Next Article in Journal
Effects of Biochar Application on Nitrogen Fixation and Water Use Efficiency of Understorey Acacia Species as well as Soil Carbon and Nitrogen Pools in a Subtropical Native Forest
Previous Article in Journal
Characterization of Litter and Topsoil Under Different Vegetation Cover by Using a Chemometric Approach
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Forests
Volume 16
Issue 8
10.3390/f16081348
forests-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

UAV-Based Multispectral Assessment of Wind-Induced Damage in Norway Spruce Crowns

by
Endijs Bāders
,
Andris Seipulis
,
Dārta Kaupe
,
Jordane Jean-Claude Champion
,
Oskars Krišāns
and
Didzis Elferts
*
Latvian State Forest Research Institute ‘Silava’, LV-2169 Salaspils, Latvia
*
Author to whom correspondence should be addressed.
Forests 2025, 16(8), 1348; https://doi.org/10.3390/f16081348
Submission received: 30 June 2025 / Revised: 7 August 2025 / Accepted: 15 August 2025 / Published: 19 August 2025
(This article belongs to the Section Forest Inventory, Modeling and Remote Sensing)
Browse Figures
Versions Notes

Abstract

Climate change has intensified the frequency and severity of forest disturbances globally, including windthrow, which poses substantial risks for both forest productivity and ecosystem stability. Rapid and precise assessment of wind-induced tree damage is essential for effective management, yet many injuries remain visually undetectable in the early stages. This study employed drone-based multispectral imaging and a simulated wind stress experiment (static pulling) on Norway spruce (Picea abies (L.) Karst.) to investigate the detectability of physiological and structural changes over four years. Multispectral data were collected at multiple time points (2023–2024), and a suite of vegetation indices (the Normalised Difference Vegetation Index (NDVI), the Structure Insensitive Pigment Index (SIPI), the Difference Vegetation Index (DVI), and Red Edge-based indices) were calculated and analysed using mixed-effects models. Our results demonstrate that trees subjected to mechanical bending (“Bent”) exhibit substantial reductions in the near-infrared (NIR)-based indices, while healthy trees maintain higher and more stable index values. Structure- and pigment-sensitive indices (e.g., the Modified Chlorophyll Absorption Ratio Index (MCARI 2), the Transformed Chlorophyll Absorption in Reflectance Index/Optimised Soil-Adjusted Vegetation Index (TCARI/OSAVI), and RDVI) showed the highest diagnostic value for differentiating between damaged and healthy trees. We found the clear identification of group- and season-specific patterns, revealing that the most pronounced physiological decline in Bent trees emerged only several seasons after the disturbance.

1. Introduction

Climate change has led to an increase in forest disturbances globally, including extreme weather events and their associated impacts, such as wildfires and insect outbreaks [1,2]. In Europe, wind is the most damaging natural disturbance agent, causing significant economic [3,4] and ecological losses [5]. Historical analyses indicate a growing trend in storm-induced forest damage between 1850 and 2000 [6], driven by both climate dynamics and changes in forest stand structure [7].
Rapid and accurate assessment of wind damage is crucial for effective forest management and post-disturbance recovery planning. Recent advancements in remote sensing, particularly in Unmanned Aerial Vehicle (UAV)-based multispectral imaging, may offer practical tools for detecting early signs of tree stress [8]. By identifying trees showing early signs of stress, forest owners can implement timely mitigation strategies to reduce further loss and enhance stand resilience. Norway spruce (Picea abies (L.) Karst.) is one of the dominant coniferous species in temperate and boreal forests across Europe, widely cultivated for its rapid growth and valuable timber [9,10,11]. In Latvia, Norway spruce covers approximately 19% of the total forest area, making its monitoring and health assessment a national priority [12]. Remote sensing technologies, especially UAV-mounted multispectral sensors, have become increasingly accessible for health monitoring. These technologies enable a detailed assessment of canopy traits, such as transparency, reflectance, and colour changes, which are linked to physiological stress [8,13,14,15]. Vegetation indices, such as NDVI, Green NDVI (GNDVI), and Red Edge-based indices, are commonly used to quantify chlorophyll content, leaf water status, and canopy structure [16,17,18,19,20]. NDVI, in particular, has been widely adopted for assessing tree vitality, with threshold values (e.g., NDVI < 0.45) indicating stress or decline [21].
Wind stress can result in internal stem damage that is not immediately visible but can be detected through changes in canopy reflectance. During storms, trees are subjected to oscillatory movement, which creates mechanical strain within both the trunk and root system. This repeated swaying can lead to structural damage such as microcracks in the wood, root plate displacement, or damage to the vascular tissues responsible for water and nutrient transport [22]. Although the canopy may appear visually unaffected in the short term, these internal injuries can impair physiological processes and lead to delayed stress symptoms. Such changes, particularly reductions in chlorophyll content or alterations in leaf water status, often result in modified canopy reflectance, detectable through multispectral imagery [23,24,25]. Tree-pulling experiments have demonstrated that mechanical stress can induce subtle yet measurable physiological changes over time [26]. Therefore, multi-year observations are essential for detecting the delayed effects of storm-induced damage and validating remote sensing approaches.
This study aims to evaluate the capacity of UAV-based multispectral imagery as a tool for the early detection of structurally damaged yet visually healthy Norway spruce trees. Specifically, the objectives are (1) to develop and demonstrate a UAV multispectral data acquisition and analysis workflow for detecting canopy changes induced by mechanical wind stress and (2) to assess the ability of vegetation indices derived from drone imagery to identify early physiological stress-related spectral changes before visible symptoms emerge, thereby supporting timely forest management responses. We hypothesise that (1) UAV-based multispectral imaging enables the detection of stress-induced changes in canopy reflectance associated with mechanically simulated wind damage at high spatial and temporal resolution. (2) Vegetation indices will reveal distinct spectral patterns indicative of stress responses to structural damage before the emergence of visible symptoms. (3) Several years of UAV-based multispectral monitoring enable the detection of both delayed stress responses and potential recovery trajectories in individual trees, thereby supporting more robust forest health assessment and management decisions.

2. Materials and Methods

2.1. Study Area

This research was conducted as a continuation of a previous study [27] and took place in the same geographic region, specifically in Norway spruce stands at the Forest Research Station Kalsnava in the middle-eastern part of Latvia (Figure 1). In total, nine permanent study sites were established on drained oligotrophic mineral soils (Myrtillosa mel.) during different years (2020, 2023, and 2024). The region features a moderately continental climate. The mean annual air temperature is +6.0 °C, with the warmest month being July (17.5–18 °C) and the coldest months being January and February (−4.5 °C). Annual precipitation typically ranges from 670 to 700 mm, with some local variations noted. The month with the most rainfall is October, while April is the driest month. Average wind speeds are highest during autumn and winter (3.7–4 m/s), with a historic maximum of 48 m/s recorded in October 1967. Notably, a prolonged drought event occurred between late May and late June 2021, partially overlapping with an extreme heatwave in mid-June, during which average daily temperatures reached 21 °C and mean solar radiation peaked at 256 W/m2.

