UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Modelling nitrogen mineralization from biosolids Rowell, Douglas Murray 1999

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_1999-0091.pdf [ 3.99MB ]
Metadata
JSON: 831-1.0089135.json
JSON-LD: 831-1.0089135-ld.json
RDF/XML (Pretty): 831-1.0089135-rdf.xml
RDF/JSON: 831-1.0089135-rdf.json
Turtle: 831-1.0089135-turtle.txt
N-Triples: 831-1.0089135-rdf-ntriples.txt
Original Record: 831-1.0089135-source.json
Full Text
831-1.0089135-fulltext.txt
Citation
831-1.0089135.ris

Full Text

M O D E L L I N G NITROGEN MINERALIZATION F R O M BIOSOLIDS by DOUGLAS M U R R A Y ROWELL B.Sc. (Forestry) Hons., The Australian National University, 1992 A THESIS SUMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE F A C U L T Y OF G R A D U A T E STUDIES (Department of Forestry) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH C O L U M B I A April 1999 © Douglas Murray Rowell, 1999 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. The University of British Columbia Vancouver, Canada Department DE-6 (2/88) A B S T R A C T Decomposition and net nitrogen mineralization from four biosolids, wheat straw, paper fines and Douglas-fir foliar litter were described and modelled. The initial chemical composition of these materials was characterized in terms of organic matter, carbon, nitrogen, proximate fraction analysis and solid-state 1 3 C nuclear magnetic resonance spectroscopy. Patterns of decomposition and net nitrogen mineralization over time were described in three incubation trials, one in the greenhouse and two in the field. Patterns were modelled based on the initial chemical characterization. Rates of decomposition were strongly related to the lignocellulose index and the carbon to organic matter ratio. The decomposition model extrapolated well to two field sites when site-specific correction factors were applied. Net nitrogen mineralization was most effectively predicted by initial organic nitrogen concentration and the phenolic content of the materials. While the mineralization models extrapolated less well to the field sites, the variables employed in the greenhouse model were relevant in the field and would be a useful starting point for further modelling of field nitrogen mineralization. Among biosolids there was a strong correlation between organic N concentration and N M R indices of protein, supporting other studies which have found positive correlations between protein content, organic N content and N mineralization in biosolids. Protein indices as described by N M R appear to be of quantitative value and may prove useful in predicting N mineralization from biosolids. TABLE OF CONTENTS Page ABSTRACT ii T A B L E OF CONTENTS Hi LIST OF TABLES vi LIST OF FIGURES '. viii A C K N O W L E D G E M E N T S x INTRODUCTION 1 1. N MINERALIZATION PROCESSES F R O M ORGANIC RESIDUALS 3 1.1 DECOMPOSITION 3 1.1.1 Processes of Decomposition 3 1.1.2 Patterns of Decomposition 5 1.2 IMMOBILIZATION 7 1.2.1 Biological Immobilization 7 1.2.2 Chemical Immobilization 8 2. PREDICTING N MINERALIZATION 9 2.1 T H E EMPIRICAL APPROACH 9 2.2 T H E PROCESS-BASED A P P R O A C H 11 2.2.1 Climate 11 2.2.1.1 Temperature 12 2.2.1.2 Moisture 12 2.2.2 Substrate Chemistry 13 2.2.2.1 C: Nratio , 14 2.2.2.2 Carbon Chemical Structure 14 2.2.3 Site 17 iii 2.3 SCOPE OF THE STUDY 19 3. METHODS : 20 3.1 OVERVIEW 20 3.2 ORGANIC RESIDUALS 21 3.3 PRE-INCUBATION ANALYSES 23 3.4 INCUBATION TRIALS 26 3.4.1 Greenhouse 26 3.4.2 Field Sites 28 3.5 POST-INCUBATION ANALYSES 30 3.6 MODELLING • 30 3.7 FIELD VALIDATION 32 4. RESULTS 34 4.1 CHEMICAL CHARACTERIZATION OF RESIDUALS 34 4.1.1 Chilliwack Biosolids 36 4.1.2 Whistler Biosolids 37 4.1.3 Annacis Island Biosolids 37 4.1.4 Lionsgate Biosolids 38 4.1.5 Douglas-fir Foliar Litter 38 4.1.6 Wheat Straw 39 4.1.7 Paper Fines 40 4.2 DESCRIBING N MINERALIZATION 40 4.2.1 Decomposition 40 4.2.1.1 Greenhouse 40 4.2.1.2 Field. 43 4.2.2 Net N Mineralization 48 4.2.2.1 Greenhouse 49 4.2.2.2 Field. 51 iv 4.3 MODELLING N MINERALIZATION 54 4.3.1 Decomposition 54 4.3.2 Net N Mineralization 57 4.3.2.1 All Residuals 57 4.3.2.3 AUBiosolids 62 4.3.2.2 Secondary Biosolids 65 4.4 FIELD VALIDATION 68 4.4.1 Decomposition 68 4.4.2 Net N mineralization 71 4.4.2.1 All Residuals 77 4.4.2.2 Biosolids ••• 73 4.4.2.3 Secondary Biosolids 74 5. DISCUSSION . 77 5.1 DESCRIBING N MINERALIZATION 77 5.1.1 Organic Residuals 77 5.1.2 Decomposition 78 5.1.3 Net N Mineralization 80 5.2 MODELLING N MINERALIZATION 85 5.2.1 Modelling Decomposition 85 5.2.2 Modelling Net N Mineralization 87 5.3 FIELD VALIDATION 89 5.3.1 Decomposition Model 90 5.3.2 Net N Mineralization Models 90 6. CONCLUSIONS ." 92 REFERENCES 94 APPENDIX Solid-state 1 3 C nuclear magnetic resonance spectra of the seven organic residuals 100 v L I ST O F T A B L E S Page Table 1. Organic residuals used in the study 21 Table 2. Incubation methods and replicates at each sampling period for three parallel trials 26 Table 3. Ranges of sample dry weight (g) 27 Table 4. Dependent and independent variables tested in the multiple linear models 31 Table 5. Initial chemical characterization of the seven organic residuals 35 Table 6. Solid-state 1 3 C nuclear magnetic resonance (NMR) spectroscopy, CPMAS analysis of the seven organic residuals. Indices and ratios of spectral regions 36 Table 7. Net N mineralization (% of initial organic N lost) after incubating for 391 days in the greenhouse and at two field sites 49 Table 8. Net N mineralization (g N lost / kg O M applied) during a 391-day incubation in the greenhouse 50 Table 9. Difference in net N mineralization (g N lost / kg O M applied) among three sites (greenhouse, interior forest and coastal forest), during 391-day incubations of seven organic residuals. „ 52 Table 10. Difference in net N mineralization (g N lost / kg O M applied) among seven organic residuals during 391-day incubations at three sites (greenhouse, interior forest and coastal forest) 53 Table 11. Correlation of independent variables over time with decomposition (% mass loss of organic matter) of organic residuals during a 391-day greenhouse incubation 55 Table 12. Correlation of independent variables and time with net N mineralization (g N lost / kg O M applied) from organic residuals during a 391-day greenhouse incubation 58 vi Table 13. Correlation of independent variables and time with net N mineralization (g N lost / kg O M applied) from four biosolids during a 391-day greenhouse incubation 63 Table 14. Correlation of independent variables and time with net N mineralization (g N lost / kg O M applied) from three secondary or part-secondary biosolids during a 391-day greenhouse incubation 66 Table 15. Validation of the greenhouse-based model of decomposition (organic matter loss)at two field sites 69 Table 16. Corrected greenhouse model of decomposition (% organic matter loss) employed at two field sites 70 Table 17. Validation of the greenhouse-based model of net N mineralization (g N lost / kg O M applied) from all residuals, at two field sites 71 Table 18. Corrected greenhouse model of net N mineralization (g N lost / kg O M applied) from all residuals, applied to two field sites 72 Table 19. Validation of the greenhouse-based model of net N mineralization (g N lost / kg O M applied) from the four biosolids, at two field sites 73 Table 20. Corrected greenhouse model of net N mineralization (g N lost / kg O M applied) from four biosolids, applied to two field sites 74 Table 21. Validation of the greenhouse-based model of net N mineralization (g N lost / kg O M applied) from secondary or partial secondary biosolids, at two field sites •. 75 Table 22. Corrected greenhouse model of net N mineralization (g N lost / kg O M applied) from secondary and part-secondary biosolids, applied to two field sites 76 LIST OF F I G U R E S Page Figure 1. Organic matter loss (%) during a 391 -day incubation in the greenhouse 41 Figure 2. Weight loss (%) during a 391-day incubation in the greenhouse 41 Figure 3. Organic matter loss (%) of biosolids and Douglas-fir litter during a 3 91 -day incubation in the greenhouse 42 Figure 4. Weight loss (%) of biosolids and Douglas-fir litter during a 391-day incubation in the greenhouse 43 Figure 5. Weight loss (%) of residuals during 391-day incubations in the greenhouse, interior forest and coastal forest 44 a: Annacis Island biosolids 44 b: Chilliwack biosolids 44 c: Lionsgate biosolids 45 d: Whistler biosolids 45 e: Paper fines 46 f: Wheat straw -..46 g: Douglas-fir foliar litter 47 Figure 6. Weight loss (%) of each residual during a 391-day incubation in a coastal forest 47 Figure 7. Weight loss (%) of each residual during a 391-day incubation in an interior forest 48 viii Figure 8. Net N mineralization (g N lost / kg O M applied) of each residual during a 391-day incubation in the greenhouse 51 Figure 9. Correlation between organic N concentration and the alkyl signal in the seven organic residuals 59 Figure 10. Correlation between organic N concentration and the alkyl to O-alkyl ratio in the seven organic residuals 60 Figure 11. Correlation between organic N concentration and the methoxyl signal in the seven organic residuals 60 Figure 12. Correlation between the lignin:Norganic ratio and net N mineralization (g N lost / kg O M applied) from the seven organic residuals after incubating for 391 days '. 61 Figure 13 Correlation between organic N concentration and lignocellulose index in the four biosolids 64 Figure 14. Correlation between organic N concentration the carboxyl signal in the three secondary or part-secondary biosolids 67 A C K N O W L E D G E M E N T S Thanks to: Dr Cindy Prescott for supervision, and academic and financial support Dr Hamish Kimmins for financial support Dr Les Lavkulich and Dr Tim Ballard for providing laboratory facilities Carol Dyck, Anh-Toan Tran and Keren for training and assistance in the lab Dr Caroline Preston for N M R analysis and advice on interpretation Graeme Hope for additional analyses Mike van Ham and Jonn Braman for background information and advice on biosolids Jen, Karen, Pascalle, David, Wendy, John and Todd Reading for their help in the field or lab David Blevins, Leandra Blevins and Dr Valerie Le May for help with SAS Dr Valerie Le May, Dr Tony Kozak and Lisa Zabek for statistical advice The B.C. Ministry of Forests, Kamloops District and Squamish District, for research sites. I N T R O D U C T I O N Biosolids is the term given to treated sewage sludge that can be beneficially applied to land as an organic amendment or as a fertilizer. The term was coined to distinguish those sludges that were sufficiently stabilized, and had sufficiently low concentrations of contaminants for land application, from the wide spectrum of industrial and municipal sludges, some of which may not be suitable for land application. Beneficial application of biosolids to land avoids the need for conventional waste disposal options such as landfills, river and ocean outfalls or incineration, which are becoming increasingly costly and less socially or environmentally acceptable. Biosolids has been demonstrated to be an effective forest fertilizer in a wide range of forest types and on a wide range of sites (Henry, 1986; Nguyen et al, 1992; de Montigny and Zumrawi, 1996; Turner et al, 1996). In order to facilitate the use of biosolids as a fertilizer option it is important to be able to predict the amount and rate of nutrient mineralization from these materials. The focus of this study was on nitrogen (N) because it is the most commonly limiting nutrient in many temperate forest ecosystems (Binkley, 1986; Nason and Myrold, 1992). 1 The objectives of the study were: To examine the N mineralization patterns of biosolids, comparing them with other organic residuals To model the N mineralization from biosolids and other organic residuals, and To carry out field validation of the models on independent data sets. 1. N MINERALIZATION PROCESSES FROM ORGANIC RESIDUALS Mineralization is a microbially mediated process by which organically bound nutrients become inorganic. It is the consequence of a complex system of microbially or biochemically mediated processes, including decomposition and immobilization. Net N mineralization is the amount of N released during decomposition (gross mineralization) minus the amount of nutrient subsequently immobilized in the microbial biomass or in organic complexes (immobilization). It is used as an indicator of N availability to plants, because they primarily use inorganic forms of N . 1.1 D E C O M P O S I T I O N Decomposition is the process by which organic matter is broken down from complex to simpler forms, mineralizing nutrients in the process. In aerobic soil environments it is dominated by three processes; leaching, comminution and catabolism (Swift et al, 1979). 1.1.1 Processes of Decomposition Leaching loss of soluble and suspended substances can occur by mass flow as precipitation and soil water percolate through the organic material. 3 Comminution is the breakdown of the macro structure of the organic material. This is carried out by macrofauna such as earthworms, dung beetles, and larger mesofauna including springtails, mites and nematodes. Catabolism is the process whereby nutrients are ultimately mineralized or immobilized. It is a microbially mediated process whereby complex organic structures are enzymatically degraded into simpler organic structures, and the carbon therein ultimately respired as CO2. The microbes utilize energy released during respiration of the carbon. Some N mineralized from the organic structure during the catabolic process is re-immobilized by the microbial biomass (see 1.2 IMMOBILIZATION). The remainder generally remains in mineral form and can be an important source of N for plants. As catabolism proceeds and more carbon is respired as CO2, the carbon to nitrogen (C:N) ratio of the material gradually declines. Organic matter decomposes at different rates according to its carbon chemical structure. Labile (rapidly decomposing) organic matter consists of simple organic structures with molecular bonds which are readily degraded, for a large net energy gain; thus it supports a larger microbial biomass. Recalcitrant (slowly decomposing) organic matter consists of more complex molecular structures with a higher density of more stable chemical bonds. It is degraded more slowly for a smaller (in the extreme, negative) net energy gain, and so supports a smaller microbial biomass. The preferential decomposition of labile organic matter over more recalcitrant organic matter results in a rapid rate of decomposition in early decay, gradually slowing down as the size and the level of activity of the microbial biomass diminishes. For example, it has been shown that composting (controlled decomposition) of biosolids results in the preferential loss of the lower molecular weight proteins (Piotrowski et al, 1984; Garcia et al, 1992). A similar pattern has been observed in composting organic matter (Inbar et al, 1989). Cogle et al. (1989) also found decomposition of wheat straw resulted in preferential loss of carbohydrates (labile organic matter) over aromatic structures (recalcitrant organic matter). After a prolonged period of decomposition, when all the labile organic matter has been respired or mineralized, the most recalcitrant forms (such as polyphenols and other aromatic carbon structures) undergo a slow process of condensation, forming large, stable, heterogeneous molecular structures. This amorphous mass of organic structures is collectively termed humus. While this pool does undergo gradual turnover of organic matter and nutrients, it is very slow, and the fertilizer value of the organic material is negligible at this stage. 1.1.2 Patterns of Decomposition In the early stages of decay while the Cmutrient ratios are high and the organic matter is relatively labile, decay is often limited by lack of nutrients (Berg, 1986). This is because the microbial biomass tends to be larger, and has a higher requirement for nutrients (see 1.2 IMMOBILIZATION). A S decay proceeds, the C: nutrient ratio declines and the microbial biomass is diminished; so nutrients become less limiting. At the same time the residual carbon becomes progressively more recalcitrant. Thus in the later stages of decay, it is the decomposition rate of this recalcitrant carbon which limits decomposition (McClaugherty et al, 1985; Berg, 1986). Melillo et al. (1989) demonstrated the importance of the recalcitrant organic fraction in the later stages of decomposition of red pine litter. They measured the progression of the lignocellulose index (LCI) throughout the decomposition of red pine litter. The LCI is defined 5 as the ratio of acid-insoluble organic matter ("lignin") to the sum of acid-insoluble and acid-soluble ("cellulose") fractions. They found the LCI increased through early decay to about 0.7 as labile carbohydrates were preferentially decomposed over lignin. At a LCI of between 0.7 and 0.8 the index levelled out and became fairly constant. Their explanation was that in late decay, the cellulose embedded within the matrix of lignin was dependent on the simultaneous decomposition of the lignin, and was therefore limited by the rate of decay of the lignin. The boundary between early and late stage decay, then, is the point at which the carbon to nutrient ratio is sufficiently low that nutrients are no longer limiting. Up to this point the rate of decomposition is limited by nutrient availability and characterized by net immobilization of nutrients (usually N or phosphorus). In late decay the rate of decomposition is determined by the decomposition of recalcitrant compounds (e.g. lignin), and is characterized by net mineralization of nutrients. The two-stage pattern of decay described by Berg (1986) for forest litters is much more accentuated in organic residuals that have undergone biological digestion, such as biosolids. The organic matter in biosolids appears to occur in two distinct pools; a labile pool, and a recalcitrant pool. The turnover of these pools is dramatically different, the latter having a decay constant (k) one or even two orders of magnitude lower than the former (Boyle and Paul, 1989; Lerch et al, 1992). The labile pool is comprised mostly of dead microbial cells remaining after the digestion process (Lerch et al, 1992), while the recalcitrant pool is comprised of residual organic matter (Boyle and Paul, 1989). Zibilske (1997) observed a similar two-phase pattern of decomposition in primary and secondary paper mill sludges. 6 Decomposition in late decay has also been correlated with other variables. For example, Johansonn (1994) found a positive correlation between calcium content and lignin degradation in very late decay, which she attributed to the satisfying of calcium requirements of lignolytic fungi. It has also been shown that NHV" and amine (NH2) groups can inhibit the production of lignolytic enzymes in fungi (Keyser, 1978; Fenn et al, 1981). 1.2 IMMOBILIZATION Immobilization of nutrients during decomposition of organic residuals can be either biological, within the microbial biomass, or it can be chemical, a function of the composition of the organic material itself. 1.2.1 Biological Immobilization Microbes require a certain amount of nutrients for biosynthesis of their own structure, and for basic metabolic functions. They immobilize these from the nutrient pool mineralized during the decay of the organic material. A high nutrient demand and low nutrient supply (such as in early decay) will result in net immobilization. That is, nutrients released from the organic material are fully utilized by the microbial biomass as it expands in response to a carbon energy source. At very high nutrient demand the microbial biomass may also immobilize available nutrients from the adjacent soil to augment the supply from the decomposing organic material. When supply of nutrients exceeds demand by the microbial biomass, net mineralization occurs, and carbon, rather than nutrients, becomes limiting to microbial growth. 7 1.2.2 Chemical Immobilization Organic residuals contain varying amounts of polyphenolics such as lignin and tannin. These compounds have been shown to immobilize proteins and amine groups, by condensation reactions, to form phenolic-protein complexes (Handayanto et al, 1997). Thus litters with high lignin content, tannery wastes or well-humified materials such as composts can chemically immobilize significant quantities of N . Furthermore, polyphenolics tend to be quite resistant to enzymatic decay, and any N so immobilized is likely to have a slow turnover. 8 2. P R E D I C T I N G N M I N E R A L I Z A T I O N To facilitate the use of biosolids as a fertilizer, it is important to be able to predict the amount and rate of N mineralization. There are two broad approaches. The first is an empirical approach. It assumes that decomposition of organic materials and mineralization of organic nutrients from within those materials follows some mathematical function; and that defining this function (without necessarily defining the processes behind it) is sufficient for prediction. The second approach is process-based. Process-based models are based upon an understanding of factors which influence decomposition or immobilization processes. 2.1 THE EMPIRICAL APPROACH Stanford and Smith (1972) first derived mathematical functions for modelling the mineralization of N from soils. Using sequential aerobic incubations (up to 30 weeks) of a variety of soil types, they concluded that N mineralization approximated first order kinetics, and that the total N mineralized at any point in time could be defined as N t =No(l-e k t ), where, N t= cumulative total of N mineralized at time t No = initial amount of potentially mineralizable N k= a first-order reaction constant: rate of N mineralization per unit amount of mineralizable N currently present t= time from the beginning of the incubation No and k are specific to the given soil (or organic material). First order kinetics described N mineralization from a range of different organic materials including biosolids (Hsieh et al, 1981; Chae and Tabatabai, 1986; Serna and Pomares, 1992), animal manures (Chae and Tabatabai, 1986) and composts (N'Dayegamiye et al, 1997). It has also been used to model decomposition of proteins in biosolids (Lerch et al, 1992). Molina et al (1980) demonstrated that Stanford and Smith's original model was a poor predictor of N mineralization from soils during the first twelve weeks of incubation. They improved upon the model by fitting two exponential curves to the data rather than just one. They attributed this better fit to separate representation of two distinct fractions of organic N , one rapid and one slow. This resulted in the function N t =N s ( l -e k t ) + N r(l-e- h t), where, N t= cumulative total of N mineralized at time t N s = initial amount of potentially mineralizable N in the slow fraction N r = initial amount of potentially mineralizable N in the rapid fraction k= decay constant for the slow fraction h = decay constant for the rapid fraction t= time since the beginning of decomposition 10 Lindemann and Cardenas (1984) found N mineralization from soils amended with anaerobically digested biosolids was also better modelled using the two parallel exponential functions. Gilmour et al. (1996) found decomposition in biosolids-soil mixtures was better modelled using the double function. Various other empirical functions have been used to model decomposition or nutrient mineralization from organic materials with some success. Whitmore (1996) demonstrated that decomposition of crop residues can be adequately modelled using a function following second order kinetics (hyperbolic). The study found that the second order function was an improvement on a single first order kinetics function but was still not as good as two parallel first order curves. Empirical modelling is quite site-specific and substrate-specific. For example, Chae and Tabatabai (1986) reported five different patterns of N release depending on the type of organic amendment measured and the soil to which it was added. This limits the confidence to which results can be extrapolated to other materials, other soils and other climates. 2.2 T H E PROCESS-BASED APPROACH The other approach to modelling nutrient mineralization is process-based; to include in a model those factors which influence decomposition and immobilization. These variables can be grouped into three broad categories; climate, substrate chemistry and site factors. 11 2.2.1 Climate Climate is the overriding factor controlling rates of decomposition at large scales (Aerts, 1997; Kochy and Wilson, 1997). Meentemeyer (1978) found actual evapotranspiration (AET) described more than half the variation in litter decomposition rates in climates ranging from warm temperate to sub-polar. The influence of climate is especially apparent in early decay (Meentemeyer, 1978; Berg, 1986; Johansson 1994). The effect of climate can be divided into the effects of moisture and the effect of temperature. 2.2.1.1 Temperature Temperature affects both the size of the microbial biomass and the metabolic activity of individual organisms. Both of these affect the rate of decomposition of organic substrates. Decomposing organisms appear to be much more sensitive to changes in temperature than changes in moisture. Edmonds (1980) found that temperature was the most important variable in describing variation in N mineralization from forest litter at a range of field sites in western Washington. Sierra (1997) found N mineralization from soil cores in the laboratory was more sensitive to changes in temperature than changes in moisture. Miller (1974) found that soil temperature was one of the major variables controlling decomposition of biosolids in the soil, in a laboratory incubation study. 2.2.1.2 Moisture Microorganisms maintain an internal moisture content of about 80% over a wide range of external water potentials (Swift et al, 1979). Thus, as mentioned, they are not as sensitive to changes in moisture as changes in temperature. Nevertheless extreme conditions in the soil 12 can have a dramatic effect on decomposition rates. Growth of fungi in the soil drops significantly below matric potentials of about -150 Bar. Bacteria are generally less resistant to desiccation than fungi. They become inactive in the soil below about -15 Bar (Swift et al., 1979). Also important, is the potential at which mass flow of water through the soil pores ceases. At this moisture content (perhaps as little as -1 Bar), translocation of decomposing microorganisms from one substrate to another can be inhibited, particularly in soils where organic substrate occurs in discrete "islands" (Griffin, 1972). This is particularly important for bacteria, which rely upon soil water for movement, though not as important for fungi, which can bridge dry soil pores with their mycelium. At the other extreme, very wet conditions can lead to anaerobiosis. Anaerobic decomposition is much slower and less efficient than aerobic decomposition, and the process is slowed dramatically under these conditions (Swift et al, 1979). 2.2.2 Substrate Chemistry Substrate chemistry factors include carbon to nutrient ratios and carbon chemical structure. Substrate chemistry (particularly carbon chemical structure) has been found to exert more influence on decomposition than climate at local scales (Kochy and Wilson, 1997). Scott and Binkley (1997) found that, even on a sub-continental scale, substrate chemistry (the lignin:N ratio) explained more of the variation in net N mineralization than climate factors. 13 2.2.2.1 C : N ratio The C:N ratio has been shown to be a useful index of both decomposition and net nutrient mineralization particularly in early decay when decomposition is nutrient limited. Edmonds (1980) found a strong negative correlation between rate of decomposition and the C:N ratio of forest litters, where the C:N ratio varied widely. A negative correlation has been found between N mineralization from biosolids-amended soils and the C:N ratio of the biosolids/soil mixture (Barbarika et.. al, 1985; Hattori and Mukai, 1986). N'Dayegamiye et al. (1997) found that more mature composts, with lower C:N ratios and less labile organic matter, contributed more to soil N , presumably because of less microbial immobilization. 2.2.2.2 Carbon Chemical Structure Labile materials generally have fewer stable molecular bonds and a smaller, simpler structure. Recalcitrant materials tend to have more stable molecular bonds and a larger more complex structure. Therefore indices which reflect the carbon chemical structure of a material often have a strong correlation to its rate of decomposition. The carbon chemical structure of organic materials has most often been measured using the ratios of fractions yielded by proximate analysis. Other technologies have enabled carbon chemical structure to be described at a more detailed molecular level. Proximate analysis of organic materials is an analytical procedure used to divide the material into components of differing solubilities. The proportions of each fraction and ratios thereof have proven to be a good index of rate of decomposition (Agren and Bosatta,1996). The acid-insoluble fraction, also known as Klason lignin, or just lignin, has proven a 14 particularly useful index in the study of litter decomposition. While the acid-insoluble fraction does not always include all true lignin (some is dissolved), and often includes other acid-insoluble compounds (such as tannin and cutin), it nevertheless remains a good measure of the recalcitrant fraction of the organic material (Preston et al, 1997). Schlesinger and Hasey (1981) found that lignin content had a strong effect on decomposition of shrub foliar litter (Ceanothus megacarpus and Salvia mellifera), in the Californian chapparal, and was a better predictor of decomposition than the C:N ratio. They inferred from this that the system was not N-limited. Meentemeyer (1978) found lignin content to be a highly significant factor (second to AET) in determining the rate of decomposition of litters at the global scale. Hattori and Mukai (1986) found the variation in carbon mineralization in biosolids/soil mixtures was best described using a combination of the acid-insoluble fraction and inorganic matter content. Melillo et al (1982) combined the C:N ratio (an index of early decomposition) and lignin content (an index of later decomposition) into the lignin:N ratio, to model the whole decay continuum. They found this composite index was good for a wide range of hardwood litters as well as woody material decomposing in a stream. They noted that lignin alone was not as good an index of decomposition, especially in environments where exogenous N was low (i.e. that were N-limited). Scott and Binkley (1997) demonstrated a strong linear correlation between net N mineralization and the lignin:N ratio for a wide range of forest litters in a wide range of age classes and climate. Aerts and Caluwe (1997) also found strong correlations between the decomposition rate in later decay and combined carbon chemistry and nutrient ratios (including lignin:N, phenolic:N and phenolic:P). 15 Ratios combining phenolic content and N content have proved useful in modelling N mineralization. The use of the phenolic content in the ratio takes account of the effect of chemical immobilization by polyphenolics and phenol groups from degraded lignin. Fox et al. (1990) found the (lignin+polyphenol):N ratio to be a good index for predicting N mineralization from incorporated legume litter. A number of technologies have been developed that indirectly measure aspects of molecular structure. The proportion of various bonds (more stable or less stable) in an organic material can be used to make predictions of the rate of decomposition of the material. The most widely used of these for organic matter studies is nuclear magnetic resonance (NMR) spectroscopy. The N M R spectrum is a spectrum of signals given off by the nuclei of isotopes ( 1 3 C or 1 5 N) rotated at very high speed (typically 4000 to 5000 Hz). The resonance varies depending on the nature of the molecular bonds involved. For example, carbon in a saturated aliphatic chain gives a distinctly different signal to that of carbon in an aromatic ring. Materials with a high density of a given chemical structure will reflect strong signals of the associated magnetic resonance. Changes in chemical structure due to decomposition will be reflected by changes in the N M R spectrum (Baldock and Preston, 1995). N M R spectroscopy has been used to describe changes in chemical structure of decomposing organic matter (Inbar et al, 1989), of biosolids proteins (Piotrowski et al, 1984; Garcia et al, 1992), of wheat straw (Cogle et al, 1989), of plant litter (Wershaw et al, 1996; Preston et al, 1997), and of fish and crab scrap composts (Preston et al, 1986). 16 Infrared (IR) spectroscopy employs a similar principle, detecting the electromagnetic resonances in the infrared spectrum. Piccolo et al. (1992) used IR spectroscopy to compare humic acids formed by manure amendments to soil with those formed by biosolids amendments. They found that humic acid structure reflected the structure of the original amendment materials; biosolids humic acid being higher in peptides (protein fragments) while manure humic acids were higher in polysaccharides (cellulose fragments). Molecular weight (MW) distribution uses a variety of methods to separate organic matter into high molecular weight and low molecular weight fractions, based on some critical size. Specific methods can include ultra-fine membrane filters, gel filters, or use of electrochemical gradients (gel electrophoresis and isoelectric focussing). Lerch et al. (1993) used M W analysis on the protein fraction in biosolids, and found them to be relatively low weight (i.e. labile). Piccolo et al. (1992) found humic acids from the biosolids-amended soils had higher molecular weights than those from the manure-amended soils, suggesting a higher degree of humification in soils following biosolids amendment. 2.2.3 Site Decomposition of organic residuals is affected by the site to which they are applied. Prescott (1996) observed that forest floor type affected the rate of decomposition of identical litter types under identical climatic environments. McTiernan et al. (1997), and Fyles and Fyles (1993) found interactions between litter types in mixtures in terms of their decomposition rates. The mechanisms of these effects are not clear but they suggest that one litter facilitates (or inhibits) decomposition of the other through nutrient dynamics, supply of polyphenols, or exposure to different soil fauna and microorganisms. 17 Barbarika et al. (1985) found significant correlations between N mineralization from biosolids-amended soil and both the N content of the soil only (a positive correlation) and the C:N ratio of the soil only (a negative correlation). Anaerobic conditions in waterlogged soils can dramatically reduce rates of decomposition. Because so much of the process of nutrient mineralization is mediated by microorganisms, it follows that the inherent nature and function of the microbial biomass will have its own fundamental effect. In a general model of the carbon and N interactions in soils, Bosatta and Agren (1985) proposed two microbial biomass variables: microbial decomposition efficiency and microbial growth rate. Both can vary with the relative lability of the material, but are nevertheless specific to the residual microbial biomass. Hyvonen et al. (1996) tested this model on a range of organic materials including sawdust, straw, peat, manures and biosolids with reasonable success. Their sensitivity analysis of the model indicated that the microbial efficiency variable was second after substrate quality in terms of sensitivity. The term quality refers to the relative decomposability of the material; labile materials are high quality and recalcitrant materials are low quality The role of larger soil fauna in facilitating comminution has been described by Swift et al. (1979). Differences in populations of soil fauna between sites, notably the absence of earthworms in mor humus types may afford different decomposition dynamics (Berg, 1986). 