Title of your Paper – Mind the Uppercase Letters
Cartographic and Hydraulic Modelling Based on Terrestrial Laser Scanning Data of Kuzlovec TorrentDejan Grigillo, Simon Rusjan, Anja Vre?ko, Elvira D?ebo, Klemen Kozmus Trajkovski, Tilen Urban?i?, Mojca Kosmatin Fras, Matja? Miko?, Du?an Petrovi?University of Ljubljana, Faculty of Civil and Geodetic Engineering, Jamova cesta 2, SI-1000, Ljubljana, SloveniaAbstract. The article describes the cartographic and hydraulic modelling of the Kuzlovec torrent in the Polhov Gradec Hills. The torrent is situated in forest covered area, formed by steep slopes, rocks and numerous cascades. Terrestrial laser scanning was used to collect the data for the DTM production, which was used as a model for the mathematical modelling, prepared using two hydraulic models, and to study the influence of the large woody debris on the water flow. The demanding field work, the DTM production, hydraulic modelling and presentation of DTM changes, caused by high water floods in the period of our measurements are explained and discussed. Keywords: digital terrain model, torrential hydraulics, large woody debris, torrents, TLS, cartographic presentationIntroductionMathematical modelling of water flow with free surface in hydraulic models requires the most accurate and correct data about the surface. One of the most commonly used mathematical models of the Earth's surface, which is used in hydraulic models, is the DTM, organized as high values in grid network. Its use is widespread because it makes it easy to deal with the direction of movement of water flows along the ground. DTM can be made from data obtained from different sources: laser scanning (lidar), terrestrial or aerial photogrammetry, radar and optical satellite imagery, GNSS and terrestrial topographic surveying, contour lines on a topographic map. Lidar is a technology that enables the capture of accurate data on the ground in large areas and also areas covered with forest. In addition, laser scanning data is an useful data source in forestry applications for estimation of biomass, of crown transparency and of other forest inventory parameters such as tree height and diameter at breast height (DBH) ADDIN Docear CSL_CITATIONTHIS CITATION DATA SHOULD NOT BE MANUALLY MODIFIED!!!{ "schema": "", "citationID": "{1410cdf6-00c8-4933-ae35-2dbc199e120a}", "properties": { "noteIndex": 0 }, "citationItems": [ { "id": "Thies2004", "item": { "id": "Thies2004", "type": "article-journal", "author": [ { "family": "Thies", "given": "M" }, { "family": "Spiecker", "given": "H" } ], "volume": "2", "title": "Evaluation and future prospects of terrestrial laser scanning for standardized forest inventories", "container-title": "Forest", "number": "2.2", "page": "1", "page-first": "1", "issued": { "date-parts": [ [ 2004 ] ] } } } ]}(Thies and Spiecker 2004; ADDIN Docear CSL_CITATIONTHIS CITATION DATA SHOULD NOT BE MANUALLY MODIFIED!!!{ "schema": "", "citationID": "{3787b347-f857-4f19-8a62-6bc13a78d0da}", "properties": { "noteIndex": 0 }, "citationItems": [ { "id": "Simonse2003", "item": { "id": "Simonse2003", "type": "paper-conference", "author": [ { "family": "Simonse", "given": "Merlijn" }, { "family": "Aschoff", "given": "Tobias" }, { "family": "Spiecker", "given": "Heinrich" }, { "family": "Thies", "given": "Michael" } ], "volume": "2003", "title": "Automatic determination of forest inventory parameters using terrestrial laserscanning", "container-title": "Proceedings of the ScandLaser Scientific Workshop on Airborne Laser Scanning of Forests", "page": "252-258", "page-first": "252", "issued": { "date-parts": [ [ 2003 ] ] } } } ]}Simonse et al. 2003).Problems in the production of DTM’s in steep forested terrain from the aerial laser scanning (ALS) have described by Kobler (2011). He presented a new algorithm, called REIN, where filtering the cloud points on steep slopes covered with heterogeneous trees uses several possible estimates of the height of the ground at points in rectangular or triangular grid interpolated from surrounding ground points (Kobler et al 2007). The nature of the terrestrial laser scanning (TLS) data is different from the ALS data (larger incidence scanning angles, higher, but inhomogeneous point cloud density). Several authors have presented the TLS data based DTM creation in forested areas. In most cases, DTM is used as one of sources to obtain data for forest inventory (Simons et al 2003; Bienert et al 2006) and for 3D modeling of standing trees (Thies et al 2004; Moskal and Zheng 2012).One of the goals of our research was also to estimate large woody debris (LWD) recruitment potential in torrential watersheds, and to describe the interactions of wood and torrential sediments in a cascading forested torrential channel (Grigillo et all 2015). Area and terrain measurementsThe Kuzlovec Torrent in the Grada??ica watershed (central Slovenia), was chosen as a study area (mouth at 394m a.s.l. and the highest point at 800m a.s.l). The TLS measurements covered the lower reach with slope ~10° between approx. 500m a.s.l. to the torrent mouth at 394m a.s.l. with slopes of ~30° (figures 1). The torrent is formed by steep slopes, rocks and numerous cascades, a huge amount of dead wood create an addition obstacle for water flow. Entire torrent area is covered mainly with beech forests.The first data acquisition was carried out in three days in spring 2013 by a terrestrial laser scanner (model RIEGL VZ-400) and digital camera Nikon D700. The discharge during the field surveying was measured by a dilution method and was estimated to ~10 l/s. In total 29 scan stations were distributed along the torrent to assure the optimal coverage of the torrential channel. The scans captured from different scanning station were registered into a common coordinate system (CCS) using tie points, which were signalized with Riegel’s cylindrical retro-reflected targets with the height of 10 cm (figure 2 left). The applied resolution of terrestrial laser scanning was 8 cm at 50 m distance for the torrential channel and the neighbouring slopes, and 4 cm at 50 m distance for the narrow torrential channel with channel structures (step-pool sequences, rocks & boulders). 23426316286500317514541500Figure 1. Forrest covered area of torrent Kuzlovec (left), TLS surveyed area and geodetic network (right).Ten control points along the torrent were stabilized and signalized with cylindrical retro-reflected targets during scanning so that their coordinates in a CCS could be obtained. Coordinates of these control points in national coordinate system were determined as coordinates of points in a traverse measured with tachymetry (figure 1 right). The traverse started and ended at the points with the coordinates obtained from the GNSS (global navigation satellite system) measurements. Accuracies of coordinates of control points were very heterogeneous due to difficult field conditions (steep slopes, occluded sights, varying temperature, etc.). Finally, five control points with an appropriate spatial distribution and with the optimal accuracies were used for geo-referencing. The relative accuracy of points in point cloud estimated as a std. deviation of residuals from registration was 1.5 cm, while absolute accuracy estimated as a std. deviation of residual from a geo-referencing was 3.5 cm (Grigilo et all 2014).233743511303000-98915684500Figure 2. Scanning in demanding steep slopes (left), coloured point cloud (right).The point cloud consists of several million points. The points were coloured with images made with digital camera to support visual interpretation of the captured objects (figure 2 right).In 2014 two natural disasters heavily affected the area of Kuzlovec torrent. In the February 2014 ice storm caused a lot of additional falling trees which could change the hydraulic behaviour rapidly. In April 2014, we therefore made an overview of area to verify the changes maturing trees due to ice storm. As expected, a lot of additional trees have fallen from the steep slopes into the watershed, but due to strait torrent channel most of the trunks fell above the riverbed having no immediate affection on fluidity and speed of the stream. The impact will be increased until the years when the trunks due to the thick will fall directly into the watercourse.In the late summer very locally concentrated heavy raining triggered high waters floods and huge changes in watershed (figures 3). Changes in topography encouraged us to make another campaign of TLS measurements, which were done in two days in September 2014 mostly following procedure form 2013, with the exception of using digital camera.DTM creationCaptured TLS data were used to create digital terrain model (DTM) of the torrential channel area. Data were processed using different software packages. We used Riegel’s RiSCAN PRO for manipulating the point cloud, LAStools for DTM production, MeshLab for computing normal vectors for point sets and Matlab to perform the calculation of inventory parameters. The point cloud was translated by 446,000 m and 104,000 m by Y and X 13302-592690023613424699000-2540857250023465487747000Figure 3. Situation in Kuzlovec watershed before (left) and after ice storm and floods (right). Upper images show significant deepness of watershed and many additional falling trees on slopes, while lower ones show accumulation of material. Neighbouring trees can be used as a reference points.axis respectively to avoid large numbers in computations. In order to obtain the appropriate amount of data for processing on common PC, the point cloud was divided in 16 sectors composed of the scans from 2 to 4 neighbouring scanning stations. Each sector was exported as LAS file and processed within LAStools. Lasground tool was used for bare-earth extraction. Within lasground we used extra_fine parameter to intensify the search for initial ground points. We also used step parameter with value 0.5, spike parameter with threshold 0.1 m at which spikes get removed and offset 0.1 to which points above the current ground estimate get included. The values of the parameters were obtained empirically based upon visual examination of the results. Ground points were imported into RiSCAN PRO and visually inspected. Some non-ground points which occurred mostly on the edges of the sectors were removed manually. Each sector of ground points was filtered using octree filter with increment of 0.02 m. All 16 sector’s ground points were merged into single file and DTM with the 0.05 m grid-cell size was interpolated using las2dem tool within LAStools, as seen at figure 4 (Grigillo et all 2015).-19051651000Figure 4. DTM of Kuzlovec torrent watershed.Furthermore, the point cloud was used for recognition of fallen trees and dead wood, and the results were compared with manual dead wood survey. Large woody debris (LWD) was defined as wood pieces of diameter >10cm and longer >1m. A 595.5 longitudinal meters of the torrential channel was scanned with TLS during the project. Several algorithms were tested for automatic LWD detection based upon geometric and/or radiometric values of the point cloud (figure 5 left). None of the algorithms proved to be efficient when applied to the larger area of the torrent channel. The main problems for the automatic extraction presented varying conditions along the torrent channel such as changes in the terrain shape, different and steep slopes, changing light conditions, laser beam shadows etc. Therefore LWD was extracted manually in RiSCAN PRO. 574 wood pieces that correspond to the LWD definition were found in the torrential channel and on the neighbouring slopes (we limited the LWD extraction to 15 horizontal meters on both sides of the torrent). The complete dead tree volume (LWD) was calculated to be 48.66m3 (Grigillo et all 2015).The lengths and diameters of fallen trees were measured manually for validation purposes. Manual measurements were performed in two smaller test areas along the torrent. A measuring tape was used to measure a diameter, while the length of each tree was calculated as a distance between points measured with tachymetry at both ends of a trunk. Position of a tree was calculated from a tachymetric data as well as from a georeferenced point cloud and was used to match corresponding pairs of trees measured manually and in the point cloud. Additionally, the level of decomposition of fallen trees was estimated by a forestry expert (Grigillo et all 2015).237109018224500-190516319500Figure 5. Extraction of LWD from TLS point cloud and recognized LWD indicated on DTM.Some 20% of the LWD was not recognised using TLS if compared to the manual survey. The amount of LWD was estimated to ~0.1m3 per running meter of the torrential channel, and dead tree volume (~30m3/ha) was compared to the average forest stand in Slovenia (346m3/ha) – LWD found close to the torrential channel represents ~8% of the wooden biomass. (Grigillo et all 2015).Hydraulic analyzesModelling the movement of water flows on the basis of the DTM was carried out by using two hydraulic models. For 2D hydraulic analysis of water flow, we used a computer program Flo-2D (O'Brien 2009). Flo-2D is a two-dimensional model, designed to calculate the movement of water and mass flows. Other mathematical model Tis ISAT allows modelling of water flow with free surface method SPH (Smoothed Particle hydrodynamics) in both two and three dimensions. The model was developed at the Chair of Fluid Mechanics with Laboratory as the first SPH model developed in Slovenia. Both model modes, 2D (Petkov?ek et al 2010; Petkov?ek et al 2010b) and 3D (D?ebo et al 2013), were tested in a number of laboratory cases. 3D model used in this paper has been tested on real topographic data (D?ebo et al 2014). SPH is a relatively new Lagrangian method, which constructed area tile by means of moving mass particles and by its nature is very suitable tiles for modelling the rapid changes in water levels, typical of steep torrential channel. The method was developed in 1977 (Gingold and Monaghan 1977; Lucy 1977) and was intended for modelling compressible gases in astrophysics. Through the years the method have been developed and updated and became popular for use in other areas. In 1994 it was proposed to model the flow of water free surface (Monaghan 1994) and within a relatively short time became very popular for such purposes (Grigillo et al 2014).In our application Flo-2D and Tis ISAT applications were used to study the impact of large tree residues (thicks) on water flow in the steep torrential river bed Kuzlovec. Calculation of changes in water flow with the program Flo-2D in two dimensions is carried out with a computational algorithm with respect to computer-based approximate finite difference methods using the numerical solution of the Navier-Stokes equations. The model simulates the progression of the flood wave by a system of square grid elements, which presents the topography of the surface in the same manner as observed in the DTM of the terrain. The program takes into account the movement of water flow within 8 directions just as it was reflected in the D8 algorithm for the account of the direction of the depositing surface water flows. At the hydraulic calculation the roughness of the surface have to be considered by the Manning roughness coefficient for shallow surface flow and numerical parameters, which regulates the deviations in maintaining the mass balance of water in the model (Grigillo et al 2014).In a 3D mathematical model Tis ISAT the size of the particles used was d = 0.05 m, the volume of water in the simulated range V = 2.88 m 3, the parameter of viscosity between the particles and the wall bvis = 12:01. Parameter bvis in the SPH method has same function as a friction coefficient (e.g. Manning’s friction coefficient ng) in conventional methods. More information about the SPH method are available in the literature (Violeau 2012; Liu and Liu 2003).-304802413000-311155485700-3296217960600Razlika DTM po floodsFigure 6. DTM changes, mostly caused by floods in summer 2014. The upper image shows DTM based on TLS survey in 2013, the second one DTM based on 2014 survey and the bottom one indicates changes in DTM, where green colour indicate loss and red accumulation of materialCartographic presentationsDifferent cartographic presentations have been created for many purposes, to picture the situation for better general overview of test area (figures 6), for easier recognition of specific forms (figures 7) and, primarily, for presenting results of hydraulic analyses.-1905-133794500Figure 7. Shaded DTM with (right) and without LWD (left)Figure 8 shows the results of the flow rate modelling using hydraulic model Flo-2D. DTM, resampled to cell size of 10 cm was used, which allows considering the tree trunks, as part geometry torrential channel. 2667014287500Figure 8. Modelled flow velocity of torrent Kuzlovec in the program Flo-2D: based on the DTM without tree trunks (left), DTM considering tree trunks (right).For a shallow stream in a torrential river bed the value of Manning roughness coefficient 0.05 was used. The small size of computational cells causes an extension of computational time, in our case modelling an event, real-time lasting 6 hours and in which the peak of the flow rises to 100 l/s, needs 52 hours. The flow rates in both models (with and without considering tree trunks in the channel) were mostly below 2 m/s. The results of the modelling indicate visible impact of tree trunks on the dynamics of the flow velocity in the bed of the torrent. Tree trunks create dams and slowing the flow of water on the upstream side; while in the overflow over t he dams the water flow increases again (the flow rate in this case exceeds 3 m/s).Figure 9 shows flow velocity in the short section of Kuzlovec torrent, modelled by model Tis ISAT. As we expected, modelling results without considering tree trunks show a slightly higher flow rates than the modelling results carried out on the topography with timbers. Even the flow of water due to the impact of tree trunks chose slightly different path.-1905-171259500Figure 9. Modelled flow velocity of torrent Kuzlovec using program Tis ISAT: based on the DTM without tree trunks (left), DTM considering tree trunks (right).ConclusionA significant part of torrents in Slovenia and other countries with comparable forest cover takes place in the areas where dead wood (fallen trees due to various abiotic phenomena such as avalanche, breakage, ice-storm, coarse woody debris that remain in the forest due to forest disturbance and the when they are not removed after forest works) quickly ends up in the trenches of erosion and flood channels. This occurs profession question of how dead wood as part of torrential channel affects the flow conditions in the river. To analyse the impact of dead wood we selected a small area of torrential Kuzlovec Grada??ica in the basin (south-west of Ljubljana) and performed field measurements before the big ice storm in February 2014 and heavy floods in August 2014 and after both events. In this paper some possible results and analyses of huge amount of data, captured with terrestrial laser scanning technology.AcknowledgementThe research work was done within the framework of European Alpine Space project SedAlp "Sediment management in Alpine basins: integrating sediment continuum, risk mitigation and hydropower" (2012-2015) and was partially founded by the Slovenian Research Agency.ReferencesBienert, A, Scheller, S, Keane E, Mullooly, G, Mohan, F. 2006: Application of terrestrial laser scanners for the determination of forest inventory parameters. Image engineering and vision metrology. Dresden.D?ebo, E., ?agar, D., ?etina, M., Petkov?ek, G. 2013: Reducing the computational time of the SPH method with a coupled 2-D/3-D SPH model. Journal of Mechanical Engineering 59-10. Ljubljana. DOI:10.5545/sv-jme.2013.944D?ebo, E., ?agar, D., Krzyk, M., ?etina, M., Petkov?ek, G. 2014: Different ways of defining wall shear in SPH simulations of a dam-break wave. Journal of Hydraulic Research – sprejeto v objavo. Abingdon. DOI:10.1080/00221686.2013.879611Gingold, R.A., Monaghan, J.J. 1977: Smoothed particle hydrodynamics, Theory and application to non-spherical stars. Monthly Notices of the Royal Astronomical Society 181. London. Grigillo, Dejan, Rusjan, Simon, Vre?ko, Anja, D?ebo, Elvira, Kozmus Trajkovski, Klemen, Urban?i?, Tilen, Petrovi?, Du?an, Miko?, Matja?. 2014. Digitalni model reliefa struge hudournika Kuzlovec in matemati?no modeliranje toka vode. V: CIGLI?, Rok (ur.), PERKO, Drago (ur.), ZORN, Matija (ur.). Digitalni prostor, (GIS v Sloveniji, ISSN 1855-4954, 12). Ljubljana: Zalo?ba ZRC, 2014, str. 35-42, ilustr.)Grigillo, Dejan, Vre?ko, Anja, Miko?, Matja?, Gvozdanovi?, Toma?, An?ur, Andreja, Vezo?nik, Rok, Petrovi?, Du?an. 2015. Determination of large wood accumulation in a steep forested torrent using laser scanning. V: LOLLINO, Giorgio (ur.). Engineering Geology for Society and Territory - Volume 3 : River Basins, Reservoir Sedimentation and Water Resources. Cham: Springer International Publishing, 2015, str. 127-130, ilustr. , A. 2011: Nove metode za obdelavo podatkov letalskega laserskega skenerja za monitoring gozdnih ekosistemov. Doktorska disertacija, Fakulteta za gradbeni?tvo in geodezijo Univerze v Ljubljani. Ljubljana.Kobler, A., Pfeifer, N., Ogrinc, P., Todorovski, L., O?tir, K., D?eroski, S. 2007: Repetitive interpolation: A robust algorithm for DTM generation from Aerial Laser Scanner Data in forested terrain. Remote Sensing of Environment 108. Elsevier.Liu, G.R., Liu, M.B. 2003: Smoothed Particle Hydrodynamics, A mesh free particle method. Singapore.Lucy, L.B. 1977: A numerical approach to the testing of the fission hypothesis. The Astronomical Journal 82. New York. DOI: 10.1086/112164Monaghan, J.J. 1994: Simulating Free Surface Flows with SPH. Journal of Computational Physics 110-2. Philadelphia. DOI: 10.1006/jcph.1994.1034Moskal, M., Zheng, G. 2012: Retrieving forest inventory variables with terrestrial laser scanning (TLS) in urban heterogeneous forest. Remote Sensing 4. Basel.O?Brien, J. S. 2009: Flo-2D, Reference Manual, Version 2009. Nutrioso.Petkov?ek, G., D?ebo, E., ?etina, M., ?agar, D. 2010a: Simulacija laboratorijske poru?itve pregrade po metodi SPH z analizo modela trenja. Zbornik del Kuhljevi dnevi. Ljubljana.Petkov?ek, G., D?ebo, E., ?etina, M., ?agar, D. 2010b: Application of Non-Discrete Boundaries with Friction to Smoothed Particle Hydrodynamics. Journal of Mechanical Engineering 56-5. Ljubljana. DOI: , M., Aschoff, T., Spiecker, H. and Thies, M. 2003: Automatic determination of forest inventory parameters using terrestrial laserscanning. ScandLaser Scientific Workshop on Airborne Laser Scanning of Forests. Ume?.Thies, M., Pfeifer, N., Winterhalder, D., Gorte, B.G.H. 2004: Three-dimensional reconstruction of stems for assessment of taper, sweep and lean based on laser scanning of standing trees. Scandinavian Journal of Forest Research 19-6. London.Violeau, D. 2012: Fluid Mechanics and the SPH Method, Theory and Applications. Oxford. ................
................
In order to avoid copyright disputes, this page is only a partial summary.
To fulfill the demand for quickly locating and searching documents.
It is intelligent file search solution for home and business.
Related searches
- the importance of your name
- words with the following letters scrabble
- what is the purpose of your existence
- letters of friendship that touch the heart
- grade your paper for free
- edit your paper online free
- calculate value of your paper savings bonds
- title for research paper generator
- calculate the value of your savings bond
- title of paper examples
- title for research paper examples
- calculated value of your paper bonds