Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Modelling the effects logging roads on the streamflow of a mountainous, snow dominated watershed Calvert, Paula Panarat 2003

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

Item Metadata

Download

Media
831-ubc_2003-0387.pdf [ 16.54MB ]
Metadata
JSON: 831-1.0075093.json
JSON-LD: 831-1.0075093-ld.json
RDF/XML (Pretty): 831-1.0075093-rdf.xml
RDF/JSON: 831-1.0075093-rdf.json
Turtle: 831-1.0075093-turtle.txt
N-Triples: 831-1.0075093-rdf-ntriples.txt
Original Record: 831-1.0075093-source.json
Full Text
831-1.0075093-fulltext.txt
Citation
831-1.0075093.ris

Full Text

• MODELLING THE EFFECTS LOGGING ROADS ON THE STREAMFLOW OF A MOUNTAINOUS, SNOW DOMINATED WATERSHED by PAULA PANARAT CALVERT B.Sc, The U n i v e r s i t y of Arizona, 1996 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTERS OF SCIENCE i n THE FACULTY OF GRADUATE STUDIES Department of Forest Resources Management; Fac u l t y of Forest We accept t h i s t h e s i s as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA August 2003 © Paula Panarat C a l v e r t , ,2003 In presenting this thesis in partial fulfillment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of Forest KtSoofCCS, MM*^>MUA The University of British Columbia Vancouver, Canada A B S T R A C T Logging roads have the ability to alter streamflow by converting subsurface flow paths to surface flow paths. The two primary mechanisms by which this occurs are the interception of subsurface flow at road cuts and production of infiltration excess overland flow on road surfaces. The impact of subsurface flow interception in a snow dominated regime, at basin and sub-basin scales, was explored in this study with the application of the hydrologic model, Distributed Hydrology-Soil-Vegetation Model (DHSVM). Sensitivities of the model to road cut depth, increased stream density and increased road density were also examined in addition to the comparison of modelled versus observed tributary flow. Redfish Creek and two of its sub-basins are the subject watersheds located in the Kootenay Mountains of the British Columbia interior. At the 26 km 2 scale of Redfish Creek basin, no substantial alteration in annual yield or annual peak flow occurred resulting from the presence of 16 km of logging roads. There was also no appreciable change due to modelled minimal stream extension to culverts connected to the stream via surface flow or the addition of skid trails and an abandoned road both modelled in the form of increased road density. Greater changes occurred at the sub-basin scale. Redfish Upper Tributary, in a 1 km 2 sub-basin, experienced decreases of less than 5% in annual yield and annual peak flow. The largest changes occurred for the South tributary, in a 0.8 km sub-basm. The annual yield increased by 19.6% and the annual peak flow increased by 11.3%. Redfish Creek model sensitivities to the refinements in representing logging roads and their effects are greater at smaller scales. As compared against a roadless model, the addition of logging roads to Redfish Upper tributary did not noticeably improve the fit of modelled flows with observed flows. n TABLE OF CONTENTS Abstract Table of Contents List of Figures List of Tables ix Acknowledgements Chapter 1: Introduction 1 1.1 Overview 1.2 Runoff generation and the effects of forest roads 1.3 Modelling effects of forest operations on watershed hydrology 1.3.1 Evaluation of available models 5 1.3.2 Previous application of the Distributed Hydrology-Soil-Vegetation Model (DHSVM) to Redfish Creek 5 1.3.3 Incorporating road effects in D H S V M 6 1.4 Specific objectives and approach of this study 9 Chapter 2: Methods 10 2.1 Redfish Creek site description 10 2.2 Distributed Hydrology-Soil-Vegetation Model 10 2.3 Hydrologic model of Redfish Creek 12 2.3.1 Data requirements 12 2.3.2 Redfish Creek model calibration 19 2.4 Incorporation of roads in the Redfish Creek model 21 2.4.1 Methods overview 21 2.4.2 Field data collection 22 2.4.3 Application of field data 23 2.4.4 Preparation of input file sets for the existing road layout 24 2.4.5 Modelling logging road impact 27 2.4.6 Comparison with Redfish Upper observed flows 28 iii Chapter 3: Results 29 3.1 Overview of progressive modelling refinements 29 3.2 Run B - Addition of primary logging roads and model sensitivity to cut depth 30 3.2.1 Redfish Creek streamflow 30 3.2.2 Redfish Upper tributary flow 35 3.2.3 South tributary flow 37 3.3 Runs Bb, C and D - Comparison of the effect of primary logging roads, stream network expansion, and skid trails and abandoned roads 40 3.3.1 Redfish Creek streamflow 42 3.3.2 Redfish Upper tributary flow 44 3.3.3 South tributary flow 47 3.4 Redfish Upper tributary modelled and observed flows 50 Chapter 4: Discussion 54 4.1 Overview 54 4.2 Streamflow response to the addition of logging roads at the basin and sub-basin scale 54 4.3 Model sensitivity to cut depth variations 56 4.4 Effect of surface flow connections from culverts to the stream network 56 4.5 Hydrologic influence of skid trails and abandoned roads 57 4.6 Enhancement of sub-basin model with road features 57 Chapter 5: Conclusion 59 5.1 S ummary o f main findings 59 5.2 Suggestions for further research 60 References 61 Appendix: Tabulated and graphed results for 1995 65 iv LIST O F FIGURES Figure 1. West Arm Demonstration Forest 6 Figure 2. Locations of tributary gauge and climate stations in Redfish Creek watershed 7 Figure 3. Topographic map with locations of harvest and roads 11 Figure 4. Vegetation types for Redfish Creek watershed 15 Figure 5. Genetic material classification used to determine soil types for Redfish Creek 17 Figure 6. Soil depth for Redfish Creek watershed 18 Figure 7. Discharge differences for Redfish Creek due to the varying road cut depths of Runs Ba, Bb, and Be during water year 1995 32 Figure 8. Discharge differences during the annual snowmelt peak for Redfish Creek due to the varying road cut depths of Runs Ba, Bb, and Be in 1995 34 Figure 9. Discharge differences during a rain on snow event for Redfish Creek due to the varying road cut depths of Runs Ba, Bb, and Be in 1995 34 Figure 10. Discharge differences for Redfish Upper Tributary due to the varying road cut depths of Runs Ba, Bb, and Be during water year 1995.....36 Figure 11. Discharge differences during annual snowmelt peak for Redfish Upper Tributary due to the varying road cut depths of Runs Ba, Bb, and Be in 1995 38 Figure 12. Discharge differences during a rain on snow event for Redfish Upper Tributary due to the varying road cut depths of Runs Ba, Bb, and Be in 1995 38 Figure 13. Discharge differences for South Tributary due to the varying road cut depths of Runs Ba, Bb, and Be during water year 1995 39 Figure 14. Discharge differences during the annual snowmelt peak for South Tributary due to the varying road cut depths of Runs Ba, Bb, and Be in 1995 .41 Figure 15. Discharge differences during a rain on snow event for South Tributary due to the varying road cut depths of Runs Ba, Bb, and Be inl995 41 Figure 16. Discharge differences for Redfish Creek due to the varying simulation scenarios of Runs Bb, C, and D during water year 1995 43 v Figure 17. Discharge differences during the annual snowmelt peak for Redfish Creek due to the varying simulation scenarios of Runs Bb, C, and D in 1995 45 Figure 18. Discharge differences during a rain on snow event for Redfish Creek due to the varying simulation scenarios of Runs Bb, C, and D in 1995 45 Figure 19. Discharge differences for Redfish Upper Tributary due to the varying simulation scenarios of Runs Bb, C, and D during water year 1995 46 Figure 20. Discharge differences during the annual snowmelt peak for Redfish Upper Tributary due to the varying simulation scenarios of Runs Bb, C, and D in 1995 48 Figure 21. Discharge differences during a rain on snow event for Redfish Upper Tributary due to the varying simulation scenarios of Runs Bb, C, and D i n ! 995 48 Figure 22. Discharge differences for South Tributary due to the varying simulation scenarios of Runs Bb, C, and D during 1995 49 Figure 23. Discharge differences during the annual snowmelt peak for South Tributary due to the varying simulation scenarios of Runs Bb, C, and D in 1995 51 Figure 24. Discharge differences during a rain on snow event for South Tributary due to the varying simulation scenarios of Runs Bb, C, and D in 1995 51 Figure 25. Modelled Redfish Upper Tributary response to logging roads compared with the modelled roadless scenario and observed flow for water year 1995 52 Figure 26. Modelled Redfish Upper Tributary response to logging roads compared with the modelled roadless scenario and observed flow for water year 1996 52 Figure 27. Modelled Redfish Upper Tributary response to logging roads compared with the modelled roadless scenario and observed flow for water year 1997 53 Figure 28. Discharge differences for Redfish Creek due to the varying road cut depths of Runs Ba, Bb, and Bc during water year 1993 67 v i Figure 29. Discharge differences for Redfish Creek due to the varying road cut depths of Runs Ba, Bb, and Be during water year 1994 68 Figure 30. Discharge differences for Redfish Creek due to the varying road cut depths of Runs Ba, Bb, and Be during water year 1996 69 Figure 31. Discharge differences for Redfish Creek due to the varying road cut depths of Runs Ba, Bb, and Be during water year 1997 70 Figure 32. Discharge differences for Redfish Upper Tributary due to the varying road cut depths of Runs Ba, Bb, and Be during water year 1993 72 Figure 33. Discharge differences for Redfish Upper Tributary due to the varying road cut depths of Runs Ba, Bb, and Be during water year 1994 73 Figure 34. Discharge differences for Redfish Upper Tributary due to the varying road cut depths of Runs Ba, Bb, and Be during water year 1996 74 Figure 35. Discharge differences for Redfish Upper Tributary due to the varying road cut depths of Runs Ba, Bb, and Be during water year 1997 75 Figure 36. Discharge differences for South Tributary due to the varying road cut depths of Runs Ba, Bb, and Be during water year 1993 77 Figure 37. Discharge differences for South Tributary due to the varying road cut depths of Runs Ba, Bb, and Be during water year 1994 78 Figure 38. Discharge differences for South Tributary due to the varying road cut depths of Runs Ba, Bb, and Be during water year 1996 79 Figure 39. Discharge differences for South Tributary due to the varying road cut depths of Runs Ba, Bb, and Be during water year 1997 80 Figure 40. Discharge differences for Redfish Creek due to the varying simulation scenarios of Runs Bb, C, and D during water year 1993 82 Figure 41. Discharge differences for Redfish Creek due to the varying simulation scenarios of Runs Bb, C, and D during water year 1994 83 Figure 42. Discharge differences for Redfish Creek due to the varying simulation scenarios of Runs Bb, C, and D during water year 1996 84 Figure 43. Discharge differences for Redfish Creek due to the varying simulation scenarios of Runs Bb, C, and D during water year 1997 85 Figure 44. Discharge differences for Redfish Upper Tributary due to the varying simulation scenarios of Runs Bb, C, and D during water year 1993 87 vii Figure 45. Discharge differences for Redfish Upper Tributary due to the varying simulation scenarios of Runs Bb, C, and D during water year 1994 88 Figure 46. Discharge differences for Redfish Upper Tributary due to the varying simulation scenarios of Runs Bb, C, and D during water year 1996 89 Figure 47. Discharge differences for Redfish Upper Tributary due to the varying simulation scenarios of Runs Bb, C, and D during water year 1997 90 Figure 48. Discharge differences for South Tributary due to the varying simulation scenarios of Runs Bb, C, and D during 1993 92 Figure 49. Discharge differences for South Tributary due to the varying simulation scenarios of Runs Bb, C, and D during 1994 93 Figure 50. Discharge differences for South Tributary due to the varying simulation scenarios of Runs Bb, C, and D during 1996 94 Figure 51. Discharge differences for South Tributary due to the varying simulation scenarios of Runs Bb, C, and D during 1997 95 vin LIST O F T A B L E S Table 1. Vegetation class descriptions 14 Table 2. Soil class descriptions 16 Table 3. Series of Model Runs Performed 25 Table 4. Extent of Modelled Features 29 Table 5. Changes to 1995 annual yield for Redfish Creek, Redfish Upper Tributary, and South Tributary 31 Table 6. Changes to 1995 annual peak flow for Redfish Creek, Redfish Upper Tributary, and South Tributary 31 Table 7. Redfish Creek annual yield changes resulting from road cut depth variation 66 Table 8. Redfish Creek annual peak flow changes resulting from road cut depth variation 66 Table 9. Redfish Upper Tributary annual yield changes resulting from road cut depth variation 71 Table 10. Redfish Upper Tributary annual peak flow changes resulting from road cut depth variation 71 Table 11. South Tributary annual yield changes resulting from road cut depth variation 76 Table 12. South Tributary annual peak flow changes resulting from road cut depth variation 76 Table 13. Redfish Creek annual yield changes resulting from stream extension, skid trails and abandoned roads 81 Table 14. Redfish Creek annual peak flow changes resulting from stream extension, skid trails and abandoned roads 81 Table 15. Redfish Upper Tributary annual yield changes resulting from stream extension, skid trails and abandoned roads 86 Table 16. Redfish Upper Tributary annual peak flow changes resulting from stream extension, skid trails and abandoned roads 86 Table 17. South Tributary annual yield changes resulting from stream extension, skid trails and abandoned roads 91 Table 18. South Tributary annual peak flow changes resulting from stream extension, skid trails and abandoned roads 91 ix ACKNOWLEDGEMENTS This research would not have been possible without the help and support of many people. I would firstly like to thank my advisor, Dr. R. Daniel Moore, who saw me through the completion of this thesis. I greatly appreciate his patience equally as much as his invaluable guidance. Additional thanks go out to Dr. Roy Sidle for his input and support. I am also thankful for Dr. Younes Alila's belief in me and the role he played in my acceptance to the graduate program. Adam Stonewall is especially deserving of recognition. He has been a very important part of my life and has supported me with love through all the successes and supposed catastrophes. He fills many roles in this accomplishment: field assistant, technical advisor, lab mate and confidant. This work would not have been possible without the technical assistance of those that are or were associated with the University of British Columbia: Jos Beckers, Jerry Maedel, Erik Schiefer, Harry Verwoerd, and Andrew Whitaker. Emilee Fanjoy, Peter Jordan, and Dave Toews of the Ministry of Forests were of great help. Valuable support was also provided by the Civil Engineering Department at the Univeristy of Washington namely Laura Bowling, Jonathan La Marche, Pascal Stork, and James Van Shaar. Thanks goes to the field assistance of Dino Gnoto without whom I'd still be trying to finding my way out of Redfish Creek watershed. The camaraderie of my fellow computer lab mates, Takashi Gomi, Joyce Kim, David Luzi, Ahmed Mtiraoui, Julie Orban, and Yuzhang Wang, made my work even more enjoyable. My appreciation goes out to my best friends of many years Zulma Biddle and Jeanie Comstock for knowing that I would eventually finish. Thanks also to each member of my family who hoped I would eventually finish. Many thanks to the Research and Development Department of Stormwater Management Inc. who have been very supportive and understanding through the final stages of my work. x C H A P T E R 1: INTRODUCTION 1.1 Overview The effects of forestry activities on streams are of great interest to those researching and preserving stream health. Logging roads may alter streamflows in different ways than forest harvesting. Removing the forest cover alters vertical exchange processes such as interception loss, snowmelt and transpiration (Bosch and Hewlett, 1982; Gray and Prowse, 1993). However, road construction more drastically affects the lateral movement of water downslope to stream channels. Roads intercept subsurface flows at road cuts and route them more rapidly as surface flow via ditches, culverts, and natural drainage features (Wemple, 1994, Croke and Mockler, . 2001.). In addition, constructed haul roads are typically highly compacted, and the resulting low permeability can inhibit the infiltration of rainfall on road surfaces, thereby producing Hortonian overland flow (Gardner and Chong, 1990). In these ways, logging roads may affect peak flow timing and magnitude, annual yield quantities, and low flow levels. There is currently vigorous debate about the hydrologic effects of logging roads and their significance in relation to downstream flooding (Jones and Grant, 1996; La Marche and Lettenmaier, 1998). This debate is fueled, in part, by the paucity of field experiments that have focused on forest roads, in isolation from the effects of forest removal and natural catchment processes. Given the considerable cost involved and time required to conduct properly controlled catchment-scale experiments, it is unlikely that this situation will be rectified in the near future. An alternative approach to field experiments to assess relative impacts is the application of computer simulation models. The general objective of this research was to apply a physically based, spatially distributed hydrologic model to examine the effect of logging roads on streamflow at the basin and sub-basin scales in Redfish Creek, a mountainous, snowmelt-dominated watershed located in the West Kootenay region of British Columbia, Canada. The remainder of this chapter reviews runoff generation in forested catchments, with attention to the effects of forest roads. It then introduces the application of the Distributed Hydrology-Soil-Vegetation Model (DHSVM) for predicting the hydrologic effects of forest practices, then ends with the specific objectives of the research. 