Search

Land surface phenology (LSP) can improve the monitoring of forest areas and their change processes. The aim of this study was to characterize the temporal dynamics in Mediterranean pines and evaluate the potential of LSP for species discrimination. We used 661 mono-specific plots for five different Pinus species (Pinus halepensis, P. pinea, P. pinaster; P. sylvestris, P. nigra) and the MOD13Q1-NDVI time series (2000–2016) to perform the analyses. The time series were smoothed to extract the phenological parameters and calculate multi-temporal metrics, to synthesize the inter-annual variability. The potential of LSP for discriminating between Pinus species was evaluated by the application of the Random Forest (RF) classifier from different subsets of explanatory variables: i) the smooth time series; ii) the multi-temporal metrics; and iii) the multi-temporal metrics plus the auxiliary physical variables. This latter subset was also used as input to a Classification and Regression Tree (CART) algorithm to better explain the differences between Pinus species regarding LSP parameters and other environmental drivers. The analysis showed two different patterns: an important NDVI decrease during the summer for P. halepensis, P. pinea, and P. pinaster; and lower NDVI variation along the year for P. sylvestris. P. nigra showed a heterogeneous intra-specific behavior, having locations with different patterns. We distinguished Pinus species plots with a global accuracy of 0.82, when we used multi-temporal metrics of LSP and auxiliary physical variables. More generally, the Mediterranean Pinus species could be differentiated considering the 23rd of July as the start of season and 179 km and 1100 m as distance to the coastline and elevation, respectively.

4 Pág. Centro de Investigación Forestal (CIFOR)., In 2019, 100 years had elapsed since the first National Forest Inventory (NFI) was established in Norway. Motivated by a fear of over-exploitation of timber resources, NFIs today enable informed policy making by providing data vital to decision support at international, national, regional, and local scales. This Collection of articles celebrates the 100th anniversary of NFIs with a description of past, present, and future research aiming at improving the monitoring of forest and other terrestrial ecosystems., The Beijing Forestry University waived the Open Access fees of all studies published in this Collection.

Background The increasing availability of remotely sensed data has recently challenged the traditional way of performing forest inventories, and induced an interest in model-based inference. Like traditional design-based inference, model-based inference allows for regional estimates of totals and means, but in addition for wall-to-wall mapping of forest characteristics. Recently Light Detection and Ranging (LiDAR)-based maps of forest attributes have been developed in many countries and been well received by users due to their accurate spatial representation of forest resources. However, the correspondence between such mapping and model-based inference is seldom appreciated. In this study we applied hierarchical model-based inference to produce aboveground biomass maps as well as maps of the corresponding prediction uncertainties with the same spatial resolution. Further, an estimator of mean biomass at regional level, and its uncertainty, was developed to demonstrate how mapping and regional level assessment can be combined within the framework of model-based inference. Results Through a new version of hierarchical model-based estimation, allowing models to be nonlinear, we accounted for uncertainties in both the individual tree-level biomass models and the models linking plot level biomass predictions with LiDAR metrics. In a 5005 km2 large study area in south-central Sweden the predicted aboveground biomass at the level of 18 m ×18 m map units was found to range between 9 and 447 Mg ·ha−1. The corresponding root mean square errors ranged between 10 and 162 Mg ·ha−1. For the entire study region, the mean aboveground biomass was 55 Mg ·ha−1 and the corresponding relative root mean square error 8%. At this level 75% of the mean square error was due to the uncertainty associated with tree-level models. Conclusions Through the proposed method it is possible to link mapping and estimation within the framework of model-based inference. Uncertainties in both tree-level biomass models and models linking plot level biomass with LiDAR data are accounted for, both for the uncertainty maps and the overall estimates. The development of hierarchical model-based inference to handle nonlinear models was an important prerequisite for the study.

Background Forests are an important component of the global carbon (C) cycle and can be net sources or sinks of CO2, thus mitigating or exacerbating the effects of anthropogenic greenhouse gas emissions. While forest productivity is often inferred from national-scale yield tables or from satellite products, forest C emissions resulting from dead organic matter decay are usually simulated, therefore it is important to ensure the accuracy and reliability of a model used to simulate organic matter decay at an appropriate scale. National Forest Inventories (NFIs) provide a record of carbon pools in ecosystem components, and these measurements are essential for evaluating rates and controls of C dynamics in forest ecosystems. In this study we combine the observations from the Swiss NFIs and machine learning techniques to quantify the decay rates of the standing snags and downed logs and identify the main controls of dead wood decay. Results We found that wood decay rate was affected by tree species, temperature, and precipitation. Dead wood originating from Fagus sylvatica decayed the fastest, with the residence times ranging from 27 to 54 years at the warmest and coldest Swiss sites, respectively. Hardwoods at wetter sites tended to decompose faster compared to hardwoods at drier sites, with residence times 45–92 and 62–95 years for the wetter and drier sites, respectively. Dead wood originating from softwood species had the longest residence times ranging from 58 to 191 years at wetter sites and from 78 to 286 years at drier sites. Conclusions This study illustrates how long-term dead wood observations collected and remeasured during several NFI campaigns can be used to estimate dead wood decay parameters, as well as gain understanding about controls of dead wood dynamics. The wood decay parameters quantified in this study can be used in carbon budget models to simulate the decay dynamics of dead wood, however more measurements (e.g. of soil C dynamics at the same plots) are needed to estimate what fraction of dead wood is converted to CO2, and what fraction is incorporated into soil.

Non-native trees enhance services that are fundamental for human well-being. Yet, the extensive use of non-native trees has the potential of causing environmental and socio-economic harm. Eucalyptus and Pinus are the most widely distributed and extensively planted tree genera worldwide, because their rapid growth allows profitable production of timber and pulp. Their naturalization is causing severe effects on the environment, but the relative importance of underlying factors determining their demographic performance is not well known. Thus, our aim was to evaluate the relative importance of biotic, abiotic, and anthropogenic factors driving demographic changes of Eucalyptus and Pinus at the regional scale. We compiled environmental variables and demographic data for Eucalyptus globulus, Eucalyptus camaldulensis, and Pinus radiata across 6388 permanent forestland plots surveyed in the Spanish Forest Inventory (SFI). We used the second (1986?1996), third (1997?2007), and fourth (2008?2017) SFI datasets to quantify annual changes in basal area per plot between consecutive inventories (?BA; m2 ha?1 year?1). We also quantified the components of ?BA: tree ingrowth (transitions from juvenile to adult trees), growth, and mortality. We evaluated juvenile recruitment with in-situ regeneration (No. juvenile trees ha?1) within plots already occupied by the focal species in the previous inventory, and with natural colonization of plots where the focal species was absent at the beginning of the time interval. We found that the structure of the biotic community was especially important to explain demographic performance of non-native trees growing in benign environments (E. globulus and P. radiata), whereas abiotic factors were particularly important regulating basal area increments of E. camaldulensis, which occurs in harsher environments. Basal area increments decreased with species and functional richness, heterospecific density, mean annual temperature, and increased with soil capacity to retain nutrients and water. Colonization of new plots increased with propagule availability in the surrounding landscape. Tree cutting was beneficial for P. radiata. Collectively, our results suggest that non-native trees perform better in forests with high propagule pressure, low biotic resistance, favourable abiotic conditions, and human management., Agencia Estatal de Investigación, Universidad de Alcalá, Ministerio de Ciencia e Innovación, Comunidad de Madrid

Tree heights are one of the most important aspects of forest mensuration, but data are often unavailable due to costly and time-consuming field measurements. Therefore, various types of models have been developed for the imputation of tree heights for unmeasured trees, with mixed-effects models being one of the most commonly applied approaches. The disadvantage here is the need of sufficient sample size per tree species for each plot, which is often not met, especially in mixed forests. To avoid this limitation, we used principal component analysis (PCA) for the grouping of similar plots based on the most relevant site descriptors. Next, we compared mixed-effects models with height-diameter models based on artificial neural networks (ANN). In terms of root mean square error (RMSE), mixed-effects models provided the most accurate tree height predictions at the plot level, especially for tree species with a smaller number of tree height measurements. When plots were grouped using the PCA and the number of observations per category increased, ANN predictions improved and became more accurate than those provided by mixed-effects models. The performance of ANN also increased when the competition index was included as an additional explanatory variable. Our results show that in the pursuit of the most accurate modelling approach for tree height predictions, ANN should be seriously considered, especially when the number of tree measurements and their distribution is sufficient.

Random forest (RF) models were developed to estimate the biomass for the Pinus radiata species in a region of the Basque Autonomous Community where this species has high cover, using the National Forest Inventory, allometric equations and low-density discrete LiDAR data. This article explores the tuning for RF hyperparameters, obtaining two models with an R2 higher than 0.7 using 2-fold cross-validation. The models selected were applied in Orozko, a municipality with more than 5000 ha of this species, where the model predicts a biomass of 1.06–1.08 Mton, which is between 16–18 % higher than the biomass predicted by the Basque Government. The work reported in this paper was partially supported by FEDER funds for the MINECO project TIN2017-85827-P and project KK-202000044 of the Elkartek 2020 funding program of the Basque Government. Additional support comes from grant IT1284-19 of the Basque Autonomous Community.

National forest inventories (NFI) provide estimates of forest resources at the national and regional level but are also increasingly used as basis for mapping forest resources based on remotely sensed data. Such maps procure local estimates of forest resources but may also improve precision of national and regional estimates. Supported by a countrywide airborne laser scanning (circa 2014) and a national land-use map (circa 2014), direct (DI), model-assisted (MA), and model calibrated (MC) estimates of wood volume (V) and aboveground biomass (AGB) densities in forest areas derived from the Danish NFI (2012–2016) are presented. Nonlinear models with three LiDAR metrics are used to predict V and AGB in forested areas. According to these models, the predicted values of V and AGB in sample plots missed in the field inventory was lower than in those visited in the field; we therefore opted for estimation with multiple (stochastic) imputations. MA estimates for the country suggested a 2% lower level of both V and AGB densities with errors 45% lower than estimated errors in DI results. National MC estimates were close to the DI estimates with an error approximately 40% lower than errors in DI estimates yet 5% greater than the MA estimates of error. Multiple imputations had the strongest impact on DI estimates, but only a weak impact on MA and MC results.

Large-scale surveys, such as national forest inventories and vegetation monitoring programs, usually have complex sampling designs that include geographical stratification and units organized in clusters. When models are developed using data from such programs, a key question is whether or not to utilize design information when analyzing the relationship between a response variable and a set of covariates. Standard statistical regression methods often fail to account for complex sampling designs, which may lead to severely biased estimators of model coefficients. Furthermore, ignoring that data are spatially correlated within clusters may underestimate the standard errors of regression coefficient estimates, with a risk for drawing wrong conclusions. We first review general approaches that account for complex sampling designs, e.g. methods using probability weighting, and stress the need to explore the effects of the sampling design when applying logistic regression models. We then use Monte Carlo simulation to compare the performance of the standard logistic regression model with two approaches to model correlated binary responses, i.e. cluster-specific and population-averaged logistic regression models. As an example, we analyze the occurrence of epiphytic hair lichens in the genus Bryoria; an indicator of forest ecosystem integrity. Based on data from the National Forest Inventory (NFI) for the period 1993–2014 we generated a data set on hair lichen occurrence on >100,000 Picea abies trees distributed throughout Sweden. The NFI data included ten covariates representing forest structure and climate variables potentially affecting lichen occurrence. Our analyses show the importance of taking complex sampling designs and correlated binary responses into account in logistic regression modeling to avoid the risk of obtaining notably biased parameter estimators and standard errors, and erroneous interpretations about factors affecting e.g. hair lichen occurrence. We recommend comparisons of unweighted and weighted logistic regression analyses as an essential step in development of models based on data from large-scale surveys.