18 2 . 3 S C O P E O F T H E S T U D Y Empirical models are generally substrate-specific, each material having its own rate constant (k). Therefore they are of limited use for modelling a range of materials as potential fertilizer options. They are also of little use in exploring the processes behind mineralization of N from these materials. For these reasons this study employed a process-based approach to modelling. Substrate chemistry is the most important factor determining N mineralization from organic residuals on a regional to sub-continental scale. It is also the factor that can be most easily controlled by the forest manager, assuming the site is already chosen. For these reasons the influence of substrate chemistry was the main focus of the study. Nevertheless climate and site factors are important. Many authors have emphasized that decomposition of a substrate is affected by both the climate (temperature and moisture) and the chemical structure of the substrate (Meentemeyer, 1978; Edmonds, 1980; Aerts, 1997; Scott and Binkley, 1997). Therefore climate and site factors should be included as far as practicable in a model based on substrate chemistry, to facilitate its transferability into the field. A range of biosolids and other organic residuals were collected and characterized in terms of substrate chemistry. Materials were then incubated and their rates of decomposition and net N mineralization measured. Models of decomposition and net N mineralization were developed based on the initial chemical characterization, and then validated in the field. In lieu of modelling climate and site factors, field correction factors were developed for the models, to facilitate their use at different field sites. 19 3. M E T H O D S 3.1 OVERVIEW Seven different organic residuals were selected for the study. These included four biosolids, two high C:N ratio materials (paper fines and wheat straw), and a representative forest litter (Douglas-fir foliar litter). The chemistry of the materials was characterized by analyzing organic matter content, carbon content, and organic N content. Two analyses were also carried out to assess the carbon chemical structure of the selected materials: proximate analysis and solid-state 1 3 C nuclear magnetic resonance (NMR) spectroscopy. Selected ratios of the different analyses were also determined to assist in the chemical characterization. Samples of each material were incubated for a range of periods up to one year. There were three parallel incubation trials; a microcosm trial in the greenhouse, a field trial in an interior forest of British Columbia (B.C.), and a field trial in a coastal forest of B.C. At the end of each sampling period replicates of the material were destructively sampled, and analyzed for weight loss, organic matter loss, carbon loss and organic N loss. Patterns of decomposition of the different residues, at the different sites, over time, were described in terms of organic matter mass loss (%) and total mass loss (%). Patterns of 20 net N mineralization from the different residues at the different sites over time were described in terms of organic N mass loss. Models for decomposition and net N mineralization were based on the greenhouse data. Multiple linear regression was employed to derive models for predicting the decomposition and net N mineralization, using independent variables based on the initial chemical characterization of substrates. Field validation of the predictive models was carried out using the two field trials as independent data sets. Correction factors were devised for the models to facilitate their extrapolation into the field. 3.2 ORGANIC RESIDUALS Seven different organic residuals were selected for the study, including four biosolids, wheat straw, paper fines and Douglas-fir foliar litter (Table 1). Table 1. Organic residuals used in the study Residual Description Chilliwack Chilliwack STP biosolids (mesophilic, anaerobic, secondary) Annacis Annacis Island STP biosolids (thermophilic, anaerobic, 70:30 primary to secondary) Whistler Whistler STP biosolids (ATAD: thermophilic, aerobic, secondary) Lionsgate Lionsgate STP biosolids (thermophilic, anaerobic, primary) Douglas-fir Douglas-fir foliar litter (CWH, Haney, B.C.) Wheat straw Straw (including 10% chaff) from Buckerfields stockfeed merchants, Burnaby, B.C. Paper fines unbleached thermo-mechanical pulp from Scott™ Paper, Burnaby, B.C. 21 The biosolids came from three sewage treatment plants (STPs) in the Greater Vancouver Region, and from the STP servicing the township of Whistler, B.C.. A l l the biosolids were dewatered i.e. water was removed from them to reduce them to a moisture content of about 75%. Otherwise the sewage treatment processes at these STPs were quite different. Chilliwack biosolids was mesophilic, anaerobic, secondary; i.e. the sewage was biologically digested at a mesophilic temperature (35°C to 37°C), in anaerobic digestion tanks, and was supplemented with the dead microflora from the digestion tanks after their demise (waste-activated sludge). The detention time in the digestion tanks was standard, about 15-18 days. Annacis Island biosolids was thermophilic, anaerobic and 30% secondary; i.e. the sewage was biologically digested at high temperatures (55°C to 57°C), in anaerobic digestion tanks. Thirty per cent of this came from the secondary stream and was therefore supplemented with waste-activated sludge, while 70% was from the primary waste stream and therefore did not include waste-activated sludge. The detention time in the digesters was much longer, because the plant was running under capacity. Detention time was unspecified but was considerably longer than the normal 15-18 day period (J. Braman, pers. comm., 1998) Lionsgate biosolids was thermophilic, anaerobic, primary; i.e. the sewage was biologically digested at thermophilic temperatures (55°C to 57°C), in anaerobic digestion tanks, and was not supplemented with waste-activated sludge. Detention time was very long for this 22 material, because the plant was built to accommodate a much larger throughput than the prevailing one, and may have been as long as 40-45 days (J. Braman, pers. comm., 1998). Whistler biosolids was autothermophilic, aerobic, i.e. it was biologically digested in aerobic digestion tanks which passed through mesophilic temperatures, ultimately reaching thermophilic temperatures (up to 60°C). It was effectively secondary biosolids because the digestive microflora were retained in the sludge as they were dewatered. Detention time in digestion tanks is very short in this process (4-5 days), (C. Jennings, pers. comm., 1999). However aerobic digestion is much more efficient than anaerobic digestion and the detention times are not directly comparable. Two high C:N ratio materials (paper fines and wheat straw) were selected as a contrast to the low C:N ratio biosolids. The paper fines came from the Scott Paper mill, and were a waste fibre from unbleached thermo-mechanical pulping. The wheat straw was a commercial product. It included about 10% chaff (glumes and other wheat-head material). Douglas-fir foliar litter was selected to compare the behaviour of biosolids with material more typical of decomposition studies. 3.3 PRE-INCUBATION ANALYSES Prior to incubation the materials were characterized in terms of chemical composition to identify variables for modelling decomposition and net N mineralization. This characterization 13 included organic matter content, carbon content, organic N content, proximate analysis, and C N M R . 23 Moisture content was determined by mean mass loss of 12-16 samples, following oven drying for 24 hours at 70 °C. For all other analyses samples were prepared by oven-drying, grinding and sieving (to 16-mesh or 1mm size) using a Tecator™ centrifuge mill. Organic matter content was determined by mass loss of a 0.5g sample following ignition in a muffle furnace, using the method described in Black (1965). Samples were preheated for one hour at 375° C to minimize charring. Heat was then increased to 550° C for 20 hours to complete the ashing. Carbon content was determined by C 0 2 release from a O.lg sample following total combustion in a Leco Analyzer (Laboratory Equipment Corporation, 1959). Organic N was defined as the difference between total N and extractable N determined on the adjacent samples. Total N was determined by colorimetric analysis of N H / ion concentration in a digested O.lg sample using a Lachat™ auto-analyzer (USEPA, 1983). Samples were digested in preparation for this analysis using a modified micro-Kjeldahl acid digestion (Bremner, 1996). This method included a salicylic acid step to reduce NO3" ions to N H / ions, to enable quantification of total N . Extractable N was determined by colorimetric analysis of N H / and NO x " ion concentrations in an extracted 0.5g sample using the Lachat™ auto-analyzer. In preparation for this analysis, samples were extracted for one hour with lOmL 24 of 2M KC1 solution, then filtered through No. 42 Whatman™ filter paper. The filter paper was previously washed with KC1 solution to remove possible residual N H / ions (Mulvaney, 1996). Proximate analysis was carried out by the British Columbia Ministry of Forests' laboratory at Glyn Road, Victoria using the forest fiber method described by Ryan et al. (1990). Samples were extracted with a mixture of 0.5M H2SO4 and cetyltrimethylammonium bromide (CTAB). Dry mass loss after the extraction was quantified and defined as the extractable fraction. The residue was then digested with 72% H 2 S O 4 . Dry mass loss after the digestion was quantified and defined as the acid-soluble fraction ("cellulose"). The residue from this process was then ashed for eight hours at 450°C. Dry mass loss after ignition was defined as the acid-insoluble fraction ("lignin" or Klason lignin). The lignocellulose index (LCI) was approximated by using the acid-insoluble fraction as a measure of lignin, and the acid-soluble fraction as a measure of cellulose (after Melillo et al, 1989)(see 1.1.2 Patterns of Decomposition). Solid-state C Nuclear Magnetic Resonance Spectroscopy (NMR) analysis was carried out by Dr Caroline Preston at the Pacific Forestry Research Centre (Victoria, B.C.). Samples were dried and ground, and about 300 mg was analyzed in a Bruker M S L 100 spectrometer at a rotation speed of approximately 4700 Hz. Two spectra were generated for each material. They were a cross-polarized spectrum with magic angle spinning (CPMAS), and a dipolar dephase (DD) spectrum (see APPENDIX). 25 CPMAS was used for quantifying the relative signal strength of different spectral regions (Preston et al, 1997). Each spectral region was quantified by measuring the ratio of the area under the spectral peaks in the region, to the area under the spectral peaks for the whole spectrum. Five regions were quantified, including alkyl C (0-48 ppm), methoxyl C (48-60 ppm), O-alkyl C (60-93 ppm), phenolic C (140-160 ppm) and carboxyl C (165-190 ppm). The DD spectrum amplifies signals at the extreme ends of the spectrum (aromatics and alkyls). It tends to suppress signals from the middle of the spectrum (polysaccharides) and therefore is not a quantitative spectrum (Preston et al, 1997). 3.4 INCUBATION TRIALS The materials were incubated in three parallel trials over a range of periods up to one year (Table 2). Day 0 was 7th July, 1997, i.e. mid-summer. Day 391 was 2nd September, 1998, i.e. late summer. Table 2. Incubation methods and replicates at each sampling period for three parallel trials § j t e Incubation Method Incubation Period (days) (no. of replicates) 14 33 62 92 145 209 271 323 391 Greenhouse Microcosms 6 6 6 6 6 6 6 6 6 Coastal Litter bags 8 8 - - 8 8 8 Interior Litter bags - 8 8 - 4 4 8 8 26 3.4.1 Greenhouse The greenhouse trial was conducted using microcosms, which consisted of 250 mL clear, plastic tubs. The tub bottoms were cut out and replaced with a nylon 'noseeum-mesh' screen (0.25 mm mesh size) to facilitate free drainage of irrigation water. The tubs were covered with a thin "produce bag" polyethylene film to limit moisture loss while allowing some aeration. They were laid out adjacent to each other in seedling trays lined with coarse industrial sand, to facilitate drainage. Organic residual type and sampling period were completely randomized within the trays. Replicated samples of each material were placed in the microcosms. Initial sample weights of each material varied (Table 3). They were selected to approximate typical operational application rates prevalent under operational field conditions. This was between 15 and 30 dry tonnes per hectare in the case of biosolids. Samples of the other materials were of approximately equivalent volume to the biosolids, resulting in considerably smaller sample weights, because their bulk densities were much lower. Table 3. Ranges of sample dry weight (g) Organic Residual Greenhouse Field Minimum Maximum Minimum Maximum Annacis 14.13 16.94 10.04 18.49 Chilliwack 12.06 14.18 7.58 12.82 Lionsgate 11.76 14.96 10.39 21.55 Whistler 13.68 16.96 13.45 18.47 Paper fines 5.07 7.21 5.89 8.02 Wheat straw 1.11 5.24 1.83 5.48 Douglas-fir 2.87 4.50 2.70 5.45 27 The microcosms were inoculated once prior to incubation with a water extraction of a forest floor. This was to introduce a microbial decomposer community similar to that in the field. The forest floor was from a Douglas-fir (Pseudotsuga menziesii), western hemlock (Tsuga heterophylla) and western redcedar (Thujaplicata) forest in the Coastal Western Hemlock biogeoclimatic zone in Pacific Spirit Park, Vancouver B.C. The suspension consisted of one part forest floor to five parts water. It was filtered through coarse filter paper. Ten millilitres of the extract were added to each sample. The microcosms were irrigated as required (approximately weekly), with a handheld watering can, to maintain sufficient moisture for decomposition to proceed, while avoiding water-logging. The greenhouse temperature ranged between an absolute minimum of 15 °C and an absolute maximum of 41 °C. Daily minimums were between 15 °C and 20 °C all year round, while most daily maximums were in the 20s in winter, and the low 30s in the summer. 3.4.2 Field Sites For the field incubations, samples were placed into 12 cm by 14 cm litter bags, made from nylon 'noseeum-mesh' (0.25 mm mesh size). Bags were pinned to the surface of the forest floor with geo-textile pins to simulate surface application of biosolids (typical of forest applications). Bags were laid out in a one by one metre grid. Residual type and sampling period were completely randomized within the grid at each site. The field sites consisted of two different forests. The first forest was a Douglas-fir (Pseudotsuga menziesii) dominated stand in the Interior Douglas-fir (IDF) biogeoclimatic zone 28 in the Mud Lake Research Forest, approximately 20 km northwest of Kamloops, B.C. Mean annual precipitation measured at Red Lake (15km to the west of the site and at a similar altitude) between 1974 and 1992 was 500 mm. Mean summer rainfall (July to September) totalled 147mm. Mean annual snowfall was 2.1 metres. Mean daily temperature recorded at Kamloops between 1951 and 1990 was -4.8 °C in January and 20.8 °C in July. The study site was approximately 800 m higher in elevation (c. 1150 m a.s.l.) than the climate station at Kamloops, so temperatures at the site were probably an estimated 4.8 °C cooler (based on an estimate of 0.6 °C cooler per 100 m elevation higher). The soils of the area were classified as brunisolic gray luvisols. The site was well drained. The stand had been high-graded and insufficiently regenerated about 60 years ago. Some regeneration had occurred, but the canopy remained fairly open. The understorey consisted of grasses and a sparse layer of low shrubs. The forest floor was not continuous, but where it occurred it consisted of a thin mor humus layer (0-2 cm) The other forest was a stand of Douglas-fir planted in 1977 after the clearfelling of the original Douglas-fir, western hemlock, western redcedar stand. It was situated in the moist submaritime Coastal Western Hemlock (CWHsm) biogeoclimatic zone in the Whistler Demonstration Forest, about 7km south of Whistler, B.C. It was a more mesic site than the interior forest, and somewhat milder than the interior forest in the winter. Mean annual precipitation recorded at Whistler township between 1965 and 1991 was 1291mm. Mean summer rainfall (July to September) totalled approximately 200 mm. Mean annual snowfall was 5.1 metres. Mean daily temperatures recorded at Whistler in the same period were -3.5 °C in January and 15.6 °C in July. Despite the relatively high precipitation the area suffered a soil moisture deficit during the summer (Zhong, 1998). The soils of the area are classified as duric 29 humo-ferric podzols. The site was well drained. The stand was thinned and pruned to 3 metres in 1991. There was a dense shrub layer dominated by Vaccinium spp. (V. alaskaense and V. ovalifolium) about one metre high. The forest floor was not continuous and had been disturbed. Where it occurred it was shallow to moderate depth mor humus (0 to 5 cm) 3.5 POST-INCUBATION ANALYSES At the end of each sampling period replicates of each material were collected, dried and weighed. Samples from selected harvests were analyzed for organic matter content, carbon content and organic N content, as per the methods described in 3.3 PRE-INCUBATION A N A L Y S E S . In the greenhouse the selected harvests were 0, 14, 145 and 371 days. In the field the selected harvests were 0 and 371 days. 3.6 MODELLING Models were based on the data from the greenhouse trial. The data from the two field trials were used as independent data sets for field validation. Dependent and independent variables modelled in the analysis are described in Table 4. Simple, two-term linear models (including day and each independent variable) were used to assess the whole group of independent variables. The most effective variables from this group were then used in a more comprehensive multiple linear modelling exercise. 30 Table 4. Dependent and independent variables tested in the multiple linear models Model Dependent variable Independent variables applied to all four models N,otai (%dry weight) Decomposition Organic matter loss (% initial mass) Norganic (%dry weight) carbon (% dry weight) organic matter (% dry weight) C:N ratio (total N) C:N ratio (organic N) NetN mineralization (all residuals) Organic N loss (g N lost / kg O M applied) C:OM extractable fraction acid-soluble fraction acid-insoluble fraction LCI lignin:N (total N) NetN mineralization (biosolids only) Organic N loss (g N lost / kg O M applied) lignin:N (organic N) alkyl index methoxyl index O-alkyl index phenolic index NetN mineralization (2ndary biosolids only) Organic N loss (g N lost / kg O M applied) carboxyl index alkyl:O-alkyl ratio phenolic:N (total N) phenolic N (organic N) Multiple linear models were constructed using the SAS™ general linear model procedure (proc glm). The assumption of normality was tested using the SAS™ procedure 'proc univariate' to test for skewness of the data. The assumption of homoscedasticity was tested by plotting model residuals against the predicted value of the dependent variable, using the SAS™ procedure 'proc plot'. Lack of fit was also assessed using the plot of residuals against the predicted values of the dependent variable. 31 Each of the preferred multiple linear models was selected on the basis of a maximum or near maximum r2 with as few independent variables as possible. 3.7 FIELD VALIDATION Data from the two field trials were used to test the models constructed with the greenhouse data set. The greenhouse models were assessed for their ability to describe the variation in the field data from both forest sites, using the coefficient of determination (I ). This was defined as: 2 (vi - yhat)2 I2= 1 -2 (yi - ybar)2 where, yi is the observed value in the field yhat is the value predicted by the greenhouse model ybar is the mean of the observed field values The greenhouse models were also assessed for bias, by comparing predicted values with observed values at both field sites. Bias was determined for the models as a whole and for the individual organic residuals. It was defined as: 2 (y; - yhat) Bias = 32 where, yi is the observed value in the field yhat is the value predicted by the greenhouse model n is the number of observations A site-specific correction factor was applied to the greenhouse models to reduce the overall bias of the model to zero at each field site. This was an attempt to quantify the effect of field site conditions on rates of decomposition and net N mineralization, and thus facilitate the transfer of the models into the field. I 2 of the adjusted models was recalculated as a corrected I 2. The r 2 values of the best-fit models based on the field data from each site were used as a 2 2 benchmark against which to compare the corrected I values. A concurrence ofther z (field) and the corrected I 2 for that site implied the variables employed in the greenhouse model were relevant in the field, and that the model was satisfactorily adjusted for the site. 33 4. R E S U L T S 4.1 CHEMICAL CHARACTERIZATION OF RESIDUALS The biosolids were characterized by high N content and relatively low carbon and organic matter contents (Table 5). As a result they had lower C:N ratios and lower lignin:N ratios than the other materials. They had remarkably uniform carbon to organic matter (C:OM) ratios. They had relatively high extractable fractions and low acid-insoluble fractions, indicating substantial non-structural organic matter. They had lower LCIs than the plant fibre materials due to their relatively low acid-insoluble fraction. The N M R spectra of the biosolids (Table 6) were characterized by strong methoxyl and carboxyl signals which, in view of the conspicuous absence of a methoxyl peak (55ppm) in the DD spectra and the high organic N content of the biosolids, was indicative of substantial quantities of protein in these materials. They were also characterized by relatively high alkyl indices, indicating significant presence of waxes and fatty acids, and generally low O-alkyl indices reflecting their low content of structural organic matter (polysaccharides). 34 CO OJ o c > c/3 <D o c o • i-H N •c o S3 j ) o "3 o s o "e3 lignin : N o r g a n i c lignin : N t 0 , a i LCI acid insoluble (Klason lignin) % acid soluble % extractable % C : O M C • No rg a njC C : N t o t a i Organic Matter % C % o^rganic % Total N % CO 3 T 3 o 2.47 2.36 3.91 4.82 29.52 7.13 16.70 2.32 1.88 3.66 4.43 29.33 7.02 16.34 \o r~ ro •5 m ro o in ro >n o © O d d d d d 9.45 7.46 11.49 12.08 29.62 6.53 4.09 SO CN oo o CN o in r~ CN SO CN CN d ro ro CN oo OS CN SO in o OS ro OS m Os in SO in r- OS m d in oo ro ro m rO "0. m in in in m d d d d d d d OS o CN ro CN in o m o 00 00 oo CN CN m OS so r~ 4-<D OJ) T3 <+-. o X CO •<t >n CN O ro CN o SO . oo SO ro so d SO in SO CN SO Os sd OS t~ OS •*— o T3 u ca o 00 oo OS p OS CN O in ro o ro ro ro in ro ro m CN m in •t 4— CO xi X) u T3 T3 CN 00 r- OS in © O CN Os CN ro . ro CN CN d d +~ co eo X> o •a r-o 00 Os ro f- p ro OS <n CN ro CO CN d d o CO U •a c CO CO 00 c o I "ob 3 o Q CO i s <L> C (L) a. co (X Table 6 Solid-state C nuclear magnetic resonance (NMR) spectroscopy, CPMAS analysis of the seven organic residuals. Indices and ratios of spectral regions 3 CD T3 -a =r CD ' O £ a o fa & hen 3 o_ Organic residual 3 o -< o_ o' alkyl ethoxyl 1 peak -jectrum O-alkyl henolic arboxyl O-alkyl o ' z o sL z O (fo Pi 3 o' Chilliwack 0.264 0.091 no 0.309 0.041 0.086 0.85 0.010 0.011 Whistler 0.194 0.078 no 0.330 0.058 0.078 0.59 0.015 0.018 Annacis Island 0.194 0.079 no 0.370 0.046 0.074 0.53 0.015 0.016 Lionsgate 0.190 0.055 yes 0.550 0.050 0.060 0.35 0.018 0.020 Douglas-fir 0.196 0.057 yes 0.381 0.062 0.072 0.51 0.061 0.062 Wheat straw 0.093 0.050 yes 0.564 0.036 0.021 0.17 0.039 0.039 Paper fines 0.050 0.038 yes 0.662 0.031 0.019 0.08 0.124 0.129 4.1.1 Chilliwack Biosolids Chilliwack biosolids had the highest N content of the biosolids and of all the materials and therefore the lowest C:N and lignin:N ratios. It had the highest methoxyl and carboxyl indices of all materials. The strength of these indices, the high N content and the conspicuous absence of a methoxyl peak on the DD spectrum suggested high levels of protein (Dr C. Preston, pers. comm., 1999). It probably had the highest protein content of all materials. It also had the highest alkyl index of all materials. It had the lowest acid-soluble fraction and a low acid-insoluble fraction suggesting its organic matter was largely non-structural. This was borne out by the low O-alkyl index, the lowest of all materials, which was indicative of a low polysaccharide content. 36 4.1.2 Whistler Biosolids Whistler biosolids had the second highest total N and organic N contents after Chilliwack biosolids, the second highest carboxyl index, a high methoxyl index and the second lowest C:N ratios. Owing to the high methoxyl index and the lack of a methoxyl peak in the DD spectrum, it was probably second in terms of protein content. It had the highest extractable fraction of all materials, a low acid-soluble fraction and a low acid-insoluble, suggesting its organic matter was also largely non-structural. Whistler biosolids had a remarkably high phenolic index, the second highest of all materials which, due to its low acid-soluble fraction and acid-insoluble fraction, was probably mostly non-structural i.e. tannins and smaller polyphenolics rather than lignin. Whistler biosolids had the second highest alkyl index after Chilliwack biosolids, indicating a relatively high content of lipids and fatty acids. 4.1.3 Annacis Island Biosolids Annacis Island biosolids had the third highest N content of all materials, similar to Whistler in terms of organic N but considerably less in terms of total N . It had the third highest carboxyl index and a methoxyl index equal to or slightly higher than Whistler biosolids. Owing to the slight methoxyl peak in the DD spectrum and the relatively high acid-insoluble fraction, there was probably some contribution of lignin in the methoxyl region, in addition to strong protein presence, and therefore Annacis Island was probably third in terms of protein content. It had the third lowest C:N ratios. Annacis Island biosolids was equal second (with Whistler biosolids) in terms of alkyl index, indicating a substantial presence of lipids and fatty acids. 37 4.1.4 Lionsgate Biosolids Lionsgate biosolids had the lowest total and organic N content among the biosolids. Its carboxyl and methoxyl indices were substantially lower than those of the other three biosolids. Furthermore it had a conspicuous methoxyl peak in the DD spectrum, indicating a lignin contribution to the signal in the methoxyl region. This suggests Lionsgate biosolids contained substantially less protein than the other three biosolids. It had the highest C:N ratios among the biosolids. It had the lowest extractable fraction of the biosolids. It had a substantially greater acid-soluble fraction and a substantially higher O-alkyl index than the other biosolids, indicating a relatively large polysaccharide content. Its organic matter was therefore probably distinctly more structural than the other biosolids. It had the lowest alkyl index among the biosolids, suggesting a fairly modest content of lipids and fatty acids. Lionsgate biosolids had one of the higher phenolic indices, third of all materials after Douglas-fir litter and Whistler biosolids. Its relatively high acid-insoluble fraction among the biosolids and the conspicuous methoxyl peak in the DD spectrum, suggest the phenolics of Lionsgate biosolids were dominated more by lignin than were those of Whistler biosolids. 4.1.5 Douglas-fir Foliar Litter Douglas-fir foliar litter was characterized by a high acid-insoluble fraction, a high phenolic index and a high carbon content. Its acid-insoluble fraction was substantially greater than any of the other materials. Consequently it had the highest lignin:N ratios and the highest LCI of all materials. It had the highest phenolic index of all materials. Owing to its high acid-insoluble fraction and the conspicuous methoxyl peak in the DD spectrum, most of this was probably lignin. However it also had a very conspicuous double peak at 145ppm and 155ppm in the phenolic region, which is characteristic of tannins (Preston et al., 1997), and is therefore 38 indicative of a substantial tannin content. It had a high organic matter content, but also the highest carbon content of all materials. This resulted in the highest C :OM of all materials, slightly higher than that of the biosolids. It had a moderate N content which resulted in a mid-range C:N ratio. It had a similar alkyl index to the biosolids, probably dominated more by waxes and cutins than fatty acids, owing to its plant foliar origins. It had a similar carboxyl index to the biosolids which, considering its high lignin content and modest N content was probably due more to presence of cutins (Preston et al, 1997) or quinone carboxyls than the protein amides and fatty acids dominant in biosolids (Preston, pers. comm., 1999). 4.1.6 Wheat Straw Wheat straw was characterized by a high acid-soluble fraction, a low acid-insoluble fraction and a high extractable fraction. It had a moderate carbon content yet a comparatively high organic matter content resulting in the second highest C :OM of all materials. It had the second highest acid-soluble fraction and the second strongest O-alkyl signal, indicating a high polysaccharide content, and therefore a highly structural organic matter. It had the second lowest acid-insoluble fraction and as result it had the second lowest LCI. It had a low phenolic index and a relatively low methoxyl index indicating a low lignin content. It had a moderate N content resulting in a mid-range C:N ratio. It had a relatively low alkyl index indicating few waxes, cutins or fatty acids. 4.1.7 Paper Fines This material was almost pure cellulose. It had the highest acid-soluble fraction, the highest O-alkyl index and the lowest extractable fraction of all materials, indicating a strongly 39 structured organic matter dominated by polysaccharides. It had the lowest acid-insoluble fraction of all materials indicating little or no lignin. As a result it had the lowest LCI. It had a very low N content resulting in the lowest C:N ratio of all materials. It had the lowest alkyl index and the lowest carboxyl index indicating little or no waxes, cutins or proteins. 4.2 DESCRIBING N MINERALIZATION 4.2.1 Decomposition 4.2.1.1 Greenhouse A l l the organic residuals, including the various biosolids, exhibited a curvilinear pattern of decomposition over time, in the greenhouse. This was consistent whether decomposition was expressed as organic matter loss (Figure 1), or mass loss (Figure 2). At the final sampling (391 days) there was insufficient sample for more than one organic matter analysis on the paper fines. Therefore the organic matter content of that one sample was extrapolated to all replicates to estimate the mass of organic matter lost over time. 40 —• • • Annacis Island Incubation Period (days) Figure 1. Organic matter loss (%) during a 391-day incubation in the greenhouse 1 —• • • Annacis Island Incubation Period (days) Figure 2. Weight loss (%) during a 391-day incubation in the greenhouse Decomposition was fastest in paper fines, followed by wheat straw, followed by the four biosolids and the Douglas-fir litter. The rate of decomposition of the four biosolids was relatively homogenous, (Figure 3) and generally slightly higher than the Douglas-fir litter. When expressed as % weight loss, decomposition rates of the biosolids were lower than the other materials, and became indistinguishable from the Douglas-fir litter (Figure 4). Incubation Period (days) Figure 3. Organic matter loss (%) of biosolids and Douglas-fir litter during a 391-day incubation in the greenhouse 42 50 100 150 200 250 300 350 400 Incubation Period (days) Figure 4. Weight loss (%) of biosolids and Douglas-fir litter during a 391-day incubation in the greenhouse 4.2.1.2 Field Initially, decomposition was significantly (p<0.05) slower in the coastal forest than the greenhouse, and significantly (p<0.05) slower in the interior forest than the coastal forest, for all materials (Figures 5a to 5g). Later, the decomposition rate of the more recalcitrant materials (the biosolids and Douglas-fir litter) in the greenhouse became indistinguishable from their decomposition rate in the coastal forest, though they both maintained a significantly (p<0.05) faster rate of decomposition than that in the interior forest. The rate of decomposition of the more labile materials, paper fines (Figure 5e) and wheat straw (Figure 5f) remained significantly (p<0.05) higher in the greenhouse than at both sites for the duration of the incubation. 43 Figure 5. Weight loss (%) of residuals during 391-day incubations in the greenhouse, interior forest and coastal forest greenhouse • o • interior forest - o - coastal forest Incubation Period (days) a: Annacis Island biosolids b: Chilliwack biosolids 44 • Greenhouse O Interior forest — O • • Coastal forest 400 Incubation Period (days) Lionsgate biosolids Greenhouse Interior forest — o - Coastal forest Incubation Period (days) Whistler biosolids Wheat straw 100 j 90 -Greenhouse • o - Interior forest - o - Coastal forest 0 50 100 150 200 250 300 350 400 Incubation Period (days) g: Douglas-fir foliar litter Despite adverse climatic conditions in the field, the influence of substrate chemistry remained strong even there. The organic residuals ranked in the same order in terms of decomposition rate in both the coastal forest (Figure 6) and the interior forest (Figure 7) as in the greenhouse (Figure 2). i 1 1 1 1 1 1 1 1 0 50 100 150 200 250 300 350 400 Incubation Period (days) Figure 6. Weight loss (%) of each residual during a 391-day incubation in a coastal forest 47 Annacis Island Lionsgate Douglas-fir Paper fines Wheat straw Chilliwack Whistler T 1 1 1 1 r 0 50 100 150 200 250 300 350 400 Incubation Period (days) Figure 7. Weight loss (%) of each residual during a 391-day incubation in an interior forest. 4.2.2 Net N Mineralization Net N mineralization loss from the seven organic residuals during one year, expressed as percentage of initial organic N lost is summarized in Table 7. However, for purposes of comparison of the different organic residuals, net N mineralization for each material was expressed as grams of organic N lost per kilogram of organic matter applied. This unit facilitated the direct comparison of the organic residuals in terms of their release of N , irrespective of their initial N content or their ash content. 48 Table 7 Net N mineralization (% of initial organic N lost) after incubating for 391 days in the greenhouse and at two field sites. Greenhouse Coastal forest Interior forest Mean Mean Mean Organic residual (st. dev.) (st. dev.) (st. dev.) Chilliwack 46.9 aT 31.4 aT 30.3 aT (13.2) (16.0) (15.8) Annacis Island 21.9 a 16.7 a 25.1 a (8.1) (13.4) (23.7) Lionsgate 12.8 a 33.4 a* 15.4 a (10.2) (16.0) (20.6) Whistler 15.0 a 6.4 b 4.0 b (5.8) (19.2) (8.8) Wheat straw 8.6 b 40.0 a 31.9 a (12.4) (16.7) (17.5) Paper fines -138.0 a -46.6 a -88.0 a (70.9) (46.3) (46.6) Douglas-fir -38.7 a -19.2 a -27.3 a (59.4) (19.4) (22.2) t Within a row, means with the same letter are not significantly different (p<0.05) { Significant at p=0.1, but not at p=0.05 4.2.2.1 Greenhouse In the greenhouse, Chilliwack biosolids exhibited the highest net N mineralization by year's end (Table 8). The other three biosolids mineralized significantly less N than Chilliwack biosolids but were not significantly different from one another. The materials of lower N content mineralized significantly less N than all the biosolids (with the exception of wheat straw and Lionsgate biosolids). Wheat straw mineralized a small amount of N , while Douglas-fir litter and paper fines immobilized substantial amounts of N . Table 8. Net N mineralization (g N lost / kg O M applied) during a 391 -day incubation in the greenhouse Organic residual Mean Chilliwack 28.20 at Annacis Island 9.89 b Whistler 7.86 b Lionsgate 5.17 be Wheat straw 0.82 cd Paper fines -3.50 d Douglas-fir -4.10 d t means with the same letter are not significantly different at p<0.05 The patterns of net N mineralization over time were quite distinct among the organic residuals, and even varied somewhat among the biosolids (Figure 8). Paper fines immobilized N until about day 145, followed by gradual net mineralization. Wheat straw exhibited neither much net immobilization nor net mineralization. Douglas-fir litter exhibited constant immobilization of N over the year. Two of the biosolids (Lionsgate and Whistler) exhibited rapid and large net immobilization of N within the first two weeks of the incubation. This was followed by net mineralization over the remainder of the year. Annacis Island exhibited modest immobilization until about day 145, and then gradually mineralized N . Chilliwack biosolids exhibited a small temporary immobilization in the first two weeks of the incubation, followed by net mineralization for the remainder of the year. 50 40 -30 1 , , , , , , , 1 0 50 100 150 200 250 300 350 400 Incubation Period (days) Figure 8. Net N mineralization (g N lost / kg O M applied) of each residual during a 391-day incubation in the greenhouse There was only sufficient sample for one simultaneous analysis of organic matter and N , of paper fines and wheat straw, at the 391-day sampling period. Therefore the organic matter content of that one sample was extrapolated to all replicates to estimate the net N mineralization. 4.2.2.2 Field Net N mineralization rates in the field were estimated from loss of organic N during one year. There was considerable variation in net N mineralization rates in the both the field and the greenhouse. This made differences among the three sites difficult to detect. Indeed, when substrates were grouped together, there was no significant effect of site on net N mineralization. When substrates were examined separately, only Whistler biosolids and the wheat straw exhibited significantly different net N mineralization in the field compared with the greenhouse (Table 9). 51 Table 9. Difference in net N mineralization (g N lost / kg OM applied) among three sites (greenhouse, interior forest and coastal forest), during 391-day incubations of seven organic residuals. Organic residual Greenhouse Coastal forest Interior forest Mean (st. dev.) Mean (st. dev.) Mean (st. dev.) Chilliwack 28.20 aT 18.87 a 18.21 a (7.97) (9.64) (9.53) Annacis Island 9.89 a f 5.41 a 11.36a (3.68) (7.55) (10.7) Lionsgate 5.17 a f 13.47 a* 6.20 a (4.11) (6.44) (8.32) Whistler 7.86 a f 3.37 b 2.09 b (3.05) (10.07) (4.59) Wheat straw 0.82 b f 3.29 a 3.02 a (1.17) (1.12) (1.66) Paper fines -3.50 a f -1.19a -2.21 a (1.78) (1.30) (1.17) Douglas-fir -4.10 a f -2.04 a -2.90 a (6.30) (2.06) (2.35) All substrates 5.97 a f 6.34 a 5.31 a (11.16) (9.57) (9.37) t Within a row, means with the same letter are not significantly different (p<0.05) t Significant at p=0.1, but not at p=0.05 Notwithstanding the weakly expressed differences between greenhouse and field, there were some differences worth noting, as well as some common trends. 52 As in the greenhouse, Chilliwack biosolids consistently mineralized the most N of all the materials at both field sites, (though in the interior forest Annacis Island biosolids mineralized nearly as much, and in the coastal forest Lionsgate biosolids mineralized nearly as much), (Table 10). Whistler biosolids mineralized significantly less N in the field. Mineralization of N from wheat straw was significantly greater at both field sites, than in the greenhouse. Paper fines and Douglas-fir litter appeared to immobilize less N in the field than in the greenhouse, though this was not statistically significant (Table 9). Table 10 Difference in net N mineralization (g N lost / kg O M applied) among seven organic residuals during 391-day incubations at three sites (greenhouse, interior forest and coastal forest) Organic residual Greenhouse Coastal forest Interior forest All Sites Chilliwack 28.20 af 18.87 af 18.21 af 21.38 a Annacis Island 9.89 b 5.41 be 11.36 ab 8.89 b Lionsgate 5.17 be 13.47 ab 6.20 be 8.28 b Whistler 7.86 b 3.37 c 2.09 cd 4.01 c Wheat straw 0.82 cd 3.29 c 3.02 bed 2.52 c Paper fines -3.50 d -1.19 c -2.21 cd -2.35 d Douglas-fir -4.10 d -2.04 c -2.90 d -3.08 d f Within a column, means with the same letter are not significantly different (p<0.05) 53 4.3 MODELLING N MINERALIZATION Models of decomposition and net N mineralization were based on the greenhouse data, as this was the more comprehensive data set. Therefore they predicted decomposition and net mineralization under near optimum environmental conditions, and were not subject to the climatic variations that prevailed in the field. The field data sets were kept as independent data sets with which to assess the models developed on the greenhouse data (see 4.4 FIELD VALIDATION) . 4.3.1 Decomposition Each independent variable was tested with time (day) in a two-term linear model of the rate of decomposition expressed as loss of organic matter (Table 11). The rate of decomposition over time was most highly correlated with the C :OM ratio. Other variables which were strongly correlated included LCI, alkyl index and carboxyl index. The correlation of decomposition with the C:N ratios (total N and organic N) were reasonable but unremarkable. The lignin:N ratios were poor predictors of decomposition rate. 54 Table 11 Correlation of independent variables over time with decomposition (% mass loss of organic matter) of organic residuals during a 391-day greenhouse incubation. T , , ^ • , , coefficient of Independent variables , A . , , ,x determination (all models include day) ^ 2^  day 0.42 Ntotal% 0.54 o^rganic % 0.56 c% 0.44 organic matter % 0.54 C:Ntotai 0.62 C .Ngrganic 0.63 C : O M 0.73 extractable % 0.55 acid-soluble % 0.65 acid-insoluble % 0.56 LCI 0.72 lignin : Ntotai 0.42 lignin: N o r g a r iic 0.42 alkyl 0.68 methoxyl 0.56 O-alkyl 0.61 phenolic 0.62 carboxyl 0.68 alkyl: O-alkyl 0.61 phenolic : Ntotai 0.58 phenolic : Norganic 0.59 55 The multiple linear model for decomposition expressed as % loss of organic matter was selected on the basis of a near maximum r 2 with as few independent variables as possible. The best multiple linear model was based on time and the C :OM ratio. It was: y = -12.7823 + 2.2528x> - 0.0037x2+ 31.2713x3- 3.9633x4+ 0.0067x5 (r2= 0.94) where, y = organic matter loss (%) over period xi xi = number of days X 2 = (number of days) x 3 = C : O M ratio X 4 = ( C O M ratio)*(number of days) x$ = (C:OM ratio)*(number of days) This model employing the C :OM ratio had an equal or higher coefficient of determination (r2) than similar multiple linear models employing LCI (r2=0.93), alkyl index (r2=0.88) and carboxyl index (r2=0.88). The model of decomposition did not satisfy the assumption of normality. Transformation of the data using natural logs and arcsine did not make the distribution of the data more normal. Most of the outliers reflected real variability and could not be justifiably removed. It was decided to proceed with the regression for two reasons. First the plot of regression residuals against predicted values of the dependent variable suggested the data were nearly normal. The distribution of the residuals about the predicted values was reasonably even over the whole range. Second, the models were being constructed to predict treatment 56 means, which is generally more forgiving of non-normal data than predictions such as minimums and maximums (Dr A. Kozak, pers. comm., 1998). 4.3.2 Net N Mineralization The variability of net N mineralization rate within an organic residual was much wider than the variability of decomposition rate. Therefore the correlations were not as strong and the models were not as definitive. Three models were developed for net N mineralization. The first was a model based on all seven organic residuals. The second was a model based on the four biosolids. The third was a model based on the three secondary or partially secondary biosolids (Chilliwack, Whistler and Annacis Island). All three models satisfied the assumptions of normality and heteroscedasticity. 4.3.2.1 All Residuals Each independent variable was tested with day in a two-term linear model of the net N mineralization, expressed as grams of organic N lost per kilogram of organic matter applied (Table 12). 57 Table 12 Correlation of independent variables and time with net N mineralization (g N lost / kg O M applied) from organic residuals during a 391-day greenh incubation. Independent variables (all models include 'day') coefficient of determination (i2) day 0.25 N t o t a , % 0.33 N o r g a n i c % 0.35 C% 0.29 organic matter % 0.29 C:N t o t a , 0.30 CiN o rg a njC 0.29 C : O M matter 0.26 extractable % 0.30 acid-soluble % 0.30 acid-insoluble % 0.25 LCI 0.28 lignin : Ntotai 0.30 lignin: N o r g a n i c 0.31 alkyl 0.34 methoxyl 0.37 O-alkyl 0.31 phenolic 0.31 carboxyl 0.30 alkyl: O-alkyl 0.37 phenolic : Ntotai 0.31 phenolic : Norganic 0.31 The rate of net N mineralization over time was most highly correlated with the alkyl to O-alkyl ratio or the methoxyl index. Other good predictors included organic N concentration and the alkyl content. The C:N ratios and the lignin:N ratios were unremarkable predictors. There was a positive correlation between organic N concentration and the alkyl index (Figure 9) and the alkyl to O-alkyl ratio (Figure 10) over the full range of residuals, though Douglas-fir was an outlier in these correlations. There was also a correlation between organic N concentration and the methoxyl index (Figure 11) over the full range of residuals. 0.30 0.25 0.15 0.10 0.05 1 2 3 organic N concentration Figure 9. Correlation between organic N concentration and the alkyl signal in the seven organic residuals 59 1.0 0.8 0.6 • (0 6 0.4 0.2 0.0 • organic N concentration Figure 10. Correlation between organic N concentration and the alkyl to O-alkyl ratio in the seven organic residuals 0.10 i 0.09 -0.08 -X IU ind 0.07 -5? o 0.06 -E 0.05 -0.04 -0.03 -e e 0 1 2 3 4 organic N concentration Figure 11. Correlation between organic N concentration and the methoxyl signal in the seven organic residuals 60 While the lignin:N ratio was a poor predictor of net N mineralization over time, it exhibited a reasonable curvilinear correlation wdth net N mineralized at the end of the incubation (day 391) (Figure 12). 40 30 20 10 -10 -20 e e e e e • 8 oo 8 ° | •* • O 0 o a e g 8 V e e . 10 15 20 lignin:N ratio 25 30 35 Figure 12. Correlation between the lignin:Norganic ratio and net N mineralization (g N lost / kg O M applied) from the seven organic residuals after incubating for 391 days. The best net N mineralization model based on all seven residuals used time, the organic N concentration of the residual, the phenolic index of the residual and the organic matter concentration of the residual. It was : y = -50.0490 + 5.2103xi - 0.0172x2+ 0.0234x3 - 173.0513x4+ 0.5344x5 (r2 = 0.73) where, y = grams of organic N lost per kilogram of organic matter applied in the biosolids xi = initial organic N concentration of the biosolids (%) X 2 = number of days 61 X 3 = (initial organic N concentration (%)) * (number of days) X 4 = phenolic index of the biosolids . X 5 = organic matter concentration (%) of the residual This model predicted N mineralization substantially better than similar multiple linear models employing the alkyl to O-alkyl ratio (r2=0.66), the methoxyl index (r2=0.64) and the alkyl index (r2=0.65). 4.3.2.3 A l l Biosolids Each independent variable was tested with day as a second term in a multiple linear model of the net N mineralization from the four biosolids (Table 13). Net N mineralization from the four biosolids was most highly correlated with the alkyl index. The phenolic:Norganic ratio, the alkyl to O-alkyl ratio, LCI and the initial organic N concentration were also good predictors. The C:N (total N) ratio and the lignin:N ratios were fairly poor predictors of net N mineralization. 62 Table 13 Correlation of independent variables and time with net N mineralization (g N lost / kg O M applied) from four biosolids during a 391-day greenhouse incubation. Independent variables coefficient of (all models include 'day') determination } (r ) day 0.50 N l o t a i % 0.59 Norganic % 0.68 C % 0.51 organic matter % 0.53 C :N t o t a , 0.56 C.N o rg a nj c 0.63 C O M 0.51 extractable % 0.51 acid-soluble % 0.59 acid-insoluble % 0.51 LCI 0.70 lignin : N t o t a i 0.54 lignin: Norganic 0.57 alkyl 0.74 methoxyl 0.63 O-alkyl 0.57 phenolic 0.66 carboxyl 0.63 alkyl : O-alkyl 0.70 phenolic : Ntotai 0.68 phenolic : Norganic 0.72 There was a correlation between organic N concentration and the alkyl index (Figure 9) and the LCI (Figure 13) among the four biosolids. 0.48 0.46 0.44 0.42 0.38 0.34 3 4 organic N concentration Figure 13 Correlation between organic N concentration and lignocellulose index in the four biosolids The selected net N mineralization model based on the four biosolids used time, the initial organic N concentration of the material and the initial phenolic index of the material. It was: where, y = 11.7735 + 1.3808xi - 0.0507x2+ 0.0341x3 - 508.7792x4 (r2 = 0.78) y = grams of organic N lost per kilogram of organic matter applied in the biosolids xi = initial organic N concentration of the biosolids (%) X 2 = number of days x 3 = (initial organic N concentration (%)) * (number of days) X 4 = phenolic index of the biosolids 64 Phenolic index and N concentration as separate variables performed as well as, or perhaps slightly better, than the combined phenolic:N ratio (r =0.76). The model performed equally well or better as a predictor of N mineralization than similar multiple linear models employing the alkyl to O-alkyl ratio (r2=0.78), the alkyl index (r2=0.78) or the LCI (r2=0.73). 4.3.2.2 Secondary Biosolids Each independent variable was tested with day as a second term in a multiple linear model of the net N mineralization from the three secondary or part-secondary biosolids i.e. Chilliwack biosolids, Whistler biosolids and Annacis Island biosolids (Table 14). Net N mineralization from the three secondary biosolids was most highly correlated with the methoxyl index, the alkyl index and the phenolic:N ratios (total N and organic N). Other variables that correlated well included the, LCI, the alkyl to O-alkyl ratio, organic N concentration and the carboxyl content. The C:N (organic N) was reasonable, but unremarkable. Again, the C:N (total N) ratio and the lignin:N ratios were fairly poor predictors of N mineralization. 65 Table 14 Correlation of independent variables and time with net N mineralization (g N lost / kg O M applied) from three secondary or part-secondary biosolids during a 391-day greenhouse incubation. Independent variables coefficient of (all models include 'day') determiiiati<m (r ) day 0.56 Nt0,al % 0.62 Norganic % 0.76 C% 0.57 organic matter % 0.57 C:N t o t a , 0.59 C'Norganic 0.68 C O M 0.56 extractable % 0.57 acid-soluble % 0.65 acid-insoluble % 0.56 LCI 0.78 lignin : Ntotai 0.57 lignin: Norgariic 0.59 alkyl 0.80 methoxyl 0.80 O-alkyl 0.67 phenolic 0.72 carboxyl 0.75 alkyl : O-alkyl 0.77 phenolic : Ntotai 0.77 phenolic : Norganic 0.80 There was a correlation between organic N concentration and the carboxyl content in the three secondary biosolids (Figure 14). 66 0.088 0.086 0.084 0.080 o •e g 0.078 0.076 0.074 0.072 2.8 3.0 3.2 3.4 3.6 organic N concentration 3.8 4.0 Figure 14. Correlation between organic N concentration the carboxyl signal in the three secondary or part-secondary biosolids The selected net N mineralization model based on the three secondary biosolids used time, the initial organic N concentration of the residual and the initial phenolic index of the biosolids. It was: where, y = 0.3990 + 3.6762x, - 0.0722x2 + 0.0404x3 - 436.5899x4 (r2 = 0.84) y = grams of organic N lost per kilogram of organic matter applied in the biosolids X | = initial organic N concentration of the biosolids (%) X 2 = number of days x 3 = (initial organic N concentration (%)) * (number of days) X 4 = phenolic index of the biosolids 67 Again, phenolic index and N concentration as separate variables performed as well as or perhaps slightly better than the combined phenolic:N ratio (r2=0.82). The model predicted N mineralization from the three secondary biosolids as well as or slightly better than similar 2 2 multiple linear models employing the methoxyl index (r =0.82), the alkyl index (r =0.82), LCI (r2=0.81), the alkyl to O-alkyl ratio (r2=0.83) and the carboxyl index (r2=0.84). 4.4 FIELD VALIDATION Field data from both field trials was used to validate the decomposition and net N mineralization models developed with the greenhouse data set. 4.4.1 Decomposition Results of the validation of the decomposition model are summarized in Table 15. 68 Table 15: Validation of the greenhouse-based model of decomposition (organic matter loss) at two field sites. Interior forest Coastal forest I2 -1.41 -0.76 Bias O V E R A L L -19.99 -16.13 Annacis Island -4.48 -0.74 Chilliwack -6.26 -4.75 Lionsgate -7.72 2.15 Whistler -11.48 -5.53 Paper fines -15.49 -14.35 Wheat straw -29.60 -23.68 Douglas-fir -9.03 -0.54 I2 = r 2 of the greenhouse model used with the field data A negative Bias means decomposition is overestimated The low I 2 values indicate a poor ability of the model to describe the variation in the field data at both sites. The negative values are an indication of a strong bias of this model, when employed at both sites. Bias is presented for each of the organic residuals separately and for the model as a whole. A negative value indicates the model has predicted a larger value than that observed in the field. The greenhouse model generally overestimated the rate of decomposition at both sites. The larger bias for some materials (e.g. paper fines and wheat straw) reflects their greater rate of decomposition. Bias was generally greater for the interior forest site, than the coastal forest site. 69 A n iterative process was used to define a correction factor for each site that reduced the overall bias of the greenhouse model to zero on each site (Table 16). Table 16: Corrected greenhouse model of decomposition (% organic matter loss) employed at two field sites. Interior forest Coastal forest correction factor 0.5786 0.6602 corrected I2 0.71 0.78 r 2 (field) 0.79 0.87 Bias O V E R A L L 0.00 0.00 Annacis Island 2.38 0.89 Chilliwack 2.11 -1.57 Lionsgate -4.38 3.06 Whistler -1.56 -0.80 Paper fines 10.64 6.29 Wheat straw -7.63 -6.60 Douglas-fir -1.58 -1.17 corrected I2 = r 2 of the adjusted greenhouse model used with the field data A negative Bias means field decomposition is overestimated r 2 (field) = the best r 2 of a model based on the field data from each site The corrected I 2 values show that, when corrected for site, the greenhouse decomposition model can be extrapolated reasonably well into the field. Corrected I of the adjusted greenhouse model is reasonably close to the r 2 (field), attained from optimized models of the field data at each site. 70 4.4.2 Net N mineralization 4.4.2.1 A l l Residuals Results of the validation of the net N mineralization model based on all residuals are summarized in Table 17. Table 17: Validation of the greenhouse-based model of net N mineralization (g N lost / kg O M applied) from all residuals, at two field sites. Interior forest Coastal forest I2 0.31 0.18 Bias O V E R A L L -1.33 -0.53 Annacis Island -0.81 -4.97 Chilliwack -6.70 -6.51 Lionsgate -0.38 9.30 Whistler 2.34 -7.72 Paper fines 4.22 4.85 Wheat straw 1.07 1.11 Douglas-fir 8.91 0.93 I2 = r 2 of the greenhouse model used with the field data A negative Bias means decomposition is overestimated The I 2 values of the N mineralization model based on all residuals are quite low at both sites, indicating a poor prediction by the model in the field. Bias of the model at the two field sites is mixed. Chilliwack biosolids appeared to be overestimated at both sites. Whistler and Annacis Island biosolids were overestimated at the coastal site. Lionsgate biosolids was substantially underestimated at the coastal site. The paper fines N mineralization was somewhat underestimated at both sites. These biases are not consistent with the biases observed in the decomposition model. 71 The results of the corrected greenhouse model are summarized in Table 18. Table 18: Corrected greenhouse model of net N mineralization (g N lost / kg O M applied) from all residuals, applied to two field sites. Interior forest Coastal forest correction factor 0.7993 0.9540 corrected I2 0.45 0.17 r 2 (field) 0.56 0.57 Bias O V E R A L L 0.00 0.00 Annacis Island 1.64 -4.06 Chilliwack -1.70 -4.89 Lionsgate 3.11 9.79 Whistler -6.70 -7.13 Paper fines 2.93 4.94 Wheat straw 1.46 1.42 Douglas-fir -0.88 0.36 corrected I2 = r 2 of the adjusted greenhouse model used with the field data A negative Bias means decomposition is overestimated r 2 (field) = the best r 2 of a model based on the field data from each site The correction factor improved the model prediction at the interior site somewhat, but not dramatically. The corrected I 2 of the model used at the coastal site remained low, suggesting that adjustment of the greenhouse model did not improve its ability to predict N 2 2 mineralization at this site at all. Corrected I values were still substantially less than the r (field) values of the optimum field models, especially at the coastal site. 72 4.4.2.2 Biosolids Results of the validation of the net N mineralization model based on the four biosolids are summarized in Table 19. Table 19: Validation of the greenhouse-based model of net N mineralization (g N lost / kg O M applied) from the four biosolids, at two field sites. Interior forest Coastal forest I 2 0.02 -0.23 Bias O V E R A L L -3.37 -1.93 Annacis Island -0.28 -4.09 Chilliwack -9.12 -8.46 Lionsgate 2.77 10.04 Whistler 2.34 -5.56 I 2 = r 2 of the greenhouse model used with the field data A negative Bias means decomposition is overestimated 2 The I values of the greenhouse N mineralization model based on the four biosolids were low at both sites, indicating a poor prediction in the field. As with the model based on all residuals, Chilliwack biosolids N mineralization was overestimated at both sites. N mineralization from Whistler and Annacis Island biosolids were overestimated at the coastal site, while the Lionsgate biosolids N mineralization was substantially underestimated on the coast site. The results of the corrected model are summarized in Table 20. 73 Table 20: Corrected greenhouse model of net N mineralization (g N lost / kg O M applied) from four biosolids, applied to two field sites. Interior forest Coastal forest correction factor 0.7375 0.8505 corrected I 0.26 -0.05 r 2 (field) 0.37 0.39 Bias O V E R A L L 0.00 0.00 Annacis Island 2.78 -2.35 Chilliwack -1.95 -4.37 Lionsgate 3.67 10.55 Whistler -4.49 -4.22 corrected I2 = r 2 of the adjusted greenhouse model used with the field data A negative Bias means decomposition is overestimated r 2 (field) = the best r 2 of a model based on the field data from each site The corrected I 2 value was a little higher than the I 2 for the interior site, indicating the adjusted model is a somewhat better predictor of net N mineralization than the original model at this site. The corrected I 2 value of the coast site was lower than the I 2, suggesting that the adjustment of the model did not improve its predictive ability at this site. Again, corrected I values were still substantially less than the r 2 (field) values of the optimum field models, especially at the coastal site. 4.4.2.3 Secondary Biosolids Results of the validation of the net N mineralization model based on all residuals are summarized in Table 21. 74 Table 21: Validation of the greenhouse-based model of net N mineralization (g N lost / kg O M applied) from secondary or partial secondary biosolids, at two field sites. Interior Coastal forest forest -0.07 0.01 O V E R A L L -4.88 -5.73 Annacis Island 2.17 -1.64 Whistler -6.34 -5.06 Chilliwack -10.47 -9.81 I2 = r2 of the greenhouse model used with the field data A negative Bias means decomposition is overestimated The I 2 values of the greenhouse N mineralization model based on the three secondary or part-secondary biosolids were low at both sites, indicating a poor prediction in the field. The model overestimated net N mineralization from Chilliwack biosolids and Whistler biosolids. The results of the corrected greenhouse model are summarized in Table 22. Corrected I 2 values were substantially higher for both field sites, indicating that the correction factors applied to the greenhouse model substantially improved its predictive ability at these sites. The corrected I value for the interior site is somewhat lower than the r (field) of the optimum 2 * • 2 model at the interior site. The corrected I for the coastal site, however, is very similar to the r (field) of that site, suggesting the predictive ability of the adjusted greenhouse model is as good as the optimum model based on the field data at this site. I2 Bias 75 Table 22: Corrected greenhouse model of net N mineralization (g N lost / kg O M applied) from secondary and part-secondary biosolids, applied to two field sites. Interior forest Coastal forest correction factor 0.6838 0.6375 corrected I2 0.28 0.38 r 2 (field) 0.41 0.41 Bias O V E R A L L 0.00 0.00 Annacis Island 5.08 1.69 Whistler -1.40 -2.00 Chilliwack -3.67 0.59 corrected I 2 = r 2 of the adjusted greenhouse model used with the field data A negative Bias means decomposition is overestimated r 2 (field) = the best r 2 of a model based on the field data from each site 76 5. DISCUSSION 5.1 DESCRIBING N MINERALIZATION 5.1.1 Organic Residuals Among the biosolids there was a positive correlation between N concentration, indices of protein (methoxyl index and carboxyl index) and the alkyl index. The highest N content biosolids (Chilliwack) had the highest methoxyl and carboxyl indices and the highest alkyl index. Whistler and Annacis Island biosolids were fairly similar in terms of organic N . They had similar protein indices and similar alkyl indices. Lionsgate biosolids was ranked last with significantly lower organic N content, much lower evidence of protein presence and a lower alkyl index than the other three biosolids. The strong positive relationship between organic N concentration and protein indices in biosolids suggested that most of the organic N in biosolids occurred in a protein pool. This is consistent with Hattori and Mukai (1986) who estimated 40-50% of the organic N in a range of six different sewage sludges was amino acid-N. The levels of organic N and the magnitude of the protein indices appeared to be strongly related to the sewage treatment process of each type of biosolids. Lionsgate biosolids, the only totally primary biosolids, had the lowest organic N and the lowest protein indices. Primary biosolids does not contain the waste-activated sludge (spent digestive microflora) in the biosolids (Vesilind, 1979) that secondary biosolids does and would therefore be expected to 77 have less proteinaceous material. Furthermore Lionsgate biosolids was detained in digestion tanks for the longest period of all biosolids. Detention time has a strong negative influence on the levels of volatile solids (labile organic matter such as protein) in treated sewage (Parkin and Owen, 1986). Annacis Island had the third strongest indices of protein content. It was 30% secondary, 70% primary and was also detained in digestion tanks for an extended period. Whistler had the second highest N content and the second strongest indices of protein, and the shortest detention time of the thermophilic biosolids. Chilliwack biosolids, which had the highest N and highest protein indices not only had a short detention time, but was the only mesophilic biosolids in the study. Digestion temperature has a strong negative influence over the levels of labile organic matter such as proteins in biosolids (Parkin and Owen, 1986). The alkyl index was probably also related to treatment process. The higher alkyl index of Chilliwack biosolids reflected higher lipid and fatty acid content, which are prevalent in biosolids. Mesophilic digestion is less efficient at digesting these relatively recalcitrant products than thermophilic digestion (Parkin and Owen, 1986). 5.1.2 Decomposition The curvilinear decomposition pattern of biosolids supports the assumption that that they decompose similarly to other organic litters, despite their atypical origins. Curvilinear decomposition is approximated by organic materials because of the diminishing carbon source supporting the decomposer community (Berg, 1986; Lerch et al, 1992; Whitmore, 1996). The slightly different results for decomposition of biosolids expressed as organic matter loss and expressed as weight loss relative to the other materials was due to the relatively low 78 organic matter content of biosolids. In most litter studies mass loss is synonymous with decomposition because the vast majority of the mass of litter is organic and subject to catabolism. Biosolids, however, have a high ash content, which is largely microbially inert. Therefore loss of the organic fraction over time is a more precise measure of decomposition than weight loss, when comparing materials of widely varying organic matter content. It is not surprising that decomposition was often slower in the field, particularly in the interior forest, which was probably subject to more adverse environmental conditions for microbial activity, than the coastal forest or the greenhouse. Nevertheless the observation that the rates of decomposition of the different materials were ranked in the same order in the greenhouse as on both the field sites suggests the strong influence of substrate chemistry on decomposition prevailed in the field. The effect of substrate chemistry probably predominates decomposition in all but the most extreme climates (Meentemeyer, 1978; Kochy and Wilson, 1997). After the first month the decomposition rate of biosolids and Douglas-fir in the greenhouse became indistinguishable from their decomposition rate in the coastal forest. This suggests that, after the first month, the decomposer communities in these materials were less able to take advantage of more favourable environmental conditions in the greenhouse. By this time, decomposition was probably limited more by recalcitrant compounds in the substrate than the environmental conditions. 79 5.1.3 Net N Mineralization The net N mineralization rates of biosolids were quite variable as is typical of biosolids N mineralization studies (Hsieh et al, 1981; Garau et al, 1986; Serna and Pomares, 1992). The rates of net N mineralization of a given source of biosolids are only strictly attributable to the material on the day of sampling, and cannot be used to generalize about all materials from that source. The very different patterns of net N mineralization from the different organic residuals (compared with their similar patterns of decomposition) illustrate the added level of complexity of net N mineralization over gross N mineralization i.e. the influence of N immobilization. A l l the biosolids exhibited a period of immobilization prior to net mineralization. This was indicative of a period of increased microbial growth. In Chilliwack biosolids, the initial immobilization period was small and temporary, indicating less change in microbial biomass. The immobilization was probably associated with a small microbial flush following the disturbance associated with sample preparation (Binkley and Hart, 1989), or as the microbial community turned over from anaerobes to aerobes in the aerobic, post-treatment environment. In the cases of Lionsgate and Whistler biosolids N immobilization was immediate and dramatic. Net N immobilization has been observed in some operational field applications of Lionsgate biosolids, causing temporary chlorosis of pastures (M. van Ham, pers. comm., 1998). The remarkably high extractable N content of Whistler biosolids (0.81%, or approximately four times that of Lionsgate biosolids) probably compensates for the initial immobilization of N in this material. Some immobilization due to the turnover of microbial biomass might be 80 expected in these materials as the decomposer community changes from predominantly thermophilic to predominantly mesophilic, in the mesothermic, post-treatment environment. However this large immobilization was not observed in Annacis Island biosolids which was also thermophilically digested. Both Lionsgate and Whistler have high phenolic contents which may exacerbate immobilization of N via condensation of amines and N H / " ions onto aromatic ring structures of phenolic compounds (Handayanto et al, 1997). The significantly lower net N mineralization of the Whistler biosolids in the field was not necessarily related to the lower rates of decomposition observed there. First, the reduced rate of decomposition was not significant in the coastal forest, yet the reduced rate of N mineralization was quite marked. Second, other biosolids that exhibited apparently lower decomposition in the field (Annacis Island and Lionsgate) did not exhibit reduced mineralization of N . Much of the Whistler biosolids' substantial phenolic content may be comprised of tannins and other non-structural polyphenolics. If these compounds are readily extractable (Preston et. al, 1997) it follows that a material like Whistler may have been a weaker immobilizer in the greenhouse, where successive irrigations may have removed more non-structural phenolics, yet a stronger immobilizer in the field, where somewhat drier conditions may have resulted in less leaching of these compounds. Mineralization of N from biosolids was strongly correlated with the initial organic N concentration and the indices of protein. Thus Chilliwack biosolids mineralized significantly more N than the other biosolids in the greenhouse and substantially more (albeit not always significant) in the field. Proteins form a particularly labile fraction of the organic pool (Lerch et al, 1992). If it is accepted that organic N in biosolids is dominated by proteins, as borne out 81 by the strong correlation between organic N concentration and proteins indices (methoxyl index and carboxyl index), then it follows that N mineralization from biosolids was driven by respiration of this labile pool and not by decomposition of the materials as a whole. Hattori and Mukai (1986) found a correlation between mineralization of organic N from biosolids/soil mixtures and crude protein in the biosolids. Hattori (1988) found a correlation between carbon and N mineralization in biosolids/soil mixtures and the level of proteinase activity in the soil. Lerch et al. (1992) found a correlation between N mineralization and the concentrations of low molecular weight amines (assumed to be proteins) in biosolids/soil mixtures. It is possible that the apparent immobilization of N from the surrounding environment was exaggerated in biosolids relative to the other three materials. While the microbial biomass in biosolids immobilized significant amounts of N in the first few weeks (especially Lionsgate and Whistler), the N was not necessarily from the external environment. The biosolids contained significant quantities of inorganic N within their matrix, which might be immediately available for biological or chemical immobilization. Because of the drying of the samples prior to chemical analysis it was impossible to accurately quantify the amount of mineral N that was available in the fresh materials at the start of the incubation. Even air-dried biosolids can lose up to 90% of NH4+ by volatilization as NH 3 ( g ) (M. van Ham, pers. comm., 1997). Nevertheless, using the difference between total N and organic N as a very conservative estimate of mineral N at the start of the incubation of Whistler biosolids, for example, would potentially contribute 0.81%, or 13.4 g N / kg O M , to the external environment, and could reduce net N immobilization by that amount. It is important to quantify the addition of extractable N in a biosolids application, for the purpose of determining its true fertilizer value. 82 The higher mineralization of N from Lionsgate biosolids at the coastal site (significant at p<0.1) was puzzling. It was probably not an effect of increased mass loss as this was not significantly greater in the coastal forest than in the greenhouse. It was not due to outliers as the result was apparent in several samples. It was not an effect of placement of the litter bags as the design was randomized. It was probably not an effect of the litter bag method as no such effect was observed in Lionsgate biosolids at the interior site. It appears to have been a specific interaction between material and site, exacerbating the loss of organic N, only in Lionsgate biosolids and only at the coastal site. The correction factors for the mineralization models were more effective for the coastal site after Lionsgate biosolids was removed from the data set, as demonstrated by the secondary biosolids mineralization model. The large sustained immobilization of N by the Douglas-fir litter may be due to its relatively high phenolic content. It had the highest phenolic index of all materials (a substantial amount of it apparently tannin) and it immobilized the most N, yet exhibited only a modest rate of decomposition. Douglas-fir litter immobilized somewhat less N in the field. This may be explained to some degree by reduced decomposition in the field, and therefore lower N requirements of a less active decomposer community. Decomposition of wheat straw was dominated by fungi rather than bacteria as evidenced by the conspicuous mycelium on wheat straw samples. The fungal hyphae were probably more resistant to desiccation than bacteria (Swift et al, 1979) in this material, which had poorer moisture retention than the other materials. Wheat straw exhibited a relatively modest immobilization of N compared with its rate of decomposition. Fungi immobilize N as chitin in hyphal cell walls, and are generally more efficient utilizers of the nutrient for the 83 amount of biomass they produce (Paul and Clark, 1989). At the same time relatively little N was mineralized from the wheat straw over the year. This may have been because fungal mycelium is somewhat hydrophobic and therefore resistant to leaching, and also probably resistant to decay (Wessen and Berg, 1986). Thus it is probably a relatively persistent immobilizer of N . Wheat straw mineralized significantly more N in the field than in the greenhouse although decomposition was significantly less. It is possible that, in the sub-optimum climatic environment of the field, a smaller fungal community was less effective at immobilizing N released from the straw. In light of the decreased rate of decomposition it is possible this N was simply leached from the wheat straw which had a relatively high extractable fraction. Wessen and Berg (1986) noted leaching as a significant loss pathway of N in decomposing barley straw. The paper fines, which had a very high C:N ratio immobilized substantial amounts of N from the surrounding environment, probably via precipitation from above or percolation from beneath. This suggests the paper fines contained insufficient N to meet the requirements of the expanding microbial biomass. By about day 145 decomposition slowed and immobilization of N occurred to a lesser degree. This suggests the more labile products were preferentially respired and, because the microbial biomass was declining, its N requirements were also declining. In the field, paper fines appeared to immobilize somewhat less N . This was probably because of sub-optimum environmental conditions for the decomposer community, hence the significantly slower decomposition, and presumably lower N requirements. 84 5.2 M O D E L L I N G N MINERALIZATION 5.2 .1 Modelling Decomposition It was not possible to assess the effectiveness of indices of decomposition rate of the biosolids alone, due to their very similar decomposition rates. The effectiveness of the C:OM ratio as an index of decomposition was of interest. It was a measure of the content of organic elements other than carbon (i.e. oxygen and hydrogen), in the material. In the plant materials it probably reflected the size of the pool of labile polysaccharides (cellulose and hemicellulose) which consist of a high proportion of these elements. Intuitively the pool of polysaccharides should have less effect on decomposition of the biosolids, which probably had a large labile protein pool (Lerch et al, 1992), and are in any case relatively low in polysaccharides, as evidence by the generally weaker O-alkyl signals. Nevertheless, the C:OM is remarkably constant among the biosolids and correlated well with their rate of decomposition when they were included in the data set with the other residuals. The laboratory procedures for determining the C:OM ratio are considerably easier and cheaper than those for either the LCI or the alkyl to O-alkyl ratio. If it continues to prove its effectiveness, it may be a useful tool in organic matter decomposition studies. LCI was also an effective index of decomposition rate. In the plant materials, which had higher acid-soluble fractions and strong O-alkyl (polysaccharide) signals, the LCI was probably a good reflection of carbon chemical structure. In the biosolids (particularly the secondary biosolids) the LCI was more a product of a low acid-soluble fraction than a high acid-insoluble fraction. The effect of the lignin cellulose matrix on the decomposition of materials with a large non-structural, extractable fraction would intuitively be less significant. 85 Nevertheless LCI was a reasonable predictor of the decomposition rate of biosolids within the larger data set. The effectiveness of the alkyl index reflected the proportion of more recalcitrant compounds in each of the materials. In the biosolids, it was probably dominated by fatty acids and lipids (Parkin and Owen, 1986). In Douglas-fir liter, the other high alkyl index material, it was probably dominated by cutins (Preston et al, 1997). Both fatty acids and cutins are relatively recalcitrant compared with proteins and carbohydrates. The effectiveness of the carboxyl index probably reflected the relative concentrations of recalcitrant cutins in the some of the materials. The low carboxyl index of paper fines reflected a low content of cutin. Conversely the high carboxyl index of Douglas-fir litter probably reflected a high content of cutin. In the case of biosolids the relationship was questionable. In secondary biosolids in particular, the high carboxyl index probably reflected protein content, which is relatively labile. The unremarkable performance of the C:N ratio is probably due to the very different carbon chemical structure of the materials. For example, wheat straw and Douglas-fir litter had approximately the same C:N ratio, yet the organic matter in wheat straw was dominated by polysaccharides, and was therefore much more labile than the organic matter in Douglas-fir litter, which had a substantial lignin content. 86 The poor performance of the lignin:N ratio was probably due to the wide variety of N contents in the material. For example, the paper fines had the second highest lignin:N ratio (because of its very low N content), yet it was the most labile of all materials. The effectiveness of the day2 variable is indicative of the non-linear relationship between time and decomposition (Figures 3 and 4). The C:OM*day and C:OM* day variables are indicative of an interaction between time and substrate chemistry on decomposition, i.e. decomposition proceeds at different rates for substrates of different relative recalcitrance. 5.2.2 Modelling Net N Mineralization The initial N concentration of an organic residual was the single most important factor in determining net N mineralization in all three mineralization models. In the case of the biosolids it probably reflected the organic N in the labile pool. Hence there was a strong correlation between N concentration, the strength of the carboxyl signal and the strength of the methoxyl signal in the biosolids. The phenolic index, while not a strong predictor of net N mineralization in its own right, made a substantial contribution to the multiple linear model when used in concert with organic N concentration. Although the phenolic:N ratio was the most effective single predictor of N mineralization in each of the biosolids models, the use of N concentration and the phenolic index as separate terms in the multiple linear model explained more of the variation. Organic matter content was only significant in the model based upon all residuals. Whereas N concentration is based on total dry weight, the N concentration of the organic pool 87 was probably more relevant. The inclusion of the organic matter content as a term may have indirectly adjusted for this, and was therefore important when comparing materials of widely varying organic matter concentration. The effectiveness of the alkyl to O-alkyl ratio and the alkyl signal in the mineralization models was probably due to the correlation between these variables and the organic N concentration. As discussed, in the biosolids, the alkyl signal (reflecting lipid content) and the organic N concentration (reflecting protein content) were probably both affected by the type of sewage treatment process at the STP of origin. The effectiveness of the LCI as an index of N mineralization in the two biosolids data sets was probably due to the correlation between LCI and organic N concentration in biosolids. Biosolids with a higher N content tended to have a higher extractable fraction and a lower acid-insoluble fraction, and therefore a higher LCI. The unremarkable performance of the C:N ratio as an index of N mineralization in all three data sets was probably due to the wide variety of carbon chemical structures. In biosolids N mineralization appeared to be driven by dynamics of a distinct labile pool of organic matter (protein) and it was clearly more related to the dynamics of that pool than the total C:N ratio. Among the other materials the C:N ratio appeared to have had some effect on net N mineralization, but it was probably masked by the other factors, e.g. the high phenolic content in Douglas-fir litter, or the nature of the decomposer community in wheat straw. 88 The poor performance of the ligninrN ratio as an index of N mineralization in all three models is probably due to the strong immobilization of N by Lionsgate and Whistler biosolids early in the incubation. However the ratio did exhibit a reasonable correlation with net N mineralization at the end of the incubation (day 391), when the transient immobilization of N in the two biosolids had passed. The lignin:N ratio at this time was a better reflection of degree of N mineralization in the seven residuals. It is noteworthy that the phenolic :N ratio which includes the non-structural phenolics such as tannin and smaller polyphenolics (in addition to lignin) was a much better index of net N mineralization in biosolids which had relatively low Klason lignin contents (acid-insoluble fractions). 5.3 F I E L D V A L I D A T I O N The use of the greenhouse data for model construction resulted in models of decomposition and net N mineralization under near optimum environmental conditions. These models were theoretically free of the adverse environmental effects of a field site, and therefore hopefully provided a stable and favourable environmental background with which to assess the effects of substrate chemistry on decomposition and N mineralization. It was not expected that the materials at the field sites would decompose at the same rates or mineralize/immobilize N to the same degree. Nevertheless it was important to assess the degree to which these models could be extrapolated into the field, to determine their robustness and identify their biases. 5 .3 .1 Decomposition Model The poor performance of the greenhouse decomposition model in the field was presumably due to the effect of sub-optimal environmental conditions in the field. As 89 previously discussed rate of decomposition was generally slower than predicted by the greenhouse-based model. This was borne out by the overall negative bias of the model at both field sites. The correction factors for each site resulted in reasonable corrected I values at both sites, after adjustment, suggesting that the climate and site factors at each location had a fairly consistent effect on all the materials. Therefore the model can be used with some degree of confidence to predict decomposition of these materials at these field sites, and might be extrapolated to other sites in coastal or interior B.C., with similar climates and site conditions. To model decomposition across a wider range it would be necessary to include climate variables in a decomposition model. Meentmeyer (1978) found lignin and actual evapotranspiration (AET) were the most significant terms in a decomposition model for forest litter on a global scale. Prescott et al, (In Review) found degree-days (number of days above a defined temperature) and lignin content to be the best predictors of leaf litter decomposition across B.C. Both AET and degree-days reflect changes in moisture as well as temperature, and are therefore useful general climatic indices. 5.3.2 Net N Mineralization Models The poor performance of the mineralization models at the field sites was probably due to a combination of wide variation in N mineralization, and interactions of site and substrate on N mineralization, such as the Lionsgate biosolids-coastal site interaction. The correction factors for the coastal site did not improve the performance of any of the mineralization models with the exception of the secondary biosolids model from which Lionsgate biosolids was 90 omitted. The I 2 values of all three models at the interior site and of the secondary biosolids model at the coastal site were somewhat improved by the correction factors applied. This suggests that there was some site and climate factors which consistently affected N mineralization from a range of materials in the same way, and that they can be somewhat adjusted for with a correction factor for each site. However, the corrected I values were generally low, reflecting the wide variation of net N mineralization in the field, even within an organic residual. Thus the models cannot be used with great confidence in the field, even with the appropriate correction factor. Nevertheless, the reasonable similarity between the corrected I 2 and the r 2 (field) of the optimum field model, particularly with the secondary biosolids model at the coastal site, suggests the variables employed in the greenhouse were also relevant in the field. Therefore organic N content and the phenolic index are probably a good starting point for further field based modelling of N mineralization from these materials. 91 6. CONCLUSIONS The initial organic N concentration and the phenolic index (as defined using NMR) consistently emerged as either the most effective model or among the most effective models for predicting net N mineralization for all seven residuals, for the biosolids only and for the secondary biosolids only. In the model for all residuals, the organic matter content also emerged as an important term. The transferability of these models into the field was limited by the wide variation in N mineralization, even within a single residual, as well as by some interactions between site and substrate. Nevertheless the variables most important in the greenhouse models maintained their relevance in the field and would be a good starting point for further field modelling of N mineralization from biosolids. There was a strong relationship between organic N content of biosolids and NMR-derived indices of protein, the carboxyl index and the methoxyl index. This supported findings of other studies that the organic N pool in biosolids is predominantly associated with protein, and that therefore the dynamics of the labile protein pool control N mineralization in biosolids. This suggests the NMR-derived protein indices may be of value as quantitative indices of N mineralization in biosolids. There was some evidence that the original treatment process of sewage controlled protein content and organic N content and therefore ultimately the N mineralization potential of the biosolids. 92 There was a weak relationship between net N mineralization and decomposition. While rates of decomposition had some effect on net N mineralization, particularly in the low C:N ratio materials, the correlation was probably clouded by other factors. Among the biosolids net N mineralization was probably more strongly related to mineralization of the labile protein pool than to the rate of decomposition of the materials as a whole. Notwithstanding the apparent lack of importance of decomposition in describing net N mineralization, the study nevertheless found strong correlations between indices of carbon chemical structure and the rate of decomposition in organic residuals. The LCI and the C:OM ratio proved the most effective predictors. Despite relatively low organic matter content and atypical origins, decomposition in biosolids were predictable by the same indices of decomposition as predicted decomposition of more typical organic materials. The unadjusted greenhouse decomposition model overestimated rates of decomposition in the field, but responded well to site-specific correction factors. This suggested the model could be extrapolated with some degree of confidence to similar sites in coast or interior B.C. Extrapolation to other sites would require a model that explicitly accounted for the effects of climate on decomposition. 93 REFERENCES AertsR. (1997). Climate, leaf litter chemistry and leaf litter decomposition in terrestrial ecosystems: a triangular relationship. Oikos 79: 439 - 449. Aerts R. and de Caluwe H. (1997). Nutritional and plant-mediated controls on leaf litter decomposition of Carex species. Ecology 78(1): 244 - 260. Agren, G.I. and Bosatta, E. (1996). Quality: a bridge between theory and experiment in soil organic matter studies. Oikos 76: 522 - 528. Baldock J.A. and Preston C M . (1995). Chemistry of Carbon Decomposition Processes in Forests Revealed by Solid State CI 3 Nuclear Magnetic Resonance Spectroscopy. Chapter 6 in McFee W.W. and Kelly J .M. (1995). Carbon Forms and Functions in Forest Soils. Soil Science Society of America, Madison, Wisconsin. Barbarika A. , Sikora, L.J. , Colacicco D. (1985). Factors affecting the mineralization of nitrogen in sewage sludge applied to soils. Soil Sci. Soc. Am. J. 49: 1403 - 1406. Berg, B. (1986). Nutrient release from litter and humus in coniferous forest soils - a mini review. Scand. J. For. Res 1: 359-369 Binkley D. and Hart S.C, (1989). The Components of Nitrogen Availability Assessments in Forest Soils in Advances in Soil Science, Volume 10, pp. 57-112, Springer-Verlag New York, Inc. Binkley D., (1986). Forest Nutrition Management. Wiley Interscience Publications, New York. Black C.A., Evans D.D., White J.L., Ensminger L.E. and Clark F.E. (1965). Methods of Soil Analysis. Part 2. Agronomy Monograph 9. A S A , Madison, Wisconsin. Bremner J .M. (1996). Nitrogen-total in Sparks D.L, Page, A . L . , Helmke P.A., Loeppert R.H., Soltanpour P.N., Tabatabai, M.A. , Johnston C T . and Sumner M.E. (eds.) (1996). Methods of Soil Analysis, Part 3 - Chemical Methods. Soil Science Society of America, Madison, Wisconsin. Bosatta, E. and Agren, G.I. (1985). Theoretical analysis of decomposition of heterogeneous substrates. Soil Biol. Biochem. 17(5): 601-610. Boyle, M . and Paul E.A. (1989). Carbon and nitrogen mineralization kinetics in soil previously amended with sewage sludge. Soil Sci. Soc. Am. J. 53: 99- 103. 94 Chae Y . M . and Tabatabai M . A . (1986). Mineralization of nitrogen in soils amended with organic wastes. J. Environ Qual. 15(2): 193 - 198. Cogle, A . L . , Saffigna P.G. and Barron P.F. (1989). The use ofnC-NMRfor studies of wheat straw decomposition. Plant and Soil 113: 125 - 128. de Montigny L. and Zumrawi A. (1996). Whistler Biosolids Forest Fertilization Program: second year growth and yield progress report (internal). Ministry of Forests, Victoria, B.C.. Edmonds, R.L. (1980). Litter decomposition and nutrient release in Douglas-fir, red alder, western hemlock, and Pacific silver fir ecosystems in western Washington. Can. J. For. Res. 10: 327-337. Fenn P., Chois and Kirk T.K. (1981). Lignolytic activity of Phanerochaete chrysosporium: Physiology and suppression by NH/and L-glutamate. Arch Microbiol. 130: 66-91. Fox R.H., Myer R.J.K. and Vallis I. (1990). The nitrogen mineralization rate of legume residues in soil as influenced by their polyphenol, lignin and nitrogen contents. Plant and Soil 129: 251-9 Fyles J.W. and Fyles I.H. (1993). Interaction of Douglas-fir with red alder and salal foliage litter during decomposition. Can. J. For. Res. 23: 358 - 361. Garau M.A. , Felipo M.T. and Ruiz de Vil la M.C. (1986). Nitrogen mineralization of sewage sludges in soils. J. Environ. Qual. 15(3): 225 - 228. Garcia C , Hernandez T., Costa F., Ceccanti B. and Dell'Amico C. (1992). Characterization of the organic fraction of an uncomposted and composted sewage sludge by isoelectric focussing and gel filtration. Biol. Fertil. Soils 13:112-118. Gilmour J.T, Clark M.D. and Daniel S.M. (1996). Predicting long-term decomposition of biosolids with a seven-day test. J. Environ. Qual. 25: 766 - 770. Griffin D . M . (1972). Ecology of Soil Fungi. Chapman and Hall, London. Handayanto E., Giller K .E . and Cadisch G. (1997). Regulating N release from legume tree prunings by mixing residues of different quality. Soil Biol. Biochem. 29(9/10): 1417 -1426. Hattori H. and Mukai S. (1986). Decomposition of sewage sludges in soil as affected by their organic matter composition. Soil Sci. Plant Nutr. 32(3): 421 -432. Henry C.L. (1986). Growth response, mortality and foliar nitrogen concentrations of four tree species treated with pulp and paper and municipal sludges. Chapter 23 in Cole D.W., Henry C.L. and Nutter W.L. (1986). The Forest Alternative for Treatment and Utilization of Municipal and Industrial Wastes. University of Washington Press, Seattle, Washington. 95 Hsieh Y.P. , Douglas L .A. and Motto H.L. (1981). Modeling sewage sludge decomposition in soil: II. Nitrogen transformations. J. Environ. Qual: 10(1), 59-64 . Hyvonen R., Agren G.I. and Andren O. (1996). Modelling long-term carbon and nitrogen dynamics in an arable soil receiving organic matter. Ecological Applications 6(4): 1345 - 1354. Inbar Y. , Chen Y. , and Hadar Y . (1989). Solid-state Carbon-13 Nuclear Magnetic Resonance and Infrared Spectroscopy of composted organic matter. Soil Sci. Soc. Am. J. 53: 1695- 1701. Johansson M.-B. (1994). Decomposition rates of Scots pine needle litter related to site properties, litter quality, and climate. Can. J. For. Res. 24: 1771 - 1781. Keyser P. Kirk T.K. and Zeikus J.G. (1978). Lignolytic enzyme system of Phanerochaete chrysosporium: synthesized in the absence of lignin in response to nitrogen starvation. J. Bacteriol. 135: 790-797. Kochy M . and Wilson S.D. (1997). Litter decomposition and nitrogen dynamics in Aspen forest and mixed-grass prairie. Ecology 78(3): 732 - 739. Laboratory Equipment Corporation (1959). Instruction manual for operation of Leco Carbon Analyzers. Saint Joseph, Michigan. Lerch R.N., Barbarick K . A . , Sommers L.E. and Westfall D.G. (1992). Sewage sludge proteins as labile carbon and nitrogen sources. Soil Sci. Soc. Am. J. 56: 1470 - 1476. Lerch R.N., Azari P., Barbarick K . A . , Sommers L.E. and Westfall D.G. (1993). Sewage sludge proteins: II. Extract characterization. J. Environ. Qual. 22: 625 - 629. Lindemann, W.C. and Cardenas M . (1984). Nitrogen mineralization potential and nitrogen transformations of sludge-amended soil. Soil Sci. Soc. Am. J. 48: 1072 - 1077. McClaugherty C.A., Pastor J. and Aber J.D. (1985). Forest litter decomposition in relation to soil nitrogen dynamics and litter quality. Ecology 66(1): 266 - 275. McTiernan K.B. , Ineson P. and Coward P.A. (1997). Respiration and nutrient release from tree leaf litter mixtures. Oikos 78: 527-538. Meentemeyer V . (1978). Macroclimate and lignin control of litter decomposition rates. Ecology 59(3): 465 - 472. Melillo J .M., Aber J.D. and Muratore J.F. (1982). Nitrogen and lignin control of hardwood leaf litter decomposition dynamics. Ecology 63(3): 621 -626. 96 Melillo J.M., Aber J.D., Linkins A.E. , Ricca A. , Fry B. and Nadelhoffer K.J . (1989). Carbon and nitrogen dynamics along the decay continuum: Plant litter to soil organic matter. Plant and Soil 115: 189- 198. Miller R.H. (1974). Factors affecting the decomposition of an aerobically digested sewage sludge in soil J. Environ. Qual. 3(4): 376 - 380. Molina J.A.E., Clapp C E . and Larson W.E. (1980). Potentially mineralizable nitrogen in soil: the simple exponential model does not apply for the first 12 weeks of incubation. Soil Sci. Soc. Am. J. 44: 442-444. Mulvaney R.L. (1996). Nitrogen - Inorganic Forms in Sparks D.L, Page, A .L . , Helmke P.A., Loeppert R.H., Soltanpour P.N., Tabatabai M.A. , Johnston C T . and Sumner M.E. (eds.) (1996). Methods of Soil Analysis, Part 3 - Chemical Methods. Soil Science Society of America, Madison, Wisconsin. Nason G.E. and Myrold D.D. (1992). Nitrogen Fertilizers: Fates and Environmental Effects in Forests. Chapter 6 in Chappell G.N. Weetman G.F and Miller R.E. (1992). Forest Fertilization: Sustaining and improving nutrition and growth of western forests. University of Washington, Seattle, WA. N'Dayegamiye A. , Royer, R. and Audesse, P. (1997). Nitrogen mineralization and availability in manure composts from Quebec biological farms. Can. J. Soil Sci. 77: 345 - 350. Nguyen P.V., Hart J.B. Jnr and Merkel D . M . (1992). Municipal sludge fertilization on oak forests in Michigan: Short-term nutrient changes and growth responses. Chapter 26 in Cole D.W., Henry C.L. and Nutter W.L. (1986). The Forest Alternative for Treatment and Utilization of Municipal and Industrial Wastes. University of Washington Press, Seattle, Washington. Parkin G.F. and Owen W.F. (1986). Fundamentals of anaerobic digestion of wastewater sludges J. Environ. Engin. 112(5): 867-921 Paul E.A. and Clark F.E. (1989). Soil Microbiology and Biochemistry. Academic Press, Inc. San Diego, California. Piccolo A. , Zaccheo P. and Genevini P.G. (1992). Chemical characterization of humic substances extracted from organic-waste-amended soils. Biores. Technol. 40: 275 -282. Piotrowski, E.G., Valentine, K . M . and Pfeffer, P.E. (1984). Solid-state, 13C, cross-polarization, "magic angle " spinning, NMR Spectroscopy studies of sewage sludge. Soil Sci. 137(3): 194-203. Prescott C E . (1996). Influence of forest floor type on rates of litter decomposition in microcosms. Soil Biol. Biochem. 28(10/11): 1319- 1325. 97 Prescott C.E., Seely, B. and Taylor B. (In Review). Modelling forest management effects on organic matter decomposition in British Columbia. For. Ecol. Managem. Preston C M . , Ripmeester J.A., Mathur S.P. and Levesque M . (1986). Application of solution and solid-state NMR to a peat-based composting system for fish and crab scrap. Can. J. Spectroscopy 31(3): 63 - 69. Preston C M . , Trofymow J.A., Sayer B.G and Niu J. (1997). C nuclear magnetic resonance spectroscopy with cross-polarization and magic-angle spinning investigation of the proximate-analysis fractions used to assess litter quality in decomposition studies. Can. J. Bot. 75: 1601 - 1613. Ryan M.G. , Melillo J .M. and Ricca A . (1990). A comparison of methods for determining proximate carbon fractions of forest litter. Can. J. For. Res. 20: 166-171. Schlesinger W.H. and Hasey M . V . (1981). Decomposition of chaparral shrub foliage: Losses of organic and inorganic constituents from deciduous and evergreen leaves. Ecology 62(3): 762-774. Scott N . A . and Binkley D. (1997). Foliage litter quality and annual net N mineralization: comparison across North American forest sites. Oecologia 111: 151-159. Serna M.D. and Pomares F. (1992). Nitrogen mineralization of sludge-amended soil. Biores. Technol. 39: 185-290. Sierra J. (1997). Temperature and soil moisture dependence of N mineralization in intact soil cores. Soil Biol. Biochem. 29(9/10): 1557- 1563. Stanford G. and Smith S.J. (1972). Nitrogen mineralization potential of soils. Soil Sci. Soc. Am. J. 41: 465-472. Swift M.J. , Heal O.W. and Anderson J.M. (1979). Decomposition in Terrestrial Ecosystems -Studies in Ecology Volume 5. University of California Press. Turner J., Lambert M.J. and Rowell D. (1996). Biosolids use and management in forest plantations in Osborne G.J., Parkin M.L. , Michalk D.L. and Grieve A . M . (eds.) (1996). Biosolids Research in New South Wales. Proceedings of the Biosolids Summit, a research conference on the use of biosolids in agriculture and forestry. Rydalmere, New South Wales, Australia. NSW Agriculture Organic Waste Recycling Unit. Wershaw R.L., Leenheer J.A., Kennedy K.R. and Noyes T.I. (1996). Use ofl3C NMR and FTIR for elucidation of degradation pathways during natural litter decomposition and composting: I. Early stage leaf degradation. Soil Sci. 161(10): 667-679. Wessen B. and Berg B. (1986). Long term decomposition of barley-straw: Chemical changes and ingrowth of fungal mycelium. Soil Biol. Biochem. 18: 53-59. 98 Whitmore A.P. (1996). Describing the mineralization of carbon added to soil in crop residues using second-order kinetics. Soil Biol. Biochem. 28(10/11): 1435 - 1442. US-EPA (1983). Methods for Chemical Analysis of Water and Wastes EPA-600 / 4-79-020 Method 351.2 in Lachat Autoanalyzer methods manual (unpublished). Vesilind, P. A . (1979). Treatment and Disposal of Wastewater Sludges (revised edition). Ann Arbor Science Publishers Inc., Michigan. USA. Zhong A . (1998). Efficacy of secondary biosolids fertilization in a managed Douglas-fir forest in the Coastal Western Hemlock Zone of British Columbia. Ph.D. thesis, University of British Columbia, Vancouver, B.C. Zibilske L . M . (1997). Temperature effects on the decomposition of paper mill sludges in soil. Soil Sci. 162(3): 198-204. 99 A P P E N D I X Solid-state 1 3 C nuclear magnetic resonance spectra of the seven organic residuals 100 CHILLIWACK BS, 47:11 HZ. AUG. 1 9 / 9 8 CHILLIWACK BIOSOLIDS HOWELL.009 PPG: CPCYCLX.PC OATE 19-8-98 SF 75.468 01 8550.000 SI 2048 TD 512 SW 26315.789 AQ 9.728M CPMAS NS 5368 RO 0 DW 19.0 FW 31600 02 6200.000 DP 5H DO CHILLIWACK BS. DD-TOSS 47:11 HZ. AUG. 1 9 / 9 8 R0WELL.010 PPG: TOSSNQS.PC DATE 19-8-98 SF 75.468 01 8550.000 SI 2048 TD 512 SW 26315.7B9 AQ 9.728M NS 5995 RO 0 DW 19.0 FW 31600 02 6200.000 DP 5H DO DO 2.000S Dl 3.900U D2 7.400U D3 46.000U D5 1500.OOOU D7 10.000M D l l 4.100U DD LB GB CX CY SR 40.000 0.0 25.00 4.50 1016.50 101 200 150 100 PPM 50 WHISTLER BS, 4709 HZ, AUG. 27/98 • WHISTLER BIOSOLIDS R0WELL.019 PPG: CPCYCLX.PC DATE 2 7 - 8 - 9 8 SF 7 5 . 4 6 8 01 8 5 5 0 . 0 0 0 S I 2048 TD 512 SW 2 6 3 1 5 . 7 8 9 AQ 9 . 7 2 8 M NS 7337 RO 0 DW 1 9 . 0 WHISTLER BS, DD-TOSS 4709 HZ, AUG. 27./HE B W J ^ R ROWELL.020 PPG: TOSSNQS.PC DATE 2 7 - 8 - 9 8 SF 7 5 . 4 6 B 01 8 5 5 0 . 0 0 0 S I 2 0 4 8 TD 5 1 2 SW 2 6 3 1 5 . 7 8 9 AQ 9 . 7 2 8 M NS 13572 RO 0 DW 1 9 . 0 FW 3 1 6 0 0 02 6 2 0 0 . 0 0 0 DP 5H DO DO 2 . 0 0 0 S D l 3 . 9 0 0 U D2 7 . 4 0 0 U D3 46.OOOU D5 1500.OOOU D7 1 0 . 0 0 0 M D l l 4 . 1 0 0 U LB 4 0 . 0 0 0 GB 0 . 0 CX 2 5 . 0 0 CY 4 . 5 0 SR 1 0 5 9 . 8 5 1 0 2 ANACIS IS BS, 4720 HZ, AUG. 20./9H ANNACIS ISLAND BIOSOLIDS HOWELL.013 PPG: CPCYCLX.PC DATE 2 0 - 8 - 9 8 I I I I I I I I i 1 I I I I i I L i I i I L i I I I I 1 1 1 1 1 1 200 150 100 50 0 PPM ANACIS I S BS, OD-TOSS 4719 HZ, AUG. 2 0 / 9 8 R0WELL.014 PPG: TOSSNQS.PC DATE 2 0 - 8 - 9 8 SF 7 5 . 4 6 8 01 8 5 5 0 . 0 0 0 S I 2048 TD 512 SW 2 6 3 1 5 . 7 8 9 AQ 9 . 7 2 8 M NS 7367 RO 0 DW 1 9 . 0 FW 3 1 6 0 0 02 6 2 0 0 . 0 0 0 DP 5H DO DO 2 . 0 0 0 S D l 3 . 9 0 0 U D2 7 . 4 0 0 U D3 4 6 . 0 0 0 U D5 1 5 0 0 . 0 0 0 U D7 1 0 . 0 0 0 M D11 4 . 1 0 0 U LB 4 0 . 0 0 0 GB 0 . 0 CX 2 5 . 0 0 CY 4 . 5 0 SR 1 0 1 6 . 5 0 200 150 100 PPM SO 103 LIONS GATE BS, 4712 HZ. AUG. 1 0 / 9 0 LIONSGATE BIOSOLIDS ROWELL.001 PPG: CPCYCLX.PC DATE 1 1 - B - 9 B SF 7 5 . 4 6 8 0 1 8 5 5 0 . 0 0 0 S I 2048 TD 5 1 2 SW 2 6 3 1 5 . 7 8 9 AQ 9 . 7 2 8 M NS 6000 RO 0 DW 1 9 . 0 FW 3 1 6 0 0 02 6 2 0 0 . 0 0 0 DP 5H DO DO 2 . 0 0 0 S D l 3 . 9 0 0 U D2 6 . 8 0 0 U D3 10.OOOU D5 1500.OOOU D7 1 0 . 0 0 0 M D l l 4 . 1 0 0 U LB 3 0 . 0 0 0 GB 0 . 0 CX 2 5 . 0 0 CY 1 0 . 0 0 SR 1 0 1 6 . 5 0 CPMAS LOINS GATE BS, DD-TOSS 4713 HZ, AUG. 1 0 / 9 B R0WELL.002 PPG: TOSSNQS.PC DATE 1 0 - 8 - 9 8 SF 7 5 . 4 6 8 0 1 8 5 5 0 . 0 0 0 S I 2 0 4 8 TD 5 1 2 SW 2 6 3 1 5 . 7 8 9 AQ 9 . 7 2 8 M NS 8000 RO 0 DW 1 9 . 0 FW 3 1 6 0 0 02 6 2 0 0 . 0 0 0 DP 5H DO DO 2 . 0 0 0 S D l 3 . 9 0 0 U D2 7 . 4 0 0 U D3 48.OOOU D5 1500.OOOU D7 1 0 . 0 0 0 M D l l 4 . 1 0 0 U LB 4 0 . 0 0 0 GB 0 . 0 CX 2 5 . 0 0 CY 4 . 0 0 SR 1 0 1 6 . 5 0 104 DF L ITTER, TC 1 MS, A I , AUG. 2 6 / 9 0 DOUGLAS-FIR LITTER R0WELL.017 PPG: CPCYCLX.PC DATE 2 6 - B - 9 8 SF 7 5 . 4 6 8 0 1 8 5 5 0 . 0 0 0 S I 2048 TD 5 1 2 SW 2 6 3 1 5 . 7 8 9 AQ 9 . 7 2 8 M NS 4936 RO 0 DW 1 9 . 0 FW 31600 02 6 2 0 0 . 0 0 0 DP 5H DO . DO 2 . 0 0 0 S D l 3 . 9 0 0 U D2 6 . 8 0 0 U D3 10.OOOU D5 1000.OOOU D7 1 0 . 0 0 0 M D l l 4 . 1 0 0 U LB 3 0 . 0 0 0 GB 0 . 0 CX 2 5 . 0 0 CY 1 0 . 0 0 SR 1 0 5 9 . 8 5 CPMAS J DF L ITTER, 4 7 1 1 HZ, AUG. 2 6 / 9 8 R0WELL.016 PPG: TOSSNQS.PC DATE 2 6 - 8 - 9 8 SF 7 5 . 4 6 8 01 8 5 5 0 . 0 0 0 S I 2 0 4 8 TD 512 SW 2 6 3 1 5 . 7 8 9 AQ 9 . 7 2 8 M NS 5 1 3 5 RO . 0 DW 1 9 . 0 FW 3 1 6 0 0 02 6 2 0 0 . 0 0 0 DP 5H DO DO 2 . 0 0 0 S D l 3 . 9 0 0 U D2 7 . 4 0 0 U D3 46.OOOU 105 STRAW, 4713 HZ, AUG. 1 7 / 9 8 WHEAT STRAW ROWELL.003 PPG: CPCYCLX.PC DATE 17-3-98 SF 75.468 01 8550.000 SI 2048 TD 512 . SW 26315.789 AQ 9.728M NS 7172 RO 0 DW 19.0 FW 31600 02 6200.000 DP 5H DO DO 2.000S Dl 3 .900U D2 6.800U D3 10.OOOU D5 1500.OOOU D7 10.000M D l l 4 .100U LB 30.000 GB 0.0 CX 25.00 CY 11.00 SR 1016.50 CPMAS 200 ISO 100 PPM STRAW DD-TOSS 48U, AUG. 1 7 / 9 8 R0WELL.004 PPG: TOSSNQS.PC DATE 18-8-98 SF 75.468 01 8550.000 SI 2048 TD 512 SW 26315.7B9 AQ 9.728M NS 21600 RO 0 DW 19.0 FW 31600 02 6200.000 DP 5H DO DO 2.000S Dl 3 .900U D2 7.400U D3 48.OOOU D5 1500.OOOU D7 10.OOOM D l l 4.100U LB 40.000 GB . 0.0 CX 25.00 CY 4.00 SR 1016.50 DD 106 PULP SLUDGE, 4717 HZ, AUG. 1 8 / 9 6 PAPER FINES R0WELL.006 PPG: CPCYCLX.PC DATE 18-8-98 SF 75.468 01 8550.000 SI 2048 TD 512 SW 26315.789 AQ 9.728M NS 3774 R0 0 DW 19.0 FW 31600 02 6200.000 DP 5H DO DO 2.000S Dl 3.900U D2 6.800U D3 10.000U D5 1500.OOOU D7 lO.OOOM Dll 4.100U LB 30.000. GB 0.0 CX 25.00 CY 11.00 , SR lOifi^ n y PULP SLUDGE, DD-TOSS 45US, 4715 HZ, AUG, 1 8 / 9 8 R0WELL.007 PPG: TOSSNQS.PC DATE 1B-8-9B SF 75.468 01 8550.000 SI 2048 TD 512 SW 26315.789 AQ 9.728M NS 6737 RO 0 DW 19.0 FW 31600 02 6200.000 DP 5H DO DO 2.000S Dl 3.900U D2 7.400U 03 45.OOOU D5 1500.OOOU D7 lO.OOOM Dll 4.100U LB 40.000 GB 0.0 CX 25.00 CY 4.50 SR 1016.50 DD 200 ISO 100 PPM M 1 1 1 1 i i 50 •J 107 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0089135/manifest

Comment

Related Items