British Columbia Mine Reclamation Symposium

Streamflow trends in northwest BC : implications for closure water balance modelling Jackson, S. I. 2016

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata


59367-Jackson_S_I_Streamflow_trends_modelling.pdf [ 435.97kB ]
JSON: 59367-1.0354671.json
JSON-LD: 59367-1.0354671-ld.json
RDF/XML (Pretty): 59367-1.0354671-rdf.xml
RDF/JSON: 59367-1.0354671-rdf.json
Turtle: 59367-1.0354671-turtle.txt
N-Triples: 59367-1.0354671-rdf-ntriples.txt
Original Record: 59367-1.0354671-source.json
Full Text

Full Text

154  STREAMFLOW TRENDS IN NORTHWEST BC – IMPLICATIONS FOR CLOSURE WATER BALANCE MODELLING   S.I. Jackson, M.Sc., P.Geo.  Hydrologist Lorax Environmental Services Ltd., Vancouver, Canada  ABSTRACT The increasing complexity of long-term predictions of water quantity and its variability for mine closure planning requires that the non-stationarity of streamflow is taken into account. In particular and given the high probability of climate-induced changes to streamflow conditions in many regions, regulatory decisions based on historical data may not necessarily reflect future conditions.  From a closure planning perspective, therefore, consideration to ongoing changes in streamflow may be necessary. The analysis described here focuses on the Skeena Region in north-western British Columbia, an area characterized by: 1) large scale industrial development; 2) a high degree of spatial variability in stream flow; and 3) large scale changes in predicted future flow conditions associated with climate change. Trends were calculated for up to 55 Water Survey of Canada hydrometric stations, and for four time periods of varying duration. Trends were not uniformly significant for all water bodies considered, emphasizing the importance of acquiring site-specific information.  Nevertheless, on the basis of cases where statistically significant trends were identified, some broad patterns were identified and included: • Trends in average annual flows are mixed but show some tendency toward increases; • There is a general but not universal tendency toward increasing winter minimum flows and declining summer flows; • Most glaciated basins are experiencing reductions in median August flows; • Evidence is observed for a general trend toward earlier freshets;  and • Most variables maintained the same trend direction at a given station for all periods analysed, although the statistical significance of the trends varied depending on the period. This paper focuses on variables of interest for future water balance estimates required for engineering design, permitting, effluent discharge and/or water treatment scheduling.   KEY WORDS Annual flows, low flows, peak flows, glaciers, trends, north-western British Columbia   155  INTRODUCTION  Many operational and regulatory decisions related to water quantity, and secondarily water quality (e.g., loading calculations) at mine sites are based on limited site specific streamflow data. These data are often linked to regional hydrometric stations with longer record periods to provide estimates of pertinent streamflow variables (e.g., mean annual runoff, 7-day low flows, and peak flow magnitudes). This requires two important assumptions: 1) the regional streamflow record(s) are transferable to the site being studied; and 2) the regional (and site specific) data are of sufficient length to adequately represent long-term variability. The latter assumption often implies that the streamflow records being utilized are statistically stationary. In other words, a metric calculated from the historic record will retain the same value when re-calculated from future data. As the understanding of anthropogenic and climatic influences on streamflow improves, it is apparent that this assumption is often not valid. Therefore, an assessment of the potential influence of long-term change in streamflow metrics should be undertaken to ensure that estimates of future streamflow conditions are as robust as possible.  This analysis focuses on the Skeena Region in north-western British Columbia, an area where the number of large scale industrial developments is currently increasing – most of which requires highly detailed knowledge of the regional streamflow regime. The intent of this analysis is to present the current trends in several (n = 23) streamflow variables of interest. Four time periods of increasing length are examined to determine the effect of record length on the identified trend magnitude, direction and significance. Trends are grouped by median basin elevation, basin area, percent glacier coverage and median glacial elevation (within the basin of interest).   