2.2. Tree Pulling and Sampling Design

The experiment began in early 2020, when statistical sampling of trees was performed in three mature Norway spruce stands to simulate wind damage through static pulling using a manual cable winch system with a dynamometer for reading the applied loading. A steel cable was anchored to an opposing tree, located 30–40 m away, and fastened to the target tree at half its height. The pulling force was applied gradually until the maximum loading resistance of the tree was reached, as indicated by the initiation of a decrease in the applied loading, ensuring the occurrence of primary failure (internal wood damage that is not visibly detectable [28,29]). Trees that withstood the applied loading without collapsing or any visible damage, such as cracks, were included in further analysis and were classified as “Bent” trees. In total, 30 bent trees were identified in the 2020 sites. An additional 30 visually healthy trees, similar in size and age, were selected from the same stands and classified as the “Healthy” group. In 2023, the research design was expanded to six additional Norway spruce stands. Static pulling tests were conducted on five trees per stand, and all trees around the Bent tree (in the immediate vicinity) were selected for analysis. These adjacent trees were visually assessed and subsequently classified into three categories: Healthy (visibly unaffected), Crown Damage (showing partial defoliation or crown asymmetry), and Dead (completely desiccated or with needle loss consistent with mortality). In 2024, one of the 2023 stands was further intensified to expand the dataset. An additional 60 Bent trees were subjected to static pulling tests (Table 1). Only the most vigorous individuals, as determined by NDVI index values in the October 2023 orthomosaic, were selected.
These additional new trees followed the same sampling protocol: precise GPS coordinates of each tree position were mapped using a Leica GS16 receiver (Leica Geosystems AG, Sankt Gallen, Switzerland) and a Leica Flexline TS06 total station (Leica Geosystems AG, Sankt Gallen, Switzerland), height was measured with a Haglöf Vertex IV (Haglöfs, Bromma, Sweden), and diameter at breast height (DBH) was measured with a Haglöf Mantax Blue calliper (Haglöfs, Bromma, Sweden).

2.3. UAV Image Acquisition

The multispectral imagery used in this study was acquired between 2023 and 2024 using a DJI Phantom 4 Multispectral unmanned aerial vehicle (UAV) (DJI Technology Co., Ltd., Shenzhen, China). The Phantom 4 Multispectral drone is equipped with multispectral sensors, capturing the following spectral regions: Blue (450 ± 16 nm), Green (560 ± 16 nm), Red (650 ± 16 nm), Red Edge (730 ± 16 nm), and Near-Infrared (NIR) (840 ± 26 nm).
Flights were conducted at an altitude of 110 m above ground level, maintaining a consistent ground speed of approximately 4 ± 1 m/s, which ensured suitable spatial resolution and an overlap of 70% both laterally and vertically for photogrammetric processing. Flight campaigns were scheduled at least twice per growing season, typically at the beginning of the vegetation season, in the mid-season, and at the end of the growing season. In 2023, imagery was captured in April, July, and October. In 2024, data acquisition was conducted in August and October.

2.4. Image Processing and Tree Crown Delineation

A 5-band multispectral orthomosaic was produced in Agisoft Metashape Professional (version 1.6.4) using Structure-from-Motion (SfM) photogrammetry methods [30,31]. The near-infrared (NIR) band was used for image alignment, building a dense cloud, and generating a digital surface model (DSM). The original orthomosaics spatial resolution varied from 4 to 6 cm/pixel between different UAV campaigns. We resampled all images to 10 cm/pixel resolution using the Resample tool in ArcMap 10.5 (ESRI, Redlands, CA, USA) to ensure spatial consistency and thereby reduce the overall volume of raster data for analysis.
To delineate tree crowns, we used a hybrid approach combining automated classification and manual correction following the approach described in Desaine et al. [27]. Initially, the Iterative Self-Organising Data Analysis Technique Algorithm (ISODATA) clustering algorithm was applied using the Iso Cluster Unsupervised Classification tool in ArcMap 10.5. This unsupervised classification tool effectively grouped pixels in the multispectral imagery into several spectral clusters corresponding to distinct object classes (e.g., high vegetation (tree crowns), low vegetation, and ground). The cluster representing tree crowns was then isolated and exported as a new raster layer. This raster was subsequently converted into vector polygons representing individual tree crowns. However, due to overlapping crowns, manual adjustments were necessary in some instances to separate adjacent trees.
Tree crown polygons were used as spatial masks to extract individual crown areas from multispectral orthomosaics. The resulting raster cells were converted into point features, where each pixel was represented as a vector point. Using the Extract Multi Values to Points tool in ArcMap 10.5, reflectance values from all spectral bands were assigned to each point from the corresponding orthomosaic layers. The resulting attribute tables were then exported and imported into R 4.5.0 [32] for vegetation index calculation and further statistical analysis.

2.5. Vegetation Indices and Statistical Analysis

Multispectral reflectance data were used to calculate vegetation indices, particularly focusing on those sensitive to various aspects of canopy vitality, such as greenness, chlorophyll content, and physiological stress. The selection of these indices was based on existing literature on vegetation remote sensing, emphasising formulations known to detect structural and physiological changes in coniferous crowns. A relatively broad set of indices was tested, as many of them were originally developed for satellite or airborne platforms and have not been systematically validated for UAV-based multispectral imagery in dense coniferous forest stands. This comprehensive approach enabled us to identify which indices were most responsive and reliable under our specific conditions and research aims. A complete list of index formulas and spectral band combinations is provided in Appendix A (Table A1). Reflectance values were averaged per tree crown, and shadowed pixels were excluded to avoid radiometric distortion [33]. All statistical analyses were performed in R version R 4.5.0 [32], using the packages lme4 [34], lmerTest [35], emmeans [36], and glmmTMB [37]. For each vegetation index, a separate linear mixed-effects model was fitted. To evaluate index differences across tree groups and seasons, the fixed effects included tree group, season, and their interaction. The random effects structure accounted for repeated measurements and spatial clustering by including stand ID and tree ID as random intercepts, and a first-order autocorrelated error structure was added to account for the serial correlation between observations between seasons.
y = µ + Tij + Sij + Tij × Sij + (Yij) + εij
where Tij—the effect of tree group; Sij—the effect of season; Tij × Sij—the effect of tree group by season interaction; (Yij)—the random effect (intercept) of a tree nested within a stand; εij—the first-order autocorrelated errors.
Post hoc comparisons between seasons within each group were performed using Tukey’s HSD test based on estimated marginal means (emmeans). To assess relative index change, the difference between each index value and its corresponding value in summer 2023 (July 2023) was calculated. These delta values were then modelled using a separate linear mixed effects model per index, with tree group as the fixed effect and the same random structure. Group-level comparisons of these changes were again tested using Tukey-adjusted post hoc tests. All models were evaluated for residual normality and homoscedasticity. Notably, in the analysis of temporal changes in vegetation indices, some trees underwent changes in group classification between seasons. Specifically, trees initially classified as Healthy or Crown Damage in 2023 that fully desiccated by 2024 were reassigned to a separate transitional group labelled “Crown Damage 2023” or “Crown Damage 2024”, based on the season in which the mortality symptoms first became apparent. This allowed index dynamics to be analysed in terms of progressive decline, supporting the detection of pre-mortality spectral changes. These trees were thus included in both seasonal comparisons and within-group temporal change models as partially repeated observations.
To ensure meaningful differentiation and avoid redundancy in the analysis, only those spectral indices that exhibited statistically significant differences between the Bent and Healthy tree groups in at least one observation season were selected. Furthermore, indices were retained for further evaluation only if their pairwise correlation coefficients remained below a threshold of r2 = 0.9. This criterion minimised multicollinearity, thereby strengthening the physiological interpretability and statistical robustness of the derived relationships.
The Pearson correlation coefficient between indexes using data from all years was calculated separately for Healthy and Bent trees. Then, the difference between the identical pairs of indices of Healthy and Bent trees was calculated. To assess the significance of the difference, the randomisation test was applied as follows: the tree was randomly assigned to the Healthy or Bent trees group (keeping the group size the same as for the original data and using the same group for all observations of the same tree), and then correlations and differences were calculated. This process was repeated 999 times. The absolute values of the differences were compared to the original absolute difference of correlation coefficients. If the difference was larger than or equal to the original one in fewer than 5% of cases, the difference was assumed to be statistically significant.

3. Results and Discussion

3.1. Simulation of Wind Damage and Differences in Vegetation Indices in Norway Spruce Stands

In our study, we simulated wind-induced damage on Norway spruce trees by conducting mechanical pulling tests on Norway spruce (Picea abies) trees, complemented by remote sensing imagery. In the experiment, substantial differences were observed in several spectral indices, indicating significant physiological and structural changes in the plants. This finding aligns with the results of Brovkina et al. [38], who demonstrated that time-series airborne hyperspectral and satellite data can effectively track Norway spruce decline and that vegetation indices (including NDVI and the Plant Senescence Reflectance Index (PSRI)) very sensitively indicate health deterioration under stress conditions.
The groups of indices that showed statistically significant differences between Bent and Healthy trees primarily reflected leaf structural damage (e.g., the Anthocyanin Content Index (ACI), TCARI/OSAVI, the Modified Chlorophyll Absorption in Reflectance Index (MCARI2)) and changes in pigment composition (the Normalised Difference Red Edge (NDRE), SIPI) (Figure 2). Similar structural and pigment-driven index shifts are described by Carter and Knapp [39], who emphasise that differences in reflectance around 700 nm (NIR) are universal stress indicator related to a decrease in chlorophyll concentration and leaf structural damage. Junttila et al. [40] found that a reduction in NIR reflectance is one of the primary indicators of cell structural damage and reduced water content following damage caused by bark beetles or other stressors, also in spruce species. In our analysis, as revealed by post hoc Tukey’s HSD tests (p < 0.05), we found that MCARI and TCARI/OSAVI respond particularly well to changes in leaf structure. At the same time, SIPI indicates pigment dynamics, such as an increase in carotenoids and anthocyanins during stress (Figure 2). Examples of images illustrating the spatial and temporal dynamics of vegetation indices for different crown groups analysed in this study are provided in Appendix A (Figure A1).
A marked decrease in NIR reflectance in the Bent trees, as indicated by several of these indices (e.g., Datt, NDRE, the Modified Anthocyanin Reflectance Index (MARI), and MCARI), points to damage to the cell structure, which reduces radiation scattering and indicates decreased leaf integrity or water status. Such physiological effects were consistently detected in both field and lab conditions in studies using NIR-based indices for stress monitoring [39,41,42,43]. Similarly, pigment indices (SIPI) show an increasing proportion of stress pigments, consistent with patterns observed under crown disruption and oxidative stress in both conifers and broad-leaved species.
Since Norway spruce is a species with high shade tolerance and is well-adapted to stable micro-environmental conditions, sudden structural changes, such as stem cell damage, can significantly impact its photosynthetic balance and pigment dynamics. Rautiainen et al. [42] review emphasises that the physiological specificity of spruce needles and trees determines a unique spectral response to stress factors, making these species particularly suitable for spectral monitoring. Moreover, recent UAV-based studies provide further evidence that such spectral responses are not static but shift seasonally and structurally [25,43,44]. Desaine et al. [27] demonstrated that vegetation indices, particularly those utilising RedEdge and NIR bands, exhibited improved separability between healthy and damaged spruce in early autumn, which is in line with our observation that pigment and structure-sensitive indices are most effective at the stressed periods.
Although SIPI values in the early observation periods showed a similar distribution across tree groups, this suggests a relatively stable chlorophyll and carotenoid ratio, indicating that the observed differences are more likely related to canopy structural properties and illumination conditions rather than physiological stress changes. Moreover, the changes in values of NIR and Red-based indices observed in October 2024, particularly in trees that were bent earliest (Bent 2020 and Bent 2023) (Figure 2). For instance, DATT index values markedly increased in damaged tree classes, which in theory would indicate higher chlorophyll content and active photosynthesis. However, such increases in DATT are more plausibly attributable to structural changes in needles (e.g., increased leaf area index) rather than an actual rise in chlorophyll content as suggested by the SIPI index. Jones and Vaughan [45] declared that tree crown structural traits can affect optical properties, while Haboudane et al. [46] showed that in sparser canopies, light penetrates deeper and can lead to misleading interpretations of chlorophyll variation. Additionally, saturation effects in the red spectral band [47], in our study, particularly in healthy trees, may result in stable index values across certain seasons or reduced differences between groups, regardless of their actual vitality status. These effects may be further amplified by seasonal illumination factors, such as canopy exposure and lower solar elevation in autumn, thereby masking fundamental differences and producing spurious statistical similarity between groups.
Overall, our results demonstrate that the physiological condition of spruce is substantially disrupted following static tree-pulling experiments, as reflected in the spectral indices. Furthermore, these indices can serve as a non-invasive tool for assessing the impact of wind damage in forest ecosystems. Similarly, Brovkina et al. [38] and Lausch et al. [25] note that vegetation indices, especially multivariate models, exhibit high accuracy in monitoring tree health and enable the distinction of nuances in physiological responses across different seasons and under various stressors. Moreover, this allows forest managers to spatially identify areas where trees are likely to be physiologically stressed, even before visible symptoms, and adjust thinning regimes. For example, indices such as NDRE, SIPI, and MCARI2, which showed consistent differences between Bent and Healthy trees, can be used to identify individuals and carry out silvicultural interventions more effectively by prioritising trees to be removed.

3.2. Correlation Shifts Between Spectral Indices Reveal Physiological Reorganisation After Mechanical Damage to Norway Spruce Stems

The comparison of correlation coefficients between spectral indices in Bent and Healthy trees (r = 0.91 vs. 0.60, respectively) revealed substantial shifts in inter-index relationships, reflecting distinct physiological reactions (Figure 3). For instance, the correlation between Datt and SIPI indices, which represent chlorophyll absorption characteristics and pigment balance, respectively, increased significantly in Bent trees (Δr + 0.31). This trend suggests that mechanical damage may intensify the coupling between pigment degradation and chlorophyll absorption processes, particularly in the Red and near-infrared spectral bands. Potentially reflecting stress-related changes in NIR could be linked to internal water content and structural disintegration. Similar patterns have been observed in previous studies, which reinforce the notion that conifer crowns under stress exhibit tighter linkages between chlorophyll degradation and pigment balance. These shifts in spectral coupling support the use of inter-index correlations as sensitive proxies for crown condition. Notably, Huo et al. [43] reported a comparable reinforcement of pigment–chlorophyll linkages in Picea abies using multispectral UAV imagery, where indices sensitive to pigments and water content demonstrated heightened spectral responsiveness under bark beetle-induced physiological stress.
The relationship between MARI and Datt also reversed; while these indices were positively correlated in Healthy trees, they became negative in Bent ones. This suggests a decoupling of the chlorophyll signal from pigment structural modulation during structural damage. Such shifts can result from divergent responses to biotic stress, e.g., one index reflecting pigment dispersion, while another responds to dehydration stress. Similar patterns were documented by Junttila et al. [40], who found that bark beetle damage altered relationships between chlorophyll-related indices in conifers. Conversely, MARIs reduced correlation with structural and pigment index (e.g., ACI, MCARI) in Bent trees suggested chlorophyll absorption efficiency, as captured by MCARI, and aligns more closely with vegetation density indicators like DVI when foliage remains intact, while structural disruption weakens this association. A comparison of MCARI and DVI correlation strength between Bent (r = 0.61) and Healthy trees (r = 0.81) indicated altered spectral responses due to structural damage. This aligns with findings from Brovkina et al. [38], who reported that under canopy degradation driven by bark beetle infestation, energy-balance-related indices become increasingly separate from pigment-sensitive metrics. These observed correlation dynamics emphasise that spectral index relationships are not static but evolve in response to physiological and biomechanical changes. Furthermore, the variability of conifer reflectance under stress underscores the diagnostic value of inter-index correlation analysis for early detection of structural damage and functional disruption [42].

