Search

Managers need to know how to mitigate rising stream water temperature (WT) due to climate change. This requires identifying the environmental drivers that influence thermal regime and determining the spatial area where interventions are most effective. We hypothesized that (i) extreme thermal events can be influenced by a set of environmental factors that reduce thermal sensitivity and (ii) the role played by those factors varies spatially. To test these hypotheses, we (i) determined which of the environmental variables reported to be the most influential affected WT and (ii) identified the spatial scales over which those environmental variables influenced WT. To this end, the influence of multi-scale environmental variables, namely land cover, topography (channel slope, elevation), hydromorphology (channel sinuosity, water level, watershed area, baseflow index) and shade conditions, was analyzed on the three model variables (day thermal sensitivity, night thermal sensitivity, and non-convective thermal flux) in the model developed by Georges et al. (2021) of the temporal thermal dynamics of daily maximum WT during extreme events. Values were calculated on six spatial scales (the entire upstream catchment and the associated 1 km and 2 km circular buffer, and 50 m wide corridors on each side of the stream with the associated 1 km and 2 km circular buffer). The period considered was 17 extreme days during the summer identified by Georges et al. (2021) based on WT data measured every 10 min for 7 years (2012–2018) at 92 measurement sites. Sites were located evenly throughout the Wallonia (southern Belgium) hydrological network. Results showed that shade, baseflow index (a proxy of the influence of groundwater), water level and watershed area were the most significant variables influencing thermal sensitivity. Since managers with finite financial and human resources can act on only a few environmental variables, we advocate restoring and preserving the vegetation cover that limits solar radiation on the watercourse as a cost-effective solution to reduce thermal sensitivity. Moreover, management at small spatial scale (50 m riparian buffer) should be strategically promoted (for finance and staffing) as our results show that a larger management scale is not more effective in reducing thermal sensitivity to extreme events. [ABSTRACT FROM AUTHOR]

Norway spruce (Picea abies (L.) Karst.) is a major production tree species for the European wood industry. However, it is highly sensitive to bark stripping (BS) by red deer (Cervus elaphus L.), which causes large timber losses. Because the red deer population has increased over the last decades, a better understanding of the underlying causes driving BS is urgently needed. BS outbreaks are multifactorial: winter food shortage, local and regional deer abundance and stand properties (thermal and visual cover, and scarcity of herbaceous layers) have been found to drive BS rates. Although the influence of each of these factors is well studied, there is still no consensus on their relative contributions to BS or what mitigation strategy is best. In this study, we made use of a long-term, large-scale BS inventory (13 years over 2570 km2) to fill these gaps. We modeled winter and summer BS rates in response to the most explanatory variables, selected from a large set of environmental variables. Our results highlight the prevalence of deer abundance combined with high sensitivity of coniferous thickets and pronounced habitat selection by deer for these stands. Our findings show that the control of red deer populations is a key measure for reducing BS in Norway spruce stands. [ABSTRACT FROM AUTHOR]

Remote sensing can be used to collect information related to forest management. Previous studies demonstrated the potential of using multispectral satellite imagery for classifying tree species. However, methods that can map tree species in mixed forest stands on a large scale are lacking. We propose an innovative method for mapping the proportions of tree species using Sentinel-2 imagery. A convolutional neural network was used to quantify the per-pixel basal area proportions of tree species considering the neighbouring environment (spectral–spatial deep learning). A nested U-shaped neural network (UNet++) architecture was implemented. We produced a map of the entire Wallonia Region (southern Belgium). Nine species or groups of species were considered: Spruce genus, Oak genus, Beech , Douglas fir , Pine genus, Poplar genus, Larch genus, Birch genus, and remaining species. The training dataset for the convolutional neural network model was prepared using a map of forest parcels extracted from the public forest administration's geodatabase of Wallonia. The accuracy of the predicted map covering the region was independently assessed using data from the regional forest inventory of Wallonia. A robust assessment method for tree species proportions maps was proposed for assessing the (1) majority species, (2) species composition (presence or absence), and (3) species proportions (proportion values). The achieved value of indicator OA m a j (0.73) shows that our approach can map the majority tree species in mixed and pure forest stands. Indicators MS (0.89), MPS (0.72) and MUS (0.83) support that the model can predict the species composition in most cases in the study area. Spruce genus, Oak genus, Beech , and Douglas fir achieved the best results, with PAs and UAs close to or higher than 0.70. Particularly, high performance was achieved for detecting Oak genus and Beech in low area proportions: PAs and UAs higher than 0.70 from the 0.4 proportion. Predicted proportions had a R a d j 2 of 0.50. The proposed method, which uses spectral–spatial deep learning to map the proportions of tree species, is innovative because it was adapted to the complexity of mixed forests and spatial resolution of current satellite imagery. Additionally, it optimises the use of available forest data in the model conception by considering all pixels from pure stands to highly mixed forest stands. When forest inventories are available in a broad sense, that is, georeferenced areas with the proportions of tree species, this method is highly reproducible and applicable at a large scale, offering potential for use in forest management. • Tree species proportions were predicted using spectral–spatial deep learning. • Proportions were mapped over a large and diversified geographical area. • A CNN was trained using georeferenced areas from forest administration's geodatabase. • Majority class, composition, and proportions were evaluated at the NFI plot level. [ABSTRACT FROM AUTHOR]

