Data and Methods

All data is displayed in the NAD_1983_BC_Environment_Albers projection. A Multi-Criteria Analysis (MCE) of suitable areas for geothermal power plant sites was conducted. Using the AHP program from the Canadian Conservation Institute ((http://www.cci-icc.gc.ca/tools/ahp/index_e.asp), three different weighting schemes were determined: one emphasizing geothermal potential in terms of heat content, one emphasizing geothermal potential in terms of permeability, and one equal weight scheme. All data and shapefiles were processed and normalized prior to the MCE, and the processing steps for each type of information are given below.

After the steps below, all the layers except the geology layer had all null values reassigned as zero, first by using the ‘Is Null’ tool to identify all NoData values within the extent of the layer, then using the ‘Con’ tool to set all nulls to zero. Setting all NoData values to zero was necessary because the ‘Weighted Sum’ overlay tool discarded any raster cells at positions that had NoData values, meaning previously, only cells with data from all six layers would be shown in the resulting raster.

Temperature Gradient

The average temperature gradient of the earth is 30 degrees Celsius per Kilometer (Gupta and Roy, 2006). A higher temperature gradient means the geothermal potential of a site is higher, since higher temperatures can be accessed with the same depth of drilling compared to sites with lower temperature gradients.

An Excel spreadsheet of analyzed borehole temperature gradient data provided by Sarah Kimball was converted into a shapefile using ArcMap 10 using latitude and longitude data within the Excel file. A rotary borehole shapefile from the BC Ministry of Energy, Mines and Petroleum Resources that contained temperature gradient data was also used.

A 30km buffer was first made around all boreholes, then the two datasets were converted to raster’s that contained temperature gradient data in degrees Celsius per Kilometer. The two temperature gradient raster’s were then merged using the ‘mosaic to new raster’ tool to create a raster with temperature gradient data from both raster’s. Finally, the new temperature gradient raster was reclassified and normalized using the large classification option in the overlay ‘fuzzy membership’ tool, with a midpoint of 40 degrees Celsius per Kilometer and a spread of 5. Higher gradient temperatures have a higher weight.

Heat Flow

The average heat flow of the earth is 80 mW/m^2 (Kimball 2010). Heat flow is a product of the thermal conductivity and the temperature gradient in the ground, making heat flows a good indicator of geothermal potential.

An excel spreadsheet of borehole heat flow data from the “Geothermal Resources of British Columbia” map by Fairbank & Faulkner (1992) provided by Sarah Kimball was used to input and digitize the data. However, the positioning of the Excel heat flow data was off compared to the other layers (boreholes were slightly off position, resulting in some boreholes being in water or outside of BC), so the Excel heat flow data was rubber sheeted to the borehole shapefile from the BC Ministry of Energy, Mines and Petroleum Resources, which contained plotted coordinates and points of drilled boreholes from the “Geothermal Resources of British Columbia” map by Fairbank & Faulkner (1992), but was missing heat flow data and certain boreholes.

A buffer of 30km was put around the borehole and and then converted to a raster with heat flow data. The large classification option in the overlay ‘fuzzy membership’ tool was used to reclassify and normalize the raster borehole heat flow data such that higher heat flow values have a higher weight. A midpoint of 90 mW/m^2 and a spread of 5 was used.

Hot Springs

In one study in Northern Japan, over 97% of geothermal wells in the area were within 4km of hot springs, meaning that hot springs are heavily correlated with geothermal potential (Noorollahi et al, 2007).

The hot springs shapefile from the BC Ministry of Energy, Mines, and Petroleum Resources contained temperature data, but not all values were given. For the values that were given, some springs were designated ‘WARM’ and ‘HOT’ and the ones that contained numbers were in string format. A new column with the temperatures in short integers format was created and populated with the string numbers, now in short integer format. The hot springs that did not have an original temperature were omitted to not skew results, while the ones designated ‘WARM’ were given a temperature equivalent to the mean of the the hot spring temperatures (40 degrees Celsius), and the hot springs designated ‘HOT’ were given a temperature equivalent to the mean plus one standard deviation (58 degrees Celsius).

