Search

• Development of the nonstationary stochastic paired watershed approach. • Detecting, attributing, and quantifying forest harvest effects on floods. • Forest harvesting affects small and large flood events at large watershed scales. • Probabilistic physics reveals high sensitivity of floods to forest disturbance. • Using probabilistic physics to evaluate the environmental controls on floods. Drawing on advances in nonstationary frequency analysis and the science of causation and attribution, this study employs a newly developed nonstationary stochastic paired watershed approach to determine the effect of forest harvesting on snowmelt-generated floods. Moreover, this study furthers the application of stochastic physics to evaluate the environmental controls and drivers of flood response. Physically-based climate and time-varying harvesting data are used as covariates to drive the nonstationary flood frequency distribution parameters to detect, attribute, and quantify the effect of harvesting on floods in the snow-dominated Deadman River (878 km2) and nested Joe Ross Creek (99 km2) watersheds. Harvesting only 21% of the watershed caused a 38% and 84% increase in the mean but no increase in variability around the mean of the frequency distribution in the Deadman River and Joe Ross Creek, respectively. Consequently, the 7-year, 20-year, 50-year, and 100-year flood events became approximately two, four, six, and ten times more frequent in both watersheds. An increase in the mean is posited to occur from an increase in moisture availability following harvest from suppressed snow interception and increased net radiation reaching the snowpack. Variability was not increased because snowmelt synchronization was inhibited by the buffering capacity of abundant lakes, evenly distributed aspects, and widespread spatial distribution of cutblocks in the watersheds, preventing any potential for harvesting to increase the efficiency of runoff delivery to the outlet. Consistent with similar recent studies, the effect of logging on floods is controlled not only by the harvest rate but most importantly the physiographic characteristics of the watershed and the spatial distribution of the cutblocks. Imposed by the probabilistic framework to understanding and predicting the relation between extremes and their environmental controls, commonly used in the general sciences but not forest hydrology, it is the inherent nature of snowmelt-driven flood regimes which cause even modest increases in magnitude, especially in the upper tail of the distribution, to translate into surprisingly large changes in frequency. Contrary to conventional wisdom, harvesting influenced small, medium, and very large flood events, and the sensitivity to harvest increased with increasing flood event size and watershed area. [ABSTRACT FROM AUTHOR]

We examine the applicability of predicting the daily flow–duration curve (FDC) using mean monthly runoff represented in its stochastic form (MM_FDC) to aid in predictions in ungauged basins, using long-term hydroclimatic data at 73 catchments of humid climate, in the eastern USA. The analysis uses soil hydrological properties, soil moisture storage capacity and the predominant runoff generation mechanism. The results show that MM_FDC did not distinguish the shapes of the upper and lower thirds of the FDC. The upper third is where the precipitation pattern and the antecedent moisture conditions are dominant, while the lower third is where drought-induced low flows and the evapotranspiration effect are prevalent. It is possible to use the MM_FDC to predict the middle third of the FDC (exceedence probabilities between 33% and 66%). The method is constrained by the catchment flow variability (slope of FDC), which changes in accordance with landscape properties and the predominant runoff generation mechanism. [ABSTRACT FROM AUTHOR]

The need for predictions of flow time-series persists at ungauged catchments, motivating the research goals of our study. By means of the Sacramento model, this paper explores the use of parameter transfer within homogeneous regions of similar climate and flow characteristics and makes comparisons with predictions from a priori parameters. We assessed the performance using the Nash-Sutcliffe (NS), bias, mean monthly hydrograph and flow duration curve (FDC). The study was conducted on a large dataset of 73 catchments within the eastern US. Two approaches to the parameter transferability were developed and evaluated; (i) the within homogeneous region parameter transfer using one donor catchment specific to each region, (ii) the parameter transfer disregarding the geographical limits of homogeneous regions, where one donor catchment was common to all regions. Comparisons between both parameter transfers enabled to assess the gain in performance from the parameter regionalization and its respective constraints and limitations. The parameter transfer within homogeneous regions outperformed the a priori parameters and led to a decrease in bias and increase in efficiency reaching a median NS of 0.77 and a NS of 0.85 at individual catchments. The use of FDC revealed the effect of bias on the inaccuracy of prediction from parameter transfer. In one specific region, of mountainous and forested catchments, the prediction accuracy of the parameter transfer was less satisfactory and equivalent to a priori parameters. In this region, the parameter transfer from the outsider catchment provided the best performance; less-biased with smaller uncertainty in medium flow percentiles (40%–60%). The large disparity of energy conditions explained the lack of performance from parameter transfer in this region. Besides, the subsurface stormflow is predominant and there is a likelihood of lateral preferential flow, which according to its specific properties further explained the reduced efficiency. Testing the parameter transferability using criteria of similar climate and flow characteristics at ungauged catchments and comparisons with predictions from a priori parameters are a novelty. The ultimate limitations of both approaches are recognized and recommendations are made for future research. [ABSTRACT FROM AUTHOR]

Summary: This study examines the effect on stand and catchment scale snow processes due to widespread forest disturbance by the Mountain Pine Beetle (MPB; Dendroctonus ponderosae) infestation, for the 1570km2 Baker Creek catchment in the B.C. interior where all healthy mature lodgepole pine (Pinus contorta) stands which dominated the catchment up until 2000 have since died or been salvage logged. Measurements in 2008 and 2009 indicate that this net canopy reduction has reduced peak snowpack and melt rate differences between remaining stands (including large clearcuts, younger regenerating stands and dead mature pine stands), relative to a healthy forest canopy with smaller clearcuts. The Distributed Hydrology Soil Vegetation Model (DHSVM) was calibrated and run at 200m resolution across the catchment, and simulated snowmelt rates and snow-covered area compared relatively well to distributed satellite or ground-based measurements. Snowpack and ablation rates were 10–20% higher during the 2008 and 2009 winters than when running the model for the pre-MPB landscape in 2000 using the same input meteorological data, resulting in almost no difference in the snowcover period. A greater snowpack volume which enters the stream network faster inevitably has implications for streamflows, flood risks and water resources, and these are assessed during the next stage of this research project. [Copyright &y& Elsevier]

Abstract: Scenarios of climate change were generated for the Okanagan Basin, a snow-driven semi-arid basin located in the southern interior region of British Columbia. Three global climate models (GCMs) were used to generate high (A2) and low (B2) emission scenarios; the canadian global coupled model (CGCM2), the Australian developed CSIROMk2, and the HadCM3 model developed at the Hadley Centre in the United Kingdom. The three time periods simulated were 2010–2039 (2020s), 2040–2069 (2050s) and 2070–2099 (2080s). An increase in winter temperature of 1.5–4.0°C and a precipitation increase of the order of 5-20% is predicted by the 2050s. Modelled summer precipitation is more variable with predicted change ranging from zero to a 35% decrease depending on the GCM and emission scenario. Summer temperatures were simulated to increase by approximately 2–4°C. The UBC Watershed Model was used to model the hydrologic response of gauged sub watersheds in the basin under the altered climates. All scenarios consistently predicted an early onset of the spring snowmelt, a tendency towards a more rainfall dominated hydrograph and considerable reductions in the annual and spring flow volumes in the 2050s and 2080s. Of the three climate models, the CGCM2 model provided the most conservative predictions of the impacts of climate change in Okanagan Basin. Simulations based on the CSIROMk2 climate model suggested greatly reduced snowpack and flow volumes despite a sizeable increase in the winter precipitation. The scenarios raise questions over the availability of future water resources in the Okanagan Basin, particularly as extended periods of low flows into upland reservoirs are likely to coincide with increased demand from agricultural and domestic water users. [Copyright &y& Elsevier]

Spatially distributed numerical hydrologic models are useful tools for examining the long-term impact of forest harvesting in mountainous basins on streamflow regime properties. Such models require the input of long-duration subdaily meteorological time series data that are not routinely available in mountainous headwater basins. A relatively simple method is presented for extending short-duration records by using a combined stochastic–empirical technique, and the approach is demonstrated using the Redfish Creek in British Columbia, Canada. Synthetic hourly precipitation, precipitation gradient, air temperature, temperature lapse rate, wind speed, relative humidity, solar beam and diffuse irradiance, and downward longwave irradiance for two station locations are generated in a three-step process: 1) hourly precipitation is generated using a clustered rectangular pulse point process, 2) daily meteorology is generated using a multivariate first-order autoregressive process, and 3) final hourly nonprecipitation meteorology is derived by disaggregating daily meteorology. Seasonal and annual precipitation means are reproduced to within 10% of observed. The skill of the generated nonprecipitation meteorological data to reproduce the statistical properties and diurnal structure of the observed data ranged from good to poor, with bias ranging from 0% to 500% and efficiency ranging from -76 to 0.93. Despite discrepancies in the generated meteorology, a comparison of annual hydrologic fluxes, spatial distribution of winter snow accumulation, flow duration, and average hydrographs, simulated using the Distributed Hydrology Soil Vegetation Model (DHSVM), indicates that model skill shows negligible response to the use of the generated subdaily meteorology. [ABSTRACT FROM AUTHOR]

Evaluates the validity of stream culvert discharge design of the Forest Practices Code of British Columbia. Analysis of annual maximum flows; Economic and environmental impact of flood; Estimation of return period of instantaneous peak flow.

This paper advances the physical understanding of the flow duration curve (FDC) regional variation. It provides a process-based analysis of the interaction between climate and landscape properties to explain disparities in FDC shapes. We used (i) long term measured flow and precipitation data over 73 catchments from the eastern US. (ii) We calibrated the Sacramento model (SAC-SMA) to simulate soil moisture and flow components FDCs. The catchments classification based on storm characteristics pointed to the effect of catchments landscape properties on the precipitation variability and consequently on the FDC shapes. The landscape properties effect was pronounce such that low value of the slope of FDC (SFDC)—hinting at limited flow variability—were present in regions of high precipitation variability. Whereas, in regions with low precipitation variability the SFDCs were of larger values. The topographic index distribution, at the catchment scale, indicated that saturation excess overland flow mitigated the flow variability under conditions of low elevations with large soil moisture storage capacity and high infiltration rates. The SFDCs increased due to the predominant subsurface stormflow in catchments at high elevations with limited soil moisture storage capacity and low infiltration rates. Our analyses also highlighted the major role of soil infiltration rates on the FDC despite the impact of the predominant runoff generation mechanism and catchment elevation. In conditions of slow infiltration rates in soils of large moisture storage capacity (at low elevations) and predominant saturation excess, the SFDCs were of larger values. On the other hand, the SFDCs decreased in catchments of prevalent subsurface stormflow and poorly drained soils of small soil moisture storage capacity. The analysis of the flow components FDCs demonstrated that the interflow contribution to the response was the higher in catchments with large value of slope of the FDC. The surface flow FDC was the most affected by the precipitation as it tracked the precipitation duration curve (PDC). In catchments with low SFDCs, this became less applicable as surface flow FDC diverged from PDC at the upper tail (> 40% of the flow percentile). The interflow and baseflow FDCs illustrated most the filtering effect on the precipitation. The process understanding we achieved in this study is key for flow simulation and assessment in addition to future works focusing on process-based FDC predictions. [ABSTRACT FROM AUTHOR]

Summary Hydrological models for headwater catchments have typically excluded deep groundwater flow based on the assumption that it is a negligible component of the water budget. This study tests this assumption using a coupled surface water–groundwater model to explore the potential contribution of deep groundwater recharge to the bedrock in a snowmelt-dominated headwater catchment (Upper Penticton Creek 241) in the Okanagan Basin, British Columbia. Recharge to the bedrock is estimated at ∼27% of the annual precipitation over the period 2005–2010, recognizing the uncertainty in this estimate due to data limitations, parameter uncertainty and calibration errors. A specified outward flux from the catchment boundary within the saturated zone, representing ∼2% of the annual water budget, was also included in the model. This outward flux contributes to cross-catchment flow and, ultimately, to groundwater inflow to lower elevation catchments in the mountain block. This modeling exercise is one of the first in catchment hydrologic modeling within steep mountainous terrain in which the bedrock is not treated as impermeable, and in which recharge to the bedrock and discharge to the surrounding mountain block were estimated. [ABSTRACT FROM AUTHOR]

Abstract: The highly diverse shifting-mosaic of forest patches of an alluvial forest within the Carmanah River valley on the west coast of Vancouver Island, British Columbia was studied to examine the hydrogeomorphic disturbance regime that structures it. We used a landscape-scale analysis to quantify historical channel migrations and changes in the extent of specific forest types. This GIS-based analysis using a 70-year aerial photographic record was complemented by field-based research. Thirty-eight plots containing 4509 trees were sampled for forest structure, age, and elevation above the contemporary channel. These data, including a vegetation chronosequence spanning over 500years, were used to examine channel and landscape dynamics. Our findings support a general conceptual model that describes cycles of patch development and destruction in unconfined alluvial forests of the Pacific Coastal Ecoregion. Over the past century, Carmanah River has eroded nearly 30% of the alluvial forest in this study area, and approximately 65% over the past 500years. At least 80% of the 2007 channel was forested area within the past 70years. Younger landforms were disturbed more frequently than mature forest patches, which suggest that as biogeomorphic succession progresses the likelihood of future disturbance decreases. Estimated half lives of landforms ranged from 24years for pioneer bars to over 1500years for old growth terraces. Years of regional high magnitude floods resulted in a net loss of floodplain forest area indicating that disturbance was climate driven in this pluvial watershed, whereby rain events result in flood disturbance that converted forests to channel. These events initiate a subsequent course of vegetation succession and geomorphic development, and often result in the deposition of large wood that modifies the channel environment and contributes to channel avulsion and further hydrogeomorphic disturbance. The composition of the landscape is a reflection of the balance between the disturbance rate and successional development. We also observed a relationship between landscape composition and watershed size. Specifically, the ratio of mature to developing alluvial forests was higher in this smaller watershed compared to larger watersheds in the region. Results imply that larger flood events predicted to occur with climate change may change the disturbance regime of floodplain forests and alter landscape composition. [Copyright &y& Elsevier]

Summary: The intricacies of hydrometeorological process interactions with basin physiography in snow-dominated catchments may only be acknowledged through the recognition and study of the spatiotemporal heterogeneity of runoff generation and groundwater dynamics. Subsurface flow mechanisms have a direct influence on stream runoff generation. In order to better understand the coupling of these processes, the linkages between stream runoff generation and groundwater dynamics need to also be considered. This study collectively examines the spatiotemporal variability of stream runoff generation and groundwater dynamics in a snow-dominated catchment in south-central British Columbia by monitoring basin outlet flow, continuous groundwater fluctuations at 9 locations, and instantaneous flows at 42 locations throughout the stream network. Observed groundwater levels were found to be more responsive to inputs in the upper hillslopes with flow accumulation in the lower slopes, and strong groundwater level correlations with stream runoff generally decreased with increasing distances from streams. The specific discharge of reach contributing areas was found to be significantly correlated to various physiographical parameters including contributing area, an index of average hillslope flow velocity, reach length, elevation, and slope. Nested subcatchment stream runoff scaled positively in a power law relation with contributing area (r2 =0.74–0.88), and sub-basin discharge was found to scale near linearly with drainage areas ranging 0.9ha–4.74km2. The findings of this study are in general agreement with the concept of the transmissivity feedback mechanism; lateral inflows to streams contributed relatively new water during periods of greater runoff contribution, whereas during the low-flow summer period, hillslopes in a hydrologically dry state produced waters that experienced longer mean residence times. In order to better understand the coupling of subsurface flow mechanisms and stream runoff generation, it is suggested that future work should emphasize spatially-intensive continuous measurements of stream network flows in combination with comprehensive groundwater level monitoring. [Copyright &y& Elsevier]

Monitoring continuous changes in snowpack dynamics and its meteorological drivers is critical for understanding key aspects of water resources, climate variability, and ecology. While manual snow surveys have traditionally been used to evaluate snow processes, their high costs and discrete measurements can lead to biased estimations of accumulation and ablation rates. Ultrasonic range sensors offer an alternative to continuously monitor snow depth but their widespread employment has been limited because of high prices. This paper describes the development of an inexpensive prototype ultrasonic sensor suite characterized by a ready-to-use stand-alone design and flexibility to incorporate additional meteorological instruments. The performance of 48 units was tested during a winter season in central British Columbia, recording snow depth and air temperature data consistent with those from nearby weather stations and manual measurements. Despite a relatively small underestimation of snow depth due to known, repairable reasons, the sensor system demonstrated reliability for research and operations. [ABSTRACT FROM AUTHOR]

Summary: Modeling the effect of forest clear-cutting with a distributed hydrological model can be used to detect hydrologic changes as an alternative to paired-catchment studies, and also to estimate the hydrologic sensitivity of a catchment to assist in forest management decisions. To model the effect of clear-cutting in a snow dominated forested watershed, a model needs to be able to simulate effects of all of the main controls on snow accumulation and melt. However, most studies that used a distributed modeling approach relied on one or a few sites for model testing. In this study, we employ a stratified nested sample layout that was specifically designed to test distributed models, involving measurement of snow water equivalent (SWE) under forest and in clearcuts over a range of elevations and aspects. To test the ability of a model to simulate the main controls on the spatial distribution of SWE, spatial gradients of observed and simulated SWE in relation to topographic and vegetation controls are computed using regression analysis. Comparison of observed and simulated gradients helps to highlight model weaknesses. The approach is applied to evaluate the snow algorithms in the distributed hydrology soil and vegetation model (DHSVM) using data collected in Cotton Creek, a snow dominated forested watershed in south-eastern British Columbia. SWE measurements were made from 2005 to 2008, covering peak snow accumulation and snow melt. Albedo decay and canopy transmittance were found to be the two processes that DSHVM version 3.0 did not simulate well enough to predict basin average differences between forests and clearcuts properly. After replacing the internal albedo decay functions with functions obtained from snow albedo measurements and changing the canopy transmittance function in the model, DHSVM was able to reproduce the major spatial patterns derived from snow surveys. Model performance is better during winter up to the peak snow accumulation than during snow melt. Spatial patterns of peak snow accumulation in the snow-rich year 2006 can be modeled better than those after the warm winter of 2005. The influence of aspect on snow accumulation and snowmelt is underestimated by DHSVM. While we have focused specifically on DHSVM, the methods developed in this study should be generally useful for model testing purposes and important in the context of interpretation of modeling results, in particular when dealing with large spatial datasets. [Copyright &y& Elsevier]

Summary: Snow accumulation and melt are the two most important processes contributing to the spring freshet, the main hydrological event in snow-dominated watersheds. The magnitude and timing of the spring freshet is determined by the spatial variability of melt rates and peak snow water equivalent (SWE). Although it is known that the spatial variability of SWE is controlled by topography and vegetation, their combined influence on snow accumulation and melt at the watershed-scale has only been addressed in a few studies. The high small-scale variability of a snowpack combined with the classical sampling technique along transects makes it difficult to estimate how much of the observed spatial variability in SWE can be attributed to topographic and vegetative controls such as elevation, aspect, wind, and forest type, and how much of it is local variability. A nested stratified sampling design was developed for the snow courses in this study, where small-scale variability at the plot-scale was integrated by 12 snow density and 60 snow depth measurements, while the variability at the watershed-scale was captured by 19 strata, defined by elevation, aspect, and forest cover (either clearcut or forest). Snow course sampling was carried out during the spring of 2005, a year with 25% below average snow accumulation, and in spring of 2006, a year with 25% above average snow accumulation for the region. Snow accumulation in both years and snowmelt for a period in the spring of 2006 are analysed with multiple linear regressions and spatial autocorrelation of SWE with variogram analysis. The combination of elevation, aspect, and forest cover explained about 80–90% of the large-scale variability in snow accumulation for both years. At the plot-scale the coefficient of variation was higher under the forest canopy than in clearcuts, but did not relate to topographic controls. Elevation had the greatest influence on SWE in both years, with a steeper gradient in the snow rich year, whereas aspect had a similar effect in both years. In the mild winter of 2005, forests accumulated 39% less snow than clearcuts, while in 2006, forests had 27% less snow than clearcuts. Relative to the mean, aspect had a greater influence on snow accumulation than forest cover in the mild winter of 2005. Elevation played the most important role in snowmelt, while both aspect and forest cover had a comparable effect with slightly less influence than elevation. Almost all variograms show a clear sill, confirming that in the majority of cases, a 60×60m plot size was adequate in capturing the variability of the plot-scale. In general, the degree of spatial autocorrelations is persistent in the snowmelt season for both years. This study has shown that it is possible to separate large-scale variability created by topography and forest cover from plot-scale variability through specifically designed snow surveys. The methods and data set of this study are useful in testing spatially-distributed snowmelt models and in developing new algorithms that reflect the relationships between the factors controlling the spatial variability of SWE. [Copyright &y& Elsevier]

The research activity described in this report is a comprehensive regional assessment of the impacts of climate change on water resources and options for adaptation in the Okanagan Basin. The ultimate goal of the project is to develop integrated climate change and water resource scenarios to stimulate a multistakeholder discussion on the implications of climate change for water management in the region. The paper describes two main objectives: (a) providing a set of research products that will be of relevance to regional interests in the Okanagan, and (b) establishing a methodology for participatory integrated assessment of regional climate change impacts and adaptation that could be applied to climate-related concerns in Canada and other countries. This collaborative study has relied on field research, computer-based models, and dialogue exercises to generate an assessment of future implications, and to learn about regional views on the prospects for adaptation. Along the way, it has benefited from strong partnerships with governments, researchers, local water practitioners, and user groups. Building on the scenario-based study components, and a series of interviews and surveys undertaken for the water management and adaptation case study components, a set of stakeholder dialogue sessions were organized which focused on identifying preferred adaptation options and processes for their implementation. Rather than seeking consensus on the "best" option or process, regional interests were asked to consider a range of available options as part of an adaptation portfolio that could address both supply side and demand side aspects of water resources management in the Okanagan. [ABSTRACT FROM AUTHOR]