Miesner, Timon, Herzschuh, Ulrike, Pestryakova, Luidmila A., Wieczorek, Mareike, Zakharov, Evgenii S., Kolmogorov, Alexei I., Davydova, Paraskovya V., and Kruse, Stefan
We compile a data set of forest surveys from expeditions to the north-east of the Russian Federation, in Krasnoyarsk Krai, the Republic of Sakha (Yakutia) and the Chukotka Autonomous Okrug (59-73° N, 97-169° E). The region is characterized by permafrost soils, and forests dominated by larch (Larix gmelinii RUPR., Larix cajanderi MAYR). Our dataset consists of a plot data base describing 226 georeferenced vegetation survey sites, and of a tree data base with information about all trees on these plots. The tree data base contains information on height, species and vitality of 40,289 trees. A subset of the trees was subject to a more detailed inventory, recording stem diameter at base and at breast height, crown diameter and height of the beginning of the crown. We recorded heights up to 28.5 m (median = 2.5 m) and stand densities up to 120,000 trees per ha (median = 1197 ha-1), both values tending to be higher in the more southerly areas. Observed taxa include Larix MILL., Pinus L., Picea A.DIETR., Abies MILL., Salix L., Betula L., Populus L., Alnus MILL. and Ulmus L.. In this study, we present the forest inventory data aggregated per site. Additionally, we connect it with different remote sensing data products to find out how accurately forest structure can be predicted from such products. Allometries were calculated to obtain the diameter from height measurements for every species group. For Larix, the most frequent of ten species groups, allometries depend also on the stand density, as denser stands are characterized by thinner trees, relative to height. The remote sensing products used to compare against the inventory data include climate, forest biomass, canopy height, and forest loss or disturbance. We find that the forest metrics measured in the field can only be reconstructed from the remote sensing data to a limited extent, as they depend on local properties. This illustrates the need for ground inventories like those data we present here. The data can be used for studying the forest structure of north-eastern Siberia, and for the calibration and validation of remotely sensed data. [ABSTRACT FROM AUTHOR]