3.3. The Temporal Dynamics of Vegetation Indices as an Effect of Stem Pulling Experiments

The UAV imagery acquisitions were obtained at key intervals before and following the tree-pulling experiment, specifically in July 2023 (baseline, immediately post-pulling), October 2023 (end of vegetation season), August 2024 (middle of vegetation season), and October 2024 (end of vegetation season). This temporal framework allowed us to monitor the progression of canopy reflectance changes over more than one year, thus allowing for the assessment of the potential effect of the stem-pulling experiment. The dynamics of vegetation index values across different seasons showed that several indices exhibited significant differences in estimated marginal means between Bent and Healthy tree groups (Figure 4), confirming the well-established sensitivity of spectral indices to the physiological status of undisturbed, photosynthetically active canopies [48]. A notable decline in spectral indices following breakage was not immediate, as most of the spectral indices did not show any changes after pulling tests (July 2023 vs. October 2023 or July 2023 vs. August 2024); the most notable variation in normalised spectral indices values were found between July 2023 vs. October 2024, especially for Bent 2020 trees. Although the Renormalised Difference Vegetation Index (RDVI) and MARI index demonstrated the most significant contrasts between Bent and Healthy tree groups already in the comparison of “July 2023 vs. October 2023”. We found that the predicted mean values of the RDVI index for Bent 2023 were decreased by 8.44, compared to 7.44 in Healthy trees. Similar trends were observed for MARI and the Ratio Analysis of Reflectance Spectra (RARSc) indices, where index values significantly reduced in one year following the tree-pulling experiment. Namely, when comparing July 2023 vs. October 2024, we found that for Bent 2023, the MARI index decreased by 1.20 compared to a 0.77 reduction in Healthy trees. However, in other observation periods, no statistically significant differences between these groups were detected. This suggests that trees can quickly compensate for stress, perhaps by mobilising resources or enhancing photosynthetic efficiency in their tissue compartments. Over time, these resources can be depleted due to the cumulative stress effects, especially when additional environmental stressors such as drought or insect infestation are present, as also described in recent studies of canopy-scale reflectance fluctuations, particularly foliar diseases, which manifest subtle changes in reflectance before visual symptoms arise [49,50,51,52]. These findings align with previous studies that have demonstrated the sensitivity of red-edge and near-infrared-based indices (such as RDVI and MARI) to vegetation stress and decline, particularly under biotic disturbance [43,53]. For example, Gamon and Surfus [41] and Huo et al. [52] reported that reflectance-based indices could distinguish between defoliated and healthy trees at early disease stages, confirming their utility in operational forest health assessments. For example, by integrating into forest management workflows as part of an early warning system, where threshold changes in indices, such as a marked drop in RDVI or MARI values, may trigger further inspection or preventive action. UAV-based index mapping can thus support operational decision-making, such as marking high-risk trees for removal or defining management zones with increased wind vulnerability. As a result, managers can optimise monitoring frequency, prioritise felling, and reduce the likelihood of secondary disturbances, facilitating more resilient stand-level planning. Nevertheless, one of the primary limitations of UAV-based approaches is their restricted spatial coverage, which presents a challenge for monitoring large-scale forest areas. To address this, UAV data could be strategically integrated with satellite-borne imagery, such as Sentinel-2, which offers broad spatial coverage and frequent revisit cycles. However, Sentinel-2 spatial resolution (10 m) limits its suitability for detecting early damage signals at the individual tree level, particularly in dense coniferous stands. Therefore, current applications of satellite imagery may be more appropriate for identifying areas with widespread or advanced damage.
Interestingly, the estimated marginal means differences for Bent 2020 trees did not differ significantly from those of Healthy trees for any of the analysed indices. These results support our hypothesis that multi-year UAV-based multispectral monitoring enables the detection of both delayed stress responses and potential recovery trajectories in individual trees. However, our data reveal spectral signatures of stress in tree crowns as decreases in vegetation indices values following mechanical damage; the precise physiological mechanisms underlying these responses remain only partially understood. It is still unclear whether the observed changes in needle reflectance are attributable solely to wind-induced structural damage or whether additional biotic and abiotic factors, such as drought or insect activity, also play a significant role, as suggested by the decrease in index values observed even in the Healthy tree groups, although to a lesser extent than in Bent trees. Future studies should focus on more detailed measurements of individual tree physiological parameters, such as changes in water flow, as well as on analyses of needle pigment composition and degradation, to elucidate the pathways linking physical damage to changes in crown reflectance. Such integrative approaches would improve our understanding of tree responses to stress and could facilitate the early detection of tree stress in multispectral imagery.

4. Conclusions

Overall, our findings confirm that multispectral drone imagery, combined with advanced vegetation index analysis, enables the early detection and long-term monitoring of wind-induced damage in Norway spruce. Mechanical stem pulling, simulating wind stress, caused significant physiological and structural disruptions that were reliably detected through reductions in Red, RedEdge, and NIR reflectance indices. The best identification of early signs of stress was observed in indices sensitive to chlorophyll content and leaf structural integrity, such as MCARI, TCARI/OSAVI, and SIPI. The decline in vegetation indices among damaged trees was most pronounced not immediately after disturbance but several seasons later, underscoring the cumulative nature of physiological stress and the delayed manifestation of canopy decline.

Author Contributions

Conceptualization, E.B.; methodology, E.B., O.K., A.S., and J.J.-C.C.; software, D.E.; investigation, E.B., D.E., and D.K.; resources, E.B.; data curation, A.S. and J.J.-C.C.; writing—original draft preparation, D.K., E.B.; writing—review and editing, E.B., O.K., and D.E.; visualization, E.B.; supervision, E.B.; project administration, E.B. All authors have read and agreed to the published version of the manuscript.

Funding

The study was prepared under the framework and funded by the “Effect of climate change on forestry and associated risks” project, Latvia’s State Forests, No 5-5.9.1_007p_101_21_78 and by the project nr 23-A01612-000005 “Development of automatic system for growth, health and weeds monitoring of tree saplings,” which is implemented under sub-measure 16.12 of measure 16, “Cooperation,” of the Latvian Rural Development Programme for 2014–2020: “Support for the implementation of the EIP Agricultural Productivity and Sustainability Working Group project.”

Data Availability Statement

Data are contained within the article.

Conflicts of Interest

The authors declare no conflicts of interest.