Stand volume growth varies with age, site index, and stand density. Analyses of changes in volume growth across stand ages need to consider the non-linear effects of all independent variables and their interactions, as is typically done in forest growth models. This is particularly important for unbalanced data, like the national forest inventory data. Stokland’s (2021) analysis did not use this approach, limiting any insights about age effects on volume growth of old stands. Other details of the national forest inventory data in Norway will need special attention in the analysis of age trends in old stands: a lack of fully stocked stands, different management regimes and sites for older and younger stands, and rough estimates of stand ages for old stands.

International organizations increasingly require estimates of forest parameters to monitor the state of and changes in forest resources, the sustainability of forest practices and the role of forests in the carbon cycle. Most countries rely on data from their national forest inventories (NFI) to produce these estimates. However, because NFI survey years may not match the required reporting years, techniques for updating NFI-based estimates are necessary. The main aim was to develop an unbiased method to update NFI estimates of mean growing stock volume (m 3 /ha) using models to predict annual plot-level volume change, and to estimate the associated uncertainties. Because the final large area volume estimates were based on plot-level model predictions rather than field observations, hybrid inference was necessary to accommodate both model prediction uncertainty and sampling variation. Specific objectives were to compare modelling approaches, to assess the utility of Landsat data for increasing model prediction accuracy, to select the most accurate method, and to compare model-based and design-based uncertainty components. For four monospecific forest types, data from the 2nd and 3rd Spanish NFI surveys together with site variables and Landsat images were used to construct models to predict NFI information for the year of the 4th NFI survey. Data from the 3rd and 4th surveys were used to assess the accuracy of the model predictions at both plot-level and large area spatial scales. The most accurate method used a set of three models: one to predict the probability of volume removals, one to predict the amount of removed volume, and one to predict gross annual volume. Incorporation of Landsat-based variables in the models increased prediction accuracy. Differences between large area estimates based on plot-level field observations for the 4th NFI survey and estimates based on the model predictions were minimal for all four forest types. Further, the standard errors of the estimates based on the model predictions were only slightly greater than standard errors based on the field observations. Thus, model predictions of plot-level growing stock volume based on field and satellite image data as auxiliary information can be used to update large area NFI estimates for reporting years for which spectral data are available but field observations are not. Finally, variances of means are under-estimated unless hybrid inferential methods are used to incorporate both model prediction uncertainty and sampling variation. For the two forest types for which the two sources of uncertainty were of the same order of magnitude, the under-estimation was non-negligible.

We used national scenario analyses to examine the effects of harvesting intensity on the development of forest resources, timber supply, carbon balance, and biodiversity indicators of Finnish forestry in nine 10-year simulation periods (90-year simulation period) under the current climate. Data from the 11th National Forest Inventory of Finland were used to develop five even-flow harvesting scenarios for non-protected forests with the annual harvest ranging from 40 to 100 million m3. The results show that the highest annual even-flow harvest level, which did not decrease the growing stock volume over the 90-year simulation period, was 73 million m3. The total 90-year timber production, consisting of harvested volume and change in growing stock volume, was maximized when the annual harvest was 60 million m3. Volume increment increased for several decades when harvested volume was less than the current volume increment. The total carbon balance of forestry was the highest with low volume of harvested wood. Low harvested volume increased the values of biodiversity indicators, namely volume of deciduous trees, amount of deadwood and area of old forest., final draft, peerReviewed

This study documents volume increment and natural mortality in 1379 old boreal forests plots during four consecutive inventory cycles in the Norwegian national forest inventory. The stands age up to 100 years beyond recommended rotation length (close to economical optimal rotation length) and comprise a wide range of site productivity classes in both pine- and spruce-dominated forests. The annual gross volume increment was stable and nearly constant up to 50–100 years beyond economically optimal rotation length. In parallel, there was very low natural mortality (0.22–0.66% of standing volume) with minimal risk of stand collapse. Stands with satisfactory stocking had volume increment equal to or higher than the reference volume increment in managed stands harvested at recommended rotation length, while poorly stocked stands had inferior volume increment. From a climate change mitigation perspective, it seems to be a good strategy to extend the rotation length beyond what is currently recommended, provided that the stands have satisfactory stocking.

International audience; For forest ecosystems, studies dealing with diversity-productivity relationships are often based on diameter increment observations. Studying how height growth is affected by species interactions can provide new insights on this issue. We studied the mixture effect on dominant height growth in order to answer two questions. Do species interactions in mixed forest modify the dominant height growth of species? Does the diversity effect on diameter found in previous studies correspond to actual overyielding, or rather to an effect on allocation of growth between diameter and height? We used the French National Forest Inventory (NFI) data to model the mixture effect on dominant height. We included biophysical factors in the models to compare the dominant height of mixed and monospecific stands, all other parameters being equal. We studied five target species – Quercus petraea (Matt.) Liebl., Fagus sylvatica L., Picea abies (L.) Karst., Abies alba Mill., and Pinus sylvestris L. – in association with sixteen other species. Mixture effects on dominant height were weak, though often significant. They were either positive or negative according to species association. We showed that mixture effect on dominant height corresponds to a leveling process between species: the taller one limits its growth while the smaller one’s growth increases. Furthermore, most of the time, mixture effects on dominant height are in the same direction as those found on diameter, though with a lower magnitude. Our results confirm that tree diversity results in overyielding rather than in a different allocation of volume between the parts of the tree.

Almost 50% of the land in Galicia (NW Spain) is covered by forest (minimum area of land of 1 ha with tree crown cover of more than 10%) and the region is the top timber–producing area in Spain. Regional changes in forest live tree biomass and carbon stocks were estimated from National Forest Inventory (NFI) data covering almost 40 years, i.e. between NFI-1 (1972) and NFI–4 (2009). Carbon stocks were estimated from biomass of different tree species and components, including roots. Biomass equations were applied to the tree stock data for different species and diameter classes. The biomass increased in all species or groups for the period considered. However, although the biomass of softwood species (mainly Pinus spp.) increased, because of the increase in the number of trees belonging to the largest diameter classes (DC), the balance of the number of small trees (DC

14 Pág., The quantification of forests available for wood supply (FAWS) is essential for decision-making with regard to the maintenance and enhancement of forest resources and their contribution to the global carbon cycle. The provision of harmonized forest statistics is necessary for the development of forest associated policies and to support decision-making. Based on the National Forest Inventory (NFI) data from 13 European countries, we quantify and compare the areas and aboveground dry biomass (AGB) of FAWS and forest not available for wood supply (FNAWS) according to national and reference definitions by determining the restrictions and associated thresholds considered at country level to classify forests as FAWS or FNAWS. FAWS represent between 75 and 95 % of forest area and AGB for most of the countries in this study. Economic restrictions are the main factor limiting the availability of forests for wood supply, accounting for 67 % of the total FNAWS area and 56 % of the total FNAWS AGB, followed by environmental restrictions. Profitability, slope and accessibility as economic restrictions, and protected areas as environmental restrictions are the factors most frequently considered to distinguish between FAWS and FNAWS. With respect to the area of FNAWS associated with each type of restriction, an overlap among the restrictions of 13.7 % was identified. For most countries, the differences in the FNAWS areas and AGB estimates between national and reference definitions ranged from 0 to 5 %. These results highlight the applicability and reliability of a FAWS reference definition for most of the European countries studied, thereby facilitating a consistent approach to assess forests available for supply for the purpose of international reporting., This research was supported by the Specific contract n. 18 “Use of National Forest Inventories data to estimate area and above ground biomass in European forests not available for wood supply” in the context of the Framework contract for the provision of forest data and services supporting the European Forest Data Centre 2012/ S 78-127532 of 21/04/2012 of the Joint Research Centre of the European Commission; the EG-013-72 agreement of the Ministry of Agriculture, Fisheries and Food (MAPA) and the INIA belonging to the Spanish Ministry of Science and Innovation (MICINN); and the project No.APVV-15-0265 granted by the Slovak Research and Development Agency.

Spatial predictions of forest variables are required for supporting modern national and sub-national forest planning strategies, especially in the framework of a climate change scenario. Nowadays methods for constructing wall-to-wall maps and calculating small-area estimates of forest parameters are becoming essential components of most advanced National Forest Inventory (NFI) programs. Such methods are based on the assumption of a relationship between the forest variables and predictor variables that are available for the entire forest area. Many commonly used predictors are based on data obtained from active or passive remote sensing technologies. Italy has almost 40% of its land area covered by forests. Because of the great diversity of Italian forests with respect to composition, structure and management and underlying climatic, morphological and soil conditions, a relevant question is whether methods successfully used in less complex temperate and boreal forests may be applied successfully at country level in Italy. For a study area of more than 48,657 km2 in central Italy of which 43% is covered by forest, the study presents the results of a test regarding wall-to-wall, spatially explicit estimation of forest growing stock volume (GSV) based on field measurement of 1350 plots during the last Italian NFI. For the same area, we used potential predictor variables that are available across the whole of Italy: cloud-free mosaics of multispectral optical satellite imagery (Landsat 5 TM), microwave sensor data (JAXA PALSAR), a canopy height model (CHM) from satellite LiDAR, and auxiliary variables from climate, temperature and precipitation maps, soil maps, and a digital terrain model. Two non-parametric (random forests and k-NN) and two parametric (multiple linear regression and geographically weighted regression) prediction methods were tested to produce wall-to-wall map of growing stock volume at 23-m resolution. Pixel level predictions were used to produce small-area, province-level model-assisted estimates. The performances of all the methods were compared in terms of percent root mean-square error using a leave-one-out procedure and an independent dataset was used for validation. Results were comparable to those available for other ecological regions using similar predictors, but random forests produced the most accurate results with a pixel level R2 = 0.69 and RMSE% = 37.2% against the independent validation dataset. Model-assisted estimates were more precise than the original design-based estimates provided by the NFI.

Angle-count sampling (ACS) is an established method in forest mensuration and is implemented in different National Forest Inventories (NFI). However, due to the lack of fixed reference areas of the inventory plots, these ACS-based field data are seldom used as training data for wall-to-wall mapping applications at forest enterprise level. In this paper, we demonstrate an approach to overcome this shortcoming. For a study area in northern Bavaria, Germany, we used ACS-based NFI data for model training to generate wall-to-wall maps of growing stock for broadleaf, conifer and mixed forest stands. Both spectral and height information from the very high resolution WorldView-2 (WV2) satellite were used as auxiliary information and the non-parametric Random Forests (RF) algorithm was chosen as modeling approach. The growing stock predictions were validated using out-of-bag (OOB) samples and further verified at the plot and stand level using additional data. For validation, field plots from a Management Forest Inventory (MFI) and delineated forest stands were used. Compared to stand-level aggregations based on field plots from the MFI, our approach explained 56% of the variability in the growing stock ( R 2 ) with a relative RMSE of 15% at the stand level ( n = 252). As expected, the scatter was higher at the plot-level ( n = 3973). Nonetheless, the models still achieved acceptable performance measures ( R 2 = 0.44; RMSE = 34%).

