Heat Tracing as a Tool to Bring the Streambed into Focus
Through the ages tools have clarified and even helped defined scientific fields. The telescope exploration of the skies and seismic reflections of the sea floor elevated astronomy and tectonics to new levels of existence. The streambed may be a humble cousin to these grand realms, but arguably more important to daily health, functioning as the membrane layer separating streams and aquifers. Detailed three-dimensional descriptions of streambeds have been limited by a lack of understanding of streambed stratigaphy and its impact on spatiotemporal streambed flow patterns. Invasive characterization is challenged by both the overlying stream and extremely high spatial variance relative to traditional stratigraphic characterizations. Heat tracing through detailed sediment temperature monitoring has emerged as a powerful exploration tool of streambeds. Analysis of thermal patterns in stream sediments affords the opportunity to spatially delineate the streambed from the streambank, the adjacent subsoil, and underlying alluvial aquifer, through comparison of temporal thermal patterns within each of these regions to temporal stream temperature characteristics. Time-series analysis of thermal patterns produces temporal characterization of changing streambed fluxes. As a result, heat tracing can identify the spatiotemporal presence/absence of longitudinal streambed flow (variously referred to as substream, 'hyporheic', interstitial, under or enter flow), as well as precisely locate areas of ground-water discharge and recharge through streambeds over instantaneous to inter-annual timescales. Analytical and numerical analysis of temperatures provide accurate flux estimates of spatiotemporal streambed flow critical to water budget estimates. Groundbreaking research from field observations at the base of San Gabriel Mountains (CA) by Troxler in the 1930s and benchmark Alaska sandbox studies by Vaux in the 1960s to the present day maturity of thermal technologies, demonstrates heat's unique ability to delineate the extent of streambeds and their key function in quantifying the importance of streambeds to such issues as stream ecology, fishery habitat, and optimal water-resources management.
Low-Altitude and Land-Based Infrared Thermography to Identify Types of Groundwater Discharge in NWT Streams
In tributaries of the Mackenzie River in the Northwest Territories (NWT), Canada, groundwater discharge provides critical fish habitat for Dolly Varden and bull trout populations by maintaining base flows, creating thermal refugia in winter, and providing stable riverbed temperatures for spawning. Where temperature contrasts exist between surface water and groundwater, infrared thermography can use heat as a tracer to locate groundwater discharge areas. Thermal images acquired from satellites and high altitude airplanes tend to be expensive, lack the resolution necessary to identify small discharge locations, and do not allow real time decisions to investigate and ground truth identified temperature anomalies. Therefore, a system was developed using a handheld FLIR ThermaCam P25 infrared camera, visual video camera, infrared video capture system, and GPS in a low flying helicopter and on the ground. The advantage of the system was its ability to inexpensively and efficiently characterize several kilometer long reaches of river and identify springs and seeps on a sub-meter scale and in real time. The different types of groundwater discharge that can occur in these streams include: deep geothermally heated groundwater; shallow groundwater; and active zone water, but differentiating them can be difficult because observed thermal anomalies can be non-unique functions of the initial groundwater temperature, magnitude of discharge, air and surface water temperatures, and temporal variations. Work performed in March and September easily detected spring and seeps of deep groundwater (8 to 13 ° C) at Smith Creek, Gibson Creek, Gayna River, and Little Fish Creek. Shallow groundwater discharge was detected (1 to 3 ° C) at White Sand Creek, Canyon Creek, and Fish Creek, but was more difficult to identify. Subtle variations from surrounding temperatures (<1 ° C) at some sites suggested seeps from the hyporheic zone or possibly the active zone. The limitations of infrared thermography include only being able to measure temperatures of surfaces and difficulty differentiating spatial anomalies from possible temporal influences. Overall, the handheld system was a useful reconnaissance tool for identifying surficial expressions of different types of ground water discharge.
