The response of water chemistry to hydrological conditions and flow paths under different forest cover types in small subalpine catchments

A study was undertaken to investigate the hydrologic processes involved in the generation of stream water quality in four small subalpine drainages in the Upper Penticton Creek watershed. The watersheds have different forest cover types including mature spruce-fir, lodgepole pine stands 80-140 years of age, and regenerating clear-cut areas with about 15 years of mixed pine, spruce and fir regeneration. This provides the opportunity to study the effects of forest cover type and hydrologic flow paths on streamflow and water quality production in headwater basins. The knowledge gained in this study will provide input to modeling procedures designed to predict the effects that forest manipulation might have on water flow and quality in the dry cold Engelmann Spruce - Subalpine Fir (ESSFdc) subzone of the southern interior of British Columbia. Streamflow, groundwater levels, precipitation, air temperature and physical conditions of melting snow were monitored throughout the study. Samples were collected of the different components of runoff for chemical analysis; streamflow, groundwater, precipitation (including rainfall, fresh snowfall and stratified snowpack samples) and seepage that could be interpreted as either groundwater or interflow outflow. The most important chemical components in terms of their chemical concentrations were found to be calcium, sodium, silica, sulphate, bicarbonate and nitrate. Nitrate was an important constituent of precipitation, and bicarbonate only a minor component, whereas in streamflow, groundwater and interflow components, bicarbonate was the major anion and nitrate was a minor component, only significant during events that flushed nitrate through the groundwater system into the stream. Water chemistry of streamflow was related to stream discharge using regression analysis, revealing highly significant log-normal relationships. The groundwater sites were found to be classed into four distinct groundwater zones representing recharging hillsides, bedrock, permanent seepage sites and streamside sites that were alternatively recharge or discharge sites depending on the flow regime (i.e., low flow vs high flow). Average chemical concentrations of groundwater were found to vary inversely with the logarithm of hydraulic conductivity, but also depended on the hydrologic zone. Hydrologic zones were defined primarily as streamside and hillslope areas. A model to predict water chemistry in groundwater was developed relating concentrations to total head and hydraulic conductivity, using dummy variables to represent hydrological zone' and forest cover. The forest cover type appeared to govern groundwater chemistry, particularly when the regenerating clearcut was compared to the mature spruce-fir dominated catchment; soil analysis revealed similarity between those two sites, whereas groundwater and streamflow chemistry differed significantly. This and other lines of reasoning led to the conclusion that the regenerating clear cut site was unrecovered in terms of its water chemistry. Comparison between the lodgepole pine and spruce-fir dominated watersheds revealed significant differences in water chemistry regimes (the spruce-fir site produced much lower concentrations in streamflow and groundwater) but the effects of soil and forest cover on water chemistry could not be separated. Chemical analysis of ripe snowpacks revealed that during periods when rapid melt was not occurring, solutes appeared to concentrate at the surface of the snowpack. During periods of rapid melt, those solutes were then flushed to the bottom of the snowpack. Relationships were developed that relate the chemistry of the base of the snow pack to daily snow melt. The concentration of sulphate and sodium at the base of the snow pack during rapid melt approached that of surface runoff, but in the case of bicarbonate, calcium and silica the concentrations were much lower than that of surface runoff, and for nitrate, much higher. The chemistry of snow at the base of the snowpack may be controlled in part by contact with soil. Measurements of soil temperature revealed that soils freeze before permanent snowpacks begin to develop, and do not thaw again until the snow has melted. Nevertheless, groundwater levels rise in response to snowmelt at all sites. This would provide local pathways for water to enter the soil to contribute to interflow or groundwater outflow, while the largely frozen soil would direct part of the snowmelt to direct runoff as overland flow. The conclusion is that while snowmelt produces a large volume of direct runoff, the chemistry of that direct runoff is not related to the chemistry of the snowpack, but is modified by contact with the soil. The UBC watershed simulator was calibrated to the study watersheds to determine the relative contributions of direct runoff, interflow and groundwater. Hydrologic conditions that govern groundwater chemistry were related to the calibrated upper and lower groundwater outflows from the UBC model. The relationships between groundwater chemical content and hydrologic conditions were then used to model the chemistry of the various groundwater components. Direct runoff was modeled as the chemistry of the base of the snowpack. A direct link between modeled components and the components measured in the field was successful at modeling streamflow chemistry. Some possible improvements to the simulator are noted, particularly in groundwater response in its relation to water chemistry simulation, but generally the model did an acceptable job of simulating water chemistry, and shows promise as a platform to develop a more comprehensive water quality simulator.Forestry, Faculty ofGraduat