Appendix A

Table A1. Indices and formula for calculation.
Table A1. Indices and formula for calculation.
Spectral IndexFormulaReference
ACIGreen/NIR[54]
ARI(1/Green) − (1/RedEdge)[55]
Carter2Red/RedEdge[56]
CCI(Green − Red)/(Green + Red)[57]
CI(NIR/RedEdge) − 1[54]
CRI11/Blue − 1/Green[55]
CRI21/Blue − 1/RedEdge[55]
CRI31/Blue − 1/Green × RedEdge[58]
CRI41/Blue − 1/RedEdge × RedEdge[58]
Datt(NIR − RedEdge)/(NIR − Red)[59]
Datt4Red/(Green × RedEdge)[60]
Datt6NIR/Green × RedEdge[60]
DVINIR − Red[61]
DWSI4Green/Red[62]
EVI2.5 × ((NIR − Red)/(NIR − 6 × Red − 7.5 × Blue + 1))[63]
Gitelson1/RedEdge[58]
Gitelson2(NIR/Red) − 1[64]
GNDVI(NIR − Green)/(NIR + Green)[65]
MACINIR/Green[54]
MARI(1/Green) − (1/RedEdge) × NIR[58]
MCARI(RedEdge − Red) − 0.2 × (RedEdge − Green) × (RedEdge/Red)[66]
mCARI_11.2 × ((2.5 × RedEdge − Red) − 1.3 × (RedEdge − Green))[46]
MCARI2((RedEdge − Red) − 0.2 × (RedEdge − Green)) × (RedEdge − Red)[66]
MPRI(Blue − Green)/(Blue + Green)[67]
MTVI1.2 × (1.2 × (NIR − Green) − 2.5 × (Red − Green))[46]
MTVI_11.2 × (1.2 × (1.2 × (RedEdge − Green) − 2.5 × (Red − Green)))[46]
MTVI_21.5 × (1.2 × (RedEdge − Green) − 2.5 × (Red − Green))/sqrt ((2 × RedEdge + 1)2 − (6 × RedEdge − 5 × sqrt(Red) − 0.5))[46]
NCPI(Red − Blue)/(Red + Blue)[68]
NDRE(NIR − RedEdge)/(NIR + RedEdge)[69]
NDVI(NIR − Red)/(NIR + Red)[61]
NLI/(NIR2 + Red/NIR2 − Red)[70]
OSAVI(1 + 0.16) × (NIR − Red)/(NIR + Red + 0.16)[71]
PSRI(Red − Blue)/RedEdge[72]
RARScNIR/Blue[56]
RDVI(RedEdge − Red)/sqrt(RedEdge + Red)[73]
REGI(RedEdge − Green)/(RedEdge + Green)[56]
RERI/NDV_761(RedEdge − Red)/(RedEdge + Red)[56]
RGIRed/Green[74]
SAVI(NIR − Red) × 1.5/(NIR + Red + 0.5)[61]
SAVI1(1 + 0.5)/(NIR − Red)/(NIR + Red + 0.5)[75]
SIPI(NIR/Blue)/(NIR − Red)[76]
SPVI0.4 × 3.7 × (NIR − Red) − 1.2 × ((Green − Red) × 2) × 0.5[64]
SRNIR/Red[77]
SR4RedEdge/Red[77]
SR8Blue/Green[77]
TCARI3 × ((RedEdge − Red) − 0.2 × (RedEdge − Green) × (RedEdge/Red))[46]
TCARI/OSAVITCARI/OSAVI[46]
TGI−0.5 × (190 × (Red − Green) − 120 × (Red − Blue))[78]
TVI0.5 × (120 × (RedEdge − Green) − 200 × (Red − Green))[46]
Forests 16 01348 g0a1
Figure A1. Examples of images illustrating the spatial and temporal dynamics of vegetation indices for different crown groups analysed in this study. Each panel represents a different year of bending, along with associated canopy damage and areas of healthy vegetation. The top row of images in each panel displays an orthomosaic showing the specific group locations, while subsequent rows visualize various vegetation indices.
Figure A1. Examples of images illustrating the spatial and temporal dynamics of vegetation indices for different crown groups analysed in this study. Each panel represents a different year of bending, along with associated canopy damage and areas of healthy vegetation. The top row of images in each panel displays an orthomosaic showing the specific group locations, while subsequent rows visualize various vegetation indices.
Forests 16 01348 g0a1

References

  1. Anderegg, W.R.; Wu, C.; Acil, N.; Carvalhais, N.; Pugh, T.A.; Sadler, J.P.; Seidl, R. A climate risk analysis of Earth’s forests in the 21st century. Science 2022, 377, 1099–1103. [Google Scholar] [CrossRef]
  2. Patacca, M.; Lindner, M.; Lucas-Borja, M.E.; Cordonnier, T.; Fidej, G.; Gardiner, B.; Hauf, Y.; Jasinevičius, G.; Labonne, S.; Linkevičius, E. Significant increase in natural disturbance impacts on European forests since 1950. Glob. Chang. Biol. 2023, 29, 1359–1376. [Google Scholar] [CrossRef]
  3. Hanewinkel, M.; Cullmann, D.; Schelhaas, M.J.; Nabuurs, G.J.; Zimmermann, N.E. Climate change may cause severe loss in the economic value of European forestland. Nat. Clim. Change 2013, 3, 203–207. [Google Scholar] [CrossRef]
  4. Schwarzbauer, P.; Rauch, P. Impact on industry and markets—Roundwood prices and procurement risks. In Living with Storm. Damage to Forests—What Science Can Tell Us; Gardiner, B., Schuck, A., Schelhaas, M.J., Orazio, C., Blennow, K., Nicoll, B., Eds.; European Forest Institute: Joensuu, Finland, 2013; p. 130. [Google Scholar]
  5. Thürig, E.; Hagedorn, F.; Lindroth, A. The Influence of Storm Damage on the Forest Carbon Balance. In Living with Storm Damage to Forests; Gardiner, B., Schuck, A., Schelhaas, M.J., Orazio, C., Blennow, K., Nicoll, B., Eds.; European Forest Institute: Joensuu, Finland, 2013; pp. 47–54. [Google Scholar]
  6. Schelhaas, M.J.; Fridman, J.; Hengeveld, G.M.; Henttonen, H.M.; Lehtonen, A.; Kies, U.; Krajnc, N.; Lerink, B.; Dhubhain, A.N.; Polley, H.; et al. Actual European forest management by region, tree species and owner based on 714,000 re-measured trees in national forest inventories. PLoS ONE 2018, 13, e0207151. [Google Scholar] [CrossRef]
  7. Seidl, R.; Schelhaas, M.-J.; Lexer, M. Unraveling the drivers of intensifying forest disturbance regimes in Europe. Glob. Chang. Biol. 2011, 17, 2842–2852. [Google Scholar] [CrossRef]
  8. Turkulainen, E.; Honkavaara, E.; Näsi, R.; Oliveira, R.A.; Hakala, T.; Junttila, S.; Karila, K.; Koivumäki, N.; Pelto-Arvo, M.; Tuviala, J.; et al. Comparison of Deep Neural Networks in the Classification of Bark Beetle-Induced Spruce Damage Using UAS Images. Remote Sens. 2023, 15, 4928. [Google Scholar] [CrossRef]
  9. Klimo, E.; Hager, H.; Kulhavý, J. Spruce Monocultures in Central Europe—Problems and Prospects; EFI Proceedings; European Forest Institute: Joensuu, Finland, 2000; p. 33.
  10. Spiecker, H.; Hansen, J.; Klimo, E.; Skovsgaard, J.P.; Sterba, H.; von Teuffel, K. Norway Spruce Conversion—Options and Consequences; EFI Research Report 18; Brill: Leiden, The Netherlands; Boston, MA, USA; Köln, Germany, 2004; p. 320. [Google Scholar]
  11. Synek, M.; Janda, P.; Mikoláš, M.; Nagel, T.A.; Schurman, J.S.; Pettit, J.L.; Trotsiuk, V.; Morrissey, R.C.; Bače, R.; Čada, V.; et al. Contrasting patterns of natural mortality in primary Picea forests of the Carpathian Mountains. For. Ecol. Manag. 2020, 457, 117734. [Google Scholar] [CrossRef]
  12. Jansons, Ā.; Bāders, E.; Zeltiņš, P.; Gailis, A.; Šņepsts, G.; Katrevičs, J. Meža selekcijas potenciāls egļu audžu ražības paaugstinašanā. In Vienvecuma Egļu Meži Latvijā; Jansons, J., Ed.; Dauagvpils Universitātes Akadēmiskais Apgāds “Saule”: Daugavpils, Latvia, 2019; pp. 99–118. [Google Scholar]
  13. Johnstone, D.; Moorea, G.; Tauszband, M.; Nicolas, M. The measurement of plant vitality in landscape trees. Arboric. J. Int. J. Urban For. 2013, 35, 18–27. [Google Scholar] [CrossRef]
  14. Lechner, A.M.; Foody, G.M.; Boyd, D.S. Applications in remote sensing to forest ecology and management. One Earth 2020, 2, 405–412. [Google Scholar] [CrossRef]
  15. Barta, V.; Hanus, J.; Dobrovolny, L.; Homolova, L. Comparison of field survey and remote sensing techniques for detection of bark beetle-infested trees. For. Ecol. Manag. 2022, 506, 119984. [Google Scholar] [CrossRef]
  16. Glenn, E.P.; Huete, A.R.; Nagler, P.L.; Nelson, S.G. Relationship between remotely-sensed vegetation indices and plant physiological processes: What vegetation indices can and cannot tell us about the landscape. Sensors 2008, 8, 2136. [Google Scholar] [CrossRef]
  17. Jin, Z.; Prasad, R.; Shriver, J.; Zhuang, Q. Crop model- and satellite imagery-based recommendation tool for variable rate N fertilizer application for the US Corn system. Precis. Agric. 2017, 18, 779–800. [Google Scholar] [CrossRef]
  18. Magney, T.S.; Eitel, J.U.H.; Vierling, L.A. Mapping wheat nitrogen uptake from RapidEye vegetation indices. Precis. Agric. 2017, 18, 429–451. [Google Scholar] [CrossRef]
  19. Xue, J.; Su, B. Significant remote sensing vegetation indices: A review of developments and applications. J. Sens. 2017, 2017, 1353691. [Google Scholar] [CrossRef]
  20. Chuvieco, E. Fundamentals of Satellite Remote Sensing: An Environmental Approach, 3rd ed.; CRC Press Taylor & Francis Group: Boca Raton, FL, USA, 2020. [Google Scholar]
  21. Nazeer, M.; Sing Wong, M.; Yu, X.; Yin Tung Kwok, C.; Peng, Q.; Dai, Y. Urban tree health assessment using multifaceted remote sensing datasets: A case study in Hong Kong. Remote Sens. Appl. Soc. Environ. 2024, 36, 101347. [Google Scholar] [CrossRef]
  22. Gardiner, B.; Berry, P.; Moulia, B. Review: Wind impacts on plant growth, mechanics and damage. Plant Sci. 2016, 245, 94–118. [Google Scholar] [CrossRef]
  23. Carter, G.A. Responses of leaf spectral reflectance to plant stress. Am. J. Bot. 1993, 80, 239–243. [Google Scholar] [CrossRef]
  24. Buddenbaum, H.; Stern, O.; Paschmionka, B.; Hass, E.; Gattung, T.; Stoffels, J.; Hill, J.; Werner, W. Using VNIR and SWIR field imaging spectroscopy for drought stress monitoring of beech seedlings. Int. J. Remote Sens. 2015, 36, 4590–4605. [Google Scholar] [CrossRef]
  25. Lausch, A.; Erasmi, S.; King, D.J.; Magdon, P.; Heurich, M. Understanding forest health with remote sensing-part I—A review of spectral traits, processes and remote-sensing characteristics. Remote Sens. 2016, 8, 1029. [Google Scholar] [CrossRef]
  26. Baders, E.; Romāns, E.; Desaine, I.; Krišāns, O.; Seipulis, A.; Donis, J.; Jansons, Ā. An Integration of Linear Model and ‘Random Forest’ Techniques for Prediction of Norway Spruce Vitality: A Case Study of the Hemiboreal Forest, Latvia. Remote Sens. 2022, 14, 2122. [Google Scholar] [CrossRef]
  27. Desaine, I.; Bāders, E.; Romāns, E.; Krišāns, O.; Seipulis, A.; Jansons, Ā. The Detection of Disturbance Effects of Norway Spruce (Picea abies (L.) KARST.) in UAV Multispectral Imagery. ISPRS Ann. Photogramm. Remote Sens. Spat. Inf. Sci. 2022, 3, 689–695. [Google Scholar] [CrossRef]
  28. Detter, A.; Rust, S.; Rust, C.; Maybaum, G. Determining strength limits for standing tree stems from bending tests. In Proceedings of the 18th International Non-Destructive Testing and Evaluation of Wood Symposium, Madison, WI, USA, 24–27 September 2013; Ross, R.J., Wang, X., Eds.; U.S. Department of Agriculture, Forest Service, Forest Products Laboratory: Madison, WI, USA, 2013; p. 226. [Google Scholar]
  29. Krisans, O.; Matisons, R.; Rust, S.; Burnevica, N.; Bruna, L.; Elferts, D.; Kalvane, L.; Jansons, A. Presence of Root Rot Reduces Stability of Norway Spruce (Picea abies): Results of Static Pulling Tests in Latvia. Forests 2020, 11, 416. [Google Scholar] [CrossRef]
  30. Verhoeven, G. Taking computer vision aloft-archaeological three-dimensional reconstructions from aerial photographs with photoscan. Archaeol. Prospect. 2011, 18, 67–73. [Google Scholar] [CrossRef]
  31. Westoby, M.J.; Brasington, J.; Glasser, N.F.; Hambrey, M.J.; Reynolds, J.M. ‘Structure-from-Motion’ photogrammetry: A low-cost, effective tool for geoscience applications. Geomorphology 2012, 179, 300–314. [Google Scholar] [CrossRef]
  32. R Core Team. R: A Language and Environment for Statistical Computing; R Foundation for Statistical Computing: Vienna, Austria, 2025; Available online: https://www.R-project.org/ (accessed on 14 August 2025).
  33. Zhang, L.; Sun, X.; Wu, T.; Zhang, H. An Analysis of Shadow Effects on Spectral Vegetation Indexes Using a Ground-Based Imaging Spectrometer. IEEE Geosci. Remote Sens. Lett. 2015, 12, 11. [Google Scholar] [CrossRef]
  34. Bates, D.; Mächler, M.; Bolker, B.; Walker, S. Fitting linear mixed-effects models using lme4. J. Stat. Softw. 2015, 67, 1–48. [Google Scholar] [CrossRef]
  35. Kuznetsova, A.; Brockhoff, P.B.; Christensen, R.H. lmerTest package: Tests in linear mixed effects models. J. Stat. Softw. 2017, 82, 1–26. [Google Scholar] [CrossRef]
  36. Lenth, R.V. emmeans: Estimated Marginal Means, aka Least-Squares Means. 2024, R package version 1.10.5-0900001. Available online: https://CRAN.R-project.org/package=emmeans (accessed on 14 August 2025).
  37. Brooks, M.E.; Kristensen, K.; van Benthem, K.J.; Magnusson, A.; Berg, C.W.; Nielsen, A.; Skaug, H.J.; Maechler, M.; Bolker, B.M. glmmTMB Balances Speed and Flexibility Among Packages for Zero-inflated Generalized Linear Mixed Modeling. R J. 2017, 9, 378–400. [Google Scholar] [CrossRef]
  38. Brovkina, O.; Cienciala, E.; Zemek, F.; Lukeš, P.; Fabianek, T.; Russ, R. Composite indicator for monitoring of Norway spruce stand decline. Eur. J. Remote Sens. 2017, 50, 550–563. [Google Scholar] [CrossRef]
  39. Carter, G.A.; Knapp, A.K. Leaf optical properties in higher plants: Linking spectral characteristics to stress and chlorophyll concentration. Am. J. Bot. 2001, 88, 677–684. [Google Scholar] [CrossRef]
  40. Junttila, S.; Näsi, R.; Koivumäki, N.; Imangholiloo, M.; Saarinen, N.; Raisio, J.; Honkavaara, E. Multispectral imagery provides benefits for mapping spruce tree decline due to bark beetle infestation when acquired late in the season. Remote Sens. 2022, 14, 909. [Google Scholar] [CrossRef]
  41. Gamon, J.A.; Surfus, J.S. Assessing leaf pigment content and activity with a reflectometer. New Phytol. 1999, 143, 105–117. [Google Scholar] [CrossRef]
  42. Rautiainen, M.; Lukeš, P.; Homolová, L.; Hovi, A.; Pisek, J.; Mõttus, M. Spectral Properties of Coniferous Forests: A Review of In Situ and Laboratory Measurements. Remote Sens. 2018, 10, 207. [Google Scholar] [CrossRef]
  43. Huo, L.; Lindberg, E.; Bohlin, J.; Persson, H.J. Assessing the detectability of European spruce bark beetle green attack in multispectral drone images with high spatial-and temporal resolutions. Remote Sens. Environ. 2023, 287, 113484. [Google Scholar] [CrossRef]
  44. Honkavaara, E.; Nasi, R.; Oliveira, R.; Viljanen, N.; Suomalainen, J.; Khoramshahi, E.; Hakala, T.; Nevalainen, O.; Markelin, L.; Vuorinen, M.; et al. Using Multitemporal Hyper- and Multispectral UAV imaging For Detecting Bark Beetle Infestation on Norway Spruce. Int. Arch. Photogramm. Remote Sens. Spat. Inf. Sci. 2020, 43, 429–434. [Google Scholar] [CrossRef]
  45. Jones, H.G.; Vaughan, R.A. Remote Sensing of Vegetation: Principles, Techniques, and Applications; OUP: Oxford, UK, 2010. [Google Scholar]
  46. Haboudane, D.; Miller, J.R.; Pattey, E.; Zarco-Tejada, P.J.; Strachan, I.B. Hyperspectral vegetation indices and novel algorithms for predicting green LAI of crop canopies: Modeling and validation in the context of precision agriculture. Remote Sens. Environ. 2004, 90, 337–352. [Google Scholar] [CrossRef]
  47. Zarco-Tejada, P.J.; Miller, J.R.; Morales, A.; Berjón, A.; Agüera, J. Hyperspectral indices and model simulation for chlorophyll estimation in open-canopy tree crops. Remote Sens. Environ. 2004, 90, 463–476. [Google Scholar] [CrossRef]
  48. Fassnacht, F.E.; Latifi, H.; Stereńczak, K.; Modzelewska, A.; Lefsky, M.; Waser, L.T.; Ghosh, A. Review of studies on tree species classification from remotely sensed data. Remote Sens. Environ. 2016, 186, 64–87. [Google Scholar] [CrossRef]
  49. Korpela, I.; Heikkinen, V.; Honkavaara, E.; Rohrbach, F.; Tokola, T. Variation and directional anisotropy of reflectance at the crown scale—Implications for tree species classification in digital aerial images. Remote Sens. Environ. 2011, 115, 2062–2074. [Google Scholar] [CrossRef]
  50. Minařík, R.; Langhammer, J.; Lendzioch, T. Automatic Tree Crown Extraction from UAS Multispectral Imagery for the Detection of Bark Beetle Disturbance in Mixed Forests. Remote Sens. 2020, 12, 4081. [Google Scholar] [CrossRef]
  51. Drechsel, J.; Forkel, M. Remote sensing forest health assessment—A comprehensive literature review on a European level. Cent. Eur. For. J. 2025, 71, 14–39. [Google Scholar] [CrossRef]
  52. Huo, L.; Matsiakh, I.; Bohlin, J.; Cleary, M. Estimation of Tree Vitality Reduced by Pine Needle Disease Using Multispectral Drone Images. Remote Sens. 2025, 17, 271. [Google Scholar] [CrossRef]
  53. Waite, O.J.; Coops, N.C.; Grubinger, S.; Isaac-Renton, M.; Degner, J.; King, J.; Liu, A. Responses of spectral indices to heat and drought differ by tree size in Douglas-fir. Sci. Remote Sens. 2025, 11, 100193. [Google Scholar] [CrossRef]
  54. Gitelson, A.A.; Gurlin, D.; Moses, W.J.; Barrow, T. A bio-optical algorithm for the remote estimation of the chlorophyll-a concentration in case2 waters. Environ. Res. Lett. 2009, 4, 045003. [Google Scholar] [CrossRef]
  55. Gitelson, A.; Merzlyak, M.; Zur, Y.; Stark, R.; Gritz, U. Non-Destructive and Remote Sensing Techniques for Estimation of Vegetation Status. Pap. Nat. Resour. 2001, 3543, 327–335. [Google Scholar]
  56. Blackburn, G.A. Quantifying chlorophylls and carotenoids at leaf and canopy scales: An evaluation of some hyperspectral approaches. Remote Sens. Environ. 1998, 66, 273–285. [Google Scholar] [CrossRef]
  57. Gamon, J.A.; Huemmrich, K.F.; Wong, C.Y.; Ensminger, I.; Garrity, S.; Hollinger, D.Y.; Peñuelas, J. A remotely sensed pigment index reveals photosynthetic phenology in evergreen conifers. Proc. Natl. Acad. Sci. USA 2016, 113, 13087–13092. [Google Scholar] [CrossRef]
  58. Gitelson, A.A.; Keydan, G.P.; Merzlyak, M.N. Three-band model for noninvasive estimation of chlorophyll, carotenoids, and anthocyanin contents in higher plant leaves. Geophys. Res. Lett. 2006, 33, L11402. [Google Scholar] [CrossRef]
  59. Datt, B. Remote sensing of water content in Eucalyptus leaves. Aust. J. Bot. 1999, 47, 909–923. [Google Scholar] [CrossRef]
  60. Datt, B. Remote sensing of chlorophyll a, chlorophyll b, chlorophyll a+b, and total carotenoid content in eucalyptus leaves. Remote Sens. Environ. 1998, 66, 111–121. [Google Scholar] [CrossRef]
  61. Lyon, J.G.; Yuan, D.; Lunetta, R.S.; Elvidge, C.D. A change detection experiment using vegetation indices. Photogramm. Eng. Remote Sens. 1998, 64, 143–150. [Google Scholar]
  62. Apan, A.; Held, A.; Phinn, S.; Markley, J. Formulation and assessment of narrow-band vegetation indices from EO-1 Hyperion imagery for discriminating sugarcane disease. In Proceedings of the Spatial Sciences Institute Biennial Conference (SSC 2003): Spatial Knowledge Without Boundaries, Canberra, Australia, 22–27 September 2003. [Google Scholar]
  63. Huete, A.R.; Liu, H.Q.; Batchily, K.V.; Van Leeuwen, W.J.D.A. A comparison of vegetation indices over a global set of TM images for EOS-MODIS. Remote Sens. Environ. 1997, 59, 440–451. [Google Scholar] [CrossRef]
  64. Main, R.; Cho, M.A.; Mathieu, R.; O’Kennedy, M.M.; Ramoelo, A.; Koch, S. An investigation into robust spectral indices for leaf chlorophyll estimation. ISPRS J. Photogramm. Remote Sens. 2011, 66, 751–761. [Google Scholar] [CrossRef]
  65. Gitelson, A.A.; Kaufman, Y.J.; Merzlyak, M.N. Use of a green channel in remote sensing of global vegetation from EOS-MODIS. Remote Sens. Environ. 1996, 58, 289–298. [Google Scholar] [CrossRef]
  66. Daughtry, C.S.; Walthall, C.; Kim, M.; de Colstoun, E.B.; McMurtrey, J. Estimating corn leaf chlorophyll concentration from leaf and canopy reflectance. Remote Sens. Environ. 2000, 74, 229–239. [Google Scholar] [CrossRef]
  67. Yang, Z.; Willis, P.; Mueller, R. Impact of Band-Ratio Enhanced AWIFS Image on Crop Classification Accuracy. 2008. Available online: https://www.asprs.org/a/publications/proceedings/pecora17/0041.pdf (accessed on 30 June 2025).
  68. Kawashima, S.; Nakatani, M. An algorithm for estimating chlorophyll content in leaves using a video camera. Ann. Bot. 1998, 81, 49–54. [Google Scholar] [CrossRef]
  69. Hunt, E.R., Jr.; Daughtry, C.S.T.; Eitel, J.U.; Long, D.S. Remote sensing leaf chlorophyll content using a visible band index. Agron. J. 2011, 103, 1090–1099. [Google Scholar] [CrossRef]
  70. Pu, R.; Gong, P.; Yu, Q. Comparative Analysis of EO-1 ALI and Hyperion, and Landsat ETM+ Data for Mapping Forest Crown Closure and Leaf Area Index. Sensors 2008, 8, 3744–3766. [Google Scholar] [CrossRef]
  71. Rondeaux, G.; Steven, M.; Baret, F. Optimization of soil-adjusted vegetation indices. Remote Sens. Environ. 1996, 55, 95–107. [Google Scholar] [CrossRef]
  72. Merzlyak, M.N.; Gitelson, A.A.; Chivkunova, O.B.; Rakitin, V.Y. Non-destructive optical detection of pigment changes during leaf senescence and fruit ripening. Physiol. Plant. 1999, 106, 135–141. [Google Scholar] [CrossRef]
  73. Roujean, J.-L.; Breon, F.-M. Estimating PAR absorbed by vegetation from bidirectional reflectance measurements. Remote Sens. Environ. 1995, 51, 375–384. [Google Scholar] [CrossRef]
  74. Ceccato, P.; Flasse, S.; Tarantola, S.; Jacquemoud, S.; Grégoire, J.M. Detecting vegetation leaf water content using reflectance in the optical domain. Remote Sens. Environ. 2001, 77, 22–33. [Google Scholar] [CrossRef]
  75. Huete, A.R. A soil-adjusted vegetation index (SAVI). Remote Sens. Environ. 1988, 25, 295–309. [Google Scholar] [CrossRef]
  76. Penuelas, J.; Baret, F.; Filella, I. Semi-Empirical Indices to Assess Carotenoids/Chlorophyll-a Ratio from Leaf Spectral Reflectance. Photosynthetica 1995, 32, 221–230. [Google Scholar]
  77. Ehammer, A.; Fritsch, S.; Conrad, C.; Lamers, J.; Dech, S. Statistical derivation of fPAR and LAI for irrigated cotton and rice in arid Uzbekistan by combining multi-temporal RapidEye data and ground measurements. In Remote Sensing for Agriculture, Ecosystems, and Hydrology XII; SPIE: Toulouse, France, 2010; Volume 7824, pp. 66–75. [Google Scholar]
  78. Hunt, E.R., Jr.; Doraiswamy, P.C.; McMurtrey, J.E.; Daughtry, C.S.; Perry, E.M.; Akhmedov, B. A visible band index for remote sensing leaf chlorophyll content at the canopy scale. Int. J. Appl. Earth Obs. Geoinf. 2013, 21, 103–112. [Google Scholar] [CrossRef]
Forests 16 01348 g001
Figure 1. Location of the studied sites in Latvia.
Figure 1. Location of the studied sites in Latvia.
Forests 16 01348 g001
Forests 16 01348 g002
Figure 2. Estimated marginal means (EMMs) ± standard error (SE) of vegetation indices across different seasons and groups.
Figure 2. Estimated marginal means (EMMs) ± standard error (SE) of vegetation indices across different seasons and groups.
Forests 16 01348 g002
Forests 16 01348 g003
Figure 3. A Heatmap showing differences in Pearson correlation coefficients (Δr) between vegetation index pairs for Bent and Healthy trees. Each cell in the matrix displays the correlation difference (Δr), with statistically significant changes annotated with * (p < 0.05).
Figure 3. A Heatmap showing differences in Pearson correlation coefficients (Δr) between vegetation index pairs for Bent and Healthy trees. Each cell in the matrix displays the correlation difference (Δr), with statistically significant changes annotated with * (p < 0.05).
Forests 16 01348 g003
Forests 16 01348 g004
Figure 4. Spectral indicators of tree health: seasonal dynamics and tree stress discrimination by using vegetation indices. (Statistically significant differences between groups in one season are indicated by different lowercase letters. If two means share the same letter, there is no statistically significant difference between them. If the letters differ, the difference is statistically significant).
Figure 4. Spectral indicators of tree health: seasonal dynamics and tree stress discrimination by using vegetation indices. (Statistically significant differences between groups in one season are indicated by different lowercase letters. If two means share the same letter, there is no statistically significant difference between them. If the letters differ, the difference is statistically significant).
Forests 16 01348 g004
Table 1. Measured tree parameters.
Table 1. Measured tree parameters.
SiteThe Month of the Pulling TestH *, mDBH *, cmTree Crown Area, m2Bent TreeHealthy TreesCrown-Damaged TreesDead TreesStand Area, ha
Stand 12020 0124.0 ± 1.2322.0 ± 1.828.11 ± 3.681011--4.40
Stand 22020 0124.7 ± 1.5124.2 ± 2.5811.51 ± 9.801110--1.97
Stand 32020 0124.2 ± 1.8324.7 ± 2.1411.44 ± 6.13910--1.23
Stand 42023 08,
2024 05–07		20.28 ± 2.0720.64 ± 4.427.46 ± 5.4365460-22.97
Stand 52023 0818 ± 3.0617.71 ± 4.124.94 ± 2.505399171.74
Stand 62023 0822.23 ± 3.6126.26 ± 7.387.72 ± 3.9851255182.98
Stand 72023 0926.51 ± 3.4324.32 ± 4.456.79 ± 3.045102630.82
Stand 82023 0822.97 ± 3.3527.24 ± 7.7111.65 ± 6.4852472241.83
Stand 92023 0820.85 ± 5.7821.64 ± 8.036.26 ± 4.055885201.07
* H-mean tree height; DBH-mean diameter at breast height.
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Bāders, E.; Seipulis, A.; Kaupe, D.; Champion, J.J.-C.; Krišāns, O.; Elferts, D. UAV-Based Multispectral Assessment of Wind-Induced Damage in Norway Spruce Crowns. Forests 2025, 16, 1348. https://doi.org/10.3390/f16081348

AMA Style

Bāders E, Seipulis A, Kaupe D, Champion JJ-C, Krišāns O, Elferts D. UAV-Based Multispectral Assessment of Wind-Induced Damage in Norway Spruce Crowns. Forests. 2025; 16(8):1348. https://doi.org/10.3390/f16081348

Chicago/Turabian Style

Bāders, Endijs, Andris Seipulis, Dārta Kaupe, Jordane Jean-Claude Champion, Oskars Krišāns, and Didzis Elferts. 2025. "UAV-Based Multispectral Assessment of Wind-Induced Damage in Norway Spruce Crowns" Forests 16, no. 8: 1348. https://doi.org/10.3390/f16081348

APA Style

Bāders, E., Seipulis, A., Kaupe, D., Champion, J. J.-C., Krišāns, O., & Elferts, D. (2025). UAV-Based Multispectral Assessment of Wind-Induced Damage in Norway Spruce Crowns. Forests, 16(8), 1348. https://doi.org/10.3390/f16081348

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop