Modeling of Forest Communities’ Spatial Structure at the Regional Level through Remote Sensing and Field Sampling: Constraints and Solutions

Kotlov, Ivan; Chernen’kova, T. V.

doi:10.3390/f11101088

Cited by 13 publications

(16 citation statements)

References 40 publications

(59 reference statements)

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Overall accuracy for an independent test sample (20% of the selected points) was 0.63 (Kappa = 0.47) which is moderate [94]. This is consistent with our study of modeling the MR forest formations [84]. The confusion matrix is shown in Table 3.…”

Section: The Results Of Retrospective Modeling 30 Years Agosupporting

confidence: 87%

“…We generalized the accumulated experience of numerous studies carried out in the MR over the past years with respect to the typological composition of coniferous communities [23,28,[95][96][97]. According to our estimates [84], the area of spruce forests amounts 21.7% (including spruce-small-leaved forests) and the area of pine forests amounts 18.5% (including spruce-pine forests). This proportion is close to official statistics [98].…”

Section: Discussionmentioning

confidence: 99%

“…Six forest formations were modeled using the training sample: 4 formations of coniferous forests, the deciduous forests and non-forest areas. The deciduous forests include 2 categories-small-leaved and broad-leaved forests [84].…”

Section: Use Of Remote Sensing Datamentioning

confidence: 99%

“…The spatial distribution of forest formations in 2020 is taken from our previous study [84]. It uses the data on forest cover for the 2020 [84].…”

Section: Quality Assessment and Additional Sources Of Datamentioning

confidence: 99%

“…The spatial distribution of forest formations in 2020 is taken from our previous study [84]. It uses the data on forest cover for the 2020 [84]. It this study we used multiseasonal mosaics of Landsat 8 (March, May, June, July, and September), spectral indices for each season, DEM with morphometric variables and HH/HV polarization of Palsar radar image.…”

Section: Quality Assessment and Additional Sources Of Datamentioning

confidence: 99%

See 4 more Smart Citations

Spatiotemporal Modeling of Coniferous Forests Dynamics along the Southern Edge of Their Range in the Central Russian Plain

et al. 2021

Self Cite

View full text Add to dashboard Cite

Forests with predominance of Norway spruce (Picea abies (L.) H. Karst.) and Scots pine (Pinus sylvestris L.) within the hemiboreal zone are considered as secondary communities formed under long-term human activity (logging, plowing, fires and silviculture). This study raises the question—how stable is current state of coniferous forests on the southern border of their natural distribution in the center of Eastern Europe using the example of the Moscow region (MR)? The object of the study are spruce and pine forests in different periods of Soviet and post-Soviet history within the Moscow Region (MR). The current proportion of spruce forests is 21.7%, and the proportion of pine forests is 18.5% from total forest area according to our estimates. The direction and rate of forest succession were analyzed based on current composition of populations of the main forest-forming species (spruce, pine, birch, aspen, oak, linden, and ash) based on ground-based research materials collected in 2006–2019. This allowed to develop the dynamic model (DM) of forest communities with the participation of Norway spruce and Scots pine for several decades. Assessment of the spatial distribution of coniferous communities is based on field data and spatial modeling using remote sensing data—Landsat 8 mosaic for 2020. In parallel, a retrospective model (RM) of the spatial-temporal organization of spruce and pine forests for a 30-year period was developed using two Landsat 5 mosaics. For this, nine different algorithms were tested and the best one for this task was found—random forest. Geobotanical relevés were used as a training sample combined with the 2006–2012 mosaic; the obtained spectral signatures were used for modeling based on the 1984–1990 mosaic. Thus, two multi-temporal spatial models of coniferous formations have been developed. Detailed analysis of the structure of spruce and pine forests based on field data made it possible to track trends of successional dynamics for the first time, considering the origin of communities and the ecological conditions of habitats. As a result, ideas about the viability of spruce and pine cenopopulations in different types of communities were formulated, which made possible to develop a dynamic model (DM) of changes in forest communities for future. Comparison of the areas and nature of changes in the spatial structure of coniferous formations made possible to develop the RM. Comparison of two different-time models of succession dynamics (DM and RM) makes possible to correct the main trends in the transformation of coniferous forests of natural and artificial origin under the existing regime of forestry. A set of features was identified that indicates risk factors for coniferous forests in the region. A further decrease of the spruce and pine plantations and increase of the spruce-small-leaved and deciduous formations are expected in the study area. The proportion of pine-spruce forests does not exceed 3% of the area and can be considered as the most vulnerable type of forest.

show abstract

Section: The Results Of Retrospective Modeling 30 Years Agosupporting

confidence: 87%

Section: Discussionmentioning

confidence: 99%

Section: Use Of Remote Sensing Datamentioning

confidence: 99%

“…The spatial distribution of forest formations in 2020 is taken from our previous study [84]. It uses the data on forest cover for the 2020 [84].…”

Section: Quality Assessment and Additional Sources Of Datamentioning

confidence: 99%

Section: Quality Assessment and Additional Sources Of Datamentioning

confidence: 99%