EXPERIMENTAL ADVANCED AIRBORNE RESEARCH LIDAR B ...
Extended Abstract
th
11 ISE 2016, Melbourne, Australia
EXPERIMENTAL ADVANCED AIRBORNE RESEARCH LIDAR B: REVEALING THE RIVER BATHYMETRY DANIELE TONINA Center for Ecohydraulics Research, University of Idaho Boise, 83702, Idaho, USA JAMES A. MCKEAN Rocky Mountain Research Station, US Forest Service Boise, 83702, Idaho, USA ROHAN M. BENJANKAR Center for Ecohydraulics Research, University of Idaho Boise, 83702, Idaho, USA The Experimental Advanced Airborne Research Lidar B (EAARL-B) is a new airborne aquatic-terrestrial sensor that allows simultaneous high resolution surveying in both environments over spatial domains of up to several hundred kilometers of stream length. Here we compare detailed ground-survey bathymetry with EAARL-B survey derived bathymetric data to quantify the performance of EAARL-B to describe river morphology. We then test whether EAARL-B survey derived bathymetries can support two-dimensional hydraulic models to study flow hydraulics and aquatic habitat modeling. Preliminary results show that maps derived from EAARL-B survey can be used to quantify micro-habitat quality at the meter scale.
1
INTRODUCTION
Three-dimensional (3D) models are too complex for routine use in riverine systems because of their high requirements of computational time and hydraulic and topographical information, which limit their application to specialized cases (e.g., flow around objects such as bridge piers and turbines) with steady state flows [1]. However, one-dimensional (1D) models cannot capture flow structures in complex meandering streams at biological meaningful scales. Thus, two-dimensional (2D) hydraulic modeling is becoming the most common tool to study riverine systems [1]. Two-dimensional modeling has the advantages of providing 1) planimetric maps of flow characteristics, like vertically averaged velocities, bottom shear stresses, water depths and other derived quantities, e.g., stream energy gradient, vorticity and circulation [2, 3], 2) vector longitudinal and transversal components, and 3) transversal structure of flow features and water surface elevation. Twodimensional models such as the multi-dimensional surface water model system (MD-SWMS) developed by the US Geological Survey [4] with secondary flows are fundamental tools in studying streambed evolution, erosion and deposition and aquatic habitats [5-11]. However, their application in riverine systems has been limited to short reaches representative of the entire system because of the lack of continuous survey data [12]. This limitation could be overcome by the EAARL-B sensor, which in contrast to discrete-return near-infrared lidars has the capacity to map in one integrated mission both the aquatic and terrestrial topographies. The EAARL-B has been developed between 2011 and 2013 and it flew the main stem of the Lemhi River on October 26th and 27th 2013. Here, we show preliminary results on the capability of EAARL-B to detect streambed bathymetry and to support 2D numerical modeling. 2
METHODS
At the end of August 2013, we surveyed a 250m long reach of the Lemhi River at a 0.6 by 0.6 meter resolution, This high resolution field survey is rare in most application [1, 6, 13] because it is highly time consuming (Figure 1). The uncertainty on the ground-survey points is 0.015m. The reach is morphologically complex with pools, riffles, submerged and emerged bars and runs. The survey included (1) a sub-meter resolution streambed and floodplain topography, (2) identification of streambed material (e.g., gravel, sand and pebbles) patches, (3) pebble count for each patch, (3) identification of the morphological units (e.g., pools, riffles, and bars), (4) discharge measurement, (5) water surface measurement along the center line and at the edge of the water and (6) velocity measurements at different locations along the reach. The EAARL-B system surveyed the same reach in
October 2013. Both surveys were conducted during the summer-fall low flow and there had been no high discharges between the surveys. We developed and calibrated MD-SWMS at low flow with EAARL-B-generated and ground-surveyed topographies (Table 1). The root mean square error, RMSE, between predicted and measured water surface elevations and velocity are within those reported in literature and expected in such complex systems [1, 12]. We compared velocity and depth distributions between models with EAARL-B-generated and ground-surveyed topographies and analyzed the results for EAARL-B applications in evaluating aquatic habitat quality. Table 1. Root mean square error (RMSE) for water surface elevation and velocity between numerically predicted and measured values. Predicted Weighted Usable Area (WUA) values of aquatic habitat are also reported for both maps. RMSE Water Surface WUA RMSE Velocity Calibration Elevation Calibration 2D model EAARL supported 1070.15 0.05 0.27 2D model Ground supported 1214.06 0.027 0.16
Figure 1: Study site: (a) aerial photograph (b) field survey map, (c) EAARL-B derived map and (d) DEM of difference between field and EAARL-derived map. 3
RESULTS
The comparison between ground survey and EAARL-B-generated digital elevation models, DEM, shows good agreement (Figure 1). The EAARL-B survey captures all the main features of the river and its floodplain. The lidar-based fine scale topography is more irregular than that of the field survey (c.f., Figure 1b and c), as the EAARL-B system has about 50 cm horizontal uncertainty and a vertically RMSE within the channel of 0.13 m, which is typical of all airborne lidars. The difference between field and lidar survey increases to 0.3 m at the steep banks, where topographical changes are abrupt. Vertical banks are difficult to survey with a scanning airborne system, regardless of sensor precision.
Figure 2: Comparison of numerical modeling results; first row predicted depth, center row velocity and last row predicted aquatic habitat for Chinook salmon spawning based on the biological model reported by Carnie et al. [14]. Hydraulic properties predicted with numerical models supported by EAARL-B-generated and ground survey maps also are very similar (Figure 2). Aquatic habitat suitability values were quantified by velocity and depth predicted with the numerical model supported by EAARL-B-generated and ground survey maps. The habitat suitability distributions present a good match with nearly 50% of the streambed having identical quality class distribution and with 70% of the streambed areas within ±1 class.
4
CONCLUSION
Our analysis shows that the EAARL-B system is capable of surveying both terrestrial and submerged topographies with an accuracy adequate to support hydrodynamic modeling. This allows studying riverine systems with unprecedented details at the network scale to understand river morphology, sediment transport and how aquatic organisms use streams and interact with their physical environment. REFERENCES [1] Tonina, D. and K. Jorde, "Hydraulic modeling approaches for ecohydraulic studies: 3D, 2D, 1D and non-numerical models", in Ecohydraulics: An integrated approach, I. Maddock, P.J. Wood, A. Harby, and P. Kemp, Editors, WileyBlackwell (2013), p. 31-66. [2] Nestler, J.M., R.A. Goodwin, D.L. Smith, J.J. Anderson, and S. Li, "Optimum fish passage and guidance designs are based in the hydrogeomorphology of natural rivers", River Research and Applications, Vol. 24, No. 2, (2008), pp 148168. [3] Shen, Y. and P. Diplas, "Application of two- and three-dimensional computational fluid dynamics models to complex ecological stream flows", Journal of Hydrology, Vol. 348, (2008), pp 195-214. [4] McDonald, R.R., J.M. Nelson, and J.P. Bennett, Multi-dimensional surface-water modeling system user’s guide. 2005, U.S. Geological Survey. p. pp. 136. [5] Pasternack, G.B., "2D Modeling and Ecohydraulic Analysis", (2011) University of California at Davis. [6] Pasternack, G.B. and A. Senter, 21st Century instream flow assessment framework for mountain streams. 2011, Public Interest Energy Research (PIER), California Energy Commission. p. 399. [7] Pasternack, G.B., C.L. Wang, and J.E. Merz, "Application of a 2d hydrodynamic model to design of reach-scale spawning gravel replenishment on the Mokelumne river, California", River Research and Applications, Vol. 20, No. 2, (2004), pp 205-225. [8] Maturana, O., D. Tonina, J.A. Mckean, J.M. Buffington, C.H. Luce, and D. Caamaño, "Modeling the effects of pulsed versus chronic sand inputs on salmonid spawning habitat in a low-gradient gravel-bed river", Earth Surface Processes and Landforms, (2013). [9] McKean, J.A., D. Nagel, D. Tonina, P. Bailey, C.W. Wright, C. Bohn, and A. Nayegandhi, "Remote sensing of channels and riparian zones with a narrow-beam aquatic-terrestrial LIDAR", Remote Sensing, Vol. 1, (2009), pp 1065-1096. [10] McKean, J.A. and D. Tonina, "Bed stability in unconfined gravel-bed mountain streams: With implications for salmon spawning viability in future climates", Journal of Geophysical Research: Earth Surface, Vol. 118, (2013), pp 1-14. [11] McKean, J.A., D. Tonina, C. Bohn, and C.W. Wright, "Effects of bathymetric lidar errors on flow properties predicted with a multi-dimensional hydraulic model", Journal of Geophysical Research: Earth Surface, (2014). [12] Tonina, D., J.A. McKean, C. Tang, and P. Goodwin. "New tools for aquatic habitat modeling". 34th IAHR World Congress 2011, Brisbane, Australia, (2011), pp 3137-3144. [13] Hassan, M.A., D. Tonina, and T.H. Buxton, "Does small-bodied salmon spawning activity enhance streambed mobility?" Water Resources Research, Vol. 51, (2015). [14] Carnie, R., D. Tonina, J.A. McKean, and D.J. Isaak, "Habitat connectivity as a metric for aquatic microhabitat quality: Application to Chinook salmon spawning habitat", Ecohydrology, (2015), doi:10.1002/eco.1696.
th
11 ISE 2016, Melbourne, Australia
EXPERIMENTAL ADVANCED AIRBORNE RESEARCH LIDAR B: REVEALING THE RIVER BATHYMETRY DANIELE TONINA Center for Ecohydraulics Research, University of Idaho Boise, 83702, Idaho, USA JAMES A. MCKEAN Rocky Mountain Research Station, US Forest Service Boise, 83702, Idaho, USA ROHAN M. BENJANKAR Center for Ecohydraulics Research, University of Idaho Boise, 83702, Idaho, USA The Experimental Advanced Airborne Research Lidar B (EAARL-B) is a new airborne aquatic-terrestrial sensor that allows simultaneous high resolution surveying in both environments over spatial domains of up to several hundred kilometers of stream length. Here we compare detailed ground-survey bathymetry with EAARL-B survey derived bathymetric data to quantify the performance of EAARL-B to describe river morphology. We then test whether EAARL-B survey derived bathymetries can support two-dimensional hydraulic models to study flow hydraulics and aquatic habitat modeling. Preliminary results show that maps derived from EAARL-B survey can be used to quantify micro-habitat quality at the meter scale.
1
INTRODUCTION
Three-dimensional (3D) models are too complex for routine use in riverine systems because of their high requirements of computational time and hydraulic and topographical information, which limit their application to specialized cases (e.g., flow around objects such as bridge piers and turbines) with steady state flows [1]. However, one-dimensional (1D) models cannot capture flow structures in complex meandering streams at biological meaningful scales. Thus, two-dimensional (2D) hydraulic modeling is becoming the most common tool to study riverine systems [1]. Two-dimensional modeling has the advantages of providing 1) planimetric maps of flow characteristics, like vertically averaged velocities, bottom shear stresses, water depths and other derived quantities, e.g., stream energy gradient, vorticity and circulation [2, 3], 2) vector longitudinal and transversal components, and 3) transversal structure of flow features and water surface elevation. Twodimensional models such as the multi-dimensional surface water model system (MD-SWMS) developed by the US Geological Survey [4] with secondary flows are fundamental tools in studying streambed evolution, erosion and deposition and aquatic habitats [5-11]. However, their application in riverine systems has been limited to short reaches representative of the entire system because of the lack of continuous survey data [12]. This limitation could be overcome by the EAARL-B sensor, which in contrast to discrete-return near-infrared lidars has the capacity to map in one integrated mission both the aquatic and terrestrial topographies. The EAARL-B has been developed between 2011 and 2013 and it flew the main stem of the Lemhi River on October 26th and 27th 2013. Here, we show preliminary results on the capability of EAARL-B to detect streambed bathymetry and to support 2D numerical modeling. 2
METHODS
At the end of August 2013, we surveyed a 250m long reach of the Lemhi River at a 0.6 by 0.6 meter resolution, This high resolution field survey is rare in most application [1, 6, 13] because it is highly time consuming (Figure 1). The uncertainty on the ground-survey points is 0.015m. The reach is morphologically complex with pools, riffles, submerged and emerged bars and runs. The survey included (1) a sub-meter resolution streambed and floodplain topography, (2) identification of streambed material (e.g., gravel, sand and pebbles) patches, (3) pebble count for each patch, (3) identification of the morphological units (e.g., pools, riffles, and bars), (4) discharge measurement, (5) water surface measurement along the center line and at the edge of the water and (6) velocity measurements at different locations along the reach. The EAARL-B system surveyed the same reach in
October 2013. Both surveys were conducted during the summer-fall low flow and there had been no high discharges between the surveys. We developed and calibrated MD-SWMS at low flow with EAARL-B-generated and ground-surveyed topographies (Table 1). The root mean square error, RMSE, between predicted and measured water surface elevations and velocity are within those reported in literature and expected in such complex systems [1, 12]. We compared velocity and depth distributions between models with EAARL-B-generated and ground-surveyed topographies and analyzed the results for EAARL-B applications in evaluating aquatic habitat quality. Table 1. Root mean square error (RMSE) for water surface elevation and velocity between numerically predicted and measured values. Predicted Weighted Usable Area (WUA) values of aquatic habitat are also reported for both maps. RMSE Water Surface WUA RMSE Velocity Calibration Elevation Calibration 2D model EAARL supported 1070.15 0.05 0.27 2D model Ground supported 1214.06 0.027 0.16
Figure 1: Study site: (a) aerial photograph (b) field survey map, (c) EAARL-B derived map and (d) DEM of difference between field and EAARL-derived map. 3
RESULTS
The comparison between ground survey and EAARL-B-generated digital elevation models, DEM, shows good agreement (Figure 1). The EAARL-B survey captures all the main features of the river and its floodplain. The lidar-based fine scale topography is more irregular than that of the field survey (c.f., Figure 1b and c), as the EAARL-B system has about 50 cm horizontal uncertainty and a vertically RMSE within the channel of 0.13 m, which is typical of all airborne lidars. The difference between field and lidar survey increases to 0.3 m at the steep banks, where topographical changes are abrupt. Vertical banks are difficult to survey with a scanning airborne system, regardless of sensor precision.
Figure 2: Comparison of numerical modeling results; first row predicted depth, center row velocity and last row predicted aquatic habitat for Chinook salmon spawning based on the biological model reported by Carnie et al. [14]. Hydraulic properties predicted with numerical models supported by EAARL-B-generated and ground survey maps also are very similar (Figure 2). Aquatic habitat suitability values were quantified by velocity and depth predicted with the numerical model supported by EAARL-B-generated and ground survey maps. The habitat suitability distributions present a good match with nearly 50% of the streambed having identical quality class distribution and with 70% of the streambed areas within ±1 class.
4
CONCLUSION
Our analysis shows that the EAARL-B system is capable of surveying both terrestrial and submerged topographies with an accuracy adequate to support hydrodynamic modeling. This allows studying riverine systems with unprecedented details at the network scale to understand river morphology, sediment transport and how aquatic organisms use streams and interact with their physical environment. REFERENCES [1] Tonina, D. and K. Jorde, "Hydraulic modeling approaches for ecohydraulic studies: 3D, 2D, 1D and non-numerical models", in Ecohydraulics: An integrated approach, I. Maddock, P.J. Wood, A. Harby, and P. Kemp, Editors, WileyBlackwell (2013), p. 31-66. [2] Nestler, J.M., R.A. Goodwin, D.L. Smith, J.J. Anderson, and S. Li, "Optimum fish passage and guidance designs are based in the hydrogeomorphology of natural rivers", River Research and Applications, Vol. 24, No. 2, (2008), pp 148168. [3] Shen, Y. and P. Diplas, "Application of two- and three-dimensional computational fluid dynamics models to complex ecological stream flows", Journal of Hydrology, Vol. 348, (2008), pp 195-214. [4] McDonald, R.R., J.M. Nelson, and J.P. Bennett, Multi-dimensional surface-water modeling system user’s guide. 2005, U.S. Geological Survey. p. pp. 136. [5] Pasternack, G.B., "2D Modeling and Ecohydraulic Analysis", (2011) University of California at Davis. [6] Pasternack, G.B. and A. Senter, 21st Century instream flow assessment framework for mountain streams. 2011, Public Interest Energy Research (PIER), California Energy Commission. p. 399. [7] Pasternack, G.B., C.L. Wang, and J.E. Merz, "Application of a 2d hydrodynamic model to design of reach-scale spawning gravel replenishment on the Mokelumne river, California", River Research and Applications, Vol. 20, No. 2, (2004), pp 205-225. [8] Maturana, O., D. Tonina, J.A. Mckean, J.M. Buffington, C.H. Luce, and D. Caamaño, "Modeling the effects of pulsed versus chronic sand inputs on salmonid spawning habitat in a low-gradient gravel-bed river", Earth Surface Processes and Landforms, (2013). [9] McKean, J.A., D. Nagel, D. Tonina, P. Bailey, C.W. Wright, C. Bohn, and A. Nayegandhi, "Remote sensing of channels and riparian zones with a narrow-beam aquatic-terrestrial LIDAR", Remote Sensing, Vol. 1, (2009), pp 1065-1096. [10] McKean, J.A. and D. Tonina, "Bed stability in unconfined gravel-bed mountain streams: With implications for salmon spawning viability in future climates", Journal of Geophysical Research: Earth Surface, Vol. 118, (2013), pp 1-14. [11] McKean, J.A., D. Tonina, C. Bohn, and C.W. Wright, "Effects of bathymetric lidar errors on flow properties predicted with a multi-dimensional hydraulic model", Journal of Geophysical Research: Earth Surface, (2014). [12] Tonina, D., J.A. McKean, C. Tang, and P. Goodwin. "New tools for aquatic habitat modeling". 34th IAHR World Congress 2011, Brisbane, Australia, (2011), pp 3137-3144. [13] Hassan, M.A., D. Tonina, and T.H. Buxton, "Does small-bodied salmon spawning activity enhance streambed mobility?" Water Resources Research, Vol. 51, (2015). [14] Carnie, R., D. Tonina, J.A. McKean, and D.J. Isaak, "Habitat connectivity as a metric for aquatic microhabitat quality: Application to Chinook salmon spawning habitat", Ecohydrology, (2015), doi:10.1002/eco.1696.
