Designing erosion management plans in Lebanon using remote sensing, GIS and decision-tree modeling
(2008) In Landscape and Urban Planning 88(2-4). p.54-63- Abstract
Soil erosion by water represents a serious threat to the natural and human environment in Mediterranean countries, including Lebanon, which represents a good case study. This research deals with how to use Geographic Information Systems (GIS), remote sensing, and, more specifically, structural classification techniques and decision-tree modeling to map erosion risks and design priority management planning over a representative region of Lebanon. The structural classification organization and analysis of spatial structures (OASIS) of Landsat TM satellite imagery (30 m) was used to define landscapes that prevail in this area and their boundaries, depending on their spectral appearance. The landscape map produced was overlaid sequentially... (More)
Soil erosion by water represents a serious threat to the natural and human environment in Mediterranean countries, including Lebanon, which represents a good case study. This research deals with how to use Geographic Information Systems (GIS), remote sensing, and, more specifically, structural classification techniques and decision-tree modeling to map erosion risks and design priority management planning over a representative region of Lebanon. The structural classification organization and analysis of spatial structures (OASIS) of Landsat TM satellite imagery (30 m) was used to define landscapes that prevail in this area and their boundaries, depending on their spectral appearance. The landscape map produced was overlaid sequentially with thematic erosion factorial maps (i.e., slope gradient, drainage density, rainfall quantity, vegetal cover, soil infiltration, soil erodibility, rock infiltration and rock movement). The overlay was visual and conditional using three visual interpretation rules (dominance, unimodality and scarcity conservation), and landscape properties were produced. Rills and gullies were measured in the field, and a decision-tree regression model was developed on the landscapes to statistically explain gully occurrence. This model explained 88% of the variability in field gully measurements. The erosion risk map produced corresponds well to field observations (accuracy of 82%). The landscapes were prioritized according to anti-erosive remedial measures: preventive (Pre), protective (Pro), and restorative (Res). This approach seems useful in Lebanon, but can also serve in other countries with similar geo-environmental conditions or those lacking detailed geospatial data.
(Less)
- author
- Kheir, Rania Bou ; Abdallah, Chadi ; Runnstrom, Micael LU and Martensson, Ulrik LU
- organization
- publishing date
- 2008-12-01
- type
- Contribution to journal
- publication status
- published
- subject
- keywords
- Decision-trees, Erosion maps, GIS, Landscape units, Structural image classifications
- in
- Landscape and Urban Planning
- volume
- 88
- issue
- 2-4
- pages
- 10 pages
- publisher
- Elsevier
- external identifiers
-
- scopus:54149091193
- ISSN
- 0169-2046
- DOI
- 10.1016/j.landurbplan.2008.08.003
- language
- English
- LU publication?
- yes
- id
- 664962a8-d954-495f-b8a6-260767617e81
- date added to LUP
- 2019-06-05 10:33:45
- date last changed
- 2022-01-31 21:31:01
@article{664962a8-d954-495f-b8a6-260767617e81,
abstract = {{<p>Soil erosion by water represents a serious threat to the natural and human environment in Mediterranean countries, including Lebanon, which represents a good case study. This research deals with how to use Geographic Information Systems (GIS), remote sensing, and, more specifically, structural classification techniques and decision-tree modeling to map erosion risks and design priority management planning over a representative region of Lebanon. The structural classification organization and analysis of spatial structures (OASIS) of Landsat TM satellite imagery (30 m) was used to define landscapes that prevail in this area and their boundaries, depending on their spectral appearance. The landscape map produced was overlaid sequentially with thematic erosion factorial maps (i.e., slope gradient, drainage density, rainfall quantity, vegetal cover, soil infiltration, soil erodibility, rock infiltration and rock movement). The overlay was visual and conditional using three visual interpretation rules (dominance, unimodality and scarcity conservation), and landscape properties were produced. Rills and gullies were measured in the field, and a decision-tree regression model was developed on the landscapes to statistically explain gully occurrence. This model explained 88% of the variability in field gully measurements. The erosion risk map produced corresponds well to field observations (accuracy of 82%). The landscapes were prioritized according to anti-erosive remedial measures: preventive (Pre), protective (Pro), and restorative (Res). This approach seems useful in Lebanon, but can also serve in other countries with similar geo-environmental conditions or those lacking detailed geospatial data.</p>}},
author = {{Kheir, Rania Bou and Abdallah, Chadi and Runnstrom, Micael and Martensson, Ulrik}},
issn = {{0169-2046}},
keywords = {{Decision-trees; Erosion maps; GIS; Landscape units; Structural image classifications}},
language = {{eng}},
month = {{12}},
number = {{2-4}},
pages = {{54--63}},
publisher = {{Elsevier}},
series = {{Landscape and Urban Planning}},
title = {{Designing erosion management plans in Lebanon using remote sensing, GIS and decision-tree modeling}},
url = {{http://dx.doi.org/10.1016/j.landurbplan.2008.08.003}},
doi = {{10.1016/j.landurbplan.2008.08.003}},
volume = {{88}},
year = {{2008}},
}