Dataset Information

Estimation of deracinated trees area in temperate forest with satellite images employing machine learning methods.

ABSTRACT: Climate change can increase the number of uprooted trees. Although there have been an increasing number of machine learning applications for satellite image analysis, the estimation of deracinated tree area by satellite image is not well developed. Therefore, we estimated the deracinated tree area of forests via machine-learning classification using Landsat 8 satellite images. We employed support vector machines (SVMs), random forests (RF), and convolutional neural networks (CNNs) as potential machine learning methods, and tested their performance in estimating the deracinated tree area. We collected satellite images of upright trees, deracinated trees, soil, and others (e.g., waterbodies and cities), and trained them with the training data. We compared the accuracy represented by the correct classification rate of these methods, to determine the deracinated tree area. It was found that the SVM and RF performed better than the CNN for two-class classification (deracinated and upright trees), and the correct classification rates of all methods were up to 93%. We found that the CNN and RF performed significantly higher for the four- and two-class classification compared to the other methods, respectively. We conclude that the CNN is useful for estimating deracinated tree areas using Landsat 8 satellite images.

SUBMITTER: Doi H

PROVIDER: S-EPMC8384040 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:Large-diameter trees dominate the structure, dynamics and function of many temperate and tropical forests. Although both scaling theory and competition theory make predictions about the relative composition and spatial patterns of large-diameter trees compared to smaller diameter trees, these predictions are rarely tested. We established a 25.6 ha permanent plot within which we tagged and mapped all trees ≥1 cm dbh, all snags ≥10 cm dbh, and all shrub patches ≥2 m(2). We sampled downed woody debris, litter, and duff with line intercept transects. Aboveground live biomass of the 23 woody species was 507.9 Mg/ha, of which 503.8 Mg/ha was trees (SD = 114.3 Mg/ha) and 4.1 Mg/ha was shrubs. Aboveground live and dead biomass was 652.0 Mg/ha. Large-diameter trees comprised 1.4% of individuals but 49.4% of biomass, with biomass dominated by Abies concolor and Pinus lambertiana (93.0% of tree biomass). The large-diameter component dominated the biomass of snags (59.5%) and contributed significantly to that of woody debris (36.6%). Traditional scaling theory was not a good model for either the relationship between tree radii and tree abundance or tree biomass. Spatial patterning of large-diameter trees of the three most abundant species differed from that of small-diameter conspecifics. For A. concolor and P. lambertiana, as well as all trees pooled, large-diameter and small-diameter trees were spatially segregated through inter-tree distances <10 m. Competition alone was insufficient to explain the spatial patterns of large-diameter trees and spatial relationships between large-diameter and small-diameter trees. Long-term observations may reveal regulation of forest biomass and spatial structure by fire, wind, pathogens, and insects in Sierra Nevada mixed-conifer forests. Sustaining ecosystem functions such as carbon storage or provision of specialist species habitat will likely require different management strategies when the functions are performed primarily by a few large trees as opposed to many smaller trees.

Dataset Information

Estimation of deracinated trees area in temperate forest with satellite images employing machine learning methods.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets