Variability of Intermittent Headwater Streams in a Boreal Landscape Examensarbete vid Institutionen för geovetenskaper ISSN 1650-6553 Nr 248 Influence of different discharge conditions Tum Nihm Dynamic expansions and contractions of stream networks can play an important role for hydrologic processes as they can connect different parts of the landscape to the stream channels. However, we know little about the temporal and spatial variations of stream networks during different flow and wetness conditions. This study focuses on the contraction and expansion of stream networks during different flow conditions in the boreal Krycklan catchment, located in Northern Sweden. The stream network and initiation points were extracted from a gridded digital elevation model (DEM) of 5-meter resolution, and then compared with the stream network initiation points (heads) observed during the spring flood (freshet) period in 2012. From the results of the study, it was clearly seen that the observed stream heads and the stream heads appearing in the stream network map extracted from DEM did not agree very well. 49% of the total observed stream heads (49) fell onto the low order stream branches and headwater streams derived from the DEM. Only few of them exactly matched the modeled stream heads. Moreover, the modeled stream network was much denser than the observed stream network, and so the simple raster based dynamic model developed could not well represent the dynamic stream network extension in the real system. Most headwater streams in the study catchment were man-made ditches, which were dug to drain water wetlands and to increase forest productivity. The majority of observed stream heads were formed by seepage from the saturated surrounding soils, while only a few of them were formed by saturation overland flow. On the other hand, the dynamic stream network derived from the DEM suggested that the number of streams of lower order and their lengths was sensitive to change in streamflow, especially during the high flow episode. Uppsala universitet, Institutionen för geovetenskaper Examensarbete E1, Hydrologi, 30 hp ISSN 1650-6553 Nr 248 Tryckt hos Institutionen för geovetenskaper, Geotryckeriet, Uppsala universitet, Uppsala, 2012. Variability of Intermittent Headwater Streams in a Boreal Landscape Influence of different discharge conditions Tum Nihm Examensarbete vid Institutionen för geovetenskaper ISSN 1650-6553 Nr 248 Variability of Intermittent Headwater Streams in a Boreal Landscape Influence of different discharge conditions Tum Nihm Copyright © Tum Nihm and the Department of Earth Sciences Uppsala University Published at Department of Earth Sciences, Geotryckeriet Uppsala University, Uppsala, 2012 Abstract Variability of intermittent headwater streams in a boreal landscape: Influence of different discharge conditions Tum Nhim Department of Earth Sciences, Uppsala University Villavägen 16, SE-752 36 Uppsala, Sweden. Dynamic expansions and contractions of stream networks can play an important role for hydrologic processes as they can connect different parts of the landscape to the stream channels. However, we know little about the temporal and spatial variations of stream networks during different flow and wetness conditions. This study focuses on the contraction and expansion of stream networks during different flow conditions in the boreal Krycklan catchment, located in Northern Sweden. The stream network and initiation points were extracted from a gridded digital elevation model (DEM) of 5-meter resolution, and then compared with the stream network initiation points (heads) observed during the spring flood (freshet) period in 2012. From the results of the study, it was clearly seen that the observed stream heads and the stream heads appearing in the stream network map extracted from DEM did not agree very well. 49% of the total observed stream heads (49) fell onto the low order stream branches and headwater streams derived from the DEM. Only few of them exactly matched the modeled stream heads. Moreover, the modeled stream network was much denser than the observed stream network, and so the simple raster based dynamic model developed could not well represent the dynamic stream network extension in the real system. Most headwater streams in the study catchment were manmade ditches, which were dug to drain water wetlands and to increase forest productivity. The majority of observed stream heads were formed by seepage from the saturated surrounding soils, while only a few of them were formed by saturation overland flow. On the other hand, the dynamic stream network derived from the DEM suggested that the number of streams of lower order and their lengths was sensitive to change in streamflow, especially during the high flow episode. Keywords Dynamic stream network, intermittent streams, headwater stream, stream heads, wetness conditions, streamflow i Referat Variabilitet av periodiskt återkommande bäckar i ett borealt landskap: Betydelse av olika avrinningsnivåer Tum Nhim Institutionen för geovetenskaper, Uppsala universitet Villavägen 16, 752 36 UPPSALA Utbredningsdynamiken av bäckars flödesnätverk kan spela en viktig roll för hydrologiska processer eftersom de kan ansluta olika delar av landskapet till bäckar och älvar. Vi vet dock väldigt lite om de temporala och spatiala variationerna i utbredningunder olika flödes- och fuktförhållanden. Denna studie fokuserade på utbredningsdynamiken av flödesnätverk under olika flödesförhållanden i Krycklans avrinningsområde, beläget ca 50 km nordväst om Umeå i Västerbotten. Bäckarnas nätverk- och initieringspunkter identifierades från en digital höjdmodell (DHM) med 5-meters upplösning och jämfördes sedan med samma bäckars initieringspunkter (källor) observerade i fält under vårfloden 2012. Från resultaten av studien var det tydligt att de observerade och simulerade initierings punkterna inte stämde bra överens med varandra. 49% av de totalt observerade initieringspunkterna (49) utgjordes av bäckar med låg strömordning och på källsflöden hämntat från DHMet. Endast ett fåtal av dem överenstämde exakt med de simulerade initieringspunkterna Dessutom var det simulerade flödesnätverket mycket tätare än det observerade, och den enkla raster-baserada dynamiska modellen representerde dåligt det dynamiska flödesnätverket som observerades i fält. De flesta källflöden i avrinningsområdet är diken som grävts för att dränera våtmarker och för att öka skogsbrukets produktivitet. Majoriteten av de observerade initieringspunkter har bildats av läckage från den omgivande mättade marken, medan endast ett fåtal av dem har bildats av ytavrinning på mättad mark. Å andra sidan visade det simulerade dynamiska flödesnätverket att antalet bäckar av lägre ordning och deras utbredningslängd var känsliga för förändringar i flödesförhållanden, speciellt under högflödesepisoder. Sökord Dynamiska flödesnätverk, periodiskt återkommande bäckar, källflöden, källor, vätaförhållanden, avrinning, flöde ii Abbreviations DIC DOC DEM GPS LDD m.a.s.l Dissolve Inorganic Carbon Dissolved Organic Carbon Digital Elevation Model Global Positioning System Local Drain Direction Meter Above Sea Level iLL Table of contents 1. Introduction ................................................................................................................................. 1 1.1. Research objectives .......................................................................................................... 1 1.2. Key research questions ..................................................................................................... 1 2. Background.............................................................................................................................. 2 3. Materials and Methods ................................................................................................................ 4 3.1. Study location ...................................................................................................................... 4 3.2. Derivation of terrain indices and flow directions ................................................................ 6 3.3. Development of a dynamic stream network model ............................................................. 7 3.4. Field investigation for mapping stream heads ................................................................... 10 3.5. Testing of model parameters’ value ................................................................................... 12 4. Results ....................................................................................................................................... 12 4.1. Stream network and frequency of surface flow generating ............................................... 13 4.2. Stream network and geology of the study catchment ........................................................ 17 4.3 Modeled stream length and stream orders during different flow conditions ...................... 19 4.4. Field observations .............................................................................................................. 23 4.5. Comparison of observed and predicted stream heads ........................................................ 25 4.6. Slope-area relationship........................................................................................................... 34 5. Discussions ............................................................................................................................... 35 5.1. Comparison of observed and simulated stream heads ....................................................... 35 5.2. Variation of stream length and stream orders at various flow conditions ......................... 35 5.3. Model uncertainties ............................................................................................................ 36 6. Concluding remarks .................................................................................................................. 36 7. Suggestions for future work ...................................................................................................... 37 8. Acknowledgements ................................................................................................................... 38 References ..................................................................................................................................... 39 Appendix 1: Field protocol ............................................................................................................. a Appendix 2: Dynamic stream network model’s scripts .................................................................. c Appendix 3: Coordinates and descriptions of observed stream heads ............................................ h Appendix 4: Field data records ........................................................................................................ j List of tables and figures List of figures Figure 1: DEM and geology class map of the Krycklan catchment ............................................... 4 Figure 2: Krycklan catchment and flow gaging sites...................................................................... 5 Figure 3: LDD code for direction of runoff .................................................................................... 6 Figure 4: Average daily flow series at site 7................................................................................... 9 Figure 5: Investigated subcatchments of site 9, 7, and 1 .............................................................. 11 Figure 6: >90% frequency map of surface flow generating ......................................................... 13 Figure 7: >90% frequency map of connecting surface flow ......................................................... 14 Figure 8: Stream network map at maximum streamflow.............................................................. 15 Figure 9: Stream network map at minimum streamflow .............................................................. 16 Figure 10: >90% frequency map and geology of site 9 ................................................................ 17 Figure 11: >90% frequency map and geology of site 1 ................................................................ 18 Figure 12: > 90% frequency map and geology of site 7 ............................................................... 19 Figure 13: Relationship between total stream length and flow..................................................... 20 Figure 14: Relation between total stream length and corresponding streamflow ......................... 21 Figure 15: Total stream length versus different stream orders ..................................................... 21 Figure 16: Total stream length versus different stream orders ..................................................... 22 Figure 17: Stream head with seepage erosion............................................................................... 24 Figure 18: Stream head with saturated overland flow .................................................................. 24 Figure 19: Stream head with seepage from saturation .................................................................. 25 Figure 20: Observed stream heads within subcatchment 7 ........................................................... 26 Figure 21: Observed stream heads in catchment of site 9 ............................................................ 27 Figure 22: Observed stream head in catchment of site 1 .............................................................. 29 Figure 23: Observed stream heads in other subcatchments .......................................................... 30 Figure 24: Observed headwater streams ....................................................................................... 32 Figure 25: Relationship of local slope/source area of stream head for man-made ditches........... 34 Figure 26: Relationship of local slope/source area of stream head for natural stream ................. 34 List of tables Table 1: Flow gaging sites in Krycklan catchment......................................................................... 6 Table 2: Statistical characteristics of the flow series .................................................................... 10 Table 3: Model calibration for b value ......................................................................................... 12 Table 4: Model calibration for Zmin value ..................................................................................... 12 Table 5: Modeled total stream length during highest and lowest flow conditions ....................... 20 Table 6: Observed stream heads ................................................................................................... 23 Table 7: Formation of stream heads.............................................................................................. 23 Table 8: Number of surveyed stream heads falling within subcatchment of site 7 ...................... 26 Table 9: Number of stream heads falling within catchment of site 9 ........................................... 28 Table 10: Number of stream heads falling within catchment of site 1 ......................................... 30 Table 11: Number of stream heads falling within other subcatchments ....................................... 31 Table 12: Characteristics of headwater streams............................................................................ 33 1.Introduction Stream heads are important transition points since, similar to riparian zones, they are at the interface between hillslopes and stream channels (Montgomery & Dietrich, 1989). Connecting the stream heads to the downslope stream networks, headwater streams maintain many hydrological and biogeochemical processes that influence, for example, stream chemistry, flow and wetness conditions in the catchment. More importantly, these controlling processes may vary according to topographic conditions, geological characteristics and different landuse in a basin. One challenging aspect in the study of where the stream initiates is that little is known about the dynamic temporal and spatial variations of the stream networks since they often occur at relatively small scales, and in mapped parts of the stream network that are seldom mapped. There might be large differences in the number of active zero or first order streams, with varying flow conditions. For example, during high flow episode, the stream network might considerably expand resulting in many active low-order streams, which in turn would be likely to increase transport of sediments, organic matter, or other stream solutes. At low flow conditions, on the other hand, the stream network would contract and, as a result, the connectivity between landscape and surface waters would decrease. Many studies use DEMs to extract and examine the stream networks such as O'Callaghan & Mark (1984), Mackay & Band (1998), McMaster (2002), and Hancock & Evans (2006). However, because of different resolutions of DEMs and variable topography and landscape, the stream initiation points may be inaccurately mapped. In this study, to better understand the controlling factors in headwater streams and dynamic extension of the stream networks, DEMderived stream heads were compared to stream heads that were mapped during field visits. The relationship between local slope and source area of the stream heads was also discussed regarding different formations and geology to which those stream heads belong. 1.1. Researchobjectives The spatial extent of stream networks can play an important role in controlling the stream chemistry as it connects different parts of the landscape to the streams. However, relatively little is known about temporal variability of the extent of the stream network during different flow and wetness conditions. The aim of this study was to determine to what extent the stream network expands at different flow conditions and to investigate if the variable extend could be reproduced using a raster based stream network model. To reach this aim, several steps were accomplished: o Establishing a dynamic stream network model based on flow-groundwater table relationship, using gridded digital elevation data of 5 meter resolution; o Mapping stream initiation points from the stream network extracted from the digital elevation data; o Determining criteria for selecting stream heads; o Mapping stream head locations in the terrain during a field survey 1.2. Keyresearchquestions In this thesis, some key research questions were addressed upon completing the stated objectives: o Would model simulation of the stream network agree with the observed stream network in the real terrain? 1 o To what extent, would the stream network expand at different flow conditions? o Could the variable extent of stream network be reproduced using a simple raster based stream network model? 2. Background Many studies demonstrate the importance of study on the stream network expansion and the locations where the streams begin. Lyon et al. (2004) mention that knowing the locations of source areas is useful for understanding non-point source pollution. O'Callaghan & Mark (1984) and Hancock & Evans (2006) show the importance of stream network extraction from digital elevation data on the study of geomorphology and hydrology. Rayburg et al. (2005) draws the attention to the impact of landuse change on the channel network evolution. Their study further concludes that better understanding about the evolution of the water delivery networks would make it possible to make better predictions of the impact of landuse change on the hydrologic functioning. Montgomery & Dietrich (1988) define channel initiation points as the points that are located closest to the upslope drainage divide with the presence of channelized morphology. Montgomery & Dietrich (1989) also suggest that channel heads can be gradual or abrupt and that the channel reach may touch the channel network immediately downslope or discontinuously expand. Connecting the stream head locations to the downslope stream channels, headwater streams maintain many inducing processes that control, for example, stream chemistry and flow conditions in the catchment. Jaeger et al. (2007) presents physical characteristics of the headwater streams. A headwater stream, for instance, can be continuous or discontinuous. Usually, it has convergent topography and can be found at the foot of a valley. In addition, to be considered as a headwater stream, it should have channel length of at least 5 meters. To get more knowledge on the governing processes in the headwater streams, it is interesting to understand the relationship between source area and slope. Some studies discuss the source area/slope relations to better understand the inducing processes in the stream head initiation points on downslope channel network. For instance Montgomery & Dietrich (1988) suggest that, to know where the stream channels begin, the relationship of source area/slope should be tested and that the channel head source areas could be selected based on field accessibility, ranges of slope values, and climatic and geological conditions. In a more recent study, Jaeger et al. (2007) investigated the source area/slope relationships of headwater channels in sandstone and basalt lithology of forested landscape by using the source area of channels, delineated in digital elevation model (DEM) and comparing DEM derived heads to heads that were mapped in the field using Global Positioning Systems (GPS). The results of the study showed, however, that there was a poor relationship between the source area and slope since the source areas vary lithologically and that the source area derived from DEM and mapped in the field did not agree well. Another important factor controlling the headwater stream is the formation of the stream heads. To understand more about the sources of the stream heads, field observations are very valuable. Channel heads can be formed by several processes such as overland flow (saturation excess or return flow) or subsurface flow seepage or landslide failure (Jaeger et al., 2007). In their study on channel initiation, Montgomery & Dietrich (1988) collected 71 points of stream heads in forested landscape underlain by folded Paleocene and Eocene sandstones. The results of the study show that, of all the 71 head locations, the formation of the steep-slope stream heads was either through subsurface flow seepage or through landslide failure, while that of gentle-slope channel heads was overland flow. The study done by Montgomery & Dietrich (1988) further 2 shows that the initiation points of the channel heads on steep gradient may be formed by subsurface flow erosion. The field investigation also depicts that the initiation point of abrupt channel heads on low slope is formed by seepage erosion, while that of gradual channel head is controlled by overland flow saturation Many studies rely on DEMs to extract flow directions of the stream networks. There are, however, some uncertainties due to either low DEM resolution (McMaster, 2002) or difficulties in deriving flow directions in gentle slope areas (Mackay & Band, 1998). Unfortunately, there is general a lack of field data on channel initiation locations and the processes that form and control those transition points (Montgomery & Dietrich, 1988). 