?
Валидация оценки индекса листовой поверхности по данным MODIS для редкостойных лесов Кольского полуострова с использованием материалов съёмок беспилотных летательных аппаратов
This paper presents the results of validation of the MODIS LAI product developed at the Space Research Institute of the Russian Academy of Sciences (the IKI MODIS LAI product). Validation was performed over a sparse spruce-birch mixed forest stand (~6.7 km2), located off the southern slopes of the Khibiny mountains in the central part of the Kola Peninsula, Russia. Field measurements were performed in June 2019. Validation was performed according to the methodology to scale local point measurements to the low resolution of MODIS data (230 m) via the intermediate link — high resolution Sentinel-2 MSI data (10 m). Special feature of this work is that local measurements of forest stand were performed using UAV technology. Such an approach is especially efficient in terms of mapping spatial variability of canopy structure at large areas. Using photogrammetric techniques, the UAV data were processed to create 3D model of canopy, and convert it to map of canopy closure. Using UAV-based estimate of canopy closure and Sentinel NDVI (Normalized Difference Vegetation Index) data and applying statistical and semi-empirical methods Sentinel LAI map was derived to serve as a baseline for MODIS LAI product accuracy assessment. Mean calculated product value at the study region was LAI = 1.4 and product accuracy 7.1 % (RMSE = 43.1 %). We also report that the IKI MODIS LAI product may have low data coverage at the northern high latitude of Russia: over the extent of validation site the coverage was ~40 % in 2019. Key reason is a (potentially) excessive cloud shadow filtering at low solar zenith angles at high northern latitudes. We also highlight the fact that the sparse northern forests have a well-developed two-level structure (1 — trees and 2 — unified understory levels, which includes forest undergrowth, grass-shrub and moss-lichen), which needs to be taken into account in the future research on the LAI algorithm development and product validation.