The objectives of this analysis are as follows: 1. Calculate trends in various streamflow variables for the Skeena Region; 2. Determine whether they are statistically significant; 3. Identify the existence of spatial patterns in the trends (related to median basin elevation, basin area, percent of basin with glacier cover, and median elevation of glaciated portion of each basin).  METHODS  Streamflow records of average daily discharge were downloaded from HYDAT (WSC, 2013). Out of 206 hydrometric stations in the study area, 55 stations met the following criteria (Jackson, 2014): • Station active as of 2010; • Record period of >10 consecutive years; 156  • Basin is non-regulated (i.e., no major diversions or impoundments). • Missing data or incomplete records were accounted for using the following methods: • Discharge time-series were truncated to ensure annual time-series was continuous (i.e., a minimum of 5 years of data was present before and after a missing year, gaps greater than one year were disallowed); • Where a month had > 5 days of missing data, this month was removed from the time-series, and; • No missing data were estimated or infilled.  • Basin areas and median elevations were taken from the Inventory of Streamflow in the Skeena Region (Ahmed and Jackson, 2013. The region has extensive glacier cover.  In general, glaciers can be an important control on watershed hydrology and hydroclimatic change (e.g., Meier, 1969; Jansson et al., 2003; Moore et al., 2009; O'Neel et al., 2014).  In particular, glaciers have been shown to influence trends in annual flow volume in northwest BC and southwest Yukon (Fleming and Clarke, 2003) and summer baseflows throughout BC (Stahl and Moore, 2006). Glacial polygons were obtained for the study area from the Global Land Ice Measurements from Space (GLIMS) database (Raup et al., 2007; Bolch et al., 2010; GLIMS, 2014). Median glacier elevation was calculated from the overlay of the polygon on the BC TRIM DEM data (25 m cell size). The glacier polygons were overlaid on the updated basin polygons to calculate percent glacier cover for all basins.  Table 1  Streamflow metrics analysed for trends Variable1 Description MIN_Q Minimum annual average daily discharge AVG_Q Average annual average daily discharge (calendar year) MED_Q Median annual average daily discharge (calendar year) AVG_WAT_YR Average water year discharge (Oct. 1 - Sept. 30) MED_WAT_YR Median water year discharge (Oct. 1 - Sept. 30 PULSE_DATE Date of freshet pulse initiation DATE_CM Date of hydrograph centre of mass - calendar year ANN_7Q_MIN Annual minimum 7-day average low flow ANN_30Q_MIN Annual minimum 30-day average low flow JUN-SEP_7Q_MIN Annual June-September minimum 7-day average low flow 157  JUN-SEP_30Q_MIN Annual June-September minimum 30-day average low flow Each Month Median daily discharge 1All variables calculated from the mean daily discharge values provided in the HYDAT database. The following streamflow timing metrics were calculated in addition to the various discharge metrics: PULSE_DATE – The date of freshet initiation calculated as the day when the cumulative departure from that years mean annual discharge is most negative (Cayan et al., 2001). The cut-off date is set as August 31, to ensure that autumn rain events in mixed rain/snow regimes are not inadvertently counted as the freshet date. DATE_CM – Date of centre of hydrograph mass (calendar year), calculated following Stewart et al. (2005).  In the context of mine effluent management, the regulatory criteria are often based on concentrations of potential contaminants of concern (PCOC) in the effluent and the receiving environment. Concentration of PCOCs is a result of a given load of a contaminant in a given volume of water, and thus reductions in the volume of water without a concurrent reduction in loadings will result in higher concentrations. This means that periods of sustained low flows in the receiving environment are often the constraining factor in an effluent discharge regime. Low flows are most commonly assessed for two periods; annual (usually representative of winter flows in nival and glacial regimes) and June-September. These indices are commonly based on a rolling 7-day average of discharge (7Q) or, less commonly, a rolling 30-day average (30Q) (Smakhtin, 2001). The minimum 7- or 30-day averages are then tabulated on an annual basis, and recurrence interval analyses conducted (e.g., 7Q10). Time series of these four low flow indices were examined, to determine whether water balance estimates may need to consider potential shifts in these metrics over time.  Many of the stations that represent small, high-elevation catchments also have the shortest record periods, relative to the larger drainages. These small catchments are of greatest interest when trying to find surrogate catchments for mine influenced headwater streams. In order to determine whether trends calculated for larger watersheds (with longer record periods) were still present at the same temporal scale as the small watersheds, the records were split into the following four periods: • 1961-2010 (19 stations) • 1971-2010 (32 stations) • 1981-2010 (37 stations) • 1997-2010 (55 stations)  Due to periods of hydrometric network expansion and contraction, there is a significant step change in the number of records that extend prior to 1961. For similar reasons, and due to the expansion of the WSC network into smaller headwater basins in the late 1990s, most of the records for basins less than 100 km2 158  in area begin in 1997 (9 stations out of 13). The hydrometric stations examined here are listed in Jackson (2014).  While every attempt has been made to ensure that the identified trends are reflective of long-term changes in streamflow, trends can be influenced by several factors, including, but not limited to: • Land cover change (e.g., urbanization, forestry, linear development); • Large scale climate cycles (e.g., Pacific Decadal Oscillation [PDO], El Niño Southern Oscillation [ENSO]; Cayan and Peterson, 1989; Stewart et al., 2005), and; • Changes in measurement techniques, QA/QC practices, rating curves, etc.  Annual time-series of these parameters were analysed for trend using the ‘zyp’ package in R (Bronaugh and Werner, 2013). Initial trend slopes were estimated using the Theil-Sen approach. To address serial correlation effects on statistical hypothesis tests for trend, if a trend was noted, the time-series was detrended using the slope (Yue et al., 2002). The trend and residuals are then blended, and the Mann-Kendall’s test for trend significance (p-value) is then applied. The zyp package reinflates the values that trend significance are calculated from by dividing by (1-AR[1]). All trends are presented as annual values (i.e., rate of change per year in the variable of interest). However, Fleming and Weber (2010) found that the basic conclusions drawn from analyses of water supply trends in British Columbia were not strongly sensitive to the methodology employed.  Discharge trends (calculated in m3/s) were converted to unit runoff (in L/s/km2) to account for the influence of basin area on relative trend magnitude.   RESULTS  The full results are presented in table, figure and map form in the full report, which is available here:, along with the data in Excel format to allow ease of use. Selected results are presented here to illustrate the general patterns in streamflow trends.  Trends were not uniformly significant for all Skeena-region rivers considered, and site-specific evaluation is necessary.  Nevertheless, on the basis of cases where statistically significant trends were identified, we can identify some broad patterns - examples of which are presented in the following sub-sections. In all figures below, all trends are presented as an annual increment in L/s/km2/year, and those that are significant at p<0.10 are shown in bold. The zero, or no trend line is included on the plots for reference. 159     AVERAGE ANNUAL FLOW  Trends in average flows are somewhat mixed but show some tendency toward increases, consistent with other work in southern BC. Note that the proportion of stations with an identified significant trend varies depending on the period selected for analysis. The 1981-2010 period largely overlaps the last positive phase of the PDO, which resulted in warmer and drier conditions in British Columbia (Stahl et al., 2006). Neal et al. (2002) examined streamflows in southeast Alaska to determine whether PDO phase had an effect on average annual and seasonal discharge. The authors found no significant difference in average annual streamflows between the warm and cold phases of the PDO, but did note higher winter flows and lower summer flows during the warm-PDO, and vice-versa for the cold-PDO.   Figure 1 Average Annual Daily Discharge  LOW FLOWS  There is a general but not universal tendency toward increasing winter minimum flows and declining summer flows, consistent with prior work in BC (e.g., Burn and Hag Elnur, 2002). The results do not indicate that the directions or magnitudes of the identified trends are related to basin area or median elevation. However, it is worth noting that while the direction of trend for annual low flows is consistent between periods (Figures 2 and 3), the direction of trend in summer low flows appears to be dependent on the period selected for analysis. The 1961-2010 period shows a general tendency towards lower flows for 160  the significant trends (Figure 4), while the 1981-2010 period indicates the opposite (increasing flows; Figure 5).     