Skip to Main content Skip to Navigation
Conference papers

Sensitivity analysis of simulated hydrographic LIDAR waveforms according to sensor and water parameters variability

Abstract : LiDAR (Light Detection and Ranging) can be used as a ranging system using electromagnetic waves in the optical domain. LiDAR airborne or satellite sensors are promising techniques for river bathymetry and water surface altimetry considering its potential accuracy, its high spatial density and resolution. When considering physics of LiDAR, many factors coming from sensor characteristics and optical phenomena interacting during the signal transfer at media interfaces are conditioning LiDAR signals, i.e. waveforms. Waveforms are the registered signals in LiDAR from which, hydrological variables, as surface water altimetry, bathymetry, are retrieved. To move forward in the LIDAR airborne and satellite sensors capabilities for river monitoring, a modelling tool extending different existing radiative transfer models has been developed. This model allows the simulation of LiDAR data from a set of instrumental parameters and a representative collection of fluvial target for laser beam (water turbidity, river bottom reflectivity, etc). Due to the large numbers of parameters in the modelling and their natural range of variation, one of the questions is which of those parameters have the most impact, and those who have a negligible effect on returned waveform? And consequently, which of them most impact the accuracy of the retrieved hydrological variables from waveforms ? In order to assess the robustness of the proposed model, to look for parsimony and to identify the key sensor parameters, a study of the model sensitivity to different media characteristics (surface reflection, absorption, scattering, bottom Albedo, etc) and sensor parameters (wavelength, pulse width, transmitted power, etc) is performed. Due to the complexity of the LiDAR waveform modelling, we used a sensitivity analysis method based on variance decomposition (Sobol) and Latin hypercube random sampling design in factor variation domains. As output model are waveforms, i.e. temporal signal, we used the specific Sobol framework proposed by Lamboni (2010). Results are the synthetic Sobol indices showing which factors are highly conditioning waveforms and others that can be neglected in the modelling. By pursuing the sensitivity analysis up to the waveform inversion (Gaussian fitting) to retrieve bathymetry, sensor and river parameters which are the more impacting bathymetry accuracy can also been hierarchically identified in this framework.
Complete list of metadata
Contributor : Migration Irstea Publications <>
Submitted on : Friday, May 15, 2020 - 5:52:27 PM
Last modification on : Friday, April 16, 2021 - 3:29:24 AM


  • HAL Id : hal-02594028, version 1
  • IRSTEA : PUB00030118


H. Abdallah, Jean-Stéphane Bailly, N. Baghdadi, Nathalie Saint Geours. Sensitivity analysis of simulated hydrographic LIDAR waveforms according to sensor and water parameters variability. AGU 2010, Dec 2010, San Francisco, United States. pp.37. ⟨hal-02594028⟩



Record views