A comparative ecotoxicity impact assessment of hydraulic fracturing chemicals used in oil and gas and coal-bed methane wells

doi:10.21203/rs.3.rs-4356156/v1

In this study, we examine the ecotoxicity impacts of chemicals used in hydraulic-fracturing (HF) operations for shale-based oil and gas (O&G) and coal-bed methane (CBM) wells in Colorado. The chemical constituents of HF fluids were analyzed for 40 O&G and 10 CBM wells located in Weld and Las Animas counties. Integrating toxicity data, physical properties, and fate and transport parameter predictions with the USEtox human and ecotoxicological impact model, we developed new fresh water ecotoxicity characterization factors for 184 different chemicals that are commonly used in fracking operations. Further, we also estimated and analyzed the overall ecotoxicity impacts associated with the potential release of these chemicals into the environment. We found that CBM HF wells, on an average, resulted in 12.6% greater ecotoxicity impacts from chemical release than O&G wells. Ecological Impacts were highly-sensitive to the fraction of HF fluid that returns to the surface as part of flow back.

Hydraulic Fracturing

Fracking Fluid

Fresh Water Ecotoxicity

Impact Assessment

Fate & Transport Modeling

The exploitation of unconventional gas deposits that involves the use of hydraulic fracturing has been an intensive topic of research. Hydraulic fracturing (HF) is a process that involves underground injection of a base fluid (water), a proppant (almost always sand), and various chemicals under high pressure (681 atmospheres for the Marcellus Shale, Stringfellow et al., 2014) to stimulate fractures in the rock so that gas and/or oil can flow easily to the surface, where it can be collected, exported, and processed. Used in conjunction with directional drilling (also called horizontal drilling), hydraulic fracturing has enabled previously unreachable or economically unviable deposits of oil and gas to be exploited; these deposits are more commonly referred to as unconventional oil and gas and include shale, tight formations, and coal-bed. This study focuses on the environmental impacts – specifically freshwater ecotoxicity impacts – associated with the release of chemicals present in the fracking fluid used for shale and coal-bed deposits into the atmosphere.

Although shale and coal-bed are both unconventional sources that use hydraulic fracturing, there are important distinctions in the geology that differentiates the two and govern which chemicals are required, the quantities of flowback returned, and produced water generated. For example, the Niobrara Formation (shale) has an average thickness of 240 to 330 feet, located at an average depth of 6,800 feet ranging from 3,000 to 8000 feet (Higley et al., 1995). For shale, once fracturing is completed the oil and gas begins to flow to the surface and not much is left to do at the well until site–closure. On the other hand, coal-bed occurs at shallower depths of 500 to 4100 feet (US EPA, 2004), meaning that less volume of water is required to inundate the entire volume of the borehole emanating from the well at the surface and the thickness is thinner at 10 to 140 feet (US EPA, 2004). Fracking is necessary for shale but not for coal-bed methane, although it is used in most cases. CBM might already have existing fractures, but HF aids to increase their size; shale deposits require fractures to be formed. Gas is not structurally-trapped in coal-bed, instead it is adsorbed within the coal. In CBM the fracking fluid is injected at increasingly high pressures until the coal cannot withstand the pressure. At that point, the produced water and some fracking fluid is pumped to the surface, along with the methane that sorbs out of the coal. CBM produced wet gas as a result of the large quantities of water produced in coal-bed deposits (US EPA, 2004).

The national chemical disclosure registry for hydraulically-fractured oil and gas wells in the United States, FracFocus.org, lists 15, separate chemical categories that are delineated by their chemical function. Chemicals are chosen based on their function, with some chemicals sharing multiple functions. Acids breakdown and initiate the formation of fissures in the rocks. Corrosion inhibitors prevent the pipe and casings from corroding. Biocides kill bacteria and microorganisms that can cause corrosion in the pipes. The base fluid carries all the proppant and chemicals and represents the vast majority of mass of the fluid. Breakers delay breakdown of gels at key points in the fracking process. Clay and shale stabilization and control reduce the amounts of underground clay that gets mixed into the fluid. Crosslinkers maintain the viscosity with increasing temperatures. Friction reducers reduce the friction caused by fluid transport which can increase pump strain and other damaging factors. Gels thicken the water to better suspend the proppant as it moves through the pipe. Non-emulsifiers break apart or otherwise separate the oil and water mixtures. A pH adjusting agent or buffer maintains the effectiveness of other chemicals by reducing the fluids acidity. The proppant aids in maintaining the structural integrity of the newly-formed fractures. Scale inhibitors prevent scale build-up in downhole. Surfactants reduce surface tension of the fluid to increase fluid recovery of the flowback fluid (FracFocus.org), which is the fraction of the injected fracking fluid that returns to the surface.

There is not an industrial consensus on the exact category names. In addition to the FracFocus.org categories, the US Environmental Protection Agency lists emulsifiers, foaming agents, iron control agents, resin curing agents, and solvents. Emulsifiers enable the dispersion of two immiscible fluids into one another. Foaming agents generate and stabilize foam-based fracturing fluids, which are common in coal-bed deposits. Iron precipitation is controlled by the iron controlling agent. When the temperatures downhole are too low, resin curing agents are used to lower the proppant activation temperature. Solvents controls the wetness of contact surfaces and/or they prevent or break apart emulsions (US EPA, 2015). There are some functions which work against each other because different properties are desired when the fluid is being transported underground prior to the well fracture compared to when it is returning or the well is producing.

Produced water is a term often used synonymously or confused with flowback water; indeed, many definitions exist for both terms and how the exact definitions are employed depends on the specific study, sources, agency, etc. (US EPA, 2015). Produced water is the naturally-occurring underground water that is not sourced from the injected fluid. It is impossible to differentiate between the two water streams that is emptied into the disposal pits, so for simplicity they are often lumped together and called produced water. For purposes of this paper they will be mentioned separately because the focus is on the chemicals in the fracking fluid, which entirely returns in the flowback water, but in the real world they are all part of the same water stream (American Water Works Association, 2013).

Geological formations do not adhere to geopolitical boundaries, but, fortunately, in Colorado Weld County sits on top of the largest shale formation, the Denver-Julesburg (DJ) Basin, almost exclusively. Weld County has 37% of all hydraulically-fractured wells in the state, and nearly half of all producing wells, and for this reason Weld County was chosen as the focus for the shale wells. Las Animas County sits on top of the Raton Basin, which is a coal-bed deposit with 6.5% of all wells in Colorado and was chosen as the location of the coal-bed analysis (Hamm, 2015). Much of the fracking fluid initially injected underground remains there, but the exact figures vary widely between wells and by local geology. The wells used for the assessment in Weld County were drilled into the Niobrara play, which is in the DJ Basin. Flowback volumes for the Niobrara play are approximately between 8 and 27% (US EPA, 2015) of that which is injected, averaging 17.5%; average values for the DJ Basin as a whole range between 15 and 30% (Boschee, 2014). Flowback data for coal-bed deposits is more difficult to find, and we relied on data from Puri et al. (1991) which states that 61% of the injected volume returns as flowback over a 19 day period. Most injection periods do not last long, approximately ten (Mantell, 2013; Clark et al., 2012) to fourteen (Jiang et al., 2014) days. These two values – 17.5% for Weld County and 61% for Las Animas County – were used as the specific flowback values for wells located in those respective counties throughout the assessment.

