Browsing by Author "Hovi, Aarne"

Filter results by typing the first few letters
Now showing 1 - 20 of 50
  • Assessment of a photon recollision probability based forest reflectance model in European boreal and temperate forests
    (2022-02) Hovi, Aarne; Schraik, Daniel; Hanuš, Jan; Homolová, Lucie; Juola, Jussi; Lang, Mait; Lukeš, Petr; Pisek, Jan; Rautiainen, Miina
     A1 Alkuperäisartikkeli tieteellisessä aikakauslehdessä
    We report a new version and an empirical evaluation of a forest reflectance model based on photon recollision probability (p). For the first time, a p-based approach to modeling forest reflectance was tested in a wide range of differently structured forests from different biomes. To parameterize the model, we measured forest canopy structure and spectral characteristics for 50 forest plots in four study sites spanning from boreal to temperate biomes in Europe (48°–62°N). We compared modeled forest reflectance spectra against airborne hyperspectral data at wavelengths of 450–2200 nm. Large overestimation occurred, especially in the near-infrared region, when the model was parameterized considering only leaves or needles as plant elements and assuming a Lambertian canopy. The model root mean square error (RMSE) was on average 80%, 80%, 54% for coniferous, broadleaved, and mixed forests, respectively. We suggest a new parameterization that takes into account the nadir to hemispherical reflectance ratio of the canopy and contribution of woody elements to the forest reflectance. We evaluated the new parameterization based on inversion of the model, which resulted in average RMSE of 20%, 15%, and 11% for coniferous, broadleaved, and mixed forests. The model requires only few structural parameters and the spectra of foliage, woody elements, and forest floor as input. It can be used in interpretation of multi- and hyperspectral remote sensing data, as well as in land surface and climate modeling. In general, our results also indicate that even though the foliage spectra are not dramatically different between coniferous and broadleaved forests, they can still explain a large part of reflectance differences between these forest types in the near-infrared, where sensitivity of the reflectance of dense forests to changes in the scattering properties of the foliage is high.
  • Avoimia spektrikirjastoja Suomen metsistä
    (2022) Rautiainen, Miina; Hovi, Aarne; Kuusinen, Nea; Juola, Jussi; Forsström, Petri; Salko, Sini-Selina; Schraik, Daniel; Burdun, Iuliia
     A1 Alkuperäisartikkeli tieteellisessä aikakauslehdessä
  • Boreaalisten metsien albedosta
    (2020) Rautiainen, Miina; Kuusinen, Nea; Hovi, Aarne; Majasalmi, Titta
     A2 Katsausartikkeli tieteellisessä aikakauslehdessä
    Metsien albedon vaihtelun selvittäminen antaa tietoa energiavirroista maan ja ilmakehän välillä sekä tarkennettuja lähtötietoja ilmastomallinnukseen. Siihen, pitäisikö albedon vaihtelu erityyppisten metsien välillä ottaa huomioon ilmastopolitiikassa tai metsänhoidon päätöksenteossa, ei pystytä vastaamaan ennen kuin kaikki metsien ilmastoon vaikuttavat tekijät on arvioitu samanaikaisesti. Boreaalisten metsien albedolla on omat erityispiirteensä verrattuna muiden kasvillisuusvyöhykkeiden albedoon; metsiemme albedo vaihtelee muun muassa lumitilanteen, puulajikoostumuksen, metsän rakenteen sekä aluskasvillisuuden mukaan. Tässä katsausartikkelissa kerromme mikä albedo on, miksi siitä ollaan kiinnostuneita, miten metsän albedoa on mahdollista arvioida ja mitkä tekijät vaikuttavat siihen. Tarkastelumme painottuu suomalaisiin metsiin. Lisäksi tarkastelemme Suomen metsien albedon kehitystä viime vuosikymmenten aikana ja pohdimme mitä seikkoja boreaalisten metsien albedosta ei vielä ymmärretä.
  • Classification of tree species based on hyperspectral reflectance images of stem bark
    (2023) Juola, Jussi; Hovi, Aarne; Rautiainen, Miina
     A1 Alkuperäisartikkeli tieteellisessä aikakauslehdessä
    Automatization of tree species identification in the field is crucial in improving forest-based bioeconomy, supporting forest management, and facilitating in situ data collection for remote sensing applications. However, tree species recognition has never been addressed with hyperspectral reflectance images of stem bark before. We investigated how stem bark texture differs between tree species using a hyperspectral camera set-up and gray level co-occurrence matrices and assessed the potential of using reflectance spectra and texture features of stem bark to identify tree species. The analyses were based on 200 hyperspectral reflectance data cubes (415–925 nm) representing ten tree species. There were subtle interspecific differences in bark texture. Using average spectral features in linear discriminant analysis classifier resulted in classification accuracy of 92–96.5%. Using spectral and texture features together resulted in accuracy of 93–97.5%. With a convolutional neural network, we obtained an accuracy of 94%. Our study showed that the spectral features of stem bark were robust for classifying tree species, but importantly, bark texture is beneficial when combined with spectral data. Our results suggest that in situ imaging spectroscopy is a promising sensor technology for developing accurate tree species identification applications to support remote sensing.
  • Comparison of contemporaneous Sentinel-2 and EnMAP data for vegetation index-based estimation of leaf area index and canopy closure of a boreal forest
    (2024-11-27) Juola, Jussi; Hovi, Aarne; Rautiainen, Miina
     A1 Alkuperäisartikkeli tieteellisessä aikakauslehdessä
    Data from the new hyperspectral satellite missions such as EnMAP are anticipated to refine leaf area index (LAI) or canopy closure (CC) monitoring in conifer-dominated forest areas. We compared contemporaneous multispectral and hyperspectral satellite images from Sentinel-2 MSI (S2) and EnMAP and assessed whether hyperspectral images offer added value in estimating LAI, effective LAI (LAIeff), and CC in a European boreal forest area. The estimations were performed using univariate and multivariate generalized additive models. The models utilized field measurements of LAI and CC from 38 forest plots and an extensive set of vegetation indices (VIs) derived from the satellite data. The best univariate models for each of the three response variables had small differences between the two sensors, but in general, EnMAP had more well-performing VIs which was reflected in the better multivariate model performances. The best performing multivariate models with the EnMAP data had ~1–6% lower relative RMSEs than S2. Wavelengths near the green, red-edge, and shortwave infrared regions were frequently utilized in estimating LAI, LAIeff, and CC with EnMAP data. Because EnMAP could estimate LAI better, the results suggest that EnMAP may be more useful than multispectral satellite sensors, such as S2, in monitoring biophysical variables of coniferous-dominated forests. © 2024 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group.
  • Contribution of woody elements to tree level reflectance in boreal forests
    (2021) Kuusinen, Nea; Hovi, Aarne; Rautiainen, Miina
     A1 Alkuperäisartikkeli tieteellisessä aikakauslehdessä
    Spectral mixture analysis was used to estimate the contribution of woody elements to tree level reflectance from airborne hyperspectral data in boreal forest stands in Finland. Knowledge of the contribution of woody elements to tree or forest reflectance is important in the context of lea area index (LAI) estimation and, e.g., in the estimation of defoliation due to insect outbreaks, from remote sensing data. Field measurements from four Scots pine (Pinus sylvestris L.), five Norway spruce (Picea abies (L.) Karst.) and four birch (Betula pendula Roth and Betula pubescens Ehrh.) dominated plots, spectral measurements of needles, leaves, bark, and forest floor, airborne hyper-spectral as well as airborne laser scanning data were used together with a physically-based forest reflectance model. We compared the results based on simple linear combinations of measured bark and needle/leaf spectra to those obtained by accounting for multiple scattering of radiation within the canopy using a physically-based forest reflectance model. The contribution of forest floor to reflectance was additionally considered. The resulted mean woody element contribution estimates varied from 0.140 to 0.186 for Scots pine, from 0.116 to 0.196 for birches and from 0.090 to 0.095 for Norway spruce, depending on the model used. The contribution of woody elements to tree reflectance had a weak connection to plot level forest variables.
  • Corrigendum to ‘A dataset composed of multiangular spectral libraries and auxiliary data at tree, leaf, needle, and bark level for three common European tree species’ [Data in Brief 35 (2021) 106820] (Data in Brief (2021) 35, (S2352340921001049), (10.1016/j.dib.2021.106820))
    (2021-08) Hovi, Aarne; Forsström, Petri R.; Ghielmetti, Giulia; Schaepman, Michael E.; Rautiainen, Miina
     Comment/debate
    The authors regret that Eq. 6 in the article is incorrect. The term GT should be subtracted from the quotient of DNleaf,T/DNWR_leaf,T (not from DNleaf,T in the numerator). The correct equation is: [Formula presented] The error was only in the article text and therefore did not influence the data published. The authors would like to apologise for any inconvenience caused.
  • Crown level clumping in Norway spruce from terrestrial laser scanning measurements
    (2021-01-15) Schraik, Daniel; Hovi, Aarne; Rautiainen, Miina
     A1 Alkuperäisartikkeli tieteellisessä aikakauslehdessä
    The clumping of coniferous needles into shoots is widely acknowledged as a structural feature that cannot be ignored in radiation regime models of coniferous forests. However, higher level clumping, i.e. the aggregation of leaves and shoots in tree crowns and forest stands, is still rarely accounted for in the models. Clumping reduces the light interception of and increases the light penetration depth in a plant stand. To improve forest radiation regime models with respect to this forest structural parameter, we propose a method that can quantify clumping at different hierarchical levels by estimating the silhouette to total area ratio from point clouds acquired by laser scanners. Our method is based on estimating attenuation coefficients in a voxel grid, and subsequently computing the total leaf area and spherically averaged silhouette area of a tree crown or forest stand. We tested our method with empirical data in young Norway spruce trees, where we compared leaf area and silhouette area to destructive and photogrammetric reference measurements. The accuracy of leaf area estimates depended strongly on the voxel size, with voxel sizes below 10 cm side length exhibiting up to 100% higher estimates than the reference leaf area, and large voxels with 90 cm side length being closest to the reference measurements due to crown clumping. The silhouette area estimates varied less with voxel size and were slightly higher than the reference estimates. We analyzed possible error sources and point out ways to improve the measurements of leaf and silhouette area for conifer trees using laser scanning data.
  • A dataset composed of multiangular spectral libraries and auxiliary data at tree, leaf, needle, and bark level for three common European tree species
    (2021-04) Hovi, Aarne; Forsström, Petri R.; Ghielmetti, Giulia; Schaepman, Michael E.; Rautiainen, Miina
     Data Article
    This article describes a dataset of multiangular scattering properties of small trees (height = 0.38–0.7 m) at visible, near-infrared, and shortwave-infrared wavelengths (350–2500 nm), and provides supporting auxiliary data that comprise leaf, needle, and bark spectra, and structural characteristics of the trees. Multiangular spectra were measured for 18 trees belonging to three common European tree species: Scots pine (Pinus sylvestris L.), Norway spruce (Picea abies (L.) H. Karst), and sessile oak (Quercus petraea (Matt.) Liebl.). The measurements were performed in 47 different view angles across a hemisphere, using a laboratory goniometer and a non-imaging spectrometer. Leaf and needle spectra were measured for each tree, using a non-imaging spectrometer coupled to an integrating sphere. Bark spectra were measured for one sample tree per species. In addition, leaf and needle fresh mass, surface area of leaves, needles, and woody parts, silhouette area, and spherically averaged silhouette to total area ratio (STAR) for each tree were measured or derived from the measurements. The data are useful for modeling the shortwave reflectance characteristics of small trees and potentially forests, and thus benefit climate modeling or interpretation of remote sensing data.
  • Direct estimation of photon recollision probability using terrestrial laser scanning
    (2020-09-15) Wang, Di; Schraik, Daniel; Hovi, Aarne; Rautiainen, Miina
     A1 Alkuperäisartikkeli tieteellisessä aikakauslehdessä
    Photon recollision probability p is a spectrally invariant structural parameter and a powerful tool to link canopy optical properties at any wavelengths to model reflectance, transmittance, or absorption of vegetation canopies. The concepts of the p-theory have been reported and examined at the shoot and canopy scales, but not yet for the crown level. Currently, the p-value is estimated indirectly, such as converted from the spherically averaged silhouette to total area ratio (STAR¯) or canopy transmittance measurements. In this work, we first validate the theoretical considerations of the p concept at the crown level (e.g., its relationship with STAR¯), and then provide the first method to directly estimate photon recollision probability using Terrestrial Laser Scanning (TLS) data. The proposed geometric method is data-driven and avoids explicit reconstructions of tree structures. The p-value estimated here is the average recollision probability over spatial locations. We showed that the average recollision probability can be interpreted as the local spherical openness on phytoelement (leaf or needle) surfaces, which enabled a simple visibility calculation by avoiding explicit ray tracing. The developed method was tested on synthetic crowns of needle-leaved tree species, for which the reference p-values were known. Results confirmed the validity of the p-STAR¯ relationship at the crown level, and showed that p-values can be accurately estimated from TLS point clouds with a relative root measure square error of less than 10%. This study displays the distinct advantage of TLS in delineating detailed tree crown structures and highlights its potential in studies of forest reflectance modeling.
  • Empirical validation of photon recollision probability in single crowns of tree seedlings
    (2020-11) Hovi, Aarne; Forsström, Petri; Ghielmetti, Giulia; Schaepman, Michael E.; Rautiainen, Miina
     A1 Alkuperäisartikkeli tieteellisessä aikakauslehdessä
    Physically-based methods in remote sensing provide benefits over statistical approaches in monitoring biophysical characteristics of vegetation. However, physically-based models still demand large computational resources and often require rather detailed informative priors on various aspects of vegetation and atmospheric status. Spectral invariants and photon recollision probability theories provide a solid theoretical framework for developing relatively simple models of forest canopy reflectance. Empirical validation of these theories is, however, scarce. Here we present results of a first empirical validation of a model based on photon recollision probability at the level of individual trees. Multiangular spectra of pine, spruce, and oak tree seedlings (height = 0.38–0.7 m) were measured using a goniometer, and tree hemispherical reflectance was derived from those measurements. We evaluated the agreement between modeled and measured tree reflectance. The model predicted the spectral signatures of the tree seedlings in the wavelength range between 400 and 2300 nm well, with wavelength-specific bias between −0.048 and 0.034 in reflectance units. In relative terms, the model errors were the smallest in the near-infrared (relative RMSE up to 4%, 7%, and 4% for pine, spruce, and oak seedlings, respectively) and the largest in the visible wavelength region (relative RMSE up to 34%, 20%, and 60%). The errors in the visible region could be partly attributed to wavelength-dependent directional scattering properties of the leaves. Including woody parts of tree seedlings in the model improved the results by reducing the relative RMSE by up to 10% depending on species and wavelength. Spectrally invariant model parameters, i.e. total and directional escape probabilities, depended on spherically averaged silhouette to total area ratio (STAR) of the tree seedlings. Overall, the modeled and measured tree reflectance mainly agreed within measurement uncertainties, but the results indicate that the assumption of isotropic scattering by the leaves can result in large errors in the visible wavelength region for some tree species. Our results help increasing the confidence when using photon recollision probability and spectral invariants -based models to interpret satellite images, but they also lead to an improved understanding of the assumptions and limitations of these theories.
  • Estimating cover fraction from tls return intensity in coniferous and broadleaved tree shoots
    (2021) Schraik, Daniel; Hovi, Aarne; Rautiainen, Miina
     A1 Alkuperäisartikkeli tieteellisessä aikakauslehdessä
    Terrestrial laser scanning (TLS) provides a unique opportunity to study forest canopy structure and its spatial patterns such as foliage quantity and dispersal. Using TLS point clouds for estimating leaf area density with voxel-based methods is biased by the physical dimensions of laser beams, which violates the common assumption of beams being infinitely thin. Real laser beams have a footprint size larger than several millimeters. This leads to difficulties in estimating leaf area density from light detection and ranging (LiDAR) in vegetation, where the target objects can be of similar or even smaller size than the beam footprint. To compensate for this bias, we propose a method to estimate the per-pulse cover fraction, defined as the fraction of laser beams’ footprint area that is covered by vegetation targets, using the LiDAR return intensity and an experimental calibration measurement. We applied this method to a Leica P40 single-return instrument, and report our experimental results. We found that conifer foliage had a lower average per-pulse cover fraction than broadleaved foliage, indicating an increased number of partial hits in conifer foliage. We further discuss limitations of our method that stem from unknown target properties that influence the LiDAR return intensity and highlight potential ways to overcome the limitations and manage the remaining uncertainty. Our method’s output, the per-beam cover fraction, may be useful in a weight function for methods that estimate leaf area density from LiDAR point clouds.
