Swedish National Forest Inventory Research Articles

Computation of prediction intervals for forest aboveground biomass predictions using generalized linear models in a large-extent boreal forest region

Abstract Remotely sensed data have an important application for estimation of forest variables, e.g. height, volume, and aboveground biomass (AGB). The increased use of remotely sensed data implemented along with model-based inference has shown improved efficiency in prediction and mapping of such forest variables. In this study, plot-level airborne laser scanning data and Swedish National Forest Inventory field reference data were used to predict AGB using generalized linear models (GLMs) assuming Gamma and Tweedie distributions for the field observed AGB. The GLMs were selected considering the convenience of not correcting transformation bias as it is required in other regression models with transformed response variable. To overcome the challenge in providing reliable uncertainty estimates for the estimated forest variable map products at individual pixel-scale, we focused on computing 95% prediction intervals (PIs) for Gamma and Tweedie GLMs with a square root link function. The relative uncertainties were computed as the ratio between the half-width of the PIs and the predicted AGBs. The AGB-airborne laser scanning models were developed with root mean square error values of 22.6 Mgha−1 (26%) and 21.7 Mgha−1 (25%), respectively, for the Gamma and Tweedie GLMs. Two methods were applied to compute PIs for the Gamma GLM, one using the R package ‘ciTools’ and another derived through asymptotic theory. It was found that the 95% PIs computed using ‘ciTools’ had the most accurate coverage probability in comparison to the other method. An extended version of these PIs was also utilized for the Tweedie GLMs. The range of PIs associated with the prediction of AGB were narrower for lower predicted AGB values compared with the length of higher predicted AGB values. Comparing the two fitted models, the Gamma GLM showed lower relative uncertainties for the lower range of predicted AGBs, whereas the Tweedie GLM showed lower relative uncertainties for the higher range of predicted AGBs. Overall, the Tweedie GLM provided a better model fit for AGB predictions.

Availability and mobilization of forest resources in Sweden

The available amount of wood supply is essential for national strategic planning and evaluation of forestry in Sweden. Since Sweden holds a large part of the forests in the European Union and plays a significant role in the global trade of wood-based products, a precise estimate of the potential of the Swedish forest resource is also important in regional and global outlook studies. In this study, we analyse factors influencing the availability and mobilization of wood supply. By comparing data from the Swedish National Forest Inventory with the stand registers of the five largest forest owners in Sweden, we estimate the productive forest area not included in the forest owners' stand databases. Our results show that 0.4 million hectares, or 5% of these large-scale forest owners productive forest area, is outside their stand registers and therefore neither included in their long-term harvesting plans nor in their nature conservation plans. For small-scale forest owners, we analyse the final felling rate during 2004–2020 using satellite imagery to estimate the proportion of properties that abstain from final fellings and thereby could affect the potential mobilization of wood supply. During this period, 32% of the forest properties owned by small-scale forest owners have not done any final felling. These forest estates hold in total 1.1 million hectares of productive forest land or 9% of the area owned by small-scale forest owners. This implies a gap between the potential and realistic estimates for Forest Available for Wood Supply.

Diversity of forest structures important for biodiversity is determined by the combined effects of productivity, stand age, and management.

In forests, the amount and diversity of structural features with high value for biodiversity, such as large trees and dead wood, are affected by productivity, stand age, and forest management. For efficient conservation of forest biodiversity, it is essential to understand the combined effects of these drivers. We used data from the Swedish National Forest Inventory to study the combined effects of productivity, stand age, and management for wood production on structures with high value for biodiversity: tree species richness, large living trees, dead wood volume, and specific dead wood types. Forest management changed the relationship between productivity and amount or diversity of some of the structures. Most structures increased with productivity and stand age, but decreased due to management. The negative effect of management was greatest for structures occurring mainly in high-productivity forests, such as deciduous dead wood. Thus, biodiversity conservation should target high-productivity forests to preserve these structures.

Estimation of plant density based on presence/absence data using hybrid inference

Monitoring of plant populations has become more and more important, especially in the current context of environmental change. In this paper, we propose methods to estimate plant density from presence/absence surveys, wherein the presence or absence of each species is recorded on sample plots. Presence/absence sampling is a useful and relatively simple method for monitoring state and change of plant communities. Moreover, it has advantages compared to traditional plant cover assessment, the latter being more prone to observer bias. We present a hybrid estimation framework, that combines model- and design-based inference features, in which a generalised linear model (for binary presence/absence data) and an inhomogeneous Poisson model (for plant locations) are used to estimate plant density in a region of interest. We look at two different cases, the first one with a known area and the second one where the area is unknown and must be estimated. Our methods are applied to real data on Vaccinium vitis-idaea from the Swedish National Forest Inventory as well as simulated data to assess the performance of our estimators of plant density and corresponding variance estimators. The results obtained are promising and indicate that this method has a potential to add considerable analytic strength to monitoring programmes that collect presence/absence data.

Mapping site index in coniferous forests using bi-temporal airborne laser scanning data and field data from the Swedish national forest inventory

Recent advancements in remote sensing of forests have demonstrated the capabilities of three-dimensional data acquired by airborne laser scanning (ALS) and, consequently, have become an integral part of enhanced forest inventories in Northern Europe. In Sweden, the first national laser scanning revolutionised forest management planning through low-cost production of large-scale and spatially explicit maps of forest attributes such as basal area, volume, and biomass, compared to the earlier practice based on field survey data. A second scanning at the national level was launched in 2019, and it provides conditions for the estimation of height growth and site index. Accurate and up-to-date information about site productivity is relevant for planning silvicultural treatments and for the prognosis of forest status and development over time. In this study, we explored the potential of bi-temporal ALS data and other auxiliary information to predict and map site productivity by site index according to site properties (SIS) of Norway spruce (Picea abies (L.) Karst) and Scots pine (Pinus sylvestris L.) in even-aged stands in Sweden. We linked ground survey data of SIS from more than 11,500 plots of the Swedish National Forest Inventory (NFI) to bi-temporal ALS data to predict and map site index using an area-based method and two regression modelling strategies: (1) a multiple linear regression (MLR) model with an ordinary least-squares parameter estimation method, and (2) a non-parametric random forests (RF) model optimised for hyper-parameter tuning. For model development, permanent plots were used, whereas the validation was done on the temporary plots of the Swedish NFI and an independent stand-level dataset. Species-specific models were developed, and the root mean square error (RMSE) metric was used to quantify the residual variability around model predictions. For both species, the MLR model gave precise and accurate estimates of SIS. The RMSE for SIS predictions was in the range of 1.96 – 2.11 m, and the relative RMSE was less than 10 % (7.68 – 9.49 %) of the reference mean value. Final predictors of site index include metrics of 90th percentile height and annual increment in the 95th percentile height, altitude, distance to coast, and soil moisture. Country-wide maps of SIS and the corresponding pixel-level prediction errors at a spatial resolution of 12.5 m grid cells were produced for the two species. Independent validations show the site index maps are suitable for use in operational forest management planning in Sweden.

Forest inventory tree core archive reveals changes in boreal wood traits over seven decades

Boreal forests play an important role in the global carbon (C) cycle, and there is great interest in understanding how they respond to environmental change, including nitrogen (N) and water limitation, which could impact future forest growth and C storage. Utilizing tree cores archived by the Swedish National Forest Inventory, we measured stemwood traits, including stable N and C isotope composition which provides valuable information related to N availability and water stress, respectively, as well as N and C content, and C/N ratio over 1950–2017 in two central Swedish counties covering an area of ca. 55,000 sq. km (n = 1038). We tested the hypothesis that wood traits are changing over time, and that temporal patterns would differ depending on alternative dendrochronological reconstruction methods, i.e. the commonly applied “single tree method” (STM) or a conceptually stronger “multiple tree method” (MTM). Averaged across all MTMs, our data showed that all five wood traits for Picea abies and Pinus sylvestris changed over time. Wood δ15N strongly declined, indicating progressive nitrogen limitation. The decline in δ13C tracked the known atmospheric δ13CO2 signal, suggesting no change in water stress occurred. Additionally, wood N significantly increased, while C and C/N ratios declined over time. Furthermore, wood trait patterns sometimes differed between dendrochronological methods. The most notable difference was for δ15N, where the slope was much shallower for the STM compared to MTMs for both species, indicating that mobility of contemporary N is problematic when using the STM, resulting in substantially less sensitivity to detect historical signals. Our study indicates strong temporal changes in boreal wood traits and also indicates that the field of dendroecology should adopt new methods and archiving practices for studying highly mobile element cycles, such as nitrogen, which are critical for understanding environmental change in high latitude ecosystems.

Effects of cluster plot design parameters on landscape fragmentation estimates: A case study using data from the Swedish national forest inventory

Forest fragmentation is commonly characterized using indices derived from analyses of classified land cover maps. An alternative is to use data obtained from sampling, such as those from a national forest inventory (NFI). The main objective of the current study is fill knowledge gaps on the performance of sample-based forest fragmentation metrics calculated with different cluster plot designs and under different forest conditions. A set of NFI cluster plot designs, each with different geometric properties, was created from Swedish NFI data. Each member of the set was used to calculate the fragmentation metrics mean patch size (MPS) and perimeter-area ratio (PA). Impacts of plot design parameters on metric estimates and their precision were assessed.Important differences in metric values were observed both within and between regions under different plot design scenarios; within regions, ranges of PA and MPS values were large, and confidence intervals for the minimum and maximum metric values did not overlap. Weighted least squares regression significance testing results suggest that subplot separation distance was an impactful design factor whereas number of subplots and cluster shape were less important. However, cluster plots with more and widely-separated subplots yielded estimates that were more precise (lower relative sampling errors) than smaller, more compact clusters. We suggest that care should be taken when interpreting the physical meaning of the metrics under study.

Forest Area and Structural Variable Estimation in Boreal Forest Using Suomi NPP VIIRS Data and a Sample from VHR Imagery

Our objective was to develop a method for the assessment of forest area and structural variables for cases in which the availability of representative ground reference data is poor and these data are not collected from the whole area of interest. We implemented two independent approaches to the estimation of the forest variables of a European boreal forest: (i) the computation of wall-to-wall estimates using moderate- to low-resolution VIIRS imagery from the Suomi NPP mission; and (ii) the visual interpretation of plots of samples from very high resolution (VHR) satellite data obtained via a two-stage design. Our focus was on the statistical comparison of forest resources at a country or larger level. The study area was boreal forest ranging from Norway to the Ural Mountains in Russia. We computed a seamless mosaic from 111 VIIRS images. From the mosaic, we computed predictions for the forest area, growing stock volume, height of the dominating tree layer, proportion of conifers and broadleaved trees, site fertility class, and leaf area index. The reference data for the VIIRS imagery were national forest inventory (NFI)-based raster maps from Finland. The first stage sample of VHR data included 42 images; of these, a second stage sample of 2690 plots was visually interpreted for the same variables. The forest area prediction from VIIRS for the whole study area was 1.2% higher than the VHR-based result. All other structural variable predictions using VIIRS fitted within the 95% confidence intervals computed from the VHR sample except for estimates of the main tree species groups, which were outside the limits. A comparison of VIIRS-based forest area estimates using Finnish and Swedish NFI data indicated overestimations of 10.0% points and 4.6% points, whereas the total growing stock volumes were overestimated by 8% and underestimated by 3.4%, respectively. The correlation coefficients between the VIIRS and VHR image predictions at the 42 VHR image locations varied from 0.70 to 0.85. The VIIRS maps strongly averaged the local predictions due to their coarse spatial resolutions. Based on our findings, the approach using two independent estimations yielded similar figures for the central forest variables for the European boreal forest. A model computed using reference data from a small part of the area of interest can provide satisfactory predictions for a much larger area with a similar biome. Therefore, our concept is applicable to the estimation and overall mapping of a forest area and central structural variables at regional to national levels.

Modelling managed forest ecosystems in Sweden: An evaluation from the stand to the regional scale

Incorporation of a forest management module in the dynamic vegetation model LPJ-GUESS has allowed the study and predictions of management treatment effects on the carbon cycle and on forest ecosystem structure. In this study, LPJ-GUESS is evaluated at the regional scale against observational data from the Swedish National Forest Inventory. Simulated standing volume is compared against observations for the four most common forest types in the country. Furthermore, eddy-covariance flux measurements from the Integrated Carbon Observation System (ICOS) are used to evaluate model predictions of net ecosystem exchange (NEE), gross primary productivity (GPP) and ecosystem respiration (Reco) at the site scale. The model results suggest an adequate representation of standing volume in monocultures of Norway spruce and Scots pine for regional simulations in southern and central Sweden, after an updated parameterization of the species. For northern Sweden, the standing volume in Norway spruce monocultures was overestimated with the updated parameter values. At the stand scale, the model produced mixed results for carbon fluxes when evaluated against eddy-covariance data for two sites, one in central and one in southern Sweden. The interannual variation of GPP was well captured for the central Swedish site, but the modelled average GPP for the period 2015–2019 was overestimated by 9%. For the southern Swedish site, GPP was underestimated by 15% for the corresponding period and the simulated interannual variation was half of the observed. The seasonal estimates of modelled net ecosystem exchange (NEE) deviated from observations and the simulated standing volume was underestimated by 25% for both sites. The results highlight further potential to perform species-specific calibration to capture latitudinal gradients in key ecosystem properties, and to incorporate additional characteristics of site quality which could benefit model accuracy at the scale of individual forest stands, both regarding simulated carbon fluxes and forest stand variables.

Taller and slenderer trees in Swedish forests according to data from the National Forest Inventory

Changes over time in annual basal area growth and mean height for Scots pine (Pinus sylvestris L.) and Norway spruce (Picea abies (L.) Karst.) over the period, 1983–2020 were studied using sample tree data from temporary plots recorded in the Swedish National Forest Inventory. The annual basal area growth was derived from the last measured full ring on increment cores. Using 20 to 60-year-old dominant trees, the mean height and annual basal area growth were examined as functions of tree, stand and site conditions, and trends were assessed mainly using residual analyses over time. A significant increase in mean height at a given age was found for both species, but the annual basal area growth level remained stable over the 38-year period. Currently, at a given age of 50 annual rings at breast height, the mean heights of pines and spruces increased on average by 10.1% (i.e. ∼2 m), compared to 50 year-old pines and spruces in the 1980s, and the increase was similar in the different regions. The results suggest that trees have become taller and slenderer in Swedish forests. Increasing tree height over time at a given age in Northern Europe has been documented in several reports and many causes have been suggested, such as changed forest management, increasing temperatures and nitrogen deposition. We suggest that elevated CO2 in the air and improved water-use efficiency for the trees might also be strong drivers.

Increase in dead wood, large living trees and tree diversity, yet decrease in understory vegetation cover: The effect of three decades of biodiversity-oriented forest policy in Swedish forests

In Sweden, the majority of forest area has been altered by industrial forestry over the decades. Almost 30 years ago, a shift towards biodiversity-oriented forest management practices occurred. Here we took advantage of long-term data collected by the Swedish National Forest Inventory to track developmental changes in forest structural components over this time. We assessed changes in structural components that play an important role in biodiversity (dead wood, large living trees, tree species composition, and understory vegetation) in four forest types with descending tiers of biodiversity protection: protected areas, woodland key habitats, low-productivity forests and production forests. Overall, we found a positive trend in the volumes of dead wood and large living trees, as well as in tree species diversity, while there was a general decline in understory vegetation coverage. Most observed changes were consistent with the intended outcomes of the current forest policy, adapted in the early 1990s. The implementation of retention forestry is likely driving some of the observed changes in forest structural components in the south. In contrast, we observed no changes in any of the focal structural components in the north, which could be attributed to the ongoing clear-cutting of forests previously managed less intensively. Dead wood and large living trees increased not only in managed, but also in unmanaged forests, likely reflecting historical management. The increased tree species diversity can be explained through current forest management practices that encourages maintenance of additional tree species. Decreasing understory vegetation coverage in both dense managed and unmanaged forests suggests that factors other than forestry contribute to the ongoing changes in understory vegetation in Swedish forests. Overall, the observed increase in structural components has not yet been reflected in documented improvements for red-listed forest species, which may be due to delays in species responses to small improvements, as well as a lack of detailed monitoring. Similarly, the increased availability of forest structural components might still be insufficient to meet the specific habitat requirements of red-listed species.

Water Limitation in Forest Soils Regulates the Increase in Weathering Rates under Climate Change

Climate change is generally expected to have a positive effect on weathering rates, due to the strong temperature dependence of the weathering process. Important feedback mechanisms such as changes in soil moisture, tree growth and organic matter decomposition can affect the response of weathering rates to climate change. In this study, the dynamic forest ecosystem model ForSAFE, with mechanistic descriptions of tree growth, organic matter decomposition, weathering, hydrology and ion exchange processes, is used to investigate the effects of future climate scenarios on base cation weathering rates. In total, 544 productive coniferous forest sites from the Swedish National Forest Inventory are modelled, and differences in weathering responses to changes in climate from two Global Climate Models are investigated. The study shows that weathering rates at the simulated sites are likely to increase, but not to the extent predicted by a direct response to elevated air temperatures. Besides the result that increases in soil temperatures are less evident than those in air temperature, the study shows that soil moisture availability has a strong potential to limit the expected response to increased temperature. While changes in annual precipitation may not indicate further risk for more severe water deficits, seasonal differences show a clear difference between winters and summers. Taking into account the seasonal variation, the study shows that reduced soil water availability in the summer seasons will strongly limit the expected gain in weathering associated with higher temperatures.

Comparing basal area growth models for Norway spruce and Scots pine dominated stands

Models that predict forest development are essential for sustainable forest management. Constructing growth models via regression analysis or fitting a family of sigmoid equations to construct compatible growth and yield models are two ways these models can be developed. In this study, four species-specific models were developed and compared. A compatible growth and yield stand basal area model and a five-year stand basal area growth model were developed for Scots pine ( L.) and Norway spruce ( (L.) Karst.). The models were developed using data from permanent inventory plots from the Swedish national forest inventory and long-term experiments. The species-specific models were compared, using independent data from long-term experiments, with a stand basal area growth model currently used in the Swedish forest planning system Heureka (Elfving model). All new models had a good, relatively unbiased fit. There were no apparent differences between the models in their ability to predict basal area development, except for the slightly worse predictions for the Norway spruce growth model. The lack of difference in the model comparison showed that despite the simplicity of the compatible growth and yield models, these models could be recommended, especially when data availability is limited. Also, despite using more and newer data for model development in this study, the currently used Elfving model was equally good at predicting basal area. The lack of model difference indicate that future studies should instead focus on model development for heterogeneous forests which are common but lack in growth and yield modelling research.Pinus sylvestrisPicea abies

Dynamics of the Swedish forest carbon pool between 2010 and 2015 estimated from satellite L-band SAR observations

Estimation of carbon stocks in forests and tracking their dynamics over time requires repeated and spatially dense observations. Sweden is one of the major forested countries on Earth with a carbon pool of approximately 1 PgC. Statistics on forest resources are regularly quantified by data collected at field inventory plots by the Swedish National Forest Inventory (NFI). Nonetheless, the sampling by the inventory does not provide detailed spatial knowledge of carbon stored in forests. Thanks to one of the worldwide densest datasets of L-band satellite synthetic aperture radar (SAR) images collected over the years, we assessed the estimation of carbon density on two years, i.e., 2010 and 2015, at pixel (25 m), landscape, county and national level, and investigated the reliability of changes over five years. While the maps of carbon density were obtained with a well-established approach, the uncertainty of the estimates were obtained with a novel procedure that accounted for observational variances and co-variances. The high spatial resolution of the maps revealed the small-scale patterns of carbon density in Swedish forests. However, the substantial uncertainties at the pixel level (25%–62% of the estimated value) prevented from a reliable estimate of carbon density difference at the pixel level. Average values at landscape and county scale were instead unbiased and in strong agreement with corresponding numbers from the Swedish NFI. The uncertainty at county level was between 3% and 12% of the reference mean value. Overall, the net carbon balance for Sweden was positive, with a magnitude comparable to estimates derived from data collected by the NFI. Between 2010 and 2015, Swedish forests acted as a carbon sink, with a net increase of carbon stocks by 2.5%. The high level of confidence in our results indicate that satellite L-band observations are potentially suitable for tracking large-scale forest carbon dynamics in boreal forests provided that multiple observations are available per year and that they are acquired with the same viewing geometry.

Predicting bilberry and cowberry yields using airborne laser scanning and other auxiliary data combined with National Forest Inventory field plot data

The increasing availability of wall-to-wall remote sensing datasets in combination with accurate field data enables the mapping of different ecosystem services more accurately and over larger areas than before. The provision of wild berries is an essential ecosystem service, and berries are the most used non-wood forest products in Nordic countries. The aim of the study was to 1) develop general prediction models for bilberry and cowberry yield based on metrics derived from airborne laser scanning (ALS) data and other existing wall-to-wall data and 2) to identify laser-based structural features of forests that can be linked to locations of the highest berry yields. We used the indirect approach where the correlation between forest structure described by the ALS data and the berry yields are utilized. Berry data collected in the Swedish National Forest Inventory (NFI) 2007–2016 were used for training the models and ALS data from 2009 to 2014 from the national ALS campaign of Sweden. Berry yields were modelled using generalised linear mixed models (GLMMs), and forest structural differences were demonstrated in histograms of presence/absence data.The ALS-based canopy cover was an important variable both in bilberry and cowberry models. Other significant variables were ALS-based height variance, shrub cover, height above sea level, slope, soil wetness and terrain ruggedness, satellite-based species-specific volume and percentage, seasonality of temperature and precipitation and annual precipitation, inventory year, soil type and land use class. In addition, the time difference between the inventory day and the Julian day when berries were expected to be ripe showed a 1.5% decrease for bilberry and a 1.1% decrease for cowberry yield per day during the season. The highest bilberry yield was identified in forests with a canopy cover of 50% and the highest cowberry yield in forests with a canopy cover close to zero. The canopy height of 15 m reflected the highest bilberry yield, whereas a canopy height close to 0 m resulted in the highest cowberry yield. The shrub cover was close to zero both with highest bilberry and cowberry yields.This is the first study combining ALS metrics with other wall-to-wall variables and NFI field data to model bilberry and cowberry yields. Prediction models can be used to produce maps showing the most potential locations for berry picking. Further, the models may, in the future, be imported into forest planning systems to obtain stand-level prognoses of berry yield development under different forest management strategies.

Rapid Changes in Ground Vegetation of Mature Boreal Forests—An Analysis of Swedish National Forest Inventory Data

The boreal forest floor vegetation is critical for ecosystem functioning and an important part of forest biodiversity. Given the ongoing global change, knowledge on broad-scale changes in the composition and abundance of different plant species and species groups is hence important for both forest conservation and management. Here, we analyse permanent plot data from the National Forest Inventory (NFI) on changes in the vegetation over a 10-year period in four regions of Sweden. To limit the direct and relatively well-known effects of forest management and associated succession, we only included mature forest stands not influenced by forestry during the 10 years between inventories, and focused on vegetation change mainly related to other factors. Results show strong decrease among many species and species groups. This includes dominant species such as Vaccinimum myrtillus and Deschampsia flexuosa as well as several forest herbs. The only species increasing are some mosses in the southern regions. Our data do not allow for a causal interpretation of the observed patterns. However, the changes probably result from latent succession in combination with climate change and nitrogen deposition, and with time lags complicating the interpretation of their relative importance. Regardless of the cause, the observed changes are on a magnitude that suggest impacts on ecosystem functioning and hence highlight the need for more experimental work.

Forest fragmentation assessment using field-based sampling data from forest inventories

ABSTRACT Forest fragmentation has a relevant impact on biodiversity. An interesting alternative to estimate these indices is to use sampling data. This study aims to estimate aggregation index (AI) and the degree of clumping of forested landscape based on AI. The assessment was conducted using different point distances, inventory regions and cardinal directions. For this purpose, a dataset from one five-year periods (2007–2011) of the Swedish National Forest Inventory (NFI) was used. The estimation of AI from field-based inventory can give us a general picture of the current status of forest landscape. The results also show that the estimated AI is a distance dependent function. The corresponding estimated variance of the index is smaller for longer distances. The obtained results indicate that the estimated variance depends on both sample size and pair point distances. Estimated AI showed different values in different cardinal directions. To compare two regions or a given region over time, a given point distance should be used. The main advantage of the applied procedure is that a range of AI values can be produced rather than a single number. Furthermore, in field-based inventory, the obtained results are more reliable, because one works implicitly with a single forest definition only.

Keeping mixtures of Norway spruce and birch in production forests: insights from survey data

ABSTRACT Admixtures of birch in Norway spruce plantations are being promoted as a means to increase habitat and species diversity. The implications of this mixture were analysed with regional survey data from southern Sweden. Permanent sample plots from the Swedish National Forest Inventory (NFI), with Norway spruce and admixture of birch, were used to describe the temporal trends in the admixture, regarding species composition and competitive strength. Observations from thinned plots show a higher harvest removal in birch (35%) than for Norway spruce (19%). Observations without thinnings in the period before measurement showed that individual birch tree growth was lower compared to Norway spruce and it decreased even more with increasing stand age and competition. In addition, a complementary field survey, with multiple distributed sample plots in each stand, was used to detect within-stand variation of species composition and density. Although within-stand heterogeneity was larger in mixed stands in terms of species composition, it was not different from Norway spruce monocultures in terms of stand density. These two surveys show that the admixture of birch, for several reasons, decreases over stand age and although birch increases tree species diversity, it does not necessary imply a change in density.

Combining TanDEM-X and Sentinel-2 for large-area species-wise prediction of forest biomass and volume

In this study, data from the satellite sensors TanDEM-X and Sentinel-2 were combined with national field inventory data to predict forest above-ground biomass (AGB) and stem volume (VOL) over a large area in Sweden. The data sources were evaluated both separately and in combination. The study area covers approximately 20,000,000 ha and corresponds to about 70% of the Swedish forest area. The study area was divided into tiles of 2.5 × 2.5 km2, which were processed sequentially. The field plots were inventoried on 7 m and 10 m circular plots by the Swedish National Forest Inventory, and plot AGB and VOL at the year of the satellite data were estimated based on a 10-year period of field data. The AGB and VOL were modelled using the k nearest neighbor (kNN) algorithm, with k = 5 neighbors. The combined use of two data sources with different scene extents enabled the generation of seamless AGB and VOL maps. Moreover, the kNN algorithm provided the VOL divided per tree species, which was used for classification of the dominant tree species at stand-level. The overall accuracy for the dominant tree species classification was 77%. The predicted AGB and VOL rasters were evaluated using 549 field inventoried forest stands distributed over Sweden. The RMSE for the predictions based on both data sources were 31.4 t/ha (29.1%) for AGB, and 59.0 m3/ha (30.2%) for VOL. By estimating and removing the variance due to sampling (the stand values were estimated from sample plots), the RMSE was improved to 18.0 t/ha (16.6%). The evaluated approach of using kNN was suitable for estimating forest variables from a combination of different satellite sensors, provided sufficient field reference data are available. The TanDEM-X data were most important for the AGB and VOL predictions, while Sentinel-2 data were essential to map the tree species.

Integration of allometric equations in the water cloud model towards an improved retrieval of forest stem volume with L-band SAR data in Sweden

Much attention is paid to the estimation of forest biomass-related variables (stem volume and above-ground biomass) with synthetic aperture radar (SAR) backscatter images because of the increasing number of sensors in space providing global and repeated coverage and the sensitivity of the backscattered intensity to forest properties. One of the most popular models used to estimate a biomass-related variable from SAR backscatter observations is the Water Cloud Model (WCM) because of its simplicity allowing for a straightforward retrieval. Nonetheless, a common feature of these estimates is the tendency to over- or underestimate specific ranges due to simplifying assumptions in the model. In this study, the WCM has been revisited by exploring pathways for a physically-based, Light Detection and Ranging (LiDAR)-aided, model parameterization at larger scale with the overall aim to reduce systematic retrieval errors associated with empirical assumptions in the model. The study was undertaken in Sweden where repeated observations of backscatter by the Advanced Land Observing Satellite (ALOS) Phased Array-type L-band Synthetic Aperture Radar (PALSAR) were available. The integration was prototyped in Sweden thanks to detailed allometries relating forest variables in the WCM. These were derived from spatially dense estimates of canopy density and vegetation height from observations by the Ice, Cloud and land Elevation Satellite (ICESat) Geoscience Laser Altimeter System (GLAS) and measurements of height and stem volume from the Swedish National Forest Inventory (NFI). The SAR backscatter predicted by the revisited WCM was in strong agreement with the observations. When evaluated against stem volumes estimated from the NFI data, the SAR-based stem volumes presented strong dispersion at the pixel level. Average stem volume at the level of five or more pixels, i.e., for an area larger than 0.3 ha, were instead unbiased and similar to the average values obtained from the NFI data (relative root mean square error: 21.4%, estimation bias: 0.9 m3/ha and coefficient of determination: 0.67). This study demonstrates that the integration of allometries in the WCM effectively reduces estimation errors. The method here prototyped in Sweden qualifies to provide large-scale estimates of biomass-related variables using multiple observations of L-band backscatter with potential application worldwide.