Relatively little research has been done to quantify the impacts (National Toxics Network, 2013) from the chemicals used in the hydraulic fracturing fluid due to the numerous chemicals¹⁴ employed in each well’s fracking fluid. Many fracking studies focus on the greenhouse gas emissions for the oil and gas well life-cycle, with some focusing only on methane leakage^15–18. Less emphasis is on non-shale sources of unconventional fossil fuels and their comparative impacts with one another; most assessments compare a specific unconventional source of oil or gas (usually shale) with already well-researched sources of conventional oil or gas¹⁹. The Marcellus shale is the most well-researched deposit in the United States, resulting in the most available and easily accessible data than other locations. Since most data are generic, and are not specific to areas most findings have to be interpreted at a relatively macro-level for screening ²⁰. Higher resolution impact assessments are relatively sparse due to lack of specific local data and educated assumptions required, multiple data sources, or tedious workload involved.

Even the studies¹ that do focus on the chemicals, usually only assess chemicals that are already known through lists and reports put together by other organizations^5,14, and most do not take empirical fluids and break them down to their constituents. In many cases, when there exists a list of chemicals, many of the chemicals have not been studied, or assessed: an Australian study found only two out of 23 of the chemicals have been researched in any detail²¹.

There is no standardized source that contains all the necessary property data, and they have to be collated from multiple sources^22–24 and prediction models need to be used if data is not available²⁵. A proper analysis involves finding each chemical’s applicable physio-chemical, human health, and ecotoxicity properties is a laborious process that depends on the data availability for the chemicals. If no experimental data exists, which is the case for many chemicals, the data must be predicted through modeling efforts.

The goal of this work is to devise a standardized source for all the data required to undergo chemical impact assessments and to develop, for the first time, characterization factors using USEtox that can be utilized within the existing impact assessment methodology such as TRACI²⁶ with an aim to carry out impact assessment on any well level. This study focuses on the development of freshwater ecotoxicity characterization factors using USEtox. Further, another goal is to utilize these developed characterization factors to perform a comparative freshwater ecotoxicity assessment for fracking and coal bed methane wells.

Few studies focus on Colorado fracking operations and their impact, so the standardized methodology was applied to oil and gas wells in Weld County and coal-bed methane wells in Las Animas County to (1) generate novel categorization factors for chemicals, (2) determine the average impact per well in each county, and (3) determine which of these unconventional sources has a larger impact per unit of energy generated from an average well.

The main goal of this study was to understand, quantify, and compare the freshwater ecotoxicity impacts of fracking chemicals used in Oil and Gas wells and Coal-Bed methane wells. Towards this end we first focused on generating ecotoxicity characterization factors for 140 different chemicals using the USEtox model. Next, we estimate the mass of different chemicals likely released into the different media during the storage and treatment of the produced water. Finally, by integrating the mass release with the characterization factors we were able to predict the ecotoxicity impact associated with these chemicals. The approach used to generate the characterization factors (Fig. 1) was devised around the data availability of the chemicals, since lack of reliable data for physio-chemical and toxicity properties was the main bottleneck;

Looking through the TRACI and USEtox chemical databases: 44 chemicals were present in both TRACI and USEtox, five were only in USEtox, and the remaining 135 existed in neither. Of those 135, we developed new characterization factors for 82 chemicals. 34 chemicals either lacked enough data to be reliably able to create characterization factors (CF’s), were called something ambiguous, or were named by their category (i.e. surfactant, 3rd party additive, proprietary component, etc.); for these we applied the average of all the known CF’s. The 21 residual chemicals had substitutes deemed to be the same chemical with a different name but otherwise the same chemical and simplified molecular-input line-entry system (SMILES) structure.

PubChem²³ and ToxNet²² were used for finding experimental data (in green). SMILES notation, found through Chemicalize²⁴, was used in the Estimation Program Interface Suite (EPI Suite)²⁵ for predicting physio-chemical and environmental fate data and in the ECOlogical Structure Activity Relationships (ECOSAR)²⁷ (in orange). Once all applicable data were acquired or predicted, they were arranged in a format conducive to input in USEtox (in orange).

2.1 Generating physio-chemical property data

The octanol-water partitioning coefficient $\left({K}_{OW}\right)$, vapor pressure $\left({v}_{p}\right)$, water solubility at 25°C $\left(So{l}_{25}\right)$, and to a lesser extent molecular weight $\left(MW\right)$ are required for USEtox to use its imbedded equations and must be found externally. Using the EPI Suite: ${K}_{OW}$ was predicted via the Fragment method; for predicting ${v}_{p}$ the modified grain method was used. The $MW$ was calculated using the chemical structure or SMILES to sum the constituent atomic weights.

Our spreadsheet includes all equations, experimental and predicted data, and assumptions for each chemical (labeled Chemicals, Properties in the supplementary data). Experimental data, if available, was always used before using predictive models.

Henry’s constant, k _H

$${k}_{H}\left[\frac{Pa\bullet {m}^{3}}{mol}\right]=\left({p}_{v}\left[Pa\right]\times \frac{MW\left[\frac{g}{mol}\right]}{So{l}_{25}\left[\frac{mg}{L}\right]}\right)\left(\left[1000\frac{mg}{g}\right]\left[0.001\frac{{m}^{3}}{L}\right]\right)$$

1

Degradation rate in air, k_degA

Estimated using the OVERALL OH Rate Constant (kOH) under AopWin in EPISuite, which is multiplied by the constant for hydroxide 1.5 x 10⁶.

$${k}_{degA} \left[{s}^{-1}\right]=1.5\times {10}^{6}\times kOH \left({s}^{-1}\right)$$

2

Degradation rates of water, soil, and sediment, k_deg

In EPI Suite, neglecting the values, under Biowin3, we used the time range to match the assigned rate constant in Table 1. Once the rate constant was identified the individual formula for water, soil, and sediment was applied

Table 1

Assigned rate constants associated with BioWin3
Biowin3 Output	Assigned Rate Constant (s^− 1)
Hours	4.7 x 10^− 5
Hours to Days	6.4 x 10^− 6
Days	3.4 x 10^− 6
Days to Weeks	9.3 x 10^− 7
Weeks	5.3 x 10^− 7
Weeks to Months	2.1 x 10^− 7
Months	1.3 x 10^− 7
Recalcitrant	4.5 x 10^− 8

Degradation rate in water, k _degW

$${k}_{degW} \left[{s}^{-1}\right]=Assigned Rate Constant$$

3

Degradation rate in soil,k _degSi

$${k}_{degSi} \left[{s}^{-1}\right]=Assigned Rate Constant/2$$

4

Degradation rate in sediment,k _degSd

$${k}_{degSd} \left[{s}^{-1}\right]=Assigned Rate Constant/9$$

5

Organic carbon partitioning, K_oc

In EPI Suite predict using the Molecular Connectivity Index (MCI) or use

$${K}_{OC}=1.26{K}_{OW}+0.81$$

6

Soil-water partitioning, K _d (kg/L): if experimental data does not exist the following formula was used; f_oc is the fraction of organic carbon and is assumed to be 0.02

$${K}_{d}\left[\frac{kg}{L}\right]={f}_{oc}{K}_{oc}=0.02{K}_{oc}$$

7

Air-water partitioning, K _aw (dimensionless): if experimental data does not exist the following formula was used

$${K}_{aw}[-]=\frac{{K}_{H}\left[\frac{Pa {m}^{3}}{mol}\right]}{R\left[\frac{{m}^{3} Pa}{mol K}\right]{ T}_{env}\left[K\right]}=\frac{0.1203{K}_{H}}{{T}_{env}}$$

8

Partition coefficient between dissolved organic carbon and water,K _doc

$${K}_{doc}=0.08{K}_{OW}; for {K}_{OW}<7.5$$

9

Bio-transfer factor for meat ²⁸,BTF_meat

$$BT{F}_{meat} \left[\frac{d}{k{g}_{milk}}\right]={10}^{-7.735+1.033\text{log}\left({K}_{OW}\right)}$$