1 1.2 Runoff Generation and the Effects of Forest Roads Fundamental to analyzing the potential results of forest management practices on the hydrologic regime is the understanding of physical mechanisms operating in the geographic area of interest. Many physical processes are involved in the conversion of rainfall and snowmelt to streamflow. Most pertinent to this study is runoff generation through subsurface flow, the dominant mechanism of temperate, forested mountainous terrain. Research has contributed to advancing theories of watershed response to runoff. Since the infiltration capacity for the undisturbed forest floor is typically much higher than rates of input from snowmelt and rainfall, Hortonian overland flow is not often observed. Betson (1964) suggested a partial area concept in which a relatively constant portion of the basin contributes to direct runoff. Hewlett and Hibbert (1967) promoted the variable source area concept, in which direct runoff is produced by rapidly draining soils which would extend the length of stream channels and expand contributing riparian zones. Areas contributing to direct runoff are not constant but expand and contract with the progression of rainfall events. More recently a hydrogeomorphic concept of runoff generation has been proposed based on studies in Japan (Sidle et al., 2000). This model suggests non-linear responses in subsurface flow and discharge from geomorphic hollows dependent on antecedent moisture (Sidle et al., 2000, 2001; Tsuboyama et al., 2000). For all of the modern runoff concepts, runoff delivery to forest streams is dominated by subsurface flow (Bonell, 1993). Aspects of forest subsurface characteristics allow for the rapid conveyance of snowmelt and rainfall to stream channels. It has been shown that macropores, as a preferential flow pathway, are integral to subsurface flow of forested watersheds (Chamberlin, 1972; Mosley, 1982; Sidle et al., 2001). Macropores that form preferential flow networks in the soil matrix are associated with live and decayed plant roots and borrowed paths of small animals. The macropore presence contributes to an effective hydraulic conductivity higher than the hydraulic conductivity of the soil matrix alone (Germann and Beven, 1981; Buttle and House, 1997). Mosley (1982) showed that macropores are capable of transmitting a large portion of subsurface flow in forest soils in New Zealand. However, he found it difficult to characterize the temporal and spatial contribution of macropores in transmitting stormflow. At a hillslope segment approximately 10 m wide, Hutchinson and Moore (2000) found that macropores can shift water flow across a slope, and dominate over the often assumed topographic controls on subsurface 2 flow. Sidle et al. (2000; 2001) noted that preferential flow networks may develop and 'self-organize' in response to increasing antecedent moisture. Preferential flow paths also influence snowmelt related throughflow. At a site in coastal . British Columbia, Kim et al. (2003) observed unsteady state subsurface flow response to snowmelt. The activation of preferential flow paths was suggested as the cause of the throughflow response. Roberge and Plamondon (1987) show that pipe flow contribution can be a significant component of snowmelt throughflow when activated over the short time periods when the saturation level reaches the mineral-organic soil contact at a site 80 km north of Quebec City. Both studies report fluctuations in throughflow rates that mimic the diurnal patterns of snowmelt. Kim et al. (2003) also found that the timing of hillslope throughflow peaks either matched or preceded streamflow peaks during diurnal snowmelt cycles. The presence of logging roads interrupts the natural basin hydrology through two primary mechanisms: Hortonian overland flow operating on road surfaces and the interception of subsurface flow at road cuts. Due to compaction, the infiltration capacities for road surfaces are much less than those found on undisturbed forest soils. The reduction of hydraulic conductivity, which controls infiltration, indicates a reduction in infiltration capacities. Ziegler and Giambelluca (1997) measured median saturated hydraulic conductivities of 146.1 mm/h for forested surfaces and 2.3 mm/h for road surfaces in the mountains of Thailand. For a range of road surface textures in three western US states, Luce and Cundy (1994) estimated hydraulic conductivities ranging from 5 x 10"5 to 8.82 mm/h. The reduced infiltration capacities of roads produce Hortonian runoff that is routed either to a ditch or the fill slope. Water carried down the road, to the ditch and through a culvert, has the potential to travel quickly as surface flow. This may result in shortened travel time to the stream channel and higher peak flows. In most paired catchment experiments, forest cutting has occurred either at the same time as or shortly following road construction. It is therefore not possible in most studies to separate the two effects statistically. In the few studies where there was a delay between road construction and harvesting, the period of road-only treatment was short, typically one or two years. In a study in the Oregon coast range, the only significant change in peak flow was in a 0.4 km 2 watershed in which roads occupied 12% of the total area (Harr et al., 1975). No statistically significant change in peak flow occurred in the other study watersheds, where roads occupied only 3-5% of the watershed areas. Ziemer (1981) did not detect a change in streamflows resulting from the addition of logging roads that covered approximately 5% of a 3 4.24 km 2 basin in northern California. King and Tennyson (1984) found a significant increase (30.5%) in the 25% exceedence flow for an 83.8 ha watershed in northern Idaho where roads occupied 3.9% of the area. However, their statistical approach placed their analysis under question (Teti and Northway, 1986). The hydrologic influence of flow interception depends on two factors: (1) the proportion of downslope flow that is intercepted, and (2) the connectivity between the road drainage system and the stream network. The proportion of intercepted flow will depend on soil depth, geology and the depth of the road cut. Megahan (1972) found that troughs installed along road cuts in two first-order watersheds in the Idaho batholith collected a significant amount of intercepted flow. He also estimated that approximately 65% of the subsurface flow passed below the road possibly by way of fractures in the bedrock. Whereas Hutchinson and Moore (2000) showed that essentially all of the downslope flow from a hillslope segment underlain by compacted till was intercepted at a road cut. King and Tennyson (1984) found that a midslope road was capable of intercepting 67% of the subsurface flow from the watershed. About 61% of the road length had cut slopes greater than 6 m in vertical height. Several studies have estimated the increase of stream density due to stream channel extension caused by surface flow connectivity of road systems (Montgomery, 1994; Wemple, 1994; Croke and Mockler, 2001). Ditchflow draining directly to a stream-crossing culvert and to drainage relief culverts that flow to an incised gully creates an efficient method of flow delivery to the stream channel. Increasing the speed of flow to the stream channel may result inJa general desynchronization of stream flows or shorter time to peak. In the H.J. Andrews Experimental Forest in the western Cascades of Oregon, the estimated drainage density increase due to stream extension in Lookout Creek was 48% and the increase in Blue River was 50% (Wemple et al., 1996). In an examination of the effect of road drainage concentration, Montgomery (1994) found that a 1.2 km 2 basin located in southern Sierra Nevada, California had an increased drainage density from 2.8 to 4.4 km/km2 due to roads. 4 1.3 Modelling Effects of Forest Operations on Watershed Hydrology 1.3.1 Evaluation of Available Models Modelling anthropogenic effects, such as the presence of logging roads and land use change, on streamflow requires the application of a distributed physically based model. A distributed model allows for spatial variation of watershed parameters such as soil and vegetation. For example, modelling the impact of logging roads can be accomplished by applying variables that rerouting subsurface flow in specified areas of the basin after model calibration. Whereas variable changes in a lumped model affect the entire watershed, a physically based, distributed model allows for a physical and spatial representation of hydrologic mechanisms. Without a physically based approach to modelling, calibration would be based on developing empirical or statistical relations between parameters and generated streamflow. Empirical models may be effective in predicting streamflow given meteorological inputs. However, the model would not be capable of accurately predicting effects of land use on streamflow since it does not rely on physically based governing equations. The D H S V M (Wigmosta et al, 1994) is well suited for the objective of modelling the impacts on basin hydrology due to spatial variation of logging practices in a snowmelt dominated regime (Whitaker et al., 1998). As a distributed, physically based model, D H S V M contains the subroutines required to meet the study objectives. The subroutines simulate snowmelt, precipitation interception by vegetation, evapotranspiration, sublimation, subsurface percolation, saturated subsurface flow, overland flow and channel flow. Most importantly, a road algorithm is contained within the model that allows for an analysis of road influences on streamflows. 1.3.2 Previous Application of D H S V M to Redfish Creek A snowmelt dominated watershed has been modelled using D H S V M Whitaker et al. (2002, 2003). Located in the Kootenay Mountains near Nelson, British Columbia, Redfish Creek * 2 drains a 26 km south facing watershed. The experimental watershed is contained in the West Arm Demonstration Forest controlled by the Ministry of Forests, Nelson District (Figure 1). Four climate stations are located in Redfish Creek watershed. Alpine, Burn, Cabin, 5 and Ross climate stations are shown in Figure 2. Stream gauge locations for Redfish Creek and a Redfish Upper Tributary are also shown. Figure 1. West Arm Demonstration Forest (courtesy of Ministry of Forests) The first study applied to Redfish Creek using D H S V M involved the evaluation of peak flows changes due to harvesting within varying elevation bands (Whitaker et al., 2002). It was found that the greatest increases to peak flows occurred for vegetation removal within the upper 80-40% of the watershed, in other words, excluding the lower 20% and upper 40% of the watershed area. Whitaker et al. (2003) conducted a subsequent study assessing the ability of D H S V M to model internal hydrologic processes such as tributary runoff generation and the spatial and temporal distribution of snow accumulation and melt. 1.3.3 Incorporating Road Effects in D H S V M The original version of DHSVM, which was applied by Whitaker et al. (2002, 2003), did not incorporate algorithms to represent the hydrologic effects of forest roads. A more recent 6 Figure 2. Locations of tributary gauge and climate stations in Redfish Creek watershed 7 version of D H S V M has been developed, which includes such algorithms (Wigmosta and Perkins, 1997). Two studies focusing on the use of the road algorithm were performed in the Department of Civil Engineering at the University of Washington (Bowling and Lettenmaier, 1997; LaMarche and Lettenmaier, 1998). The modelling objective of Bowling and Lettenmaier (1997) was to test the ability of DHSVM's road algorithm to simulate logging road effects on runoff response. D H S V M was applied to the adjacent watersheds of Hard and Ware Creeks, located in the Deschutes River basin of Western Washington. Hard Creek is a 2.31 km watershed containing 11.4 km of roads. Streamflow extension caused by the road presence resulted in a 63.5% increase in stream channel density to 5.9 km/km in Hard Creek. Ware Creek is a 2.83 km 2 watershed containing 10.7 km of roads. Streamflow extension caused by the road presence resulted in a 52.3% increase in stream channel density to 5.6 km/km2 in Ware Creek. A 2 ha contributing area was selected as the threshold for channel initiation for a 30 m resolution D E M (Bowling and Lettenmaier, 1997). A 2-hour time step was applied. Data used to test the model during calibration include monitored peak culvert discharge for six culverts in each basin. The model tended to overpredict culvert discharge for the largest stormflows. The occurrence of culvert discharge as simulated by the model tended to agree with field observations. Following calibration, a series of scenarios were run. According to DHSVM, the addition of the road network to the existing vegetation resulted in a 10% and 8% increase in the 10-year return period flow in Hard and Ware Creeks, respectively. Adding the existing road network and vegetation to pristine watersheds resulted in 18% and 22% increases in the 10-year flows in Hard and Ware Creeks, respectively. The modelling objective of LaMarche and Lettenmaier (1998) was to simulate the effect of logging roads and forest harvesting on peak flows. Through modelling, they also examined the effect of vegetation and topography on road-generated runoff. The study site is the Upper Deschutes River basin that encompasses Hard and Ware Creeks of the Bowling and Lettenmaier (1997) study. Fourteen named and eight unnamed tributaries (totaling 15 km2) are included in the 149 km drainage area of the Upper Deschutes. The road densities for the 14 named tributaries range from 3.2 km/km2 to 5.0 km/km2. The drainage density due to roads in the tributary watersheds increased from a range of 3.3 km/km2 to 4.1 km/km2 to a range of 3.5 km/km2 to 4.4 km/km2. As with the Bowling and Lettenmaier (1997) study, a 2 ha contributing area was used as the channel initiation threshold for a 30 m resolution D E M (LaMarche and Lettenmaier, 1998). 8 However, a 3-hour step was used. Initial model results over-predicted culvert response. After calibration, percent changes in the 10-year return flow were examined by comparing simulations of the current vegetation state with and without roads. Changes ranged from 2.5% to 10.1% for the 14 named tributaries. Percent changes in the 10-year return flow were then examined through simulations of a pristine vegetation state with and without roads. Changes ranged from 2.6% to 12.2% for the 14 named tributaries. In this case, the modelled road effects appeared to be relatively independent of the vegetation state. Modelled comparisons of similar sub-basins with differences in the percent of connected culverts showed a difference in the percentage increase in the 2.3-year return flow. 1.4 Speci f ic Object ives a n d A p p r o a c h of this S t u d y This study presents the first application of DHSVM, including its road algorithms, in a snowmelt-dominated watershed. The specific objectives were: 1. To examine the sensitivity of streamflow to the addition of roads at the scale of the entire Redfish Creek basin, as well as sub-catchments, including changes to peak flows and annual yield. 2. To explore the sensitivity of the model to variations in road cut depth. 3. To examine the effect of surface flow connections from culverts to the steam network. 4. To examine hydrologic influence of skid trails and abandoned roads. 5. To contrast the effect of logging roads on tributary flow against observed flow 9 C H A P T E R 2: M E T H O D S 2.1 Redfish Creek Site Description Redfish Creek watershed was previously glaciated. The soils overlie Jurassic, granitic bedrock of the Nelson Batholith. The upper elevations fall within the Engleman spruce -subalpine fir biogeoclimatic zone. The lower elevations are within the interior cedar - hemlock biogeoclimatic zone. The Burn climate station, located at an elevation of 1290 m, recorded an average annual precipitation of 1040 mm based on data for water years 1993-1997. Approximately 15% of the precipitation fell as snow assuming a rain-snow threshold of 0 C. The Cabin climate station, located at an elevation of 1735 m, recorded an average annual precipitation of 1580 mm of which approximately 45% fell as snow. In addition to four total climate stations within the watershed (Figure 3), an additional climate station, Seed Tree, is located approximately 1 km east of the basin outlet. Runoff peaks are generated by snowmelt and rain on snow events. Forest harvesting occurred during the years 1966, 1969-1974, 1987-1988, and 1990, with road building taking place approximately a year prior to harvesting each area. By 1993, harvested areas occupied 10% of the watershed area. Of the roads in the watershed, 16 km are active and 3 km are unused and overgrown. 2.2 Distributed Hydrology-Soil-Vegetation Model (DHSVM) Developed at the University of Washington, D H S V M (Wigmosta et al, 1994) explicitly solves the water and energy balance for each grid cell specified by a digital elevation model (DEM). Meteorological parameters from up to five base stations are adjusted by the model according to lapse rates for application to different elevations in the basin. Precipitation interception is simulated through use of a two layer canopy model. Snow accumulation and ablation calculations for intercepted snow are controlled by a single layer energy balance model. A two layer energy balance model is applied to ground snow. Shortwave and longwave radiation budgets are utilized for each energy balance. The Penman-Monteith formula is used to calculate evapotranspiration. Model conditions are updated according to user defined time steps. 10 2 0 2 Ki lometers F i g u r e 3. Features o f R e d f i s h C r e e k watershed 11 All throughfall and snowmelt enter the soil. Vertical movement of soil water in the unsaturated zone is calculated using Darcy's law. Grid cell water table is recharged by this flux. Lateral subsurface flow (slope parallel) relies on hydraulic gradients applied to cells based on the local ground surface slope derived from the D E M . The model assumes lateral saturated hydraulic conductivities decrease exponentially with depth as suggested by Beven (1982). Subsurface flow is intercepted by the stream channel for routing to the basin outlet. The stream channel is generated by the D E M based on the number of contributing upslope cells. Subsurface flow contribution to streamflow is determined by the water table location relative to the channel bottom. The Muskingum-Cunge routing scheme is used to route water downstream. The road algorithm operates similarly to the channel routing algorithm (Wigmosta and Perkins, 1997). Roads are essentially treated as an extension of the stream network. Interception of subsurface flow at road cuts is determined by the water table location relative to the cut depth. Ditch flow is routed according to the Muskingum-Cunge method. Once flow enters the ditch, it will not be allowed to reinfiltrate through ditch bottoms. Flow in ditches will either be added to streamflow at stream crossings or reinfiltrated at cross drain culvert locations. Extending the stream network to a cross drain location simulates flow released by culverts that directly connects to streams via overland channels. The road surface is modelled as crowned and impermeable. The model allows for various output options. Streamflow can be assessed at user specified tributaries, culverts and ditch segments. Basin maps of calculated variables can be generated. An example of map outputs include precipitation, snow water equivalent, water table depth, and saturated subsurface flow. The model can also be run for a chosen cell in the basin to examine internal physical processes without investing time to run the model using the entire basin. 