Numerical modelling of the impacts of water abstraction for hydraulic fracturing on groundwater–surface water interaction: a case study from northwestern Alberta, Canada

Figures & data

Figure 1. Location of the study area in Alberta, Canada.

Figure 2. Surface water monitoring stations in the study area. Only one groundwater monitoring well is shown here as a water abstraction location for hydraulic fracturing.

Table 1. Details of input data for the coupled MIKE-SHE and MIKE-11 model for the study area.

Data type	Data and format	Source
Watershed	Canadian Digital Elevation Data of 17.77 m grid	Natural Resources Canada
	Channel geometry	Digitizing digital elevation data and Google maps
	Land use/land cover of 30 m by 30 m grid for year 2000	Natural Resources Canada
	Soil	Alberta Environment and Parks
Hydrological	Daily streamflow from 2000 to 2012	Water Survey of Canada
	Groundwater level from 2000 to 2012	Alberta provincial groundwater monitoring wells database
Climate	Observed precipitation, temperature and reference evapotranspiration from 1986 to 2014	14 weather stations in the study area from Alberta Agroclimatic Information Service, and Alberta Environment and Sustainable Resource Development
Vegetation characteristics	Leaf area index and rooting depths of each land-use type	Task Committee on Hydrology Handbook (1996), Allen et al. (1998), Zeng (2001), Myneni et al. (2003), Kim et al. (2005), Wijesekara and Marceau (2009)

Table 2. Monthly number of hydraulically fractured wells and water use in hydraulic fracturing in 2013 and 2014.

	Number of wells		Water used (m^3)
Month	2013	2014	2013	2014
January	18	20	43,714	66,865
February	29	19	97,466	43,309
March	10	35	15,535	78,858
April	3	9	5249	56,039
May	7	13	67,317	91,354
June	9	25	27,857	109,275
July	20	13	64,243	53,732
August	19	22	94,626	68,282
September	15	25	37,067	102,478
October	22	25	72,300	99,122
November	18	20	408,756	81,356
December	16	21	63,161	98,261

Table 3. Different types of water abstraction scenarios for hydraulic fracturing activities.

Scenario	Water abstracted for hydraulic fracturing	Type of source water	Recycling of water
Baseline scenario	No hydraulic fracturing	-	No
Scenario 1	Water abstracted 1 km away from the SW3 station to the upstream direction from the Smoky River	Surface water (river) (100%)	No
Scenario 2	Water abstracted 1 km away from the SW3 station to the upstream direction from the Smoky River	Surface water (river) (85%)	Yes, 15%
Scenario 3	Water abstracted 1 km away from the SW3 station to the upstream direction from the Smoky River, and the GW1 well	Surface water (84%) and groundwater (16%)	No
Scenario 4	Water abstracted 1 km away from the SW3 station to the upstream direction from the Smoky River, and the GW1 well	Surface water (71.4%) and groundwater (13.6%)	Yes, 15%

Table 4. Performance statistics using observed and simulated streamflow, stream (i.e. river) water level and groundwater level data at various monitoring stations and wells during calibration and validation periods. R²: coefficient of determination; NSE: Nash-Sutcliffe efficiency criterion.

	Calibration			Validation
Monitoring station (measuring parameter)	R^2	NSE		R^2	NSE
SW1 (streamflow)	0.88		0.76	0.92	0.89
SW2 (river water level)	0.48		0.56	0.52	0.59
SW3 (streamflow)	0.92		0.63	0.9	0.75
SW3 (river water level)	0.91		0.71	0.86	0.65
GW1 (groundwater level)	0.71		0.69	0.75	0.81

Figure 3. Comparison of observed and simulated streamflow by the developed model at the SW1 station (outlet of the study area), and observed and simulated river water levels at the SW3 station during (a) calibration and (b) validation periods.

Figure 4. Comparison of observed and simulated groundwater levels by the developed model at the GW1 well during (a) calibration and (b) validation periods.

Figure 5. Increase in mean monthly groundwater contributions to streamflow under various types of water abstraction scenarios for hydraulic fracturing in (a) 2013 and (b) 2014 compared to the baseline scenario.

Table 5. Mean annual streamflow, groundwater discharge and surface runoff generated in the study area under the baseline scenario and various scenarios of water abstraction for hydraulic fracturing in 2013 and 2014. Values in parentheses are absolute changes between various scenarios of water abstraction for hydraulic fracturing and the baseline scenario, in which “–” indicates a decrease with respect to the baseline scenario. Note, there is no change in surface runoff.

	2013			2014
Scenario	Mean annual streamflow (m^3/d)	Mean annual groundwater discharge (m^3/d)	Mean annual surface runoff (m^3/d)	Mean annual streamflow (m^3/d)	Mean annual groundwater discharge (m^3/d)	Mean annual surface runoff (m^3/d)
Baseline	2,310,941	1,179,014	1,131,927	2,032,906	1,152,835	880,071
Scenario 1	2,307,744 (–3197)	1,179,006 (–8)	1,131,927	2,029,882 (–3024)	1,152,825 (–10)	880,071
Scenario 2	2,308,224 (–2717)	1,179,009 (–5)	1,131,927	2,030,335 (–2570)	1,152,829 (–6)	880,071
Scenario 3	2,308,255 (–2685)	1,178,826 (–188)	1,131,927	2,030,365 (–2540)	1,152,435 (–400)	880,071
Scenario 4	2,308,658 (–2283)	1,178,838 (–176)	1,131,927	2,030,746 (–2159)	1,152,449 (–386)	880,071

Table 6. Mean annual groundwater contribution to streamflow under the baseline scenario and various scenarios of water abstraction for hydraulic fracturing in 2013 and 2014. Values in parentheses are absolute changes between various scenarios of water abstraction in hydraulic fracturing and the baseline scenario, in which positive values indicate an increase with respect to the baseline scenario.

Year	Baseline scenario	Scenario 1	Scenario 2	Scenario 3	Scenario 4
2013	51.02%	51.09% (0.07%)	51.08% (0.06%)	51.07% (0.05%)	51.06% (0.04%)
2014	56.71%	56.79% (0.08%)	56.78% (0.07%)	56.76% (0.05%)	56.75% (0.04%)

Figure 6. Mean monthly river water level declines compared to the baseline scenario at the (a) SW3 and (b) SW2 stations under various types of water abstraction scenarios for hydraulic fracturing in 2013.

Table 7. Mean annual river water level declines compared to the baseline scenario at the SW3 station under various types of water abstraction scenarios for hydraulic fracturing in 2013 and 2014.

Year	Mean annual river water level decline (mm)
	Scenario 1	Scenario 2	Scenario 3	Scenario 4
2013	4.31	3.63	3.61	3.06
2014	4.57	3.88	3.86	3.29

Figure 7. Mean monthly river water level declines compared to the baseline scenario at the (a) SW3 and (b) SW2 stations under various types of water abstraction scenarios for hydraulic fracturing in 2014.

Table 8. Relative declines in environmental flow under various scenarios of water abstraction for hydraulic fracturing during low-flow periods in 2013 and 2014 with respect to the environmental flow under the baseline scenario.

Month	Scenario 1	Scenario 2	Scenario 3	Scenario 4
February 2013	1.66%	1.41%	1.4%	1.19%
November 2013	4.17%	3.54%	3.51%	3.0%
September 2014	1.58%	1.34%	1.32%	1.13%
October 2014	1.17%	1.05%	1%	0.9%

Figure 8. Mean monthly groundwater level drawdown compared to the baseline scenario at the GW1 well under various types of water abstraction scenarios for hydraulic fracturing in (a) 2013 and (b) 2014.

Table 9. Comparison of mean monthly river water level decline at the SW3 station and monthly water abstraction under Scenario 1 between 2013 and 2014.

Month	Mean monthly river water level decline under Scenario 1 (mm)		Difference in mean monthly river water level decline (mm) ^a	Monthly water abstraction from the Smoky River under Scenario 1 (m^3)		Difference in monthly water abstraction (m^3) ^b
	2013	2014		2013	2014
January	3.0	3.8	–0.8	43,714	66,865	–23,151
February	8.6	2.5	6.0	97,466	43,309	54,157
March	1.2	4.3	–3.0	15,535	78,858	–63,323
April	0.1	2.5	–2.3	5,249	56,039	–50,790
May	0.8	3.1	–2.3	67,317	91,354	–24,037
June	0.3	4.9	–4.5	27,857	109,275	–81,418
July	1.4	3.2	–1.8	64,243	53,732	10,511
August	2.3	4.4	–2.1	94,626	68,282	26,344
September	1.5	7.7	–6.2	37,067	102,478	–65,411
October	3.3	6.6	–3.4	72,300	99,122	–26,822
November	24.2	5.1	19.1	408,756	81,356	327,400
December	4.5	6.4	–1.9	63,161	98,261	–35,100

^amean monthly river water level decline in 2013 minus mean monthly river water level decline in 2014.

^bmonthly water abstraction from the Smoky River in 2013 minus monthly water abstraction from the Smoky River in 2014.

Task Committee on Hydrology Handbook, 1996. Hydrology handbook prepared by the task committee on hydrology handbook of management group D of the American Society of Civil Engineers. 2nd ed. New York: ASCE.

 Google Scholar

Allen, R.G., et al., 1998. Crop evapotranspiration – guidelines for computing crop water requirements – FAO irrigation and drainage paper 56. Rome (Report): FAO – Food and Agriculture Organization of the United Nations.

 Google Scholar

Zeng, X., 2001. Global vegetation root distribution for land modeling. Journal of Hydrometeorology, 2 (5), 525–530. doi:10.1175/1525-7541(2001)002<0525:GVRDFL>2.0.CO;2

 Web of Science ®Google Scholar

Myneni, R., et al., 2003. User’s guide FPAR, LAI (ESDT: MOD15A2) 8-day composite NASA MODIS land algorithm, FPAR. LAI User’s Guide, Terra MODIS Land Team (Report).

 Google Scholar

Kim, S.H., et al., 2005. Analysis of temporal variability of MODIS leaf area index (LAI) product over temperate forest in Korea. International Geoscience and Remote Sensing Symposium (IGARSS), 6, 4343–4346.

 Google Scholar

Wijesekara, G.N. and Marceau, D.J., 2009. Integrating a land-use cellular automata (CA) model with a hydrological model (MIKE-SHE) to simulate the impact of land use changes on water resources in the elbow river watershed in Southern Alberta. Report Submitted to Alberta Environment.

 Google Scholar