10

Bio-transfer factor for milk ²⁸, BTF_milk

$$BT{F}_{milk} \left[\frac{d}{k{g}_{milk}}\right]={10}^{-8.1+\text{log}\left({K}_{OW}\right)}$$

11

Bioaccumulation factor in fish, BAF_fish (L/kg fish)

$$BA{F}_{fish} \left[\frac{L}{k{g}_{fish}}\right]=0.05{K}_{OW};for {K}_{OW}<9$$

12

Dissipation rates in above-ground plant tissues, k_dissP (1/s): if left blank USEtox automatically calculates internally using the formula

$${k}_{dissP}=0.1{k}_{degSi}$$

13

For inorganic chemicals USEtox assumes ${K}_{H}$, ${v}_{p}$, ${k}_{deg,A}$, ${k}_{deg,W}$, ${k}_{deg,sd}$, and ${k}_{deg,si}$ are all $1\times {10}^{-20}$

2.2 Calculating aqueous-based and evaporative chemical masses

Once the flowback fluid, with the chemicals, returns to the surface it is temporarily stored in an open-air pit, where the chemical masses can evaporate into the atmosphere, percolate into the soil, or remain in the water. Environmental regulations state that the pits must be lined if built after 01 May 2009, by which we assume the operator are actually abiding, so no chemicals should be entering the soil²⁹. This simplified the calculations to just water and air phases.

Residence time $\left(\tau \right)$ was required to find the duration the chemicals existed in the pits, which can then be used to calculate the evaporative and water-based losses. Shale and CBM each needed a residence time calculation. The pits were assumed to be at capacity, meaning flowrate in equaled flowrate out, because this would maximize profits for the operators.

$$\tau =\frac{{V}_{p}}{{Q}_{out}}=\frac{{V}_{p}}{{Q}_{in}}=\frac{{V}_{p} {t}_{I}}{f {V}_{HFF}}$$

14

Where: ${V}_{p}$ is the volume of the pit; ${t}_{I}$ is the time required to inject all the fracking fluid; $f$ is the fraction of fracking fluid that returns as flowback; ${V}_{HFF}$ is the volume of the injected fluid. Inputting the county-specific values into Eq. 14 yields a residence time of 30.3 and 41.2 days for Weld and Las Animas counties, respectively, Table 2.

Table 2

Residence times and required values by county
	Weld County	Las Animas County
${V}_{p} \left({m}^{3}\right)$	2775.03	389.95
${t}_{I} \left(days\right)$	14	19
$f (-)$	0.175	0.61
${V}_{HFF}\left({m}^{3}\right)$	7332.79	295.09
$\tau \left(days\right)$	30.3	41.2

To simply calculate the evaporative masses, we used the equation derived by Mackay et al. correlating chemical evaporation rates with vapor pressure³⁰. This equation was developed for only organic chemicals, but USEtox assumes all inorganic chemicals to have a vapor pressure of $1\times {10}^{-20}$, so inorganic molecules evaporation can be assumed to be insignificant. The equation from Mackay et al. ³⁰ was adjusted to fit the conditions of our assessment, yielding Eq. 15.

$${E}_{mass}\left[kg\right]=3.67\times {10}^{-5} {p}_{v} MW {A}_{s,i} {\tau }_{i}$$

15

Where: subscript $i$ denotes the county; ${A}_{s} \left(\frac{{V}_{p}}{depth}\right)$ is the surface area of the pit

Table 3

Values required for calculating evaporative masses by county
	Weld County	Las Animas County
${A}_{s} \left({m}^{3}\right)$	1041.2	212.2
$\tau \left(days\right)$	30.3	41.2
${p}_{v} \left(Pa\right)$	chemical-specific
$MW \left(\raisebox{1ex}{$g$}\!\left/ \!\raisebox{-1ex}{$mol$}\right.\right)$	chemical-specific

Applying the values in Table 3 to Eq. 15 we generated the county-specific formulae that were used for each chemical to calculate the mass evaporated into the atmosphere: Equations 16 and 17 for Weld and Las Animas counties, respectively.

$${E}_{mass,W}\left[kg\right]=1.11 {p}_{v} MW$$

16

$${E}_{mass,LA}\left[kg\right]=0.31 {p}_{v} MW$$

17

Any mass that did not evaporate was assumed to remain in the pit in the chemical’s aqueous-phase, estimated by Eqs. 18 and 19 respectively. Considering the mass of each chemical lost from the fraction of fluid that remains underground, the formulae for the chemical mass that remains in the pit are shown in equations 18 and 19.

$${W}_{mass,W} \left[kg\right]=f{m}_{x}-{E}_{mass,W}=f{m}_{x}-1.11 {p}_{v,x} M{W}_{x}$$

18

$${W}_{mass,LA} \left[kg\right]=f{m}_{x}-{E}_{mass,LA}=f{m}_{x}-0.31 {p}_{v,x} M{W}_{x}$$

19

Where subscript $x$ denotes the chemical; $m$ is the chemical mass.

2.3 Characterization factors and impact assessment

Once all the required properties and data were estimated, we utilized the USEtox model to generate characterization factors of the chemicals. USEtox is a multimedia fate and transport model based on matrix formulation that allows prediction of overall characterization factors as a product of fate factors (FF), exposure factors (XF), and effect factors (EF). The fate factor is equal to the residence time (in days) of a chemical in an environmental medium. This residence time depends on the properties of the chemical as well as the medium from which the chemical is released and the medium that receives the chemical. This parameter accounts for the removal and intermediate transport processes in the environment. The exposure factors depend on different partition coefficients of a given chemical. For example, freshwater ecotoxicity represents the fraction of a chemical dissolved in freshwater medium which in turn depends on the partition coefficient between freshwater and other medium such as air and soil. The effect factor relates to the inherent toxicity of the chemical of interest and is estimated as a relationship between the concentration of a chemical in each medium and the potentially affected fraction of aquatic organisms.

Combining the specific CFs generated to their chemical masses in the air and water phases with energy output and quantity of wells assessed yields the average well impact per unit energy generated from chemicals for the average well in each county with units of (PDF.m3.day)/MJ or more simply, comparative toxicity units per MJ (CTU/MJ).

$$Impact per average well per unit energy=\frac{\sum \left(C{F}_{RuralAir}{E}_{mass}+C{F}_{Freshwater}{W}_{mass}\right)}{\sum \left(M{J}_{i}\right)\times {n}_{i}}$$

20

Where: $C{F}_{x}$ is the categorization factor; subscript $x$ denotes the environmental compartment; ${n}_{i}$ is the quantity of wells; $M{J}_{i}$ is the total energy output of all well; and subscript $i$ denotes the county

Table 4 depicts the average fracking fluid composition for the selected shale oil and gas wells in Weld County, and Table 5 shows the same for the selected coal-bed methane wells in Las Animas County. Using FracFocus, the selected wells were identified, their chemical data extracted, and the values converted from ‘maximum ingredient concentration in HF fluid (% by mass)’ into mass (kg). Each chemical’s chemical abstract service (CAS) number and purpose were extracted to aid in identifying ambiguously name chemicals and to group them into their categories, respectively. The chemical purpose was usually identical or similar to existing selected chemical categories, but when they were not we either made educated guesses as to where the chemical should be placed or put that chemical into the unknown category. For some chemicals that were unknown the exact chemical was listed under one known category for another well; all instances of these chemicals were then categorized under the known category. Hydrochloric acid was listed as unknown in one instance but was categorized as an acid/solvent under all others, so the unknown occurrence was assumed to be acid. For isopropanol and other similar chemicals, unknown was left unchanged since the category could not be assumed with any accuracy. For each instance of chemical was identified in the FracFocus data sheets an associated category under which it falls was listed, but some chemicals had multiple categories; isopropanol was one such versatile chemical: biocide, corrosion inhibitor, emulsifier, non-emulsifier, unknown, and surfactant.

Category	Average Mass in Fracking Fluid
Category	(kg)	(%)
Base Fluid	7,331,793.469	98.50920%	86.63257%
Proppant	1,005,130.115	98.50920%	11.87663%
Friction Reducer	40,820.972	1.49080%	0.48234%
Gel	26,660.891		0.31503%
Surfactant	17,554.821		0.20743%
Unknown	10,677.651		0.12617%
Breaker	6,958.211		0.08222%
Crosslinker	6,599.517		0.07798%
Buffer	6,152.945		0.07270%
Acid/Solvent	4,797.758		0.05669%
Biocide	1,580.297		0.01867%
Clay stabilizer	1,323.675		0.01564%
Additive	1,292.660		0.01527%
Corrosion Inhibitor	1,260.027		0.01489%
Activator	309.634		0.00366%
Emulsifier	120.082		0.00142%
Non-Emulsifier	45.185		0.00053%
Concentrate	13.161		0.00016%
Iron control	0.480		0.00001%
Total	8,463,092	100%

Table 4 Average hydraulic fracturing fluid for shale oil and gas wells by category

Category	Average Mass in Fracking Fluid
Category	(kg)	(%)
Base Fluid	295,090.655	60.27667%
Proppant	122,362.348	24.99434%
Foaming Agent	16,579.404	14.56234%
Gel	445.789	0.16665%	0.09106%
Acid	226.885		0.06025%
Unknown	33.126		0.01150%
Breaker	7.372		0.00361%
Corrosion Inhibitor	0.548		0.00009%
Scale Inhibitor	0.301		0.00007%
Iron Control	1.631		0.00007%
Total	434,748	100%

Table 5 Average hydraulic fracturing fluid for CBM wells by category

184 unique chemicals were identified in all 50 wells, 182 not including water (base fluid) and sand (proppant). 31 chemicals appeared in the 10 coal-bed methane wells, and 170 chemicals were found in the 40 wells from shale. Of the 182 chemicals: 12 were unique to CBM, 170 were unique to shale, and 17 were shared by both. 13 other wells were looked at on a preliminary basis after the initial 50 were analyzed, but it was decided to omit them and focus on the initial group because they would have only added 10 new chemicals to the already substantial 182 chemicals.

The two largest differences between the shale and coal-bed methane wells were the total mass of each fluid and their composition. The fluid injected in shale was 19.5 times more than for coal-bed. Shale wells used more chemicals per fracturing job, averaging 30 versus 15 in coal-bed deposits. In addition to chemical quantity, the relative uses of each chemical in relation to the entire fluid masses were greater with shale, with chemicals representing 1.49% for shale and 0.17% for CBM of the entire fracking fluid: a difference of 8.9 times. CBM included a large mass of a foaming agent (99.7% Nitrogen by mass), which represented 14.6% of the entire mass of the CBM fracking fluid. CBM proppant was also used in relatively larger quantities, compensated by a lower base fluid use.

To accurately compare shale with coal-bed methane the results needed to be normalized in relation to a functional unit; the primary function of oil and gas extraction operations is to provide energy in the form of fuel, so we selected the functional unit as megajoules (MJ). The total lifetime output of each well was required to find the energy output for the average well in each county. The data was sourced from the Colorado Oil and Gas Conservation Commission.

Unfortunately, nine of the 40 wells in Weld County and only one of the ten wells in Las Animas County were finished production. We built a model to estimate the total output of each well. Regardless of when their first year of operation was, all wells were standardized by production year, so that according to the model all wells began operating in the same year. Weld and Las Animas counties were modeled separately; then within each county, produced water, oil, and gas, were also analyzed separately.

The average annual growth rate for the empirical data was applied to the unknown data. If the average growth rate in energy output grew by 5% between production years one and two, then wells with only data for the first year (i.e. they began production one year ago) had their second-year output increase by 5% over their first year. The lifetime of the wells was assumed to be the same as the longest producing well – 17 years for Weld County, and 11 years for Las Animas County.

Table 6

Energy and Produced Water Output
		Weld County		Las Animas County
		Produced Water (gals)	Energy (MJ)	Produced Water (gals)	Energy (MJ)
Empirical Data	Average	725,630	449,149,939	18,588,961	484,653,113
	Total	29,025,192	21,586,484,011	185,889,606	4,846,531,134
	Annual Change	100.52%	101.05%	107.90%	138.40%
Data Filling	Average	1,152,521	1,034,841,937	24,719,357	1,493,937,492
	Total	46,100,851	41,393,677,463	247,193,569	14,939,374,916
	Annual Change	102.38%	102.32%	102.32%	140.01%
Data Derived from Model		37.04%	47.62%	24.80%	67.56%

CBM wells generate more water than shale, so produced water output was also analyzed, but in terms of gallons^3,31. Weld County averaged 1,152,521 and Las Animas averaged 24,719,357 gallons per well of produced water, a ratio of 21.5 times more produced water was generated for CBM wells than regular oil and gas wells. Weld County wells generated an average of 1,034,841,937 MJ of energy; Las Animas wells generated 1,493,937,492, 1.44 times greater. These findings were on par with the literature – Weld County has a rate of 0.0011 gal/MJ compared to 0.0165 gal/MJ for Las Animas. Las Animas generates 14.9 times more produced water for the same energy output.

Table 7 shows all the chemicals with new characterization factors, as well as the five chemicals (in orange) that already existed in USEtox but not in TRACI; chemicals in green are where chemicals with substitutes that were close or identical. The column in blue are the CFs – Rural Air, Continental and Freshwater, Continental – used in our assessment. Eq. 20, above, yields the average impact per average well for shale versus coal-bed methane with the resulting impacts compared in Table 8.

We expected that HF fluids used in CBM wells would have a lower ecological impact than O&G wells due to the use of fewer chemicals, lower quantity of chemicals used, and higher average energy output per well. However, our analysis reveals that, on average, the impact associated with CBM wells was 12.6% greater than that of O&G wells. However, the results are highly sensitive to energy production per well, residence time in ponds, surface area, and especially flowback fraction, as shown in Table 9. Flowback fraction is highly variable between different geological regions and specific wells. A slight change in the flowback fraction for either shale or CBM generates significant changes in the impact. In fact, if the average flowback fraction for Weld County increased to just 19.7% from 17.5%, the impacts would be equal. Conversely, if the flowback fraction dropped to 54.2% for Las Animas County, the impacts would equate. For CBM the 61% flowback fraction was derived using a 19-day injection duration, whereas for shale the duration was 10 to 14-day. Flowback increases with time, so with equal injection duration shale might have a larger impact than CBM. A higher flowback fraction meant more chemicals were removed from the well. Any chemicals remaining underground were omitted from the impact assessment.

The impacts from the chemicals’ evaporative mass were small for shale and negligible for CBM wells, so most of the attention for reducing the chemical impact should be focused on the mass of chemicals remaining in the water. Pits can be used as temporary storage for the wastewater emanating from the well as well as a conduit for treatment through evaporation. Our findings suggest that mass loss from evaporation is negligible compared to the mass that remains behind in the pit, so evaporation has a negligible impact on environmental impact reductions; environmental impacts from evaporative masses were 1.13% that of the impact of chemicals in the water for shale and 0.007% for CBM.

