Die-off of Fecal Indicator Organisms Following Land Application of Poultry Manure

Land-applied animal manure can be a major contributor of agricultural nonpoint source pollution. For surface-applied manures the time period prior to the first runoff is of greatest concern, because most enteric species initially exhibit rapid die-off rates. The objectives of this investigation were to follow the die-off of indicator organisms after manure application to a bar? soil surface and to ascertain if dieoff could be accurately modeled by first-order-kinetics. Poultry manure was surface-applied at approximately 36.5 and 164 metric tons/ha on Davidson clay loam and Norfolk sandy loam plots in a controlled environment chamber at a constant temperature (24.5°C). Soil samples were analyzed at intervals during a 30-day period for fecal coliforms and fecal streptococci. Modeling die-off of fecal coliform during the first 7 days gave good agreement with experimental data showing an average die-off rate constant of 0.29/day. After 7 days, regrowth of fecal coliform populations was seen on all plots. Fecal streptococcal numbers declined steadily after the first 3 days of the 30-day study and exhibited an average first-order rate constant of 0.093/day. Neither soil type nor manure application rate seemed to influence the decline in organism populations. First-order kinetics did not explain all aspects of die-off, but it appears this model could be successfully employed as a first approximation for estimating bacterial kinetics in a larger nonpoint source pollution model for agricultural lands. ' Paper No. 6068 of the Journal Series of the North Carolina Agric. Res. Service, Raleigh, N.C. The use of trade names in this publication does not imply endorsement by the North Carolina Agric. Res. Service of the products named, nor criticism of similar ones not mentioned. Received 16 July 1979. 2 Research Associate, Associate Professor, and Associate Professor, respectively, Biological and Agricultural Engineering Department, North Carolina State Univ., Raleigh, NC 27650. Additional Index Words: fecal coliforms, fecal streptococci, firstorder kinetics, modeling, nonpoint source, animal waste, poultry manure. Crane, S. R., P. W. Westerman, and M. R. Overcash. 1980. Die-off of fecal indicator organisms following surface poultry manure application. J. Environ. Qual. 9:531-537. The methodology of animal waste disposal has received increasing study over the past decade after the realization that this material may be a major contributor to agricultural nonpoint source pollution. Because of prohibitive costs associated with advanced waste treatment methods and facilities, land application of animal wastes is the only economic alternative for most producers who must now comply with no-discharge regulations or face closure. Land-applied animal wastes are a major source of microorganisms including many pathogenic species. This presents a potential health hazard from animal or human exposure via the runoff from these disposal areas. For surface-applied wastes, the first runoff event after waste application is often of greatest concern. Since most enteric microorganisms exhibit a rapid initial die-off rate, the number of viable pathogens and, hence, their availability to be lost in runoff can greatly decrease within a few days after waste appliction. It is, therefore, important that the die-off rate of these microorganisms be determined immediately following their application J. Environ. Qual., Vol. 9, no. 3,1980 531 to land so that a realistic estimate can be made of the environmental hazard they present. Animal wastes have the potential to carry and spread many diseases; however, the actual incidence of such disease is extremely low. A list of 25 significant diseases and the organisms involved was reported by Azevedo and Stout (1974). The important variables involved the survival of enteric organisms following application to soil were presented by Morrison and Martin (1977) and Ellis and McCalla (1976). In summary, they were: (i) the physiological state of the organism, (ii) the physical and chemical nature of the receiving soil, (iii) atmospheric conditions, (iv) biological interaction with other organisms, and (v) the waste application method. The soil provides a natural filtering action and adsorption site for the removal of bacteria and virus. Gerba et al. (1975) found 92 and 97~/0 of the bacteria were removed in the top 1 cm of soil. Due to the removal processes, a majority of the organisms are stranded at the soil surface, thus increasing the likelihood of losses during runoff. The continued application of waste on one area could also lead to extended pathogen survival and build-up (Dazzo et al., 1973). Difficulties and expenses involved in testing for pathogens have generally led to the use of indicator organisms of enteric origin to estimate the die-off of pathogens in the soil. Although it is generally accepted that indicator organisms will be affected in the same manner as pathogens in the soil, few studies have been made comparing the survival of these indicators in the soil with known pathogens. One investigation of this kind was performed by Dazzo et al. (1973) who found that die-off of Salmonella enteritidis was related to fecal coliform die-off under similar conditions. The indicator organisms generally accepted, with standard tests for determination, are fecal coliforms and fecal streptococci (Standard Methods, 1976; Van Donsel et al., 1967; Geldreich et al., 1962; Cohen and Schuval, 1973). Morrison and Martin (1977) concluded that even though the use of indicator organisms may not give data comparable to the die-off of all pathogens, it is probably true that they are more resistant to die-off than are pathogens. This makes any error on the side of public safety. The use of models to describe pathogenic and indicator organisms die-off in soil is very limited. Klock (1971) used first-order kinetics to describe the die-off coliforms in lagoon waters, and related the die-off rates to various environmental conditions. Smallbeck and Bromel (1975) also used first-order kinetics to estimate the die-off of Eschericha coli and S. typhimurium in seeded animal lagoon effluent applied to a clay soil. They found the death rate for these organisms was rapid after waste application and that the first-order model provided an accurate description of the data. The study reported here was initiated to ascertain the die-off rates of fecal coliforms and fecal streptococci following surface poultry manure application to a bare soil. The objective was to determine if this die-off could be satisfactorily modeled by first-order kinetics. This information is needed to better evaluate the environmental hazards presented by the land disposal of animal waste as related to modeling of nonpoint source contributions. MATERIALS AND METHODS Experimental Arrangement A laboratory study utilizing a controlled environment was chosen for this investigation to minimize the interaction of climatic factors found naturally in the field. The temperature was 24.5 42oc and the relative humidity of the air was 70 ± 5% as measured continuously by a hygrothermograph. Two soils were chosen for the experiment as common soil types with substantially different properties. The first soil, typical of the Coastal Plains, was a Norfolk loamy fine sand. The Norfolk Series is a member of the fine-loamy, siliceous, thermic family of Typic Paleudults. It is characterized by a surface layer of loamy sand 23 cm thick, is well drained with a high infiltration rate, and is usually on level to slightly sloping land (North Carolina Soil Conservation Service, personal communication). The second soil used was Davidson clay loam from the Piedmont region. The Davidson Series is a member of the clayey, kaolinitic, thermic (oxidic) family of Rhodic Paleudults. It is characterized by a surface horizon of loam 18 cm thick with a first sublayer of clay 46 cm thick. The average slope ranges from 2 to 12% with moderate permeability. It is well drained with little flood hazard (North Carolina Soil Conservation Service, personal communication). Soil from surface horizons (0to 15-cm deptl~ of the plow layer) had as much vegetative matter removed as possible and then was air dried, ground, passed through a 6.5-mm sieve, and mixed well to produce a more homogeneous material. It was stored at 15°C for later use. Only the surface horizon was used, because this is the soil that comes into direct contact with surface-applied wastes under field conditions. The experimental design was a 2 x 3 factorial. Two types of soil and three levels of waste application were used (including a 0 level or control). The experimental apparatus consisted of six boxes (plots--I plot/treatment) made from sheet metal, with each having an open-end surface area of 0.0929 m2 and a depth of 40.6 cm. The boxes were coated with an epoxy-based paint to prevent corrosion and contamination from zinc contained in the galvanized metal. A soil base 20.3 cm deep, composed of the previously dried surface soil, was placed in the bottom of each box and packed to the bulk densities determined at the soil collection sites. This soil base was used to better simulate field conditions by allowing a more natural moisture flow in the soil during the experiment. A 3.2-mm mesh screen was cut to size and placed over the soil base to insure