Spectral-ALS data fusion for different roughness parameterizations of forested floodplains

G. Forzieri, L. Guarnieri, Enrique Vivoni, F. Castelli, F. Preti

Research output: Contribution to journalArticle

19 Citations (Scopus)

Abstract

Natural river floodplains and adjacent wetlands grow typically a diverse and heterogeneous combination of herbs, shrubs and trees, which play an essential role in determining the total flow resistance. Hydrodynamic effects of trees in forested floodplains can provide the majority of flow resistance during flood events. Nevertheless, ground-based techniques to acquire vegetation parameters are expensive and difficult to apply over long reaches. This paper presents a novel method of automated roughness parameterization of riparian woody vegetation by fusion of Quickbird multi-spectral image with airborne laser scanning (ALS) data. The data fusion approach includes: individual tree detection and estimation of vegetation metrics from light detection and ranging (LiDAR) data, assessment of predictive models for the vegetation parameters and spatial mapping of the vegetation parameters for the forest plants in the riparian corridor. The proposed method focuses on estimation of plant density (d), crown diameters (D C), tree height (h), stem diameter (D S), crown base height (cbh) and leaf area index (LAI). The procedure is tested along a 14-km reach of the Sieve River (Tuscany, Italy) characterized by high woody plant density. Due to the complex study area, the data fusion approach explains with variable reliability the local vegetation properties (R 2(D C)=0.14, R 2(h)=0.84, R 2(D S)=0.25, R 2(cbh)=0.66). The generated structural parameter maps represent spatially explicit data layers that can be used as inputs to hydrodynamic models used to analyse flow resistance effects in different submergence conditions of vegetation. A simple flow resistance model was applied over a test area comparing the results of the proposed method and a traditional ground-based approach. The modelling results showed that the new method is able to provide accurate output data to describe the interaction between water levels and bio-mechanical characteristics of vegetation. The proposed methodology provides a fast, repeatable and accurate way of obtaining floodplain roughness, which enables regular updating of vegetation characteristics.

Original languageEnglish (US)
Pages (from-to)826-840
Number of pages15
JournalRiver Research and Applications
Volume27
Issue number7
DOIs
StatePublished - Sep 2011

Fingerprint

Data fusion
Parameterization
roughness
floodplain
parameterization
laser
Surface roughness
Scanning
Lasers
vegetation
Hydrodynamics
Rivers
hydrodynamics
QuickBird
multispectral image
Sieves
submergence
Wetlands
woody plant
Water levels

Keywords

  • Flow resistance
  • Hydraulic roughness
  • LiDAR
  • Quickbird
  • Remote sensing
  • Riparian corridor
  • Sieve River
  • Woody vegetation

ASJC Scopus subject areas

  • Environmental Science(all)
  • Environmental Chemistry
  • Water Science and Technology

Cite this

Spectral-ALS data fusion for different roughness parameterizations of forested floodplains. / Forzieri, G.; Guarnieri, L.; Vivoni, Enrique; Castelli, F.; Preti, F.

In: River Research and Applications, Vol. 27, No. 7, 09.2011, p. 826-840.

Research output: Contribution to journalArticle

Forzieri, G. ; Guarnieri, L. ; Vivoni, Enrique ; Castelli, F. ; Preti, F. / Spectral-ALS data fusion for different roughness parameterizations of forested floodplains. In: River Research and Applications. 2011 ; Vol. 27, No. 7. pp. 826-840.
@article{700021b34c4749098941fa660f29dc1e,
title = "Spectral-ALS data fusion for different roughness parameterizations of forested floodplains",
abstract = "Natural river floodplains and adjacent wetlands grow typically a diverse and heterogeneous combination of herbs, shrubs and trees, which play an essential role in determining the total flow resistance. Hydrodynamic effects of trees in forested floodplains can provide the majority of flow resistance during flood events. Nevertheless, ground-based techniques to acquire vegetation parameters are expensive and difficult to apply over long reaches. This paper presents a novel method of automated roughness parameterization of riparian woody vegetation by fusion of Quickbird multi-spectral image with airborne laser scanning (ALS) data. The data fusion approach includes: individual tree detection and estimation of vegetation metrics from light detection and ranging (LiDAR) data, assessment of predictive models for the vegetation parameters and spatial mapping of the vegetation parameters for the forest plants in the riparian corridor. The proposed method focuses on estimation of plant density (d), crown diameters (D C), tree height (h), stem diameter (D S), crown base height (cbh) and leaf area index (LAI). The procedure is tested along a 14-km reach of the Sieve River (Tuscany, Italy) characterized by high woody plant density. Due to the complex study area, the data fusion approach explains with variable reliability the local vegetation properties (R 2(D C)=0.14, R 2(h)=0.84, R 2(D S)=0.25, R 2(cbh)=0.66). The generated structural parameter maps represent spatially explicit data layers that can be used as inputs to hydrodynamic models used to analyse flow resistance effects in different submergence conditions of vegetation. A simple flow resistance model was applied over a test area comparing the results of the proposed method and a traditional ground-based approach. The modelling results showed that the new method is able to provide accurate output data to describe the interaction between water levels and bio-mechanical characteristics of vegetation. The proposed methodology provides a fast, repeatable and accurate way of obtaining floodplain roughness, which enables regular updating of vegetation characteristics.",
keywords = "Flow resistance, Hydraulic roughness, LiDAR, Quickbird, Remote sensing, Riparian corridor, Sieve River, Woody vegetation",
author = "G. Forzieri and L. Guarnieri and Enrique Vivoni and F. Castelli and F. Preti",
year = "2011",
month = "9",
doi = "10.1002/rra.1398",
language = "English (US)",
volume = "27",
pages = "826--840",
journal = "River Research and Applications",
issn = "1535-1459",
publisher = "John Wiley and Sons Ltd",
number = "7",

}

TY - JOUR

T1 - Spectral-ALS data fusion for different roughness parameterizations of forested floodplains

AU - Forzieri, G.

AU - Guarnieri, L.

AU - Vivoni, Enrique

AU - Castelli, F.

AU - Preti, F.

PY - 2011/9

Y1 - 2011/9

N2 - Natural river floodplains and adjacent wetlands grow typically a diverse and heterogeneous combination of herbs, shrubs and trees, which play an essential role in determining the total flow resistance. Hydrodynamic effects of trees in forested floodplains can provide the majority of flow resistance during flood events. Nevertheless, ground-based techniques to acquire vegetation parameters are expensive and difficult to apply over long reaches. This paper presents a novel method of automated roughness parameterization of riparian woody vegetation by fusion of Quickbird multi-spectral image with airborne laser scanning (ALS) data. The data fusion approach includes: individual tree detection and estimation of vegetation metrics from light detection and ranging (LiDAR) data, assessment of predictive models for the vegetation parameters and spatial mapping of the vegetation parameters for the forest plants in the riparian corridor. The proposed method focuses on estimation of plant density (d), crown diameters (D C), tree height (h), stem diameter (D S), crown base height (cbh) and leaf area index (LAI). The procedure is tested along a 14-km reach of the Sieve River (Tuscany, Italy) characterized by high woody plant density. Due to the complex study area, the data fusion approach explains with variable reliability the local vegetation properties (R 2(D C)=0.14, R 2(h)=0.84, R 2(D S)=0.25, R 2(cbh)=0.66). The generated structural parameter maps represent spatially explicit data layers that can be used as inputs to hydrodynamic models used to analyse flow resistance effects in different submergence conditions of vegetation. A simple flow resistance model was applied over a test area comparing the results of the proposed method and a traditional ground-based approach. The modelling results showed that the new method is able to provide accurate output data to describe the interaction between water levels and bio-mechanical characteristics of vegetation. The proposed methodology provides a fast, repeatable and accurate way of obtaining floodplain roughness, which enables regular updating of vegetation characteristics.

AB - Natural river floodplains and adjacent wetlands grow typically a diverse and heterogeneous combination of herbs, shrubs and trees, which play an essential role in determining the total flow resistance. Hydrodynamic effects of trees in forested floodplains can provide the majority of flow resistance during flood events. Nevertheless, ground-based techniques to acquire vegetation parameters are expensive and difficult to apply over long reaches. This paper presents a novel method of automated roughness parameterization of riparian woody vegetation by fusion of Quickbird multi-spectral image with airborne laser scanning (ALS) data. The data fusion approach includes: individual tree detection and estimation of vegetation metrics from light detection and ranging (LiDAR) data, assessment of predictive models for the vegetation parameters and spatial mapping of the vegetation parameters for the forest plants in the riparian corridor. The proposed method focuses on estimation of plant density (d), crown diameters (D C), tree height (h), stem diameter (D S), crown base height (cbh) and leaf area index (LAI). The procedure is tested along a 14-km reach of the Sieve River (Tuscany, Italy) characterized by high woody plant density. Due to the complex study area, the data fusion approach explains with variable reliability the local vegetation properties (R 2(D C)=0.14, R 2(h)=0.84, R 2(D S)=0.25, R 2(cbh)=0.66). The generated structural parameter maps represent spatially explicit data layers that can be used as inputs to hydrodynamic models used to analyse flow resistance effects in different submergence conditions of vegetation. A simple flow resistance model was applied over a test area comparing the results of the proposed method and a traditional ground-based approach. The modelling results showed that the new method is able to provide accurate output data to describe the interaction between water levels and bio-mechanical characteristics of vegetation. The proposed methodology provides a fast, repeatable and accurate way of obtaining floodplain roughness, which enables regular updating of vegetation characteristics.

KW - Flow resistance

KW - Hydraulic roughness

KW - LiDAR

KW - Quickbird

KW - Remote sensing

KW - Riparian corridor

KW - Sieve River

KW - Woody vegetation

UR - http://www.scopus.com/inward/record.url?scp=80051965402&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=80051965402&partnerID=8YFLogxK

U2 - 10.1002/rra.1398

DO - 10.1002/rra.1398

M3 - Article

AN - SCOPUS:80051965402

VL - 27

SP - 826

EP - 840

JO - River Research and Applications

JF - River Research and Applications

SN - 1535-1459

IS - 7

ER -