Scenario analysis for assessing the impact of hydraulic fracturing on stream low flows using the SWAT model

Figures & data

Figure 1. Location map with climate and USGS stations in the Muskingum watershed. WETN: herbaceous wetland; WETF: wetland forest; WATR: open water; URMD: urban medium density; URLD: urban low density; URHD: urban high density; UIDU: industrial; SWRN: bare rock or sandy or clays; RNGE: grassland; RNGB: shrubland; HAY: hay; FRST: mixed forest; FRSE: evergreen forest; FRSD: deciduous forest; and AGRR: agriculture.

Figure 2. Utica shale wells in the State of Ohio, USA, from January 2011 to September 2013, indicating more wells in Eastern Ohio.

Figure 3. Tuscarawas sub-watershed (shaded) in the upper part of the Muskingum watershed, where drilling for hydraulic fracking has been increasing significantly.

Figure 4. Change in number of Utica wells in Ohio from January 2011 to May 2014.

Table 1. Data used for SWAT model development in the Muskingum watershed.

Data type	Data	Source
GIS	30-m DEM	USGS National Geospatial Program (NGP) http://viewer.nationalmap.gov/viewer/
	Land use and land cover 2006	USGS National Geospatial Program (NGP), National Land Cover Dataset (NLCD), http://viewer.nationalmap.gov/viewer/
	Soil data	Geographic STATSGO soil map (scale of 1:250 000)
Climate	Rainfall and temperature	NOAA National Climatic Data Center (NCDC) http://www.ncdc.noaa.gov/cdo-web/
Hydrology	Streamflow	USGS http://waterdata.usgs.gov/usa/nwis/sw
	Reservoir outflow	US Army Corps of Engineers (USACE)
Stream networks/water bodies	Stream and flow direction, reservoirs	USGS National Geospatial Program (NGP) National Hydrograph Dataset (NHD)
Water use (surface and ground water)	Irrigation, public, power, mineral extraction, industries and golf courses	Ohio Department of Natural Resources (ODNR), Ohio Environmental Protection Agency (OEPA)
Point sources	Flow discharge	OEPA
Oil and natural gas	Wells for hydraulic fracking	ODNR
	Sources of drilling water	ODNR
	Drilling water estimate per well and future drilling trend	FracFocus (http://fracfocus.org) (National hydraulic fracturing chemical registry)

Table 2. Water withdrawal for hydraulic fracking per well in the Muskingum watershed, 2011–2013.

Year	Average vertical depth (m)	Freshwater (m^3)	Recycled water (%)
2011	7 720	11 450	13.8
2012	7 730	13 010	4.4
2013	10 900	16 680	3.7

Table 3. List of predictors used in the SDSM model to downscale the NCEP re-analysis data.

Station parameter	Predictor variable
Precipitation	Zonal velocity component at surface level
	Zonal velocity component at 850 hpa
	Geostrophic airflow velocity at 850 hpa
	Vorticity at surface level
	Vorticity at 850 hpa
	Sea level pressure
	Specific humidity at 500 hpa
	Specific humidity at 850 hpa
Temperature	Mean temperature
	Near surface specific humidity

Table 4. Statistical criteria measuring the daily performance of the SWAT model for the calibration (2002–2009) and validation (1995–2001) periods.

USGS gauge station	Calibration			Validation
	R^2	NSE	PBIAS (%)	R^2	NSE	PBIAS (%)
3117000	0.42	0.42	0	0.45	0.45	−2
3124500	0.43	0.40	−16	0.53	0.47	−10
3139000	0.51	0.49	−4	0.63	0.60	−3
3136500	0.47	0.46	9	0.41	0.40	18
3129000	0.57	0.56	3	0.54	0.47	21
3140500	0.63	0.63	2	0.69	0.65	12
3146500	0.42	0.40	11	0.43	0.42	12
3142000	0.55	0.47	13	0.51	0.49	−2
3150000	0.65	0.65	0	No data

Table 5. Statistical criteria measuring the monthly performance of the watershed model for the calibration (2002–2009) and validation (1995–2001) periods.

USGS gauge station	Calibration			Validation
	R^2	NSE	PBIAS (%)	R^2	NSE	PBIAS (%)
3117000	0.89	0.89	0	0.90	0.86	7
3124500	0.59	0.53	−16	0.63	0.61	−9
3139000	0.64	0.64	−4	0.72	0.71	−3
3136500	0.50	0.49	9	0.63	0.56	18
3129000	0.68	0.66	3	0.67	0.55	21
3140500	0.68	0.67	1	0.76	0.69	13
3146500	0.79	0.72	11	0.76	0.71	12
3142000	0.71	0.69	13	0.77	0.68	−1
3150000	0.73	0.72	0	No data

Table 6. Characteristics of the various scenarios used for analysis.

Scenario	Climate conditions	Hydraulic fracking
Baseline (without fracking) scenario	Current	No
Current (with fracking) scenario	Current	Yes
Future scenario on current climate period	Current	Projected fracking
Future scenario on future climate period	Future	Projected fracking

Figure 5. The 7-day monthly minimum flow during the low-flow period for baseline and future scenarios analysed for the current climate period in the Muskingum watershed.

Figure 6. Percentage difference in 7-day minimum flows between the baseline and future hydraulic fracking scenarios simulated using current climate data in the Muskingum watershed.

Figure 7. Percentage difference in annual average flows between the baseline and future hydraulic fracking scenarios simulated using current climate data in the Muskingum watershed.

Figure 8. The 7Q10 flows for the baseline and future projected hydraulic fracking scenarios using the 30-year projected climate dataset generated by the SDSM.

Figure 9. Percentage difference in 7Q10 between the baseline and projected future hydraulic fracking scenarios using the 30-year projected climate dataset generated by the SDSM.

Figure 10. The 4B3 and 1B3 flows for the baseline vs future fracking scenarios using the 30-year climate dataset generated by the SDSM.

Figure 11. Percentage difference in 4B3 and 1B3 between the baseline and future hydraulic fracking scenarios using the 30-year climate dataset generated by the SDSM. The dotted line represents the trend in 4B3 and solid line represents the trend in 1B3.

Figure 12. Flow duration curve for the current and future hydraulic fracking scenarios using the 30-year climate dataset generated by the SDSM in the Muskingum watershed.