Based on an extensive survey of woodland visitors in Wallonia, south Belgium, we examined a wide range of individual-, residential- and destination-level variables for their associations with the distance travelled for woodland leisure on foot, by bicycle and by car. For each transport mode, explanatory bivariate analyses were conducted firstly to identify potential correlates of the distances travelled. Then, cross-classified multilevel analysis was performed to build estimation models for the trip distance. The results showed that, amongst the multilevel variables selected, walking trip distance was only associated with individual trip behaviour, while cycling and car-borne trip distance could also be associated with individual socio-economic profile as well as a large range of residential and destination attributes on land use, land cover and visitor support services. The final estimation model for (i) walking trip distance included trip duration as the only explanatory variable, for (ii) cycling trip distance included variables on trip duration, proportion of woodland area at residence and presence of service facilities at destination, and for (iii) car-borne trip distance included variables on trip duration, visitor's employment status, whether the trip is on weekend or in summer, proportion of woodland area at residence and remoteness of destination from urban area. Despite being simple in form, all multilevel estimation models show a satisfactory explanatory power and a better performance than the ordinary single-level models. Our results add new empirical evidences on the key factors associated with the transport mode-specific travel distance, in particular, for woodland leisure. The cross-classified multilevel analysis used in our study provides new methodological insights for the estimation of individual trip distance, which could benefit future modelling of woodland recreation demand. [ABSTRACT FROM AUTHOR]

This paper is the first to report data on wildlife-vehicle collisions (WVC) in Wallonia, southern Belgium, characterised by one of the densest road network worldwide. With the collaboration of police we identified 3965 accidents involving "free ranging animal" between 2003 and 2011. We observed that these accidents with free ranging animals result in 13% of cases in injuries for the drivers or passengers, and in less than 1% of cases in fatalities (death). 78% of these casualties involve wild animals, among which wild boar take the largest part (39% ). During the covered period we observed an annual increase of WVC of 21%. For wild boar and red deer, this increase was significantly correlated with hunting statistics, used as an index of population density. The temporal analysis demonstrated an increase of WVC during night time with peak of accidents at dusk and dawn. Monthly distribution revealed the role of breeding, dispersal and hunting in shaping temporal patterns of accidents. Spatial analysis, focusing on wild boar, roe deer, red deer and red fox demonstrated clustering of accidents for all these species, until scale between 20 to 70 km. Mapping of accidents via Kernel density analysis permitted us to highlight areas with high risk of WVC risk. Our study suggests that the problem of car accidents due to wildlife is an increasing concern in Wallonia but results on spatial and temporal patterns should help for setting up mitigation measures in the most sensible areas. Moreover we suggest that police data source should be used for nationwide analysis and for comparison between countries. [ABSTRACT FROM AUTHOR]

Bark peeling by red deer ( Cervus elaphus L.) has become a serious issue for productive forests in western Europe. The damage is particularly severe on species such as spruce, as these become vulnerable to fungus attacks that result in considerably depreciated timber. This article presents a monitoring setup for recent bark peeling damage in spruce plantings in Wallonia (southern part of Belgium). This setup implies to collect data annually from a survey involving cluster sampling. It has been employed since 2004 in coniferous stands amounting to 165,000 ha of land, where Norway spruce is the predominant species. The study area was divided into blocks delineated along features preventing deer migrations. A set of indicators was computed either at the whole study area level or at block level. The resulting set of indicators enabled forest managers to follow up debarking intensity in productive forests. Additional analyses were carried out in order to assess the relationship between the social position of trees in the stand and the debarking probability on the one hand, and the relationship between the variation in damage magnitude and seasonality, namely summer versus winter, on the other hand. [ABSTRACT FROM AUTHOR]

Copyright of Biotechnologie, Agronomie, Societe et Environnement is the property of Les Presses Agronomiques de Gembloux and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Abstract: The Walloon forest taken as a whole can be regarded as typical of a rural area, although its proximity to densely populated areas gives it a peri-urban character. It is visited by the local population as well as by tourists (including from neighboring countries). To provide spatial information on the level of Wallonia woodland visitation, a survey was conducted among managers of Forest Service districts (also called “cantonnements”). The aim was to map the woodlands to show spatial patterns of visitation levels, and analyze them qualitatively to determine the influence factors. This map identifies regional hubs of recreation woodlands and, conversely, areas where the recreational function is much less important. The level of visitation was scaled in four levels ranging from low to very high. The mapping was supplemented with a statistical analysis of data collected from the forest managers and also from different GIS-layers (slope, hydrology, land use, etc.). An equation using the level of visitation as a dependent variable was fitted to a set of characteristics of the woodland with an ordered Logit model. The results show that type of ownership, type of forest, and recreational facilities significantly influence the level of woodland visitation. They also show that woodland visitors prefer any type of forest to mainly coniferous woodland. These results and the analysis of the map are particularly useful for developing forest policy and tourism as well as managing the forest. [Copyright &y& Elsevier]

An individual distance-independent girth increment model for pure stands of Douglas-fir (Pseudotsuga menziesii (MIRB.) FRANCO), comprising two equations, is presented. The data used to fit the model were collected from 1007 trees in 42 plots installed in regularly stocked and even-aged stands located in Wallonia (southern Belgium). Both equations predict girth increment from individual girth, dominant height, basal area per hectare, stand mean girth and variables linked to site fertility. These last variables are the site index H50 in the first equation, and a combination of mean annual rainfall and altitude in the second. The coefficient of determination ranges from 0.434 to 0.481 and the root mean square error ranges from 0.7857 to 0.8194cmyear−1. Estimated increments of 224 Douglas-fir trees in 12 different and independent stands were used to validate the model, which is expected to provide reliable predictions for most of the pure Douglas-fir stands located in the study area. [Copyright &y& Elsevier]

The leisure use of forests is a fast growing sector in European tourism. The difficulties in monitoring real usage pattern have hindered communications in tourism, conservation and public health in forested areas. In this study, we present an agent-based model to approximate forest visit patterns at the regional level, based on the visitor, residential and forest attributes. The model was then adapted locally for Wallonia, south Belgium, to predict the daily distribution of forest visits and visitor flows. The results suggest high visit rates in urban forests. In highly attractive forests that are distant from major cities, the visit rates may greatly respond to the visitor-level and environmental changes. Future empirical investigations are encouraged to build a cross-contextual understanding on visitors' decision-making mechanisms and to identify how these mechanisms may be influenced by environmental factors operating at different spatio-temporal levels. [ABSTRACT FROM AUTHOR]

In forestry, top height is a common parameter used as indicator of the stand development stage. It can be used to estimate the potential production of monospecific even-aged stands. However, accurate field estimation of top height is time-consuming and expensive. Since the last two decades, LiDAR has proven to be very useful in estimating forest heights. In Wallonia, a low density LiDAR dataset (0.8 points /m² on ground-level) is available for the whole territory. This paper outlines a tool, based on a predictive model of top height from airborne LiDAR data, to help forest management decision-making. The estimations provided by the model are associated with top height growth models to update top height over time and then estimate Site Index. The model has been validated for Norway spruce (Picea abies (L.) H. Karst.) and Douglas-fir (Pseudotsuga menziesii (Mirb.) Franco) stands in the entire Wallonia area (Belgium). In order to facilitate access to these models, the process has been implemented as a plugin of the open source GIS software QGIS. Free and user-friendly, it is aimed to be used by forest managers and scientists. [ABSTRACT FROM AUTHOR]