ANALYSIS OF METHODOLOGIES FOR CARBON STOCK ESTIMATION IN FORESTS
| Autor: | O.O. Kavats, D.A. Khramov, K.L. Sergieieva, V.V. Vasyliev |
|---|---|
| Rok vydání: | 2022 |
| Předmět: | |
| Zdroj: | System technologies. 4:56-73 |
| ISSN: | 2707-7977 1562-9945 |
| DOI: | 10.34185/1562-9945-4-141-2022-05 |
| Popis: | Current approaches to carbon stock estimation in forest ecosystems are discussed. Datasets containing biomass and carbon stock estimates that can be used for training/validation in machine learning are described. Examples of applying the remote approach to assessing forest biomass over large areas are analyzed. To estimate the forest carbon stocks in Ukraine, the most promising is the remote approach, which combines ground-based and satellite measurements for forest classification and statistical modeling of carbon stocks. For training and validation of machine learning algorithms, it is proposed to use the GEDI Biomass Map covering most of the territory of Ukraine — from the southern borders to the latitude of Chernihiv in the north. A prototype of forest biomass estimating product in Ukraine can be based on publicly available MODIS NBAR data, SRTM DEM, ECMWF climate data and use the Random Forest machine learning method. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|