Figure 2 Annual Minimum 7-day Average Discharge (1961-2010)   Figure 3 Annual Minimum 7-day Average Discharge (1981-2010)  161   Figure 4 June-September Minimum 7-day Average Discharge (1961-2010)   Figure 5 June-September Minimum 7-day Average Discharge (1981-2010)   GLACIATED BASINS  Most glaciated basins examined in this study are experiencing reductions in median August flows. Figure 6 shows the influence of median basin elevation and glaciated area on identified trend. Discharge trends become less pronounced (less negative) as the median elevation of a basin increases. This is not surprising, given the strong dependence of air temperature on elevation. Basins with less than 5% glaciated area, have consistently negative trends in median August streamflow. This relationship reverses above 5% glaciated area, and two basins have experienced a significant increase in median August discharge.  162   Figure 6 Median August Daily Discharge (1971-2010)  Both stations have glacier cover on greater than 8% of the total area, suggesting that the recent melt has resulted in increases in August flows. The basin with the strongest trend is Surprise Creek near the Mouth (08DA005), which is 17% glaciated by area. Figure 7 shows an aerial photograph from 1997 showing the present ice front, and the previously glaciated down-valley area. The glacier has retreated ~3 km and down wasted ~300 m since its maximum extent at the end of the Little Ice Age (~1880 AD). The rate of retreat has increased in recent years, and the ice front has been retreating at a rate of 25 m/year from 1986-2004 (Jackson et al., 2008). This has resulted in a significant additional discharge during the summer months, and is reflected in the strong positive trend identified for August streamflows.   FRESHET TIMING  This study found strong evidence for a general trend toward earlier snowmelt and consequently initiation of the spring freshet. This is consistent with prior work across much of western North America, and has been linked to increasing spring temperatures (Cayan et al., 2001; Stewart et al., 2005). For the longest record period examined, all identified trends were negative (indicating earlier initiation of freshet over time) (Figure 8). Most of the significant trends (at p<0.10) were in basins with median elevations less than 1300 masl.   163   Figure 7 Surprise Glacier Ice Front (1997).    Figure 8 Date of Freshet Initiation (1961-2010)    164  IMPLICATIONS FOR CLOSURE WATER BALANCE MODELLING  While the trends identified in this paper (and in the complete report) are interesting in their own right, these findings have specific application to the issue of water management at a closed mine site. In general, the two main issues arising from potential changes in flow volumes over time are: • Mine site – timing and volume of contact water requiring management (i.e., diversion, storage and treatment), and; • Receiving environment – timing and volume of natural catchment streamflows that provide dilutive capacity for mine site effluent.  Average Annual Flow  At a closed mine site with an active water management regime, an increase in the average annual flow volume would likely lead to increased storage and treatment requirements. Depending on the site conditions, it could potentially lead to an increase in the available dilution as well, but the specifics would require examination on a seasonal basis, as presented below.  Low Flows  Since there are two low flow periods (annual or winter, and summer) in the Skeena region, the relative importance of an identified trend will depend on the site specific conditions. However, similar considerations apply – an increase in receiving stream winter low flows may provide additional dilution, or increasing winter flows at site may result in additional water management challenges. For example, there may be reduced capacity in diversion and storage infrastructure due to ice cover, or an increase in contact water without a concurrent increase in receiving environment flows. On the other hand, increasing summer low flows may provide a net positive effect, in that additional receiving environment flows could increase during a biologically sensitive time of year. Increases in mine site flows during the summer are likely to be lower magnitude than increases at other times of year, and may not result in appreciable changes to the water management regime for this season.  Glaciated Basins  Given the apparent threshold of 5% glacier area, and the relationship between trend magnitude and elevation, it is possible that as the glaciers continue to melt, a threshold will be passed where the 165  remaining ice is sufficiently high enough in the watershed that the rate of melt will slow considerably. The same result would be expected as the glaciated proportion of a given basin’s area decreases. Therefore, these factors should be considered when producing water balance predictions for the closure phase of a mine located in a watershed with significant glacial cover. The potential effects of glacial retreat on summer streamflows, and the potential for a threshold in the melt augmentation of these flows should be incorporated into the future water balance scenarios.   Two examples of how this phenomena may be relevant to the management of a closed mine site are: • Design capacities of water management infrastructure (e.g., diversions, settling and storage ponds, water treatment rates and seasonal hydrograph matching), and; • Receiving environment water quality predictions, as the streamflow regime shifts, so will the dilution ratios. This will require adjustments to the water management plan, and potentially permit amendments, and/or updates to the water treatment schedule. Glacial melt augmentation of streamflow provides additional dilution for effluent during a season that is otherwise subject to lower flows, particularly in freshet dominated areas. Water treatment plants are often not sized to handle the peak flows generated by freshet, and thus it is common practice to store water from the freshet and release it over the following months and if the melt regime changes as a result of lower ice volumes at higher elevations in the future, this could have implications for the management of water quality during a biologically important time of year.  Date of Freshet Initiation  Most of the identified trends towards an earlier freshet were identified in basins with a median elevation less than 1,300 masl. This has implications for water management planning, including the need to treat larger volumes of contact water earlier in the year and ensuring that snow has been removed from diversion ditches. An additional consideration is that an earlier freshet means that snowmelt will be occurring during a period of reduced solar radiation (earlier in the year, therefore lower solar zenith angle and shorter days). In some cases, this may prolong the freshet, which from a water management perspective might be beneficial. Finally, as snowmelt is primarily a function of temperature, an alteration of the timing of melt may have an elevational signature as well. If a mine component that generates loads that must be treated and/or diluted (e.g., waste rock dump) is located at a lower elevation, and melts earlier than it did historically, the runoff may be out of phase with the dilutive capacity provided by snowmelt from higher elevations. This in turn may require adjustments to the water management plan, treatment, diversion and storage infrastructure, or discharge permit limits.   CONCLUSIONS 166   The results of this analysis clearly demonstrate the non-stationarity of many common streamflow metrics used to characterize the water balance of a mine-site. During operations, these changes would likely be insignificant in comparison to the water balance alterations resulting from changes in water management and mine scheduling. In many cases, using conservative criteria selected during the permit application phase and adaptively managing the site discharges would likely suffice to remain within permit limits. However, when planning for closure, the practitioner would be prudent to consider the implications of long-term changes in the regional streamflow regime with respect to planning and management of a closed mine.   The results of this analysis are in agreement with many other similar studies published in North America, and also dovetail with some of the predicted changes to streamflow regimes that would be expected to result from climate change. While caution must be exercised in extrapolating a trend line too far into the future, the consequences of current trends continuing (e.g., decreased summer flows, and the potential for a change point in glaciated basins once the source glacier has melted) may be significant enough to warrant changes in water balance assumptions, water management infrastructure and water management plans.  It is also worth emphasizing that serial correlation within the data can have a potentially significant effect on trend characterization. Therefore, the appropriate steps should be taken to ensure that the slope (or significance) of a trend is not an artefact of persistence in the basin being analysed (e.g., storage effects on low flows that vary on greater than annual time-scales).  ACKNOWLEDGEMENTS  Comments on the original draft report (on which this paper is based) were provided by: • David Campbell (BC River Forecast Centre) • Michael Dabiri (Klohn Crippen Berger Ltd.) • Colin Fraser (Lorax Environmental Services Ltd.) • R. Dan Moore (University of British Columbia) • Robin Pike (BC Ministry of Environment) • Markus Schnorbus (Pacific Climate Impacts Consortium) • Hamish Weatherly (BGC Engineering Inc.) • Arelia Werner (Pacific Climate Impacts Consortium) 167   Final peer review of the report was provided by Sean Fleming (Environment Canada) and Markus Schnorbus. Dave Amirault (BC Ministry of Forests, Lands and Natural Resource Operations) conducted the analysis to determine the glaciated proportion of each basin and produced all trend maps.     REFERENCES  Ahmed, A. and Jackson, S. 2013. Inventory of Streamflow in the Skeena Region. October 2013. Knowledge Management Branch, British Columbia Ministry of Environment, Victoria, B.C.  Bolch, T., Menounos, B. and Wheate, R. 2010. Landsat-based inventory of glaciers in western Canada, 1985-2005. Remote Sensing of the Environment, Vol. 114, pp. 127-137.  Bronaugh, D. and Werner, A. 2013. Package ‘zyp’. V.0.10-1.   Burn, D.H. and Hag Elnur, M.A. 2002. Detection of hydrologic trends and variability. Journal of Hydrology, Vol. 255, pp. 107-122.  Cayan, D.R. and Peterson, D.H. 1989. The influence of North Pacific atmospheric circulation on streamflow in the West, in Aspects of Climate Variability in the Pacific and the Western Americas, Geophysical Monograph Series, Vol. 55, pp. 375-397.  Cayan, D.R., Kammerdiener, S.A., Dettinger, M.D., Caprio, J.M. and Peterson, D.H. 2001. Changes in the onset of spring in the western United States. Bulletin of the American Meteorological Society, Vol. 82, pp. 399-415.  168  Fleming, S.W. and Clarke, G.K.C. 2003. Glacial control of water resource and related environmental responses to climatic warming: empirical analysis using historical streamflow data from northwestern Canada. Canadian Water Resources Journal, Vol. 28, pp. 69-86.  Fleming, S.W. and Weber, F.A. 2010. Detection of long-term change in hydroelectric reservoir inflows: bridging theory and practice. Journal of Hydrology, Vol. 471, pp. 36-54.  GLIMS (Global Land Ice Measurements from Space). 2014. Bolch, T. (submitter); Bolch, T. (analyst), 2007. GLIMS Glacier Database. Boulder, CO. National Snow and Ice Data Center. Last accessed January 2014.  Jackson, S.I., Laxton, S.C. and Smith, D.J. 2008. Dendroglaciological evidence for Holocene glacial advances in the Todd Icefield area, northern British Columbia Coast Mountains. Canadian Journal of Earth Sciences, Vol. 45, pp. 83-98.  Jackson, S. 2014. Streamflow Trends in the Skeena Region. BC Ministry of Environment. 75 pp.  Jansson, P., Hock, R. and Schneider, T. 2003. The concept of glacier storage: a review. Journal of Hydrology, Vol. 282, pp. 116–129.  Meier, M.F. 1969. Glaciers and water supply. Journal of the American Water Works Association, Vol. 61, pp. 8-12.  Moore, R.D., Fleming, S.W., Menounos, B., Wheate, R., Fountain, A., Stahl, K., Holm, K. and Jakob, M. 2009. Glacier change in western North America: influences on hydrology, geomorphic hazards and water quality. Hydrologic Processes, Vol. 23, pp. 42–61.  Neal, E.G., Walter, M.T. and Coffeen, C. 2002. Linking the pacific decadal oscillation to seasonal stream discharge patterns in Southeast Alaska. Journal of Hydrology, Vol. 263, pp. 188-197.  O’Neel, S., Hood, E., Arendt, A. and Sass, L. 2014. Assessing streamflow sensitivity to variations in glacier mass balance. Climatic Change, doi 10.1007/s10584-013-1042-7.  169  Raup, B.H., Racoviteanu, A., Khalsa, S.J.S., Helm, C., Armstrong, R. and Arnaud, Y. 2007. The GLIMS Geospatial Glacier Database: a New Tool for Studying Glacier Change. Global Planetary Change, Vol. 56, pp. 101-110. (doi:10.1016/j.gloplacha.2006.07.018)  Smakhtin, V.U. 2001. Low flow hydrology: a review. Journal of Hydrology, Vol. 240, pp. 147-186.  Stahl, K. and Moore, R.D. 2006. Influence of watershed glacial coverage on summer streamflow in British Columbia, Canada. Water Resources Research, Vol. 42,  Stewart, I.T., Cayan, D.R., and Dettinger, M.D. 2005. Changes toward Earlier Streamflow Timing across Western North America.  Journal of Climate, Vol. 18, pp. 1136-1155.  Water Survey of Canada. 2013. Environment Canada Data Explorer (HYDAT). Last accessed August 16, 2013.  Yue, S., P. Pilon, B. Phinney and G. Cavadias. 2002. The influence of autocorrelation on the ability to detect trend in hydrological series. Hydrological Processes, Vol. 16, pp. 1807-1829. 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items