| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Dep. of Land Resource Science, Univ. of Guelph, Guelph, ON, Canada N1G 2W1
b School of Engineering, Univ. of Guelph, Guelph, ON, Canada N1G 2W1
c Bell Canada, 250 Fieldway Rd., Toronto, ON, Canada M8Z 3L2
Corresponding author (sausma{at}uoguelph.ca)
Received for publication April 21, 2000.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSFIELD SITE DESCRIPTIONSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: db, dry basis • DOY, day of year • FID, flame ionization detector • THC, total hydrocarbon • THD, total hydrocarbon detector
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSFIELD SITE DESCRIPTIONSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Landfarming has been widely used to treat a variety of industrial wastes and contaminated soils, including petroleum refinery wastes and heavily oil-contaminated soils. Within the petroleum refining industry, landfarming is used to treat dissolved air flotation sludge, slop oil emulsion solids, cooling water sludge, wastes from tank cleaning, sediments from holding tanks, separators, and sewage treatment plants, and petroleum-contaminated soils (Bartha and Bossert, 1984). In southwestern Ontario, wastes applied to landfarms typically range from 0.1 to 50% (w/w) oil and grease, and 13 to 99% (w/w) water.
Additionally, soils can become contaminated with petroleum products accidentally during product transfer and transportation. Often, low-technology treatment methods involving biodegradation of the contaminant are used.
Losses to the atmosphere can include the release of either volatile compounds or particulate matter. Volatile compounds can impair local air quality through the production of ground-level ozone in urban areas or through the release of odorous or irritating compounds. For humans, air is a significant route of exposure to toxic pollutants introduced into the environment (Hwang, 1982). Therefore, the compounds of most concern tend to be those that have a potential to volatilize and are either persistent or toxic. At landfarms, the highest volatile hydrocarbon emission rates are likely to occur immediately after waste application and before the biodegradation process can dominate (Thibodeaux and Hwang, 1982). They can also be released during waste handling prior to application, during waste application, and during cultivation.
Concern over the emission of volatiles and semivolatiles to the atmosphere from landfarm activities has been expressed in the literature (Cadena et al., 1990; Dupont, 1986; McNicoll and Baweja, 1995). As well, government agencies in both Canada and the USA have been implementing guidelines that require trapping, collecting, and treating emissions from the landfarming of petroleum products (Coover and Walker, 1990; Environment Canada, 1993). Coover and Walker (1990) state that air emissions from landfarms are one of the most important issues facing petitioners hoping to obtain no migration variances in the USA for the treatment of hazardous waste. Wastes that are high in light components, such as oil fractions containing compounds in the gasoline and kerosene boiling range, are especially susceptible to volatilization.
Several researchers have concluded that volatilization may be an important loss mechanism during hydrocarbon degradation in soil (McNicoll and Baweja, 1995; Siegell et al., 1991; Wetherold et al., 1981). Laboratory scale studies have demonstrated hydrocarbon volatilization rates from 0.1 to 41% (w/w) (Bossert and Bartha, 1984; Schwendinger, 1968; Wetherold et al., 1981). Additionally, volatile hydrocarbons may continue to be emitted during tilling of a landfarm for a period of years after landfarming operations have been discontinued (Streebin et al., 1984).
A number of methods are available for measuring the air surface exchange of trace gases from soils, including micrometeorological (Fowler and Duyzer, 1989; Lenschow, 1995) and chamber techniques (Mosier, 1989; Hutchinson and Livingston, 1993). Micrometeorological methods are continuous, unobtrusive techniques in which the trace gas flux is quantified over an area of emission. The average emissions from an area can be determined. There are few published applications demonstrating micrometeorological methods to monitor the emission of total hydrocarbons from petroleum-contaminated soils (Ausma et al., 1999), though they have been extensively applied to the measurement of other trace gas fluxes from a variety of surfaces including soils (Edwards et al., 1994; Simpson et al., 1997; Fowler and Duyzer, 1989; Hicks et al., 1989; Wagner-Riddle et al., 1996; Wesely, 1988).
The principle of chamber operation is to isolate a surface from the environment and measure concentration changes in the overlying air with time. Use of chambers can be limiting because their presence alters the local environment under study (Baldocchi et al., 1988; Denmead and Raupach, 1993). Chamber methods have been commonly used to quantify hydrocarbon emissions from refinery landfarm surfaces (American Petroleum Institute, 1989a,b; Coover and Walker, 1990; Dupont and Reineman, 1986; Streebin et al., 1984; Wetherold and Balfour, 1986). The values from these studies cannot be directly compared with values obtained in the studies described here due to the variety of compounds used in reporting the fluxes. Emissions from soils containing petroleum products are complex mixtures of many hydrocarbons; the adoption of a common simplified reporting structure would facilitate meaningful intercomparisons between studies and discussions about the effect of landfarming emissions.
This paper summarizes the development of a unique approach to implementing a robust flux gradient micrometeorological method to measure total hydrocarbon (THC) fluxes from landfarms treating petroleum–hydrocarbon products. Results from three field studies will be used to highlight the technique's performance and versatility. The primary objective of these studies was to quantify peak emissions during typical landfarm activities, such as waste application and site cultivation, that would result in elevated THC fluxes. The studies demonstrate the efficacy of the method to perform measurements during routine activities.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSFIELD SITE DESCRIPTIONSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
 | [1] |
C/z (µg C m-3 m-1) is the vertical concentration gradient. Figure 1 presents a schematic of the system developed to estimate K, measure the concentration gradient, and monitor environmental parameters.
|
Total Hydrocarbon Gradient Measurement
The vertical concentration gradient is determined by performing THC air concentration measurements above a landfarm at two levels separated by 
z using a custom-built real-time total hydrocarbon detector (THD) equipped with a flame ionization detector (FID) (Fitzgerald-Hubble, 1997; Fulton et al., 1998). Rugged, high-quality components and stainless steel fittings were used to build the gradient measurement system, which requires minimal maintenance and is capable of repeated long-term implementation under less than ideal conditions.
Air from above the landfarm soil surface is drawn through stainless steel intakes 5 cm in diameter mounted on a mast positioned on the landfarm (Fig. 1). To accomplish rapid sample delivery from the intakes to the THD, an in-line needle valve (Whitey Model SS-3NRM4; Swagelok Canada Ltd., Hamilton, ON, Canada) positioned downstream of the intakes restricts the flow to create a vacuum within the upstream tubing, and a vacuum pump (MPU 751-NO35; KNF Neuberger, Trenton, NJ) draws air through the intakes and tubing at a maximum flow rate of 30 L min-1. This ensures rapid, turbulent plug flow and minimizes sorption effects. The volumetric flow rate is low enough to not affect the local flux of THCs from the soil surface (maximum linear velocity through intakes = 0.25 m s-1). A 4.7-cm stainless steel in-line filter holder (Gelman Science, Ann Arbor, MI) with a 2-µm Teflon filter (Gelman Science) prevents particulate matter from damaging downstream instrumentation. The sampled air is drawn through 30.5 m of black polyethylene tubing (0.375 in [9.5 mm] o.d.) before delivery to the THD. A second pump (KNF NO5; KNF Neuberger) draws a subsample at 3 to 5 L min-1 off of the exhaust port of the larger pump to deliver sampled air directly to the THD. A FID sample flow of 40 mL min-1 is accomplished by pressurizing capillary tubing through a pressure relief valve (Model SS-RL324; Swagelok).
Although eddy correlation methods would provide a direct measurement of the THC flux, the THD's response is not fast enough to support this technique close to the soil surface. However, high-frequency sampling of the THD does enable resolution of very small concentration differences, which are typical when monitoring trace gas fluxes. The THD has a minimum detection limit of 140 µg C m-3 and a minimum resolvable flux of 0.4 µg C m-2 s-1 based on 30 min of sampling and K = 0.04 m2 s-1 (K depends on meteorological conditions and measurement height, a typical value was selected). The THD is calibrated for its response to propane.
Figure 2 illustrates the sampling and averaging technique used to calculate the concentration gradient on a half-hour basis. The method is described in detail below. In essence, sampling from the upper and lower intakes is alternated at 15-s intervals, vertical concentration gradients are calculated after each 15-s sampling interval, summed, and ultimately averaged after 30 min of data has been collected.
|
After each 15-s sampling interval a vertical concentration difference is obtained by first calculating the average concentration, 
S 
at the sampled level (S):
 | [2] |
S. It also maximizes the number of gradient estimates that can be performed in a period of time (tp) since the intake switching interval can be minimized. High-frequency sampling also contributes to lower noise levels since noise is proportional to n-1/2.
These average concentrations are then used to calculate the vertical concentration difference, CS (µg m-3):
 | [3] |
By measuring the difference between the average of two concentrations calculated at the same height and the concentration calculated at the second height, long-term signal drift is filtered from the gradient measurement since the time scale for intake switching (15 s) is much shorter than the time scale of drift. Through the concentration difference calculations, the gradient technique also removes background concentrations measured by the system in well-mixed environments.
The acquisition system averages the measured concentration difference every 30 min (tp):
 | [4] |
C is the half-hour averaged concentration difference used to calculate the flux in Eq. [1] and j is the number of differences calculated during tp.
Micrometeorological Experimental Approach
The micrometeorological system includes instrumentation to measure the energy balance, turbulence, temperature, humidity, wind speed, and wind direction. Figure 1 describes the integrated measurement system. Redundant measurements of most parameters are performed.
Sensible and latent heat flux measurements are performed to calculate stability and density corrections. The sensible (H) and latent (
) heat fluxes are determined via eddy correlation (Eq. [5] and [6]) using a sonic anemometer (Kaijo Corp., Tokyo, Japan) and lyman alpha hygrometer (Model AIR-LA-1; Atmospheric Instrumentation Research, Boulder, CO), respectively:
 | [5] |
is the air density (kg m-3), cp is the specific heat at constant pressure (J kg-1 K-1), w is the vertical component of the wind velocity (m s-1), and T is the potential temperature (K). The overbar denotes the time average of the instantaneous covariance of w and T:
 | [6] |
is the latent heat flux (W m-2), Lv is the latent heat of vaporization (J kg-1), and v is the water vapor density (kg m-3). A net radiometer (Radiation Energy Balance Systems, Seattle, WA) is used to measure the net all-wave radiation. Wind direction is measured using a wind vane (Model 05103; R.M. Young Company, Traverse City, MI). Four cup anemometers (Model F460; Climatronics Corp., Bohemia, NY) positioned on a mast provide a wind speed profile. Soil thermocouples are buried at several depths below the landfarm surface. The integrated system is constructed to be weatherproof and durable enough to withstand less than ideal transportation conditions and repeated assembly and disassembly for various field studies.
A variety of approaches have been developed to estimate the eddy diffusivity, required to calculate the THC flux through Eq. [1], and the associated stability corrections (Businger et al., 1971; Dyer, 1974; Dyer and Hicks, 1970; Paulson, 1970; Arya, 1981; Wyngaard et al., 1971). The approach used for this research follows the methods of Businger et al. (1971). Similarity is assumed between heat and THC transfer from the landfarm surface to the atmosphere. Equation [7] is used to calculate the THC eddy diffusivity (K) as a function of the friction velocity (u*), the sample intake heights (z), and the integrated stability function for heat transfer (
H) for both sampling heights (Businger et al., 1971):
 | [7] |
H is represented by the following expressions, depending on atmospheric stability:
 | [8] |
 | [9] |
 | [10] |
 | [11] |
 | [12] |
 | [13] |
 | [14] |
M is the momentum transfer stability correction as determined through Eq. [15] through [17]:
 | [15] |
 | [16] |
 | [17] |
 | [18] |
An estimate of u* is obtained from the wind speed profile using Eq. [14] and setting M = 0. Next, the Monin–Obukhov length is determined (Eq. [12]) and the stability ratio (z/L) is calculated. Stable conditions are defined as z/L > 0.0001, unstable conditions by z/L < -0.0001, and neutral conditions by -0.0001 
z/L 0.0001. The stability correction (M) at each measurement height is determined using Eq. [15] through [17], depending on atmospheric stability, and a new estimate for u* is obtained. An iterative process is used between the calculation of u*, L, and M until the values of u* converge.
The factor of 1.35 is applied to K as a correction factor to account for a systematic underestimation of fluxes, which has been found to occur when using methods other than Bowen ratio techniques (Simpson et al., 1997; Lee, 1998; Twine et al., 2000; Wagner-Riddle et al., 1996; G.W. Thurtell, personal communication, 1999).