2.3 Hydrologic Model of Redfish Creek 2.3.1 Data Requirements A large amount of data is required to operate D H S V M due to its high degree of physical representation. A range of spatial and temporal data constitutes the input files. Data relating to the distribution of elevation and parameters relating to soils, vegetation, roads and streams are 12 needed. A range of meteorological data is also required to drive the model. The data required to develop a hydrologic model of Redfish Creek was acquired through the Ministry of Forests, Nelson District. Meteorological input establishes quantity of inputs and the partitioning within the water balance of the basin. Data required by the model include air temperature, wind speed, humidity and incoming longwave and shortwave radiation. Hourly climate data for October 1992 to December 1997 from the Burn and Cabin climate stations were utilized for the initial Redfish Creek modelling applications (Whitaker et al., 2002, 2003). The climate data were quality checked only for 1997. Gaps in the climate data were resolved by Whitaker and Alila (2000) through estimations based on climate data from Nelson and Castlegar, located approximately 30 km southeast and 65 km southeast, respectively, of Redfish Creek watershed. Additional temporal data are available for comparison against model outputs. Snow course and hourly streamflow data were used to calibrate and test the model. Streamflow data for the Redfish Upper Tributary are also available for the comparison of model performance on a smaller scale. Hourly data for the 1 km 2 tributary spans late April through October for 1995-97. Of the spatial data required by the model, the D E M is one of the most important components. Information derived from the D E M is applied to a range of model routines. For example, topography is a controlling factor in the distribution of precipitation. Lateral water movement also depends on topography. The 25 m resolution D E M used for Redfish Creek was derived from a 1:20,000 TRIM (Terrain Resource Information Mapping). Ninety percent of the elevations fall within 10 m of actual elevations. The rectangular D E M consists of 100,800 pixels. The Redfish Creek drainage comprises 41,331 of these pixels for a drainage area of 25.83 km 2 (Whitaker and Alila, 2000). A map containing the distribution of vegetation classes along with the attributes associated with each class are used by the model to determine vegetative effects on meteorological input and water movement. Modelled effects of vegetation include precipitation interception, evapotranspiration, and the attenuation of solar radiation and wind speed. The vegetation attributes include trunk space, constants for aerodynamic and radiation attenuation, height of vegetation layers, summer and winter leaf area index (LAI), stornatal resistance ranges, albedo, the percentage of roots within soil layers and the depth of soil layers, or root zones. The forest cover map used in the original Redfish Creek modelling project was derived from a Ministry of Forests operational forest cover map consisting of 53 vegetation classes based on 13 species, presence or absence of overstory and understory, vegetation height, stand age and crown closure. Whitaker and Alila (2000) derived additional data required by the model, such as LAI, from available literature. The 53 vegetation classes were merged into 14 classes to allow easier calibration and shorter model run time. Table 1 and accompanying Figure 4 show the distribution of the applied vegetation classes. Table 1. Key characteristics of the fourteen vegetation types (courtesy of Whitaker and Alila, 2000) Veg. Description / Species Stand Age Overstory Canopy Leaf Area Type (years) Height (m) Closure (0-1.0) Index (LAI) 1 Vegetation absent (tundra-rock) N/A N/A N/A N/A 2 SubAlpine (subalpine-BS) 200 12 0.2 2.5 3 Balsam-Spruce (BS-20-3) 180 21 0.3 3.0 4 Balsam-Spruce XBS-20-5) 140 22 0.5 3.5 5 Balsam-Pine (BPL-20-8) 130 20 0.8 5.0 6 Balsam-Spruce (BS-30-6) 170 30 0.6 4.5 7 Balsam-Spruce (BS-36-4) 260 36 0.4 4.0 8 Cedar-Hemlock (CWH-30-7) 160 27 0.65 7.0 9 Cedar-Hemlock (CWH-40-7) 170 39 0.7 7.0 10 Douglas-fir-Larch-Pine 115 29 0.7 6.0 (FDLPL-30-7) 11 Clearcut-no overstory 8 0 0 0 12 Clearcut-early regeneration 10 2 0.1 2.5 13 Clearcut-shelterwood 10 15 0.1 2.5 14 Regenerated clearcut 38 15 0.2 3.0 Soil attributes are obtained through soil depth and soil type maps. Information provided by these maps and their associated attributes contribute to calculations of soil moisture content, percolation and lateral subsurface flow. Attributes required for each soil type include lateral saturated hydraulic conductivity and its exponential decrease with depth, maximum infiltration rate and soil surface albedo. Information pertaining to each soil layer specified by the root zones include porosity, pore size distribution, bubbling pressure, field capacity, wilting point, bulk density, vertical conductivity, thermal conductivity and thermal capacity. Soils and terrain information originating from Kutenai Nature Investigation Ltd. were developed to create the \ 14 Vegetation Type | 1 Tundra-Rock P H SubAlpine-BS j j BS-20-3 [~1 BS-20-5 r j T J BPL-20-8 « BS-30-6 BS-36-4 CWH-30-7 CWH-40-7 FDLPL-30-7 Clearcut Clearcut regen Shelterwood Re-growth / \ / Sub-basin Boundaries 2 Kilometers Figure 4. Vegetation types for Redfish Creek watershed 15 input files. Derivation of soil parameters were based on percent sand and clay content. As with the vegetation data, the soils data were merged into seven soil types to allow easier calibration and shorter model run time. Table 2 with accompanying Figures 5 and 6 show the distribution of soil classes. Table 2. Key characteristics of the seven soil types (courtesy of Whitaker and Alila, 2000) Soil Type Coarse Fragment Vert. Sat. / Genetic Soil Texture Content >2mm Porosity Field Capacity Hydraulic Material Layer (%byvol.) Conductivity (m/s) Colluvium 1 fSL 54 0.45 0.24 l.le-5 (C) 2 cSL 66 0.4 0.17 1.6e-5 3 LS 76 0.4 0.17 3.3e-5 Colluvial- 1 SiL 40 0.49 0.33 3.8e-6 Fluvial 2 SiL 55 0.45 0.33 3.8e-6 (CF) 3 SL 65 0.4 0.21 1.2e-5 Fluvial 1 fSL 23 0.45 0.24 l.le-5 (F) 2 SL 29 0.4 0.21 1.2e-5 3 LS 36 0.4 0.17 3.3e-5 Morainal 1 L 32 0.46 0.27 7.4e-6 (M) 2 SL 48 0.4 0.21 1.2e-5 3 SL 54 0.4 0.21 1.2e-5 Weathered 1 L/SL 24 0.46 0.24 1.0e-5 Bedrock 2 cSL 29 0.4 0.17 1.6e-5 (D) 3 cSL 39 0.4 0.17 1.6e-5 Colluvium- 1 SL 61 0.45 0.21 1.2e-5 Bedrock 2 SL/LS 82 0.4 0.17 2.3e-5 (CR) 3 LS 92 0.4 0.17 • 3.3e-5 Organic 1 L 25 0.46 0.27 7.4e-6 (0) 2 SL/LS 38 0.4 0.17 2.3e-5 3 SL/LS 52 0.4 0.17 2.3e-5 Three input files are required to describe the stream network. The stream class file provides channel routing parameters such as stream width, depth and friction coefficient. Each stream segment is assigned a stream class in the stream network file. Channel routing parameters for individual segments and stream segment orders for flow routing calculations are also contained in this file. The stream map file locates stream segments in grid cells. Segment location is expressed in terms of grid cell column and row number. The straight line stream channel length in cells, stream bank height and stream channel width are also specified by the stream map file. 16 N Genetic Material Bedrock | | Colluvial | ] Colluvial-Fluvial | | Fluvial g Morainal p~^ ~] Organic l^sp Weathered Bedrock /\J Sub-basin Boundaries 2 Kilometers Figure 5. Genetic material classification used to determine soil types for Redfish Creek 17 /\J Sub-basin Boundaries 1 0 1 2 Kilometers Figure 6. Soil depth for Redfish Creek watershed 18 As stated previously, the road network is treated as an extension of the stream network. The input needed to operate the road algorithm is similar to that of the streams. The road class file provides channel routing parameters such as ditch width, depth and friction coefficient. The crown type may also be specified but has not been incorporated into model calculations. Each road segment is assigned a road class in the road network file. Also included in this file are channel routing parameters for individual segments and road segment orders for ditch flow routing calculations. The road map file correlates road segments with cells for the quantification of intercepted flow. Segment location is expressed in terms of grid cell column and row number. 2.3.2 Redfish Creek Model Calibration The calibrated model used in this study is described in Whitaker et al. (2001). Calibration of the Redfish Creek model is summarized in the following paragraph and detailed by Whitaker et al. (2003), who made further applications of the calibrated model, including a slight enhancement to the calibration, using an updated version of DHSVM. Although neither the final calibration nor updated D H S V M version was used in this study, the calibration information of Whitaker et al. (2003) is relevant to the calibrated model used for this study. The calibration process addressed annual streamflow yield through adjustment of precipitation input to the system; occurrence of seasonal peak flows through separation of rain and snow using a temperature threshold, and by adjustment of albedo curves; effect of vegetation on snow cover through analysis of snow water equivalence distribution; and hydrograph fit through designating values for parameters controlling lateral hydraulic conductivity. A 4-year period from October 1992 to September 1996 was used for the initialization of basin conditions and calibration. As the first six months are very sensitive to the initial conditions applied to the start of the model, they were not used for calibration. The last year of available data (October 1996 to December 1997) was used to test model output response to parameter changes during calibration. Included among the variables controlling snow accumulation are those involving the partitioning of precipitation into snow and rain. The separation was set such that precipitation below -0.5 °C falls as snow. Above 0.5 °C precipitation occurs as rain. A combination of snow and rain occurs between the thresholds. Due to the high elevation of the Redfish Creek, snow 19 water equivalent is not very sensitive to these thresholds since snow typically falls at temperatures much lower than freezing (Whitaker et al., 2003). Hourly temperature arid precipitation lapse rates were calculated to model approximate precipitation distribution and, in turn, snow pack distribution (Whitaker and Alila, 2000). Temperature lapse rates were determined based on the available records for a pair of four climate stations located in the Redfish Creek watershed. In addition to the Burn and Cabin climate stations, partial records from the Alpine station, located in the northern part of the basin at an elevation of 2050 m, and Seed Tree station, located just south of the basin at an elevation of 850 m, were used. For those time periods in which multiple pairs of overlapping information exist, the pair with the greatest elevation difference was used to determine the temperature lapse rate. Precipitation lapse rates were first defined according to a monthly time scale. The monthly rates were then applied to hourly precipitation data to calculate hourly precipitation lapse rates. An upward adjustment of the precipitation lapse rates corrected deficiencies in snow accumulation at the Cabin and Burn sites based on comparisons of pixel scale simulated snow-water equivalence with snow course data at those locations (Whitaker et al, 2001). To adjust timing of seasonal melt, albedo curves were redefined within the D H S V M code. Snow albedo determines the absorption of solar radiation by the snow surface. Albedo values applied by D H S V M are based on number of days from the last snow event (Wigmosta et al, 1994). Initially, early season melt was under predicted and late season melt was over predicted. The albedo curves were lowered to better simulate timing of melt. Melt calibration based on albedo curve changes were performed by contrasting modelled SWE for pixels at the Burn and Cabin locations to available snow surveys. Maps of snow cover extent, developed from air photos, were also compared to model generated maps of snow water equivalent (Whitaker and Alila, 2000). Vegetation related variables affecting snow accumulation and melt in forested areas were determined based on available literature as referenced by Whitaker et al. (2003). Comparisons of simulated and measured SWE for the clear-cut and forested areas of Upper Burn, located north of the Burn climate station at an elevation of 1435 m. The variables controlling saturated lateral hydraulic conductivity were the most sensitive variables controlling subsurface water movement (Whitaker and Alila, 2000). Two components involved in modelling this parameter are maximum saturated hydraulic conductivity in the top of the soil profile and its exponential decay with depth (Beven, 1982). These components were 20 assumed constant throughout the basin. The calibration process involved maximum saturated hydraulic conductivities ranging from 2 x 10"4 to 2 x 10"3 m/s and exponential decay coefficients from 1.0 to 2.0 (Whitaker et al, 2001). Calibration resulted in a maximum saturated hydraulic conductivity of 5 x 10"4 and an exponential decay coefficient of 2.0. Generated maps of water table depth were examined in addition to comparisons between simulated and observed streamflows to determine that a realistic distribution of saturated depth was attained. A fairly good fit exists between simulated and observed streamflows at the basin mouth. However, there are discrepancies between simulated and observed flows for the Upper Redfish Tributary. Modelled tributary flows have a much flashier response to snowmelt and rainfall. The flows also tend to overestimate observed flows. These differences may be attributed to less precise delineation involved with smaller drainage areas, the sensitivity of small basins to the spatial variability of soil properties, spatial variability of snowmelt runoff, or the presence of roads in the tributary that where not included during calibration (Whitaker and Alila, 2000; Whitaker et al, 2001; Whitaker et al., 2003). An additional factor could be the inability of D H S V M to capture flow generation processes at smaller scales. 2.4 Incorporation of Roads in the Redfish Creek Model 2.4.1 Methods Overview The research approach largely involved computer simulations of hydrological processes and response at the basin and sub-basin scales. Model simulations of the hydrology of Redfish Creek incorporating the road network were compared to a base simulation that did not include the road network. Field data were collected to derive input parameters required by the road algorithm. After collection, the data were used to create a database for application to a geographic information system (GIS). Three input file sets were produced which relate to the impact of logging roads and skid trails. The first set of files represent the operational logging roads alone. The second set builds on the first set by incorporating stream extensions to those culverts that are connected to the stream via surface flow. The remaining set incorporates skid roads and abandoned roads with the logging roads and stream extensions of the second set. The skid roads and the abandoned roads were represented as logging roads. These three file sets were incorporated with the input files 21 used for the original model calibration that preceded this study. Streamflow output from the model runs were contrasted against each other and the streamflow generated from the original pre-road calibration to measure the sensitivity of the model to its road algorithm. Streamflow from two sub-catchments were examined in addition to the flow at the basin mouth. Observed tributary flow was contrasted against tributary flow generated from the third input set. 2.4.2 Field Data Collection Field data were collected during a 2-week period in July and a 1-week period in October 1999. Roads were traversed twice during the initial fieldwork. Points of interest were flagged and labeled during the initial traverse of road branches. Flagged features, which were mapped using a global positioning system (GPS), include points of ditch flow diversion and conversion, culverts, cross ditches, areas where no ditch existed, seeps visible at the cut slope, and ephemeral streams. Road surveys were performed during the second traverse. The parameters measured include road width, ditch width, ditch depth, road cut length, vertical road cut height, road cut slope, and vertical depth to a confining layer. Ditch roughness was qualitatively described using a rating system of 1 (smoothest) to 4 (roughest). Values of ditch roughness ranged from 0.030 to 0.045 and were assigned based on ditch descriptions in the field notes. This range of values is typically applied for moderately rough sandy bed characteristics to rougher cobbly channels with boulders. These characteristics were observed at a smaller scale for the ditchlines. The crown types noted for the road include insloped, outsloped and crowned. Survey points were established approximately every 200 m or after a significant change appeared in one of the parameters. Areas below culverts, cross ditches, and switch back turns were examined for surface flow connectivity to the stream. Most drainage relief culverts carried little if any water at the time of field data collection due to the lack of rainfall. In addition, it was often difficult to tell whether connectivity existed due to dense vegetation that obscured views downslope. Downslope topography in combination with erosion directly below the culvert and distance to the stream were used to assess likelihood of connectivity noted as none, low, moderate or high. Culverts were considered to be connected where there was substantial erosion directly below the culvert on a continuously steep hillslope with a relatively short distance to the stream. Omitted 22 from consideration were those culverts for which, in absence of detailed field descriptions, there is a flow length of 150 m or greater to the stream. Following the second traverse, a GPS was used to map the road and feature locations that were flagged earlier. The GPS utilized was a Trimble GeoExplorer II set at a position dilution of position (PDOP) of 6.0. The PDOP is an error assessment of position reading based on of the triangulation between the GPS and satellites. If the PDOP exceeded 6.0, measurements cannot be read. The resulting map of road and feature locations is the foundation of the electronic database that is processed via GIS to generate two of the three required road input files. Additional road surveys were performed on skid trails and an abandoned road during the final field visit. Measurements were taken upon crossing switchbacks of the abandoned road segments as the hillslope was traversed from the operational logging road down to the main stem of the stream. Locations of road survey points on the abandoned road were not mapped because vegetation obstructed GPS reception of satellite signals. Skid trails located in a clearcut on the east side of the catchment behave hydrologically as roads without a ditch. Small road cuts present along the trails justify their representation as roads in the model. Road surveys were also conducted to characterize two general skid trail types: one that is wider and was likely more heavily used and one that is narrower and was less used. 2.4.3 Application of Field Data A series of GIS scripts developed at Battelle Memorial Institute facilitated the development of D H S V M road input files based only on road placement relative to a D E M along with assumed road and ditch parameters (Wigmosta and Perkins, 1997). Although field data are not required to create road input files for DHSVM, they provide more accurate input data against which outputs can be compared for the purposes of assessing general modelling accuracy. Measurements of ditch width and depth, estimates of ditch friction coefficient (Manning's n) and crown type were direct inputs to the model. Road width and cut slope information were used in conjunction with road locations and the D E M to calculate road cut depth and width within the GIS. Mapped road positions were used to overlay model cells to determine which cells are to be accessed by the road algorithm. Ditch flow directions were noted in the field to correct computer generated discrepancies of the flow directions determined through GIS assisted calculations involving road locations relative to the D E M . Inputs of accurate flow directions are 23 of greater importance than detailed road parameter measurements due to model sensitivity (James VanShaar, personal communication 1999). Culverts and cross ditches were mapped as hydraulic sink locations for model input. Evidence of connectivity based on field observations was represented through the inclusion of stream extensions to the original stream network input files. Field observations were made to assess the validity of model-generated flow paths. The model may not account for true water movement or pathways due to scale effects or model assumptions. For example, accuracy of ridge and valley locations depends on the accuracy of the D E M , and subtle topographic features in the field may not be well represented by a D E M due to the pixel size. D H S V M assumes that no infiltration occurs in ditches; however, significant loss of flow resulting from ditch infiltration was noted in the field. Additional field data collected includes road cut depth, the depth of confining layers within the exposed road cut, location of ephemeral streams, the presence of subsurface flow at road cuts, and presence or absence of a ditch. The depth of a confining layer within the road cut was compared against the soil information used in the initial model calibration. Ephemeral streams and seeps were compared against the location of DEM-generated streams that do not appear as perennial streams in the field. Seep locations were also compared to model-generated maps of water table depth. The locations of ditch absence were noted since D H S V M does not model roads without at least a minimal ditch presence. 2.4.4. Preparation of Input File Sets for the Existing Road Layout Three input file sets, developed from the field data, modelled the current roads and their stream channel extensions. With each model run, varying detail was added to modelling the logging road affect on streamflows (Table 3). Input file sets include files representing the existing logging roads, stream networks including stream channel extensions to connected culverts, abandoned roads, and skid trails. The road files were generated from field data through the use of ArcView and ArcINFO GIS. These applications also enabled the alteration of the existing DEM-based stream network to account for surface flow from selected culverts to the stream. Modified stream input files were created as a result. ArcView GIS was applied to alter input soil maps to represent areas of lowered infiltration rate at skid trail locations. 24 Table 3. Series of Model Runs Performed Model Runs A Original calibrated model B Original calibrated model + existing loging roads Ba: shallow road cuts Bb: observed road cuts Be: road cut depth = soil depth C Original calibrated model + existing loging roads (Bb) + stream extension D Original calibrated model + existing loging roads (Bb) + stream extension + abandoned roads ans skid trails (modelled as logging roads) Initial editing of raw data and creation of a road database were accomplished through Arc View GIS. GPS data relative to the existing logging roads were imported to ArcView. Using this application, roads located on the east side of the basin were shifted approximately 25 m northward to improve their fit against contours of the basin and stream crossing locations. To guide road placement, elevations of road survey points, obtained through GPS, were contrasted against the elevation contours derived from the D E M . Roads were then divided into segments for which the ArcView database associates road segments with a road class. Each road class is an integer with the associated road parameters of ditch width, ditch depth, ditch roughness (Manning's roughness coefficient), crown type road width and cut slope. Exporting pertinent data from ArcView to a text file generated the road class input file. Although road width and cut slope are not explicitly included in the road class input file, they are required to calculate parameters that must correspond to the D E M . The database was then applied within Arclnfo GIS to generate the remaining road input files. The GIS scripts used to generate the road network file utilized the D E M and the digital map of road segments to create an order for flow routing computations. Segment length was also calculated as an additional parameter to the Muskingum-Cunge routing method. The length of the operational logging roads added to the model is 15.9 km. The remaining road input file was generated though scripts that accessed information from the D E M , the digital map of road segments, and their associated parameters. The resulting road map file associates road segments with their respective grid cells. The road cut depth and 25 width, segment length within a cell, and its aspect within the cell were calculated using the D E M , road width, cut slope, ditch width and depth. The parameters contained within this file help to determine the quantity of intercepted flow for each road segment. A total of 119 sink locations, representing culverts and other features that route flow from ditches, were specified by the map file, 20 of which were stream crossing culverts. Road input files were edited prior to conducting model runs. A previous application of the road routines resulted in overestimation of intercepted subsurface flow as a result of cut depths specified as deeper than the soil depth (La Marche and Lettenmaier, 1998). Consequently, the road map input file was first edited such that any road cuts deeper than the soil were changed to equal the soil depth. The road map file was further edited to create three versions differing in general road cut depths. The first version contained original ArcINFO generated cut depths that were often equal to the ditch depth, and were therefore much shallower than the road cuts observed. The cut depths of the second version were based on the field notes. In the third version, cut depths were defined equal to soil depths to simulate a worst-case scenario of maximum subsurface flow interception. The stream network was extended to culvert locations to represent surface flow extending below culverts to the stream. Cross ditches and switch back bends were also examined for connectivity. The culverts were identified as connected based primarily on field note descriptions. For example, if a culvert appeared to drain into a perennial or ephemeral stream in the field that was not represented as a stream in the stream input files, the culvert was selected for incorporation to the stream network via surface flow. Additional culverts were selected based on interpretation of field notes in conjunction with stream network proximity and hillslope gradient. Further examination of field notes led to deletion of many culverts, that were initially designated as connected, from the stream extension process. A conservative estimate of 19 out of 99 non-stream crossing culverts, cross ditches and switch back bends were modelled as connected to the stream network via surface flow. An existing set of stream input files were edited to include stream channel extensions. The total stream length increased by approximately 3% from, 64.3 to 66.4 km, resulting in a slight increase in drainage density by 3%, from 2.49 to 2.57 km/km2. Adding further detail to modelling the existing condition involved the inclusion of skid trails and abandoned logging roads. The surveyed, abandoned road, eastern skid trails, and additional abandoned, and unsurveyed roads and skid trails were modelled as logging roads. The additional roads and skid trails were selected based on their high level of visibility on aerial 26 photos. The road database developed for the operational logging roads was expanded to include these additional roads and skid trails. The total road length increased from 15.7 to 25.7 km. This database was run through GIS to create an extended set of road input files. The stream network that includes stream extensions to connected culverts was also utilized in this model run. 2.4.5 Modelling Logging Road Impact Road effects were tested through application of the three data sets corresponding to runs B through D. Sensitivity to various modelling details was examined via the three model runs that incorporated files associated with road and skid trail impact with the original input files for the calibrated model developed by Whitaker et al (2000) that did not include roads. The calibrated data set includes the stream network, vegetation parameters, soil parameter, and climate parameter. Conclusions on road effects were based on changes in streamflows for the basin and sub-basin scales. Streamflows resulting from the varying road cut depths of runs Ba through Bc were compared to the stream flows of run A, the original calibrated model without roads. Streamflow differences between the B runs and run A were plotted for the basin mouth and two tributaries (Figure 3,). The road input corresponding to run Bb, in which road cut depths represent field conditions, was incorporated into the remaining simulations involving the existing logging roads. Output and analysis for runs C and D were conducted in a similar manner to those of runs A andB. Results of model runs Bb and C assessed the modelled impact of existing operational logging roads. Run C differed from run B by the inclusion of stream channel extensions to connected culverts. A comparison of flow simulations (Bb and C) illustrated the modelled result of incorporating stream channel extensions in the examination of road effects. Model run D was comprised of the input to run C with skid trails, abandoned roads, and unsurveyed roads modelled as logging roads. Simulated flows at the basin mouth in run D were compared to those for run C in order to examine the collective effect of all roads and trails on streamflows versus the effect of active access roads and spurs. 27 2.4.6 Comparison With Redfish Upper Observed Flows Whitaker et al. (2001) sited the omission of logging roads to be a possible reason for the modelled overestimate of Redfish Upper tributary flows. The remainder of this study contrasts observed tributary flow against tributary flow generated from the third input set, which incorporates the primary logging road, stream extensions, skid trails and abandoned roads. This was done to determine whether the addition of roads decreased the overestimated flow of the roadless scenario to provide a better approximation of the observed flows. 28 C H A P T E R 3: R E S U L T S 3.1 Overview of Progressive Modelling Refinements The following results are described in order of the model runs with a focus on annual and seasonal changes to streamflow characteristics such as annual yields snowmelt, low flows, and peak flow. Subsequent topics include examination of specific runoff events such as snowmelt and rainfall induced peaks. Induced streamflow changes can be related to the differences in modelled feature extent displayed in Table 4. The modelled road length for Run D includes skid trails, abandoned roads, and unsurveyed roads in addition to the primary logging roads. The sinks include cross-draining and stream-crossing culverts, switchback bends, and cross ditches. Stream network length does not include the length of ditch contributing intercepted flow to stream crossing culverts or culverts connected to the stream network by way of surface flow. Table 4. Extent of modelled features Redfish Creek Basin (25.8 km2) Run A Run B RunC RunD Modeled road length (km) 0 15.7 15.7 25.7 2 Road density (krn/km) 0 0.61 0.61 1.00 Number of sinks _q_ 150 150 187 Stream network length (km) 64.3 64.3 66.4 66.4 2 Stream drainage density (km/km ) 2.49 2.49 2.57 2.57 Redfish Upper Sub-Basin (1.15 km2) Modeled road length (km) 0 1.73 1.73 1.92 2 Road density (knVkm) 0 1.50 1.50 1.67 Number of sinks 0 17 I 17 17 Stream network length (km) 3.51 3.51 3.69 3.69 Stream drainage density (krn/km ) 3.05 3.05 3.21 3.21 South Tributary Sub-Basin (0.75 km2) Modeled road length (km) 0 _._018_ 0.18 0.20 2 Road density (krn/km) 0 0.24 0.24 0.27 Number of sinks 0 4 4 5 Stream network length (km) 3.34 3.34 3.37 3.37 Stream drainage density (krn/km ) 4.45 4.45 4.49 4.49 29 Results focus on water year 1995 so as not to overwhelm the results content. This year was chosen based on the simple rise and fall pattern of streamflow associated with the seasonal snow pack melt. Changes to annual yield and peak flows are respectively shown in Table 5 and Table 6, for Redfish Creek, Redfish Upper Tributary, and South Tributary. The figures referenced in this chapter, Figures 7 to 27, are located in Appendix A. Graphical and tabulated results for the remaining years are located in Appendix B, shown in Figures 28 to 51 and Tables 7 to 18. As modelled Redfish Upper tributary flows had shown discrepancies when compared to observed flows, this chapter concludes with the presentation of modelled versus observed Redfish Upper tributary flows. 3.2 Run B - Addition of primary logging roads and model sensitivity to cut depth As the alteration of streamflow characteristics resulting from the addition of primary logging roads is difficult to detect based on graphing the Run B set of streamflows against that of Run A (the base simulation without roads), the following figures display the effect of roads as differences between the simulated streamflows relative to the streamflow of Run A, which is shown at the bottom of the figures. Discharge differences of the three cut depth types are displayed in an overlain fashion in order to contrast their differences relative to Run A. 3.2.1 Redfish Creek streamflows As shown in Figure 7, the addition of primary logging roads produces little change in Redfish Creek streamflow. Discharge increases relative to the reference run begin during the early part of the melt season and persist through the peak of the melt season. Flow differences then decrease through the remainder of the melt season through the low flow period, with occasional spikes of increased flow during the period of snow accumulation. Spikes of increased and decreased flows are scattered through the year. The addition of logging roads with road cuts that represent the minimal road cuts, equal to ditch depth (Figure 7.a.) increased the annual Redfish Creek yield for 1995 by less than 0.1%. The average annual yield for 1994 to 1997 also increased by less than 0.1%. Water year 1993 was omitted from consideration, as the first six months of the simulation were sensitive to initial 30 i- -l l s 15 SO SO o oo r- 00 CN CN in 00 r f o r f •n so OS ro r f Os m t~-CN CN rf IT) rf r f t~- r- oo oo oo oo et 111 •a <L> e« 3 to U 2 3 s-~' 2 > so „ a 1 u e 3 PS S O 0 S 0 S - 0 S -so rn O O Os oo t"^  fN »—< S-S 0 S -0 S - 0 S -r-; oo ro oo d t~ so CN *—< ^P N P •>? 0 S 0 S 0 S - 0 S - 0 S -p CN r f >-< ro rf ro 5^ 0 S 0 S 0 S - 0 S 0 S -00 CN s q OO CN © ro CN ro ro SO so 00 m o ro in 1—1 Os os CN m Os oo rf SO CN r f CN o oo CN O so in CN CN i CN ro ro ro ro ro ro '— 1 1—1 r — ' 0 S -o V v • 0 S 0 S - 0 S 0 S -*—< CN »—' d d d d s o s o 6 s 0 S -A A o Os rf OS Os in r- i n CN in O Os ro in ro oo SO O r f SO so r f i n in i n SO SO SO SO SO SO ro ro ro ro ro ro CN CN CN CN CN CN •< pq pg pq U Q 2 5 •S .2 H = o -a a CM -w S H a a & Vi c •a cu rs cu cu VI C T3 cu O C a <u a. Os C3N cu DD a A U ve cu 3 H s 3 2 sP so sP 0 S o^ 0 S ro oo CN d OS 0 S - 0 S - 0 S - 0 S - 6s-CN rt d SO rf O r f r f SO O r- Is- r f so r f >—i • — i ro ro ro ro ro ro ro ro ro ro d d <£ d <=>' d> s O s O 0 S - 0 S - 0 S 0 S - 0 S Os ro CN i-i ro CN d 0 S - 0 S N O 0 S -~<= 0 S 0 S -p r f ro ro rf r f CN CN SO ^ H r f so r— r f CN oo CN ro CN CN ro CN - H CN CN r f r f r f r f r f r f d d d d d d s ° s o 2 s ? \ » 0 S • X\ 0 S 0 S o 1^ so d ' o o d A s° 0 S 0 S 0 S - 0 S 0 S -ro r f d d d d fN r f SO in ro Os os SO O CN •n oo CN CO in m Os ro ro ro ro ro ro so SO SO so so so < pa pa PQ u Q 31 0.04 7.a. Run Ba - Run A, effect of minimal road cut depths approximately equal to ditch depths -0.04 -0.08 0.04 - — i~ -• l * rv—• * 0.04 S 0 la § 8» -0.04 -0.08 7.b. Overlay of Run Bb - Run A, effect of field surveyed road cut depths . . . . . . — , . I R ~ C \ * ilwwiBi 7.c. Overlay of Run Be - Run A, effect of maximum road cut depths equal to soil depth -0.04 -0.08 -0.12 H -0.16 -0.2 ] -0.24 - T 9 o o o o o Run Ba - Run A • Run Bb -Run A • Run Be - Run A 9 o O Run A Figure 7. Discharge differences for Redfish Creek due to the varying road cut depths of Runs Ba, Bb, and Be during water year 1995 32 V conditions (Whitaker et al., 2003). Annual peak flow for 1995, and similarly the average annual peak flow for 1993 through 1997, increased by 0.1%. The addition of logging roads with field-defined road cuts resulted in a slightly greater impact on Redfish Creek streamflow compared to that of the Run Ba (Figure 7.b). The increased flows during the beginning of the melt season are higher and persist further into the melt season. The decrease in streamflow, beginning at end of the melt season, lasted approximately a month longer than the impact resulting from Run Ba. For the scenario of Run Bb, annual yield for 1995 and the average annual yield for 1994 to 1997 decreased by 0.1 % compared to that of Run A. The 1995 peak flow increased by 0.3%. However the average annual peak flow, for 1993 through 1997, decreased by less than 0.1 %. The effects of setting cut depths equal to the soil depth are similar to those modelled in Run Bb (Figure 7.c). The increased flows during the beginning of the melt season are higher than that of Run Ba and lower that that of Run Bb. These increased flows continue later into the melt season than for Runs Ba and Bb. The Run Be annual yield for 1995 decreased by 0.1 % and the average annual yield for 1994 to 1997 decreased by 0.2% compared to that of Run A. The 1995 peak flow increased by 0.4%. However the average annual peak flow, for 1993 through 1997, decreased by 0.2%. A closer look at the effect of the primary logging roads on the snowmelt generated annual peak flow for 1995 shows a sinusoidal pattern of discharge difference between the road imposed simulations and the roadless scenario (Figure 8). Run Ba yielded dominantly increased discharges over the 24-hour period of diurnal streamflow response to snowmelt. Decreased discharge occurred at the midpoint in streamflow recession. Runs Bb and Be generally increased flows compared to Run A, with Run Be having a slightly greater impact. The timing of the annual peak occurrence was not affected by the presence of primary logging roads for any year of Redfish Creek streamflows for Runs Ba, Bb, or Be. The differences resulting from the addition of primary logging roads during a peak flow resulting from a rain-on-snow event is shown in Figure 9. The decreases in streamflows, represented in Figure 9.a., fluctuate with the rainfall intensity shown in Figure 9.b. The magnitude of peak flow decreases are reduced with decreased cut depth representation. Increasing cut depth, from Run Ba to Be, was associated with a greater reduction in peak flow. 33 o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o Run A Run B a - Run A Run B b - Run A Run Be - Run A Figure 8. Discharge differences during the annual snowmelt peak for Redfish Creek due to the varying road cut depths of Runs Ba, Bb, and Be in 1995 Figure 9. Discharge differences during a rain on snow event for Redfish Creek due to the varying road cut depths of Runs Ba, Bb, and Be in 1995 3 4 3.2.2 Redfish Upper tributary flows As shown in Figure 10, the addition of primary logging roads results in an overall decrease to Redfish Upper tributary flows. The greatest decreases occur during the melt season. As with the discharge differences for Redfish Creek, there are occasional spikes of increased flow during the period of snow accumulation for Runs Bc and Bb. With the addition of minimal road cut logging roads (Figure 10.a.), the annual Redfish Upper flow yield for 1995 and the average annual yield for 1994 to 1997, decreased by 0.8% and 1.0%, respectively. Annual peak flow for 1995 and the average annual peak flow for 1993 through 1997 decreased by 1.0% and 1.1% respectively. For the case of this tributary flow, the annual peak flow is the rain-on-snow event of late June and not the peak flow resulting from the snowmelt of late May, as it is for the Redfish Creek streamflow. The addition of logging roads with field-defined road cuts produced a slightly greater impact to Redfish Upper tributary flows compared to that of Run Ba (Figure lO.b). The reductions in flow relative to Run Ba are greater through the melt season. The scenario represented by Run Bb decreased annual Redfish Upper tributary yield for 1995 and the average annual yield for 1994 to 1997 by 3.2% and 3.5%, respectively, compared to that of Run A. The 1995 peak flow decreased by 4.7%. Average annual peak flows for 1993 through 1997 decreased by less than 3.1%. The effects of the addition of primary logging roads with cut depths equal to the soil depth are similar to those caused by Run Bb, although there are occasionally larger decreases to tributary peak flows resulting from Run Bb. Run Bc annual yield for 1995 decreased by 2.6% and the average annual yield for 1994 to 1997 decreased by 3.0% compared to that of Run A. The 1995 peak flow decreased by 4.4%. For 1993 through 1997, the average annual peak flow decreased by 2.9%. The change to the snowmelt-induced streamflow peak for the Redfish Upper Tributary is displayed in Figure 11. A dip in decreased streamflows is common among the three road cut scenarios. Unlike the Redfish Creek streamflows, in which the greater changes were induced by deeper road cuts, the greatest decrease for the Redfish Upper Tributary is shown for the field-defined road cut scenario and not the scenario with the deepest road cuts. The largest decrease in discharge occurred approximately two hours following the peak flow. 35 0.005 0.000 -0.005 -0.010 -0.015 -0.020 -0.025 0.005 10.a. Run Ba - Run A, effect of minimal road cut depths approximately equal to ditch depths lO.b. Overlay of Run Bb - Run A, effect of field surveyed road cut depths £ 0.000 -H o c -0.005 --0.010 -a irge -0.015 -isch: -0.020 -a -0.025 -1 . . 1 1 1 .nm, t - 1 r H p i ^ y ^ 10.c. Overlay of Run Be - Run A, effect of maximum road cut depths equal to soil depth 9 o o o Sr O 0> 00 o Run Ba - Run A Run Bb - Run A • Run Be - Run A Run A o Figure 10. Discharge differences for Redfish Upper Tributary due to the varying road cut depths of Runs Ba, Bb, and Be during water year 1995 36 The Redfish Upper 1995 annual peak flow resulted from a rain-on-snow event. As with the response difference to Redfish Creek streamflows, the differences to Redfish Upper Tributary fluctuate with the fluctuating rainfall rates (Figure 12). Run Bb has a slightly greater impact of decreased peak tributary flow than Run Be. The timing of the annual peak occurrence was not affected by the presence of primary logging roads for any year of Redfish Upper Tributary streamflows for Runs Ba, Bb, or Be. 3.2.3 South tributary flows The addition of primary logging roads results in an overall increase to the South tributary flows, as displayed in Figure 13. The greatest increases occur during the melt season with the largest differences occurring approximately two weeks prior to the tributary peak of the snowmelt season. For the period of increased flows during the melt season, the minimal road cut scenario yielded a shorter duration than of increased flows than either of deeper road cut scenarios. There are spikes of increased discharge changes for peak flows resulting from rainfall events. The addition of logging roads with minimal road cuts, results in increased flows during the rise of tributary flows at the beginning of the melt season (Figure 13.a). The increase in flows tapers to zero with the conclusion of seasonal melt. The annual South tributary flow yield for 1995 and the average annual yield for 1994 to 1997 increased by 2.6% and 4.6%, respectively. Annual peak flow for 1995 and the average annual peak flow for 1993 through 1997 decreased by 0.1% and 5.4% respectively. As with Redfish Upper Tributary, the annual peak flow is the resulted from the rain-on-snow event of late June. The addition of logging roads with field-defined road cuts is shown in Figure 13.b. as the line showing greater change to tributary flows. Similarly to the impact of the minimal road cuts, the largest increases to flows occur during the rising limb of the seasonal snowmelt hydrograph. The scenario represented by Run Bb increased annual south tributary yield for 1995 and the average annual yield for 1994 to 1997 by 18.7% and 19.6%, respectively, compared to that of Run A. The 1995 peak flow increased by 5.5%. Average annual peak flows for 1993 through 1997 increased by 11.3%. The addition of primary logging roads with cut depths equal to the soil depth produced slightly greater changes to streamflow than those caused by Run Bb (Figure 13.c). Run Be 37 i n in i n i n </-. in in »n >n ^ ^ *n »n v-» i n vn m m m Run A Run Ba - Run A - Run Bb - Run A • Run Bc - Run A Figure 11. Discharge differences during the annual snowmelt peak for Redfish Upper Tributary due to the varying road cut depths of Runs Ba, Bb, and Bc in 1995 12.a. Discharge differences due to varying cut depths Run Ba - Run A — Run Bb - Run A • Run Bc - Run A 12.b. Corresponding rainfall and streamflow for the rain on snow event i„ l_ _^ JZ\ JS zz — : £ - - ZZ zz -o I— O N o m rl — — — — rs rs ri CO- p o in If-. >r, iv-. V N «ri ir* >/-. i/M » a a c ON a a o- Os ON O N ON O N 9 c~ ON ON O N 5> O N 0> o--f T - f -f -f -f - f -i= — T, in CJ r j C! r l r l fs (N fN) cj r 1 - 0 O •o -£ NO NO NO NO NO NO o o o © o o o O p Q o o Figure 12. Discharge differences during a rain on snow event for Redfish Upper Tributary due to the varying road cut depths of Runs Ba, Bb, and Bc in 1995 38 0.025 13.a. Run Ba - Run A, effect of minimal road cut depths approximately equal to ditch depths -0.005 0.025 13.b. Overlay of Run Bb - Run A , effect of field surveyed road cut depths -0.005 0.020 0.010 0.000 -0.010 -0.020 -0.030 -0.040 -0.050 13.c. Overlay of Run Be - Run A , effect of maximum road cut depths equal to soil depth h 0.30 0.50 0.45 0.40 0.35 o ON Q •r. • Run Ba - Run A • Run Bb - Run A • Run Be - Run A Run A Figure 13. Discharge differences for South Tributary due to the varying road cut depths of Runs Ba, Bb, and Be during water year 1995 39 annual yield for 1995 increased by 20.8% and the average annual yield for 1994 to 1997 increased by 21.3% compared to that of Run A. The 1995 peak flow increased by 6.2%. For 1993 through 1997, the average annual peak flow increased by 11.5%. The change to the snowmelt induced streamflow peak for the South Tributary is displayed in Figure 14. For the three runs, there is a fairly consistent change to streamflow over the 24-hour flow period, with a nearly indiscernible dip to the increased flows corresponding to the flow peak of Run A. Run Ba shows the least amount of change to the streamflow. The increased changes due to roads imposed in Runs Bb and Be are nearly identical with Run Be resulting in slightly greater change. The South Tributary annual peak flow resulting from a rain-on-snow event is displayed in Figure 15. Changes resulting from Run Ba show little alteration to tributary flows, unlike the changes resulting from Runs Bb and Be, which exhibit greater sensitivity to rainfall rates. Run Be has a slightly greater impact on increased peak tributary flow than Run Bb. The addition of logging roads for all B runs resulted in a 10-day advance of the South Tributary annual peak during water year 1997. 3.3 Runs Bb, C and D - Comparison of the effect of primary logging roads, stream network expansion, and skid trails and abandoned roads As in the case of the contrasting streamflow changes due to varying road cut depths, the streamflow changes resulting from expanding the stream network and adding skid trails and abandoned roads are difficult to detect based on graphing the resultant streamflows against those of Run A, the base simulation without roads. Therefore, the following figures are laid out similarly to that of the previous figures. The streamflow results from Runs Bb, C, and D are shown as differences between the simulated streamflows relative to the streamflow of Run A that is shown at the bottom of each figure. The results from Run Bb, the scenario with road cuts representing field-defined measurements, represent the effects of the primary logging roads, alone, on streamflow. Discharge differences are displayed in an overlain fashion in order to contrast their differences relative to Run A. 40 0.03 «• 0.02 5 ro ro ro ro ro ro i n i n m i n i n * n i n t n r o r o r o r o r o r o r o r o tn wi in in in in »n ro ro ro ro •Run A • Run Ba - Run A Run Bb - Run A • Run Be - Run A Figure 14. Discharge differences during the annual snowmelt peak for South Tributary due to the varying road cut depths of Runs Ba, Bb, and Be in 1995 Figure 15. Discharge differences during a rain on snow varying road cut depths of Runs Ba, Bb, and Be in 1995 41 event for South Tributary due to the 3.3.1 Redfish Creek streamflows The resulting streamflow of the three model runs exhibit a similar response pattern during the snowmelt season (Figure 16). Discharge is slightly increased during the beginning of the melt season through the melt water peak. During the recession of the annual melt flows, the three modelled discharges show slightly decreased flows. Results of model Runs C and D, however show an opposing response to rain-on-snow and rainfall events compared to Run Bb. later during the water year. The peaks resulting from these events are generally reduced for Run Bb but are increased for Runs C and Run D. Extending the stream network to represent culverts connected to streams via surface flow resulted in streamflows that are more responsive to diurnal snowmelt and rainfall events, as shown in Figure 16.b., in comparison to flows resulting from just the addition of primary logging roads with field defined road cut depths, shown in Figure 16.a. Incorporating stream extensions decreased the 1995 Redfish Creek yield by less than 0.1% in comparison to the roadless scenario, whereas the logging roads alone decreased the flows by 0.1 %. The average annual yield for 1994 through 1997 is decreased by 1.5% and 1.3% for Runs Bb and C, respectively. The 1995 peak flow increased by 0.3% and 0.7% for runs Bb and C, respectively. The average annual peak flow, for 1993 through 1997, decreased by less than 0.1 % for Run Bb and increased by 0.6% for RunC. Discharge changes resulting from the addition of skid trails and abandoned roads to primary logging roads and stream extensions are shown as the overlying graph in Figure 16.c. The rainfall induced streamflow peaks of the skid trail and abandoned road scenario show a lesser difference than the scenario including stream extensions with primary logging roads. The scenario represented by Run D decreased annual yield for 1995 and the average annual yield for 1994 to 1997 by 0.1% compared to that of Run A. The 1995 peak flow increased by 1.1% and the average annual peak flow for 1993 through 1997 increased by 0.7%. A closer look at the effect of the primary logging roads on the snowmelt generated annual peak flow for 1995, displayed in Figure 17, shows a sinusoidal pattern of discharge difference between the roadless scenario and Runs Bb, C, and D. The impact of Runs C and D are more pronounced than that of Run Bb. The results of Runs C and D exhibit the greatest amount of increased streamflow just prior to the peak and the greatest amount of decrease during streamflow decline over the 24-hour period of diurnal streamflow response to snowmelt. The 42 0.16 16.a. Run Bb - Run A , effect of field surveyed road cut depths 0.08 0.04 0.00 -0.04 -0.08 0.16 0.12 0.08 0.04 0.00 -0.04 -0.08 0.16 0.12 0.08 0.04 0 -0.04 H -0.08 -0.12 -0.16 -0.2 -0.24 -0.28 -C CO i " i • I fr '•"«T-rr~*»"' r S u , 16.b. Overlay of Run C - Run A, effect of stream density expansion I . I I I M Ml • i i ilLi X, j -1 Wf 16.c. Overlay of Run D - Run A , effect of abandoned roads and skid trails SI. '— J= u I a "3 O • Run Bb - Run A Run C - Run A Run D - Run A Run A Figure 16. Discharge differences for Redfish Creek due to the varying simulation scenarios of Runs Bb, C, and D during water year 1995 43 addition of roads, as expressed through Runs Bb, C, and D, do not impact the peak to the extent that the timing of the peak flow was altered. The differences resulting from the addition of primary logging roads during a rain-on-snow peak flow event are shown in Figure 18. The streamflow changes fluctuate with the rainfall intensity. The addition of logging roads (Run Bb) decreases the peak flow by 1.1%, while the stream network expansion and addition of skid trails and abandoned roads has the opposite effect, increasing the peak flow by 1.3% and 1.2 % respectively. 3.3.2 Redfish Upper tributary flows The addition of primary logging roads results in generally decreased flows, with slight increases to rain-on-snow events during the period of snow accumulation. Results from Run C (Figure 19.b) show that increasing the stream length by 5% from 3.51 to 3.69 km results in higher flows for rain-on-snow events than the results of Run Bb. Increasing the road length by 11%) from 1.73 to 1.92 km in Run D results in a greater sensitivity to snowmelt around the seasonal snowmelt peak. Unlike the responses to Runs Bb and C, Run D yields more variable changes to diurnal snowmelt fluctuations, with occasional increases of streamflow during the seasonal snowmelt. As with the effect of stream network extension on Redfish Creek, streamflows are more 1 responsive to diurnal snowmelt and rainfall events when the stream network is extended (Figure 19.b), in comparison to flows without stream network extensions and with primary logging roads with field-defined road cut depths (Figure 19.a). Incorporating stream extensions decreased the 1995 Redfish Upper Tributary yield by 3.8% in comparison to the roadless scenario, whereas the logging roads alone decreased the flows by 3.2%. The average annual yield for 1994 through 1997 is decreased by 3.5% and 4.2% for Runs Bb and C, respectively. The 1995 peak flow decreased by 4.7% and 2.3% for runs Bb and C, respectively. The average annual peak flow, for 1993 through 1997, decreased by 3.1% for Run Bb and decreased by 1.3% for Run C. Discharge changes resulting from the addition of skid trails and abandoned roads to primary logging roads and stream extensions are shown as the overlying graph in Figure 19.c. The response of Run D to rain-on-snow events during the period of snow accumulation is nearly identical to that of Run C. Although the snowmelt modelled by Run D shows fluctuation 44 •Run A • Run Bb - Run A Run C - Run A • Run D - Run A Figure 17. Discharge differences during the annual snowmelt peak for Redfish Creek due to the varying simulation scenarios of Runs Bb, C, and D in 1995 18.a. Discharge differences due to stream density expansion and additional roads Run Bb - Run A Run C - Run A • Run D - Run A 18.b. Corresponding rainfall and streamflow for the rain on snow event Q o C O z o Figure 18. Discharge differences during a rain on snow event for Redfish Creek due to the varying simulation scenarios of Runs Bb, C, and D in 1995 45 . . 1 i 1 r M—V >*V 1T I9.b. Overlay of Run C - Run A , effect of stream density expansion I9.c. Overlay of Run D - Run A , effect of abandoned roads and skid trails 0.010 0.005 0.000 -0.005 -0.010 Run Bb - Run A • Run C - Run A • Run D - Run A Run A Figure 19. Discharge differences for Redfish Upper Tributary due to the varying simulation scenarios of Runs Bb, C , and D during water year 1995 46 between positive and less negative differences as compared with Run C, the difference from Run A annual yield is just 0.6% more than that of Run C. For the scenario represented by i Run D, annual yield for 1995 and the average annual yield for 1994 to 1997 decreased by 3.2% and 3.4%), respectively, compared to that of Run A. The 1995 peak flow decreased by 2.3% and the average annual peak flow for 1993 through 1997 decreased by 0.2%. The change to the peak snowmelt induced streamflow for the Redfish Upper Tributary is displayed in Figure 20. The addition of the stream extensions in Run C causes a slightly greater sensitivity in decreased streamflow than the simulation of the main logging road impact of Run Bb, although the magnitude of decrease is similar. The addition of skid trails and abandoned roads decreased the negative difference to Run A streamflows. As shown in Figure 21, the addition of stream extensions resulted in a reduced change in the 1995 annual peak flow caused by a rain-on-snow event, compared to that induced by the addition of roads alone. Including the skid trails and abandoned roads only slightly increased the negative difference resulting from the combined effect of modelling the main roads and stream extensions as compared with Run A streamflows. 3.3.3 South tributary flows The streamflow responses to Runs Bb, C, and D are nearly identical. The greatest differences between the streamflows of Run A and those of Runs Bb, C, and D occur approximately two weeks prior to the seasonal snowmelt peak. Following the peak increase in streamflows, it is evident that these differences are slightly reduced from Run Bb to Run D. The increases to the rain-on-snow generated peak were also slightly reduced with the small addition of stream length on top of the main logging roads and further reduced with the addition of skid trails and abandoned roads. The expansion of the stream network contributing to South tributary flows, as shown in Figure 22.b, results in increases to both annual yield and annual peak flow. These increases are less than those induced by Run Bb, the addition of logging roads alone (Figure 22.a). Incorporating stream extensions increased the 1995 South Tributary yield by 17.1% in comparison to the roadless scenario, whereas the logging roads alone increased the flows by 18.7%. The average annual yield for 1994 through 1997 is increased by 19.6% and 18.0% for Runs Bb and C, respectively. The 1995 peak flow increased by 5.5% and 5.4% for runs Bb and 47 sz _c sz sz sz sz j2 jEl sz J3 JZ: .C sz 4= sz sz SO r> OC o cs m rf m so r-- oc Os O ci rn ri rn ~t in o o —: CS cs <S rs O 5 © © © b © in in tn in «n >n •n •n (n •n in »n in «n in >n 'n •n in •A •n in. in m m cr- cr -> Qi Cr » c> cr- & g; Cr o cr Qh 0 Cr Q^  C> cr- CT Os c- OS c- C CT- » cr f> cr os cr r> CT1 Cr QS Cr- Cr c~ Cr- C* Ss a- as Os o o o O O O O O a ©" a © a O © O rn m m m m m m m rn rn rn m rn m rn rn m m m rn m rn m rn m in <n «n «/n «n in >n »n •n in «n in m 'n «n fcn »n m •n in •n in in •n in O p. O o O o O o © o © O o © o O o © O O © o o O •Run A - Run Bb - Run A Run C - Run A • Run D - Run A Figure 20. Discharge differences during the annual snowmelt peak for Redfish Upper Tributary due to the varying simulation scenarios of Runs Bb, C, and D in 1995 0.000 -E -0.005 21. a. Discharge differences due to stream density expansion and additional roads c fa a st fa C5 -0.010 -0.015 -0.020 -I -0.025 ox fa a JZ 5 >. L a a. a. 3 B •a 0.8 0.7 0.6 0.5 0.4 0.3 -I 0.2 0.1 Run Bb - Run A Run C - Run A • Run D - Run A 21.b. Corresponding rainfall and streamflow for the rain on snow event 0 E e c io e 20 1 •15hr -16hr •17hr -18hr •19hr •20hr •21hr zz (N r j 23hr •OOhr Olhr •02hr '1995 '1995 '1995 '1995 '1995 1995 1995 1995-1995-1995-1995-1995--f cj r l - f - r CJ - r r l T -F fN t t C'1 - f C m i/-, o SO o a sc o -£ © ••6 © .£ S 5 o o o o c Figure 21. Discharge differences during a rain on snow event for Redfish Upper Tributary due to the varying simulation scenarios of Runs Bb, C, and D in 1995 48 22.a. Run Bb - Run A, effect of field surveyed road cut depths 22.b. Overlay of Run C - Run A, effect of stream density expansion 0.025 0.020 -I 0.015 0.010 0.005 0.000 -0.005 -o.oio H -0.015 -0.020 -I -0.025 -0.030 --0.035 --0.040 22.c. Overlay of Run D - Run A, effect of abandoned roads and skid trails o o o -r o Run Bb - Run A • Run C - Run A • Run D - Run A Run A 0.2 0.1 s Figure 22. Discharge differences for South Tributary due to the varying simulation scenarios of Runs Bb, C, and D during 1995 49 C, respectively. The average annual peak flow, for 1993 through 1997, increased by 11.3% for Run Bb and increased by 10.8% for Run C. The only change to peak flow timing occurred for water year 1997. For the roadless scenario, peak flow occurred June 11. The peak occurred ten days earlier on June 11 for Runs Bb and C. The addition of skid trails and abandoned roads to primary logging roads and stream extensions only slightly diminish the increased South Tributary response to the addition of main logging roads and stream extensions (Figure 22.c). The scenario represented by Run D increased annual yield for 1995 and the average annual yield for 1994 to 1997 by 16.3% and 17.0%), respectively, compared to Run A. The 1995 peak flow increased by 4.4% and the average annual peak flow for 1993 through 1997 increased by 9.2%. As compared with the roadless scenario, the annual peak flow date was approximately three weeks earlier, on May 17. The change to the peak snowmelt induced streamflow for the South Tributary is displayed in Figure 23. It appears that streamflow changes resulting from the slight expansion of the stream network, Run C, and the further addition of road length, Run D, serve to just slightly reduce streamflow increases. As shown in Figure 24, the addition of stream extensions reduced the 1995 rain-on-snow generated annual peak flow compared to the streamflow change induced by the addition of roads alone. Including the skid trails and abandoned roads only slightly reduced the increased difference resulting from the combined effect of modelling the main roads and stream extensions as compared with Run A streamflows. 3.4 Redfish Upper Tributary modelled and observed flows As modelled Redfish Upper tributary flows had shown discrepancies when compared to observed flows, this chapter concludes with the presentation of modelled versus observed Redfish Upper tributary flows. The changes modelled by Run D, representing logging roads and related features, as contrasted with observed flows are displayed in Figures 25, 26, and 27 for water years 1995 through 1997. Although the addition of the logging roads and related features appears to reduce flows as compared with the original Run A roadless model, the modelled flows remain appreciably higher than the observed flows. 50 "g 0.7 5 0.5 I °-4T £ 0 . 3 s 5 0.2 H S 0.1 H £ 0 4-9 o T 0.02 JJ 0.01 ^ 3 o.oo g £ -0.01 s u ex -0.02 « fi 5 -0.03 5 1_ ,2 u fi , c t-fi (_ fi fi — fi JE fi £ fi i— fi !— fi <- J=. ta fi fi r> 0 0 c o ro «n r> CO OS o r-i ro o i ro •n 53 cp o o o rN CN CN r s c o o O o O o «r> tA LTJ •A v i u i in *n in «n <n in in in m m in in in in in in •n >n in c^  c- o aj C N r> c-- CN. c~ C N C N C N C N C N •CN c C N C N C N c- ON C N o> C N O N •c- C N C ^ Os C N C N C N CN. Of C N O CN- C N C N ©> 'CN ON ON *—. o o o © O o o o o © o o o O © o o o ro ro ro ro ro ro CO CO ro ro ro ro <"*"> r*1 ro rO ro ro ST) ro ro ro ro >n in <n «n <n in >n in in *n in in •n tn in in iri in in in V) in in © © o o o O O o O o O o o O o © o O o o o o o O o •Run A • Run Bb - Run A Run C - Run A • Run D - Run A Figure 24. Discharge differences during a rain on snow event for South Tributary due to the varying simulation scenarios of Runs Bb, C, and D in 1995 Figure 23. Discharge differences during the annual snowmelt peak for South Tributary due to the varying simulation scenarios of Runs Bb, C, and D in 1995 51 « 0.02 e is 5 BC i_ cd j= CA s o.oo -0.02 -0.04 -0.06 -0.08 -0.10 . 1 1 . 1 1 1 ..Mi L 1 0.6 0.5 0.4 J G? 5 0.3 « 3 a o.2 g Q . -JC o.i £ o.o 9 • Run D - Run A Run D • Observed Figure 25. Modelled Redfish Upper Tributary response to logging roads compared with the modelled roadless scenario and observed flow for water year 1995 5s SO OS 5 rl Run D - Run A Run D • Observed Figure 26. Modelled Redfish Upper Tributary response to logging roads compared with the modelled roadless scenario and observed flow for water year 1996 52 — _ - o — o — — — — — — o o o m r o f ^ m o o o m o m n m o o — fN — r - i ^ v ^ i / ^ r - r - o o o — — — — o o o o o o o o o Run D - Run A Run D Observed Figure 27. Modelled Redfish Upper Tributary response to logging roads compared with the modelled roadless scenario and observed flow for water year 1997 53 C H A P T E R 4: DISCUSSION 4.1 Overview This chapter will focus on the objectives outlined in the introduction, which include the examination of modelled streamflow response to the addition of roads, road cut depth variation, stream network extension to culverts connected via surface flow, and the addition of skid trails and abandoned roads. The sub-basin model with enhanced road features will also be compared with observed streamflows. The discussion is focused primarily within the context of basin and sub-basin scale differences with references to the average annual yield and average annual peak changes. 4.2 Streamflow response to the addition of logging roads at the basin and sub-basin scale At the scale of Redfish Creek basin (25.8 km2), no substantial change to streamflow was observed in terms of annual yield and annual peak flow with the addition of primary logging roads, whereas greater changes to tributary flows were observed. The degrees to which logging roads affect these streamflows are likely due to various factors such as logging road density, expansion of the basin contributing area, and particularly, road location with respect to streams. The lack of considerable alterations to Redfish Creek streamflow is consistent with previous field studies showing insignificant change to streamflows when logging road area occupies 5% or less of the watershed area (Ziemer, 1981; Harr et al., 1975). Based on running width, logging roads occupy less than 1% of the Redfish Creek watershed area. Redfish Creek tributaries appear to be more sensitive to the effects of logging roads than the whole Redfish Creek basin. The annual yield and annual peak changes for were greater for the South Tributary than for Redfish Upper. Relative to the roadless case, the South Tributary experienced increases in annual yield and annual peak, whereas Redfish Upper showed decreases. The degrees of streamflow changes were not proportional to the percent of the sub-basin area covered by road, suggesting that, in the case of these two sub-basins, percent roaded area has little affect in determining streamflow impact at the sub-basin scale. The difference in relative changes is more likely due to road location and changes in contributing area. 54 The differing results for the adjacent sub-basins superficially suggest that subsurface flow was transferred from Redfish Upper Tributary to the South Tributary by means of the logging road, similar to the case for the DHSVM-derived model of the Carnation Creek watershed, where water was transferred between two neighbouring sub-basins via a road (Wigmosta and Perkins, 2001). However, upon examining the topography surrounding the sub-basins and the potential subsurface flow path alteration due to the logging road, it appears that subsurface flow was not directly diverted from one sub-basin to another by the logging road and thus would not be an explanation of the simulated tributary flow changes. The common boundary of the sub-basins is located well above (600 m) the road location so there is no direct interception and transfer from one sub-basin to another by way of roads. The presence of roads increased the contributing area of the sub-basins according to road altered sub-basin boundaries. These boundaries were determined using road and culvert locations and cell to cell flow paths based on surface topography. The resulting South Tributary contributing area potentially increased by approximately 20%. This may account for the increased flow of the South Tributary for the road scenarios. The increased contributing area of the South Tributary extends out onto the hillslope that separates it from the Redfish Upper sub-basin. The expansion of the South Tributary contributing area does not cause a decrease in the road-altered Redfish Upper contributing area. The slight increase in Redfish Upper contributing area is contrary to the decrease in tributary flow. This may be an artefact of the D H S V M multiple flow path routing routine (Wigmosta et al., 1994). As not all the water contained by a cell is transferred to the most downslope neighbour, but may distribute amongst all adjacent downslope cells, effective subsurface flow boundaries are larger than those based on cell to cell transfer of subsurface flow. The complexity of the multiple flow path is compounded with the presence of the switchbacks located on the hillslope between the two basins. The switchbacks are capable transferring flows between subsurface and surface paths by the conversion of subsurface flow to surface flow at road cuts and the subsequent transfer of surface flow back to subsurface flow at ditch discharge locations, thereby returning flow to the subsurface for potential interception by a lower switchback. Considering the increased potential for contributing area expansion given the multiple flow paths, it is possible that a transfer of flow did occur between the sub-basins, resulting in the decreased Redfish Upper tributary flows. 55 4.3 Model sensitivity to cut depth variations The sensitivity of modelled annual yield and annual peak flows due to variations in logging road cut depth was greater in the smaller catchments. The low road density associated with Redfish Creek watershed inhibits a sensitivity analysis of cut depth because the addition of primary logging roads results in a negligible impact on streamflow regardless of cut depth. The greater South tributary flow changes may be influenced to some degree by the expansion of the contributing area to the South Tributary. The larger the contributing area increase, coupled with increasing road cut depth, the greater the potential to intercept subsurface flows. It is intuitive that deep road cuts would intercept larger volumes of subsurface flow; such effects were demonstrated by Wigmosta and Perkins (2001). Deeper road cuts would therefore pose potentially greater affects to streamflow. However, such relationships were not consistent in Redfish Creek and its tributaries. The deepest cut depth scenario did not create the greatest impact on annual snowmelt and rain-on-snow peaks for Redfish Upper Tributary. The scenario incorporating the field-defined road cuts generated the largest change to these tributary flows. It may be that subsurface flow, which would have been intercepted and routed to Redfish Upper sub-basin in the case of the road cut depth equalling the soil depth, is passing below the road surface to either flow down the hillslope away form the Redfish Upper sub-basin or is intercepted by lower switchback road cuts for distribution away from the sub-basin. This phenomenon may, in part, explain the lesser decrease in annual yield and peak flow for the field-defined scenario as compared with the deepest cut depth scenario. 4.4 Effect of surface flow connections from culverts to the stream network The primary effect of extending the stream network to connected culverts should be an increase in streamflow responsiveness, expressed as a shortened time to peak and increased magnitude of peak flows (Wemple et al, 1996). However, for all of the catchments examined in this study, changes to either the average annual yield or peak caused by stream network extensions were less than 2% in addition to the effect of the primary logging roads with field specified road cuts. Further, although slight increases to annual streamflow peak did occur in Redfish Creek and Redfish Upper Tributary, there was no decrease in the time to peak for either snowmelt or rain-on-snow events at both scales. A decrease in the time to peak may have 56 occurred if a greater portion of downslope flow was intercepted by connected ditch-culvert systems, whereas a conservative, but field based, approach was taken to determine the connectivity of culverts. It is also possible that changes in timing of runoff peaks could be detected by employing a smaller temporal resolution instead of use of hourly time steps. 4.5 Hydrologic influence of skid trails and abandoned roads The addition of skid trails and abandoned roads produced some change in the annual yield of Redfish Upper and South tributaries, over and above the effects of adding the channel extension. As compared against the base scenario without roads, Redfish Upper annual yield increased slightly from -4.2%, for the stream extension scenario, to -3.4% after the addition of skid trails and abandoned roads. The South Tributary annual yield decreased from 18.0% for the stream extension scenario to 17.0%. The addition of an abandoned trail on the hillslope between the two sub-basins may transfer subsurface flows away from the South Tributary to the Redfish Upper Tributary. Only minimal changes to annual peaks were observed at the Redfish Creek basin scale, probably because less than 1% percent watershed is covered by primary logging roads, abandoned roads, and skid trails. Even though the low permeability of the road is not represented in the model, it is unlikely that such an addition to the model would have changed the results given the scale of the basin and the low proportion of roads. 4.6 Enhancement of sub-basin model with road features The addition of logging roads did not substantially improve the modelled representation of observed Redfish Upper tributary flows. The addition of roads reduced the streamflows that were overestimated in the unroaded simulation, but only by a negligible amount in compared with observed flows. The base scenario without roads had flashy peaks compared to observed flows. The addition of roads, more notably with channel extensions, increased modelled tributary flow responsiveness. Other sources of error that potentially contributed to the overestimation of Redfish Upper tributary flows include high natural variability of soil characteristics, hillslope flow paths that differ from DHSVM's flow path algorithm and DEM 57 inaccuracy (Whitaker et al., 2001). It is also possible that parameters controlling precipitation, snow accumulation, or snowmelt may have been inaccurately specified for the model. Whereas inaccurate soil representation in the model would likely have a greater effect on the timing of flows, the inaccurate definition of contributing area and specification of precipitation parameters affect streamflow yield. Subsurface flow routing can be controlled by the topography of the confining layer at the bottom of the soil profile (Hutchinson and Moore, 2000), as well as complex patterns of preferential flow (Sidle et al., 2000). D H S V M and other models assume that subsurface flow paths conform to surface topography. Differences between true and modelled flow paths will affect stream yield to some extent. The modelled flow path is dependent largely on the D E M , which underscores the potential impact of D E M errors. Studies have also shown substantial variability of hydrologic modelling results based on differing horizontal and vertical scales (Zhang and Montgomery, 1994; Kenward et al., 2000). Lack of sufficient topographic detail, which can only be expressed with higher resolution DEMs, can lead to inaccurate flow path representation. As seen with the potential to alter annual yield in the South Tributary, road placement relative to the D E M is also an important factor for sub-basins. Two of three switchback bends, located on an adjacent hillslope, deposit flow intercepted from the hillslope directly to the South Tributary sub-basin. The increase in South Tributary contributing area is also dependent on the location of roads and culverts. An error in the D E M was detected during the initial overlay of road locations, acquired by GPS, and stream locations, digitized from aerial photos. The east side of the Redfish Creek watershed topography appeared to be displaced northward by almost 50 m. Road placement error could have resulted from shifting the east side logging roads location to better align with the topography. 58 C H A P T E R 5: C O N C L U S I O N 5.1 Summary of main findings For the snow dominated regime of Redfish Creek watershed, the impact of logging roads on annual and peak streamflows was generally very minor; these minor changes varied with basin size and topography. The low density of logging roads produced negligible changes at the scale of Redfish Creek watershed. Redfish Upper Tributary experienced decreases of less than 5% in annual yield and peak flow, while the South Tributary experienced larger changes in the form of a 19.6% increase in annual yield and an 11.3% increase in annual peak flow following the addition of primary logging roads with field specified road cuts. A complex interaction between the multiple flow path subsurface flow routing routine and redistribution of subsurface flows at the switch back bends that were located at the edge of the sub-basin boundary likely resulted in the Redfish Upper flow reductions. An increased contributing area of approximately 20% due to the presence of logging roads likely caused the increased South tributary flows. Increased sensitivity to road cut depth accompanied those scenarios in which modelled streamflows show notable changes due to the general presence of logging roads. The greatest sensitivity was found for annual flow from the South Tributary; increases ranged from 2.6% for the minimal cut depth scenario to 20.8% for road cuts equal to soil depth. The addition of culvert connectivity through stream density expansion and skid trails and abandoned roads did not impart notable additional streamflow changes to any of the watersheds examined. The lack of change resulting from stream density expansion may be due to a small amount of stream length added to the existing stream network. However, all streamflows became more responsive to the increased surface flow efficiency resulting from the increased stream density. The addition of skid trails and abandoned roads appeared to cause the transfer of a portion of subsurface flow from the South Tributary to Redfish Upper Tributary. The addition of logging roads to Redfish Upper Tributary did not noticeably improve the simulated flows compared with observed flows. The source of tributary flow overestimation may lie in possible errors in the D E M , precipitation inputs, and/or melt rates. 59 5.2 Suggestions for further research Although this project has generated some insights into the effects of roads on streamflow in a snowmelt-dominated mountain catchment, the results are specific to the current road network in Redfish Creek, and may not be applicable to catchments with greater road densities, especially at higher elevations. Some of the issues that deserve further investigation are outlined in the next paragraph. The streamflow impact of varying road densities should be examined to explore the relationship between road density and peak flow changes. Variations in study site characteristics, such as a high plateau catchment with lower relief and a more synchronised snowmelt timing of annual snowmelt, would develop further understanding of potential road impacts in snow dominated regimes. Another avenue of continued research involves contrasting the effects of multiple and single flow path algorithms in determining the road impacts on subsurface path and contributing areas. 60 R E F E R E N C E S Betson, R.P., 1964. What is watershed runoff? Journal of Geophysical Research, 69 (8): 1541-1552. Beven, K., 1982. On subsurface stormflow: an analysis of response times. Hydrological Sciences Journal, 4: 505-521. Bowling, L.C. and Lettenmaier, D.P., 1997. Evaluation of the effects of forest roads on streamflow in Hard and Ware Creeks, Washington. Water Resources Series Technical Report No. 155, University of Washington, Department of Civil Engineering, Seattle, WA. Bonell, M . , 1993. Progress in the understanding of runoff generation dynamics in forests. Journal of Hydrology, 150: 217-275. Bosch, J.M. and Hewlett, J.D., 1982. A review of catchment experiments to determine the effects of vegetation changes on water yield and evapotranspiration. Juornal of Hydrology, 55: 3-23. Buttle, J.M. and House, D.A., 1997. Spatial variability of saturated hydraulic conductivity in shallow macorporous soils in a forested basin. Journal of Hydrology, 203: 127-142. Chamberlin, T.W., 1972. Interflow in the mountainous forest soils of coastal British Columbia. In: H.O Slaymaker and H.J McPherson (Editors), Mountain Geomorphology. Tantalus Research: Vancouver, BC, pp. 121-127. Croke, J. and Mockler, S., 2001. Gully initiation and road-to-stream linkage in a forested catchment, Southeastern Australia. Earth Surface Processes and Landforms, 26: 205-217. Gardner, B.D. and Chong, S.K., 1990. Hydrologic responses of compacted soils. Journal of Hydrology, 112: 327-334. Germann, P. and Beven, K., 1981. Water flow in soil macropores. 1. An experimental approach, Journal of Soil Science, 32: 1-13. Gray, D.M. and Prowse, T.D., 1993. Snow and floating ice. In: D.R. Maidment (Editor), Handbook of Hydrology, McGraw-Hill, USA, pp. 7.1-7.58. Harr, R.D, Harper, W.C., Krygier, J.T. and Hsieh, F.S., 1975. Changes in storm hydrographs after road building and clear-cutting in the Oregon Coast Range. Water Resources Research, 11 (3): 436-444. Hewlett, J.D. and Hibbert A.R., 1967. Factor affecting the response of small watersheds to precipitation in humid areas. In: W.E. Sopper and H.W. Lull (Editors), International Symposium on Forest Hydrology, Pergamon Press, New York, pp. 275-290. 61 Hutchinson, D.G. and Moore, R.D., 2000. Throughflow variability on a forested hillslope underlain by compacted glacial till. Hydrological Processes, 14: 1751-1766. Jones, J.A. and Grant, G.E., 1996. Peak flow responses to clear-cutting and roads in small and large basins, western Cascades, Oregon. Water Resources Research, 32 (4), 959-974. Kenward, T., Lettenmaier, D.P., Wood, E.F. and Fielding, E. , 2000. Effects of digital elevation accuracy on hydrologic predictions. Remote Sensing of Environment, 74: 432-444. Kim, H.J., Sidle, R . C , Moore, R.D. and Hudson, R., 2003. Throughflow variability during snowmelt in a forested mountain catchment, Coastal British Columbia, Canada. In Press. King, J.G. and Tennyson, L.C. , 1984. Alteration of streamflow characteristics following road construction in North Central Idaho, Water Resources Research, 20 (8): 1159-1163. LaMarche, J. and Lettenmaier, D.P., 1998. Forest road effects on flood flows in the Dechutes River basin, Washington. Water Resources Series Technical Report No. 158, University of Washington, Department of Civil Engineering, Seattle, WA. Luce, C H . and Cundy, T.W., 1994. Parameter identification for a runoff model of forest roads. Water Resources Research, 30 (4): 1057-1069. Megahan, W.F., 1972. Subsurface flow interception by a logging road in mountains of central Idaho. Paper presented at Symposium on Watersheds in Transition, June 1972, American Water Resources Association, Ft. Collins, Colorado. Montgomery, D.R., 1994. Road surface drainage, channel initiation, and slope instability. Water Resources Research, 30 (6): 1925-1932. Mosley, M.P., 1982. Subsurface flow velocities through selected forest soils, South Island, New Zealand. Journal of Hydrology, 55: 65-92. Roberge, J. and Plamondon, A.P., 1987. Snowmelt runoff pathways in a boreal forest hillslope, the role of pipe throughflow. Journal of Hydrology, 95: 39-54. Sidle, R . C , Noguchi, S., Tsuboyama, Y. and Laursen, K., 2001. A conceptual model of preferential flow systems in forested hillslopes: evidence of self-organization. Hydrological Processes, 15: 1675-1692. Sidle, R . C , Tsuboyama, Y., Noguchi, S., Hosoda, I., Fujieda, M . and Shimizu, T., 2000. Stormflow generation in steep forested headwaters: a linked hydrogeomorphic paradigm. Hydrological Processes, 14: 369-385. Tsuboyama, Y., Sidle, R . C , Noguchi, S., Murakami, S. and Shimizu, T., 2000. A zero-order basin - its contribution to catchment hydrology and internal hydrological processes. Hydrological Processes, 14: 387-401. 62 Wemple, B.C., 1994. Hydrologic integration of forest roads with stream networks in two basins, Western Cascades, Oregon. M.S. Thesis, Department of Geosciences, Oregon State University, Corvallis. Wemple, B.C., Jones, J.A. and Grant, G.E., 1996. Channel network extension by logging roads in two basins, Western Cascades, Oregon. Water Resources Bulletin, 32 (6): 1195-1207. Whitaker, A . C , Alila, Y., Calvert, P. and Toews, D.A.A., 1998. Evaluation of existing hydrological models for use in assessing the impact of forest management on peak flows. In: Y. Alila (Editor), Mountains to sea: human interaction with the hydrologic cycle: proceedings, Canadian Water Resources Association, 51st annual conference, June 1998, Victoria, B.C., pp. 76-71. Whitaker, A .C. and Alila, Y. , 2000. Distributed hydrologic modelling in Redfish Creek, West Kootenays, British Columbia. Part 1: application, calibration and verification using internal catchment data. Part 2: sensitivity of peak flows to clear cut forest harvesting across different elevation zones. Unpublished reports by Faculty of Forestry, University of British Columbia, Vancouver, B.C., submitted to B.C. Ministry of Forests, Nelson Region, Nelson, B.C. Whitaker, A., Alila, Y. and Toews, D., 2001. Modelling of peak flow change using the D H S V M Model. In: D.A.A. Toews and S. Chatwin (Editors), Watershed Assessment in the southern Interior of British Columbia, Work Paper 57/2001, Research Branch, BC Ministry of Forestry, Victoria, British Columbia, Canada, pp. 94-111. Whitaker, A., Alila, Y. and Beckers, J., 2002. Evaluating peak flow sensitivity to clear-cutting in different elevation bands of a snowmelt-dominated mountainous catchment. Water Resources Research, 3 8 (9), 11-1-11-17. Whitaker, A., Alila, Y. , Beckers, J. and Toews, D., 2003. Application of the Distributed Hydrology Soil Vegetation Model to Redfish Creek, British Columbia: model evaluation using internal catchment data. Hydrological Processes, 17: 199-224. Wigmosta, M.S. and Perkins, W.P., 1997. A GIS-Based modeling System for Watershed Analysis, Final Report to the National Council of the Paper Industry for Air and Stream Improvement (NCASI), 160 p. Wigmosta, M.S. and Perkins, W.A., 2001. Simulating the effects of forest roads on watershed hydrology. In: M.S. Wigmosta and S.J. Burges (Editors), Land Use and Watersheds: Human Influence on Hydrology and Geomorphology in Urban and Forest Areas, Water Science and Application 2, American Geophysical Union, Washington, DC, pp. 127-143. Wigmosta, M.S., Vail, L.W. and Lettenmaier, D.P., 1994. A distributed hydrology-vegetation model for complex terrain. Water Resources Research, 30 (6): 1665-1679. Ziegler, A.D. and Giambelluca, T.W., 1997. Importance of rural roads as source areas for runoff in mountainous areas of northern Thailand. Journal of Hydrology, 196: 204-229. 63 Zimmer, R.R., 1981. Storm flow response to road building and partial cutting in small streams of Northern California. Water Resources Research, 17 (4): 907-917. Zhang, W. and Montgomery, D.R., 1994. Digital elevation model grid size, landscape representation and hydrologic simulations. Water Resources Research, 30: 1019-1028. / > 64 Appendix Tabulated and Graphed Results for 1993, 1994, 1996, and 1997 65 e o .2 "E > -S +H OH cu T3 •a « o a o is DX) a cu u. CU OX) a « .a <y 2 "3 « a a « CU i-U CO TJ cu — « ON ON -•9 eg fi OO ON N t oo NO OO o\ CN r—< • * t ON IO o ON OO O NO NO NO NO NO NO NO NO CN CN CN CN N O O N ON 2 "3 G I/O I/O ON oo I/O NO CN ON I/O ro ro ro • * t NO NO ON oo o NO NO NO NO ro ro ro ro in O N O N O N O N ro O N O N | . 2 S <; ^ u e 3 Pi . - H CN © d o V ' 1 2? N§ ON N? © d • d \ 0 vO , — i o x o x d ^ c n -i o o V ' ' © © o ON t-- ON ON io CN I/O o ON ro oo NO o NO NO N * m NO NO NO NO ro ro ro ro CN CN CN CN 2? N? ON N? ON o CN rod d d ON O 00 NO NO ON NO OO o CN ro o o oo I/O N t N t T ON oo ro ro CN CN CN CN CN CN CN CN CN CN N? ON N? ON- N? ON t—< .—' d d d V V V CN ON oo oo NO NO IO NO O O O O 00 0O 00 0O pa oa M CD a> * ^ 1/1 a 1 C3 +-» O a a o OS "a *3 I NO <D .fi H 3 CD .fi o a o o o a a o a o v 03 CD J3 o CD •s o -4-» ro oa CN OJ &H ro fi ON "• ON ~ ~H ON „ o O CN ro u >-. 60 <D 03 £ ?0 CD o - 00 CN T3 CD (-< CD oo C o T3 C .o fi O O a & 03 t" <D \S S 2 ro ire ON O CJ o >H " bX) < o o § ^ .3 •*3 c j 2 * II £ u fi —<• M C 03 c o _ 03 tfc- CJ • 2 ro - S ON T3 _ o\ o S ^ 03 fi S3 « fi (U M < >< < T3 CD CD 2 fi O o c CD CD .fi (H O0 CD O > V o o ro ON ON 03 CD >N a .2 'IZ cvt > OH cu T3 « O O 0X) a s cn cu U CA <U ex a « O o C3 CU OH c a u <U <U u CA C T3 CU 0 0 s H 3-* CD D 2 — o CD CC W > 3 ft < 3 U 3 N ? N= ON ON • o o ° V V OO N t NO O o r - - I - - r - r > ro ro ro ro NO NO NO NO ro NO d 9 9 N t NO NO O ON CN O — rN rN-NO N t f N IO o o r - - r - - r - r -N? N° N= O ON o x H m \t o d d CN N t NO IO ON NO O CN CN ro io io ro ro ro ro NO NO NO NO N ? O N <^ d ^ d © ° V 1 r - N CN oo NO r > IO r ~ NO N t 1—1 N t N t N t N ? ON N ? 0 N N ? ON-CN CN d d d oo ro NO 00 IO O o r - N IO IO N t ro CN CN CN CN ON ON ON ON < pa pa PO PH ff n ON 66 28.a. Run Ba - Run A , effect of minimal road cut depths approximately equal to ditch depths 0.05 -0.05 28.b. Overlay of Run Bb - Run A , effect of field surveyed road cut depths 28.c. Overlay of Run Be - Run A , effect of maximum road cut depths equal to soil depth • Run Ba - Run A •Run Bb - Run A • Run Be - Run A •Run A Figure 28. Discharge differences for Redfish Creek due to the varying road cut depths of Runs Ba, Bb, and Be during water year 1993 6 7 0.05 29.a. Run Ba - Run A, effect of minimal road cut depths approximately equal to ditch depths -0.05 0.05 29.b. Overlay of Run Bb - Run A, effect of field surveyed road cut depths ' ^Fwnn T f r - — — -0.1 29.c. Overlay of Run Be - Run A, effect of maximum road cut depths equal to soil depth 0.05 T - r 10 -0.05 -0.1 A -0.15 -0.25 o m o a z o Run Ba - Run A • Run Bb - Run A • Run Be - Run A •Run A Figure 29. Discharge differences for Redfish Creek due to the varying road cut depths of Runs Ba, Bb, and Be during water year 1994 68 0.1 -, 30.a. Run Ba - Run A, effect of minimal road cut depths approximately equal to ditch depths 30.b. Overlay of Run Bb - Run A, effect of field surveyed road cut depths -v i*" # , r n rrrvn—-30.C. Overlay of Run Be - Run A, effect of maximum road cut depths equal to soil depth 0.05 -0.05 -0.1 J 0.1 0.05 -0.05 -0.1 0.1 0.05 _Je—3L i. ikJf^. K—IE_KB BHJ—, -0.05 -o.i H -0.15 -0.2 -0.25 o 'A a o 5 © • Run Ba - Run A • Run Bb - Run A • Run Be - Run A •Run A Figure 30. Discharge differences for Redfish Creek due to the varying road cut depths of Runs Ba, Bb, and Be during water year 1996 69 0.15 31 .a. Run Ba - Run A , effect of minimal road cut depths approximately equal to ditch depths 0.15 g" o.i a QJ E 5 0.05 Q II on ~ 0 u M ° -0.05 31 ,b. Overlay of Run Bb - Run A, effect of field surveyed road cut depths J Liu. J? 31 x . Overlay of Run Bc - Run A , effect of maximum road cut depths equal to soil depth • Run Ba - Run A - Run Bb - Run A • Run Bc - Run A Run A Figure 31. Discharge differences for Redfish Creek due to the varying road cut depths of Runs Ba, Bb, and Bc during water year 1997 70 60 ca i-i to > < 2 *cj to 1 o r-os os s© os OS OS ro I T ) r~ fN >n so SO ro o\ os SO CN o tn ro as o t— r- r-1 1 1 1 1 1 ir-es OS 1? I OS Os 2 ro OS OS 2 B 3 S = v O N= o x o N o N <-< in o >^ ro ro "•5 N? CN ro OS T-T ro CN o r-o m m <ji m o o O T-I SO CN OS so CN OO OS N? 5^ ro oo ro T-* ro ro N? o x oo CN so O ro CN os SO so oo ro tr, Os in OS OO fN o 00 SO tr, CN fN ro ro ro ro 1 - 1 T—t "<? OS p d ro ro in O o o so SO SO oo SO •* 00 ro oo SO tr, o r— CN CN CN CN ' — ' N? 5^ T-- ro OS T-i ro CN o r- CN SO 00 -3- tn -rf OS SO o m t O O O ^ M P5 M o CO Os X! Os 3 O n a S •5 •s 1 1 -.2 -a I so (L) I -e a SO o TG H o 8 » S 13 ^ a o .2 a cu •a •*•» 3 u TS C3 O S. s o M l a s w <U I . cn CO 0/3 a cs JS w o CU a "« s a a CU a a & •S is -a M H 00 SO oo CN OS oo 00 r-ro ro ro ro O O O O ro SO CN SO ro oo OS 00 r- SO SO -3-o o d o a cu 3 6 a 3 Pi \ \ \ O x 0S o x 1-4 cn <N oo CN ro d T—1 T-i 8^ CN oo ro CN N? o x c x o N o r - - - * SO — i •* SO CN oo CN ro ro CN — i <—i f 0 0 0 0 eak ange p cN cu d ro ro u i i i OS OS 00 *—< o CN CN SO SO CN CN CN CN d d d d c x o x o N "* CO O CN co ro OS oo 00 so <N oo CO «n 00 so so o^ tr, tr, i n O O d d ca pa PQ 71 0.01 0.00 -0.01 -0.02 -0.03 0.01 0.00 32.a. Run B a - Run A, effect o f min imal road cut depths approximately equal to ditch depths r \ 1 1 If 32.b. Over lay o f Run B b - Run A, effect o f f ield surveyed road cut depths & u B *S -0.01 Q 0 H a -o.o2 H -0.03 0.01 0.00 -0.01 -0.02 i -0.03 -0.04 -0.05 -0.06 32.c. Over lay o f Run B c - Run A, effect o f max imum road cut depths equal to soi l depth • B a - A B b - A • B e - A Run A Figure 32. Discharge differences for Redfish Upper Tributary due to the varying road cut depths of Runs Ba, Bb, and Bc during water year 1993 72 0.005 0.000 -0.005 i -0.010 33.a. Run Ba - Run A, effect of minimal road cut depths approximately equal to ditch depths 33.b. Overlay of Run Bb - Run A, effect of field surveyed road cut depths 0.005 0.000 -0.005 -0.010 -0.015 -0.020 -0.025 -0.030 33.c. Overlay of Run Bc - Run A, effect of maximum road cut depths equal to soil depth • B a - A B b - A B e - A Run A Figure 33. Discharge differences for Redfish Upper Tr ibutary due to the varying road cut depths of Runs Ba, Bb , and B c during water year 1994 73 0.02 -i 0.01 -0.00 --0.01 -0.02 --0.03 --0.04 -0.02 -I Ml "s 0.01 -:rences 0.00 -£ -0.01 -5 irge -0.02 -ischi -0.03 -Q -0.04 -34.a. Run Ba - Run A , effect of minimal road cut depths approximately equal to ditch depths 1 1 34.b. Overlay of Run Bb - Run A , effect of field surveyed road cut depths 0.02 0.01 0.00-34.C. Overlay of Run Be - Run A , effect of maximum road cut depths equal to soil depth -0.01 i -0.02 -0.03 -0.04 -0.05 -0.06 -0.07 --0.08 --0.09 -0.10 f^^f\f - I 1 1 f 1 1™ O N O N O N O N O cn 'M80-966L IM91-9661, '1996-OOhi IM80-966I/ '1996-16hi '1996-OOhi o ON r-i m o ro o ro o O o m o -F o >n o o Ba - A - Bb- A — — B e - A -Run A a 5 o o 3 r) s. — r 5 3S r l o Figure 34. Discharge differences for Redfish Upper Tributary due to the varying road cut depths of Runs Ba, Bb, and Be during water year 1996 74 35.a. Run Ba - Run A, effect of minimal road cut depths approximately equal to ditch depths 35.b. Overlay of Run Bb - Run A, effect of field surveyed road cut depths j 35.c. Overlay of Run Be - Run A, effect of maximum road cut depths equal to soil depth fi jc -n fi J Z fi t-r fi zn. O OO o oo NO o O — o o — ' o o 5 a s O NO r-'~ r- rA r- r-O- c C* 5* 2* ON ON ON c o ON e> c-- ON ON & 0" •c>. o> ON ON o> o~ •—. o o o o co ro ro ro o o o ro © ro ro o — fN .—. ro -r r- ON 1 — — © o o o O O o s © Run Ba - Run A • Run Bb - Run A • Run Be - RUn A Run A Figure 35. Discharge differences for Redfish Upper Tributary due to the varying road cut depths of Runs Ba, Bb, and Be during water year 1997 75 a o > a. cu T3 T3 es o s o DX) s =3 CS) CD >. CS) CU DX) a cs JS CJ 2 • • H 73 s S3 a es >, u cs 3 £ "C H o CO H r-ON ON .—i CD S . S § 1 * 6 2 2 B 3 -—-N O O N O N 1 2 |l 1*3 in O N O N 2 *a B —^  S CD CD Ml x 6 T T O N ON 2 "3 *s S fi ro O N O N N? ON ON s? ^ ^ CO ° . « 2 S Nt NJ-O IO oo ON ON NT ON o ON OO — I I O ON i-i ON CN * - i r o o o N? ON ro NO N" N" ON ON IO -H d CN CN CN 00 CN NO Nt ON T t ON T t i/o r o r -o ~ Nt O NO r— r o ON ™I CN r o N? 0 N NO N§ N ? ON ON r— oo CN 22 NO NO O oo CN CN •o oo Nt IO NO 1—i Nt r o ON IO CN Nt IO r~ oo oo Nt Nf Nt ~ Nt r-N ON —* NO r-P N= ON ON -H O O CN CN CN NO r -i-* CN r - r o r o r-~ r o Nt oo oo N§ N= ON ON Nt ON 00 ON ON O oo ON NO IO r o f N -O oo ON IO * N t N t O CN ^ CN NO NO t - - f N -< pa pa pa Si r o a -a •5 13 2 'P» o es > a. cu S u T 3 03 O u E o DX) a s cs; <U U CS5 <u DX) a CS JS tt o 53 ^ cs CU a 73 a >N S-03 s 3 O CO CN i—I CU 3 cs H J) o o M P . M 2 — C <u s s > 3 ?3 < 3 ° < 1 I3" S cj cj 3 PH -=< < O 03 CD PH C3 fi 03 03 3 "s ^ §n 3 o3 C H CD 03 3 PH -fi < U 1^ 03 CD PH 03 fi B S tf N = 2? $ <^  ^ «^  2 2 N? N= ON ON r o i o oo i o r o CN. O CN ON ON t o t — oo oo CN CN CN CN o d d © N = 6N- O N © ^ °°. ON " ^ r - O O ON O NO NO NO ON CN r o r o Nt Nt d d Nt d i—< IO CN d i o NO o Nt Nt NO r - r - Tt NO -H ^H r o r o r o r o r o r o o d d © N" ON oo N? N° ON ON. i o r o IO ^ — NO NO r o NO IO r o Nt NO NO CN CN CN CN d d d d > F O N O N CN ^ ° ° NO 2 2 O >o r o ^H V) Nt H CN — i r o i o m r o r o r o r o d d d d > CM A U pa ca pa < B 76 0.04 0.03 0.02 0.01 0.00 J5 0.04 -r 0.03 0.02 H 0.01 H 0.00 0.04 0.03 0.02 0.01 0.00 36.a. Run Ba - Run A , effect of minimal road cut depths approximately equal to ditch depths 36.b. Overlay of Run Bb - Run A , effect of field surveyed road cut depths 36.C. Overlay of Run Be - Run A , effect of maximum road cut depths equal to soil depth 0.60 -0.05 O c e Pi 9 • Run Ba - Run A • Run Bb - Run A • Run Be - Run A •Run A Figure 36. Discharge differences for South Tributary due to the varying road cut depths of Runs Ba, Bb, and Be during water year 1993 77 0.04 37.a. Run Ba - Run A, effect of minimal road cut depths approximately equal to ditch depths 37.b. Overlay of Run Bb - Run A, effect of field surveyed road cut depths 0.00 0.04 0.03 -I 0.02 0.01 0.00 -0.01 -0.02 i -0.03 -0.04 -0.05 37.c. Overlay of Run Bc - Run A, effect of maximum road cut depths equal to soil depth - - ' . / A - ~ k/4 3? 0 Run Ba - Run A • Run Bb - Run A • Run Bc - Run A •Run A Figure 37. Discharge differences for South Tributary due to the varying road cut depths of Runs Ba, Bb, and Bc during water year 1994 78 38.a. Run Ba - Run A, effect of minimal road cut depths approximately equal to ditch depths 38.b. Overlay of Run Bb - Run A, effect of field surveyed road cut depths 38.c. Overlay of Run Be - Run A, effect of maximum road cut depths equal to soil depth 0.04 i 0.60 0.50 0.40 0.30 0.20 3 o so o- sp so O v ON • Run Ba - Run A • Run Bb - Run A • Run Be - Run A Run A Figure 38. Discharge differences for South Tributary due to the varying road cut depths of Runs Ba, Bb, and Be during water year 1996 79 0.06 39.a. Run Ba - Run A, effect of minimal road cut depths approximately equal to ditch depths 39.b. Overlay of Run Bb - Run A, effect of field surveyed road cut depths 0.06 0.04 0.02 H 0.00 -0.02 -0.04 -0.06 -0.08 -0.10 39.c. Overlay of Run Be - Run A, effect of maximum road cut depths equal to soil depth -0.12 0.50 h 0.45 0.40 0.35 0.30 0.25 0.20 0.15 0.10 0.05 0.00 = c r Cr • Run Ba - Run A Run Bb - Run A • Run Be - Run A •Run A Figure 39. Discharge differences for South Tributary due to the varying road cut depths of Runs Ba, Bb, and Be during water year 1997 80 CA TJ 03 O i-TJ CU a o T J a 03 £> 03 T J a 03 03 CA a o "CA a CU X cu E 03 CU I-E o a CA CU U. CA CU DX) a 03 JS o -9 a a 03 U CU cu u. u CA tt T J cu tf r»5 03 H CD 03 > -< CD 3° 3 U 2 £ "3 3 CD V * c 3 tf \ ° \ ° \ ° o x o x o x H t N (N O O O \0 c N o x 0S- T - H T - H 5 9 9 ' V V t~- ON O O N oo oo N O O N O N O N oo CO uo CN O N oo r» NO NO NO NO NO NO NO NO C N C N ( N t N C N t N C O o d d 00 N O O N ro ro N O co O N O N — O N O N O N O N O ro ro ro ro \J3 O ON d d O i i ' V V o r-r~- O N CN) O ro N O N O T f —I T f r-~ co ro >o T f ^ CO CO " N O NO NO NO ro ro ro ro C N C N C N C N N? N= N° o o c r -CN CN fN o d d ON 00 NO NO ro C O O N O N T f T f NO >0 NO O O O 00 O O O O ro CN CN CN CN CN CN CN CN CN CN CN NP V O X O ^ ON ON 9 d d v v v CN oo O N oo oo ON co ON T f < § U CA T J CS o u T J cu a o T J ca cs JS es T J a es es *^  TJ a "CA a <U X cu E cs CU s-CA E o DX) a s CA CU i-CA CU DX) a es JS u o C es a> a 73 3 a ca cs CU CU u u CA a T J CU tf cs H M 03 CD PH ^ "3 ^ 3 S < » - H fl> S ^ Ml 3 03 0 H CD 03 1*6 •a ID PH 03 1 £: NO NO " 0S o N d O N \0 \0 T—< d ^ ^ V ° ° oo T f ro oo O H CO N f~- f N - O O O ro ro ro T f NO NO NO NO SN N? N? 5£ ^ °N ^\ ro T f Q ^ ^ T f r— r- N O N O N O O N IO O N f N - oo O N O o T - ~< K K C N ! f - : N? N= V= ON ON ON C<1 h H d d ^H' CN N O ro O N O N O C O O O CN *o r-*- O N ro ro ro ro NO NO NO NO cN NP NP T - H ON ON © CO CO i © d V oo CN O N CN t— NO ON NO NO 00 00 1 *—» *—< * - H T f T f T f T f ' >N NP v? £ ON ON-C N | _ 9 d d 00 NO f N - NO CO o T f O O •O T f NO NO CN CN CN CN ON O N ON ON ro CN ro ON ON O ro CD J3 CD oo < £ 3 o PH ° < X B ° Pi 55 3§ tf CO - -§ II & CD P § cfcl ft <u ° S3 PH •a > CD ro PH O N — O N 03 .-H 3 >-g 03 S CD 81 40.a. Run Bb - Run A , effect of field surveyed road cut depths -0.10 [ 0.15 i 40.b. Overlay of Run C - Run A , effect of stream density expansion 0.10 0.05 H 0.00 -0.05 -0.10 40.c. Overlay of Run D - Run A , effect of abandoned roads and skid trails -0.35 H i h r l ill 1 1 '1 16 14 12 10 ax-im a JS u M 5 M V g £ r-l Of a-o f*s cr-o O N o Run Bb - Run A • Run C - Run A • Run D - Run A Run A Figure 40. Discharge differences for Redfish Creek due to the varying simulation scenarios of Runs Bb, C, and D during water year 1993 82 0.15 0.10 0.05 0.00 -0.05 -0.10 0.15 0.10 0.05 0.00 -0.05 -0.10 41 .a. Run Bb - Run A, effect of field surveyed road cut depths 41.b. Overlay of Run C - Run A, effect of stream density expansion j P s l 1 | - \\ 41.c. Overlay of Run D - Run A, effect of abandoned roads and skid trails • Run Bb - Run A Run C - Run A • Run D - Run A Run A Figure 41. Discharge differences for Redfish Creek due to the varying simulation scenarios of Runs Bb, C, and D during water year 1994 83 42.a. Run Bb - Run A, effect of field surveyed road cut depths 42 .b. Overlay of Run C - Run A, effect of stream density expansion 42.c. Overlay of Run D - Run A, effect of abandoned roads and skid trails Run Bb - Run A Run C - Run A Run D - Run A Run A Figure 42. Discharge differences for Redfish Creek due to the varying simulation scenarios of Runs Bb, C, and D during water year 1996 84 43.a. Run Bb - Run A, effect of field surveyed road cut depths 43.b. Overlay of Run C - Run A , effect of stream density expansion 0.25 0.20 0.15 0.10 0.05 0.00 -0.05 -0.10 -0.25 -0.30 -0.45 43.c. Overlay of Run D - Run A, effect of abandoned roads and skid trails J 14 12 10 6 £ 0 -3 at Run Bb - Run A • Run C - Run A • Run D - Run A Run A Figure 43. Discharge differences for Redfish Creek due to the varying simulation scenarios of Runs Bb, C, and D during water year 1997 85 vi « O 1_ TS <u S o a cs TS C GS 2 a o CS u B o OX) a 1/5 CU s. cn CU OX) a CM cu 2 "3 *£> 13 s a a as > > i -CS 1-CU a c TS es H co 6 0 ca CL> > 2 •rH <Jj 3 3 o a" u OS OS -•9 3 so OS OS OS oo SO r- - s f OS I f SO ro CO oo Os CN OS SO in OS r-~ OS r- SO SO SO Os Os T OS OS 3 2 S co 2 3 60| u 43 o 3 2 S co a* 43 u r o os Os s s S = s O s O o o 6 s in CN - s f co - s f ro 5^ o s O S O S CO O S t-s; ro ro c s i ro >n O -sf — o OS o O O 00 -si" "sf s P s P s O O x o x o x O O SO SO ro - s f ro \ ? N = S = o N o x 6s-CN 00 CN co ro ro OS SD m o ro T—t OS CN m 00 so CN "sf o CN o SO CN f-H CN ro CO CO CO •"* ' 1 x ? s = s = O O o N t~- ro O ro - s f -st in o r- - s t SO SD CN - s f oo t s f m - s f 00 O "sf m © o o CN CN CN CN o CO 00 p ro • ro i ro i OS CN m r—< ro in ro CN SO O OS ro ro ro O O O »—i < £ U Q • § - a TS CU a o TS a 05 -O es TS a CO 53 Vi a 'vi a CU •w X cu £ CS cu B © OX) a cu U. Vi cu OX) a cs JS cu it © CB a "3 a a es >-> t-cs •w S •a "C H s-cu a a P ts TS NO CU </> ^ TS C3 H cs o CJ CJ CJ > < co 6 0 a ca a 3 43 M cd co CM 3 oo m SO CN 00 t— oo r- 00 oo ro ro ro ro O d d d i—i CO 2 M 6 0 | §3 ca fi co ca < U ro CN >n ro 00 m r-s oo SO oo OS - s f • s f ^ f "sf d d d d 44 C3 ca 3 4<J ca If1 co ca OH 43 u 4<S ca cj PH 1? 3 S C S Pi S O s p s p O S 0 N o s H n N c i H 6 -s? . o C > O N 5s CN CO $1 T - Q ^ H > ? S p s O °_; o~- o s 1 t o " d m s p s P s P © s o s O S r-- ro ro •sf CN Csi so - s f r- r f CN CN CN CN CO ^H CN CN - s f r f T f T f o S d d O S O O S t--5^ O S CO ^H »""H 1^ Os "sf r f CN SO SO t~s SO SO SO CN CN CN CN d d d d s O O S s P O S ^s5 o--ro oq *—< ro CN OS oo in SO CN ro so SO oo SO so in in in in d d d d < £ O Q ro CN ro Os Os Os 86 0.02 -1 0.01 -0.00 --0.01 --0.02 --0.03 -0.02 -i to 0.01 -'ereni 0.00 -s 5 -0.01 -Mb n .c -0.02 -5 -0.03 -44.a. Run Bb - Run A, effect of field surveyed road cut depths 44.b. Overlay of Run C - Run A, effect of stream density expansion 44.c. Overlay of Run D - Run A, effect of abandoned roads and skid trails I - A •—1 1 ' ' |—M] t hi m — — o Run B b - Run A • Run C - Run A • Run D - Run A o O Run A c rn o Figure 44. Discharge differences for Redfish Upper Tributary due to the varying simulation scenarios of Runs Bb, C , and D during water year 1993 87 0.02 45.a. Run Bb - Run A , effect of field surveyed road cut depths 45.b. Overlay of Run C - Run A , effect of stream density expansion 45.c. Overlay of Run D - Run A , effect of abandoned roads and skid trails Run Bb - Run A • Run C - Run A • Run D - Run A Run A Figure 45. Discharge differences for Redfish Upper Tributary due to the varying simulation scenarios of Runs Bb, C , and D during water year 1994 88 46.a. Run Bb - Run A , effect of field surveyed road cut depths 0.02 -i 0.01 -0.00 • -0.01 --0.02 --0.03 -0.04 -0.02 -1 0.01 -E 0.00 -£ Difl -0.01 -irge -0.02 -ischi -0.03 -a -0.04 -46.c. Overlay of Run D - Run A , effect of abandoned roads and skid trails 46.b. Overlay of Run C - Run A , effect of stream density expansion J I 1 ' - i • 1 — 1 a-. e> 3 -O 3 fN S P • Run Bb - Run A Run C - Run A • Run D - Run A Run A Figure 46. Discharge differences for Redfish Upper Tributary due to the varying simulation scenarios of Runs Bb, C, and D during water year 1996 89 0.02 -> 0.01 -0.00 --U.U1 -0.02 --0.03 --0.04 -M 0.02 -] E 0.01 -(U u B 0 00 V £ c -0.01 5 u -0.02 -JZ -0.03 -n 5 -0.04 -0.02 -i 0.01 -0.00--0.01 -0.02 --0.03 --0.04 --0.05 --0.06 --0.07 --0.08 --0.09 --0.10 --0.11 -0.12 -47.a. Run Bb - Run A, effect of field surveyed road cut depths 47.b. Overlay of Run C - Run A, effect of stream density expansion 47.c. Overlay of Run D - Run A , effect of abandoned roads and skid trails P •a z o Run Bb - Run A • Run C - Run A • Run D - Run A Run A Figure 47. Discharge differences for Redfish Upper Tr ibutary due to the varying simulation scenarios of Runs Bb, C , and D during water year 1997 90 T J eS O s-T3 cu C o T J a es .a cs T J a es es u T J on •N a .© a eo es cu h a e o CJD a ce <U L. ee CU DD a cs cu 2 • mm a a es >. u cs H s o <V3 -a cs H CO CO „ _ % co ca 3 > 3 ^ co .= U 2 T3 3 13 a01 C3 43 u 3 -—-3 i— B CO CO S3 ss § 2 cd S 3 2 a CO 2 13 If1 U s 3 05 S O O O o s od p--N ° X ? N P ©S 0 % O S p vq os OS I—' i n oo 2 o oo m co © 2; w OS OS o oo o O O S r-CN H^ - H oo so so oo m ro SO SO o r -o s oo s o CN CN CN t-; H m oo r-^ s o so o r-- oo CN m r f © r f SO OS CO r f OS f - CN CN m r f r f t> 00 00 00 d od CN T-I — SO © © r f ro ro r f r - OS oo r f ro —i 00 Os ro CN T—I SO 00 00 oo s p s o s o © © s O x r f Os oq od sb i n OS oo r - T—l SO ro r f © OS r f m r f r f SO OS © T—l © OS SO C~s SO g U Q in 60 CO cd S <A "c3 so T J es © u TJ CU a o T J a es X) es T J a es es m •N a JO an a cu X <u S es cu u B o OJD a cu u c/1 CU OX a es w o C es « a 9 a a es Z? es -O "C H o 00 es H CO , Ml cd Cj co cd n PH 43 •< u 44 ca co PH 13 3 2 44 j&l §3 ca g co ca ,^ PH 43 < u 1 44 M| a3 ca C co ca , p , J s 3 S? S? N ? . °° CN' - 2 s P s P s P O S o x O x CO OS CN iri r f ro oo ro ro © © os oo r f m oo oo oo CN CN CN CN © © © © SP SP SO o o o S O © i n T^I CN ©' r s © r f so © so r- T-T Os ro ro ro ro r f r f r f © © © © s p s u s p O S C X O IT) r f r f i n in r f © r f © ^H ( s r f r f -H ^H ro ro ro ro ro ro ro © © © © s p s p s o OS o s o x m i-i © CN i-H T-1 m m s o s o CN CN ro s o s o s o CN CN CN CN © © © © S? ^ °° So ~ 2 NO © ro © m i n i-H O N s o - H i n r f ro ro ro ro ro © © © © ro CN ro OS OS ••3 s & =: 91 0.04 48.a. Run Bb - Run A, effect of field surveyed road cut depths 48.b. Overlay of Run C - Run A, effect of stream density expansion 0.04 48.c. Overlay of Run D - Run A, effect of abandoned roads and skid trails CN Run Bb - Run A • Run C - Run A • Run D - Run A Run A Figure 48. Discharge differences for South Tributary due to the varying simulation scenarios of Runs Bb, C, and D during 1993 92 0.04 49.a. Run Bb - Run A , effect of field surveyed road cut depths 49.b. Overlay of Run C - Run A , effect of stream density expansion 0.00 0.04 -» 49.c. Overlay of Run D - Run A , effect of abandoned roads and skid trails -o.oi H -0.02 i -0.06 ex m m 3 3 O c o 5 • Run Bb - Run A • Run C - Run A Run D - Run A Run A Figure 49. Discharge differences for South Tributary due to the varying simulation scenarios of Runs Bb, C, and D during 1994 93 0.06 0.04 { 0.02 0.00 0.06 5 a, o.o2 Im 03 •C u ag 5 o.oo 0.06 7 0.04 0.02 0.00 -0.02 -0.04 -0.06 -0.08 H -0.10 50.a. Run Bb - Run A, effect of field surveyed road cut depths 50.b. Overlay of Run C - Run A , effect of stream density expansion 50.c. Overlay of Run D - Run A , effect of abandoned roads and skid trails 0 0.7 0.6 0.5 0.4 0.3 © 0.2 0.1 2 3 a s rn 5 5 OS Run Bb - Run A • Run C - Run A • Run D - Run A Run A Figure 50. Discharge differences for South Tributary due to the varying simulation scenarios of Runs Bb, C, and D during 1996 94 Run Bb - Run A Run C - Run A Run D - Run A Run A Figure 51. Discharge differences for South Tributary due to the varying simulation scenarios of Runs Bb, C, and D during 1997 95 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

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"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0075093/manifest

Comment

Related Items