Efforts to reduce impact should be focused on using chemicals with lower ecotoxicity values, reducing spills, and using closed-systems storage containers on a more widespread basis. There are interchangeable chemicals that could be used for the same function (i.e. category), so by identifying these chemicals, determining their impact, and then using them in place of more ecotoxic chemicals would be the easiest and most passive option because once the chemicals are chosen nothing extra would be required. This assessment should be carried out with many more chemicals to identify possible alternatives. Spills are a large pathway for environmental contamination; they occur during truck accidents during transport and at the well. Specific statistics do not exist for the accidents involving fracking fluid transportation trucks, but according to one study there are between 5,000 and 7,800 truck accidents a year spilling an average of 40% of their load, representing 0.6% of all truck trips³². At the disposal site, often underground injection wells class II, a 2% leak rate was reported³². The majority of wastewater treatment plants are not designed for removing some of the constituents in produced water so water sent to a treatment plant might not get treated to acceptable levels. Often times in these cases the effluent is still above the background or desired levels³³. A closed-system greatly reduces the opportunity chemicals have to impact the environment, but closed-systems are more complicated and tend to have more problems than open pits. This assessment also did not include the potential impact of the majority of chemicals that remain underground.

From strictly a chemical standpoint, shale and CBM are relatively close. When including other avenues for environmental impact these impacts might become more pronounced. Since CBM only produced methane, a cleaner burning fuel than petroleum, it is expected to have a lower greenhouse gas footprint, so a more detailed life-cycle analysis would be required that focusses around the hydraulic fracturing chemicals to holistically determine the impact from chemicals. Further assessments should aim at looking at the upstream chemical impacts as well as other impact categories. The methodology used in this study will work, albeit slightly altered for specific impact categories. The standardized database could easily be expanded to include all necessary information and to expand the scope of the chemical impact assessment, but this study provides a foundation to do that.

Table 7

New endpoint ecotoxicity categorization factors
Chemical	CAS Number	Endpoint Ecotoxicity Characterization Factors [PDF.m³.day/kg_emitted]
Chemical	CAS Number	Household Indoor Air	Industrial Indoor Air	Urban Air	Rural Air, Continental	Freshwater, Continental	Sea Water, Continental	Natural Soil, Continental	Agricultural Air, Continental
Average Categorization Factor for All Chemicals		1.98E + 04	2.21E + 04	2.36E + 04	1.61E + 04	6.08E + 05	2.09E + 02	4.30E + 03	4.30E + 03
1-(benzyl)quinolinium chloride	15619-48-4	4.41E-02	4.41E-02	4.41E-02	4.41E-02	1.30E-01	3.12E-10	6.52E-02	6.52E-02
2,3-Dihydroxypropyl-trimethylammonium chloride	34004-36-0	6.47E-03	6.48E-03	6.49E-03	6.45E-03	2.22E-02	1.85E-16	9.33E-03	9.33E-03
2-bromo-3-nitrilopropionamide	1113-55-9	6.19E-01	6.20E-01	6.21E-01	6.16E-01	2.21E + 00	2.28E-09	9.00E-01	9.00E-01
3-chloro-2-hydroxypropyl-trimethylazanium;chloride	3327-22-8	1.15E + 02	1.15E + 02	1.15E + 02	1.15E + 02	3.95E + 02	3.06E-11	1.66E + 02	1.66E + 02
4-nonylphenyl	127087-87-0	4.07E + 04	4.21E + 04	4.30E + 04	3.84E + 04	4.59E + 05	3.24E-06	4.54E + 04	4.54E + 04
Alcohol amine		2.28E + 00	2.29E + 00	2.29E + 00	2.27E + 00	8.07E + 00	4.88E-09	3.28E + 00	3.28E + 00
Aldehyde		2.79E + 02	2.80E + 02	2.81E + 02	2.77E + 02	1.07E + 03	4.81E-06	3.98E + 02	3.98E + 02
Alkoxylated amine		1.25E + 06	1.41E + 06	1.52E + 06	9.72E + 05	4.35E + 07	7.01E-05	8.72E + 03	8.72E + 03
Alkyl dimethyl benzyl ammonium chloride	68424-85-1	1.01E + 02	1.09E + 02	1.14E + 02	8.80E + 01	2.19E + 03	6.18E-06	6.22E + 01	6.22E + 01
Alkyl pyridine benzyl quaternary ammonium chloride	68909-18-2	5.73E-02	5.73E-02	5.73E-02	5.73E-02	1.71E-01	1.39E-10	8.46E-02	8.46E-02
Amide		1.11E + 02	1.22E + 02	1.29E + 02	9.38E + 01	2.88E + 03	2.04E-02	4.48E + 01	4.48E + 01
Amines, coco alkyl, ethoxylated	61791-14-8	1.34E + 03	1.37E + 03	1.40E + 03	1.27E + 03	1.30E + 04	2.94E-09	1.55E + 03	1.55E + 03
Amines, tallow alkyl, ethoxylated	61791-26-2	1.27E + 02	1.41E + 02	1.51E + 02	1.02E + 02	3.93E + 03	1.28E-09	2.01E + 01	2.01E + 01
Ammonium persulfate	7727-54-0	4.80E + 02	4.80E + 02	4.79E + 02	4.81E + 02	1.32E + 03	7.88E-13	7.61E + 02	7.61E + 02
Ammonium phosphite	13446-12-3	1.07E + 02	1.07E + 02	1.07E + 02	1.08E + 02	2.96E + 02	2.36E-13	1.70E + 02	1.70E + 02
Aromatic aldehyde		5.17E + 01	5.27E + 01	5.34E + 01	4.99E + 01	4.10E + 02	3.28E + 00	6.65E + 01	6.65E + 01
Chlorous acid, sodium salt	7758-19-2	1.00E-01	1.00E-01	1.00E-01	1.00E-01	2.76E-01	2.15E-16	1.59E-01	1.59E-01
Disodium ethylene diaminediacetate	38011-25-5	8.02E-06	8.06E-06	8.08E-06	7.95E-06	3.37E-05	1.38E-18	1.13E-05	1.13E-05
EDTA/copper chelate	14025-15-1	2.75E + 00	2.77E + 00	2.78E + 00	2.71E + 00	1.37E + 01	2.41E-22	3.78E + 00	3.78E + 00
EO-C7-9-iso, C8 rich-alcohols	78330-19-5	3.36E + 02	3.36E + 02	3.37E + 02	3.36E + 02	1.08E + 03	3.18E-08	4.94E + 02	4.94E + 02
EO-C9-11-iso, C10-rich alcohols	78330-20-8	1.57E + 01	1.71E + 01	1.81E + 01	1.33E + 01	3.96E + 02	2.78E + 00	7.32E + 00	7.32E + 00
Ethoxylated alcohol		2.35E + 02	2.36E + 02	2.37E + 02	2.33E + 02	9.88E + 02	9.45E-09	3.32E + 02	3.32E + 02
Ethoxylated amine		1.27E + 02	1.41E + 02	1.51E + 02	1.02E + 02	3.93E + 03	1.28E-09	2.01E + 01	2.01E + 01
Ethoxylated branched C13 alcohol	78330-21-9	1.22E + 02	1.33E + 02	1.41E + 02	1.02E + 02	3.22E + 03	1.87E + 01	4.90E + 01	4.90E + 01
Ethoxylated decyl alcohol		2.61E + 01	2.84E + 01	2.99E + 01	2.23E + 01	6.35E + 02	9.06E-10	1.23E + 01	1.23E + 01
Ethoxylated fatty acid		1.27E + 02	1.41E + 02	1.51E + 02	1.02E + 02	3.93E + 03	1.28E-09	2.01E + 01	2.01E + 01
Fatty acid tall oil amide		1.49E + 04	1.67E + 04	1.80E + 04	1.18E + 04	4.96E + 05	2.46E + 02	6.08E + 02	6.08E + 02
Fatty acids, tall oil	61790-12-3	1.49E + 04	1.67E + 04	1.80E + 04	1.18E + 04	4.96E + 05	2.46E + 02	6.08E + 02	6.08E + 02
Formaldehyde amine resin	56652-26-7	1.34E + 01	1.40E + 01	1.45E + 01	1.24E + 01	1.93E + 02	1.43E-03	1.26E + 01	1.26E + 01
Formaldehyde;2-methyloxirane;(1E,3E)-4,5,5-trimethylhexa-1,3-dien-1-ol	29316-47-0	4.40E + 02	4.48E + 02	4.54E + 02	4.27E + 02	3.26E + 03	4.46E-01	5.67E + 02	5.67E + 02
Guar gum	9000-30-0	3.67E + 03	3.67E + 03	3.67E + 03	3.68E + 03	1.04E + 04	5.74E-35	5.50E + 03	5.50E + 03
Guar gum derivative		7.61E + 00	7.60E + 00	7.60E + 00	7.62E + 00	2.17E + 01	7.37E-13	1.14E + 01	1.14E + 01
Heavy aliphatic petroleum naphtha solvent	64742-96-7	1.40E + 00	1.52E + 00	1.61E + 00	1.19E + 00	3.45E + 01	9.48E-02	7.08E-01	7.08E-01
Heavy aromatic petroleum naphtha	64742-94-5	3.09E + 01	3.28E + 01	3.41E + 01	2.76E + 01	5.70E + 02	9.98E + 00	2.39E + 01	2.39E + 01
Heavy hydrotreated petroleum naphtha	64742-48-9	1.26E + 02	1.36E + 02	1.43E + 02	1.09E + 02	2.80E + 03	1.50E + 01	7.80E + 01	7.80E + 01
Hydrotreated light petroleum distillate	64742-47-8	3.80E + 02	4.10E + 02	4.31E + 02	3.29E + 02	8.59E + 03	3.84E + 01	2.29E + 02	2.29E + 02
Hydrotreated medium petroleum distillates	64742-46-7	1.13E + 06	1.25E + 06	1.33E + 06	9.23E + 05	3.31E + 07	2.19E + 04	3.22E + 05	3.22E + 05
Inorganic salt		3.99E + 03	3.99E + 03	3.99E + 03	4.00E + 03	1.10E + 04	1.03E-11	6.33E + 03	6.33E + 03
Isopropanol	67-63-0	4.99E-02	5.26E-02	5.44E-02	4.55E-02	7.93E-01	1.19E-02	4.59E-02	4.59E-02
Isotridecanol, ethoxylated (TDA-6)	9043-30-5	4.45E + 01	4.53E + 01	4.58E + 01	4.31E + 01	3.26E + 02	1.58E-08	5.73E + 01	5.73E + 01
Lactic acid	50-21-5	2.43E-03	2.49E-03	2.52E-03	2.34E-03	2.07E-02	2.98E-10	2.94E-03	2.94E-03
Laury alcohol ethoxylate	68551-12-2	8.74E + 02	9.64E + 02	1.02E + 03	7.24E + 02	2.45E + 04	1.39E + 02	2.88E + 02	2.88E + 02
Light aromatic petroleum naphtha solvent	64742-95-6	3.97E + 01	4.22E + 01	4.39E + 01	3.55E + 01	7.32E + 02	1.28E + 01	3.07E + 01	3.07E + 01
Naphthenic acid ethoxylate	68410-62-8	6.36E + 01	6.37E + 01	6.38E + 01	6.35E + 01	2.12E + 02	1.63E-08	9.39E + 01	9.39E + 01
Nitrilotriacetate, trisodium salt (NTA)	5064-31-3	1.23E-02	1.24E-02	1.24E-02	1.22E-02	5.59E-02	5.85E-18	1.71E-02	1.71E-02
Olefin	64743-02-8	3.26E + 02	3.56E + 02	3.77E + 02	2.74E + 02	8.46E + 03	5.14E + 01	1.42E + 02	1.42E + 02
Oxirane, 2-methyl-, polymer with oxirane, monodecyl ether	37251-67-5	4.21E + 02	4.47E + 02	4.64E + 02	3.78E + 02	7.57E + 03	1.06E-02	3.26E + 02	3.26E + 02
Phenol/formaldehyde resin	9003-35-4	1.34E + 01	1.40E + 01	1.45E + 01	1.24E + 01	1.93E + 02	1.43E-03	1.26E + 01	1.26E + 01
Poly(oxy-1,2-ethanediyl),.alpha.-tetradecyl-.omega.-hydroxy	27306-79-2	4.50E + 03	5.11E + 03	5.51E + 03	3.50E + 03	1.61E + 05	1.15E + 01	1.13E + 01	1.13E + 01
Poly(tetrafluoroethylene)	9002-84-0	3.67E-01	3.87E-01	4.00E-01	3.34E-01	5.95E + 00	7.95E-02	3.32E-01	3.32E-01
Polyacrylate	79-10-7	4.96E-01	4.99E-01	5.02E-01	4.89E-01	2.43E + 00	4.15E-04	6.85E-01	6.85E-01
Polyethylene glycol	25322-68-3	2.11E + 02	2.12E + 02	2.12E + 02	2.10E + 02	7.26E + 02	4.18E-09	3.04E + 02	3.04E + 02
Polyoxyalkylenes	68951-67-7	1.90E + 03	2.10E + 03	2.24E + 03	1.55E + 03	5.59E + 04	1.97E + 02	4.73E + 02	4.73E + 02
Polysaccharide	68130-15-4	2.07E-05	2.11E-05	2.14E-05	2.00E-05	1.62E-04	1.32E-32	2.56E-05	2.56E-05
Potassium carbonate	584-08-7	1.15E-02	1.15E-02	1.15E-02	1.15E-02	3.95E-02	1.08E-22	1.66E-02	1.66E-02
Potassium hydroxide	1310-58-3	7.61E-03	7.60E-03	7.60E-03	7.62E-03	2.09E-02	1.99E-17	1.21E-02	1.21E-02
Potassium persulfate	7727-21-1	1.45E + 00	1.45E + 00	1.45E + 00	1.45E + 00	3.99E + 00	2.25E-15	2.30E + 00	2.30E + 00
Proprietary sesquiolate	8007-43-0	2.79E-02	2.75E-02	2.73E-02	2.84E-02	1.58E-03	2.07E-16	5.72E-08	5.72E-08
Quaternary ammonium compound	122-18-9	4.20E + 02	4.66E + 02	4.96E + 02	3.44E + 02	1.24E + 04	4.12E-05	1.06E + 02	1.06E + 02
Quaternary ammonium salt		4.20E + 02	4.66E + 02	4.96E + 02	3.44E + 02	1.24E + 04	4.12E-05	1.06E + 02	1.06E + 02
Silica, amorphous - fumed	7631-86-9	1.59E + 02	1.59E + 02	1.59E + 02	1.59E + 02	4.37E + 02	4.07E-13	2.52E + 02	2.52E + 02
Sodium bicarbonate	144-55-8	2.68E-02	2.68E-02	2.68E-02	2.67E-02	9.27E-02	5.54E-07	3.85E-02	3.88E-02
Sodium bromide	7647-15-6	1.45E + 02	1.45E + 02	1.45E + 02	1.45E + 02	3.99E + 02	3.10E-13	2.30E + 02	2.30E + 02
Sodium chloride	7647-14-5	3.99E + 03	3.99E + 03	3.99E + 03	4.00E + 03	1.10E + 04	1.03E-11	6.33E + 03	6.33E + 03
Sodium erythorbate	6381-77-7	1.27E + 00	1.43E + 00	1.54E + 00	9.93E-01	4.40E + 01	1.53E-21	5.88E-03	5.88E-03
Sodium hydroxide	1310-73-2	3.10E + 04	3.10E + 04	3.09E + 04	3.10E + 04	8.53E + 04	9.09E-11	4.91E + 04	4.91E + 04
Sodium hydroxyacetate	2836-32-0	7.49E-03	7.65E-03	7.76E-03	7.22E-03	6.12E-02	2.12E-08	9.19E-03	9.19E-03
Sodium hypochlorite	7681-52-9	1.03E + 01	1.03E + 01	1.02E + 01	1.03E + 01	2.82E + 01	2.30E-14	1.63E + 01	1.63E + 01
Sodium iodide	7681-82-5	1.63E + 02	1.63E + 02	1.62E + 02	1.63E + 02	4.48E + 02	3.07E-13	2.58E + 02	2.58E + 02
Sodium lactate	72-17-3	5.94E-02	5.97E-02	6.00E-02	5.89E-02	2.55E-01	1.31E-05	8.35E-02	8.35E-02
Sodium persulfate	7775-27-1	3.10E + 00	3.10E + 00	3.09E + 00	3.11E + 00	8.53E + 00	5.02E-15	4.91E + 00	4.91E + 00
Sodium sulfate	7757-82-6	3.73E + 00	3.72E + 00	3.72E + 00	3.73E + 00	1.03E + 01	7.12E-15	5.90E + 00	5.90E + 00
Sodium;prop-2-enamide;prop-2-enoate;prop-2-enoic acid	62649-23-4	3.31E + 03	3.32E + 03	3.32E + 03	3.29E + 03	1.19E + 04	5.03E-06	4.75E + 03	4.75E + 03
Sorbitan monooleate polyoxyethylene derivative	9005-65-6	1.82E + 04	1.82E + 04	1.82E + 04	1.83E + 04	5.02E + 04	2.25E-31	2.80E + 04	2.80E + 04
Sorbitan, mono-9-octadecenoate, (Z)	1338-43-8	3.56E + 03	3.95E + 03	4.22E + 03	2.91E + 03	1.06E + 05	8.31E-06	7.79E + 02	7.79E + 02
Soybean oil methyl ester	67784-80-9	4.18E + 04	4.67E + 04	4.99E + 04	3.36E + 04	1.31E + 06	4.46E + 03	6.54E + 03	6.54E + 03
Styrene acrylic copolymer	25085-34-1	6.10E + 00	6.28E + 00	6.40E + 00	5.80E + 00	6.26E + 01	2.72E-04	6.93E + 00	6.93E + 00
Sucrose	57-50-1	1.33E-02	1.33E-02	1.34E-02	1.32E-02	5.36E-02	1.35E-21	1.87E-02	1.87E-02
Surfactants		4.99E-02	5.26E-02	5.44E-02	4.55E-02	7.93E-01	1.19E-02	4.59E-02	4.59E-02
Tall oil acid diethanolamide	68155-20-4	3.04E + 03	3.37E + 03	3.59E + 03	2.49E + 03	8.92E + 04	5.62E-05	7.19E + 02	7.19E + 02
Terpenes and terpenoids	68956-56-9	3.38E-01	3.71E-01	3.93E-01	2.83E-01	9.03E + 00	2.25E-13	1.23E-01	1.23E-01
Terpenes and terpenoids, sweet orange-oil	68647-72-3	2.34E + 02	2.50E + 02	2.61E + 02	2.06E + 02	4.74E + 03	4.60E + 01	1.65E + 02	1.65E + 02
Thiourea polymer	68527-49-1	5.79E-01	5.79E-01	5.79E-01	5.79E-01	1.71E + 00	4.29E-08	8.56E-01	8.56E-01
Triisopropanolamine	122-20-3	1.68E + 00	1.81E + 00	1.89E + 00	1.47E + 00	3.57E + 01	1.21E-08	1.04E + 00	1.14E + 00
Trisodium ethylenediaminetriacetate	19019-43-3	3.42E-02	3.44E-02	3.45E-02	3.39E-02	1.44E-01	1.73E-18	4.80E-02	4.80E-02
Vinylidene chloride-methyl acrylate copolymer	25038-72-6	1.53E + 02	1.57E + 02	1.60E + 02	1.46E + 02	1.46E + 03	2.91E + 01	1.85E + 02	1.85E + 02
Zirconium, acetate lactate oxo ammonium complexes	68909-34-2	1.58E-02	1.58E-02	1.58E-02	1.58E-02	4.66E-02	7.14E-14	2.34E-02	2.34E-02