3 3.Matterialsan ndMetho ods 3.1.Sttudylocattion The ressearch was conducted d in the borreal Kryckllan study catchment, c situated ab bout 50 km m northweest of Umeåå, northern Sweden, w with a catchm ment area of 67 km2. T The catchmeent geologyy (Figure 1) is pooorly weatheered gneissiic bedrock covered with w sedimeent depositts at lowerr elevatioon and moraine (glaciaal till) at higgher elevattion. The caatchment haas a gentle topographyy with eleevations rannging from m 126 to 3772 m above sea level (m.a.s.l). ( Thhe catchmeent receivess mean annnual preciipitation of 623 mm, aand has the mean annu ual air tempperature of 1.8 oC. Thee catchmeent vegetattion cover is dominatted by foreests, coveriing 88% oof the catch hment areaa followeed by 8% off wetland, 3% of agricuultural land, and 1% of lake. (Grabbs, 2009) F Figure 1: DE EM and geoloogy class maap of the Kry ycklan catchm ment 4 Figure 2: Krycklann catchmentt and flow gaging g sites Flow gaaging sites in Kryckla an catchmeent There w were primarrily 16 flow gaging sitees (Figure 2)) in Krycklaan catchmennt, but later site 11 andd 8 were no longer considered c as a flow ggaging sites.. Site 16 is the main ouutlet of the catchment,, whereass site 7 is thhe site whosse dischargee series from m April 17,, 2007 to D December 31 1, 2008 wass used in the model. Subcatchm ments of gagging site 1, 7, and 9 arre interestinng study areeas for fieldd investiggation sincee they repressent differennt types of catchment c geology. g 5 Table 1 below show ws details about a each ssubcatchmen nt of the stu udy catchmeent. Table 1: Flow gagging sites in n Krycklan catchment SIITENO 1 2 3 4 5 6 7 8 9 10 12 13 14 15 16 SITE NAME RB SVV LMB SVE STO STB SVW FB NÄB SMB NMB LB ÅB ÖKL KL X_COORD DINATE 11694040 11692550 11692840 11692310 11691640 11692330 11692560 11692140 11693320 11692940 11694430 11693140 11692410 11695500 11697360 Y_ _COORDINA ATE CATCH HMENT AREA A (m2) 71330 030 438617 4 71332 270 126070 1 71338 830 32 2814.5 71341 150 172171 1 71342 260 643354 6 71332 200 10 084540 71332 260 465478 4 71327 790 21 193080 71317 770 28 879970 71339 910 33 315120 71322 200 53 396080 71312 220 69 900410 71304 430 137 759600 71330 080 172 285700 71276 650 649 914300 3.2.De erivationofterrain nindices andflowdirection ns D8 algoorithm by O’Callaghan O n & Mark (11984) was used u to calcu ulate flow ddirections in n PCRaster.. To com mpute the floow direction ns in PCRasster, Deurseen (1995) ussed the folloowing proceedure: Creatioon of local drain d direcction (LDD D) and pits removing r The conncept of loccal drain dirrection is baased on an analysis a of each cell inn a DEM. The T analysiss of eachh cell in thee DEM is performed p inn a 3x3 cellls window.. The slopee of a center cell to itss eight neeighboring grid g cells iss first determ mined. The material (ffor examplee, surface water) wouldd flow froom that cennter cell to th he steepest downward cell as show wn in figuree 3. This flo ow directionn is then defined as the local drrain directioon. If the lo ocal drain direction d off that cell iss found, forr instancee, the 3x3 cells window w is moved to the next cell. This reefers to a sim mple case. Fig gure 3: LDD D code for direction d of runoff d, this mean ns the centeer cell can have eitherr In contrrast, if the local drain direction iis not found equal ellevation (thhe cell is parrt of a flat) or lower eleevation (thee cell is partt of a pit) th han its eightt neighbooring cells. In this casse, special LDD codee will be asssigned andd processed d further too computte the locall drain direection. The solution fo or finding the local ddrain directiion, in thiss 6 regard, is to start with identifying two types of flat areas: flat area of type 1 and flat area of type 2. Flat areas of type 1 are defined as a set of neighboring cells of the same elevation, and at least one of those cells has lower elevation than the core cell. The local drain direction for this case can be resolved by assuming that is like a simple case, which means the water will flow from the cell under consideration to the cell of lower elevation. Flat areas of type 2 are defined as a set of neighboring cells of the same elevations, and all of these cells have exactly the same elevation as the cell under consideration (the core cell). To resolve this, each neighboring cell of the core cell is assigned a drain direction. As mentioned earlier, pits are those cells whose neighbors point towards them. Pit removing process is to assign artificial local drain direction to those depressed cells. This process is performed because there might be problems with digitizing or discretization of digital elevation data, which is not natural phenomenon. Calculation of upslope contribution area based on D8 algorithm Upslope contributing area was computed based on local drain direction derived from D8 algorithm. The calculation assumed that, for each cell, accumulated amount of material would flow out from itself to neighboring downstream cell following the local drain direction. The accumulated material is the amount of material in a cell under consideration plus the amount of material flowing from the upstream cells of the cell under consideration. Based on local drain direction network, the catchment for an outflowing cell is determined, which contains the cell itself and the upstream cells that drain into it. To get the value of upslope contributing area for each outflowing cell, the total cell area was considered. Calculation of terrain slope based on DEM Terrain slope is the ratio of vertical distance to horizontal distance (dz/dx), whose value ranges between 0 and 1, and can be turned into percentage by multiplying that ratio with 100. For each cell, computation of slope is performed in a 3x3 cell window, based on an analysis of DEM. There might be cases that the elevation of neighboring cells of the core cell is unknown. If this is the case, for each missing value cell, the elevation of the core cell is assigned with average elevation of non-missing value cells in the 3x3 cell window. 3.3.Developmentofadynamicstreamnetworkmodel Flow-Groundwater table relationship The dynamic stream network model examined the expansion and contraction of stream network during different streamflow conditions. The basic principle for model development was based on the combination of landscape analysis and field measurements of streamflow and riparian groundwater tables. The model assumes that groundwater flow can be scaled with specific contributing area and that hydraulic gradient can be estimated by terrain slope tan . Thus, total groundwater discharge at a riparian location of homogeneous soil can be related to groundwater table position based on the Darcy’s law as following (Grabs, 2010). . tan . [1] Where, : Total groundwater discharge [l/s] 7 : Width [m] : Terrain slope : Soil depth [m] : Groundwater table position [m] K(z) : Hydraulic conductivity at depth Z [l.s-1m-1] tan If the riparian soil and hydraulic conductivity vary exponentially with depth, then the equation becomes . tan . ′ . . [2] Furthermore, the total groundwater discharge can be related to streamflow measured at the catchment outlet as following: . . [3] Where : Specific hillslope contributing area derived from topography [m2/m] A : Catchment area [m2] By solving the last two equations, [2] and [3], we get ln . . , . [4] Where, o b o , o : Transmissivity shape parameter [m-1] : Streamflow measured at the catchment outlet [l/s] : Transmissivity parameter [l.s-1.m-1] The dynamic model for estimating groundwater table distribution in the whole catchment was then based on equation [4]. For the whole study period, based on the available discharge series from April 17, 2007 to December 31, 2008, the frequency of surface flow occurrence for each cell in the stream networks can be calculated by taking up the number of days that the streams may generate flow divided by the total number of observations (total time steps in the model). To increase the chance of finding the actual locations where the stream starts, a map with at least 90% chance of flow surface occurrence was created and overlaid with geology class map, and used as a referencing source for locating the stream heads. Hence, it is practical to create a frequency map of surface flow occurrence in the stream to facilitate the finding of stream head locations in the field. The frequency was calculated by assuming that if the groundwater table was at least 0.1m above the ground surface, there would be surface flow in the stream channels. Days for which the groundwater table was above 0.1m was summed up for each grid cell and divided by the total number of days The model generated the stream networks based on local drain directions, and so the stream network that went across the lake and wetland in the extracted map did not reflect reality. Also the model did not perform very well at the zone of gentle-slope terrain because groundwater level distribution in the catchment was interpolated in the sense that its distribution is the same way as the terrain topography does. Thus, the stream networks that come across lakes and wetlands in the catchment were not considered. 8 Determination of model parameters There are three main parameters in the model: a transmissivity shape parameter ( , a transmissivity parameter ( ′ ), and threshold groundwater table (Zmin) parameter, i.e., the minimum groundwater table above which surface flow occurring in the stream channels. The transmissivity parameters value ( ′ ), of the model were primarily set based on the study location geology. Since the catchment geology is sediment deposit and moraine (glacial till), then the values for these two parameters can be approximately estimated to be 0.034 (Grabs, 2010, p.27) 7.9 ′ The value of groundwater level, based on the equation [4], is negative if the considered location has a groundwater table below the surface, and is zero or higher if the groundwater table position is at or above the ground surface. Within the stream, the estimated groundwater level will take a positive value. Thus, it is important to define a threshold level above which the presence of water in the stream might generate flow. In this study, the threshold value for the stream to generate flow was primarily assumed to be Zmin=0.1 m. Terrain slope From equation [4], the groundwater table position is related to terrain slope, which is the denominator of fraction, and this value should not be zero. In the model, terrain slope was set to take the value of minimum slope (assumed to be 1/10000) for all cases where DEM-derived slopes were less than the minimum slope value. Time series of flow In the dynamic stream network model, groundwater table position for each cell in the catchment was estimated based on corresponding flow for each time step of the model. The average daily flow series from April 17, 2007 to December 31, 2008, measured at site 7 of the catchment was used in the model. Figure 4 below illustrates the flow series for each corresponding time step of the model. 80 70 Flow(l/s) 60 50 40 30 20 10 0 0 50 100 150 200 250 300 350 400 450 Timesteps Figure 4: Average daily flow series at site 7 9 500 550 600 Table 2 shows the statistical characteristic of the flow series used in the model. The maximum discharge (Qmax) occurs during the spring flood period after most of the snow melts, while the minimum discharge (Qmin) happens at the beginning of the summer. Q75% and Q90% are discharge of 75th percentile and 95th percentile respectively. From Table 2, it is observed that up to 90% of the discharge record falls below 8.59 l/s, and 75% falls below 4.91l/s, which is close to the average discharge of the total time series. Table 2: Statistical characteristics of the flow series Qmax (l/s) Qmin (l/s) Qmean (l/s) Qmedian (l/s) Q75% (l/s) Q90% (l/s) 67.3 0.4 4.6 2.6 4.9 8.5 3.4.Fieldinvestigationformappingstreamheads Field investigation was done to validate the model results, i.e. to see how the locations of stream heads identified in the map extracted from the model differ from the real head locations in the catchment. Furthermore, from field observation, it was possible to see how simple understanding represented by the model works, and to determine if this simple process needs to be revised. It is inevitable that the results of the model are to some extent not consistent with the real systems due to certain model uncertainties and assumptions in the processes, i.e. the locations of the stream heads identified in the map extracted from the model might be different from those in real catchment. In the study catchment of Krycklan, there are 16 flow gaging sites including site 16, the one at the main outlet of the catchment. In this study, three subcatchments upstream of the gaging sites 7, 9, and 1 were selected for detailed field investigations. Additionally, stream heads were also searched in other subcatchments where depending on the accessibility of the terrain. Accessibility was most limited at the onset of spring flood period when the snow layer was relatively deep at some places. The three subcatchments were selected for field observations because they have more available data and are underlain by the majority of till deposit whose transmissivity parameters of the model were based on. Another interesting aspect is that those subcatchments are also partially underlain by sediment deposits, which make it possible not only to discuss how stream network extractions from DEM differs topographically, but also how differently the sources of the stream heads are formed according to geological differences. Another advantage for choosing these subcatchments is that they are relatively easily accessible from nearby roads (Figure 5). 10 Figure 5:: Investigateed subcatch hments of sitte 9, 7, and 1 To facilitate the fiield investig gation, a haand held Gaarmin Oreg gon 450 GPPS with a base b map off stream nnetwork exttracted from m DEM of 55-meter reso olution was used. The bbase map composes off contourr lines of 5 meter m equid distance andd >90% freq quency of co onnecting suurface flow w (Figure 5).. During field obserrvation, the mainstream ms of each selected s sub bcatchmentts were follo owed up soo all headdwater streams or thee branches of the maiinstream co ould be couunted. Whether or nott observeed headwateer streams appear a in thee map extraacted from th he DEM waas verified. 11 3.5.Testingofmodelparameters’value GPS coordinates of observed stream heads were imported into ArcGis and spatially compared with the stream networks extracted from the model. Then, the model parameter values were changed as shown in Table 3 and Table 4 so the observed stream heads and the simulated stream heads were spatially matched. There are three main parameters in the model (b, K0, and Zmin), but only two parameters were used in the testing: b and Zmin. The reason is that, from equation [2], the total groundwater discharge is to the exponent of b. A small change in b value would increase the total groundwater discharge value a lot, and so increase the surface flow in the stream network since the total groundwater discharge is proportional to the stream flow (equation [1]). Case 1: Table 3: Model calibration for b value -1 -1 Ko[l.s .m ] b (m) Numbers of matched heads 0.033466 7.9024 24 0.033466 8.2975 24 0.033466 8.6926 24 0.033466 9.0878 24 Case 2: Table 4: Model calibration for Zmin value -1 -1 Ko [l.s .m ] Zmin (m) Numbers of matched heads 0.033466 ‐0.1 25 0.033466 0.0 26 0.033466 0.1 24 0.033466 0.2 25 Table 3 and Table 4 show the model parameter testing procedure for b and Zmin values. Ko was kept constant for all cases because it is less sensitive to model response. In case 1, the value of b was increase by 5% for each time, and its initial value is bolded in Table 3. For case 2, the value of threshold groundwater table, Zmin, was increased by 0.1 m and later decreased by 0.1 m instead. The bold value (Table 4) corresponds to its initial value. For each case, the observed and simulated stream heads were visually examined in the map and spatially compared. Since there might be spatial error in the GPS, it was assumed that the observed and simulated stream heads are matched if the relative error between them is less than 15 m. From each step of the model parameter testing, it can be seen that there was almost no improvement on the results, which means the number of matched stream heads did not increase even after changing the parameter values. Therefore, afterwards the parameter values from the initial assumption were used for the whole study, i.e. Zmin=0.1 m, Ko= 0.033466 [l.s-1.m-1], and b=7.9024 [m-1]. 4.Results Since modeled stream network may be inaccurately mapped, field observation was carried out to verify the model results. The results as presented in the following sections were divided into two main parts: model results and field observation. Section 4.1, 4.2, and 4.3 show the results from the model. 12 4.1.Sttreamnettworkand dfrequen cyofsurffaceflowg generatin ng Figure 6: >90% > frequuency map of o surface flow generatiing The maap of at leasst 90% frequ uency (Figuure 6) was extracted frrom the freqquency map p of surfacee flow geeneration. The T grid cellls of the loocal drain direction d map, which hhad a frequ uency valuee rangingg from 0.900 to 1.00, were w extractted to exam mine the arreas in the catchment that wouldd frequenntly have suurface flow. From this map, it can n be seen th hat most off the times many m smalll branchees of the strream netwo ork had disccontinuous surface flow w, especiallly those loccated in flatt areas. T These smalll branches can be connsidered as first or zerro-order stre reams. Som me areas nott belongiing to the stream s netw work, in add ddition, had also the presence of discontinuo ous surfacee flow. Siince the moodel assumeed that grouundwater tab ble value from 0.1 m w would generrate surfacee flow wiithin the strream, thus, it i can be infferred that the t initial groundwaterr table thresshold’ valuee in the m model assum mption has strong influuence on strream netwo ork density. If the threshold valuee were inncreased, thhere would be fewer sm mall stream m branches formed beccause of a decrease inn numberr of saturateed areas at th he ground ssurface. 13 Figure 7: >90% > frequeency map of o connectin ng surface fllow Figure 7 representts a frequen ncy map of at least 90% % chance su urface flow w would con nnect to thee c nuous surfaace flow (Fig gure 6), nott main ouutlet of the catchment. Despite maany areas wiith discontin all of thhose areas had h connectiing surface flow to the main outlett (Figure 7)). 14 Figure 8: Stream nettwork map at a maximum m streamflow w ng the stud dy period frrom April 17, 1 2007 too The maaximum vallue of time series disccharge durin Decembber 31, 20008 was 67.3 l/s. The m map in Figurre 8 illustraates the streeam network k extensionn corresponding to maximum m sttreamflow. It can be seeen that durring the higgh flow periiod most off the areaas in the cattchment waas saturated and had su urface flow. It seems thhat the streaam networkk was so dense and the t stream started s from m almost ev verywhere in n the catchm ment. Is it realistic? Itt would nnot be realistic becausse this resuult was soleely based on the modeel assumptiion that thee stream would havee surface flo ow if the grroundwater table is equ ual or higheer than 0.1m m above thee uld be much less densse than this in spite off ground surface. Inn reality, thee stream neetwork shou maximuum streamfllow becausee the saturatted areas with the preseence of surfface water would w drainn into streeam channeel and surfacce flow usuually occurs only in those existing stream chan nnels. 15 Figure 9:: Stream nettwork map at minimum m streamflow w w studyy Unlike the period of highest flow in thee catchmentt, during thee lowest floow in the whole period, the presencce of surfacce flow dram matically deecreases in most m areas (Figure 9). The modell also preedicted a seeveral disco ontinuous ppatches of surface wateer that weree not conneected to thee main strream netwoork. 16 4.2.Sttreamnettworkand dgeologyofthestu udycatchm ment Figure 10: >90% freequency maap and geolo ogy of site 9 Figure 10 shows thhe frequenccy map of sttream netw work in subccatchment oof gaging sitte 9 with att least 900% chance of surface flow generaating within n the stream ms. In overaall, it can be b seen thatt most off the time during d the sttudy periodd from Apriil 17, 2007 to Decembber 31, 2008 8, not manyy small sstream brannches exist in the catcchment. On n the otherr hand, thee catchmentt is mostlyy underlaain by till deeposit, on which w is thee model parrameters’ values v were based. In the t areas off sedimennt deposits, there were a few shorrt stream branches form med along th the mainstreeams, whilee in till ddeposit areaas, the stream networkks tended to t be longeer and havee more bran nches. Onee reason may be because b thiis sedimentt area hass less genttle slope aand located d closer too subcatcchment bordder. 17 Figure 11: >90% freequency maap and geolo ogy of site 1 Additioonally, the stream s netw work with att least 90% frequency at subcatchhment 1 was examinedd (Figure 11). Underrlain by alm most till deeposit area, most frequ uently, the sstream netw work in thiss subcatcchment seem med to havee very few small bran nches. The stream s netw work appeaaring in thiss map shhould be thee perennial streams, w which were usually fou und in the rreal terrain during thee field obbservation. 18 Figure 12 2: > 90% frrequency maap and geollogy of site 7 Subcatcchment of site s 7 (Figu ure 12) was not differeent from thee previous ccase (subcaatchment off site 1). It was undeerlain by alm most till depposits. Almost none off the small sstream brancches appearr in this map, and the t stream network prresent in th his map sho ould be perrennial streeams of thee subcatcchment becaause it is thee stream neetwork extraacted with >90% > frequuency, accorrding to thee model. One of the mainstream ms in the suubcatchmentt ends up in n the wetlannd area. In the t wetlandd area, m many small short brach hes of headdwater streaam were formed and tthese stream m networkss usually are not reealistic sincce this areaa has a verry gentle sllope, wheree the modeel does nott simulate well. 4.3Mo odeledstrreamleng gthandsttreamord dersdurin ngdiffereentflowco onditions 19 80 70 5000 60 4000 50 3000 40 30 2000 Flow (l/s) Total stream length (Km) 6000 20 1000 10 0 0 18‐Jan‐07 28‐Apr‐07 6‐Aug‐07 14‐Nov‐07 22‐Feb‐08 1‐Jun‐08 9‐Sep‐08 18‐Dec‐08 28‐Mar‐09 Total stream length (Km) Flow (l/s) Figure 13: Relationship between total stream length and flow The total length of the stream network was calculated by summing up frictional length from upstream grid cell to downstream grid cell along local drain direction (flow directions extracted from the DEM) in the stream network for each corresponding value of streamflow. The total stream length varied with the same trend as the streamflow (Figure 13). The total stream length at maximum discharge value was around 10 times longer than that at the minimum discharge value (Table 5). The total length of the stream network would be overestimated because the flow directions extracted from the DEM, which was used in the calculation, also took into account the discontinuous patches in the catchment. Table 5: Modeled total stream length during highest and lowest flow conditions Date 4‐May‐08 25‐Jun‐07 Discharge(l/s) 67.3 0.4 20 Total Stream Length (Km) 5620 509 Total stream length (Km) 6000 y = 763,29x0,4477 R² = 0,9988 5000 4000 3000 2000 1000 0 0 10 20 30 40 50 60 70 80 Flow (l/s) Figure 14: Relation between total stream length and corresponding streamflow After fitting different functions, the streamflow could be best approximated by a power relation to the total stream length (Figure 14). The power trend line seemed to fit the points very well when the flow is less than 30 l/s and somehow deviates from those points at the higher streamflow values. The line is steeper at low discharge value. This depicts that during the high flow episodes the total stream length dramatically increase, which means during this episode a small change in discharge values would increase the total stream length a lot. 3000 2750 Total stream length (Km) 2500 2250 2000 1750 1500 1250 1000 750 500 250 0 18‐Jan‐07 28‐Apr‐07 6‐Aug‐07 14‐Nov‐07 22‐Feb‐08 1‐Jun‐08 Order 1 Order 2 Order 3 9‐Sep‐08 18‐Dec‐08 28‐Mar‐09 Order 4 Figure 15: Total stream length versus different stream orders Figure 15 above shows that the stream length of the first, second and third-order stream changed very much at different flow conditions, especially during the high flow period. The stream of these low orders can be regarded as ephemeral streams, the streams that may not be channelized or that form only during the storm event and snowmelt. 21 140 Total stream length (Km) 120 100 80 60 40 20 0 18‐Jan‐07 28‐Apr‐07 6‐Aug‐07 14‐Nov‐07 22‐Feb‐08 1‐Jun‐08 Order 5 Order 6 9‐Sep‐08 18‐Dec‐08 28‐Mar‐09 Order 7 Figure 16: Total stream length versus different stream orders In contrast, the length of the fourth (Figure 15) and fifth-order stream (Figure 16) did not change very much at various flow conditions. The stream of these orders, for instance, can be classified as intermittent streams, the stream that are formed seasonally, after spring flood for example, and that could be channelized. On the other hand, the streams of higher orders, i.e. from order 6 seemed to have little change in length. At certain flow conditions, the stream length does not change even when the flow increases, usually in the low flow episodes. These high order streams may be considered as perennial streams. 22 4.4.Fieldobservations Descriptions of observed stream heads The majority of the observed stream heads were man-made ditches (Table 6). Generally, these ditches were dug to drain water away from agricultural lands, forested area, or wetland and connected to downstream channels. Furthermore, the ditches had a straight regular channel. During the spring flood period in 2012, most of the ditches had active and continuous surface flow from the heads till the first confluence with the water depth ranging from 0.02m to 0.27 m. Only few ditches were found to have discontinuous flow at the upstream channels. Of all the 49 observed stream heads, only two points were road ditches. These road ditches were generally connected to other downslope man-made ditches before evacuating the water to the mainstream. Besides this, 10 stream heads were found in the form of a natural stream. These natural streams, overally, were unchannelized and had irregular channel shape governed by the terrain topography. Table 6: Observed stream heads Types of headwater streams Natural stream Man‐made ditch Road ditch Total Numbers of observed stream heads 10 37 2 49 Water depth (m) 0.02‐0.27 Sources of headwater stream Seepage erosion/ Subsurface flow/ Saturated overland flow Stream head formations Of all the 49 investigated stream heads, it was found that generally those stream heads were formed by either subsurface flow or saturated overland flow (Table 7). Subsurface flow was seen in the form of seepage erosion and seepage from saturation at the upslope of the stream heads, while saturated overland flow was formed because certain area around the stream heads were saturated with presence of surface water. Table 7 illustrates that the majority of the stream heads were formed by subsurface flow in the form of seepage from the saturated soil at the upslope of the stream heads. Because of convergent topography or the hollow at the stream heads, the water drained downwards to the stream heads. Table 7: Formation of stream heads Formations Natural stream Seepage erosion 3 Seepage from saturation 3 Overland flow 4 Man‐made ditch 2 35 0 Road ditch 0 2 0 Total Total 5 40 4 49 Figure 17 below shows an example of the head of a ditch which was formed by seepage erosion in a hollow. Because the surrounding soil of the ditch head was saturated, it formed subsurface flow and recharged water to the hollow (usually the ditch head) before generating surface flow within the ditch channel. From observations in the field this kind of stream head had presence of debris, dead understory vegetation, and tree roots. 23 Figu ure 17: Streeam head wiith seepage erosion t picture (Figure 188) was form med by saturrated area w which had presence p off Overlannd flow in this surface water. Usuually in thiss kind of sttream head,, there was the presennce of dead understoryy vegetatiion and debbris with black color bbecause the water surfaace at this ssaturated so ource lastedd long ennough, and the t water flo ows accordiing to topog graphical ex xpression. Figure 18: 1 Stream hhead with saturated s ov verland flow w 24 Figure 19 below is another example of stream head that was formed by subsurface flow. Thiss stream hhead was loocated at thee base of hiillslope and the headwaater stream was a man--made ditchh ending with a holllow. The upslope u of this stream m head had d water graass and wass saturated.. Becausee of conveergent topog graphy of tthe head, the t water accumulated a d into the hollow h andd formed surface flow w downwarrd in the dittch channel. Figure 19: 1 Stream hhead with seepage from m saturationn 4.5.Co omparisonofobserrvedandpredicted dstreamheads In the ccomparison of observeed and preddicted stream m heads, th he observedd stream heaads and thee predicteed stream network werre mapped aand visually y examined. The observved stream heads weree consideered matcheed with the predicted oones if they fell on the simulated sstream netw work withinn 15 m distance. In I the dyn namic streeam networrk model, for each streamflow w value, a ork for thee study caatchment was w generatted. During g the fieldd corresponding streeam netwo observaation, the daaily average streamfloow was 48 l/s. The strream netwoork in this section s wass extracteed from thhe model with correesponding streamflow s value durring which h the fieldd observaation was doone. A streaam head in tthis study was w defined as a point ffrom which continuouss surface flow of at least 5 m in i the stream m channel generates. A headwateer stream, on o the otherr hand, w was regardeed as a streaam connectting the streeam head to o its first coonfluence. The stream m heads w were primarrily observeed within suubcatchmen nt of site 1, 9, and 7. T The stream heads weree also obbserved in other o subcatchments raather than these t three subcatchm ments and th hose stream m heads rrepresented almost half of the colllected sam mple (49). During D the ffield investigation, thee coordinnates for each observ ved stream heads weere recorded d with a hhand held GPS. Thee coordinnates of som me first con nfluence pooints were also record ded, while some otherrs were nott accessibble due to thhick snow accumulatio a on. 25 Subcatchment of site 7 Figure 20:: Observed stream head ds within su ubcatchmennt 7 During the field observation o , the two m mainstreams of subcattchment off site 7 werre followedd upwardds from its outlet so as to counnt all possib ble headwaater streamss connectin ng to thosee mainstrreams. The results r from m the field oobservation showed thaat only two active ditch hes that hadd surface flow conneecting to thee mainstreaam were fou und, while there t were a lot of stream brachess connectting to the mainstream m according tto the modeel prediction ns. One obseerved stream m head wass located near the border b of the t subcatcchment, and d the otherr one appeaared in weetland area,, m (Figuree 20), none of observed d and simullated stream m heads werre matched.. According to the map This illuustrates the inconsisten ncy betweenn the model simulation and the fielld observatiion. Table 8: Number N of surveyed strream heads falling with hin subcatchhment of site 7 Stream heead No. 26 Descriptio ons 34 41 Natural strea N am Man‐made d M ditch Subcatchment of site 9 Figure 21: Observedd stream heaads in catchm ment of sitee 9 Of all 15 stream heeads observ ved along thhe mainstreaams of the catchment c ass seen in Fiigure 21, 100 w were foun nd on the m map extracteed from thee headwaater streams (bold text rows in Tabble 9) below model. One headw water stream m was a rroad ditch that conneccted to othher ditches. The otherr headwaater streamss did not ap ppear in thee map becaause some ditches werre dug receently, whilee others m might not bee covered in n the DEM rresolutions.. 27 Table 9: Number of stream heads falling within catchment of site 9 Stream head No. Descriptions 4 7 1 2 3 11 37 36 39 31 32 30 41 29 28 40 10 Man‐made ditch Natural stream Man‐made ditch Man‐made ditch Man‐made ditch Man‐made ditch Man‐made ditch Man‐made ditch Man‐made ditch Man‐made ditch Road ditch Man‐made ditch Man‐made ditch Man‐made ditch Man‐made ditch Man‐made ditch Man‐made ditch 28 Subcatchment of site 1 Figure 22 2: Observedd stream heaad in catchm ment of site 1 Like suubcatchmentt of site 7, the mainsttreams of th his subcatch hment weree followed upwards too count aall possible active head dwater streaams or bracches that co onnect to thhose mainsttreams. Thee results from the field investig gation (Figuure 22) sho ow that all headwater h sstreams are man-madee ditches and only five f of them m had activee flow into the mainstreams, whiile the otherrs were dryy and paartly covereed by snow w. The streeam head No.19 (Taable 10) feell within one o of thee mainstrreams, and were w found to be dry dduring the day d of field observationns with the presence p off little suurface waterr further up of the stream m head, wh here water sttarted flowiing. 29 Tablle 10: Numb ber of stream m heads fallling within catchment of site 1 Stream heead No. Descriptio ons 22 21 19 18 20 Man‐made d M ditch Man‐made d M ditch Man‐made d M ditch Man‐made d M ditch Man‐made d M ditch Other ssubcatchments ds in other subcatchmen s ents Figure 23:: Observed sstream head 30 Table 11: Number of stream heads falling within other subcatchments Stream head No. Descriptions 42 43 44 45 13 47 48 49 14 17 16 5 6 8 9 36 12 24 27 26 25 15 23 46 33 Man-made ditch Man-made ditch Man-made ditch Natural stream Man-made ditch Natural stream Man-made ditch Man-made ditch Man-made ditch Man-made ditch Man-made ditch Natural stream Natural stream Man-made ditch Man-made ditch Man-made ditch Man-made ditch Natural stream Natural stream Natural stream Natural stream Man-made ditch Man-made ditch Man-made ditch Road ditch From Table 11, the stream heads observed in other subcatchments other than subcatchment of site 1, 7, and 9 represented half of the collected sample. More importantly, there were up to 12 simulated stream heads that matched with observed stream heads whose headwater streams were either ditches or natural stream (Figure 23). 31 Headw water stream ms Figure 24: O Observed heeadwater streams 32 Table 12: Characteristics of headwater streams No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Streamlength 13 152 8 94 26 18 49 24 20 43 27 5 284 68 154 160 162 48 62 102 397 32 86 16 14 20 18 95 5 220 Descriptions Man‐madeditch Man‐madeditch Man‐madeditch Man‐madeditch Naturalstream Man‐madeditch Naturalstream Man‐madeditch Man‐madeditch Man‐madeditch Man‐madeditch Man‐madeditch Man‐madeditch Man‐madeditch Man‐madeditch Man‐madeditch Man‐madeditch Man‐madeditch Man‐madeditch Man‐madeditch Man‐madeditch Naturalstream Man‐madeditch Man‐madeditch Naturalstream Naturalstream Naturalstream Man‐madeditch Naturalstream Man‐madeditch A headwater streams was defined as the stream that connect its head to its first confluence. During the field observation, some first confluence points of the observed headwater streams were covered by snow, and so were recorded. The length of the headwater streams presented in Table 12 was calculated in ArcGis, based on the recorded coordinates of the observed stream heads. The observed headwater stream’s length, for instance, ranged from 5m to 397m and those streams could be man-made or natural. Generally, the man-made streams tend to be longer, while the natural ones are shorter. 33 4.6.Slope‐arearelationship Slope-area relationship was defined as the relationship between the local slope and the source area (drainage area) of the stream heads. The value for local slope of a stream head was extracted from the DEM-derived slope map, while the value for the source area of the stream head was extracted from the map of upslope accumulated area, whose calculation procedure was previously described. This relationship can describe how stream channels form naturally in different landscapes. Figure 25 below shows the relationship between local slope and source area of the stream head for man-made ditches. It is no doubt that the relationship was not significant because the value of R2 was too low. Local slope of stream heads (%) 45 40 35 30 25 20 15 10 5 0 y = ‐0,0009x + 8,2731 R² = 0,0024 0 200 400 600 800 1000 1200 1400 1600 1800 Source area of stream heads (m^2) Figure 25: Relationship of local slope/source area of stream head for man-made ditches Local slope of stream heads (%) Not different from the case of man-made ditches, any relationship between the local slope and source area of the stream heads for the natural streams was not found (Figure 26). 20 15 10 y = 0,0016x + 11,339 R² = 0,0033 5 0 0 50 100 150 200 250 300 350 400 Source area of stream heads (m^2) Figure 26: Relationship of local slope/source area of stream head for natural stream 34 450 5.Discussions 5.1.Comparisonofobservedandsimulatedstreamheads Model parameter testing was performed to increase the number of matched stream heads between the model simulation and the field observation. However, there was almost no improvement on the results. There might be two reasons for this none-response to changing model parameters’ value. First, it may be because the most of the streams in the Krycklan study catchment are man-made ditches. People dug these ditches on purpose, for example, to drain the water away to and from specific areas, which alters from the natural processes. In contrast, the stream network extraction from the DEM is primarily based on the topographical expressions in the digital elevation data, which tends to be more natural processes. Second reason is that the calibrated parameters were probably less sensitive to the expansion and contraction of the stream network, but the streamflow was much more sensitive. From the results, only 24 out of 49 observed stream heads fell onto branches of the modeled stream network, which accounted for 49% of the collected sample. This result, to some extent, illustrates the inconsistency between the stream network extracted from the model and the observed one. This finding supports other studies which suggest that this inconsistency could be either due to the low resolution of DEM (McMaster, 2002) or unsound performance of DEM in gentle slope areas (Mackay & Band, 1998). For instance, from the field observation, it was observed that the real stream network of the catchments was generally formed by the network of man-made ditches that drain the water away from the agricultural land, forested area, or wetland. These ditches might not be incorporated in the gridded digital elevation data of 5 meter resolution, which was used to extract the stream network. On the other hand, all of the observed stream heads were identified to have been formed by either subsurface flow or overland flow, which is consistent with the finding from Jaeger et al., (2007) that the stream heads can be formed by subsurface flow, or overland flow, or landslide failure. In this study, the form of landslide failure was not encountered. Seepage from saturation at the upslope of the stream heads was the common form of subsurface flow formed at the heads of the stream, most of which were the man-made ditches. Other form of subsurface flow such as seepage erosion was also found. These findings illustrate human impacts on natural systems. The flow in the stream network of the study catchment was, to large extent, governed by the artificial ditch network. 5.2.Variationofstreamlengthandstreamordersatvariousflowconditions From the results (Figure 15), the modeled total stream length of the stream network appears to have a power relation to the streamflow. This suggests that the change in modeled stream length of the stream network was very sensitive to the changes in discharge. However, the power relation might only be an artifact of the model structure which used an exponential relation between groundwater tables and streamflow. Moreover, the model appeared to largely overpredict the occurrence of surface flow and therefore, the modeled total length of the stream network might not present reality. If the flow increases, more streams will be formed and the stream network would expand in the catchment. In contrast, if the flow decreases, the stream network will contract and result in more inactive streams. Thus, during the increasing trend of the flow, the stream network, especially the headwater streams that connect the surrounding stream head locations to the downstream 35 channel, potentially receives more chemical substances from those connected areas such as DIC/TOC or metals. In addition, during different flow conditions (low, medium, and high episode) the stream length for each stream order also changed. This means that if the flow increases, the number of stream orders also increases, especially the lower-order streams. This low-order stream such as first, second and third order are considered ephemeral streams, and usually the transitional streams to the higher-order stream. Thus, in the study of DIC/TOC evasion from the stream, one can assume that stream order and stream length are constant during the same flow condition. The assumption that the stream order and stream length does not change during different flow condition, for example in dry and wet condition, would more or less overestimate the amount of DIC/TOC evasion during dry period, but underestimate the amount during the wet period, especially during the high flow episode. This result is potentially significant for models such as the TRIM model (Grabs, 2010), which assumed a static stream network regardless of the discharge conditions. 5.3.Modeluncertainties The results of the study shows considerable differences between the stream network extracted from DEM and the real stream network observed in the field. This inconsistency may be caused by uncertainties and assumptions in the model structure. The raster based dynamic stream network model used in this study had assumptions and uncertainties as following: o The soil types in the study catchment were assumed to be homogeneous and the same transmissivity parameters’ value were used in the whole catchment areas; however, Grabs (2010) mentions that the threshold groundwater above which the surface flow may generate varies with different riparian zones. o The groundwater table position in the whole catchment was estimated directly from the streamflow and some terrain indices, and only the streamflow measured at site 7 of the catchment was used. This streamflow was then multiplied by the percentage of each contributing area to the total catchment area to scale up the streamflow for the other subcatchments. o Groundwater table was not monitored in the study area, but assumed that the groundwater table position would follow the topographical expression of the catchment. o The surface flow in the stream channels was assumed to occur if the groundwater table was 0.1 m above the ground surface. This assumption might not be true because the elevation of the surface flow in the stream channel may be lower than the ground surface. o The stream network in the study catchment was formed mostly by man-made ditches, some of which were recently dug. There might be chance that these ditches were not incorporated in the DEM. 6.Concludingremarks From the results of the study, it was clearly seen that the observed stream heads and the stream heads appearing in the stream network map extracted from the model did not agree very well. 49% of the observed stream heads (49) fell onto the branches/headwater streams of the modeled stream network and only few of them exactly matched the predicted stream heads. One reason for this inconsistency is because the majority of the stream network in the Krycklan study catchment is man-made ditch, most of which would not be incorporated in the DEM. This result illustrates the human intervention on the natural landscape. People dug the ditches to increase the forest productivity, but they also altered the natural formation of the stream network. 36 Furthermore, it can be seen that the developed raster based stream network model seemed to overestimate the stream network, which cannot represent the dynamic expansion of the stream network very well. For instance, the modeled stream network was so dense, forming many zero and first-order stream almost everywhere, especially during the high flow episode. The modeled stream network was somehow unrealistic because much fewer streams were observed during the field observation carried out during the spring flood of 2012. This inconsistency encourages further field surveys to be conducted in the study of where the streams begin to verify the observed and predicted stream network extracted from the DEM because the stream network extracted from the DEM can be inaccurately mapped. From model results, the number of streams of lower order and their lengths were very sensitive to change in streamflow, especially during the high flow episode. Even though the model seemed to overestimate the stream network, large difference in number of stream with active flow was observed between the peak discharge period and the decreasing discharge period during the field observation. Therefore, studies on the variation of stream-water chemistry should take into account dynamic flow and stream network conditions. The headwater streams are important transitional zones that connect the stream head and its surrounding upslope areas to the downslope stream networks. During the dry conditions such as in summer, many headwater streams may be inactive, while in the wet conditions, for example during the spring flood of 2012, many headwater streams turned to be very active and substantially influence the flow conditions in the catchment. 7.Suggestionsforfuturework Since the modeled stream network was usually not accurately mapped, field observations are always necessary for the study of where the stream starts and the field survey should be done at different flow and wetness conditions in the catchment. For example, a specific stream network belonging to certain subcatchments may be examined during and after spring flood and in late fall to see where the stream begins at different times. Specifically for the study of where the stream begins in the Krycklan study catchment, one would have a clearer picture on the stream network expansion and contraction if one tries to investigate the real stream networks, most of which are man-made ditches, by tracing all the networks that connect to the main streams and include those networks in the map. One would just record the coordinates of each confluence of the stream networks and trace the networks based on the available gridded digital elevation data, and/or the contour lines. This field observation could be done after the spring flood because during the spring flood period, there is still very thick snow accumulation in many areas of the catchment and that is a big challenge that hides the ditch networks from visual observation. 37 8.Acknowledgements I remember when I got a list of topics for master thesis in hydrology/hydrogeology from Roger Herbert. I scanned through the list many times, but still I couldn’t keep my eyes away from a line that wrote “Where does the stream begin?”, and then it became the topic for my master thesis. I would like to express special thanks to Thomas Grabs, my helpful supervisor. He gave me not only the new idea and comments regarding the research, but also a warm encouragement from the start to the end of the writing. Without his immense support, I wouldn’t have completed this thesis. I can’t forget Reinert Huseby Karlsen, who always accompanies me since the start of my master thesis research. He also gave me a lot of helpful comments on my thesis writing and assisted me in technical issues. During the field investigation, he drove me to and from the field. I’d like to say thank to Prof: Kevin Bishop, my subject reviewer, who put me into the research topic at the beginning and for his intervention in making it possible for me to conduct the field observation in Vindeln. His comments on my thesis were very helpful to improve the quality of my research. Pianpian and Anna, thank both of you for accompanying me during the field investigation. You walked with me in the quiet forest, with thick snow depth, sometimes went up to the high hillslope for searching the stream head locations with me. Last but not least, I remember Prof: Allan Rodhe, who gave me the colorful tracer that could be used to identify the movement of the water flowing in the stream channels. 