Fraser

A Changing Climate: Environmental Assessment for a Proposed Mine in Yukon, Canada Presenting Author: Colin Fraser, [email protected]

Co-authors (alphabetically by last name):

Lorax Environmental Services Ltd., 2289 Burrard Street, Vancouver, BC, Canada V6J 3H9

Jennie Gjertsen, [email protected] Goldcorp Inc., 666 Burrard St #3400, Vancouver, BC, Canada V6C 2X8 Scott Jackson, [email protected] Lorax Environmental Services Ltd., 2289 Burrard Street, Vancouver, BC, Canada V6J 3H9 James Scott, [email protected] Goldcorp Inc., 666 Burrard St #3400, Vancouver, BC, Canada V6C 2X8

1. Background This poster pertains to the construction, operation, reclamation and closure of the Coffee Gold Mine (Project), a proposed gold development project in westcentral Yukon, approximately 130 km south of Dawson, YT, Canada (Figure 1a). Major infrastructure related to mining and processing at the Project area includes: an upgraded road; a primary waste rock storage facility; several open pits; water diversion structures and storage ponds; haul roads; primary and secondary crushing facilities; heap leach facilities; a gold refinery; an accommodation complex; and an all-weather airstrip. Most of the proposed infrastructure will be situated at high elevation (~1,250 m asl).

2. Objectives of the Poster

c

a

The panels below: i) summarize steps taken to create a long-term (84-year), daily- climate record that accounts for a changing climate; and ii) introduce a water balance model (WBM) that was constructed and calibrated using GoldSim modelling software. To evaluate potential water quantity issues and risks associated with the Project, and to fulfill requirements of the environmental assessment process, the 84-year climate record was used to drive the WBM with outputs used to quantify residual streamflow changes attributable to the Project in a valued component context.

d

b

Local climate conditions for the site are typical for the region: average annual air temperature (T) is -2.6 °C and mean annual precipitation (P) is estimated to be 485 mm (65% rain, 35% snow) at elevation 1,300 m. Inspection of the instrumental record from nearest weather stations (e.g., Dawson; Mayo, YT; Pelly Ranch. YT) confirms T and P have been increasing on an annual basis since the 1960s (refer to Figure 1b, T trends are shown). Further, climate change scenario data for the Project area indicate continued change for these key water balance variables (i.e., +5°C and +20% increases by 2100; Figure 1c and 1d). Against this changing climate context, engineering and permitting studies are ongoing at the mine site.

Figure 1 (to left) : a) Map of Canada showing location of the Coffee Gold Mine; b) Temperature trends by season for the Dawson A and Pelly Ranch climate stations (1952 to 2013); c) Climate change scenario data (T) for a grid point near the Project and worst case (i.e., RCP8.5 or A2); and d) Climate change scenario data (P) for a grid point near the Project and worst case (i.e., RCP8.5 or A2).

3. Baseline Monitoring

4. Generating the Long-term Climate Record

A baseline hydrology program was initiated at the Project site in the autumn of 2010, starting with installation of 3 automated hydrometric stations. Eight additional stations were established in 2014 to further characterize the streamflow regime of headwater basins containing Project infrastructure. Monthly sampling trips are made year-round to develop rating curves and maintain the stations. Hourly discharge records are computed from rating curves and high-resolution water level data.

The procedures below were followed to generate a long-term, daily- climate record for the Coffee Gold mine site:

A high-elevation climate station was installed in summer 2012 and measures: air temperature and relative humidity (2 m); wind speed and direction (10 m); incoming solar radiation (2 m); barometric pressure; and precipitation (tipping bucket rain gauge with solid phase adapter and Alter shield). Snow course measurements are also carried out at several stations (i.e., various elevations, aspects and cover types) following YT protocols. Figure 2 shows a comparison of mine site- and regional T and P data.

•

•

•

•

•

• Figure 2: Coffee Gold baseline climate data compared to period of record for 21 regional climate stations.

Daily climate data (i.e., T and P) from the mine site and a nearby regional station (McQuesten, YT) were assembled to create monthly predictive relationships based on overlapping data. Next, the regional station (predictor) and monthly predictive relationships derived for T and P were utilized to compute a long-term synthetic climate record representative of the mine site. To span the full Project life (i.e., Construction (2018 to 2020); Operations (2021 to 2032); Closure (2033 to 2042); and Postclosure (2043 to 2100), the Coffee Gold climate record was looped three times to create an 84-year, daily- climate record. To represent a plausible future condition, climate change scenario data were downloaded from the Scenario Network for Alaska and Arctic Planning, a research collaborative that produces downscaled, historical and projected climate data for sub-Arctic and Arctic regions (SNAP, 2016). Monthly T and P predictions (2001 to 2100, CMIP3/AR4 – A2 Scenario, 2 km grid) for grid points covering the mine site extent were downloaded, averaged and used to scale the 84year daily climate record. At monthly time step, assembled climate data are shown in Figure 3.

Figure 3: Air temperature (upper panel) and precipitation (lower panel) inputs to the Coffee Gold Water Balance Model. Shading in the upper and lower panels demarcates the main phases of the Coffee Gold Mine Plan. The final climate record is a daily dataset, but shown at monthly frequency above.

5. Water Balance Model Description

6. Data Analysis and Streamflow Changes

The Coffee Gold WBM is built using GoldSim software, a highly-structured, robust platform designed for flow/contaminant tracking. The WBM is climate-driven (i.e., uses the 84-year, daily- reconstructed climate record which accounts for climate change to generate flows) and is comprised of two sub-models:

For each identified WBM node, GoldSim outputs (i.e., resultant flow series from the Natural Flow and Base Case sub-models) were compared to one another with Project-related flow changes being indexed against natural/baseline conditions. WBM model outputs were averaged to monthly flow values and predicted streamflow changes were summarized by a percent change metric as follows:

• •

Natural Flow/Baseline (No Project) – This sub-model was constructed first and calibrated to accurately predict streamflow conditions for a Baseline condition (refer to Table 1 and Figure 4a for additional detail and a representative calibration example). Base Case (With Project) – For this sub-model of the WBM, year-by-year mine footprints for open pits, the waste rock storage facility, the heap leach facility, soil and ore stockpiles and related Project infrastructure were encoded onto the Baseline sub-model. Table 1: Natural Flow (Baseline) Sub-model a

• Watershed boundaries and hypsometric outputs (i.e., curves and representative bands of elevation data) for local Catchment catchments were generated from 1:50,000 mapping data. Boundaries, Elevation Data • To encode elevation dependent climate parameterizations into the WBM, drainages were separated into three elevation bands (400-800 m, 800-1200 m and >1200 m).

Climate

Hydrology

Outputs

• For this sub-model, 84-year long predicted streamflow records are generated by the WBM. Flow records are produced for seven local tributaries and three large river nodes at a monthly time step. • The outputs per WBM node consist of 28 unique iterations (i.e., 28-year climate record is time stepped) each extending the 84-year time-period.

A suite of streamflow change characteristics (i.e., direction, magnitude, frequency and reversibility of streamflow change) are proposed to guide a detailed streamflow change assessment. These change characteristics are selected to best quantify and describe potential changes against key components of a natural flow regime (Poff et al., 1997). WBM outputs were screened using tabular and graphical formats (e.g., flow vs. percent change plots, flow duration curve plots and time series plots comparing Natural Flow vs. Base Case outputs by Project phase). Example plots for the CC-1.5 model node are shown below (Figure 5). Identified streamflow changes will ultimately be described in a valued component context for the surface hydrology (water quantity) discipline, noting the streamflow summary will also inform other water-related studies (e.g., GW, surface water quality, fish and aquatic habitat) to be described as part of the environmental assessment.

• The natural flow sub-model of the WBM was driven by a 28year precipitation, air temperature and evaporation record that was looped three times. • Precipitation and air temperature inputs were scaled by elevation using gradients ascertained from site- and regional climate data. • Monthly climate change scenario data (from the Scenario Network for Arctic Planning) for the A2 emission scenario (2-km resolution) were used to scale precipitation and air temperature inputs over the long term (Closure and Postclosure phases). • Baseline hydrology data from autumn 2010 to December 2015 were combined with regional streamflow data to generate long-term synthetic streamflow records. • The sub-model was calibrated at daily time-step using longterm, daily- synthetic streamflow data as the target.

Percent change (%) = ((Mine Altered Flow – Natural Flow)/Natural Flow) x 100 (Eqn: 1)

Climate Change and Streamflow in the Yukon As part of the streamflow assessment, decadal trends were assessed by analyzing Natural Flow sub-model outputs (Figure 6). Consistent with findings for the Yukon (e.g., Streiker, 2016), GoldSim WBM results confirm the following streamflow changes for a warmer and wetter future climate regime: • Progressively earlier onset of freshet, later occurrence of autumn freeze up, longer ice-free season. • Shifts to the proportions of P realized as rain vs. snowfall. • Increases in baseflow conditions and likelihood of mid-winter melt events. • Progressive increases in total annual discharge.

b

Building upon applicable research and modelling of Christophersen and Seip, 1982 and Seip et al., 1985, the architecture of the watershed model is predicated on the concept that natural streamflow, or runoff generated from mine footprint areas, is comprised of three types of flow as described by Maidment (1993): i) quickflow, generated by storm or snowmelt events and often resulting in peak flow events; ii) interflow, derived from near-surface, lateral movement of infiltrated meteoric water; and iii) baseflow, the portion of surface discharge derived from GW discharge. This architecture is shown in Figure 4b. Same catchment boundaries, climate and hydrology inputs described for the Natural Flow sub-model (Table 1) were used to populate undisturbed portions of watersheds in the Base Case sub-model. However, to represent the disturbed condition, mine footprints for open pits, waste rock storage facilities, the heap leach facility, soil and ore stockpiles and related Project infrastructure were encoded into the Base Case sub-model.

POSTER TEMPLATE BY: www.PosterPresentations.co m

Figure 6: Decadal averaged- monthly streamflow patterns for the CC1.5 WBM node (22 km2 drainage area). Results are shown for three time bands: 2030-39, 2060-69 and 2090-99.

References: Christophersen, N. and H.M. Seip. 1982. A model for streamwater chemistry at Birkenes, Norway. Water Resources Research. 18:4, 977-996. Poff, N.L. et al. 1997. The natural flow regime: a new paradigm for riverine conservation and restoration. BioScience 47:769-784.

Figure 4: a) Daily water balance model output (red) for a node located on Halfway Creek, compared to a reconstructed streamflow timeseries (blue) for corresponding time period (2010-2014); b) A schematic presenting an overview of the three-reservoir water balance model in conceptual format.

Seip, H.M. et al. 1985. Model of sulphate concentration in a small stream in the Harp Lake catchment, Ontario. Can. J. Fish Aquat. Sci. 42: 927-937. SNAP, 2016. Scenario Network for Alaska and Arctic Planning. Data portal for climate change scenario data: https://www.snap.uaf.edu/tools/data-downloads

Figure 5: Example plots based on Goldsim WBM outputs.

Streicker, J., 2016. Yukon Climate Change Indicators and Key Findings 2015. Northern Climate ExChange, Yukon Research Centre, Yukon College, 84 p.