Table 8

Summary of chemical impact per MJ for shale vs. CBM
		Weld County, Regular Oil and Gas Wells		Las Animas County, Coal-Bed Methane Wells		Average Ratio (LA/W)
		Average per Chemical	Average per Well	Average per Chemical	Average per Well	Average Ratio (LA/W)
Mass Removed from Well	(kg)	5.7E + 3	21.5E + 3	15.2E + 3	44.0E + 3	204.24%
Evaporative Mass	(kg)	854.3E + 0	2.8E + 3	26.2E + 0	62.9E + 0	2.28%
Mass in Water	(kg)	5.5E + 3	17.8E + 3	18.3E + 3	43.9E + 3	246.04%
Emission to Rural Air	(CTU)	703.9E + 3	2.2E + 6	229.0E + 0	549.5E + 0	0.02%
Emissions to Freshwater	(CTU)	63.5E + 6	198.4E + 6	3.4E + 6	8.2E + 6	4.11%
Total Air Impact per Unit Energy	(CTU/MJ)	17.0E-6	53.1E-6	153.3E-9	367.8E-9	0.69%
Total Water Impact per Unit Energy	(CTU/MJ)	1.5E-3	4.8E-3	2.3E-3	5.5E-3	113.87%
Average Impact per Unit Energy	(CTU/MJ)		4.857E-3		5.457E-3	112.63%

Table 9

Significant variables affecting results
		Weld County	Las Animas	Ratio (LA/W)
Energy Produced per Well	(MJ)	1.035E + 09	1.494E + 09	144%
Flowback Fraction	(%)	17.5%	61.0%	349%
Residence Time	(days)	30.3	41.2	136%
Pit Surface Area	(m²)	1041.2	212.2	20%

Author Contribution

David Zelinka (DZ): Data curation, Modeling, Analysis, Writing Original Draft.Arunprakash Karunanithi (AK): Conceptualization, Supervision, Reviewing and Editing.