38 References Deursen, V.WPA (1995). Geographical Information Systems and Dynamic Models; development and application of a prototype spatial modelling language. Doctoral dissertation, University of Utrecht, The Netherlands. Grabs, T., Seibert, J., Bishop, K., & Laudon, H. (2009). Modeling spatial patterns of saturated areas: a comparision of the topographic wetness index and a dynamic distributed model. Journal of Hydrology, 373, 15-23. Grabs, T., Bishop, K.H., Laudon, H., Lyon, S.W., Seibert J. (2010). Riprian zone processes and soil-water total organic carbon (TOC): Implication for spatial varability, upscaling and carbon exports. Manuscript. Hancock, G. R., & Evans, K. G. (2006). Channel head location and characteristics using digital elevation models. Earth Surface Processes and Landforms, 31, 809-824. Jaeger, K. L., Montgomery, D. R., & Bolton, S. M. (2007). Channel and Perennial Flow Initiation in Headwater Streams: Management Implications of Variability in SourceArea Size. Environ Manage, 40, 775-786. Köhler, S., Buffam, I., Seibert, J., Bishop, K., & Laudon, H. (2009). Dynamics of stream water TOC concentrations in a boreal headwater catchment: Controlling factors and implications for climate scenarios. Journal of Hydrology, 373, 44-56. Lyon, S. W., Walter, M. T., Gerard-Marchant, P., & Steenhuis, T. S. (2004). Using a topographic index to distribute variable source area runoff predicted with the SCS curve-number equation. Hydrological Processes, 18, 2757-2771. Mackay, D. S., & Band, L. E. (1998). Extraction and representation of nested catchment areas from digital elevation models in lake-dominated topography. Water Resources Research, 34(4), 897-901. Montgomery, D. R., & Dietrich, W. E. (1988). Where do channels begin?. Nature, 336, 232-234. Montgomery, D. R., & Dietrich, W. E. (1989). Source areas, drainage density, and channel initiation. Water Resources Research, 25, 1907-1918. McMaster, K. J. (2002). Effects of digital elevation model resolution on derived stream network positions. Water Resources Research, 38(4), 13-1. O'Callaghan, J. F., & Mark, D. M. (1984). The extraction of drainage networks from digital elevation data. Computer Vision, Graphics, and Image Processing, 28(3), 323-344. PCRaster documentation — PCRaster v3.0.1 documentation. (n.d.). PCRaster | Software for environmental modelling. Retrieved June 13, 2012, from http://pcraster.geo.uu.nl/documentation/PCRaster/html/index.html Tarboton, D. G., Bras, R. L., & Rodriguez-Iturbe, I. (1991). On the extraction of channel networks from digital elevation data. Hydrological Processes, 5, 81-100. Wallin, M., I. Buffam, M. Öquist, H. Laudon, and K. Bishop (2010), Temporal and spatial variability of dissolved inorganic carbon in a boreal stream network: Concentrations and downstream fluxes, J. Geophys. Res., 115, G02014, doi:10.1029/2009JG001100 39 Appendix1:Fieldprotocol Physical characteristics of stream heads o Begin with convergent topography; o Located at the base of a valley down slope of a hollow o Formed by landslide failure or seepage erosion or subsurface flow Physical characteristics of headwater streams o Can be continuous or discontinuous from the downstream channel to the first confluence o Can be first order or zero order stream o Having flowing water continuous for at least 5 meters o Can be channelized (man-made ditches) or unchannelized (natural streams) Headwater channels have four distinguished topographic units: o Hillslope: it has divergent or straight contour lines with unchannelized flow or dispersion flow. o Zero-order basins: there is the presence of unchannelized hollow with convergent contour lines. It can be saturated with overland flow and return flow into flood plain. o Ephemeral/transitional channels: it emerges from zero-order basins, with definable banks if the channels exist at the outlet of the basins. In addition, it has ephemeral flow and can be considered temporary storage of organic carbon, which may also contain discontinuous segments prior to entering first-order channels. o First order stream channels: it may be directly originated from zero-order basin. Field materials o o o o o o Tracer for indentifying if the water is moving Global Positioning System (GPS) device to mark the stream heads Field phones to communicate between field observers A plastic container for maintain the tracer A tape/meter to measure the water depth and stream length Waterproof field notes with back-ups Field procedures a. Identify the formation of the headwater stream - Overland flow: presence of debris, brown /killed vegetation like pastures/grass seed, or - Landslide failure: presence of recent mass wasting scars, or - Convergent subsurface flow: if the two cases above are not applied or there is the presence of observed seeping from the channel head. b. Record the coordinates of the stream heads with GPS c. Determine the stream order and identify if the stream is channelized or not. d. Determine if the water is moving in the stream In a steep slope stream, the movement of water flow can be easily visualized, but in a gentle slope stream, it was really difficult to indentify if the water is moving. In the later case, the colorful tracer was used: - Mix little amount of tracer with a cup of stream water and stir it to a solution a - Choose any point at the upper part of the headwater stream, record the time, and inject the solution to the selected point - After several minutes, measure the distance of which the tracer solution moves along with the water in the stream channels. e. Approximately find the deepest point of the stream heads and measure the depth of the water. b Appendix2:Dynamicstreamnetworkmodel’sscripts binding # ========================================= # INPUT # ========================================= # Maps #-----------------------------------------# Network DEM STREAM_DEM = dem.map; # Stream slope (Downslope Index) STREAM_SLOPE = stream_slope.map; # Stream catchment area STREAM_CATCHM_AREA = stream_ca.map; # Gaging sites GAGING_SITE = gaging_site.map; # Setting based map STREAM_LDD=stream_ldd.map; #Geology map classes GEO_CLASS=geo_class.map; # Mainstream of more than 4 ordered MAIN_STREAM=main_stream.map; #-------------------------------------------# Time series TS_FLOW_SITE = flow.tss; # [l/s] # #-----------------------------------------# CONSTANTS #-----------------------------------------# Start time (make sure it is smaller or equal END_TIME) START_TIME = 1; # End time (make sure it is smaller or equal to N_TIMESTEPS) END_TIME = 625; # Total number of time steps N_TIMESTEPS = 1; # Since the catchment geology is kind of sediment deposit and moraine b=7.9024; # [1/m] Ko=0.033466; # [l/s.m], Flow per m flow width # Site upslope area UAA_SITE=465478 ; #[m2] # Minimum slope SLOPE_MIN =0.0001; # Minimum GWLevel above which the stream generate flow Z_MIN=0.1; #[m] # Total number of time series data TOTAL_TS=625; # ========================================= c # OUTPUT # ========================================= # Maps #-----------------------------------------# Groundwater table #-----------------------------------------areamap STREAM_LDD; timer START_TIME END_TIME 1; initial ONES = STREAM_DEM/STREAM_DEM; ZEROS = STREAM_DEM*0; # Smooth dem DEM_ = lddcreatedem(STREAM_DEM,1E35,1E35,1E35,1E35); # Create LDD STREAM_LDD = lddcreate(DEM_,1E35,1E35,1E35,1E35); report stream_ldd.map=STREAM_LDD; # Calculate stream slope STREAM_SLOPE = slope(STREAM_DEM); STREAM_SLOPE=if(STREAM_SLOPE<SLOPE_MIN then SLOPE_MIN else STREAM_SLOPE); report stream_slope.map = STREAM_SLOPE; # Stream order STREAM_ORDER=streamorder(STREAM_LDD); report stream_order.map=STREAM_ORDER; #MAIN_STREAM=STREAM_ORDER>5; MAIN_STREAM = if(STREAM_ORDER>5,STREAM_ORDER); report main_stream.map = MAIN_STREAM; # Calculate stream cathcment area STREAM_CATCHMENT_AREA=accuflux(STREAM_LDD,25); report stream_ca.map=STREAM_CATCHMENT_AREA; # Catchment for the gaging site whose discharge series used in the model GAGED_CATCHMENT=catchment(STREAM_LDD,GAGING_SITE); report site_catchment.map=GAGED_CATCHMENT; # Determine upslope accumulated area at the outlet UAA_OUTLET= STREAM_CATCHMENT_AREA==mapmaximum(STREAM_CATCHMENT_AREA); # Determine the main outlet of the catchment MAIN_OUTLET = nominal(scalar(UAA_OUTLET)/scalar(UAA_OUTLET)); #Remove missing values SURFACE_FLOW_CONNECTED_SUM = ZEROS; SURFACE_FLOW_SUM = ZEROS; d dynamic # Flow at gauging site FLOW_SITE = timeinputscalar(TS_FLOW_SITE,2); # Estimate the discharge at the main outlet Q = FLOW_SITE*(STREAM_CATCHMENT_AREA/ UAA_SITE); # Estimate the groundwater table distribution within study period Zgwt=(1/b)*ln(Q/(celllength()*STREAM_SLOPE*Ko)); # Determine GWLevel above which the sream generates flow SURFACE_FLOW = if(Zgwt>Z_MIN then Zgwt); report SF=SURFACE_FLOW; # Count the number of days for which the stream may generate the flow SURFACE_FLOW_ = if(Zgwt>Z_MIN then ONES else ZEROS); SURFACE_FLOW_ = cover(SURFACE_FLOW_, ZEROS); SURFACE_FLOW_SUM = SURFACE_FLOW_ + SURFACE_FLOW_SUM; SURFACE_FLOW_SUM_CLEAN = if(SURFACE_FLOW_SUM> 0 then SURFACE_FLOW_SUM); report N_DAY_SF.map = SURFACE_FLOW_SUM_CLEAN; # Calculate the probability for which the stream might generate surface flow PROB_SF = SURFACE_FLOW_SUM_CLEAN/TOTAL_TS; PROB_SF_CLEAN = if(PROB_SF>0.1,PROB_SF); report P_SF.map = PROB_SF_CLEAN; P95_SF = if(PROB_SF_CLEAN>=0.95,PROB_SF_CLEAN); report P95_SF.map=P95_SF; P90_SF = if(PROB_SF_CLEAN>=0.90,PROB_SF_CLEAN); report P90_SF.map = P90_SF; # LDD map corresponding to the days with surface flow LDD_SF=if(Zgwt>Z_MIN then STREAM_LDD); report ldd_sf.map=LDD_SF; STREAM_ORDER_=streamorder(LDD_SF); STREAM_ORDER_SF =if(STREAM_ORDER_>=1,STREAM_ORDER_); report stream_order_sf.map =STREAM_ORDER_SF; # Catchment whose flow at the outlet connecting with surface flow CATCHMENT_FLOW=boolean(catchment(LDD_SF,MAIN_OUTLET)); report SFC=CATCHMENT_FLOW; # Compute the number of days the surface flow connecting to the outlet SURFACE_FLOW_CONNECTED = if(CATCHMENT_FLOW then ONES else ZEROS); SURFACE_FLOW_CONNECTED = cover(SURFACE_FLOW_CONNECTED, ZEROS); SURFACE_FLOW_CONNECTED_SUM = SURFACE_FLOW_CONNECTED + SURFACE_FLOW_CONNECTED_SUM; report N_DAY_SFC.map = SURFACE_FLOW_CONNECTED_SUM; # Calculate the probability for which the stream flow connect to the outlet PROB_SFC = SURFACE_FLOW_CONNECTED_SUM/TOTAL_TS; PROB_SFC_CLEAN = if(PROB_SFC>0.1,PROB_SFC); report P_SFC.map = PROB_SFC_CLEAN; P95_SFC= if(PROB_SFC_CLEAN>=0.95,PROB_SFC_CLEAN); report P95_SFC.map = P95_SFC; P90_SFC= if(PROB_SFC_CLEAN>=0.90,PROB_SFC_CLEAN); e report P90_SFC.map = P90_SFC; # Count the number of cell with surface flow connecting to the stream orders CELL_SF = areatotal(SURFACE_FLOW,MAIN_STREAM); CELL_SFC = areatotal(scalar(CATCHMENT_FLOW),MAIN_STREAM); report cell_sf=CELL_SF; report cell_sfc=CELL_SFC; # Assign stream length for all streams STREAM_LENGTH_OUTLET = downstreamdist(LDD_SF); # Calculate total total legnth of all streams in the catchment STREAM_LENGTH_SUM = maptotal(STREAM_LENGTH_OUTLET); TS_STREAM_LENGTH_OUTLET_TOTAL = timeoutput(1,STREAM_LENGTH_SUM); # Assign the longest stream of all streams STREAM_LENGTH_ = slopelength(LDD_SF,1); # Calculate the stream length for the longest stream STREAM_LENGTH_LONGEST =mapmaximum(STREAM_LENGTH_); TS_STREAM_LENGTH_LONGEST = timeoutput(1,STREAM_LENGTH_LONGEST); # Assing stream length by stream orders STREAM_LENGTH_ORDER1 = if(STREAM_ORDER_SF==1,downstreamdist(LDD_SF)); STREAM_LENGTH_ORDER1_SUM = maptotal(STREAM_LENGTH_ORDER1); TS_STREAM_LENGTH_TOTAL_ORDER1 = timeoutput(1,STREAM_LENGTH_ORDER1_SUM); STREAM_LENGTH_ORDER2 = if(STREAM_ORDER_SF==2,downstreamdist(LDD_SF)); STREAM_LENGTH_ORDER2_SUM = maptotal(STREAM_LENGTH_ORDER2); TS_STREAM_LENGTH_TOTAL_ORDER2 = timeoutput(1,STREAM_LENGTH_ORDER2_SUM); STREAM_LENGTH_ORDER3 = if(STREAM_ORDER_SF==3,downstreamdist(LDD_SF)); STREAM_LENGTH_ORDER3_SUM = maptotal(STREAM_LENGTH_ORDER3); TS_STREAM_LENGTH_TOTAL_ORDER3 = timeoutput(1,STREAM_LENGTH_ORDER3_SUM); STREAM_LENGTH_ORDER4 = if(STREAM_ORDER_SF==4,downstreamdist(LDD_SF)); STREAM_LENGTH_ORDER4_SUM = maptotal(STREAM_LENGTH_ORDER4); TS_STREAM_LENGTH_TOTAL_ORDER4 = timeoutput(1,STREAM_LENGTH_ORDER4_SUM); STREAM_LENGTH_ORDER5 = if(STREAM_ORDER_SF==5,downstreamdist(LDD_SF)); STREAM_LENGTH_ORDER5_SUM = maptotal(STREAM_LENGTH_ORDER5); TS_STREAM_LENGTH_ORDER5 = timeoutput(1,STREAM_LENGTH_ORDER5_SUM); STREAM_LENGTH_ORDER6 = if(STREAM_ORDER_SF==6,downstreamdist(LDD_SF)); STREAM_LENGTH_ORDER6_SUM = maptotal(STREAM_LENGTH_ORDER6); TS_STREAM_LENGTH_TOTAL_ORDER6 = timeoutput(1,STREAM_LENGTH_ORDER6_SUM); STREAM_LENGTH_ORDER7 = if(STREAM_ORDER_SF==7,downstreamdist(LDD_SF)); STREAM_LENGTH_ORDER7_SUM = maptotal(STREAM_LENGTH_ORDER7); f TS_STREAM_LENGTH_TOTAL_ORDER7 = timeoutput(1,STREAM_LENGTH_ORDER7_SUM); STREAM_LENGTH_ORDER8 = if(STREAM_ORDER_SF==8,downstreamdist(LDD_SF)); STREAM_LENGTH_ORDER8_SUM = maptotal(STREAM_LENGTH_ORDER8); TS_STREAM_LENGTH_TOTAL_ORDER8 = timeoutput(1,STREAM_LENGTH_ORDER8_SUM); STREAM_LENGTH_ORDER9 = if(STREAM_ORDER_SF==9,downstreamdist(LDD_SF)); STREAM_LENGTH_ORDER9_SUM = maptotal(STREAM_LENGTH_ORDER9); TS_STREAM_LENGTH_TOTAL_ORDER9 = timeoutput(1,STREAM_LENGTH_ORDER9_SUM); STREAM_LENGTH_ORDER10 = if(STREAM_ORDER_SF==10,downstreamdist(LDD_SF)); STREAM_LENGTH_ORDER10_SUM = maptotal(STREAM_LENGTH_ORDER10); TS_STREAM_LENGTH_TOTAL_ORDER10 = timeoutput(1,STREAM_LENGTH_ORDER10_SUM); g Appendix3:Coordinatesanddescriptionsofobservedstreamheads Note: Lat/Lon: X/Y: Name: Site No: Site No 1 2 3 4 5 6 7 8 9 10 11 49 35 30 36 31 32 33 26 27 28 29 34 45 46 47 48 18 19 20 21 22 23 24 25 Imported from GPS with UTM_WGS84 coordinate system Calculated in ArcGIS from imported Lat/Lon, projected to RT90_25_gon_V coordinate system Modified name of stream heads saved in GPS (012=stream head of first point, 112=stream head of 11th point) Observed stream heads No corresponding to original field data Name Lat Lon 012 022 032 042 052 062 072 082 092 102 112 122 132 142 152 162 172 182 192 202 212 222 232 242 252 262 272 282 292 302 312 322 332 342 352 64.24398 64.24662 64.24638 64.24274 64.23888 64.23891 64.24225 64.24527 64.24525 64.24774 64.24774 64.20282 64.25129 64.25738 64.24999 64.25845 64.2587 64.25658 64.25599 64.25618 64.25506 64.25202 64.25128 64.21952 64.21796 64.21774 64.2182 64.25861 64.25672 64.25553 64.25297 64.25325 64.25974 64.26015 64.26195 19.79073 19.79445 19.79125 19.79079 19.79097 19.79076 19.78672 19.77278 19.77288 19.782 19.782 19.83581 19.80894 19.80015 19.80805 19.79947 19.80013 19.80723 19.80697 19.80547 19.80724 19.80919 19.8098 19.75886 19.75526 19.75572 19.75464 19.771 19.77107 19.77454 19.77494 19.77632 19.78383 19.78057 19.77469 X Y Descriptions Geology Classes 1693046.49 1693208.16 1693055.23 1693057.97 1693093.98 1693083.51 1692864.33 1692168.23 1692173.54 1692597.46 1692597.46 1695520.23 1693877.31 1693409.08 1693843.30 1693368.60 1693398.73 1693757.12 1693748.94 1693675.04 1693768.29 1693884.32 1693918.94 1691672.59 1691508.82 1691532.80 1691477.27 1691989.73 1692006.17 1692182.07 1692219.28 1692284.51 1692602.74 1692441.88 1692144.76 7132345.48 7132650.39 7132613.98 7132208.21 7131778.94 7131781.52 7132141.32 7132434.80 7132433.00 7132737.38 7132737.38 7127902.10 7133214.51 7133866.10 7133066.89 7133982.11 7134012.16 7133797.93 7133731.05 7133747.85 7133629.02 7133296.08 7133216.46 7129526.20 7129342.06 7129318.74 7129366.69 7133914.31 7133703.71 7133581.61 7133297.89 7133333.59 7134078.88 7134114.30 7134297.30 Man-made ditch Man-made ditch Man-made ditch Man-made ditch Natural stream Natural stream Natural stream Man-made ditch Man-made ditch Man-made ditch Man-made ditch Man-made ditch Man-made ditch Man-made ditch Man-made ditch Man-made ditch Man-made ditch Man-made ditch Man-made ditch Man-made ditch Man-made ditch Man-made ditch Man-made ditch Natural stream Natural stream Natural stream Natural stream Man-made ditch Man-made ditch Man-made ditch Man-made ditch Road ditch Road ditch Natural stream Man-made ditch Sediment deposits Till deposits Till deposits Sediment deposits Sediment deposits Sediment deposits Till deposits Till deposits Till deposits Till deposits Till deposits Till deposits Till deposits Till deposits Till deposits Sediment deposits Till deposits Till deposits Till deposits Till deposits Till deposits Till deposits Till deposits Till deposits Till deposits Till deposits Till deposits Till deposits Till deposits Till deposits Till deposits Till deposits Till deposits Till deposits Wetlands h 12 13 14 15 16 17 37 38 39 40 41 42 43 44 362 372 382 392 402 412 422 432 442 452 462 472 482 492 64.24993 64.25163 64.2511 64.2511 64.25536 64.25534 64.24908 64.24833 64.2498 64.2499 64.25238 64.25372 64.25476 64.25647 19.77275 19.7749 19.77654 19.77654 19.77714 19.7772 19.80714 19.80533 19.8035 19.80352 19.80306 19.8046 19.80506 19.80509 1692134.46 1692226.67 1692310.13 1692310.13 1692309.18 1692312.24 1693805.28 1693723.17 1693624.19 1693624.68 1693585.08 1693649.89 1693664.86 1693654.28 i 7132953.15 7133149.30 7133094.51 7133094.51 7133571.01 7133568.63 7132963.59 7132874.17 7133031.98 7133043.62 7133318.36 7133471.83 7133589.20 7133778.83 Man-made ditch Man-made ditch Man-made ditch Man-made ditch Man-made ditch Man-made ditch Man-made ditch Man-made ditch Man-made ditch Natural stream Man-made ditch Natural stream Man-made ditch Man-made ditch Till deposits Till deposits Till deposits Till deposits Till deposits Till deposits Till deposits Till deposits Till deposits Till deposits Till deposits Till deposits Till deposits Till deposits Appendix4:Fielddatarecords Date Site No 16/5/2012 1 2 3 4 5 6 7 8 9 10 Head WD (m) 0.06 0.08 0.1 0.09 0.1 0.13 0.12 0.13 0.11 0.11 CD WW (m) (m) 0.3 0.3 0.3 0.4 0.4 0.4 0.35 0.4 0.6 0.6 0.7 0.7 CW (m) 0.6 0.8 0.6 0.8 0.8 1.2 1.2 j Descriptions Ditch, no flow further upward Ditch discontinuous further upward with water grass Ditch continuous flow slow flow Not channelized falling tree over Natural flow Not channelized Natural flow with many small hollows Flood plain or semi‐wetland Not channelized Ditch with continuous flow Ditch with continuous Formation Seepage from saturated flow Seepage from saturated flow Seepage from saturated flow Seepage erosion from under the tree roots Seepage erosion from under the tree roots Saturated overland flow Seepage from saturated flow Seepage from saturated WD (m) 0.05 0.09 0.1 flow Sharing head with site 17 at the highest dividing elevation Sharing head with site 16 at the highest dividing elevation flow Seepage from saturated flow Seepage from saturated flow 0.09 0.1 WW (m) 0.02 0.3 0.06 0.6 Confluence CW (m) Descriptions Connecting the other ditch Confluence with that of site 9 Spreading fast flow Natural fast disperse flow 0.7 0.8 Natural setream connecting to ditch at the upstream 1 1.1 0.35 0.4 Natural stream not channelized connecting to ditch at the upstream CD (m) 0.35 0.3 17/5/2012 11 12 13 14 15 16 17 18 19 20 0.05 0.08 0.08 0.02 0.04 0.03 0.06 0.05 0.6 0.5 0.35 0.45 0.3 0.3 0.4 0.3 0.6 0.45 0.3 0.4 0.25 0.4 0.85 0.7 1 1.2 0.8 1 0.5 0.6 1.7 Continuous flow ditch Continuous flow ditch with snow covered Ditch covered by snow the whole stream length sharing head with site 21 Ditch covered by snow the whole stream length sharing head with site 20 Covered by snow ditch Covered by snow ditch continuous flow snow melting contribute ditch continuous flow ditch continuous flow Ditch continuous flow snowmelting contribute Ditch no flow upward Seepage from saturated flow with the presence of water grass Seepage from saturated flow and snowmelting contribution Seepage from saturated flow Seepage from saturated flow Seepage from saturated flow Seepage from saturated flow Seepage from saturated flow Seepage from saturated flow k 0.08 Natural stream not channelized connecting to ditch at the upstream Natural stream not channelized covered by snow Covered by snow covered by snow Covered by snow Connecting to road ditch 18/5/12 21 22 23 24 25 26 27 28 29 30 31 0.05 0.06 0.1 0.12 0.06 0.05 0.05 0.05 0.06 0.04 0.08 0.8 0.3 0.25 0.3 0.4 0.45 0.04 0.4 0.4 0.7 0.2 0.35 0.4 0.57 0.8 0.4 0.4 1.2 0.5 0.4 0.6 1.05 0.3 1 Not channelized widen road titch Widen road ditch Covered by the snow at the end Not channelized killed grass, natural flow continuous flow Ending with the tree snow covered nearby hillslope Discontinuous upwards of the stream head presence of channel erosion Ditch with discontinuous flow upwards with water grass presence of surface water Ditch with discontinuous flow upwards Ditch with continuous flow till the end Debris, killed grass, organic matters, at the base of hillslope ditch Presence of surface water upwards of the head Ditch Dividing at the highest elevation between 39 & 40 covered by snow Saturate surface flow Saturated flow Seepage from saturated flow Seepage from saturated head Seepage from saturated head Seepage erosion Seepage from saturated head Seepage from saturated head Seepage erosion from under the tree roots l 0.5 0.05 0.05 0.05 0.3 0.45 0.4 0.8 0.7 272 19/5/12 32 33 34 35 36 37 38 39 40 41 42 0.08 0.05 0.1 0.06 0.04 0.06 0.02 0.27 0.05 0.05 0.4 0.02 0.4 0.4 0.25 0.5 0.3 0.6 0.25 0.4 0.45 0.8 1.2 0.09 0.75 0.5 0.35 1 0.7 1 1.5 0.7 0.9 0.8 0.5 Dividing at the highest elevation between 39 & 40 covered by snow Debris and organic matter no flow upwards but little surface water no flow upwards of the head with water grass No flow upwards of the head covered by snow drain to road ditch Presence of surface water upwards Organic matter, killed grass rocks, water grass in the channel Not channelized, small hollows under the rocks and roots of tree, channelized downwards Ditch with little sufrace flow saturated upwards, under falling die tree with killed grass Not channelized, old shallow Ditch with little sufrace flow and water grass Ditch covered by snow head Subsurface flow covered bys snow Seepage from saturated head Seepage from saturated head Seepage from water grass upwards Seepage from saturated Seepage from saturated head Seepage from saturated head Seepage from saturated head Seepage from saturated m 43 44 45 46 47 Note: WW WD CD CW 0.07 0.3 0.25 0.6 no flow upwards No flow upwards, but saturated with water grass Natural stream, not channelized black color, presence of killed grass and organic matters Natural flow Natural flow Natural flow CW CD WW head Seepage from saturated head Seepage erosion from under the died falling tree Overland flow Overland flow Overland flow WD Width of water surface within the channel Depth of water measured at the deepest point Channel depth Channel width n Tidigare utgivna publikationer i serien ISSN 1650-6553 Nr 1 Geomorphological mapping and hazard assessment of alpine areas in Vorarlberg, Austria, Marcus Gustavsson Nr 2 Verification of the Turbulence Index used at SMHI, Stefan Bergman Nr 3 Forecasting the next day’s maximum and minimum temperature in Vancouver, Canada by using artificial neural network models, Magnus Nilsson Nr 4 The tectonic history of the Skyttorp-Vattholma fault zone, south-central Sweden, Anna Victoria Engström Nr 5 Investigation on Surface energy fluxes and their relationship to synoptic weather patterns on Storglaciären, northern Sweden, Yvonne Kramer Nr 237 Structural Model of the Lambarfjärden Area from Surface and Subsurface Data in Connection with the E4 Stockholm Bypass Anna Vass, June 2012 Nr 238 Mechanisms Controlling Valley Asymmetry Development at Abisko, Northern Sweden and Sani Pass, Southern Africa, Carl-Johan Borg, August 2012 Nr 239 Effect of Orientation on Propagation of Pre-existing fractures, Hajab Zahra, August 2012 Nr 240 Mobility of multi-walled carbon nanotubes in unsaturated porous media, Abenezer Mekonen, August 2012 Nr 241 Re-processing of Shallow and Deep Crustal Reflection Seismic Data along BABEL Line7, Central Sweden, Hanieh Shahrokhi, August 2012 Nr 242 Usability of Standard Monitored Rainfall-Runoff Data in Panama, Juan Diaz River Basin, José Eduardo Reynolds Puga, August 2012 Nr 243 Numerical Model of a Fossil Hydrothermal System in the Southern East Pacific Rise Exposed at Pito Deep, Páll Halldór Björgúlfsson, September 2012 Nr 244 Regional Precipitation Study in Central America, Using the WRF Model Tito Maldonado, September 2012 Nr 245 Reduktion av järn, mangan och CODMn i dricksvatten – Ett pilotförsök vid Högåsens vattenverk, Tommy Olausson, September 2012 Nr 246Short-term Variations in Ice Dynamics During the Spring and Summer Period on Storglaciären, Kebnekaise, Sweden, Helena Psaros, September 2012 Nr 247 Reassessment of the ‘last’ Goniopholidid Denazinosuchus kirtlandicus Wiman, 1932 from the Late Cretaceous of New Mexico, Oskar Bremer, October2012
* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project
advertisement
© 2021