PHOSPHORUS DISTRIBUTION WITH RESPECT TO PARTICLE SIZE AND MICROBIAL ACTIVITY IN DAIRY MANURE AND IMPLICATIONS FOR MANURE MANAGEMENT by Charles Bradshaw B.Sc. Chemistry–University of Victoria, 2008 B.Sc. Biology–University of Northern British Columbia, 2015 THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE IN NATURAL RESOURCES AND ENVIRONMENTAL STUDIES (ENVIRONMENTAL SCIENCE) UNIVERSITY OF NORTHERN BRITISH COLUMBIA April 2024 © Charles Bradshaw, 2024 Abstract This work investigated the distribution of phosphorus in dairy manure as it relates to particle size and provides experimentally derived insights for improving the management of phosphorus. The data presented here illustrates some aspects of manure handling that may be optimized for controlling the distribution of phosphorus. It provides experimental data that supplements existing literature by 1) extending the limits of study regarding particle size distribution; 2) extending the limit of centrifugal forces studied; 3) demonstrating an improved pathway for phosphorus management in wastewater using a substrate mediated flocculation process; and 4) providing some experimental observations showing that the microbiota present in dairy manure can diminish stratification of suspended solids and phosphorus concentrations within separated liquid manure products. Further, based on quadratic modeling of phosphorus relative to sieve size, and the observations that accompanied progressive washing and mechanical agitation performed, a convincing argument for relating phosphorus distribution to the surface area of a substrate is made. The findings imply that much of the phosphorus present in manure is superficially bound to larger particulates, rather than an intrinsic or chemical part of the bulk material itself, and can be mediated by washing or mechanical means where the phosphorus present in coarser fractions can be shunted towards finer or liquid fractions. Approximately 80% of phosphorus was associated with particulates bound to larger substrate. Indeed, 95.5% of the phosphorus in the bulk manure was removable as particulates with cross sectional diameter less than 2mm by simple washing. Further removal of phosphorus in >2mm substrate through fermentation and leaching was also demonstrated. Some portion of this distribution in phosphorus appears to be dominated by microbes. Keywords: Phosphorus, manure management, dairy waste, eutrophication, wastewater, particle size distribution, substrate mediated flocculation, substrate mediated filtration. ii Table of Contents Abstract .......................................................................................................................................... ii Table of Contents ......................................................................................................................... iii List of Figures ............................................................................................................................... iv List of Tables ................................................................................................................................ iv List of Abbreviations .................................................................................................................... v Acknowledgements ...................................................................................................................... vi Dedication ..................................................................................................................................... vi Forward ....................................................................................................................................... vii Introduction ................................................................................................................................... 1 Materials and Methods ................................................................................................................. 4 2.1. Materials .............................................................................................................................. 4 2.2. Methods................................................................................................................................ 6 2.2.1 Analytical Procedures .................................................................................................... 6 2.2.2 Experimental Methods ................................................................................................... 7 Experimental Procedures and Results ...................................................................................... 12 3.1 Phosphorus Distribution as a Function of Particle Size ...................................................... 12 3.2 Particle Size and Phosphorus Distribution in the <63µm Liquid Fraction ......................... 24 3.3 Effects of Microbial Activity on the Distribution of Phosphorus in the Liquid Fraction ... 28 3.3 Rates of Sedimentation in <63µm Liquid Fraction and its Relation to Microbial Activity 32 3.4 Effects of Leaching and Fermentation on the Availability of Phosphorus from Solids ..... 38 3.5 Managing Phosphorus in Liquid Manure from a Microbiological Perspective .................. 41 3.5.1 Centrifugation .............................................................................................................. 42 3.5.2 Substrate Mediated Flocculation and Sedimentation ................................................... 45 3.5.3 Substrate Mediated Filtration ....................................................................................... 49 Discussion..................................................................................................................................... 53 Conclusion ................................................................................................................................... 64 References .................................................................................................................................... 65 iii List of Figures Figure 1: Manure reception pit at Vanmar Farms,Vanderhoof, BC. .............................................. 4 Figure 2: Stack assembly and manure products of “Sieved” treatment ........................................ 14 Figure 3: Manure products of “Rinsed” treatment ........................................................................ 15 Figure 4: Manure products of “Washed” treatment ...................................................................... 16 Figure 5: Relative distribution of oven dry solids in sieve fractions after each treatment ........... 17 Figure 6: Relative distribution of phosphorus in oven dry sieve fractions after each treatment .. 19 Figure 7: Relative moisture content of sieve fractions after each treatment ................................. 20 Figure 8: Oven dry phosphorus concentration in sieve fractions after each treatment ................. 21 Figure 9: Relative reduction in phosphorus in each sieve fraction after treatments ..................... 22 Figure 10: Relative distribution of phosphorus in each sieve fraction after treatment ................. 23 Figure 11: Log scale distribution of particles observed in <63µm liquid manure fraction .......... 24 Figure 12: Abundance of bacteria observed in <63µm manure matrix at 1000x magnification .. 27 Figure 13: Chemotaxis experiment for the determination of phosphorus in motile bacteria ....... 29 Figure 14: HPC of organisms in pair-controlled chemotaxis replicates ....................................... 30 Figure 15: Effect of microbial activity on stratification of TS in <63µm liquid over 8 days ....... 33 Figure 16: Effect of microbial activity on stratification of Total P in <63µm liquid over 8 days 33 Figure 17: Temporal and depth profile of phosphorus in undisturbed diluted liquid manure ...... 34 Figure 18: Effects of microbial activity on appearance of stratification and turbidity ................. 36 Figure 19: Effects of microbial activity on the stratification of total phosphorus ........................ 37 Figure 20: Leaching trials of phosphorus in >2mm “Washed” sample ........................................ 39 Figure 21: Effects of centrifugation on solids and phosphorus in supernatant ............................. 42 Figure 22: Effects of centrifugation on HPC in supernatant......................................................... 44 Figure 23: Observations of photosynthetic organisms in high-speed centrifuge supernatants ..... 45 Figure 24: Substrate mediated flocculation and filtration using different additives ..................... 46 Figure 25: Effects of additives on phosphorus in diluted liquid manure after 24hrs .................... 47 Figure 26: Effects of sediments on HPC in liquid manure supernatant after 24hrs ..................... 48 Figure 27: Phosphorus reduction by substrate mediated flocculation .......................................... 49 Figure 28: Relative ease of filtration after treatment with individual additives ........................... 51 Figure 29: Relative ease of filtration with substrate mediated flocculation ................................. 51 Figure 30: Microscopic comparison of flocs with and without substrate additive ....................... 52 List of Tables Table 1: Mean particulate size in the <63µm fraction below the 10th, 50th, & 90th percentiles .... 25 Table 2: Total P and HPC in liquid manure products discriminated based on size ...................... 26 Table 3: Results of paired chemotaxis study reflecting mean differences in HPC and Total P ... 30 Table 4: Summary of phosphorus leached from samples following incubations ......................... 39 iv List of Abbreviations Aluminum sulfate American Chemical Society American Society for Testing and Materials Centrifugal force equivalence to gravity (~9.8m/s3) Colony forming units Distribution of particulates less than a particular percentile Environmental Protection Agency Heterotrophic plate count Hour/hours Logarithmic scale of acidity Moringa oleifera Most probably number Nitrogen to phosphorus ratio Northern Analytical Laboratory Services Not applicable Oven dry Particle size analysis Plate count agar Standard Methods for the Examination of Water and Wastewater Calculated student’s t-test statistic Student’s t-distribution critical value Total phosphorus Total solids Total suspended solids University of Northern British Columbia Weight per volume basis Alum ACS ASTM g cfu Dx EPA HPC hr/hrs pH moringa MPN N:P NALS N/A OD PSA PCA SM Tcalc Tcrit Total P TS TSS UNBC (w/vol) v Acknowledgements ❖ To my supervisor, Dr. Hossein Kazemian, and committee members Dr. Jianbing Li and Dr. Mike Rutherford for their patience and constructive suggestions while completing this work. ❖ To Allan Martins (Vanmar Farms in Vanderhoof, BC) and Dave, Rich and Joe (Cedarwal / Vanderwal dairy farms in Abbotsford, BC) for kindly accommodating my request for samples and showing me their dairy operations. ❖ To the Northern Analytical Laboratory Service (NALS) team at the University of Northern British Columbia (UNBC) for their dedication in providing accurate and timely analytical results for which my work depended on (Erwin Rehl, Mya Schouwenburg, and Dorna Sobani) ❖ To Chalan Ozmen (NPower Clean Tech) whose Mitacs project inspired a further investigation into the relationship of particle size and phosphorus distribution in dairy manure and the importance of phosphorus management. Dedication ❖ To my love, Rebecca, and our darling children Cassidy, Charlotte and Julianna… vi Forward This thesis study was born out of two observations arising from a Mitacs project funded by Chalan Ozmen from NPower Clean Tech. In an investigation of nutrient capture using anionic adsorption media, it was realized that the majority of phosphorus present in liquid dairy manure was actually present as solids, which were removable by filtration. It was also apparent that solid-liquid separating processes commonly employed did not fully separate solids from liquids and that the liquid portion had dry solids content much higher in total phosphorus concentration. This implied that solid-liquid separating processes may impart differing proportions of phosphorus in either fraction depending on the efficiency of separation. One of the objectives of this study was to characterize the distribution of phosphorus in manure by particle size with varied levels of separation efficiency to discern how much room for optimization exists in a solid-liquid separating process. The second observation that arose in the project was that the amount of dissolved phosphorus present in a liquid manure sample was not fixed. Dissolved phosphorus could be depleted with an anionic adsorbent and then its concentration would increase over time once the adsorbent was removed. This implied that the species of phosphorus present were subject to change over time, likely as a consequence of microbial activity. The second objective of this study was to discern what effects microbial activity may have on the distribution of phosphorus, particularly in regard to liquid separated manure products. vii Introduction Phosphorus is disproportionally excreted in the feces fraction of manure while potassium and nitrogen are predominately excreted in urine1. The initial concentration of these mixed waste streams start appropriately proportioned as a fertilizer for crops2. However, during storage much of the nitrogen is lost to the atmosphere, which results in a somewhat excessive proportion of phosphorus relative to crop demands1–4 . Further, when used as a fertilizer, excess phosphorus not used by the plants will accumulate in the soil and can spread into groundwater and runoff2,3,5. This limits when manure can be applied and how much manure can be applied per acre. For optimum yields, supplemental nitrogen may also need to be applied to offset the imbalance2,3. These problems typically lead to increased trucking costs and other strategies for improving the ratio of N:P with exogenous fertilizers. This problem greatly affects farms limited by area for land application. An additional problem with phosphorus is that it is often the limiting nutrient responsible for eutrophication in water ways6. Small amounts can stimulate growth of photosynthetic micro- and macroorgansims6 causing nuisance blooms of different organisms and adverse consequences to the environment. Limiting phosphorus in runoff is of major importance. Many jurisdictions impose requirements on the handling, storage, and land application of manure to limit externalities of excess nutrients. For instance, in British Columbia manure management falls under the Environmental Management Act Code of Practice for Agricultural Environmental Management. Among other things, it stipulates the responsibilities of the farmers, designates high-risk areas and conditions for application, as well as the requirements of a nutrient management plan7,8. For this reason, phosphorus is a key focus of manure management. 1 Much of the literature involving manure management of phosphorus relates to its mitigation through the management of solids, the different types of phosphorus present, and the cost-benefits of implementing different technologies1,3,4,9. Bacteria and other microbes are secondary considerations, as if harboured by the same substrate, but not major contributors to the overall distribution of phosphorus themselves. This perspective is likely born out of the technologies used to manage solids, which demonstrate clear changes in phosphorus relative to visually observable changes in the distribution of solids. The common mechanical processes employed to control solids and manage nutrients in manure include screens, screw presses, belt presses, and centrifuges as well as membrane filters and dryers for removing water4,10. However, knowing that phosphorus disproportionately associates with smaller particulates, the question of how much of this macroscopic observation is associated with the bulk material itself and how much may be associated with residual manure liquids and particulates superficially retained on larger substrates is a worthwhile question to answer. One underexplored aspect of phosphorus distribution is from the effect of microbes. Obviously, microbes are the dominate pathway ruminants use to transform plant nutrients into usable components for their metabolism11,12. Most of those metabolic functions that break down plant material continue in the manure products well after excretion. For instance, Moller et al. found that the proportion of small particulates decrease over time relative to large particulates as a consequence of microbial activity during storage13. How they affect manure management in terms of phosphorus should also be considered. Do microbes constitute a considerable portion of the phosphorus present at all scales of particulate size in manure or are they occupying substrate with intrinsically higher levels of phosphorus not related to the microbes themselves? If the phosphorus is intrinsically associated 2 with the chemical and material properties of the substrate, it reinforces the paradigm of managing phosphorus through solids. Conversely, if the phosphorus present is more superficially bound to substrate, as small particulates or as microbes adhering to larger particulates, other management perspectives may be important too. This study attempts to discern how much phosphorus is present as superficially bound particulates at different scales, how much of that may be due to microbes and what management strategies might be considered from a microbiological perspective. 3 Materials and Methods 2.1. Materials The dairy manure sample used in this study to profile the distribution of phosphorus with respect to particle size was collected from Vanmar Farms in Vanderhoof, BC on September 24, 2023. The farm uses a passively distributed manure reception pit that collects the liquid scrapings from semi-confined lactating Holstein cows. Figure 1 shows the collection site directly adjacent to the barns outflow. The grab samples were collected using clean 20L buckets, which were stored outside until the sieve profiles were completed by late October 2023. Figure 1: Manure reception pit at Vanmar Farms,Vanderhoof, BC. The sieves used were Standard 8” ASTM E11 stainless steel sieves purchased from VWR/Avantor. All rinse water and waters used in analyses were Type I derived from a 4 Millipore Milli-Q IQ7000 deionized water unit. Particle size analysis (PSA) was performed using a Malvern Laser Diffraction 3000 series Mastersizer. Total phosphorus was assessed using an Agilent Technologies 5100 inductively coupled optical emissions spectrophotometer (ICPOES). Hydrochloric acid and nitric acid were trace metals grade Aristar©Plus purchased from VWR/Avantor and digestions were performed in a 15mL DigiPrep MS hot block from SCP Science. Multi-element standards were purchased from High-Purity Standards. The aluminum sulfate octadecahydrate (Alum) used for flocculation was from ThermoScientific, the sodium hydroxide from Anachemia, and potassium phosphate monobasic from VWR Life Science. Each was ACS grade. The cell counting chamber was a Neubaauer-Improved from Marienfeld, Germany. The microscope used was an Olympus CH30. Bernardin 1L mason jars were used for the flocculation and settling studies. Drying was performed using a Heratherm model oven from ThermoScientific. Masses were weighed on either a Sartorius Secura225D-1S analytical balance or an AND FX3200 top load balance verified against ASTM class 1 calibration weights. Centrifugation experiments were performed in a high speed Sorvall RC6+ centrifuge from ThermoScientific using a FIBERLite F18-12-50 rotor. Heterotrophic plate counts (HPC) were performed using HIMEDIA brand plate count agar (PCA) in standard 100mm Petri dishes purchased from VWR and Quanti2000 tray format HPC media from IDEXX Canada. Incubations were conducted in a Heratherm incubator from ThermoScientific. Sterilization of materials was performed in an Getinge Castle 500LS series steam sterilizer. The Moringa oleifera (moringa) powder, Floor-Dry brand 100% diatomaceous earth, and clay samples used in the sedimentation trials were left over materials from prior projects with unknown purity and quality. Standard 47mm glass microfiber filters for determining filterability and total suspended solids (TSS) were Whatman 934-AH, syringe filters were 25mm diameter 0.45µm nylon filters 5 from Restek, and the 47mm defined 0.2µm pore size membrane filters were Nucleopore TrackEtch Whatman brand filters, all purchased from VWR/Avantor. 2.2. Methods The analytical tests employed in this study included assessments of total phosphorus (Total P) and dissolved phosphorus; gravimetric assessments of dry matter content or total solids (TS); bacteriological counts using heterotrophic plate count (HPC) media; particle size discrimination through sieving; and profiling distributions of microscopic particles through laser diffraction; as well as manual measurements of distance, mass, volume, and time. Experiments included chemical treatments, centrifugation, dilutions, and other manipulations of liquids, as well as practices for culturing, fermenting and sterilizing microbial solutions. The methods for the pertinent analytical procedures used are described succinctly below. All analytical testing was performed at the Northern Analytical Laboratory Services (NALS) department located at the University of Northern British Columbia (UNBC). Experimental procedures were performed as described. Specific analytical tests used to generate data are described first. The broader experimental methods employed in this study are described second, in an order relating to the Experimental Procedures and Results section. All calculations and graphing were performed using Microsoft Excel 2016. 2.2.1 Analytical Procedures All assessments of phosphorus were determined by ICP-OES via EPA method 200.7 following acid digestion of representative samples sampled and prepared as per procedures described in EPA method 200.2 by qualified personnel at the NALS. All assessments of TS, TSS, and moisture content were performed gravimetrically using an analytical or top load balance by subtracting weights of containers against the containers with their prescribed 6 contents. Removal of water was performed in a lab oven held at 105oC for at least 24hrs and the mass loss was presumed to be entirely water for the purposes of calculating moisture or dry matter content. Similarly, total solids in water were assessed gravimetrically as per procedures described in the Standard Methods for the Examination of Water and Wastewater6 (SM) section 2540B. HPC was performed using two different procedures. Representative samples were collected by pipet and diluted sufficiently using phosphate buffer before being assessed on standard plate count agar (PCA) by procedures described in SM 90006. Known aliquots of sample were spread plated and the concentrations were back calculated based on the dilutions incurred once sufficient incubation time for counting colony forming units (cfu) had passed. Procedures followed SM 9215C6. However, one assessment of heterotrophic plate count (performed for the substrate mediated flocculation study) used a proprietary enzyme defined substrate counting system by IDEXX that counts fluorescent positive wells in a multi-well tray against a most probable number (MPN) table. Procedures and quantification were followed as per directions from the manufacturer. 2.2.2 Experimental Methods Phosphorus distribution as a function of particle size used standard 8” diameter soil sieves to partition bulk manure grab samples into fractions defined by square sieve hole length. This process would roughly discriminate against particles and aggregates based on their crosssectional diameter and the degree of mechanical agitation, pressure, and water applied during treatments (described more thoroughly in the Experimental Procedures and Results section). The sieve stack was arranged from largest size on top to smallest on bottom (2, 1, 0.5, 0.25, 0.125 and 0.063mm) with a lid and final collection pan for liquids. The liquid portion from sieving was regarded herein as the <63µm liquid manure fraction. Sieve treatments were applied 7 sequentially to the same sample and at each stage of treatment a small subsample was collected from each size fraction for assessment of its total solids and total phosphorus content. The sieving was done manually by oscillating the stack vigorously in a horizontal plane for 10 minutes and then periodically in successive steps as sieves were removed. Pressure used to dewater, break up aggregates and push small particles progressively through each sieve was done with a rubber spatula and modest hand pressure so as not to damage the sieves. This first stage of treatments was referred to as the “Sieved” treatment. The rinsing treatment used type I deionized water partitioned into three equal parts and the sieving proceeded as before with progressive movement of liquid and smaller particles being sieved through the stack. This was referred to as the “Rinsed” stage of treatment. The thorough washing step was different. Each sieve fraction was washed separately with a garden hose until the solution arising from that fraction ran clear. As a result, smaller particles and liquid did not transverse the subsequent sieves. To limit any contamination from phosphorus caused by the tap water used in this stage, each washed fraction was re-washed with additional portions of deionized water before excess water was pressed out. This process was called the “Washed” treatment. Three replicates of manure were processed in this way and the average oven dry solids, phosphorus concentration and moisture were related to the total oven dry matter determined gravimetrically. For subsequent tests of the sieved liquid manure product, the <63µm fraction was pooled from the three “Sieved” replicates for all subsequent tests. The <63µm fractions from the “Rinsed” and “Washed” treatments were discarded. The particle size distribution of the pooled <63µm liquid manure fraction was assessed using a Malvern laser diffraction PSA equipped with HydroLV attachment using an obscuration 8 between 4-20 and modeled for non-spherical particles with refractive indices of 1.47, no sonication, and only deionized water as a dispersant. The defined pore size filtrate was collected directly by filtering a small portion of <63µm liquid manure by vacuum through a 0.2µm membrane filter into a tared digestion tube. The 0.45µm fraction was collected simply by pressing a small portion of <63µm manure into a tared digestion tube using a 0.45µm syringe filter. Subsamples for HPC were drawn off from these samples and then spread plated immediately on PCA media. The unfiltered <63µm manure was serially diluted in 1.36g/L KH2PO4 buffer solution adjusted to pH 7 with 0.1M NaOH before being spread plated on PCA and incubated to attain the HPC of each solution. These bacteriological assessments all followed procedures described in the SM for HPC. A settling experiment to relate the stratification of total solids and phosphorus over time was done by simply pouring a homogenized <63µm manure fraction into a 50mL burette and allowing it to sit covered for a prolonged period. Defined volumes were then poured off from the bottom of the burette while the meniscus distance from the stopcock was measured to ascertain heights of the collected fractions. The samples were then dried in tared digestion tubes to determine their total solids and then acid digested in the same tube for simultaneously determining total phosphorus. The settling experiment was also repeated with a sterilized portion of liquid manure by simply autoclaving the whole assembly once filled. Note that some volume loss occurred while autoclaving due to evaporation. This minor difference was ignored but likely contributed to an increased bias in total solids and total phosphorus. To have more temporal clarity of how the liquid manure fraction may stratify over time without possible effects of particle-particle interferences that may mechanically inhibit settling rates, a settling experiment using a diluted <63µm liquid manure fraction was also proposed. A 9 mason jar was filled with a 10/1000 dilution of liquid manure in deionized water, shaken and allowed to settle undisturbed. Small samples of the supernatant were then taken using a micropipette just below the surface at periodic intervals for determining any changes in phosphorus over time. At 24hrs, a number of samples were also taken at prescribed depths with care not to disturb the solution. The experiment was continued for four days and phosphorus concentrations were plotted accordingly. The effects of microbial activity were characterized in three ways. One way included the settling study between <63µm manure and an autoclaved version meant to see how microbes might affect the distribution of phosphorus and settled solids in a stratified solution. The other experiments were meant to quantify how much phosphorus is mobile in motile bacteria and how that may affect the stratification of phosphorus and solids over time. A method for measuring chemotaxis using a capillary was adapted to simultaneously count the number of organisms passing through a solution interface and the associated phosphorus that moves with them. The volume had to be scaled larger for determination of phosphorus by ICP-OES and instead of a cell counter, the concentration of microbes was assessed by spread plating an appropriate dilution on PCA plates and counting colony forming units. To do this, a portion of diluted <63µm manure was drawn into a narrow 1mL serological pipet with sufficient room to add 100 µL of 10-3M glucose solution to the manure interface as an attractant at one end. Approximately 1hr was allowed to pass to allow for motile bacteria to move towards the attractant. Replicates with both the 100 µL of attractant removed and retained were then split in half, so that the farthest 0.5mL could act as a control while the nearest half could be assessed for an effect. A measured portion of each half was then acid digested and assessed for total phosphorus while a sub-sample was diluted sufficiently and assessed for HPC. 10 The replicates with the 100µL of attractant removed were then assessed using a paired t-test at the 95% confidence interval to see if their total phosphorus and bacteria counts were statistically different. The same was done with the replicates having their 100µl attractant retained to verify that they should not be different. The difference in phosphorus concentrations was intended to be divided by the difference in observed bacteria to arrive at a concentration of total phosphorus per motile bacteria that was discarded in the glucose attractant. However, the bacterial counts were found to not be statistically different, and the experiment was abandoned after a few attempts due to a lack of precision in the methodology. An experiment for qualitatively assessing the effects of microbial activity on phosphorus distribution was formulated using a test tube comparison. Three test tubes were filled with 10/1000 dilution of <63µm liquid manure. One was left untreated as a control while the other two were autoclaved. Immediately, after the autoclaved test tubes had cooled to room temperature, 100µL of the live control solution was inoculated back into one of the sterile tubes. The three tubes were then left undisturbed for several days and observed. The sterile tube began to clarify as suspended material at its top appeared to slowly descend. After 9 days, similar to the settling study of the undiluted <63µm manure fraction, the three test tubes were carefully sampled to determine phosphorus concentrations at depths which bracketed the perceived changes in the sterile tube. The samples were drawn carefully so as to keep the sterile tube stratified. Then, immediately after sampling, a 100µL sample from the live control was gently pipetted at the top of the sterile test tube. The test tubes were then left undisturbed and observed for two more days before re-sampling at the same depths to determine the effects of microbial activity on stratification and phosphorus distribution. 11 Experimental Procedures and Results Manure handling is managed with pragmatic and economical considerations suited to the specific demands of each farm, the equipment available, and the conditions of where its generated, stored and spread. The processes involved may impart different properties and characteristics to the manure. A sieving experiment to help evaluate the phosphorus distribution as it relates to particle size was devised where the effects of mechanical agitation and washings could be illustrated. Some follow-up experiments to then characterize the coarsest and liquid manure fractions generated were also carried out for the purpose of investigating how these manure products might be further processed for the purpose of managing phosphorus. The experimental procedures and results are described below. 3.1 Phosphorus Distribution as a Function of Particle Size A replicate of wet manure samples ranging from 1.2-1.5kg was sieved by mechanical means through progressively smaller sieves with increasing mechanical agitation and washings in three sequential stages to see how phosphorus would distribute and what insights could be gained towards optimization of phosphorus removal for on-farm practices. The treatments included 1) a moderate pressing of manure directly through sieves to represent a solid-liquid separation process with modest input pressure; 2) supplementation of the sieving process with a limited supply of water to facilitate greater separation of surface-bound particulates; and 3) excessive washings to demark the maximum extent to which aggregates and substrate bound particulates with smaller cross-sectional diameter could be removed from larger sieve fractions. The sieve sizes chosen decreased progressively in cross sectional area by a factor of four. The stack included standard 2mm, 1mm, 0.5mm, 0.25mm, 0.125mm and 0.063mm sieves as well 12 as a collection pan, regarded herein as the “<63µm fraction”. The tare weight of each sieve and the stack were determined for gravimetric calculations. A grab sample from a homogenized bucket of manure was then loaded onto the largest sieve of the stack and weighed to find the net weight of addition. The stack was then vigorously oscillated manually for ten minutes to facilitate transfer of smaller particulates progressively through the stack. Then, in intermittent steps one sieve at a time, the residual manure in the top sieve was lightly pressed with a rubber spatula to squeeze out excess moisture and push smaller particulates through the sieve. The stack was then closed and oscillated again for a few minutes before a second attempt to force moisture and particulates through the top-most sieve. Once reaching a perceived end point, the largest sieve was removed and weighed to determine the net weight of wet manure remaining in that sieve. That sieve was then removed and the processed was repeated for each remaining sieve fractions down to the final 0.063mm sieve. 13 Sieved Treatment >2mm 0.25mm 1mm 0.125mm 0.5mm 0.063mm Bulk Figure 2: Stack assembly and manure products of “Sieved” treatment Figure 2 shows the sieve stack with grab sample loaded as well as each fraction after being sieved and mechanically pressed. Note the very liquidy nature of the product in the smallest sieve used. This treatment represents a modest mechanical separation of solids according to sieve size and a liquid fraction generated without excess pressure or further washings imposed. The intent of this treatment was to replicate what might be expected from a simple mechanical separation process, like screening. The second treatment proceeded as the first on the same material, but with the addition of a small portion of water to facilitate the mechanical separation of aggregated clumps and surface bound particulates. The <63µm fraction was poured off and collected while the other sieve fractions of manure were flattened out before reassembling the sieve stack. A 333mL aliquot of 14 deionized water was then added to the top-most sieve before manually oscillating the entire assembled stack for another ten minutes. The process was repeated two more times (for a total of 1L of rinse water) with the <63µm fraction being removed after each addition. Each sieve fraction was then mechanically pressed again with a rubber spatula to push excess water and smaller particulates progressively through the sieve stack before being weighed to determine their residuals. This added rinsing was meant to represent an increase in mechanical processing applied for the purpose of reducing smaller particulates bound to larger particulates. The added processing was meant to better distribute particulates according to their cross-sectional diameter and represents a more vigorous separation process that might be attained by a mechanical press. Figure 3 shows each fraction after the additional mechanical agitation and rinsing steps. Note the changing consistency of the manure products between the successive treatments. In particular, the liquidy nature of the 0.063mm fraction was greatly diminished. Rinsed Treatment >2mm 0.25mm 1mm 0.125mm 0.5mm 0.063mm Figure 3: Manure products of “Rinsed” treatment 15 The third treatment was meant to represent the maximum degree for which separation of surface bound particulates could be attained by simple washing and/or mechanical processes. The <63µm collection pan was removed and each sieve fraction was washed successively with a garden hose until the solution from each sieve fraction ran clear. Following this thorough washing, each sieve fraction was washed three more times with a portion of deionized water and pressed with a rubber spatula to remove excess water. Once pressed, the sieves were then weighed to determine the residual wet weight of material. Expectedly, these fractions should have very few particulates with cross sections smaller than their respective sieve size and should approximate an end point for possible improvements to the distribution of manure particulates according to size. Figure 4 shows the “Washed” sieve fractions. Note the change in colour and consistency of the residuals. The coarsest of these sieve fractions may also represent a reasonable feedstock for green bedding once dry. Washed Treatment >2mm 1mm 0.25mm 0.125m m 0.5mm <0.063mm Figure 4: Manure products of “Washed” treatment 16 At each stage of the separation processes, a small portion of each fraction was collected for the purpose of determining its moisture content, total solids and total phosphorus. The results of this sieve analysis are summarized in Figures 5-10 below with the treatments described above being identified as “Sieved”, “Rinsed” and “Washed”. The error bars represent one standard deviation from the mean calculated from the three replicates performed. Note that the <63µm fraction for the “Rinsed” and “Washed” treatments were not collected as their wash volumes were not directly comparable between replicate trials due to somewhat different manure masses and the varied wash volumes. Oven Dry Weight Percent (%) 100 90 Sieved 80 Rinsed 70 Washed 60 50 40 30 20 10 N/A 0 2mm 1mm 0.5mm 0.25mm 0.125mm 0.063mm <0.063mm Sieve Size Figure 5: Relative distribution of oven dry solids in sieve fractions after each treatment 17 Most of the solids were found in the coarsest >2mm fraction. Figure 5 shows the percent distribution of dry solids decreasing with the progressive treatments imposed. The improved separation of smaller bound (or aggregated particulates) in the >2mm fraction had a cascading effect on subsequent sieve fractions leading to more solids being present in their “Rinsed” treatment compared to the “Sieved” treatment. Each fraction was then reduced further in dry solids content in the subsequent “Washed” treatment. The trend in total phosphorus relative to sieve size is somewhat different. Figure 6 shows a similar decline in phosphorus relating to sieve size down to approximately 0.25mm, but then a relative increase in the subsequent fractions. The distribution of phosphorus compared to solids is weighted much more heavily towards smaller particle sizes. Further, in relation to the coarsest fractions described in Figures 5 and 6, washing greatly reduced the abundance of phosphorus in a proportion much greater than the loss of solids that was observed. These two observations suggest that phosphorus is disproportionality associated with smaller particulates and that much of the phosphorus present in coarse fractions may actually be due to smaller particulates (or through some superficial adherence to the larger particles) capable of being washed through the remaining sieves. 18 100 Sieved 90 Rinsed Distribution of Total P (% OD) 80 Washed 70 60 50 40 30 20 10 N/A 0 2mm 1mm 0.5mm 0.25mm 0.125mm 0.063mm <0.063mm Sieve Size Figure 6: Relative distribution of phosphorus in oven dry sieve fractions after each treatment Although the end point for pressing was somewhat subjective and related to how much pressure could be applied without damaging the sieves, a clear trend in moisture content relating to particulate size was observed. Figure 7 shows the trend in moisture observed for the sieve fractions with respect to their wet weight. The data implies that smaller particulates hold more moisture when strained under similar pressures. This may have implications for controlling moisture during storage of manure managed products. 19 100 Sieved Rinsed 95 Washed Moisture Content (%) 90 85 80 75 70 65 60 55 N/A 50 Unsieved 2mm 1mm 0.5mm 0.25mm 0.125mm 0.063mm <0.063mm Sieve Size Figure 7: Relative moisture content of sieve fractions after each treatment The successive treatments caused a marked reduction of phosphorus in each sieve fraction. Figure 8 shows the oven dry concentration of phosphorus relative to the bulk manure as well as the residual phosphorus after each stage of treatment. Error bars represent one standard deviation from the mean of the three replicate trials. 