Canopy Cover Estimation from Landsat Images: Understory Impact onTop-of-canopy Reflectance in a Northern Hardwood Forest
In northern hardwood forests, light availability is considered to be the main factor limiting seedling and sapling growth. However, field measurement of this variable is time-consuming. To address this issue, we developed random forest regression models to estimate canopy cover from a Landsat 8 OLI...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Taylor & Francis Group
2018-09-01
|
| Series: | Canadian Journal of Remote Sensing |
| Online Access: | http://dx.doi.org/10.1080/07038992.2018.1533399 |
| _version_ | 1827794943293259776 |
|---|---|
| author | Stéphanie Landry Martin-Hugues St-Laurent Peter R. Nelson Gaetan Pelletier Marc-André Villard |
| author_facet | Stéphanie Landry Martin-Hugues St-Laurent Peter R. Nelson Gaetan Pelletier Marc-André Villard |
| author_sort | Stéphanie Landry |
| collection | DOAJ |
| description | In northern hardwood forests, light availability is considered to be the main factor limiting seedling and sapling growth. However, field measurement of this variable is time-consuming. To address this issue, we developed random forest regression models to estimate canopy cover from a Landsat 8 OLI image of a northern hardwood forest in northwestern New Brunswick, Canada. We then assessed the accuracy of model predictions with a canopy height model (CHM) derived from LiDAR data. We selected 2 threshold heights (1.3 and 5 m) to distinguish the understory from the overstory and to determine the impact of the understory on top-of-canopy reflectance. Our results show that the understory influenced top-of-canopy reflectance and that a 1.3 m height threshold provided the most accurate estimation of canopy cover. In contrast with studies conducted in softwood stands, we found no evidence that the shortwave infrared (SWIR1) band decreased the influence of the understory on top-of-canopy reflectance. In northern hardwood forests, the estimation of canopy characteristics, such as canopy cover and leaf area index, should be focused on the green band, as it was least influenced by understory vegetation. |
| first_indexed | 2024-03-11T18:41:07Z |
| format | Article |
| id | doaj.art-a4f7d79c3fc943deab166669ad694021 |
| institution | Directory Open Access Journal |
| issn | 1712-7971 |
| language | English |
| last_indexed | 2024-03-11T18:41:07Z |
| publishDate | 2018-09-01 |
| publisher | Taylor & Francis Group |
| record_format | Article |
| series | Canadian Journal of Remote Sensing |
| spelling | doaj.art-a4f7d79c3fc943deab166669ad6940212023-10-12T13:36:22ZengTaylor & Francis GroupCanadian Journal of Remote Sensing1712-79712018-09-0144543544610.1080/07038992.2018.15333991533399Canopy Cover Estimation from Landsat Images: Understory Impact onTop-of-canopy Reflectance in a Northern Hardwood ForestStéphanie Landry0Martin-Hugues St-Laurent1Peter R. Nelson2Gaetan Pelletier3Marc-André Villard4Université du Québec à RimouskiUniversité du Québec à RimouskiUniversity of Maine – Fort Kent, 23 University Dr.Northern Hardwoods Research Institute Inc. 165 boulevard HébertMount Allison UniversityIn northern hardwood forests, light availability is considered to be the main factor limiting seedling and sapling growth. However, field measurement of this variable is time-consuming. To address this issue, we developed random forest regression models to estimate canopy cover from a Landsat 8 OLI image of a northern hardwood forest in northwestern New Brunswick, Canada. We then assessed the accuracy of model predictions with a canopy height model (CHM) derived from LiDAR data. We selected 2 threshold heights (1.3 and 5 m) to distinguish the understory from the overstory and to determine the impact of the understory on top-of-canopy reflectance. Our results show that the understory influenced top-of-canopy reflectance and that a 1.3 m height threshold provided the most accurate estimation of canopy cover. In contrast with studies conducted in softwood stands, we found no evidence that the shortwave infrared (SWIR1) band decreased the influence of the understory on top-of-canopy reflectance. In northern hardwood forests, the estimation of canopy characteristics, such as canopy cover and leaf area index, should be focused on the green band, as it was least influenced by understory vegetation.http://dx.doi.org/10.1080/07038992.2018.1533399 |
| spellingShingle | Stéphanie Landry Martin-Hugues St-Laurent Peter R. Nelson Gaetan Pelletier Marc-André Villard Canopy Cover Estimation from Landsat Images: Understory Impact onTop-of-canopy Reflectance in a Northern Hardwood Forest Canadian Journal of Remote Sensing |
| title | Canopy Cover Estimation from Landsat Images: Understory Impact onTop-of-canopy Reflectance in a Northern Hardwood Forest |
| title_full | Canopy Cover Estimation from Landsat Images: Understory Impact onTop-of-canopy Reflectance in a Northern Hardwood Forest |
| title_fullStr | Canopy Cover Estimation from Landsat Images: Understory Impact onTop-of-canopy Reflectance in a Northern Hardwood Forest |
| title_full_unstemmed | Canopy Cover Estimation from Landsat Images: Understory Impact onTop-of-canopy Reflectance in a Northern Hardwood Forest |
| title_short | Canopy Cover Estimation from Landsat Images: Understory Impact onTop-of-canopy Reflectance in a Northern Hardwood Forest |
| title_sort | canopy cover estimation from landsat images understory impact ontop of canopy reflectance in a northern hardwood forest |
| url | http://dx.doi.org/10.1080/07038992.2018.1533399 |
| work_keys_str_mv | AT stephanielandry canopycoverestimationfromlandsatimagesunderstoryimpactontopofcanopyreflectanceinanorthernhardwoodforest AT martinhuguesstlaurent canopycoverestimationfromlandsatimagesunderstoryimpactontopofcanopyreflectanceinanorthernhardwoodforest AT peterrnelson canopycoverestimationfromlandsatimagesunderstoryimpactontopofcanopyreflectanceinanorthernhardwoodforest AT gaetanpelletier canopycoverestimationfromlandsatimagesunderstoryimpactontopofcanopyreflectanceinanorthernhardwoodforest AT marcandrevillard canopycoverestimationfromlandsatimagesunderstoryimpactontopofcanopyreflectanceinanorthernhardwoodforest |