Consideration must also be made for variations in the THC concentration due to the presence of a simultaneous water vapor and/or heat flux. Water vapor flux is measured with a lyman-alpha hygrometer. The design of the THD is such that the air sample entering the FID was maintained at a constant temperature of 105°C, eliminating the need for heat flux corrections.
The presence of moisture in the air sample entering the FID reduces the overall air density. Webb et al. (1980) developed equations to correct the measured water vapor flux for temperature and the measured scalar for variations in moisture. The respective equations, as applied to the measured THC fluxes, are:
 | [19] |
is temperature (K) measured through sonic anemometry, 
is the ratio of the mean water vapor density to the mean dry air density, µ is the ratio of the dry air molecular mass to the water vapor molecular mass, 
v is the mean moisture vapor density (kg m-3), is the mean air density (kg m-3), and:
 | [20] |
c is the mean THC density (g m-3), and a is the mean dry air density (kg m-3). In our measurements to date, the corrections have resulted in minimal differences between the raw and corrected fluxes; approximately 2 to 10% for large to small fluxes, respectively.
Integrated Experimental System
The micrometeorological approach requires the extensive use of instrumentation and the capability to acquire and store large volumes of associated data. The micrometeorological approach is executed in real-time using a personal computer and the measured signals are presented graphically on a monitor in real-time. The data acquisition and control software was developed in-house and it integrates the micrometeorological instrumentation, namely the sonic anemometer and the lyman alpha hygrometer, with the gradient measurement instrumentation, namely the THD and the intake solenoid valves (Fig. 1). The software is designed to acquire data at 18.2 Hz and supply continuous, real-time, half-hour estimates of the current meteorological conditions and the reduced THC fluxes. Through the activation of relays, the software accommodates either a two- or four-point THC concentration profile. The flexibility to perform a four-point profile allows the application of a micrometeorological mass balance technique when working with small sources (Lapitan et al., 1998).
The evaluation of fluxes and supporting data in real-time allows for fine-tuning of averaging parameters (Fig. 2), in-field quality assurance, and system troubleshooting. Experimental difficulties can be isolated and rectified, minimizing the amount of unusable data due to instrument problems. Data is stored in three forms for optimum archiving and data security: to a printer, to files, and graphically to the computer screen. The processed half-hour averages and unprocessed high-resolution raw data are stored in files, affording the opportunity, if necessary, for post-study processing and re-analysis.
The system's hardware consists of three distinct components: a data acquisition (DAQ) board, a computer, and a shielded input–output connector block. The DAQ board is a high-resolution, multifunction, analog/digital (A/D), input/output board (DAQ AT MIO 16X; National Instruments, Austin, TX). The board's 16 A/D channels and 2 D/A channels employ 16-bit sampling at 100 kHz. During each of the field studies, the DAQ board was installed in a Pentium computer. The data acquisition and control software was written in Borland C and is capable of operating on a Pentium 75 with 16 MB of RAM. The input/output connector block (SCB-68; National Instruments) provides an easy-to-use, rugged, low-noise signal termination and was selected to simplify the connection of signal cables to the DAQ board.
The data acquisition and control system is complemented with dataloggers (21x Micrologger; Campbell Scientific, Logan, UT) that are used to collect signals from the net radiometer, cup anemometers, wind vane, and soil thermocouples. Data from these instruments are collected at a frequency of 0.5 Hz and averaged every 30 min.
A novel experimental system has been developed that is capable of providing a high-resolution gradient and maximized sensitivity. This has been accomplished through design features such as rapid air sample delivery, which contributes to fast analyzer response, minimal tube delay, and optimized plug flow; high-frequency sampling of the analyzer that maximizes the data contributing to concentration estimates and minimizes noise; and the use of solenoids to alternate sampling between intakes in combination with an averaging algorithm that maximizes the number of concentration gradients that can be measured over an averaging period.