samples of uniform depth. A quantity of soil was then placed on top of this screening and packed to the bulk density of the base soil. The depth of this upper or sampling layer was 5 cm in all boxes. This depth was chosen because: (i) calculations indicated liquid contained in the poultry waste slurry would not infiltrate below 5 cm upon waste application, and (ii) it allowed adequate soil sample volumes. One pore volume of distilled water was added to all boxes 3 days before surface waste application and the soil was allowed to drain to field capacity conditions. The waste material used was raw poultry manure collected from a caged laying house with an age varying from 0 to 3 days. Two waste application rates were chosen based on the total Kjeldahl nitrogen in the manure. A summary of the treatments and corresponding manure application rates is shown in Table 1. The raw waste was diluted by volume (3 parts waste:l part H~O) and homogenized to make it more amenable to measurement and surface spreading. The advantage of greater accuracy in measurement and spreading brought about by this Table 1--Manure and total Kjeldahl nitrogen application rates.~ Manure Nitrogen Soil type Treatment loading rate loading rate kg wet manure/ha kg TKN/ha Davidson clay C-0 0 0 C-220 36,500 220 C-910 152,000 910 Norfolk sand S-0 0 0 S-220 36,500 220 S-1050 176,000 1,050 All loading rates based on weight of manure per unit surface area, rather than weight/weight basis {ppm}. 532 J. Environ. Qual., Vol. 9, no. 3, 1980 dilution was assumed to outweigh the disadvantage of not using the actual raw manure. In the average caged layer house this dilution level could easily be attained through washdown procedures or leaky watering devices. The manure was then applied to the surface of the soil and spread evenly at the two predetermined rates. It was originally planned that the highest application rate on both soils would be equal, However, through an oversight, not enough manure,was left for application on the clay at the same rate as on the sand. This resulted in the differences shown in Table I. The boxes were incubated under the conditions mentioned earlier for 30 days. Soil cores were removed from each plot for analysis at scheduled intervals during the incubation period. Soil Sampling and Analyses Procedures Bacterial die-off was studied by removing soil samples from the plots on a predetermined schedule for analysis. A soil sample consisted of the composite of two soil cores taken to a depth of 5 cm on a randomized grid from a specific treatment. This was done to reduce the effect of spatial variability across the plots and to assure an unbiased sample. The core holes were then refilled with soil of approximately the same moisture content as that removed to reduce moisture redistribution in the plots. All sample containers were placed on ice to reduce further changes that might take place before analysis, The total time between sampling and the start of bacterial analysis averaged from 1 to 2 hours. In the laboratory, the samples were mixed thoroughly and a portion removed for soil pH and moisture content determination. Soil moisture was measured on an oven-dry weight basis (Black, 1965a) and soil pH was determined in a 1:1 soil/water suspension using a glass electrode (Black, 1965b). A second portion of the fresh, well-mixed soil samples was prepared for bacterial analysis using the method outlined by Van Donsel (1967). Five grams of soil were removed from the composite soil samples, added to 50 ml of phosphate buffer solution (0.0003M K H2PO,, pH 7.2), and shaken vigorously by hand with 6 ram-glass beads to insure homogeneous mixture. These solutions were then serially diluted and bacteria removed by the membrane filter technique (APHA, 1971). All soil samples were analyzed in triplicate to increase experimental precision. For fecal coliforms, the filters were incubated on Difco M-FC Bacterial Broth for 24 hours at 44.5 ~ 0.2°C. The fecal streptococci filters were incubated on Difco M-Enterococcus Agar for 48 hours at a temperature of 35 ± 0.5 °C. All solution preparation procedures and sterilization techniques were followed directly from Standard Methods (APHA, 1971). Organism counts were made under magnification of 10 x and results expressed as organism number/gram of oven-dry soil. Several plates were taken randomly for each sampling period and colonies were confirmed by the methods outlined in Standard Methods (APHA, 1971). Data Analysis and Modeling Theory All data analysis and statistical procedures were performed using standard package programs of Statistical Analysis Systems--76 (Burr et al., 1976). First-order kinetics were assumed to follow the equation: