Can tree cover extent be estimated directly from ICESat-2 spaceborne lidar data?

We performed the estimation at two boreal forest areas in Finland, Valtimo and Pello, using two methods: post-classification and logistic regression estimation. In post-classification, dominant height and canopy cover were estimated first, and areas where dominant height was (≥ 5 m) and canopy cover...

Full description

Saved in:
Bibliographic Details
Main Authors: Hanna Haapala, Petri Varvia, Ranjith Gopalakrishnan, Lauri Korhonen
Format: Article
Language:English
Published: Elsevier 2025-08-01
Series:International Journal of Applied Earth Observations and Geoinformation
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S1569843225003395
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1839635755498995712
author Hanna Haapala
Petri Varvia
Ranjith Gopalakrishnan
Lauri Korhonen
author_facet Hanna Haapala
Petri Varvia
Ranjith Gopalakrishnan
Lauri Korhonen
author_sort Hanna Haapala
collection DOAJ
description We performed the estimation at two boreal forest areas in Finland, Valtimo and Pello, using two methods: post-classification and logistic regression estimation. In post-classification, dominant height and canopy cover were estimated first, and areas where dominant height was (≥ 5 m) and canopy cover (≥ 10%) were considered forested. In the second method, logistic regression was used to determine the areas with tree cover using ICESat-2 variables as predictors. The reference values were obtained using airborne laser scanning data and locally measured field plots. The transferability of the ICESat-2 models was also assessed.Our results indicated that ICESat-2 data are suitable for the estimation of tree cover extent over large areas. The overall accuracy ranged from 92% to 95% both for post-classification and logistic regression, depending on the study area. Although the performance of the two classification methods was similar, the classification accuracies were generally lower in Pello. The ICESat-2 models for dominant height and canopy cover had RMSE% values of 21.1–24.5% and 24.9–36.8%, respectively. We conclude that ICESat-2 data appear promising for estimation of accurate statistics on tree cover extent in the boreal biome.
format Article
id doaj-art-fbf4ea7e868143f7944d34dffb51f8bc
institution Matheson Library
issn 1569-8432
language English
publishDate 2025-08-01
publisher Elsevier
record_format Article
series International Journal of Applied Earth Observations and Geoinformation
spelling doaj-art-fbf4ea7e868143f7944d34dffb51f8bc2025-07-09T04:32:03ZengElsevierInternational Journal of Applied Earth Observations and Geoinformation1569-84322025-08-01142104692Can tree cover extent be estimated directly from ICESat-2 spaceborne lidar data?Hanna Haapala0Petri Varvia1Ranjith Gopalakrishnan2Lauri Korhonen3Corresponding author.; School of Forest Sciences, University of Eastern Finland, P.O. Box 111, Joensuu, FI-80101, FinlandSchool of Forest Sciences, University of Eastern Finland, P.O. Box 111, Joensuu, FI-80101, FinlandSchool of Forest Sciences, University of Eastern Finland, P.O. Box 111, Joensuu, FI-80101, FinlandSchool of Forest Sciences, University of Eastern Finland, P.O. Box 111, Joensuu, FI-80101, FinlandWe performed the estimation at two boreal forest areas in Finland, Valtimo and Pello, using two methods: post-classification and logistic regression estimation. In post-classification, dominant height and canopy cover were estimated first, and areas where dominant height was (≥ 5 m) and canopy cover (≥ 10%) were considered forested. In the second method, logistic regression was used to determine the areas with tree cover using ICESat-2 variables as predictors. The reference values were obtained using airborne laser scanning data and locally measured field plots. The transferability of the ICESat-2 models was also assessed.Our results indicated that ICESat-2 data are suitable for the estimation of tree cover extent over large areas. The overall accuracy ranged from 92% to 95% both for post-classification and logistic regression, depending on the study area. Although the performance of the two classification methods was similar, the classification accuracies were generally lower in Pello. The ICESat-2 models for dominant height and canopy cover had RMSE% values of 21.1–24.5% and 24.9–36.8%, respectively. We conclude that ICESat-2 data appear promising for estimation of accurate statistics on tree cover extent in the boreal biome.http://www.sciencedirect.com/science/article/pii/S1569843225003395Land coverForest areaProfiling lidarTreeline migrationCrown coverForest definition
spellingShingle Hanna Haapala
Petri Varvia
Ranjith Gopalakrishnan
Lauri Korhonen
Can tree cover extent be estimated directly from ICESat-2 spaceborne lidar data?
International Journal of Applied Earth Observations and Geoinformation
Land cover
Forest area
Profiling lidar
Treeline migration
Crown cover
Forest definition
title Can tree cover extent be estimated directly from ICESat-2 spaceborne lidar data?
title_full Can tree cover extent be estimated directly from ICESat-2 spaceborne lidar data?
title_fullStr Can tree cover extent be estimated directly from ICESat-2 spaceborne lidar data?
title_full_unstemmed Can tree cover extent be estimated directly from ICESat-2 spaceborne lidar data?
title_short Can tree cover extent be estimated directly from ICESat-2 spaceborne lidar data?
title_sort can tree cover extent be estimated directly from icesat 2 spaceborne lidar data
topic Land cover
Forest area
Profiling lidar
Treeline migration
Crown cover
Forest definition
url http://www.sciencedirect.com/science/article/pii/S1569843225003395
work_keys_str_mv AT hannahaapala cantreecoverextentbeestimateddirectlyfromicesat2spacebornelidardata
AT petrivarvia cantreecoverextentbeestimateddirectlyfromicesat2spacebornelidardata
AT ranjithgopalakrishnan cantreecoverextentbeestimateddirectlyfromicesat2spacebornelidardata
AT laurikorhonen cantreecoverextentbeestimateddirectlyfromicesat2spacebornelidardata