 56 experimental units for coir [(2 plants x 3 inoculants x 2 cure times x 4 reps) + (2 plants x 1 uninoculated coir control x 4 reps)] for a grand total of 136 experimental units.

Part A: Phytotoxicity

Digested hemp was added to pasteurized mineral soil (collected from the University of Vermont Horticulture Research and Education Center, 44°25’42.128”N 73°12’30.435”W) or coconut coir (Eco-co® Coir, Gardener’s Supply Company, Burlington, VT) at a 1:1 v/v ratio. Field soil (Adams and Windsor loamy sands) was steam pasteurized at 70°C for 4 h to destroy native pathogens and then re-inoculated with its endemic microbial community by adding 4 L of 10-µm filtered-soil extract and rested for 3 weeks to allow the microbial community to reestablish. Fresh and cured digested hemp (remaining from Question 1 experiments described above) was amended to soil or coconut coir and placed into 10.2 cm (4”) plastic orchid pots. Twenty-five seeds of garden cress or clover were planted into each pot using a customized dibble-stick to ensure 2.54 cm (1”) between each seed sown. Plant bioassays were performed in greenhouse conditions under natural day lengths and watered daily. Mean day and night temperatures were 26°C and 21°C, respectively with relative humidity of 68%. Percent germination and seedling emergence was quantified for each experimental unit after 14 days. Phytotoxicity was quantified as the net reduction in seedling emergence compared to unamended soil and coconut coir reference controls.

Part B: Disease Suppression

A plate competition method was employed to compare growth of R. solani with and without competition from living microbes in the media of interest. Briefly, two pairs of 50 mL tubes containing 50 mL of sterile water were amended with 0.5 g samples of fresh or 4-week cured digested hemp from each treatment in experiment 1 and shaken overnight. The next day, one tube from each of the pairs was autoclaved (to kill microbes) while the remaining tubes were left on the shaker to preserve living microbes. The contents of the pairs (with or without living microbes) were each added to flasks of water agar (1.5 g agar to 50 mL of deionized water) that had been autoclaved at 121°C for 25 min and the molten agar cooled to 55°C. The slurry of digested hemp in molten water agar was swirled gently to homogenize and poured into 100 mm x 15 mm plastic petri plates with both fresh and 4-week cured material replicated 4 times for each treatment. The next day plugs of R. solani growing on potato dextrose agar were transferred onto the living and autoclaved digested hemp agar plates and incubated at room temperature for 24 h. Mycelia growth was measured to the nearest 1 mm using a stereo microscope. Three of the longest hyphal strands from the center of the plug were recorded and the mean was used as a representative measure to compare suppressive potential among the different treatment groups. Suppressive potential was computed as the difference of mycelial radii of the non-autoclaved slurry subsample from the autoclaved slurry subsample (control). Negative values represented suppressive potential.

Question 3: How Do Different Microbial Formulations Affect Cannabis Plant Growth?

Experimental Design

The approach was as described for the phytotoxicity tests above with the following modifications. Experimental units were 2.4 L plastic pots containing three autoflower CBG hemp seeds (Oregon CBD, Independence, Oregon), arranged in a completely randomized design. Each treatment was replicated five times for a total of 25 experimental units. Treatments included the top performing proprietary inoculum (MI4.1) in either liquid or solid bran media without synthetic fertilizer and proprietary inoculum (liquid) with Jack’s 15-16-17 Peat-Lite water-soluble fertilizer (JR Peters Inc., Allentown, PA). These were compared to commercial and synthetic controls. The commercial control was EM bran to match a typical protocol for making bokashi (Christel 2017; Quiroz and Céspedes 2019). Synthetic control was unamended soil fertigated with Jack’s 15-16-17 Peat-Lite water-soluble fertilizer at 18 kg N per hectare (100 lbs N/acre). MI4.1 liquid inoculum fertigation treatment received an equivalent dose of 9 kg N per hectare (50 lbs N/acre) to test the interaction between digested hemp and synthetic nutrients on plant growth. Three seeds were planted into each of the five treatment combinations. Seedling emergence was recorded on day 10 when all pots were thinned to 1 plant per pot and allowed to grow an additional 20 days in which fertigation treatments received respective doses of Jack’s 15-16-17 Peat-Lite water-soluble fertilizer. Plant height, width and depth was measured periodically until the completion of the experiment. On day 30, above and belowground biomass was harvested, cleaned of any soil, dried, and weighed to the nearest 0.1 g. Subsamples of dried leaves were sent for nutrient analysis.

Germination Assays and Hemp Growth Trial

Two-week cured digested hemp was amended to steam pasteurized soil at a 1:1 v/v ratio and filled into 2.4 L (4”) plastic pots. Three hemp seeds were sown 2 cm deep into the media of each pot and five replicates were ascribed to each treatment group. Germination and growth experiments were performed in a greenhouse with a mean daily light intake of 11 mol⋅m⁻² ⋅day⁻¹ and temperature controlled at 26°C during the day and 21°C at night with relative humidity of ∼70%. Pots were arranged randomly on greenhouse benches and observed over 10 days for seedling emergence. Plants were rerandomized following each irrigation event and treatment results were expressed as percent seedling germination rate. On day 10, pots were thinned to 1 plant and watered ad libitum with corresponding nutrient solution (Jack’s 15-16-17 Peat-Lite) to saturation with a 30% leaching fraction for an additional 20 days. Plants amended with MI4.1 liquid inoculum, MI4.1 bran, and EM bran treated hemp received only water throughout the growth trial. Plant height was measured from the ground level to the leaf base of the highest fully expanded leaf and was recorded every 3 days starting on day 10. Plant width and depth measured the diameter of opposite axial leaf tips and was reported every 5 days starting on day 10. On day 30, aboveground biomass was harvested, cleaned of any soil, dried in an oven for 48 h at 60°C and weighed. Samples of the most recently matured (RML) dried leaves were sent for plant tissue analysis. Additionally, roots were washed clean of any soil, dried 48 h at 60°C and weighed. Root:shoot (R:S) ratio was expressed as g root dry weight divided by g shoot dry weight.

Hemp Leaf Tissue Analysis

At the conclusion of the grow trial, RML samples were sent for plant tissue analysis to assess nutrient deficiencies (Waters Agricultural Laboratories Inc., Camilla, GA). Samples were taken from each plant and consolidated based on treatment. Leaves were dried in an oven for 48 h at 60°C, weighed, and assayed for nutrient content. Plant P, K, Ca, Mg, S, B, Zn, Mn, Fe, Cu were measured by wet digestion (ICAP – open vessel wet digestion Digi Block 3000). Plant nitrate was assayed using a specific ion electrode technique.

Statistical Analysis

A two-way analysis of variance (ANOVA) was employed to analyze effects of inoculum and time as independent variables on decomposition and pathogen growth as dependent variables. A one-way ANOVA was employed to analyze effects of treatment on plant growth for each plant species separately. Dependent variables included plant height, width, and depth. Tukey post-hoc comparisons were performed to compare among levels of independent variables. Data were transformed to meet normality assumptions when necessary. Enzyme activities were square root transformed, decomposition and germination were arcsine of the square root of the proportion (mass loss, germination) transformed. Pathogen growth did not require transformation. R:S did not meet normality so was analyzed non-parametrically. Normality, ANOVA, and non-parametric ANOVA were performed using the UNIVARIATE, MIXED, and NPAR1WAY procedures, respectively, in SAS version 9.4 software (SAS Institute, Cary, NC). All statistical analyses were performed with transformed variables.

Results

Best Microbial Inoculum for Decomposition

To determine the ability of the inoculants to decompose lignin-rich material, various formulations were applied to hemp waste and mass was tracked over 21 days. Hemp mass loss inside the no-see-um sachet was similar (p > 0.4) across all treatment groups within each experiment on each of the sampling periods (Figure 1).

Figure 1. Mass loss displayed as percent of total hemp waste remaining in no-see-um sachets through time for experiments 1 (a), 2 (b), and 3 (c). treatments are represented by shape and fill pattern (white triangle: MI1; white square: MI2; white hexagon: MI3; black upside-down triangle: MI4; black and white diamond: EM; black circle: water control; black triangle: MI4.1; black and white upside-down triangle: MI4.2; black circle: water control; black triangle: MI4.1; black and white triangle: MI4.1 solid formulation (bran); black and white diamond: EM solid formulation (bran)). Data points represent means plus 1 standard deviation (n = 5). There were no statistical differences in mass loss among treatments.

Lignin degradation is typically carried out by oxidative enzymes such as peroxidases or phenol oxidases. Imio and EM inoculants reliably produced greater oxidative enzyme activities than biomass treated with water on days 14 and 16 of the peroxidase and phenol oxidase assays. Specifically, MI4 outperformed all others on days 16 and 21 of the peroxidase and phenol oxidase assays in experiment 1. For experiment 2, MI4.1 liquid inoculum recorded the highest activities of both peroxidase and phenol oxidase enzymes on days 14 and 21. EM bran produced greater enzyme activities on days 14 and 21 of the peroxidase assay when compared MI4.1 liquid and MI4.1 bran in experiment 3. However, MI4.1 liquid recorded greater phenol oxidase activities on days 14 and 21 when compared to MI4.1 bran and EM bran. Biomass treated with water (negative control) generally produced low oxidative enzyme activity and occasionally outperformed prototype inoculants. Overall, oxidative enzyme activity tended to increase after the first sampling event with activity peaking on day 14 or 16, followed by a reduction in activity on day 21 (Figure 2).

Figure 2. Peroxidase (a, b, c) and phenol oxidase (d, e, f) enzyme activity values throughout the 3-week fermentation period for experiments 1 (a, d), 2 (b, e) and 3 (c, f). Values are provided as μmol/h/g of substrate. Treatments are represented by bar fill pattern (solid white: water control; solid gray: EM liquid formulation; white with black speckles: MI1; gray with black checkers: MI2; gray with diagonal black stripes: MI3; solid black: MI4; solid white: water control; light gray: MI4.1; light gray with gray checkers: MI4.2; dark gray with black speckles: EM^TM solid formulation (bran); light gray: MI4.1; light gray with black speckles: MI4.1 solid formulation (bran)). Bars represent means plus 1 standard deviation (n = 5) and contrasting lowercase letters represent statistical significance among treatments within the same time (p < 0.05).

Enzyme ratios indicating relative C:N and C:P dynamics in the decomposition communities suggest C was more limiting than N and P throughout the 3-week digestion period for all experiments (Supplemental Figure 1). Demand for C was highest for all treatments on day 7 in experiments 1 and 2. C limitations in experiment 3 remained consistent throughout 3-weeks for MI4.1 liquid and EM bran treatment groups but varied greatly for MI4.1 bran with spikes of C-acquiring enzymes occurring on days 7 and 21. C:P ratios varied greatly across treatments for experiments 1 and 2 but in general MI3 and MI4 had steadily increasing demand for C while the other treatments showed little change. In experiment 2, biomass treated with water showed an increased demand for C when compared to P on days 7 and 21 but this trend was not observed with MI4.1 and MI4.2. C:P ratios in experiment 3 were similar across all treatments throughout 21 days.

Nutrient analysis was performed on the microbially treated hemp waste in experiment 3 only (Table 2). Total N and C were similar across all treatments. Extractable P, K, Mg, Mn, and Zn were greatest in MI4.1 bran and EM bran when compared to MI4.1 (liquid). However, extractable S and Fe was greatest in MI4.1 (liquid). Extractable Ca, B, Mo were similar across all treatments while Cu was greatest in MI4.1 bran.

Table 2. Properties of the microbial treated hemp waste (n = 1).

Parameter	MI4.1	MI4.1 Bran	EM Bran
Total Nitrogen (%)	1.9	2.1	2.2
Total Carbon (%)	44.2	42.2	43.3
C:N ratio	22.7	19.7	19.4
Phosphorous (%)	0.39	0.58	0.60
Potassium (%)	0.74	0.83	0.94
Calcium (%)	1.99	2.03	1.85
Magnesium (%)	0.42	0.53	0.52
Sulfur (%)	0.59	0.28	0.31
Boron (mg kg^-1)	36.5	33.4	32.5
Copper (mg kg^-1)	41.8	115.2	48.9
Iron (mg kg^-1)	661.2	545.1	463.3
Manganese (mg kg^-1)	128.9	168.0	169.0
Molybdenum (mg kg^-1)	8.1	7.8	8.1
Zinc (mg kg^-1)	68.8	106.1	98.7

Amending field soil with hemp treated with MI4.1 liquid inoculant, MI4.1 bran, and EM bran improved pH to near neutral values (Table 3). MI4.1 liquid and bran-treated hemp also increased %OM by roughly 1%. Bran formulations contributed greater extractable ammonium and nitrate, P, and K than liquid formulations. Liquid formulations resulted in greater S and Na concentrations than bran formulations. MI4.1 bran contributed the most Mg and Mn while Ca, B, Cu, and Fe values were similar across all treatment groups. Zn and Al values were greatest in unamended soil.

Table 3. Properties of soil amended with microbially treated hemp waste (n = 1 for all treatments).

Parameter	Optimal levels*	Field soil	MI4.1 + Field Soil	MI4.1 Bran + Field Soil	EM Bran + Field Soil
pH	6.2–6.8	6.2	6.8	7.4	7.4
Organic matter (%)	–	4.0	5.1	4.8	4.2
Effective CEC (meq/100g)	–	7.1	8.5	9.5	8.5
Ammonium (mg kg^-1)	–	67.0	31.0	79.0	86.0
Nitrate (mg kg^-1)	–	10.0	8.7	103.0	102.0
Phosphorous (mg kg^-1)	44–83	8.6	18.5	50.6	27.6
Potassium (mg kg^-1)	200–400	265.0	405.0	677.0	567.0
Calcium (mg kg^-1)	800–1600	999.0	1115.0	1068.0	1022.0
Magnesium (mg kg^-1)	101–200	179.0	237.0	301.0	237.0
Sulfur (mg kg^-1)	–	20.0	114.0	34.0	19.0
Aluminum (mg kg^-1)	–	40.0	25.0	25.0	26.0
Boron (mg kg^-1)	–	0.45	0.65	0.85	0.65
Copper (mg kg^-1)	–	0.95	0.75	1.05	0.8
Iron (mg kg^-1)	–	10.0	8.0	12.4	8.7
Manganese (mg kg^-1)	–	64.8	30.6	69.7	27.8
Sodium (mg kg^-1)	–	8.0	86.0	14.0	10.0
Zinc (mg kg^-1)	–	13.2	4.3	4.6	4.1

^* Optimal levels based on recommendations provided by A&L Canada Laboratories.

Phytotoxicity and Plant Disease Suppression

Longer cure times tended reduce the phytotoxic effects of the digested hemp fertilizer and improve germination rates of both plant species. Mean germination rate of garden cress planted in 4-week cured MI4 digested biomass was greater than seeds grown in uncured material in bioassay 1 (p < 0.05) (Figure 3). In bioassay 2, germination rate of garden cress and red clover varied greatly by treatment, curing media, plant species and length of cure phase (Figure 4). Mean germination was generally greater in red clover than in garden cress; however, both species performed similarly in either curing media.

Figure 3. Plant bioassay 1 garden cress germination rates in soil amended with microbially treated hemp waste expressed as percentage of 25 seeds. Solid fill bars represent means of treated hemp that was cured in soil for 4 weeks (n = 4) while speckled bars represent means of uncured hemp (n = 2). Treatments are labeled on the x-axis with corresponding code (soil: unamended field soil control; water: field soil amended with water treated hemp; EM: field soil amended with Effective Microorganisms^TM liquid inoculant treated hemp; MI1: field soil amended with Imio microbial inoculant 1 treated hemp; MI2: field soil amended with Imio microbial inoculant 2 treated hemp; MI3: field soil amended with Imio microbial inoculant 3 treated hemp; MI4: field soil amended with Imio microbial inoculant 4 treated hemp). Statistical significance between cure phases within the same treatment group are marked with an asterisk (p < 0.05). Error bars represent plus 1 standard deviation.

Figure 4. Garden cress (a, c) and red clover (b, d) germination rates in soil (a, b) and coconut coir (c, d) amended with microbially treated hemp waste expressed as a percentage of 25 seeds. Bar fill pattern represents the length of time material was cured in soil/coir before the assay (speckled fill: uncured; striped fill: 1 week; solid fill: 2 weeks). Treatments are labeled on the x-axis with corresponding code (soil: unamended field soil control; coir: unamended coconut coir control; water: field soil/coir amended with water treated hemp; MI4.1: field soil/coir amended with Imio microbial inoculant 1 treated hemp; MI4.2: field soil/coir amended with Imio microbial inoculant 4.2 treated hemp). Bars represent means plus 1 standard deviation (n = 4). There were no statistical differences between microbial treatments or cure times.

All treatments exhibited suppression of the growth of fungal pathogen Rhizoctonia solani; however, 4-week cured material had a greater suppressive effect than uncured (Figure 5). Cured material treated by MI2 and MI4 was significantly greater at inhibiting pathogen growth when compared to uncured material of the same treatment. Biomass treated with MI1 was significantly greater at suppressing pathogen growth than the negative control (p < 0.05).

Figure 5. Illustrated are means ± 1 standard deviation of increased suppressiveness of the same material with living microbes than without microbes (autoclaved). Negative values represent suppressive potential. Solid fill bars represent treated hemp that was cured in soil for 4 weeks (n = 4) while speckled bars represent uncured hemp (n = 4). Treatments are labeled on the x-axis with corresponding code (water: hemp waste treated with water; EM: hemp waste treated with Effective Microorganisms^TM liquid inoculant; MI1: hemp waste treated with Imio microbial inoculant 1; MI2: hemp waste treated with Imio microbial inoculant 2; MI3: hemp waste treated with Imio microbial inoculant 3; MI4: hemp waste treated with Imio microbial inoculant 4). Statistical differences between cure phases within the same treatment group are marked with an asterisk (p < 0.05). Statistical differences between treatments with respect to water are marked with a Delta symbol (p < 0.05). Error bars represent plus 1 standard deviation.

Effect of Digested Hemp Fertilizer on Growth of Hemp CBD Plants

Regardless of microbial treatment, cured digested hemp supported seed germination and plant growth through the first 20 days of the trial suggesting its ability to be used as a fertilizer substitute. Hemp germination rate in soil amended with 2-week soil cured microbially treated hemp waste was similar across all treatment groups (p > 0.3) (Supplemental Figure 2). Hemp growth was also similar across all treatments through day 16 of the 30-day trial. After day 16, plants growing in MI4.1 treated waste fertigated with 9 kg N/ha per acre showed a decline in height when compared to the other treatments (Supplemental Figure 3). Plants growing in only MI4.1 liquid recycled hemp paralleled plants growing in bran treated biomass and plants fertigated with 18 kg N/ha up until day 22 where height then started to plateau. On day 30, plants grown in MI4.1 bran and EM bran treated biomass, along with plants fertigated with 18 kg N/ha (control) were taller than the other treatment groups (p < 0.05).

Results of the leaf tissue analysis showed potential deficiencies in hemp grown in microbially treated waste (Table 4). Plants grown in MI4.1 treated biomass showed the lowest values of total nitrogen and sulfur suggesting the plants were deficient in these nutrients. Plants grown in MI4.1 treated biomass fertigated with 22.5 Kg N per acre also showed potential nitrogen deficiencies.

Table 4. Leaf tissue analysis results (n = 1 for all treatments).

Parameter	Optimal range*	Control	MI4.1	MI4.1+ ½ Control	MI4.1 Bran	EM Bran
Nitrogen (%)	3.3–5.0	5.7	2.2	3.3	4.7	4.6
Phosphorous (%)	0.27–0.48	0.35	0.39	0.49	0.48	0.41
Potassium (%)	1.8–2.7	3.8	2.1	2.7	3.4	3.2
Calcium (%)	1.5–2.9	2.2	3.6	4	3.8	4.3
Magnesium (%)	0.3–0.65	0.47	0.57	0.68	0.86	0.92
Sulfur (%)	0.25–0.36	0.53	0.25	0.39	0.39	0.37
Boron (mg kg^-1)	30–90	84	77	91	54	58
Copper (mg kg^-1)	5–11	17	10	11	14	13
Iron (mg kg^-1)	70–150	255	102	123	164	148
Manganese (mg kg^-1)	40–158	272	109	77	272	826
Zinc (mg kg^-1)	33–60	128	61	74	69	75

* Optimal leaf tissue nutrient ranges from Suchoff et al. (2021).

R:S ratios of hemp grown in MI4.1 liquid treated waste were greater than plants grown with synthetic fertilizers and EM bran (p = 0.06) suggesting a plant growth promotion benefit from MI4.1 (Figure 6).

Figure 6. Root:shoot ratios of hemp CBD plants grown in soil amended with microbially treated hemp waste. Treatment groups are represented by fill pattern (white: soil control fertigated with 18 Kg N per hectare; dark gray with black speckles: Effective Microorganisms^TM solid formulation (bran); light gray with black speckles: microbial inoculant 4.1 solid formulation (bran); light gray: microbial inoculant 4.1; gray with black stripes: microbial inoculant 4.1 + 9 Kg N per hectare). Error bars represent means plus 1 standard deviation (n = 5). Contrasting lowercase letters indicate statistical differences of root:shoot ratios between treatments (p = 0.06).

Discussion

Prototype inoculants and the commercial control (EM) consistently produced greater activities of oxidative enzymes than hemp waste treated with water (negative control) suggesting the added microbes successfully colonized the wasted stalks and outcompeted indigenous microbes to some degree. Fertilizer generated after 3 weeks of fermentation proved to be a significant source of plant-available nutrients but needs to cure in soil at least 2 weeks to reduce phytotoxic effects. When properly cured, the digested hemp fertilizer can serve as a substitute for synthetic nutrients and boost plant growth. Digested hemp generated through this process could also be used to suppress the soilborne fungus Rhizoctonia solani.

Lignin Degradation

The high activity of oxidative enzymes produced by Imio and EM inoculants suggest the added microbes successfully colonized the wasted stalks and initiated lignin degradation. Despite this trend, there were exceptions where the negative control (water) produced measurable and even higher activities of oxidases than the prototype inoculants. This activity can be attributed partly to the indigenous microbial community arising from storage of the post-harvest hemp material outdoors in uncontrolled environmental conditions. It is likely that the established microbes were active prior to inoculation which would explain the observed enzyme production. Regardless, these activities were much less when compared to MI4, MI4.1, and EM treatments indicating the commercial inoculants were successful at outcompeting indigenous microbes to some degree. MI4 generally produced higher oxidase activities than all others including EM in experiment 1 which supported MI4 as the microbial blend best suited for future commercialization and led to further optimization attempts resulting in MI4.1 and MI4.2.

Optimization of microbial formulation and concentration had minimal effect on improving oxidative enzyme production which could be due to the change of formulation or possibly from sampling inconsistencies between the two experiments. Inoculants in experiment 1 were sampled on days 7, 16, and 21 while the other experiments were sampled on 7-day intervals (sampling inconsistencies were due to a delay in delivery of experiment materials arising from supply chain issues created by the COVID-19 pandemic). This 2-day difference could explain the observed loss in activity between optimized and original formulations. However, further investigation is required to fully understand what caused this change.

The addition of bran tended to promote oxidative enzyme production despite reports that excess N in decomposition environments suppress oxidative enzyme production (Ryckeboer et al. 2003; Sinsabaugh 2010). It is understood that N limitation serves as a prerequisite for the release of oxidative enzymes and the subsequent degradation of lignin (Fog 1988; Keyser, Kirk, and Zeikus 1978; Waldrop and Zak 2006). However, it is possible that the addition of N via bran increased growth and competition of the microbial community leading to the release of oxidases as a competitive strategy rather than a means to obtain nutrients from lignin (Sinsabaugh 2010). This could explain the differences in activity between the liquid and bran treatment groups. In either case, more investigation is required to determine the mechanism behind this observation.

Oxidative enzyme activity tended to increase after day 7 suggesting the microbial communities were initially limited by C but became more N limited around days 14 and 16. Insufficient N could explain the increased activity of oxidative enzymes observed on the second sampling event because microbes would be producing the peroxidases and phenol oxidases required to liberate N from lignin and other recalcitrant material to meet their metabolic requirements. It would also explain the decrease in C:N enzyme activity ratios observed in experiments 1 and 2 which also suggest the communities were increasingly competing for N.

Nutrient Recovery

Nutrient analyses performed on the microbially treated hemp waste validate that digested hemp C:N ratios and nutrient values are comparable to that of finished composts (Siedt et al. 2021). The digested hemp could be used as an organic fertilizer substitute because it yields relatively high amounts of primary and secondary nutrients. The C:N ratio is also desirable from a compost perspective which ranges from 19–22 indicating the amendment is stable for fertilizer application (Azim et al. 2018; Greff et al. 2022; Harindintwali, Zhou, and Yu 2020). Despite these generalizations, liquid formulations (i.e., MI4.1) produced a slightly higher C:N ratio and lower concentrations of primary and secondary nutrients except for S when compared to bran formulations which was likely due to the nutritional and physical properties of the wheat bran itself. The use of bran as a mode of delivery would provide additional nutrients to the hemp waste and lower the C:N ratio of the resulting fertilizer. This would explain the improved fertility properties of the recycled biomass generated from bran formulations when compared to liquid. Regardless of delivery mode or microbial treatment, amending field soil with digested hemp improved soil fertility, %OM and pH.

The unamended field soil used in the plant bioassays and growth trial was nutrient poor for a nutrient-intensive crop like hemp. Ideally, pH should be near 7.0 and P levels should be 44–83 mg•kg⁻¹. However, when amended with digested hemp, field soil properties improved to near optimal levels (Table 3) indicating that recycled hemp could serve as a potential fertilizer substitute when preparing a site for cultivation. In addition, the digested hemp seemed to increase CEC and trace elements in some cases. In terms of N, amendments of MI4.1 liquid treated hemp were associated with losses of nitrate and ammonium not observed in bran formulations. This was expected because wheat bran is relatively rich in N and other nutrients with some studies reporting N-P-K values of 0.41-0.01-2.67 (Abbas et al. 2012). Overall, these findings suggest digested hemp generated through the Imio recycling process could serve as a fertilizer substitute considering its high nutrient content, desirable C:N ratio, and generally positive effect on soil properties when used as an amendment.

Phytotoxicity

Phytotoxicity is common and well-documented in immature composts and bokashi fertilizers that have not been properly cured (Selim, Zayed, and Atta 2012). In this study, longer cure phases correlated with increased germination rates of both plant species suggesting that uncured digested hemp was phytotoxic to a certain extent and the cure phase is necessary to reduce deleterious effects. The detrimental effects observed in the uncured treatment groups likely originate from high residual microbial activity and volatile organic acids that can interfere with plant root development and stunt growth. For this reason, it is generally recommended that composted material cures or matures in open air for a few weeks to many months prior to use (Pace, Miller, and Farrell-Poe 1995). Contrarily, bokashi fermented material should cure in soil for 2 or more weeks before planting (Quiroz and Céspedes 2019). Because the Imio recycling strategy most closely resembles bokashi composting, material was cured for 0-, 1-, 2- or 4 wk in this study. Our experiments suggest a 2-week cure phase seemed to be sufficient in reducing phytotoxicity which is consistent with the general recommendations for curing bokashi composts. When properly cured, the digested hemp fertilizer can provide additional benefits to the soil such as improving its physical properties.

Seeds planted in unamended filed soil (control) recorded some of the lowest germination rates throughout the three experiments which had to due to the sandy nature of the soil. The field soil’s physical properties consisted of 80% sand and tended to compact in the pots following irrigation events. Lack of oxygen and poor water holding capacity resulting from compressed mineral soil likely reduced germination. However, when this same soil was amended with digested hemp, the compaction problem was alleviated and more seeds germinated suggesting a positive impact of hemp fertilizer on the physical properties of soil. Overall, these findings indicate that the 2-week cure phase is necessary to decrease phytotoxic effects and that digested hemp could potentially improve the properties of sandy soils which has important implications for field application.

Disease Suppression

Finished composts have historically been used to control soilborne pathogens such as Pythium spp., Phytophthora spp. and R. solani through competitive biocontrol (Neher, Fang, and Weicht 2017; Noble and Coventry 2005; Tuitert, Szczech, and Bollen 1998). These pathogens are notoriously difficult to control because many survive as saprophytes and become pathogenic when conditions favor disease (Kato et al. 2013). Because there is only one pesticide currently approved for use on cannabis plants in the US, pathogen suppression by the digested hemp fertilizer would represent an innovative value-add to farmers who employ the Imio recycling strategy (Punja 2021). Plate competition assay results indicate that cured material was more suppressive against fungal pathogen growth than uncured material suggesting the cure phase is not only necessary to reduce phytotoxic effects but also stabilize decomposition processes and inhibit disease. These results agree with Chung, Hoitink, and Lipps (1988) who report that the level of decomposition of organic material is related to its antagonistic effects against R. solani with more decomposed material generally more effective at controlling pathogen growth. Despite these findings, even the hemp waste treated with water (negative control) suppressed the growth of R. solani comparable to the other treatments. This result indicates that decomposing hemp in general could be used as a method to control certain soilborne pathogens regardless of microbial treatment which is consistent with previous reports using cannabis extracts to control R. solani (Hussain et al. 2014; Kumar et al. 2013).

Plant Growth

Regardless of microbial formulation and delivery method (liquid or bran), digested hemp supported plant growth through the first 20 days of the trial suggesting its ability to be used as a fertilizer substitute. Differences in hemp plant height between liquid and bran formulations occurring after day 20 are likely due to the additional nutrients (particularly N) provided by the bran itself and the relatively small pot sizes used in this experiment. The 2.4 L pots may have limited root growth and amount of nutrients accessible to plants grown solely in digested hemp past 20 days. These factors would not be limiting in a field setting, however, suggesting that digested hemp generated using the MI4.1 liquid inoculant could serve as an organic fertilizer substitute granted enough of it is provided. Despite containing considerably less N than bran formulations, MI4.1 liquid treated biomass promoted greater R:S ratios than the other treatments.

Greater root mass when compared to shoots is an indicator for microbial interactions within the rhizosphere that induce root growth and alter root architecture in ways that benefit both microbes and plants (Pereira et al. 2020; Zahir, Arshad, and Frankenberger 2004). As a seedling, increased root mass increases the probability of roots encountering nutrients and reflects increased volume of habit for microbial interactions in soil generally supporting healthier plant growth (Beneduzi, Ambrosini, and Passaglia 2012; Kaymak 2010; Zahir, Arshad, and Frankenberger 2004). The greater R:S ratios observed in plants grown with MI4.1 digested hemp suggest successful colonization of beneficial microbes (e.g., B. subtilis) in the rhizosphere and indicate a potential growth promotion benefit of using Imio formulations. However, it is important to acknowledge that some of this effect could be due to the nature of the hemp fertilizer and organic growing techniques in general that are designed to increase root growth.

Organic growing techniques (e.g., applying organic amendments, biostimulants) are employed to promote root growth and encourage interactions with soil microorganisms (Adugna 2016; Khosro et al. 2011). It is through these interactions that organic matter can be mineralized and made available to plants. In contrast, using synthetic fertilizers typically reduces root growth because the plant need not scavenge nor partner with microbes when nutrients are provided in mineral forms (Yan and Gong 2010). This phenomenon could partially explain the reduced R:S ratios observed in the synthetically fertilized control group and the larger ratios found in plants grown organically. It is also possible that the differences are due to a combination of factors such as availability and amount of nutrients in the soil, relationship with soil microorganisms, and plant-to-plant variation to name a few. Regardless, root interactions with soil microorganisms likely caused the stunted growth observed in plants grown with both digested hemp and synthetic fertilizers.

Competition between roots and rhizosphere microbes can result in a deleterious effect on plant growth because microbes are more efficient assimilators of soil nutrients (Kuzyakov and Xu 2013). The combination of digested hemp and synthetic fertilizers may have caused a bloom in soil microorganisms due to the influx of available nutrients. Pools of mineral nutrients create a competitive environment in which microbes can outcompete plants for resources—ultimately resulting in stunted plant growth. However, this is only speculation and further experiments are needed to assess the effect of combining digested hemp fertilizer with other nutrient sources.

Overall, the digested hemp fertilizer performed comparably to the synthetic control especially when the microbial formulation included bran. Despite the benefits of bran as a plant fertilizer, digested hemp alone seems to provide enough nutrients to sustain plant growth granted enough of it is provided. Potential plant growth promotion benefits are also obtainable when growing in digested hemp treated with Imio inoculants because plants grown using MI4.1 recycled biomass had greater R:S ratios than the synthetic control. Taken together, the Imio recycling strategy could offer farmers a sustainable solution for handling their waste with the added benefits of generating an organic fertilizer source on-farm that can serve as a substitute for synthetic nutrients.

Conclusions

The current production model of medical hemp is not sustainable because it overlooks the potential for crop residues to replenish fertility and improve soil health. This study provides empirical data to support the adoption of microbial inoculants for on-farm recycling of cannabis waste to provide economic and environmental returns to farmers in the form of fertilizer and soil organic matter. Microbial digestion of cannabis biomass not only produces a more consistent output than traditional composting but also requires less labor and reduces the recycling period from 8–12 months to roughly 5 weeks. Additional benefits of biocontrol and plant growth promotion are also obtainable when recycled hemp is applied as a fertilizer.

Acknowledgments

We thank Thomas R. Weicht for technical assistance in conceptualization and design of experiments.

Disclosure Statement

Victoria I. Holden is a co-founder of Imio, Inc. (formerly Full Circle Microbes, Inc.) and holds partial ownership in the company.

Additional information

Funding

This work was support by the National Science Foundation Small Business Innovation Research under Grant #2014792 to Imio Technologies, Inc. (formerly known as Full Circle Microbes, Inc.).