|
FIELD SITE DESCRIPTIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSFIELD SITE DESCRIPTIONSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Site 1—Contaminated Soil Landfarm
Total hydrocarbon (THC) fluxes were measured above a landfarm used to bioremediate soil contaminated with diesel fuel in the First Nations community of MacDowell Lake in northern Ontario. The objective of the study was to estimate the maximum level of THC emissions to assist in evaluating the effect of hydrocarbon emissions on the environment and on the health of nearby residents and workers servicing the facility. Total hydrocarbon flux measurements were performed during two studies conducted in the summer of 1996. Data collected between between 15 and 23 June 1996 (day of year [DOY] 167 to 175) are presented here. The landfarm (5.3 x 9.4 m) was centrally located in a clearing, a sufficient distance from the forest edge to allow the use of a micrometeorological method. The landfarm and other bare soil areas comprised approximately 5% of the clearing while the remainder was covered with homogeneous short grass. The maximum length and width of the clearing was approximately 77 m. The clearing had a uniform roughness and there was adequate fetch to estimate the flux. The landfarm was located in a pristine area with no local hydrocarbon sources or sinks.
The landfarm was constructed as part of site decommissioning efforts to treat approximately 50 m3 of soil contaminated with unweathered to moderately weathered diesel fuel. Soil contamination resulted over many years through accidental spillage during refueling of on-site diesel generators. The measured level of diesel fuel contamination was 1.23% (w/w) diesel fuel (dry basis [db]), as determined from a composite soil sample collected on 19 June 1996 (DOY 171). Total hydrocarbon emissions were monitored after completion of landfarm construction and during several tilling events. Intakes were centrally located on the landfarm. Cup anemometers were positioned on a mast 6 m east of the landfarm. The regional soil was characterized as clayey silt with a surface layer consisting of organic materials. The average soil surface temperature during the study was 18.5°C.
Site 2—Refinery Landfarm
In the fall of 1997, a preliminary study was performed at a landfarm in southwestern Ontario to estimate the THC flux magnitude from an active landfarm undergoing waste applications. The THC fluxes were measured during the application of dewatered biosolids (2.85% [w/w] db oil and grease) and oily liquid waste (2.67% [w/w] db oil and grease). Measurements were made between 15 and 18 Sept. 1997 (DOY 259 to 262). The landfarm facility was approximately 2.3 ha in size and consisted of four 0.5-ha landfarm plots (33 x 152 m) laid out in a north–south direction. Dirt roadways separated the plots. The landfarm facility was surrounded by trees and was isolated from regular refinery operations that were located approximately 1 km away to the west. The prevailing wind direction during the study was from the south. The surface of the landfarm was flat, offering ideal micrometeorological fetch conditions.
The regional soil is characterized as grey silty clay. The landfarm is used for the treatment of refinery-generated biosolids from a waste water treatment plant or sludge pits, and oily liquid wastes from storage pits. In 1997, the landfarm soil had an average total organic carbon (TOC) content of 8.6%, an average pH of 6.8, an average oil and grease concentration of 6.6% (w/w), and an average water concentration of 25% (w/w). Approximately 1300 Mg of waste (87.7% [w/w] water, 9.8% [w/w] solids, 2.5% [w/w] oil and grease) are applied to the landfarm annually.
Spreading typically occurs from the access roads, unless the soil surface can support the weight of the tractor and spreader. Dewatered biosolids were applied to the landfarm using a manure spreader at an application rate of approximately 2 kg m-2, while oily liquid wastes were spread using a liquid manure spreader at an application rate of approximately 2.5 kg m-2.
The vertical THC gradient was measured from a 2-m mast placed 3.5 m from the roadway on a centrally located plot, midway along its length. Micrometeorological measurements were performed using instrumentation mounted on two masts placed in the center of the adjacent plot with four cup anemometers mounted linearly. The landfarm surface had the roughness of a frequently cultivated field. Cultivation was performed in a north–south direction. During the measurement period the winds were aligned with the cultivation furrows. The surface roughness was similar on both plots.
Air temperatures were cool (average air temperature = 17.5°C, average landfarm surface temperature = 19.8°C). There had also been substantial rainfall over the weeks preceding the study, leaving the landfarm soil saturated and covered with pools of standing water.
Site 3—Refinery Landfarm
Over the summer and fall of 1999, an additional study was performed at a second landfarm in southwestern Ontario. The goal of this study was to obtain a longer time series of THC fluxes from an active refinery landfarm. A sample of data collected between 29 Oct. (DOY 302) and 3 Nov. 1999 (DOY 307) is presented here. The entire landfarm facility occupied 3.5 ha of land and was composed of eight flat fields of variable dimensions laid out in a north–south direction. THC flux measurements were performed on a 30 by 140 m field in the northwest corner of the facility. Three masts supporting the intakes and micrometeorological instrumentation were placed 3.5 to 6 m from the eastern edge of the field and 30 m from the northern edge. An oily liquid waste pond was part of the landfarm facility and was located approximately 100 m to the southwest of the field.
The facility was in an open area surrounded by low-lying vegetation. The surface of the landfarm was flat, offering ideal micrometeorological fetch conditions. Regular refinery operations were located approximately 1 km northeast of the landfarm. The regional soil is characterized as silt loam. In the fall of 1999, the landfarm soil had a pH of 7.2, an oil and grease concentration of 3.9% (w/w) (db) and a moisture content of 15.5% (db). Approximately 7000 Mg of waste (3% [w/w] oil and grease) are applied to the 3.5-ha facility annually.
No waste was applied at the facility during the months of August and September. Intensive daily waste application and site cultivation was resumed on 22 Oct. 1999 (DOY 295). During the study, oily liquid wastes were applied by subsurface injection to a depth of approximately 15 cm at a daily application rate of 1.3 kg m-2 and water from a digester was applied to the soil surface to supply microbes at a daily application rate of 0.9 kg m-2. On a daily basis between DOY 302 through 305, 5500 kg of waste was applied to the soil subsurface, the field was cultivated three to four times, and 3700 kg of digester water was applied. Cultivation was performed immediately after wastes were applied to the soil. There was no landfarm activity on DOY 306 and 307 due to heavy rainfall.
Air temperatures were cool (average air temperature = 14°C, average landfarm surface temperature = 14°C).
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSFIELD SITE DESCRIPTIONSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
|
|
It was expected that emissions of THCs would be elevated during landfarm construction and shortly thereafter while newly exposed contaminated soil released volatile THCs, during tilling of the landfarm when buried soil would be exposed to the air, and during periods of elevated temperatures when the vapor pressures of the hydrocarbons would increase.
Total hydrocarbon (THC) fluxes peaked early in the study, shortly after landfarm construction was completed, and decreased over time. Loss of available volatile hydrocarbons and the declining temperature between DOY 169 and 175 would both contribute to the decreasing fluxes. A diurnal variation is evident in the fluxes between DOY 169 and 171, correlating to diurnal temperature variations. There is insufficient data on the following days to verify this trend; however, THC fluxes do begin to rise on DOY 175 with increasing temperatures after a minimum of both temperature and fluxes on DOY 174. Precipitation events throughout the study are expected to reduce volatile emissions to the atmosphere by limiting the transport of volatile hydrocarbons in the soil.
Tilling on DOY 171 did not result in highly elevated THC fluxes; however, rainfall occurred on the same day and would have depressed fluxes.
A subsequent static chamber study was conducted by the authors at this site to evaluate site heterogeneity with regard to THC emissions. Fluxes estimated during this study were the same order of magnitude as the flux gradient measurements presented here. The site conditions, that is, a source surrounded by a nonemitter, also allowed us to perform a simple mass balance calculation to estimate the flux. The results of this analysis were within 20% of the micrometeorological results presented here.
Figure 4 presents the THC flux estimates from the first refinery landfarm study. An oily liquid waste application was made during the early afternoon of DOY 259 and a dewatered biosolids application was made during the afternoon of DOY 261. A rain event occurred on DOY 260. The data were filtered to remove points collected when measurements may have been influenced by other THC sources or intake manipulations. One-hour and 0.5-hour periods of high flux data collected immediately after spreading oily liquid waste and dewatered biosolid waste, respectively, were lost for these reasons.
A peak during each spreading event characterized THC fluxes from the refinery landfarm (Fig. 4). Flux levels dropped to approximately 1 µg C m-2 s-1 within 30 h for the oily liquid waste and 20 h for the dewatered biosolids. This response, a peak followed by a sharp decline, is typical when refinery wastes containing hydrocarbons are spread onto a landfarm surface (American Petroleum Institute, 1989b; Coover and Walker, 1990; Evans, 1988; Streebin et al., 1984). The maximum emission rates estimated for oily liquid waste and dewatered biosolids spreading were 185 and 17 µg C m-2 s-1, respectively. A background flux measurement of 2.1 µg C m-2 s-1 was measured on the road adjacent to the landfarm. This was much lower than the fluxes measured above the landfarm soil, suggesting that the local roads and vegetation did not contribute significantly to the measured landfarm fluxes.
Figure 5 presents flux measurements collected during the refinery landfarm study performed in the fall of 1999. Ideal wind directions were limited due to the on-site oily liquid pond and limited fetch in several wind directions: data were filtered to only include measurements between 180 and 360° where the fetch was adequate. The fluxes peaked at 64, 45, 119, and 66 µg C m-2 s-1 on DOY 302 through 305, respectively. Rain began early on DOY 306 and continued through DOY 307. The magnitudes of the THC fluxes during oily liquid waste spreading are 25 to 65% of the peak fluxes measured during the first refinery study, suggesting that subsurface injection followed by cultivation does not eliminate emissions during application. Cultivating immediately after waste application does not provide adequate opportunity for wastes to become incorporated with the soil. Repeated cultivation of the soil surface results in emissions by exposing waste material.
During periods of non-activity (i.e., early morning of DOY 304 and night of DOY 305) fluxes declined from peak daytime values to 3 to 20 µg C m-2 s-1. The THC fluxes dropped late on DOY 306 to levels fluctuating between -3 and 3 µg C m-2 s-1 when temperatures dropped to less than 5°C.
Figures 3 through 5 demonstrate the ability of the designed flux gradient micrometeorological system to measure THC fluxes from both hydrocarbon-contaminated soils and refinery landfarms without interfering with routine facility operations. The technique was also able to discern emission trends from contaminated soils over time and to monitor changes in emissions as a result of landfarm manipulations such as waste application and cultivation.
Although there are no published studies to directly compare with the studies described here, the data from published chamber studies performed at refinery landfarms can be used to validate the magnitude of the measured THC fluxes. Published peak flux estimates range from approximately 64 to 6500 µg m-2 s-1 for a variety of hydrocarbon compounds depending upon the conditions of waste application (American Petroleum Institute, 1989a,b; Coover and Walker, 1990; Dupont and Reineman, 1986; Streebin et al., 1984; Wetherold and Balfour, 1986). The peak values obtained in our studies for refinery wastes, composed of 3% or less oil and grease applied at a maximum of 2.5 kg m-2, fall within the lower range of these values, which would be expected considering that the waste application rates in the published studies ranged from 5 to 13 kg m-2 for wastes composed of 10 to 45% oil and grease.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSFIELD SITE DESCRIPTIONSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
The gradient and micrometeorological measurement systems are integrated using a custom-designed and developed data acquisition and control system. Real-time measurements along with on-screen visualizations of concentrations allow fine-tuning of the averaging algorithms, isolation of instrument problems, and the assurance of high-quality data while at the field location.
The integrated system incorporates design features that provide a high-resolution gradient and maximized sensitivity. This is accomplished by source sampling under vacuum, high frequency analyzer signal sampling, using solenoid valves to provide rapid intake switching, and using an averaging algorithm to maximize gradient measurements.
Total hydrocarbon (THC) flux measurements from three active landfarm facilities treating diesel fuel–contaminated soil and refinery wastes were presented. The THC emissions peaked during liquid and solid waste application and during subsequent cultivation of the landfarms. Peak emissions during these activities fell within the range of values published for chamber studies. The field studies demonstrated the versatility of the developed technique to monitor emissions from landfarms during soil manipulations such as waste application and cultivation without interfering with routine facility activities. The ability to perform these measurements provides emissions data during activities that result in hydrocarbon bursts to the atmosphere.
In conjunction with comprehensive monitoring of soil conditions and/or the incorporation of other trace gas analyzers, the method can be a useful tool to study mechanisms, such as temperature, moisture content, and nutrient levels, which control biodegradation and the release of hydrocarbons from petroleum-contaminated soils. This information is important in assessing the effect of emissions on the environment, and on the health of workers at the facility and nearby residents. It also offers the opportunity to develop and assess management strategies that will reduce emissions while reducing hydrocarbons in the soil through biodegradation.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSFIELD SITE DESCRIPTIONSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  S. Ausma, G. C. Edwards, and T. J. Gillespie Laboratory-Scale Measurement of Trace Gas Fluxes from Landfarm Soils J. Environ. Qual., January 1, 2003; 32(1): 8 - 22. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK |
