Analysis of Regional Distribution of Tree Species Using Multi-Seasonal Sentinel-1&2 Imagery within Google Earth Engine

Accurate information on tree species is in high demand for forestry management and further investigations biodiversity environmental monitoring. Over regional or large areas, distinguishing at resolutions faces the challenges of a lack representative features computational power. A novel methodology was proposed to delineate explicit spatial distribution six dominant (Pinus tabulaeformis, Quercus mongolia, Betula spp., Populus Larix Armeniaca sibirica) one residual class 10 m resolution. Their patterns were analyzed over an area covering 90,000 km2 using analysis-ready volume multisensor imagery within Google Earth engine (GEE) platform afterwards. Random forest algorithm built into GEE used together with 20th 80th percentiles multitemporal extracted from Sentinel-1/2, topographic features. The composition natural forests plantations city county-level performed detail classification achieved reliable accuracy (77.5% overall accuracy, 0.71 kappa), revealed that outnumber (Quercus mongolia spp.) by 6% mainly concentrated northern southern regions. Arhorchin had largest 4500 km2, while Hexingten Aohan ranked first plantation area. Additionally, proportion number Karqin Ningcheng more balanced. We suggest focusing suitable areas modeling species’ models factors based results rather than field survey plots studies.