Quantification of Surface Water and Groundwater Nitrate Fluxes to two Small Estuaries in Atlantic Canada
In parts of Atlantic Canada there is currently concern that nutrient loadings from catchments are adversely affecting water quality and ecosystems in estuaries. This is especially the case in Prince Edward Island (PEI), a province in which intensive potato production has contributed to elevated nitrate concentrations in groundwater and streams, and where eutrophic or anoxic conditions occur regularly in many estuaries. Previous nutrient loading studies conducted in PEI have focused only on the contributions from surface water, although it is known that elevated nitrate concentrations exist in groundwater and that the regional fractured sandstone aquifer has relatively high hydraulic conductivity. In this research the nitrogen loadings delivered by surface water and groundwater to two small estuaries located in PEI were quantified over a two-year period. Surface waters were monitored directly, while groundwater discharge to the estuaries was estimated using a combination of airborne thermal infrared imaging, direct discharge measurements at selected shoreline spring locations, and numerical modeling of groundwater flow in the two catchments. Relatively widely spaced shoreline springs have been identified as the dominant mode of groundwater discharge, accounting for about 70% of the total groundwater flux, and this is likely a result of the fractured nature of the contributing aquifer. Focused spring discharge was sampled on six occasions and a two component mixing model based on salinity and nitrate concentrations measured in springs, streams and estuaries was applied to calculate the nitrate concentration in groundwater discharge. The nitrate loadings to both estuaries are highly correlated with freshwater discharge, and the annual nitrate fluxes are dominated by streams (approximately 80% of the total). However, groundwater contributes between 15% and 18% of the annual nitrate load which is significant when compared to other components of estuary nutrient mass balances. The total annual groundwater nitrate load to McIntyre Creek estuary (3200 kg NO3-N/yr) is more than half of the nitrate load for Trout River estuary (5900 kg NO3-N/yr), although the extent of the estuary and its catchment are about ten times smaller than the respective areas of the Trout River estuary. This is a result of the much larger extent of developed land (mostly agricultural) in the McIntyre Creek catchment. The generally poor conditions in these estuaries, including annual anoxic events, are probably driven to a large extent by the high nitrogen loads from streams and groundwater. This study demonstrates that nitrogen loads from direct groundwater discharge to estuaries should not be ignored in these and other areas with similar land use and hydrogeological conditions.
Groundwater Flux estimation from Streambed Temperatures
Temperature is an easily measured parameter which can track groundwater and surface water interactions at the stream bed interface. The application of streambed temperature mapping as a means to estimate groundwater flux was investigated for a heterogeneous streambed. A one-dimensional, quasi-steady-state analytical solution was used to calculate groundwater upwelling and flux prediction maps were generated using ArcGIS. The results were validated with a second estimation of groundwater flux through the use of mini- piezometers. Two 40m long sites were mapped in Swan Creek, a tributary of the Grand River, located in southern Ontario. Each site spanned different geomorphic units; the first site included a riffle-pool-riffle sequence while the second site contained plane bed morphology. At both sites the streambed was dominated by cobbles and contained isolated sections of organic fines. Due to the size of the streambed material temperature measurements were collected at depths less than 10cm into the ground and were subsequently impacted by diurnal warming which invalidated the quasi-steady-state assumption. Additionally, the plane-bed site contained significant lateral groundwater inputs as spring-fed groundwater seeped in along one stream bank. This consequently invalidated the one-dimensional assumption. The results were compared with the mini-piezometer results and demonstrated the utility of the method in less than ideal conditions.
Characterization of Surface Water-Groundwater Interactions in a Proglacial Moraine Using Heat and Solute Tracers
Alpine watersheds represent the headwaters of many major rivers in western North America. Understanding the groundwater systems in these watersheds is critical to understanding the timing of water release and late season stream flow, especially given the predicted shifts in precipitation patterns due to climate change. Recent water-balance studies of alpine lakes have shown the importance of groundwater, and suggest that moraines may play an important role in its storage and release. Due to challenging terrain and the inability to install wells, the use of conventional hydrogeological methods to characterize groundwater flow is not possible on most moraines. Thus, alternative methods are needed if these features are to be adequately incorporated into future physically-based modelling attempts. As part of an integrated alpine hydrology study of the Lake O'Hara research basin in the Canadian Rockies, we used a small tarn (~ 600 m3) on a partially ice-cored, proglacial moraine as a well surrogate. The lack of surface water inflows or outflows to the tarn means that tarn water level can be used to indicate the local groundwater table and that flow rates through the tarn can be used to represent local groundwater flow rates. Here we present the results of two experiments performed on the tarn to determine local flow rates. 1) A chloride-dilution tracer experiment was performed, in which the chloride decay-rate after a one-time addition of NaCl was used to determine the volumetric flow rate through the tarn. Chloride concentration was determined from daily water samples and interpolated from half hourly electrical conductivity measurements. 2) Detailed energy-balance measurements for the tarn were made and flow rates through the tarn were determined based on the advection component of the energy balance equation. Both experiments used a digital elevation model of the tarn to determine tarn volume changes, and therefore solute mass and energy storage changes based on water level data. We compare the above methods in terms of the calculated flow rates and accuracy. Our results provide a significant step forward in parameterizing future hydrological modelling attempts in challenging alpine environments.