Sustainable forest management (SFM) is many things to many people – yet a common thread is the production of forest goods and services for the present and future generations. The promise of sustainability is rooted in the two premises; first that ecosystems have the potential to renew themselves and second that economic activities and social perceptions or values that define human interaction with the environment are choices that can be modified to ensure the long term productivity and health of the ecosystem. SFM addresses a great challenge in matching the increasing demands of a growing human population while maintaining ecological functions of healthy forest ecosystems. This paper does not seek to define SFM, but rather provides analyses of key indicators for the national-scale enabling environment to gain a global insight into progress in implementing enabling and implementing SFM at the national and operational levels. Analyses of the Global Forest Resources Assessment 2015 (FRA) country report data are used to provide insights into the current state of progress in implementing the enabling conditions for SFM. Over 2.17 billion ha of the world’s forest area are predicted by governments to remain in permanent forest land use, of which some 1.1 billion ha are covered by all of the SFM tools investigated in FRA 2015. At the global scale, SFM-related policies and regulations are reported to be in place on 97% of global forest area. While the number of countries with national forest inventories has increased over that past ten years from 48 to 112, only 37% of forests in low income countries are covered by forest inventories. Forest management planning and monitoring of plans has increased substantially as has forest management certification, which exceeded a total of over 430 million ha in 2014. However, 90% of internationally verified certification is in the boreal and temperate climatic domains – only 6% of permanent forests in the tropical domain have been certified as of 2014. Results show that more work is needed to expand the extent and depth of work on establishing the enabling conditions that support SFM over the long term and suggests where those needs are greatest.

Forest ecosystem carbon (C) stocks and stock change in the United States (US) have been documented using Intergovernmental Panel on Climate Change (IPCC) procedures and guidance with 1990 as a baseline reference for all United Nations Framework Convention on Climate Change reports. In the US, estimates of forest C stocks and stock change are obtained from data collected and maintained by the Forest Inventory and Analysis (FIA) program of the US Forest Service. Over the course of the IPCC monitoring period, the FIA program made a transition from state-by-state multiyear periodic inventories selected on a rotating basis – with reporting standards largely tailored to regional requirements – to nationally consistent, annual inventories (where a proportion of plots is measured in each state each year) designed for large-scale strategic requirements. Lack of measurements on all forest land during the periodic inventories, along with plot access difficulties and misidentification of forest plots as nonforest due to poor aerial imagery, have resulted in missing data (i.e., nonresponse) throughout the FIA database. Nonresponse, which in some US states is greater than 20%, may lead to differences in estimates of forest C stock change due to the procedural transition from periodic to annual inventories. As an initial step towards rectifying the differences in estimates, we examined several strategies to compensate for missing observations using the most recent annual inventory data from the Lake States region of the US. Results varied by state in the study but given the annual reporting cycle and requirements to compile national estimates of forest C, it was deemed that techniques, where non-observed samples are removed from estimation procedures, provided the optimal combination of statistical performance and efficiency. While the initial analysis focused on the Lake States region, several compensation strategies described may be useful in bridging the gap between national C flux estimates from periodic and annual forest inventories.

tThe estimation of past and future forest carbon (C) dynamics in European countries is a challengingtask due to complex and varying silvicultural systems, including uneven-aged forest management, andincomplete inventory data time series. In this study, we tested the use of the Carbon Budget Model ofthe Canadian Forest Sector (CBM-CFS3) in Italy, a country exemplifying most of these challenges. Our objective was to develop estimates of forest carbon budgets of the Forest Management area (including allforests existing in 1990) for the period 1995–2009, and to simulate alternative scenarios of natural dis-turbance (fire) and harvest rates to 2020. A number of methodological challenges required modifications to the default model implementation. Based on National Forest Inventory (NFI) data, we (i) developed a historic library of yield curves derived from standing volume and age data, reflecting the effect of past silvicultural activities and natural disturbances, and a current library of yield curves derived from thecurrent net annual increment; (ii) reconstructed the age structure for a period antecedent to the refer-ence NFI year (2005), to compare the model results with data from other sources; and (iii) developeda novel approach for the simulation of uneven-aged forests. For the period 2000–2009, the model esti-mated an average annual sink of −23.7 Mt CO2yr−1excluding fires in Italy’s managed forests. Adding firesto the simulation reduced the sink to −20.5 Mt CO2yr−1. The projected sink (excluding all fires) for theyear 2020 was −23.4 Mt CO2yr−1assuming average (2000–2009) harvest rates. A 36% increase in harvestrates by 2020 reduced the sink to −17.3 Mt CO2yr−1. By comparing the model results with NFI data andother independent studies, we demonstrate the utility of the CBM-CFS3 both for estimating the currentforest sink in even-aged and more complex uneven-aged silvicultural systems in Italy, and for explor-ing the impact of different harvest and natural disturbances scenarios in managed forests. This studydemonstrates the utility of the CBM-CFS3 to national-scale estimation of past and future greenhousegas emissions and provides the foundation for the model’s future implementation to other European countries, JRC.H.3-Forest Resources and Climate

Retaining trees for conservation at final harvest is becoming increasingly common within forestry globally, especially connected to clearcutting. The main action is to leave single living and dead trees, tree patches and buffer strips, to benefit biodiversity and to enhance ecosystem functioning. We present the first national analysis of effects on structural components from applying the retention approach. In Sweden retention forestry has been practiced large-scale for about 25years, prescribed by the law and a requirement in certification standards. By analyzing data from the Swedish National Forest Inventory we found that the volume of dead trees (⩾100mm in diameter; single trees and trees in patches

This paper suggested simulation approaches for quantifying and reducing the effects of National Forest Inventory (NFI) plot location error on aboveground forest biomass and carbon stock estimation using the k-Nearest Neighbor (kNN) algorithm. Additionally, the effects of plot location error in pre-GPS and GPS NFI plots were compared. Two South Korean cities, Sejong and Daejeon, were chosen to represent the study area, for which four Landsat TM images were collected together with two NFI datasets established in both the pre-GPS and GPS eras. The effects of plot location error were investigated in two ways: systematic error simulation, and random error simulation. Systematic error simulation was conducted to determine the effect of plot location error due to mis-registration. All of the NFI plots were successively moved against the satellite image in 360° directions, and the systematic error patterns were analyzed on the basis of the changes of the Root Mean Square Error (RMSE) of kNN estimation. In the random error simulation, the inherent random location errors in NFI plots were quantified by Monte Carlo simulation. After removal of both the estimated systematic and random location errors from the NFI plots, the RMSE% were reduced by 11.7% and 17.7% for the two pre-GPS-era datasets, and by 5.5% and 8.0% for the two GPS-era datasets. The experimental results showed that the pre-GPS NFI plots were more subject to plot location error than were the GPS NFI plots. This study’s findings demonstrate a potential remedy for reducing NFI plot location errors which may improve the accuracy of carbon stock estimation in a practical manner, particularly in the case of pre-GPS NFI data.

Deadwood biomass is one of the five terrestrial carbon pools that are relevant for the estimation of carbon stocks and carbon stock changes under the UNFCCC and the Kyoto Protocol (IPCC Guidelines for National Greenhouse Gas Inventories). Statistics on this important ecosystem component are generally provided by national forest inventories: in the last few decades, national forest inventories in Europe have gradually included deadwood among the traditionally investigated components. However, deadwood is generally surveyed in terms of volume, which is then converted to biomass using basic density values. These are rarely specifically derived for the population under estimation; more often, values available from the literature and based on the findings of local studies are used instead. Based on a survey conducted by the Italian national forest inventory in 2008 and 2009 that specifically intended to more deeply investigate the carbon storage of forests, samples of deadwood were collected all over the country from the plots of the national forest inventory network and used to derive basic density values. This paper presents the basic densities obtained for groups of species (conifers, broadleaves), deadwood components (coarse woody debris, standing dead trees, stumps) and decay classes (using a five-class system). Fine woody debris was also surveyed, but in terms of weight and not volume. For such material, the ratio of dry weight to fresh weight for conversion to biomass is presented. In addition, the main aspects of the field protocol are discussed. Lastly, the use of less accurate, generalised basic density values is tested to evaluate its effects on deadwood biomass estimates. (C) 2013 Elsevier B.V. All rights reserved.

Although they were originally introduced for different purposes, forest models are often used today for scenario development, which includes forest production as one aspect of forest development. However, studies using an independent data set to compare different simulators are rarely found. In this study a subset of National Forest Inventory data for the whole of Austria was compared to simulations of the biogeochemistry model Biome-BGC, the hybrid gap model PICUS and a climate sensitive version of the growth and yield model PrognAus. The models were used to simulate the development of approximately 700 forest inventory sample plots over a period of 15?years. The study focussed on the models' sensitivity to varying environmental conditions; thus, the comparison was based on the mean current annual volume increment per hectare. All models showed a significant average deviation from the inventory (over- or under-estimation). The estimated year-to-year variation was best reproduced by PICUS. However, the 15?year growth trend was also shown by Biome-BGC and PrognAus. Potential model users interested in relating mean current annual volume increment to climate data will need to weigh accuracy against applicability when choosing among these models. Highlights? We compare simulations by a BGC model, a gap model and an empirical growth model. ? The comparison is based on the current annual volume increment for 1987-2001. ? All models show a significant average deviation from inventory. ? The year-to-year increment variation was best reproduced by the gap model. ? Complex models lose applicability due to extended input data needs.

National forest inventories report estimates of parameters related to forest area and growing stock volume for geographic areas ranging in size from municipalities to entire countries. Landsat imagery has been shown to be a source of auxiliary information that can be used with stratified estimation to increase the precision of estimates, although the increase is greater for estimates of forest area than for estimates of growing stock volume. The objective of the study was to assess the utility of lidar-based stratifications for increasing the precision of mean proportion forest area and mean growing stock volume per unit area. Stratifications based on nonlinear logistic regression model predictions of volume obtained from lidar data reduced variances of mean growing stock volume estimates by factors as great as 3.2 and variances of mean proportion forest area estimates by factors as great as 1.5.

An airborne profiling lidar was used to estimate total aboveground dry biomass, a surrogate for aboveground carbon stocks, for all land cover types in Hedmark County, Norway. One-hundred-five parallel profiling flight lines systematically spaced three km apart, totaling 8309 km, were acquired to measure forest canopy height and density across the 27,390 km2 County. Two sampling/estimation strategies were investigated; both use profiling lidar data to augment field data collected as part of a systematic network of Norwegian National Forest Inventory (NFI) ground plots. The two strategies include (1) a model-dependent (MD), two-phase approach which permits calculation of sampling and model error, and (2) a model-assisted (MA), probability-based, two-stage approach which utilizes asymptotically design-unbiased estimators. The lidar-augmented survey results from these two sampling strategies were compared to NFI ground estimates for the County. The MD estimate (38.88 ± 1.07 Mg ha− 1, one σ) differs from the NFI ground estimate (37.64 ± 0.94 Mg ha− 1) by + 1.24 Mg ha− 1 (+ 3.3%) and the MA (36.09 ± 1.56 Mg ha− 1) versus NFI difference is − 1.55 Mg ha− 1 (− 4.1%). As expected, as smaller geographical areas are considered, lidar–NFI differences increase, and at the individual cover class level, MD lidar–ground absolute differences average 3.1 Mg ha− 1 for the eight individual classes (8.2% of the County mean). The comparable MA lidar–ground average absolute difference for the individual cover classes is 4.3 Mg ha− 1 (11.4% of the County mean). For all of Hedmark, in general, the MD approach agreed most closely with NFI estimates. The standard errors (SE) generated using the MA approach were, in general, 2–3 times larger than the MD SEs for productive forest classes at the County level, though neither lidar sample was consistently more precise than the NFI ground estimates alone. For Hedmark County, the results indicate that profiling lidar surveys can provide estimates comparable to ground surveys but, at this particular juncture, the MD and MA estimators did not improve the precision of adequately designed (in terms of sample size), probability-based ground samples. However when smaller political units are considered, i.e., as estimates are made on smaller and smaller subunits within Hedmark County, in general, MD-SE < NFI-SE ≪ MA-SE. Of the two designs considered in this study, the MD approach is most amenable to satellite lidar surveys since it does not require that profiling lidar satellite pulses intercept ground plots already established for a regional or national forest inventory. Direct comparisons of lidar and ground SEs are complicated by two factors: (1) the influence of treating both lidar and ground surveys as random samples when, in fact, they’re systematic, and (2) by the two stage/phase structure of the lidar surveys versus the single stage structure of the NFI. These results should be viewed as preliminary, subject to validation using a Monte Carlo simulator.

This article addresses the issue of to what extent forest certification schemes contribute to the enhancement of environmental protection in Swedish forestry. Our assessment is built on the analysi ...

We tested, whether increment cores, routinely sampled and measured as a part of the Finnish National Forest Inventory (NFI), could be used for studying the progress of wood formation during growing season. Widths of incomplete current-year annual rings of the NFI cores were compared to respective measurements from small wood samples (microcores), taken from the same trees. Results of the two methods were rather similar. However, statistically significant differences between the two methods were found for trees, which had been cored near the time of the onset of radial growth. This was caused by cases, where current-year wood formation was observed from NFI cores, while no current-year tracheids were found from the microcores. The accuracy of the microcoring method was also tested by having a subset of the microcores measured by three persons. The results were consistent in the majority of cases. The results thus suggest that the NFI increment data do have potential for analyzing the timing and rate of wood formation during the growing season.

The aim of this study was to estimate the simultaneous recovery of industrial wood and raw material for energy wood from cuttings in Finland during the next 50 years. Two cutting scenarios (maximum and sustainable cuttings) and two climate scenarios (current and changing climate) were analysed to determine their impacts on harvesting potential. The analysis was carried out using sample plot and tree data from the ninth National Forest Inventory and a management-oriented large-scale forestry model (MELA) into which the transfer functions based on physiological modelling were incorporated to describe the impacts of climate change. Depending on the climate scenario, the volume of potential recovery of industrial wood in the maximum cutting scenario during the period 2003–2013 varied from 103 to 105 million m3 a−1, while the amount of potential energy wood raw material was 35 Tg a−1. During the period 2043–2053, in the current climate scenario, the potential recovery of industrial wood was 64 million m3 a−1 and energy wood raw material 22 Tg a−1, and in the changing climate scenario, 85 million m3 a−1 and 29 Tg a−1, respectively. In the sustainable cutting scenario, depending on the climate scenario used, the potential recovery of industrial wood during the period 2003–2013 varied from 74 to 76 million m3 a−1, while the amount of potential energy wood was 25 Tg a−1. During the period 2043–2053, in the current climate scenario, the potential recovery of industrial wood was 80 million m3 a−1 and energy wood raw material 26 Tg a−1, and in the changing climate scenario, 88 million m3 a−1 and 29 Tg a−1, respectively.

National forest inventory programs are tasked to produce timely and accurate estimates for a wide range of forest resource variables for a variety of users and applications. Time, cost, and precision constraints cause these programs to seek technological innovations that contribute to measurement and estimation efficiencies and that facilitate the production and distribution of an increasing array of inventory data, estimates, and derived products. Many of the recent innovations have involved remotely sensed data and related statistical estimation techniques. Current applications of remote sensing in support of national forest inventories are reviewed for three areas: (1) observation or measurement, meaning using remotely sensed data in lieu of field observations or measurements; (2) estimation, meaning calculation of traditional inventory areal estimates such as forest area or volume per unit area; and (3) mapping. Future applications focus on two areas: augmenting field measurements with remotely sensed data obtained from lidar sensors and Internet accessible map-based estimation.

An increased attention drawn to uneven-aged forests, selective cutting and natural regeneration have been seen in Norway over the past years. This means that forest managers face new challenges related to localization and prioritization of forest areas that potentially are suitable for selective cutting. A selective cutting index describing the suitability for selective cuttings based on objectively measured stand attributes related to forest structure and conditions for natural regeneration was therefore, developed. The index comprises four sub-indices describing diameter diversity, growth potential, stability of remaining trees, and conditions for natural regeneration of a stand. The index may be used as a tool in decision-support in practical management planning. It is easy to interpret and can be applied directly because all variables included are available from practical inventories. An evaluation, based on data from the national forest inventory in Norway, showed that the performance of the index agreed with general biological knowledge and previously performed research. The evaluation also showed that the suitability for selective cutting probably is more precisely determined with the index compared to a subjective assessment based on the number of vertical tree layers.

Forests and forestry in Sweden are dominated by conifers. Silviculture using mixed or broadleaved stands is often recommended, but the degree to which broadleaves regenerate naturally needs to be clarified. The Swedish National Forest Inventory is here used for a region-wide study of broadleaved saplings (1.3 m tall to 4.9 cm dbh) regenerated naturally. For 12 species (taxa) in young forests (

A multisource and multiresolution method was developed for estimating large area tree stem volume of growing stock and aboveground biomass of trees. Combined Landsat-TM data and IRS-1C WiFS data, together with field data of National Forest Inventories (NFIs), were applied. Landsat-TM data were used as an intermediate step between the field data and WiFS pixels. A nonparametric k-nearest neighbour (k-nn) estimation method was applied with Landsat-TM data and field plot data from the Swedish National Forest Inventory (SNFI). A nonlinear regression analysis was used in deriving models for volume and biomass as a function of WiFS data. The estimates were evaluated by applying independent estimates from the Finnish Multi-source National Forest Inventory (MS-FNFI): The estimates are derived using field plots from the Finnish National Forest Inventory (FNFI) and Landsat-TM images. Mean volume as estimated from the Finnish multisource data for a study area of 447 000 ha was 84.2 m3 ha−1. This compared with 87.2 m3 ha−1 as derived from the developed method presented in this paper. The corresponding estimates for aboveground tree biomass were 59.5 and 58.3 tons ha−1, respectively.

The information contained in stand compartment databases can be utilized in the context of remote sensing interpretation. The study discusses the gain to be achieved by using digital stand border maps when estimating the growing stock of forest stands from satellite images. The reference sample plot method (RSP) and National Forest Inventory (NFI) sample plots were used to estimate forest stand volumes. The stand edge information on the NFI field plots was used to estimate the areas of the stand margins. The relative improvement in accuracy in the volume estimates when stand delineation data were used was from 62% to 56%, and correspondingly, the use of satellite image segmentation resulted in a 59% RMSE. The variation between similar studies is, nevertheless, due to the data and the validation method. The mean square error estimates were too high for practical forest management planning of small stands.

Only two basic methods exist for obtaining tree height (H) increment data: felled tree measurements and remeasured height or height increment on standing forest trees. Because the former method is expensive (but reliable), and the latter has a large measurement error relative to the actual height increment, it is difficult to obtain good height increment data. The contradictory occurrence of high coefficients of determination for height increment models that are not based on felled-tree samples can only be explained by so-called height increment ‘data’ that is actually predicted from some heuristic function, usually of diameter. Such smoothed ‘data’ are not observable, not measurable, and have much variation removed. Use of smoothed data reduces the apparent problem of height increment modeling to a simplistic problem of using one function to estimate the smoothed predictions from another function. We illustrate this phenomenon with a controlled experiment. Using more than 7500 Norway spruce trees from the Austrian National Forest Inventory with remeasured heights (5 year interval), we built height increment models: (1) based on the difference in observed heights; (2) based on the difference in predicted heights using a heuristic function of diameter. Using the same model and input variables, the coefficient of determination was 3 times higher (0.44 vs. 0.14) using the smoothed increment ‘data’ than with the observed increment data. Furthermore the increment predictions based on the data sets with smoothed increment ‘data’ exhibited a significant overestimation. This demonstrates three things. First, that fit statistics measuring deviations about smoothed height increment data are misleading and strongly biased upward. Second, that the resulting models produce biased predictions that overestimate increment, especially for trees in an intermediate to suppressed social position in the stand. Third, that measurement errors in remeasured heights on standing trees are so large that the underlying height increment signal is nearly hidden (R2 = 0.14).

The long-term effect of butt rot on the diameter increment of Norway spruce, Picea abies Karst., was studied by analysing data collected by the Swedish National Forest Inventory. The 15 698 Norway spruce trees were classified as healthy or infected based on bore cores taken at breast height. Healthy and infected trees, similar in terms of age, diameter at breast height in 1954, increment from 1954 to 1963, and from sites of the same productivity class in the same region were paired. Owing to restrictive criteria used, only 329 pairs were chosen for further analysis. These trees were analysed groupwise as well as pairwise. In the period 1954–1958, trees classified as infected grew faster than the healthy controls; however, within 5 years the relationship had reversed. The degree to which the growth of infected trees was suppressed increased with time, and was 8.9% and 8.2% respectively during the last two 5 year periods. Differences in growth between healthy and infected trees within pairs were highly significant. The diameter growth of the infected trees was reduced by an average of 1.4 mm, or 9.4%, during the last 5 years, corresponding to a 9.6% reduction in the basal area growth.

Fire management activities can greatly benefit from the description of wildland fuel as a fire behaviour predictor. A forest typology developed from the Portuguese National Forest Inventory that combines cover type (the dominant overstorey species) and forest structure defined as a combination of generic stand density (closed or open) and height (low or tall) was translated into fuel models. Fire behaviour simulations that accounted for the fire environment modification induced by stand structure resulted in an objective and quantitative fire hazard classification for 19 forest types. We found that potential fire behaviour is primarily driven by stand structure, rather than by cover type. The results reveal four basic fire hazard groups. Surface fire spread rate is low to moderate, surface fire intensity is low and the crown fire hazard is insignificant in the open and tall forest types, as well as in dense and tall Quercus suber and diverse forests. Moderate surface fire potential and high to very high crowning potential characterizes dense and low woodland of deciduous oaks, Quercus suber and diverse forests, closed and tall Pinus pinaster stands, and open and low eucalypt stands. Open and low types of P. pinaster, Q. suber and Q. rotundifolia have very high or extreme spread and crowning potential, while surface fire intensity ranges from high to very high. In closed and low stands of P. pinaster, eucalypts and acacias, spread rate is moderate, but fireline intensity varies from high to extreme, and the potential for crowning is extreme. Practical applications for the hazard classification are suggested and exemplified.

Portugal with its Mediterranean climate has frequently wildfires during summertime. The knowledge of the composition and the configuration of patches that constitute the forest matrix (Godinho-Ferreira et al., in press) is important to understand the processes occurring in it, including forest wildfire probability. Forests vertical structure depicted from the third revision of the Portuguese National Forest Inventory (NFI) show different layers composed by different kinds of species occupying different percentage of cover, identifying 10 main forest types and 22 structural types (GodinhoFerreira et al., 2005). A first analysis of NFI sampling plots burned between 1999 and 2005 indicates that large and very large patches of maritime pine and eucalyptus forests have the highest rates of burning areas. This presentation is part of a first and coarse approach aiming to develop more accurate methodologies within the project "PHOENIX – Forest conversion in burned areas" (POCI/AGR/58896/2004), particularly “Task 1 – What forest types are less likely to burn?”.