20 50000 Sieved 45000 40000 Rinsed Washed Total-P (mg/kg OD) 35000 30000 25000 20000 15000 10000 Bulk Manure 5000 N/A 0 Unsieved 2mm 1mm 0.5mm 0.25mm 0.125mm 0.063mm <0.063mm Sieve Size Figure 8: Oven dry phosphorus concentration in sieve fractions after each treatment The rinsing treatment reduced total phosphorus in each fraction by nearly half and the subsequent washing stage reduced it further to less than 20% of its starting concentration for most fractions. Figure 9 shows the relative phosphorus remaining after each treatment. If the “Washed” treatment truly reflects a discrete size distribution of manure particulates, the data suggests that only about 20% of the phosphorus present is actually associated with particulates above 0.125mm and that the remainder would fall predominantly below the 63µm threshold used here. This likely has implications for phosphorus management via solids, where a desired reduction in phosphorus could be achieved by targeting a particular screen size or use of iterative progressive washings to push smaller particulates into liquid manure products. 21 100% Sieved Residual Phosphorus Following Treatments (%) 90% Rinsed 80% Washed 70% 60% 50% 40% 30% 20% 10% 0% 2mm 1mm 0.5mm 0.25mm 0.125mm 0.063mm Sieve Size Figure 9: Relative reduction in phosphorus in each sieve fraction after treatments Because of the prescribed constraints of cross-sectional diameter imposed on the sieving process, we can model the distribution of phosphorus based on the different treatments imposed as a function of sieve diameter. Note that the applicable range here is from <2mm to >0.063mm and should not be extrapolated to substrates outside the range imposed by the largest and smallest sieve sizes without caution. Figure 10 shows the plotted relationship of sieve size against the phosphorus distribution. Interestingly, the distributions are well explained by a quadratic function for each of the treatments, with a widening of the parabolas occurring with the greater mechanical agitation and washing efforts imposed. This may signify that progressive removal of phosphorus from a given sieve fraction is related to the squared function of some intrinsic property of the material present. Since much of the phosphorus present in the bulk manure appears to not be confined to the interstitial spaces within particles, but separable by 22 mechanical or washing processes, as suggested by the data here, the distribution of phosphorus in a given application may be related to the surface area of the particulates rather than some material or chemical property of the particles themselves. For basic geometric shapes, the surface area of a particulate can be approximated by some squared function of its cross-sectional diameter, radius or width that could be directly related to the hole diameter of each sieve fraction. This may explain the relationship modeled and the distribution of phosphorus observed. It also suggests that there is a limited ability to improve phosphorus distribution by simple mechanical means as the sieve size approaches the vertices of the parabola. 60 Relative Distribution of Total P (%) Sieved y = 19.768x2 - 18.839x + 7.235 R² = 1.000 Rinsed y = 6.546x2 - 4.436x + 3.684 R² = 0.996 50 Washed y = 1.629x2 - 1.286x + 0.528 R² = 0.995 40 30 20 10 0 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 Sieve Size (mm) Figure 10: Relative distribution of phosphorus in each sieve fraction after treatment 23 3.2 Particle Size and Phosphorus Distribution in the <63µm Liquid Fraction There is limited capacity for profiling the distribution of particulates in manure by mechanical means below 63µm, as smaller sieve diameters become impractical. Two alternate approaches were included here to better characterize the <63µm liquid manure fraction. Obviously, any amount of water present or added will affect phosphorus concentrations by simple dilution in the final collection pan. This precluded direct comparison of <63µm fractions collected after the “Rinsed” and “Washed” treatments where the input of water relative to manure mass was not precisely controlled. Thus, the focus of this follow-up study was only related to the <63µm fraction that passed into the collection pan from the original bulk manure collected during the “Sieved” treatment. This <63µm fraction was subjected to PSA using a Malvern Mastersizer 3000 to profile the size distribution of particles between 3200-0.01µm. The results were very interesting. Figure 11 shows the volume density of particulates present by size class in logarithmic scale. Figure 11: Log scale distribution of particles observed in <63µm liquid manure fraction 24 From this distribution profile, two things are immediately apparent. The majority of particulates, or micro-aggregates, are clustered around a mean of 13.4µm and there is almost no abundance of particulates below 0.2µm. The upper abundance was likely imposed by the use of the 0.063mm sieve, however, there is no clear reason for the lower boundary of 0.2µm. This data implies that, while the proportion of phosphorus present increases inversely with particle size discriminated by sieves, this trend does not continue indefinitely and appears to begin to diminish precipitously below ~20µm. The table below summarizes the mean particulate sizes for the proportion of particles below the 10th, 50th, and 90th percentiles for a replicate of readings taken. Table 1: Mean particulate size in the <63µm fraction below the 10th, 50th, & 90th percentiles Parameter Mean Std Dev RSD (%) Dx 10 (μm) 2.07 0.009 0.428 Dx 50 (μm) 13.4 0.09 0.641 Dx 90 (μm) 50.6 0.7 1.48 There is limited ability to directly assess the concentration of phosphorus with respect to particle size at this scale. However, filters with well defined pore diameter can offer discrete cutoff points for which the distribution of phosphorus can be further resolved. The table below summarizes the concentration of phosphorus found in the <63µm “Sieved” fraction as well as the concentration of total phosphorus found in the filtrate imposed by both a 0.45µm syringe filter and a 0.2µm membrane filter (the latter having a well-defined pore diameter). Each of these samples were also assessed for bacterial concentration using HPC media. The results are summarized in Table 2. 25 Table 2: Total P and HPC in liquid manure products discriminated based on size Size Discrimination Used 63µm sieve 0.45µm syringe filter 0.2µm pore diameter membrane Total P (mg/L) 18632 36.0 24.5 HPC Cell Count (MPN/mL) 6.9x107 <1 <1 From this data we can infer that very little of the phosphorus present in the <63µm liquid separated manure fraction was actually dissolved phosphorus. Almost all of it must be associated with particulates in some form. Speculatively, based on the size expectations of bacteria, the distribution of particles observed in Figure 11 and the large abundance of bacteria found in this liquid fraction, it may be that much of the distribution of phosphorus was governed by its association with bacteria and other microbes, rather than particulates (such as sediments or plant debris ending up in the liquid fraction). This observation likely has implications for phosphorus management where microbes or dissolved phosphorus are the principle focus of consideration. To help qualitatively assess the relative abundance of particulates versus microbes in the <63µm fraction, a drop of the liquid was diluted and inspected under an optical microscope. A Neubauer-Improved cell counting chamber was used for scale and volume with the expectation that the number of microbes and particulates in a given volume could simply be counted to estimate the proportion of particles to microbes present. This proved somewhat too difficult for a few reasons. Contents within the depth of the chamber could not all be kept in focus at the same time for counting. Secondly, only bacteria that were clearly motile could be distinguished from similar shapes which could be either abiotic/biotic debris or actual microbes that were either stationary or dead. Only a few of the larger irregular shaped vesicles could be positively 26 identified as being plant debris due to some chloroplast organelles (that had thylakoids still partially intact). However, qualitatively, it was immediately apparent that the <63µm fraction was dominated by microbes. Cell walls and cellular debris from plants that might otherwise be expected to derive from the cow’s diet were much less obvious or abundant in this fraction. This could be because of the inherent differences in cell sizes between bacteria and plants, where plants having diameters typically much greater than 10µm14 were effectively excluded by the 63µm sieve. The figure below shows a picture of the diversity of microbes and debris present at 1000X magnification. The small squares are 50µm in length and width and the group of 16 squares represent an area of 0.04mm2. Quite a diversity of microbes were visualized moving, but much of the other particulates lacked a simple means for positive identification. Diverse Range of Bacteria Plant Debris Figure 12: Abundance of bacteria observed in <63µm manure matrix at 1000x magnification 27 3.3 Effects of Microbial Activity on the Distribution of Phosphorus in the Liquid Fraction Once an abundance of microbes was identified in the <63µm manure fraction, an attempt to quantify how much of the phosphorus in this fraction was specifically associated with microbes was undertaken. By adapting a method for measuring chemotaxis in microbes to an attractant, it was thought that the directional movement of microbes could change the concentration of phosphorus near an attractant by the proportion of phosphorus present in their membranes and cytoplasm. The capillary tube experiment was scaled to a volume sufficient for simultaneously quantifying phosphorus by ICP-OES, but still small enough so that the linear dimension of the tube and surface tension would limit diffusion and mixing during solution transfers. In this modified chemotaxis experiment, 1mL of diluted <63µm manure fraction was loaded into a narrow serological pipet and drawn back from the orifice tip to allow for 100µL of a 10-3M glucose solution to be injected at the liquid interface (without considerable disturbance to the manure solution along the length of the pipet). The pipet was then laid horizontally for a period of approximately 1hr to allow a gradient of glucose to diffuse into the <63µm solution and allow for microbes to move into the glucose attractant. This experiment included ten replicates each time it was performed. After the duration allowing for chemotaxis to occur, each pipet volume was partitioned into two parts. The first half closest to the attractant was either isolated with the full volume of the attractant included or 100µL of the terminal part of the solution was drawn off before the remaining 1mL volume could be split into two equal parts. The second half of each tube was used as a paired control to normalize the relative change observed in each pipet. The intent was to see if the 0.5mL of solution that had an adjacent 100µL of glucose attractant removed, ended up with statistically lower phosphorus 28 concentrations and bacterial counts compared to its 0.5mL control. Presumably, the bacteria that moved into this attractant (with their associated membrane and cytosolic phosphorus) would be lost and the difference in phosphorus could be related back to the difference in HPC compared to their controls. This observation would be further substantiated if the replicates involving a retained glucose attractant were not statistically different form their control halves. Figure 13 shows one attempt at this experiment with five replicates of each approach and two untreated quality controls. Figure 13: Chemotaxis experiment for the determination of phosphorus in motile bacteria After partitioning each pipet volume in half, a portion of each solution was then diluted in buffer for plating on plate count agar (PCA) and another portion was acid digested for assessment of total phosphorus. The experiment had to be completed a few times to find the 29 appropriate dilutions for plates to be accurately counted. Figure 14 below shows the plate counts for one of the attempts at this experiment and Table 3 summarizes the statistical results obtained. Figure 14: HPC of organisms in pair-controlled chemotaxis replicates Table 3: Results of paired chemotaxis study reflecting mean differences in HPC and Total P Sample Attractant Mean HPC (Log10 cfu) tcalc Mean Total P (mg/L) 1st half discarded 7.45 18.8 (closest attractant) 0.46 2nd half n/a 7.37 21.8 (used as control) 1st half retained 7.56 17.9 (closest attractant) 0.59 2nd half added after 7.51 18.3 (used as control) partitioning tcrit = 2.776 with n=5 for each paired t-test using a 100-fold dilution of <63µm dairy manure Results Not Significant, Results Significant tcalc 10.65 2.88 30 Unfortunately, a paired t-test of five replicates showed no statistical difference in the means between heterotrophic plate counts at the 95% confidence interval with and without the attractants. The phosphorus concentration was statistically lower in the fraction closest to the discarded attractant, but the replicate with the attractant retained was also found to be statistically lower (albeit by a much smaller margin). Thus, a relationship between phosphorus concentration and a statistically different bacteria concentration failed to be established. There appears to be too much variance in the datasets to perceive a statistical change in bacteria by this approach nor substantiate the change in phosphorus observed. After a few attempts this experiment was abandoned. Further refinements to the experimental design might be laudable, but an alternate approach involving less dilutions and less bacteria may be necessary for reasons of experimental precision. 31 3.3 Rates of Sedimentation in <63µm Liquid Fraction and its Relation to Microbial Activity Since it was substantiated that phosphorus is predominately associated with solid particulates in the liquid fraction, sedimentation rates of suspended solids may provide another pragmatic means of managing phosphorus in liquid manure. If suspended solids in the liquid fraction will stratify over time as a result of gravity, a portion of an undisturbed water column may passively diminish in phosphorus near its surface with a corresponding increase in phosphorus at greater depths. This passive thickening mode has been demonstrated for stratifying dairy lagoon sludge for the purpose of reducing trucking volumes of water15 and was attempted here to observe its applicability to liquid separated manure products. A simple sedimentation experiment was performed to see how useful this phenomenon might be using the <63µm liquid manure fraction. This experiment was conducted by filling a 50mL burette with a portion of homogenized <63µm liquid manure and allowing it to settle undisturbed for a considerable period of time. Visually, the solution did not appear to stratify. Eight days were allowed to pass before samples were slowly drained off from the bottom of the column in a few prescribed intervals (as measured from the bottom to the top). Very little stratification in terms of phosphorus or suspended solids was observed. At a later date, after the observation that microbes constitute a considerable portion of the liquid manure, this experiment was repeated using an autoclaved burette full of homogenized <63µm liquid manure. Figures 15 and 16 show that the autoclaved liquid stratifies in relation to both its total solids and its total phosphorus composition. In contrast, the untreated liquid manure did not appear to stratify over the study duration. 32 <63um Sieve Fraction Autoclaved <63um Sieve Fraction 6 Total Solids % (w/vol) 5 4 3 2 1 0 Homogized 0 109 225 325 436 505 Height from Bottom of Column (mm) Figure 15: Effect of microbial activity on stratification of TS in <63µm liquid over 8 days <63um Sieve Fraction Autoclaved <63um Sieve Fraction Concentration of Total P (g/kg) 30 25 20 15 10 5 0 Homogized 0 109 225 325 436 505 Height from Bottom of Column (mm) Figure 16: Effect of microbial activity on stratification of Total P in <63µm liquid over 8 days 33 At the early stage of experimentation, prior to the knowledge of microbial activity in this fraction, it was thought that the solids may be interfering with sedimentation rates by physically obstructing each other in the column. A follow-up experiment was performed to better profile sedimentation rates at a 1:100 dilution where particle-particle interactions inhibiting sedimentation could be avoided and with a means for collecting samples at differing depths and at different times. Figure 17 shows the profile of phosphorus with respect to time and depth on a diluted liquid manure sample over a four-day period. The sampling regime included two triplicates as well as a depth profile conducted at 24hrs. Error bars signify one standard deviation from the mean calculated between three replicates. The small decline in concentration at the surface is minimal compared to the variance associated with the replicates. The depth profile (shown as a colour gradient) also did not show a trend with depth despite some visual differences in colour observed in the actual sample container near the bottom. 3.7 n=3 3.2 n=3 2.2 1.7 1.2 0.7 -Depth series at 24hrs 2.7 -Independantly prepared settling trials Total P of 1:100 dilution of <63um sieve fraction (ppm) 4.2 Time Series at Surface (1-10mm) Depth 1mm Depth10mm Depth 20mm Depth 40mm Depth 80mm Depth 140mm (Bottom) 0.2 -0.3 0 5 10 15 20 25 30 35 40 45 50 Settling Time (hrs) Figure 17: Temporal and depth profile of phosphorus in undisturbed diluted liquid manure 34 From these two experiments, we can reasonably assert that sedimentation of particulates will not passively stratify liquid manure fractions in terms of either solids or phosphorus in any short duration. However, some aspect of autoclaving did appear to facilitate stratification. Although, sterilizing large volumes of liquid manure to stratify phosphorus or solids would be untenable, the phenomenon was further investigated to determine if the stratification was the result of autoclaving or the loss of microbial activity more specifically. Three test tubes of liquid manure were prepared for treatment. Two of the test tubes were autoclaved to kill all microbes present and the third tube was left untreated with its microbial community intact. One of the autoclaved tubes was allowed to cool and then immediately re-inoculated with microbes from a small portion of the untreated test tube. The three tubes were then allowed to settle undisturbed for a similar duration as the burette experiment. Figure 18 shows how the three test tubes differed after 9 days. Note the reduced turbidity at the top of the sterile test tube (blue cap) compared to the autoclaved tube (red cap) that was re-inoculated with 100µL of the untreated solution (green cap) immediately after cooling. It appears that stratification of suspended material is only occurring in the sterile tube. Biofilms were also forming at the surface of the two non-sterile solutions. At this time, samples at the surface, and at approximately 15mm and 40mm depths were taken from each test tube to profile the phosphorus concentration as a result of treatment. After the samples were taken, a 100uL sample of the untreated liquid manure was then re-inoculated into the sterile test tube. Care was taken to not unduly disturb the sterile solution during sampling or re-inoculation. 35 Day 11 Day 9 transparent layer expanding in sterile tube became turbid like others after re-inoculating Figure 18: Effects of microbial activity on appearance of stratification and turbidity If a significant portion of phosphorus in the liquid is held within motile microbes, then sterilizing the solution should allow dead bacteria to settle out of the column according to density and stratify the phosphorus concentration by the amount contained within them. Phosphorus associated with other solids may also be expected to settle according to their size and density, but if bacteria are re-introduced into the undisturbed solution only the phosphorus associated with motile microbes should be able to rise against gravity. Two days following re-inoculation the sample (blue cap) became turbid like its counterpart (red cap) that was re-inoculated immediately after autoclaving. At day 11, each of the samples was re-assessed for total phosphorus at the surface and at a depth of 15mm and 40mm. Figure 19 shows that only the sterile sample stratified in terms of phosphorus. Note that there is a difference between the red and green capped tube concentration as a result of the autoclaving process but that their profiles each appear unstratified. 36 20 18 16 Total P (ppm) 14 12 10 8 6 No Treatment 4 Inoculated Immediately After Autoclaving Sterile 2 2 Days After Inoculating 0 Surface 15mm 40mm Sampling Depth Figure 19: Effects of microbial activity on the stratification of total phosphorus The phosphorus in an undisturbed and sterile solution can then be shown to increase against its depth profile following inoculation. From this experiment it appears that motile bacteria are, in some part, responsible for preventing liquid fractions of manure from stratifying in phosphorus concentration over time. One other observation worth noting is that the time needed for dead microbes to settle out of solution and passively reduce phosphorus concentrations was considerable. The turbidity associated with microbes near the surface appeared to descend in an observable layer only 30mm over the course of approximately 9 days. This implies that the inherent density of these organisms is very similar to the density of the solution they are in. 37 3.4 Effects of Leaching and Fermentation on the Availability of Phosphorus from Solids It is uncertain how much of the total phosphorus in a particular sieve fraction may be leachable or further available to biological organisms. Given that the data suggests microbes have a sizable effect on the phosphorus in the liquid fraction, their metabolic activity or adherence to larger particles may have implications on manure management of the coarser fractions too. Two processes were explored to see how leaching and biological activity might affect residual phosphorus in the coarsest sieve fraction. The “Washed” >2mm fraction was subjected to fermentation and leaching in a sealed plastic baggie held at room temperature for 1 month. The sample was then split in two and either dried the way it was or washed thoroughly again with deionized water before being oven dried. After this room temperature incubation, the rewashed and dried sample was split into six different leaching and fermentation treatments. 0.5g samples were weighed into plastic digestion tubes with 30mL of deionized water. Some of the samples were inoculated with 20µL of a 1/10 dilution of <63µm liquid manure before or after autoclaving and some were further supplemented with a small amount of urea to facilitate fermentation if nitrogen was lacking in the “Washed” sieve product. Urea was chosen as the source of nitrogen as this may reasonably represent what may happen if the “Washed” >2mm fraction was to be used as green bedding (where it may have broader applicability). Once autoclaved or supplemented as desired, the solutions were shaken thoroughly and left in an incubator at 35oC for a week to either ferment and/or leach further. Following incubation, the supernatant and the remaining washed solids were assessed for total phosphorus. Figure 20 shows the samples following incubation and summarizes the treatment of each. Note that the autoclaved samples each lost approximately 5mL in volume as a result of the process and 38 darkened somewhat during their thermal treatment. Further, samples S5 and S6 differ only in the order of operations (in which sample S5 remained sterile). The two sterile samples were visually less turbid than the others. This again, suggests that motile microbes diffuse throughout the water column. not washed washed and autoclaved washed and inoculated with <63µm triplicate of washed solids samples inoculated with <63µm manure and supplemented with urea sterile washed, inoculated, supplemented and then autoclaved washed, , supplemented autoclaved, and then inoculated with <63µm sterile Figure 20: Leaching trials of phosphorus in >2mm “Washed” sample Table 4 summarizes the amount of phosphorus in the supernatant and the residual solids with respect to the “Washed” >2mm fraction collected immediately after the sieving experiment. Sample 1 included the solution leached from the room temperature incubation, while all the others used the >2mm fraction after re-washing. That is why the sums from S2-S6 do not approximate 100%. Table 4: Summary of phosphorus leached from samples following incubations Sample ID S1 S2 S3 S4 (SD, %RSD) S5 S6 Phosphorus remaining in solids (%) 24.6 16.3 27.6 21.5 (3.3, 15.3) 23.7 22.5 Phosphorus leached into solution (%) 76.5 54.3 59.0 59.3 (2.4, 4.1) 66.6 58.9 39 From the data we can see that, indeed, much of the phosphorus present in the coarse solids is still subject to leaching and/or fermentation processes. Differences due to autoclaving, inoculating, or supplementing with an exogenous source of nitrogen are less obvious. Based on the standard deviation observed for the replicate, these treatment differences are not different enough to clearly discern an effect. However, the amount of phosphorus that transferred into solution during incubation was considerable for all treatments. The results indicate how residual phosphorus may be mobilized with time and wet storage conditions, which may have major implications for how manure products are stored for phosphorus management. 40 3.5 Managing Phosphorus in Liquid Manure from a Microbiological Perspective Great efforts are put forth handling liquid manure products for the purpose of managing phosphorus and other nutrients. From this study, it is clear that phosphorus can be shunted from coarser solid fractions into liquid manure products at the expense of increased mechanical processing and/or added water. However, this may just transfer a problem of high phosphorus in solids to a greater volume of liquid manure requiring more resources to manage later. Farmers likely choose between situations that best suit their farms. Arguably, one of the more desired outcomes for manure management would be to have a clean fibrous coarse fraction (for green bedding and easy of storage), a relatively clean aqueous component (for washing, irrigation, direct discharge or even potable water purposes), and a highly enriched and dewatered nutrient cake that can be used as a fertilizer or soil amendment (with greatly reduced water content to facilitate storage and reduce trucking costs). In this regard, removing solids and phosphorus from liquid manure products is quite desirable. However, the costs and equipment involved are prohibitive. Three technologies commonly employed at scale to address this task in liquid manure are 1) centrifugation, 2) flocculation, and 3) microfiltration. They have benefits and costs that need to be weighed and there is extensive literature discussing the anticipated outcomes of these different technologies9,16–19. Some follow-up experiments on the <63µm liquid manure product were conducted to address solids removal using some of the insights gained here. Chiefly, if the distribution of phosphorus in liquid fractions are indeed dominated by microbes, how might that suggest improvements to techniques for centrifugation, flocculation and microfiltration typically employed? Further, the follow-up experiments trialed here broadened the scope of test parameters described more thoroughly elsewhere. They were not exhaustively studied but 41 included 1) the effects of centrifugation at much higher centrifugal forces, 2) different means of affecting sedimentation rates, and 3) how the addition of additives might facilitate both filtration and flocculation processes. 3.5.1 Centrifugation With the observation that sterilized solutions stratify more easily and that the motility of microorganisms hinder stratification, a few centrifugation comparison trials were conducted to see how aspects of microbial activity might hinder or promote stratification. The trials included elevated centrifugation speeds (more typical of laboratory techniques for isolating suspended cells from cultures) as well as extreme centrifugation speeds (found to impair microbes). The trials also included sterilized liquid manure to see how microbial activity might be influencing sedimentation rates as well as re-inoculated samples to see how the solutions might change with time after treatment due to microbial activity. 250 25 Total P in <63um Total P in Re-Inoculated <63um Total P in Supernatant (mg/L) 200 20 TS in <63um TS in Autoclaved <63um 150 TS in Re-Inoculated <63um 15 100 10 50 5 0 0 5000 10000 15000 20000 25000 30000 35000 Total Solids in Supernatant (g/L) Total P in Autoclaved <63um 0 40000 Centrifugal Force per 10 Minute Application (g) Figure 21: Effects of centrifugation on solids and phosphorus in supernatant 42 Figure 21 shows the residual solids and total phosphorus content in the supernatant following ten-minute treatments at different centrifugation speeds. Centrifugation forces beyond what are considered feasible in an agricultural setting are seldom considered. However, with the observation that microbes are quite prevalent in the liquid fraction, this merited the use of high speeds here to determine if further improvements are possible. This was done on sterilized <63µm solution too. Comparatively, these elevated speeds did not improve reductions already realized at lower speeds with or without sterilization. This suggests that trying to improve sedimentation rates with higher centrifugation forces would not provide much benefit in terms of phosphorus or solids management on farms. The re-inoculated sample that was treated to extreme centrifugal forces did not differ much after 24hrs once an inoculum of microbes was reintroduced. This suggest that high centrifugation speeds might pack solids in a way that microbes are less able to re-spread throughout the water column. However, that speculation has limited data supporting it and was only carried forward for 24hrs. The supernatant was also investigated for its bacteriological count on PCA media following treatment. Figure 22 shows how the abundance of microbes changed throughout the exposure to escalating levels of centrifugation. The sterile samples proved sterile, while the HPC for the intact sample was observed to diminish by two orders of magnitude (log transformed counts of colony forming units per millilitre of sample). Error bars represent one standard deviation about the mean for the replicate that was performed at 2000x g. 43 8.0 n=3 Heterotrophic Plate Count (Log cfu)/mL 7.0 6.0 5.7 5.0 5.0 4.0 3.0 2.0 <63um Sieve Fraction <63um Sieve Fraction 1 Day After Centrifuging 1.0 0.0 0 5000 10000 15000 20000 25000 30000 35000 40000 Centrifugal Force per 10 Minute Application (g) Figure 22: Effects of centrifugation on HPC in supernatant It is interesting to note that the microbial counts did not diminish to zero (or near zero) at extreme centrifugal forces routinely employed to pellet cells from pure cultures. This was unexpected. As a result of this observation, the supernatants were investigated under a microscope. Figure 23 shows two microscope images of the microbes observed after extreme levels of centrifugation at 1000x magnification. The observed microbes were predominately round, larger and had obvious green coloured organelles present (suggesting chloroplasts). These may be algae, cyanobacteria or other photosynthetic microbes. This implies that centrifugation enriched the distribution of photosynthetic organisms in the supernatant, presumably, due to greater buoyancy as a result of gaseous oxygen generated in their cytosol which has been observed with similar types of photosynthetic microbes20,21. Instead of pelleting 44 at the bottom with other microbes, their reduced density caused them to rise. This may explain the abundance of microbes detected on the PCA plates even after extreme forces of centrifugation. The arrows point to examples of the larger green microbes found in the supernatant after centrifuging. Figure 23: Observations of photosynthetic organisms in high-speed centrifuge supernatants From these observations, centrifugation with forces beyond what is commonly employed on farms currently is unlikely to improve phosphorus or solids management further, but may influence the types and distribution of microbes present. 3.5.2 Substrate Mediated Flocculation and Sedimentation If microbes dominate the liquid component of manure, and if the reduction in phosphorus observed in the coarser fractions was simply due to the removal of microbes adhering to larger substrates, then it may be possible to promote removal of microbes (and their internalized phosphorus) by introducing additives having properties that promote sedimentation. These may 45 include flocculants that congeal suspended particles together or particles with higher density that, if interacting with microbes, might encourage settling. A jar mixing experiment using various additives with different concentrations of the <63µm liquid manure was devised to compare against untreated jars to see if sedimentation could be encouraged and if that would lead to stratification of total solids or total phosphorus. Figure 24 shows a dilution series of different treatments and control solutions included. Figure 24: Substrate mediated flocculation and filtration using different additives The additives included alum (a very effective flocculent used to treat wastewaters by removing solids and reducing phosphorus4,20,21), Moringa oleifera (moringa) powder (shown to reduce pathogens and clarify waters in some studies 22), clay (to act as a substrate for bacteria to adhere to and with sufficient density to settle out of solution) and diatomaceous earth, (which should have pores and structures optimal for associating with microbes and also dense enough to settle23). The <63µm liquid manure was diluted by 1/1000, 5/1000, 10/1000 and 100/1000 in 1L jars and mixed with their prescribed additives for a few minutes before allowing the solutions to settle for one day. The supernatants were then sampled at their surface to determine if their phosphorus declined relative to an untreated control with the same dilution factor. The 46 supernatants for the clay and the diatomaceous earth were simultaneously assessed for their heterotrophic plate counts on PCA. The results of this comparison study are summarized in Figures 25 and 26. Total P Reduction After Treatment (%) 90 40 -10 1/1000 5/1000 10/1000 100/1000 Moringa Root Powder 2g Alum 0.2g -60 Clay 2g Diatomaceous Earth 2g -110 -160 Dilution of <63um Being Treated With Additives Figure 25: Effects of additives on phosphorus in diluted liquid manure after 24hrs As confirmed from the reagent blank, the moringa had a background level of phosphorus that contributed noticeably to the supernatant at low dilution levels and did not appear to reduce levels at higher concentrations either. Its turbidity also did not seem to improve as might have been expected from other studies22. In contrast, each of the other additives reduced phosphorus in the supernatant relative to the control. Alum proved to show the greatest reduction despite being added at 1/10th the mass ratio as the other additives. 47 10 No Additive 9 Clay 2g 8 Diatomaceous Earth 2g 7 Log(HPC/mL) 6 5 4 3 2 1 0 1/1000 5/1000 10/1000 100/1000 Dilution of <63um Being Treated With Additives Figure 26: Effects of sediments on HPC in liquid manure supernatant after 24hrs From Figure 26 we can see that the treatment with clay appeared to have reduced bacteria in the supernatant compared to the control at each dilution tested, while diatomaceous earth did not. However, there were no replicates included in this trial or other means for determining if the perceived differences were statistically relevant. Further, moringa and alum treatments were not tested for HPC to compare against. These preliminary results suggest that adding exogenous substrates to manure wastewaters may reduce phosphorus and bacteria through interaction with denser particulates that promote sedimentation, with the latter being less discernable. However, the effect on total phosphorus in the supernatant was not as profound as the amount removed using alum. The question of whether a flocculent, like alum, could be combined with a denser substrate to improve sedimentation rates was proposed. To do this, the 10/1000 dilution of 48 <63µm was compared to the treatments of clay and diatomaceous earth with an equal amount of alum added. The samples were re-shaken after addition and allowed to settle. Nearly 100% of the phosphorus was removed in each of these trials. Figure 27 shows the change in phosphorus observed relative to each other. Although there was little room for improving the net removal, the substrate mediated flocculation settled much faster and more densely than alum used on its own. Further, when being decanted the sediments at the bottom of the jar partitioned more easily from the supernatant when clay and diatomaceous substrates were present. This may have broader utility for high throughput processing, where sedimentation rates using flocculants are a limiting factor or where the disturbance of the settling layer needs to be reduced. <63um with Alum <63um with Clay and Alum <63um with Diatomaceous and Alum Total P Reduction After Treatment (%) 100 95 90 85 80 75 70 65 60 55 50 10/1000 Dilution of <63um Being Treated Figure 27: Phosphorus reduction by substrate mediated flocculation 3.5.3 Substrate Mediated Filtration Some of the literature describing isolation and purification of bacterial cultures suggested the use of diatomaceous earth to facilitate filtration of large volumes rather than using batch 49 centrifugation, which can be onerous23. Their results were somewhat encouraging. For this reason, diatomaceous earth was incorporated into the flocculation and sedimentation study, where it was later evaluated here for how well its supernatant could be filtered compared to other additives and compared to untreated liquid manure controls. This comparative filtration experiment was conducted on all the additives trialed for promoting flocculation and sedimentation. Briefly, once the treated solutions were allowed to stratify for a day and tested for their total residual phosphorus concentration, each solution was then gently decanted through a standard TSS filter until all of the solution was filtered or until the filtrate slowed to an intermittent drip rate under vacuum. The figure below indicates how much volume of supernatant passed through the filters with ease. Relative to the control, the increased volume signifies some degree of improvement in filterability. The volumes differed somewhat between jars and were subsequently normalized to 1L for direct comparison. Only alum improved filtration at all manure dilutions tested. Clay was also shown to greatly improve the ease of filtration at the lowest manure concentration. The others all compared worse than the diluted liquid manure on it own. However, when the particulate substrates were combined with alum, the filterability improved even more. Figure 29 shows the improvements in filterability due to the substrate mediated flocculation normalized to 1L. Moringa was not trialed. 50 1.000 Achievable Filtration Volume (normalized to 1L) No Flocculant 0.900 Moringa Root Powder 2g 0.800 Alum 0.2g Clay 2g 0.700 Diatomaceous Earth 2g 0.600 0.500 0.400 0.300 0.200 0.100 0.000 1/1000 5/1000 10/1000 100/1000 Dilution of <63um Being Filtered Figure 28: Relative ease of filtration after treatment with individual additives Reshaken <63um <63um with Alum Achievable Filtration Volume (normalized to 1L) <63um with Clay and Alum 1.000 0.900 <63um with Diatomaceous and Alum 0.800 0.700 0.600 0.500 0.400 0.300 0.200 0.100 0.000 10/1000 Dilution of <63um Being Filtered Figure 29: Relative ease of filtration with substrate mediated flocculation 51 After this was observed the differing flocculation processes were compared under a microscope. It appears that the clay added to the flocs coalescing them into denser units compared to the flocculent particles not mediated with an added substrate. This likely caused the improved rates of sedimentation observed. flocculating particles with diffuse borders (without added substrate) denser flocculating particles mediated by clay substrate Figure 30: Microscopic comparison of flocs with and without substrate additive 52 Discussion The results of this series of experiments were derived using a single source of manure. Much of the literature discussing phosphorus concentrations in manure imply stark differences relating to species, breed, and conditions of the livestock, as well as feed, bedding, and how the manure is generated, handled, and stored3,13,16,21,24 . It is uncertain how broadly interpretations made in this study should be applied elsewhere. Some caution is warranted. For instance, differing sources of feed or bedding material may impart different attributes to the efficiency of classification or the size distribution of particles present in the manure. This manure had wood shavings introduced from the bedding. Presumably, alternate bedding could change the relative distribution of particles ending up in the reception pit and cause a different relative profile of phosphorus in the solids. However, the manure used appeared typical of dairy farms which combine the liquidy portions of manure, bedding and washings into open reception pits. The characteristics of the manure were not outside expectations based on the literature and appearance of the manure observed. From one reference, expectations for dairy liquid manure are 15 pounds of P2O5 per 1000 gallons1, which corresponds to approximately 784ppm phosphorus. The replicate grab samples used in this experiment averaged to 761ppm total phosphorus. It is likely that the broader trends and interpretations derived in this study would generalize to dairy farms with similar infrastructure and practices. From the sieve analyses, most of the oven dry solids ended up in the largest sieve used. However, the choice of the largest sieve was somewhat arbitrary. In keeping with the pattern used, the profile could have easily included 4mm, 8mm, 16mm sieves before the abundance of solids diminished precipitously. The 2mm sieve was used as the cut off because materials larger than this may end up handled as bulk materials piled by tractors in compost piles, whereas the 53 primary focus of this study was the materials that end up in reception pits and lagoons. The stack size was also limited by the number of sieves an individual could easily manage. From the trends observed, there are no indications that alternative choices of sieves would have led to different compositional outcomes. However, results may not extrapolate well beyond the largest and smallest sieve used. Further, it is important to note that the sieve diameters were only a proxy for cross-sectional size discrimination. The sieve holes were square, and the shapes of the manure particles were irregular and aggregated in ways that may not be conducive for efficient transfer though holes of minimal diameter. Indeed, we saw from Figure 5 that supplementation with a small portion of process water greatly reduced the mass of material in the 2mm sieve and led to a cascading increase of material in all subsequent fractions in the “Rinsed” treatment. The extent of pressure and nature of the mechanical agitation applied may also influence how efficiently particulates are size discriminated. Alternate approaches may lead to alternate outcomes. However, “Sieved”, “Rinsed” and “Washed” treatments, which varied greatly in the efficiency of particle size separation, each lead to well defined parabolic functions relating phosphorus to the screen size used. We may subsequently expect similar profiles for manure particles generated by other mechanical means, with their functions offset by some factor specific to the dimensions of size discrimination used by those processes or equipment. Differences in the types of size exclusion mechanisms were not studied here, but it is assumed that the cross-sectional diameters of the particles collected fits approximately within the range of the sieve sizes used. For simplicity, all results in this study used this assumption. Most of the “Sieved” manure solids ended up in the >2mm fraction. This fraction harboured 66% of the dry matter content and nearly 50% of the total phosphorus. However, after progressive treatments this reduced to 32% of the dry matter but less than 5% of the total 54 phosphorus. In the “Rinsed” treatment, the efficiency of size discrimination likely improved as mass from the >2mm fraction led to a corresponding increase in all subsequent fractions. In contrast, the phosphorus concentration declined by approximately half in all of the isolated solid fractions. This implied that the movement of solids based on increased separation efficiency does not have the same proportional change in phosphorus. In other words, the movement of matter was not intrinsically linked to similar proportions of phosphorus. Most of the phosphorus containing substrate must have had particle dimensions less than the smallest sieve size or, expectedly, the total phosphorus in at least one of the sieves would have increased. This was not observed. Because the amount of rinse water was arbitrary imposed and differed in proportion to the mass of manure between replicates, this “Rinsed” <63µm fraction was not retained for direct comparison. It is assumed, however, that the reduction in phosphorus observed must correspond to a mass transfer of phosphorus containing particles <63um for the phosphorus to leave the sieve stack. In Figures 5 and 6 we see that thorough washing further reduced solids and total phosphorus. These washings were not pushed progressively through the sieve stacks as in the “Rinsed” treatment. Instead, they represent an improvement to the size discrimination of particles already collected in each sieve and a clearer division between particles larger than the hole diameter and superficially bound particles. The residual phosphorus in these coarse fractions represents less than 20% of the initial phosphorus present. This would either be phosphorus infused or trapped within the substrate or as part of its material and chemical composition. Based on the summation of phosphorus in all sieved fractions, 80.9% of all the phosphorus isolated in the “Sieved” treatment was able to be removed by simple washing and mechanical agitation. This is very similar to observations reported by Wu and Zhong (2020) 55 that had size exclusion down to 0.5mm and saw 81.45% of total phosphorus removed as water soluble or as particulates smaller than <0.5mm in fresh dairy manure24. When the coarsest fraction was further subjected to a simple fermentation and leaching experiment, the residual phosphorus reduced by another 76.5%. Finer sieve fractions were not trialed this way, but they would have more surface area, and likely be subject to similar levels of leaching. This observation may have some pragmatic implications for iterative treatments of manure. If a screening process separated and stockpiled a coarse fraction of solids, natural leaching and fermentation might make its phosphorus content more available over time. Reprocessing that same material may improve the overall extraction efficiency of nutrients from the residual fiber and coarse substrates that persist. Collecting a clean coarse fraction of material as a value-added manure product is a desirable outcome in manure management, particularly if it is suited for on-farm re-use as green bedding. Since green bedding has both potential benefits and risks compared to alterantives25, added washing and nutrient removal may reduce those potential risks. However, increased use of water and the added costs of mechanical processing may offset any potential gains. A corresponding improvement to the value of the other manure products would also be needed. This would likely be centered around dewatering to offset any additional water applied as well as the costs of handling larger volumes (i.e. increased trucking, lagoon volumes and pump costs). For this reason, the <63µm fraction was subjected to follow-up experimentation and further characterizations of treatments described in section 3.5 to see what insights might be gained. The actual species of phosphorus present are typically only alluded to based on how they are ascertained and the extraction procedures used 3,6. They are primarily present as orthophosphates, condensed phosphates (pyro-, meta and polyphosphates) and organic phosphate 56 (found in molecules such as nucleic acids, phospholipids, lipopolysaccharides) and in various cytoplasmic solutes6,26. Phosphorus is also dogmatically divided into dissolved and suspended phosphorus by 0.45µm pore diameter filtration, while total represents the summation of all phosphorus in a given sample. Each is further divisible into three chemical types that can be described as reactive, acid-hydrolysable or organic based on the mode or lack of digestion involved in their analyses6. Other extraction schemes are also used for classification and have implication in manure management3. The actual species of phosphorus present, however impactful, are metabolized and interconverted by microbial activity26. For simplicity, the distribution of phosphorus studied here was characterized only in terms of total phosphorus as this relates easiest to concentrations in solids needing direct comparison. However, the “dissolved” portion of the <63µm fraction was assessed in the <63µm fraction. The recovery of dissolved phosphorus was less than 1/50th of the total in the <63µm fraction. This constrains the observations in the sieve analyses and implies two important things. First, of the total phosphorus that was removed from larger substrate and shunted into the <63µm liquid fraction, very little of it was actually dissolved. This substantiates that the phosphorus present in the manure and liquid fractions exists predominately as suspended solids. The second implication of this observation is that although the concentration of phosphorus increases inversely proportional to particle size discrimination, that trend must not continue to <0.45µm sized particles. When the distribution of particles in the <63µm fraction was further profiled using a laser diffraction PSA, something interesting was revealed. Almost no particles were less than 0.2µm in size. Under microscope inspection at 1000x magnification, it appeared that almost all of the particles present were either actively motile microbes or unidentified objects with appearances of microbes. Very little plant debris or other particulates were observed. It is uncertain if this 57 distribution was induced by some storage conditions, an inability to visualize other types of particulates, or a true representation of the liquid fraction of manure more generally; but it implies that much of the phosphorus containing material in the liquid fraction is associated with microbes. Likewise, much of the plant debris expected in the excretion must be >63µm. Since much of the phosphorus from the coarser solids was also shown to pass into the <63µm fraction, it may be that most of the phosphorus present in manure is actually related to the distribution of microbes which have intrinsic dimensions in the micron range. Further, when the liquid manure was filtered with a 0.2µm defined pore size membrane, this filtrate had even less phosphorus than what is considered “dissolved” by a 0.45µm filter convention. This filtrate was found to be 24.5mg/L. Since PSA suggests that there is low abundance of particles <0.2µm, 24.5mg/L is likely a better representation of the true amount of dissolved phosphorus present in the manure. A provisional attempt to directly quantify the concentration of phosphorus in motile bacteria was attempted using an adapted chemotaxis assay 27,28. However, the imprecision of the bacterial quantification appeared to be too high to reveal statistical differences against paired controls. The titre concentration in the working manure solution was likely too high to see a relative difference between treatments after incurring the serial dilution needed to bring the plate counts within a quantifiable range. The methodology had some flaws due to trying to assay both total phosphorus and microbial concentrations simultaneously. If the minimum volume for determining phosphorus concentration can be reduced substantially, it might be prudent to conduct the chemotaxis trial from the perspective of counting the bacteria moving into an attractant solution, rather than trying to count the bacteria moving out of the manure liquid (as was attempted here). The relative difference might be made more pronounced. 58 Quite a bit of literature has addressed centrifugation as a management tool for treating liquid manure. They typically focus on optimizing solids removal. Less attention has been focused on the microbes present. A recent study looked at concurrent effects of pathogenindicator reduction during centrifugation, with and without a polymer coagulant. Coliforms and Escherichia coli were observed to have a two-log10 reduction in concentration at centrifugal forces slightly higher than typically employed on farms and did not improve appreciably with the supplementation of a polymer flocculant17. Since this study was focused on microbes more generally and their relation to phosphorus, a similar centrifugation study was carried out here but with a broader account of all heterotopic bacteria present. The results were a similar two-log10 reduction in bacteria found over a similar treatment range. However, this is a bit dissimilar to results of a different high-speed centrifugation study that showed little difference in the pellet number of bacteria for pure cultures of Staphylococcus aureus ATCC12600 centrifuged between 6480-14800 x g29. The discrepancy between studies may suggest that certain types of microorganisms have a threshold acceleration needed to remove them from suspension in a timely manner, while a diverse set of microorganisms may be removed more incrementally over a broader range of centrifugal forces applied. The latter study also implied, based on altered cell adhesion and zeta potential, that elevated centrifugation speeds can cause cell surface damage and that the damage was related to shear forces against cell walls29. How shear forces experienced during centrifugation might affect a mixed culture of organisms in a complicated liquid manure matrix was unknown. Subsequently, a high-speed centrifuge was trialed to examine the effects of increased centrifugal force up to 38000x g using the <63µm liquid manure. 59 The purpose of this high-speed centrifugation study was to see what extreme forces and cell damage might impart on a complex solution of microbes and if the heterotrophic plate counts could be reduced further (irrespective of the practicality of these excessive speeds). By 15000x g no more reduction in solids was apparent. By 25000x g the HPC was reduced to its lowest, with a near three-log10 reduction. Further centrifugal forces up to 38000x g did not appear to reduce total phosphorus, total solids or HPC any further. The procedure was repeated with an autoclaved version of the <63µm manure. Its reduction profiles for total solids and total phosphorus were similar. Further, re-inoculating the sterilized solution after a 38000x g treatment did not appear to increase TS or TP substantially in the supernatant after 24hrs and the pelleted solids seemed to persist with similar consistency. This may imply that high-speed centrifugation and the associated cell damage could reduce the rate at which microbes can resuspend solids and total phosphorus into the water column. Unfortunately, however, the sterilized 38000x g treatment had to be homogenized afterwards to derive its total phosphorus concentration. This was done because the solution lost volume in the autoclave while cooling and quantifying (via sub-sampling) before centrifugation would have risked contamination. It might have been valuable to conduct a time series of measurements after re-inoculating to see how HPC, total solids and phosphorus in the supernatant might have changed due to microbial activity after re-inoculation and to see if the consistency of the pellet would change over a longer duration. From this centrifugation experiment, further improvements to total solids, phosphorus or reduced microbial concentrations are not feasible with extreme centrifugal forces. It was surprising that the HPC could not be reduced further. A visual examination of microbes in the supernatant suggested a possible reason. It appeared that the distribution of microbes changed. 60 The organisms present after centrifugation were larger, mostly round and had green interiors. The latter observation suggests the presence of chloroplast and the function of photosynthesis. It may be that gas vacuoles or air bubbles made during photosynthesis reduce buoyancy of these photosynthetic organisms (such as some blue-green alga30 or green macroalga31 that change buoyancy in this way) which prevents their removal from solution by centrifugal forces. If this were indicative of all the HPC organisms collected in the 38000 x g supernatant, this might infer an approximate 1000:1 ratio of settled microbes to buoyant photosynthetic organisms. That assertion is highly speculative, but something that could explain the results obtained here. Two other approaches used to address phosphorus removal in liquid manure are filtration and flocculation. Despite much study on the use of filtration18 and flocculations19 both approaches have cost and benefits needing to be weighed to meet the specific needs of each farm. Often, the technologies are not practical or financially viable4,9. Bacteria are notoriously difficult to filter in large volumes due to fouling23. Likewise, flocculation is complicated by the need for optimal flocs to form and the requirement of separating the consolidated particles without breaking up fragile flocs with physical forces19 . These complications can make the deployment of the technologies prohibitive. However, with the observation that much of the phosphorus (arguably microbial in nature) is superficially associated with larger particulate matter, the possibility of removing phosphorus by simply adding an exogenous source of particulates for smaller particulates (or microbes) to associate with was an interesting possibility for improving these techniques. Adding denser particulates may also improve the sedimentation rate of flocs having low density. For these reasons, exogenous sources of sediment were trialed with and without flocculent in an exploratory manner. Alum and moringa powder (the latter being a natural flocculant reported in the literature) were used as the flocculants, while 61 diatomaceous earth and clay were used as the exogenous sources of particulates. Moringa, despite expectations from literature22, performed quite poorly as a flocculent in terms of phosphorus removal on its own. So, only alum was evaluated in the substrate mediated flocculation trial. No attempts to optimize dosage were tried and the quality of the clay and the diatomaceous earth were not investigated either. Despite the superficial and cursory extent to which these variables were studied, some compelling results were obtained. When compared to different dilutions of untreated manure, both the diatomaceous earth and the clay showed a discernable reduction in total phosphorus after 24hrs of settling. In very dilute liquid manure, the clay also appeared to both improve the ease of filtration and lower the HPC. All the other treatments lead to lower filterability, despite some expectations for the prospects of the diatomaceous earth23. However, the phosphorous removal was much less than that realized for alum alone. The modest reduction of phosphorus in the supernatant with these simple treatments may still find utility on farms in some applications. Presumable the land application of these spent solids could easily fit into soil amendment practices given that both clay and diatomaceous earth are natural. However, the most promising observation from this experiment was found by combining the alum flocculent with the particulate substrates. This substrate mediated flocculation process did not reduce the supernatants phosphorus concentration much further, but the rate with which it was removed was dramatically improved. Whereas the flocculation of the alum alone took hours to near equilibrium. The substrate mediated flocculation caused sedimentation of the flocs in minutes. The interface of which also appeared much more stable. It would be tempting to imagine an optimized mixture that provides the sedimentation rate best suited for large scale treatments while giving properties to the floc that allow it to stabilize and 62 condense in a way that expels more water and resist physical disturbance. This is a phenomenon that likely has lots of practical utility and will be explored further in future works. 63 Conclusion Observations from this work support the conclusion that much of the phosphorus in manure is associated with small particulates superficially bound to larger substrates. In processes that partition manure based on size discrimination, the distribution of phosphorus can be shunted towards the smaller sieve fractions with improvements to separation efficiency of smaller particulates bound to larger particulates. Simple washing removed 95.5% of the phosphorus in the bulk manure as <2mm particulates. With simple fermentation and leaching, the phosphorus in the coarsest fraction was able to be reduced by another 76.5%. Presumably, all of this phosphorus would end up in a <63µm liquid manure fraction. In the liquid separated fraction, almost all of the phosphorus was associated with particles greater than 0.2µm and, based on microscope observations, was dominated by bacteria and other microbes rather than plant debris or other abiotic residues. The microbial activity in this fraction was shown to disrupt passive stratification of an undisturbed solution and maintain its homogeneity in terms of suspended solids and total phosphorus. In centrifugation trials, it was confirmed that much of the reduction in phosphorus and total solids content in supernatants occurs around 2000x g. High-speed centrifugation did not suggest possibilities for further improvement. Microscopic observations of supernatants showed a relative increase in large chloroplast containing microorganisms relative to bacteria, suggesting that high-speed centrifugation may simply stratify organisms based on their density rather than being able to completely remove them. Sedimentation rates and the ease of filtration of liquid manure subjected to flocculation with alum can be greatly improved with the application of exogenous substrates such as clay. The added substrate appears to increase the relative density of the flocs formed, which results in improved rates of sedimentation. 64 References (1) Lorimor, Jeff; Powers, Wendy; Sutton, Al. Manure Characteristics, Manure Management Systems Series-18, 2nd Edition, 2004. (2) Toth, J. D.; Dou, Z.; Ferguson, J. D.; Galligan, D. T.; Ramberg, C. F. Nitrogen‐ vs. Phosphorus‐based Dairy Manure Applications to Field Crops: Nitrate and Phosphorus Leaching and Soil Phosphorus Accumulation. J of Env Quality 2006, 35 (6), 2302–2312. https://doi.org/10.2134/jeq2005.0479. (3) Wang, Y.; Zhang, T.; Akinremi, O. O.; Bittman, S.; Brown, C.; Hao, X.; Hunt, D.; Li, S.; Tan, C. S.; Ziadi, N. Phosphorus Characteristics of Canada-Wide Animal Manures and Implications for Sustainable Manure Management with a Cleaner Environment. Science of The Total Environment 2022, 845, 157200. https://doi.org/10.1016/j.scitotenv.2022.157200. (4) Hallbar Consulting. Evaluation of Nutrient Recovery Technologies for Dairy Manure and Digestate, 2015. (5) Kumaragamage, D.; Akinremi, O. O. Manure Phosphorus: Mobility in Soils and Management Strategies to Minimize Losses. Curr Pollution Rep 2018, 4 (2), 162–174. https://doi.org/10.1007/s40726-018-0084-x. (6) Baird, Rodger; Eaton, Andrew; Rice, Eugene. Standard Methods for the Examination of Water and Wastewater; APHA, AWWA, APHA, 2017; pp 2540, 4500, 9000. (7) BC Ministry of Agriculture, Food and Fisheries. British Columbia Environmental Farm Plan Reference Guide, 2021. (8) Environmental Management Act; 2019; Vol. B.C. Reg. 8/2019. (9) Hullcar Situation Review, Nutrient Management Practices: Technical Report. Technical Report 2017. (10) British Columbia Ministry of Agriculture Innovation & Adaptation Services Branch. Summary of Nutrient Management Technology Options in the Context of Hullcar, 2017. (11) Edwards, J. E.; Huws, S. A.; Kim, E. J.; Lee, M. R. F.; Kingston-Smith, A. H.; Scollan, N. D. Advances in Microbial Ecosystem Concepts and Their Consequences for Ruminant Agriculture. Animal 2008, 2 (5), 653–660. https://doi.org/10.1017/S1751731108002164. (12) Dewhurst, R. J.; Davies, D. R.; Merry, R. J. Microbial Protein Supply from the Rumen. Animal Feed Science and Technology 2000, 85 (1–2), 1–21. https://doi.org/10.1016/S03778401(00)00139-5. (13) Møller, H. B.; Sommer, S. G.; Ahring, B. K. Separation Efficiency and Particle Size Distribution in Relation to Manure Type and Storage Conditions. Bioresource Technology 2002, 85 (2), 189–196. https://doi.org/10.1016/S0960-8524(02)00047-0. (14) Rezvani Moghaddam, P.; Wilman, D. Cell Wall Thickness and Cell Dimensions in Plant Parts of Eight Forage Species. J. Agric. Sci. 1998, 131 (1), 59–67. https://doi.org/10.1017/S0021859698005632. (15) Chastain, J. P.; Darby, J. A. A Thickening Process for Reducing the Cost of Utilizing Dairy Lagoon Sludge. Animal, Agricultural and Food Processing Wastes, 2000. (16) Font-Palma, C. Methods for the Treatment of Cattle Manure—A Review. C 2019, 5 (2), 27. https://doi.org/10.3390/c5020027. (17) Liu, Z.; Carroll, Z. S.; Long, S. C.; Roa-Espinosa, A.; Runge, T. Centrifuge Separation Effect on Bacterial Indicator Reduction in Dairy Manure. Journal of Environmental Management 2017, 191, 268–274. https://doi.org/10.1016/j.jenvman.2017.01.022. 65 (18) Masse, L.; Massé, D. I.; Pellerin, Y. The Use of Membranes for the Treatment of Manure: A Critical Literature Review. Biosystems Engineering 2007, 98 (4), 371–380. https://doi.org/10.1016/j.biosystemseng.2007.09.003. (19) Lee, C. S.; Robinson, J.; Chong, M. F. A Review on Application of Flocculants in Wastewater Treatment. Process Safety and Environmental Protection 2014, 92 (6), 489– 508. https://doi.org/10.1016/j.psep.2014.04.010. (20) Narasiah, K. S.; Morasse, C.; Lemay, J. Phosphorus Removal from Aerated Lagoons Using Alum, Ferric Chloride and Lime. Water Quality Research Journal 1994, 29 (1), 1–18. https://doi.org/10.2166/wqrj.1994.001. (21) Chen, S.; Liao, W.; Liu, C.; Wen, Z.; Kincaid, R. L.; Harrison, J. H.; Elliott, D. C.; Brown, M. D.; Solana, A. E.; Stevens, D. J. Value-Added Chemicals from Animal Manure; PNNL14495, 15009485; 2003; p PNNL-14495, 15009485. https://doi.org/10.2172/15009485. (22) Ueda Yamaguchi, N.; Cusioli, L. F.; Quesada, H. B.; Camargo Ferreira, M. E.; FagundesKlen, M. R.; Salcedo Vieira, A. M.; Gomes, R. G.; Vieira, M. F.; Bergamasco, R. A Review of Moringa Oleifera Seeds in Water Treatment: Trends and Future Challenges. Process Safety and Environmental Protection 2021, 147, 405–420. https://doi.org/10.1016/j.psep.2020.09.044. (23) Allahghadry, T.; Bojesen, A. M.; Whitehead, B. J.; Antenucci, F. Clarification of LargeVolume Bacterial Cultures Using a Centrifuge-Free Protocol. Journal of Applied Microbiology 2022, 133 (2), 870–882. https://doi.org/10.1111/jam.15608. (24) Wu, J.; Zhong, X. Relationship between the Particle Size and Nutrient Distribution of Feces and Nutrients in Pigs and Dairy Cows and the Efficiency of Solid-Liquid Separation. IOP Conf. Ser.: Earth Environ. Sci. 2020, 514 (5), 052035. https://doi.org/10.1088/17551315/514/5/052035. (25) Leach, K. A.; Archer, S. C.; Breen, J. E.; Green, M. J.; Ohnstad, I. C.; Tuer, S.; Bradley, A. J. Recycling Manure as Cow Bedding: Potential Benefits and Risks for UK Dairy Farms. The Veterinary Journal 2015, 206 (2), 123–130. https://doi.org/10.1016/j.tvjl.2015.08.013. (26) Hirota, R.; Kuroda, A.; Kato, J.; Ohtake, H. Bacterial Phosphate Metabolism and Its Application to Phosphorus Recovery and Industrial Bioprocesses. Journal of Bioscience and Bioengineering 2010, 109 (5), 423–432. https://doi.org/10.1016/j.jbiosc.2009.10.018. (27) Adler, J. A Method for Measuring Chemotaxis and Use of the Method to Determine Optimum Conditions for Chemotaxis by Escherichia Coli. Journal of General Microbiology 1973, 74 (1), 77–91. https://doi.org/10.1099/00221287-74-1-77. (28) Adler, J.; Hazelbauer, G. L.; Dahl, M. M. Chemotaxis Toward Sugars in Escherichia Coli. J Bacteriol 1973, 115 (3), 824–847. https://doi.org/10.1128/jb.115.3.824-847.1973. (29) Peterson, B. W.; Sharma, P. K.; Van Der Mei, H. C.; Busscher, H. J. Bacterial Cell Surface Damage Due to Centrifugal Compaction. Appl Environ Microbiol 2012, 78 (1), 120–125. https://doi.org/10.1128/AEM.06780-11. (30) Spencer, C. N.; King, D. L. Regulation of Blue-Green Algal Buoyancy and Bloom Formation by Light, Inorganic Nitrogen, C02, and Trophic Level Interactions. Hydrobiologia 1987, 144 (2), 183–191. https://doi.org/10.1007/BF00014531. (31) Cano-Ramirez, D. L.; Saskia De Fraine, T.; Griffiths, O. G.; Dodd, A. N. Photosynthesis and Circadian Rhythms Regulate the Buoyancy of Marimo Lake Balls. Current Biology 2018, 28 (16), R869–R870. https://doi.org/10.1016/j.cub.2018.07.027. 66