Constraining Groundwater Discharge in a Large Watershed: Integrated Isotopic, Hydraulic and Thermal Data from the Canadian Shield
Understanding the rate and pattern of groundwater discharge to lakes and rivers is critical for watershed budgets and for protecting the ecological integrity of lake and river ecosystems. A 900 km2 study watershed contains a river and over 3000 lakes and wetlands, mostly underlain by exposed crystalline bedrock or a thin veneer of coarse-grained sediments. The objective of this study is to constrain the rate and pattern of groundwater discharge at the watershed-scale. Groundwater discharge points were identified by conducting detailed transects of the river and lakes using temperature, conductivity and radon-222 tracers. Surface water samples from representative lakes were analyzed for ä2H, ä18O, radon-222 and chloride during three consecutive summers. Radon and chloride concentrations are used in a new steady-state advective model to determine groundwater fluxes to the representative lakes. The detailed transects identified minor and highly localized groundwater discharge locations which did not coincide with mapped geological structures or exposed bedrock fractures. Stable isotope, temperature and conductivity data identified only one subsidiary stream with significant groundwater discharge. The steady-state model indicates that the groundwater flux to lakes is generally less than 0.1 percent of the total input. This integrated thermal, chemical, isotopic and hydraulic dataset indicates that the rate of groundwater discharge to lakes in this crystalline bedrock watershed is not significant and that discharge is localized but not focused at exposed geological structures or bedrock fractures. This conclusion implies that in the watershed groundwater and surface water is largely decoupled, which has significant ecological and water management implications.
Geochemistry of Perennial Groundwater Springs in the Mackenzie Basin
Perennial groundwater flow maintains stream baseflow in winter and can provide important habitat for fish. Such groundwater flow is less common in permafrost terrain. Tributaries of the Mackenzie watershed are located in discontinuous to continuous permafrost where there is critical fish habitat which appears to be linked to this perennial groundwater discharge. The objective of this research is to use groundwater geochemistry as an indicator of where water is discharging in select arctic rivers. We hypothesized that discharging groundwater will have distinct geochemical and isotopic signatures, and that sampling of thermal and non thermal groundwater springs and the surface water of associated rivers can determine the extent of groundwater inputs. Eight rivers were selected for study based on the known occurrence of open water in winter. Sampling of water on these rivers has focused on sections of river where discharge is suspected. Analysis has focused on identifying the mixing of groundwater with surface water. Gibson Creek, Little Fish Creek, and Smith Creek were all characterized by thermal groundwater having high total dissolved solids. Non-thermal groundwater observed discharging to White Sand Creek, Hodgson Creek, Canyon Creek, Gayna River, and the Rat River have more subtle but unique geochemical signatures that reflect the different geological setting of the particular river. Groundwater discharging into White Sand Creek and Hodgson Creek had higher sodium and lower calcium concentrations than surface water. Groundwater discharging to Canyon Creek was high in both calcium and sodium. Previous researchers, who have sampled some of these same springs, suggest that many of the thermal springs originate from flow along faults, whereas the non thermal springs flow through porous rocks rich in evaporate minerals. Groundwater has constant 18O and 2H ratios through the year which can help identify groundwater contributions. Rivers which have been sampled more extensively up and downstream show an end member mixing between runoff at that time of year and groundwater. The stable isotope results confirm that the creeks are in baseflow in winter and that most creek water is from groundwater at this time of year. Although stable isotope ratios are not identical between creeks, they are very similar. This means it may be possible to use these stable isotopes as a regional groundwater tracer. Groundwater contribution in these rivers can be identified based on water chemistry. The best results have been obtained where multiple samples have been collected along a reach with repeat samplings at different times of the year. The high TDS of the thermal springs makes this type of groundwater easy to distinguish. In the non thermal systems the best indicator of groundwater is from combining the major ion geochemistry and stable isotopes.