  • Estimation of boreal forest attributes using hyperspectral remote sensing data
    (2025-02-04) Karhuvaara, Henriikka
     School of Engineering | Master's thesis
    Multispectral and hyperspectral satellite remote sensing can be applied to observe the Earth and detect changes caused by human actions or the climate. Currently there are global multispectral satellite missions that can observe a large variety of different targets on the Earth’s surface and atmosphere by measuring their reflectance spectra with a few selected spectral bands. By 2030, satellite mission providers such as NASA and ESA plan to offer enhanced observation of different targets by measuring their reflectance spectra in hundreds of narrow spectral bands with global hyperspectral satellite missions. In this work, hyperspectral satellite images collected by EnMAP (The Environmental Mapping and Analysis Program) mission were applied to estimate stem volumes and tree species proportions in five different boreal forest locations of Finland. Estimations were performed with two machine learning regression methods, and the estimation accuracies were assessed based on available field data collected by the Finnish Forest Centre. In general, the best estimation accuracies were reached for total stem volumes with RRMSE values between 52–93 %. Estimation accuracies of total stem volumes were also the most consistent between the study locations. Estimation accuracies of stem volumes and proportions related to single species were more dependent on the species distribution of the study locations, and the best accuracies were reached for the most dominant species of each location with RRMSE values between 44–143 %. Higher mean species volumes and proportions in the field data were related to better estimation accuracy also between the study locations. This study gave information on the applicability of hyperspectral satellite data to estimate selected boreal forest attributes in a few targeted locations. More reliable conclusions on the applicability of the estimations, for example, to monitor forest biodiversity can be made once the availability of hyperspectral satellite data increases.
  • Estimation of boreal forest floor lichen cover using hyperspectral airborne and field data
    (2023) Kuusinen, Nea; Hovi, Aarne; Rautiainen, Miina
     A1 Alkuperäisartikkeli tieteellisessä aikakauslehdessä
    Lichens are sensitive to competition from vascular plants, intensive silviculture, pollution and reindeer and caribou grazing, and can therefore serve as indicators of environmental changes. Hyperspectral remote sensing data has been proved promising for estimation of plant diversity, but its potential for forest floor lichen cover estimation has not yet been studied. In this study, we investigated the use of hyperspectral data in estimating ground lichen cover in boreal forest stands in Finland. We acquired airborne and in situ hyperspectral data of lichen-covered forest plots, and applied multiple endmember spectral mixture analysis to estimate the fractional cover of ground lichens in these plots. Estimation of lichen cover based on in situ spectral data was very accurate (coefficient of determination (r2) 0.95, root mean square error (RMSE) 6.2). Estimation of lichen cover based on airborne data, on the other hand, was fairly good (r2 0.77, RMSE 11.7), but depended on the choice of spectral bands. When the hyperspectral data were resampled to the spectral resolution of Sentinel-2, slightly weaker results were obtained. Tree canopy cover near the flight plots was weakly related to the difference between estimated and measured lichen cover. The results also implied that the presence of dwarf shrubs could influence the lichen cover estimates.
  • Evaluating the performance of a double integrating sphere in measurement of reflectance, transmittance, and albedo of coniferous needles
    (2020-01-01) Hovi, Aarne; Mõttus, Matti; Juola, Jussi; Manoocheri, Farshid; Ikonen, Erkki; Rautiainen, Miina
     A1 Alkuperäisartikkeli tieteellisessä aikakauslehdessä
    Leaf reflectance and transmittance spectra are essential information in many applications such as developing remote sensing methods, computing shortwave energy balance (albedo) of forest canopies, and monitoring health or stress of trees. Measurement of coniferous needle spectra has usually been carried out with single integrating spheres, which has involved a lot of tedious manual work. A small double integrating sphere would make the measurements considerably faster, because of its ease of operation and small sample sizes required. Here we applied a compact double integrating sphere setup, used previously for measurement of broad leaves, for measurement of coniferous needles. Test measurements with the double integrating sphere showed relative underestimation of needle albedo by 5–39% compared to a well-established single integrating sphere setup. A small part of the bias can be explained by the bias of the single sphere. Yet the observed bias is quite significant if absolute accuracy of measurements is required. For relative measurements, e.g. for monitoring development of needle spectra over time, the double sphere system provides notable improvement. Furthermore, it might be possible to reduce the bias by building an optimized measurement setup that minimizes absorption losses in the sample port. Our study indicates that double spheres, after some technical improvement, may provide a new and fast way to collect extensive spectral libraries of tree species.
  • Evaluation of a forest radiative transfer model using an extensive boreal forest inventory database
    (2023-12) Gopalakrishnan, Ranjith; Korhonen, Lauri; Mõttus, Matti; Rautiainen, Miina; Hovi, Aarne; Mehtätalo, Lauri; Maltamo, Matti; Peltola, Heli; Packalen, Petteri
     A1 Alkuperäisartikkeli tieteellisessä aikakauslehdessä
    The forest reflectance and transmittance model (FRT) is applicable over a wide swath of boreal forest landscapes mainly because its stand-specific inputs can be generated from standard forest inventory variables. We quantified the accuracy of this model over an extensive region for the first time. This was done by carrying out a simulation study over a large number (12,369) of georeferenced forest plots from operational forest management inventories conducted in Southern Finland. We compared the FRT simulated bidirectional reflectance factors (BRF) with those measured by Landsat 8 satellite Operational Land Imager (OLI). We also quantified the relative importance of several explanatory factors that affected the magnitude of the discrepancy between the measured and simulated BRFs using a linear mixed effects modelling framework. A general trend of FRT overestimating BRFs is seen across all tree species and spectral bands examined: up to ∼0.05 for the red band, and ∼0.10 for the near infrared band. The important explanatory factors associated with the overestimations included the dominant tree species, understory type of the forest plot, timber volume (acts as a proxy for stand maturity), vegetation heterogeneity and time of the year. Our analysis suggests that approximately 20% of the error is caused by the non-representative spectra of canopy foliage and understory. Our results demonstrate the importance of collecting representative spectra from a diverse set of forest stands, and over the full range of seasons.
  • Evaluation of Accuracy and Practical Applicability of Methods for Measuring Leaf Reflectance and Transmittance Spectra
    (2018-01) Hovi, Aarne; Forsström, Petri; Mõttus, Matti; Rautiainen, Miina
     A1 Alkuperäisartikkeli tieteellisessä aikakauslehdessä
    Leaf reflectance and transmittance spectra are urgently needed in interpretation of remote sensing data and modeling energy budgets of vegetation. The measurement methods should be fast to operate and preferably portable to enable quick collection of spectral databases and in situ measurements. At the same time, the collected spectra must be comparable across measurement campaigns. We compared three different methods for acquiring leaf reflectance and transmittance spectra. These were a single integrating sphere (ASD RTS-3ZC), a small double integrating sphere (Ocean Optics SpectroClip-TR), and a leaf clip (PP Systems UNI501 Mini Leaf Clip). With all methods, an ASD FieldSpec 4 spectrometer was used to measure white paper and tree leaves. Single and double integrating spheres showed comparable within-method variability in the measurements. Variability with leaf clip was slightly higher. The systematic difference in mean reflectance spectra between single and double integrating spheres was only minor (average relative difference of 1%), whereas a large difference (14%) was observed in transmittance. Reflectance measured with leaf clip was on average 14% higher compared to single integrating sphere. The differences between methods influenced also spectral vegetation indices calculated from the spectra, particularly those that were designed to track small changes in spectra. Measurements with double integrating sphere were four, and with leaf clip six times as fast as with single integrating sphere, if slightly reduced signal level (integration time reduced from optimum) was allowed for the double integrating sphere. Thus, these methods are fast alternatives to a conventional single integrating sphere. However, because the differences between methods depended on the measured target and wavelength, care must be taken when comparing the leaf spectra acquired with different methods.
  • Exploring the potential of SAR and terrestrial and airborne LiDAR in predicting forest floor spectral properties in temperate and boreal forests
    (2025-01-01) Mercier, Audrey; Myllymäki, Mari; Hovi, Aarne; Schraik, Daniel; Rautiainen, Miina
     A1 Alkuperäisartikkeli tieteellisessä aikakauslehdessä
    Forest floor vegetation plays a crucial role in ecosystem processes of temperate and boreal forests. Remote sensing offers a valuable tool to characterize the forest floor through reflectance spectra. While passive optical airborne and satellite data have been used to map spectral properties of forest understory, these sensors are limited by cloud cover, especially in high latitudes. To date, LiDAR and SAR have not been explored for this application even though their data are less dependent on illumination conditions and provide information on tree canopy structure and tree distribution which is connected to forest floor properties. We investigated active remote sensing techniques to establish links between forest structure and spectral properties of forest floor across European temperate, hemiboreal and boreal forest ecosystems. First, in the exploratory part, the research question was : Which forest structure metrics are connected to the spectral properties of the forest floor? Next, our predictive part focused on: What is the potential of (1) terrestrial laser scanning (TLS) data, (2) airborne laser scanning data, (3) satellite-borne SAR data, and (4) these data sources combined to predict forest floor spectral properties? Our results revealed that nine forest structure metrics were potentially associated with forest floor reflectance. We identified TLS-derived clumping index and SAR-derived VV backscatter coefficient and VH/VV ratio as significantly connected to forest floor reflectance in certain Sentinel-2 spectral bands. Overall, the active remote sensors achieved the best predictions for forest floor reflectance in red-edge, near-infrared and shortwave infrared regions. Using data from all three sensors together to predict the forest floor spectra yielded better results than using any of the sensors alone. When data from a single sensor were used, the highest prediction accuracies for forest floor reflectance in the red-edge and near-infrared regions were achieved with SAR data, and in the shortwave infrared region with either SAR or TLS data. In the future, the accuracy of predicting forest floor characteristics in temperate and boreal forests could benefit from a synergy of passive and active technologies.
  • Hyperspectral characterization of vegetation in hemiboreal, boreal and Arctic peatlands using a geographically extensive field dataset
    (2024-09) Salko, Sini-Selina; Hovi, Aarne; Burdun, Iuliia; Juola, Jussi; Rautiainen, Miina
     A1 Alkuperäisartikkeli tieteellisessä aikakauslehdessä
    Northern peatlands store up to 25% of global soil organic carbon and function as important hotspots for biodiversity. However, they are facing degradation from climate change driven by human activities as well as anthropogenic land use changes, up to the point of endangering the ecosystems' functioning and the storage of soil organic carbon entailed within them. The surface vegetation of northern peatlands is an important indicator of the ecosystem's functioning and ecohydrology, highlighting the importance of its large-scale, continuous monitoring. Approaches utilizing hyperspectral data for monitoring vegetation health and species composition can also be applied to peatland vegetation. To support the development of methods for interpreting hyperspectral satellite data from peatlands, we conducted a comprehensive in situ study of hemiboreal, boreal, sub-Arctic and Arctic peatland vegetation. We measured the reflectance spectra (350–2500 nm), soil moisture, and various vegetation-related attributes from a total of 446 vegetation plots in Estonia and Finland, from a 1500 km south-north interval. We then investigated (i) the spectral variation in surface vegetation of hemiboreal, boreal, sub-Arctic and Arctic peatlands and (ii) explored its connection to plant functional types (PFTs) and soil moisture, as well as evaluated the potential of hyperspectral data in estimating PFT cover using simple vegetation indices and partial least square (PLS) regression. Our results indicate that (i) the best spectral regions to retrieve information regarding the PFT vary greatly especially between vascular plants and bryophytes, (ii) the reflectance at an individual wavelength as well as simple vegetational index can, to some extent, predict the PFT, and that (iii) the PLS regression can predict the PFT with good accuracy. Overall, our findings demonstrate the potential of using hyperspectral data in monitoring PFTs in northern peatlands. The spectral library and the ancillary data from the peatland sites collected for this study are available as open data.
  • Hyperspectral data of understory elements in boreal forests : In situ and laboratory measurements
    (2024-12) Mercier, Audrey; Karlqvist, Susanna; Hovi, Aarne; Rautiainen, Miina
     Data Article
    Enhancing our understanding of the spectral properties of forest elements is essential for interpreting airborne and satellite-borne remote sensing data. This article presents two datasets on the spectral properties of understory elements in boreal forests collected with close-range hyperspectral measurements. We conducted two field campaigns in June and July 2023 in Finland to acquire spectral measurements at wavelengths from 350 to 2500 nm using an ASD FieldSpec 4 spectrometer for forest understory elements. We measured ferns, decaying wood, common wood sorrel and May lily in situ. In a laboratory, we measured leaves from European fly honeysuckle, alder buckthorn and common hazel. These data support the analysis of vegetation characteristics, training of classification algorithms and improvement of forest radiative transfer models, and could be used to evaluate the potential of hyperspectral data to discriminate the understory elements of boreal forest.