A buffer of 4km was created around all the hot springs, with no buffers being dissolved to ensure all buffers kept their temperature values. The hot springs buffer layer with temperature was then converted to a raster, with any overlapping parts of the buffers defaulting to having the higher temperature value. Finally, the large classification option in the overlay tool ‘fuzzy membership’ was used to reclassify and normalize the hot spring temperature buffer raster such that higher temperatures have a higher weight, with a midpoint of 40 degrees Celsius and a spread of 5 being used.

Geology

Volcanic and intrusive rocks are usually associated with geothermal resources (Noorollahi et al, 2007, Kimball, 2010). Also, younger rock formations are usually correlated with higher geothermal potential, especially in volcanic rocks (Kimball, 2010, Noorollahi et al, 2007). We will be combining and analyzing both these factors using the geology layer, as data for both rock types and rock age are available from the BC Geology shapefile from the BC Ministry of Mines, Energy, and Petroleum Resources.

First, the field ‘rock class’ in the vector ‘BC Geology’ shapefile was converted to a raster. The ‘rock class’ raster was then reclassified such that volcanic and intrusive rocks had a value of 10, while other rocks were given a value of 5 due to data on the correlation of other rock types with geothermal potential being unavailable. Finally, the linear classification option in the overlay tool ‘fuzzy membership’ was used to normalize the ‘rock class’ raster so that volcanic and intrusive rocks had a weight of 1 and other rocks had a weight of 0.5

The numbers in the field ‘max age of rock formation’ (given in millions of years) in the vector ‘BC Geology’ shapefile were in string format, so a new field ‘max age of rock formation’ was created and populated with the numbers as integers using the field calculator. The field ‘max age of rock formation’ was then converted to a raster layer. According to the metadata, any layer with a ‘max age of rock formation’ of 4000 actually has an unknown age, so the linear classification option in the overlay tool ‘fuzzy membership’ was used to reclassify the ‘max age of rock formation’ raster such that younger formations have a higher weight.

Finally, the ‘rock class’ and ‘max age of rock formation’ raster's were overlaid using the ‘fuzzy overlay’ tool with overlay type ‘SUM’, which ensured that rocks that were both young and volcanic or intrusive had a higher weight, as a combination of age and type is more important than just one of the factors.

Faults

According to one study in northern Japan, over 95% of geothermal wells in the area were within a 6km radius of active faults, possibly due to the importance of water flow through faults for permeability (Noorollahi et al, 2007). A faults shapefile from the BC Ministry of Energy, Mines, and Petroleum Resources was used.

A buffer of 6km was created around the faults. Because the relationship in fault permeability and well productivity is not well known, all the values within the 6km fault buffer have a value of 1. The buffer was then converted to a raster.

Transmission Lines

Transmission lines are neccessary to carry electricity that is generated from geothermal power plants. We will base our analysis on locations within a 50km distance of the power line, since the general trend is that 1km of transmission lines can be built for every mW of power generated by a geothermal facility (Kimball, 2010), and 40-60mW is in the mid range of typical geothermal power generated by facilities. The transmission line data from BC Hydro is a polyline and ranges from 68-500kV. We are going to ignore the maximum capacity of the power lines due to maximum capacity not always being a good indicator of how much capacity is available (Kimball, 2010). The transmission polyline was first converted to a raster; this was done because the euclidean distance tool did not function properly when used on the transmission polyline (the size of the cells could not be set properly and the extent of the created raster was too large).

Using the transmission raster and the euclidean distance tool, a euclidean raster with the distance from the transmission line was created for up to 50 km away from the transmission line. One problem encountered while creating the new euclidean raster was that part of the raster was cut off and not generated; this was later determined to be a problem with the extent of the analysis, and the extent was readjusted in the Environment Settings. Finally, the new euclidean transmission raster was normalized using the linear classification option in the overlay ‘fuzzy membership’ tool, with a minimum of 50 km and a maximum of 0 km to ensure that pixels closer to transmission lines had higher weights.