Supporting Information

Full physio-chemical properties of all chemicals
Individual well fracking fluid composition
Complete categorization factors output from USEtox

Author Information

Corresponding Author

*Telephone: (303) 315-7477 | Email: arunprakash.karunanithiucdenver.edu

Stringfellow, W. T.; Domen, J. K.; Camarillo, M. K.; Sandelin, W. L.; Borglin, S. Physical, chemical, and biological characteristics of compounds used in hydraulic fracturing. Journal of Hazardous Materials 2014, 275, 37–54.
Higley, D. K., Pollastro, R.M., and Clayton, J.L., 1995, Denver Basin Province (039), in Gautier, D. L., Dolton, G.L., Takahashi, K.I., and Varnes, K.L., ed., 1995 National Assessment of United States oil and gas resources. (https://certmapper.cr.usgs.gov/data/noga95/prov39/text/prov39.pdf)
US. EPA, Evaluation of Impacts to Underground Sources of Drinking Water by Hydraulic Fracturing of Coalbed Methane Reservoirs; EPA 816-R-04-003, 2004. (https://fracfocus.org/sites/default/files/publications/evaluation_of_impacts_to_underground_sources_of_drinking_water_by
_hydraulic_fracturing_of_coalbed_methane_reservoirs.pdf)
FracFocus.org. Chemical Use http://fracfocus.org/chemical-use (accessed Oct 19, 2015).
US EPA. Assessment of the Potential Impacts of Hydraulic Fracturing for Oil and Gas on Drinking Water Resources (External Review Draft); EPA/600/R-15/047a, 2015. (https://www.epa.gov/sites/production/files/2015-06/documents/hf_es_erd_jun2015.pdf)
American Water Works Association. Water and Hydraulic Fracturing - A White Paper From the American Water Works Association; 2013. (http://www.awwa.org/Portals/0/files/legreg/documents/AWWAFrackingReport.pdf)
Hamm, K. Colorado oil and gas wells by the numbers, 2015. http://www.denverpost.com/datacenter/ci_27519246/colorado-oil-and-gas-wells-by-numbers (accessed Jan 10, 2016).
Boschee, P. Produced and flowback water recycling and reuse. Oil Gas Facilit. 2014, 3, 16.
Puri, R.; King, G. E.; Palmer, I. D. Damage to Coal Permeability During Hydraulic Fracturing. In Low Permeability Reservoirs Symposium; Society of Petroleum Engineers: Denver, 1991.
Mantell, M. E. Summary of the Technical Workshop on Water Acquisition Modeling : Assessing Impacts Through Modeling and Other Means. In Technical Workshop on Water Acquisition Modeling: Assessing Impacts Through Modeling and Other Means; 2013; p 100.
Clark, C. E.; Burnham, A.; Harto, C.; Horner, R. Hydraulic fracturing and shale gas production: Technology, Impacts, and Policy; Argonne National Laboratory, 2012. (http://38nhoe2hyq2v19k9rp1nlpf0.wpengine.netdna-cdn.com/wp-content/uploads/2012/11/Hydraulic-Fracturing-and-Shale-Gas-Production-Technology_Impacts_Polic-.pdf)
Jiang, M.; Hendrickson, C. T.; Vanbriesen, J. M. Life cycle water consumption and wastewater generation impacts of a Marcellus shale gas well. Environmental Science & Technology 2014, 48 (3), 1911–1920.
National Toxics Network. Toxic Chemicals in the Exploration and Production of Gas from Unconventional Sources. Bangalow, Australia, 2013. (http://www.ntn.org.au/wp/wp-content/uploads/2013/04/UCgas_report-April-2013.pdf)
U.S. House of Representatives Committee on Energy and Commerce. Chemicals Used in Hydraulic Fracturing; 2011.
Howarth, R. W.; Shindell, D.; Santoro, R.; Ingraffea, A.; Phillips, N.; Ownsend-Small, A. Methane Emissions from Natural Gas Systems. National Climate Assessment 2012, 1–6.
Howarth, R. W.; Santoro, R.; Ingraffea, A. Methane and the greenhouse-gas footprint of natural gas from shale formations. Climatic Change 2011, 106 (4), 679–690.
Howarth, R. W. A bridge to nowhere: methane emissions and the greenhouse gas footprint of natural gas. Energy Science & Engineering 2014, 2 (2), 47–60.
Howarth, R. W.; Santoro, R.; Ingraffea, A. Venting and leaking of methane from shale gas development: Response to Cathles et al. Climatic Change 2012, 113 (2), 537–549.
Stephenson, T.; Valle, J. E.; Riera-Palou, X. Modeling the Relative GHG Emissions of Conventional and Shale Gas Production. Environmental Science & Technology 2011, 45 (24), 10757–10764.
Vandecasteele, I.; Marí Rivero, I.; Sala, S.; Baranzelli, C.; Barranco, R.; Batelaan, O.; Lavalle, C. Impact of Shale Gas Development on Water Resources A Case Study in Northern Poland. Environmental Management 2015, 55 (6), 1285–1299.
Lloyd-Smith, M.; Senjen, R. Hydraulic Fracturing in Coal Seam Gas Mining: The Risks to Our Health, Communities, Environmentand Climate. National Toxics Network 2011, 1–37.
ToxNet.gov http://toxnet.nlm.nih.gov/.
PubChem.gov https://pubchem.ncbi.nlm.nih.gov/.
Chemicalize.org http://www.chemicalize.org/.
US Environmental Protection Agency. Estimation Programs Interface Suite^TM for Microsoft® Windows, v4.11. United States Environmental Protection Agency. Washington, DC, USA 2016.
Bare, J. TRACI 2.0: the tool for the reduction and assessment of chemical and other environmental impacts 2.0. Clean Technologies and Environmental Policy 2011, 13 (5), 687–696.
Mayo-Bean, K.; Moran-Bruce, K. Ecological Structure Activity Relationships Predictive Model. Washington, D.C., United States 2015.
Travis, C. C.; Arms, A. D. Bioconcentration of organics in beef, milk, and vegetation. Environmental Science & Technology 1988, 22 (3), 271–274.
Colorado Oil & Gas Conservation Commission. E&P Waste Management Series 900; Colorado Oil & Gas Conservation Commission, 2015; pp 1–24.
Mackay, D.; van Wesenbeeck, I. Correlation of Chemical Evaporation Rate with Vapor Pressure. Environmental Science & Technology 2014, 48 (17), 10259–10263.
Guerra, K. L.; Dahm, K. G.; Dundorf, S. Oil and Gas Produced Water Management and Beneficial Use in the Western United States. Science and Technology Program 2011, No. 157, 129.
Rozell, D. J.; Reaven, S. J. Water Pollution Risk Associated with Natural Gas Extraction from the Marcellus Shale. Risk Analysis 2012, 32 (8), 1382–1393.
Warner, N. R.; Christie, C. A.; Jackson, R. B.; Vengosh, A. Impacts of Shale Gas Wastewater Disposal on Water Quality in Western Pennsylvania. Environmental Science & Technology 2013, 47 (20), 11849–11857.

Supplementary Data is not available with this version

No competing interests reported.

A comparative ecotoxicity impact assessment of hydraulic fracturing chemicals used in oil and gas and coal-bed methane wells

Status:

Version 1

Abstract

Figures

1 Introduction

2 Methodology

2.1 Generating physio-chemical property data

2.2 Calculating aqueous-based and evaporative chemical masses

2.3 Characterization factors and impact assessment

3 Hydraulic Fracturing Fluids Analysis

4 Energy and Produced Water Analysis

5 Results and Discussion

Declarations

Author Contribution

References

Supplementary Data

Additional Declarations

Status:

Version 1

	Weld County	Las Animas County
\({V}_{p} \left({m}^{3}\right)\)	2775.03	389.95
\({t}_{I} \left(days\right)\)	14	19
\(f (-)\)	0.175	0.61
\({V}_{HFF}\left({m}^{3}\right)\)	7332.79	295.09
\(\tau \left(days\right)\)	30.3	41.2

	Weld County	Las Animas County
\({A}_{s} \left({m}^{3}\right)\)	1041.2	212.2
\(\tau \left(days\right)\)	30.3	41.2
\({p}_{v} \left(Pa\right)\)	chemical-specific
\(MW \left(\raisebox{1ex}{$g$}\!\left/ \!\raisebox{-1ex}{$mol$}\right.\right)\)	chemical-specific