Assessing reforestation after a fire using UAVs

. In 2022, an aerial photograph of a 16-year-old burnt area was carried out along the northwestern border of the Botanical Garden of Petrozavodsk State University (Republic of Karelia, Russia). In addition to this, a field survey of the survey sites was carried out. The work was carried out to assess the state of natural forest regeneration after the fire. The results of combined studies showed that after the fire, viable young growth of mixed composition was formed with a predominance of Pinus sylvestris (71% of the total number of tree species), Betula pendula (22%), Populus tremula and Alnus incana (3.5% each). When comparing plant heights on three-dimensional point clouds with field measurements, it was only possible to reliably determine plant heights for a large generally accepted category (> 1.5 m), while the dependence of determining tree heights obtained by different methods turned out to be very significant (R 2 = 0.95). The maximum height of P. sylvestris trees in the study area was 6.2 m, the average values varied from 3.5 m to 4.5 m, which indicates the success of reforestation after a fire on rocky outcrops.


Introduction
About 60% of the area of lost forests is associated with the impact of forest fires.These circumstances make it necessary to carry out detailed assessments of areas after a fire to determine the success of reforestation, which is the most important task of forestry.
Traditional methods for assessing the state of forest regeneration [1] are associated with field surveys and (or) analysis of remote sensing data from satellites and do not sufficiently allow for a comprehensive and complete assessment of the state.
At the same time, advances in unmanned aerial vehicles (UAVs) and photogrammetric data processing techniques provide high-resolution information that can be used to qualitatively assess the condition of fire-damaged areas.
The use of UAVs for the purpose of reforestation analysis can significantly reduce labor costs in comparison with field surveys and, in addition, obtain important information and high spatial resolution images in real time [2].
The interpretation and accuracy of the estimates obtained from UAV data for the tasks of determining the state and inventory of forest resources remains poorly studied and requires verification by field surveys.
A number of works [3,4] show that when studying the state of reforestation, the most effective combination of combined research methods is the use of UAVs and field survey of the area.A comparison of the results obtained from field survey sites and high spatial resolution materials makes it possible to reliably characterize the vegetation in the study area.The purpose of the study is to give a comparative assessment of the state of natural forest regeneration after a fire in the middle taiga by comparing aerial photography data from a UAV with the results of a field survey.

Materials and methods
The object of the study is a 16-year-old burnt-out that formed on the site of burnt rock pine forests along the northwestern border of the Botanical Garden of Petrozavodsk State University (Republic of Karelia, Russia).The area of the study site is 30 hectares.Geographic coordinates of the study area: 61.84885 N, 34.39479 E.
Before the fire, a rock pine forest about 100-120 years old grew on the site.The ground cover was dominated by lichens, and the undergrowth was dominated by juniper.These indicators are taken from the 2000 taxation description.The consequences of the surface fire were massive windfall and drying out of trees due to the complete burning of organic matter at rock outcrops.At the time of the study, remains of burnt wood remained on the site.
The burn is characterized by vegetation growing in patches, which alternate with rocky outcrops of various sizes.The vegetation is currently represented by undergrowth of the main tree species, undergrowth, grass-shrub and moss-lichen layers.
When carrying out field work, a combined method was used: the classical method of assessing the success of natural reforestation [1] and analysis of aerial photography materials from a UAV [5].
When using the classical method, regardless of the density of the undergrowth, its height and condition, in all cases circular areas of 10 m2 were laid (circle radius 178.5 cm).At the same time, for further comparison with the orthophotomap of the area, preliminary marking of counting passages and centers of counting sites was carried out.In total, two survey tracks were laid along the lines of the longest diagonals of the study area.On each move, 10 platforms were laid, distributed throughout the entire move at the same distance from each other.
A Phantom 4 pro quadcopter UAV was used to conduct aerial photography.The UAV flight took place at an altitude of 110 meters from the level of the take-off platform throughout the entire study area.The original shooting resolution is 2.5 -3 cm/pixel depending on the local relief heights.The obtained data were processed in Agisoft Metashape Professional software.The data processing algorithm included the implementation of the solution proposed in the user manual (alignment of images, construction of a dense point cloud, construction of a textured model, height map and orthomosaic).

Results
As a result of photogrammetric processing of 367 photographs from a UAV, 219 thousand tie points were corrected with an average discrepancy of 0.5 pixels (reprojection error).For the study area, an orthophotomap of the territory was reconstructed with a spatial resolution of 4.6 cm/pixel (Figure 1  On the elevation map of the area (Figure 1 -below) the undergrowth is clearly visible (individual trees with a height of 1.5 m), while on the elevation map the pine forest that was not damaged by fire and has maximum height values is best identified (Figure 1, number 4).When analyzing the orthophotomap, it was calculated that rock outcrops are present throughout almost the entire study area and occupy 60% of the area.The use of UAVs made it possible to determine the nature of the placement of undergrowth.The presented materials clearly show that some of the vegetation is located along rock outcrops (Figure 1, number 3), oriented from southeast to northwest, which is probably due to the morphology of bedrock and large cracks that were formed behind account of tectonic activity.
In a quantitative analysis of the distribution of the species composition of vegetation across the territory on the orthophotomap, it was found that the easiest to recognize trees are Pinus sylvestris due to the light bluish shade of the crown and Betula pendula due to the bright green contrast in comparison with other plants.As a result of counting tree species, their quantitative distribution was determined in the ratio of 64% Pinus sylvestris and 23% E3S Web of Conferences 463, 02024 (2023) EESTE2023 https://doi.org/10.1051/e3sconf/202346302024Betula pendula.It was not possible to identify other forest-forming species to species, as well as the composition of young trees as a whole, on the orthophotomap, but when counted, their number was about 13% of the total number of plants.
Field surveys showed that 16 years after the fire, viable young growth of mixed composition with a predominance of Pinus sylvestris (71% of the total number of tree species) was formed.In addition to pine, Betula pendula (22%), Populus tremula and Alnus incana (3.5% each) participate in the formation of young trees in this burnt area.These indicators are consistent with the analysis of the orthophotomap of the area (error less than 10%).
A field examination of the undergrowth revealed that it is represented by a small number of shrub species: several species of Salix sp., Lonicera xylosteum L., Juniperus communis L., Sorbus aucuparia L. and Rosa canina L. The total number of undergrowth species does not exceed 4500 ind./ha.Sorbus aucuparia predominates.Identification of undergrowth and small undergrowth on the orthophotomap was not possible due to their small size.
To analyze the structure of young stands in height, three-dimensional clouds of points of the plant canopy and the earth's surface were constructed.When comparing the height of plants on photogrammetric point clouds with field measurements, it was possible to reliably determine only the height of plants for a large generally accepted category (> 1.5 m) and only for the main forest-forming species.Moreover, the heights measured from UAV data were in good agreement with the heights measured by the ground method (Table 1).The relationship between the heights of individual trees obtained by different methods is very significant (R 2 = 0.95), and no significant differences were found between the heights obtained by different methods (paired t test, p < 0.09).The maximum height of Pinus sylvestris trees in the study area did not exceed 6 m, and the average values varied from 3.5 m to 4.5 m, which indicates the success of reforestation on burnt rock outcrops.Differences between adolescents by age are less pronounced.The main part of the undergrowth is 12-13 years old, which indicates that the bulk of the undergrowth appeared in the burnt area after an abundant seed harvest (2-3 years after the fire).The distribution of undergrowth by height is also practically unrelated to age.
A study of 12-year-old plants taken as model plants for current growth and height showed that the height of plants depends to a greater extent on microconditions and the thickness of sedimentary rock at the site of seedling formation.Thus, plants of the same age had different absolute height increases, which made it possible to construct a series of polynomial equations of the second degree for plants of the same age (Figure 2).
It is known that the average thickness of the A 0 horizon (litter) varied from 1 to 6 cm.The beginning of the process of formation of the humus-accumulative horizon was observed only when the thickness of the litter was at least 4-5 cm.Reliability indicators for the difference for a height of more than 2 m in the same volume same age (12 years) were at the 1% significance level.The t St criterion for 12-year-old adolescents with 1 cm thick E3S Web of Conferences 463, 02024 (2023) EESTE2023 https://doi.org/10.1051/e3sconf/202346302024litter and 6 cm thick is 4.2≥4, for 2 cm and 6 cm thick it is 2.6≥2.58, that is, the differences are significant at the 5% significance level.

Discussion
Analysis of materials obtained using UAVs made it possible to accurately determine the proportion of rock outcrops and some characteristics of areas with vegetation recovering after a fire.Small undergrowth and the species composition of the grass-shrub and mosslichen layers are not identified based on data from the UAV.To solve these problems, it is likely necessary to use multispectral imaging with near-infrared (NIR) and/or far-infrared (RedEdge) spectrum channels.At the same time, the use of UAVs has its disadvantages in comparison with satellite remote sensing data: dependence on weather conditions (strong wind, rain, low temperature) and small coverage of the survey area due to limited battery life (from several minutes to 3-4 hours ).
Based on the research conducted, it has been established that the self-renewal of tree species in burnt areas, even in poor soil conditions on rocky outcrops, is quite successful.The beginning of the soil formation process was observed only in microdepressions; the height of the plants significantly depended on the thickness of the litter.The height values of Pinus sylvestris varied from 3.5 m to 4.5 m, this indicates the success of reforestation of the study area, but this figure is 20% less than the height of reforestation plants in clearings in Karelia [6].

Conclusion
Using the example of studying the burned-out territory of the Botanical Garden of Petrozavodsk State University, the applicability of photogrammetric aerial photography for assessing forest regeneration after a fire is shown.The actual data obtained during the survey of the sites allowed us to analyze the state of the vegetation.The results of combined E3S Web of Conferences 463, 02024 (2023) EESTE2023 https://doi.org/10.1051/e3sconf/202346302024studies showed that after the fire, viable young growth of mixed composition was formed with a predominance of Pinus sylvestris (71% of the total number of tree species), Betula pendula (22%), Populus tremula and Alnus incana (3.5% each).When comparing plant heights on three-dimensional point clouds with field measurements, it was only possible to reliably determine plant heights for a large generally accepted category (> 1.5 m), while the dependence of determining tree heights obtained by different methods turned out to be very significant (R 2 = 0.95).

Fig. 2 .
Fig. 2. Height of model pine plants of the same age by year of growth with different thickness of litter A 0.

Table 1 .
Comparison of Pinus sylvestris tree heights measured using UAV data with ground measurements.