This paper explores the potential of unmanned aerial system (UAS) optical aerial imagery to characterize grain roughness and size distribution in a braided, gravel-bed river (Vénéon River, French ...Alps). With this aim in view, a Wolman field campaign (19 samples) and five UAS surveys were conducted over the Vénéon braided channel during summer 2015. The UAS consisted of a small quadcopter carrying a GoPro camera. Structure-from-Motion (SfM) photogrammetry was used to extract dense and accurate three-dimensional point clouds. Roughness descriptors (roughness heights, standard deviation of elevation) were computed from the SfM point clouds and were correlated with the median grain size of the Wolman samples. A strong relationship was found between UAS-SfM-derived grain roughness and Wolman grain size. The procedure employed has potential for the rapid and continuous characterization of grain size distribution in exposed bars of gravel-bed rivers. The workflow described in this paper has been successfully used to produce spatially continuous grain size information on exposed gravel bars and to explore textural changes following flow events.
Understanding the spatial dynamics of invasive alien plants is a growing concern for many scientists and land managers hoping to effectively tackle invasions or mitigate their impacts. Consequently, ...there is an urgent need for the development of efficient tools for large scale mapping of invasive plant populations and the monitoring of colonization fronts. Remote sensing using very high resolution satellite and Unmanned Aerial Vehicle (UAV) imagery is increasingly considered for such purposes. Here, we assessed the potential of several single- and multi-date indices derived from satellite and UAV imagery (i.e., UAV-generated Canopy Height Models—CHMs; and Bi-Temporal Band Ratios—BTBRs) for the detection and mapping of the highly problematic Asian knotweeds (Fallopia japonica; Fallopia × bohemica) in two different landscapes (i.e., open vs. highly heterogeneous areas). The idea was to develop a simple classification procedure using the Random Forest classifier in eCognition, usable in various contexts and requiring little training to be used by non-experts. We also rationalized errors of omission by applying simple “buffer” boundaries around knotweed predictions to know if heterogeneity across multi-date images could lead to unfairly harsh accuracy assessment and, therefore, ill-advised decisions. Although our “crisp” satellite results were rather average, our UAV classifications achieved high detection accuracies. Multi-date spectral indices and CHMs consistently improved classification results of both datasets. To the best of our knowledge, it was the first time that UAV-generated CHMs were used to map invasive plants and their use substantially facilitated knotweed detection in heterogeneous vegetation contexts. Additionally, the “buffer” boundary results showed detection rates often exceeding 90–95% for both satellite and UAV images, suggesting that classical accuracy assessments were overly conservative. Considering these results, it seems that knotweed can be satisfactorily mapped and monitored via remote sensing with moderate time and money investment but that the choice of the most appropriate method will depend on the landscape context and the spatial scale of the invaded area.
Recurrent fires are more frequent in the Mediterranean basin due to global climate change. The effects of fire recurrence (up to more than three recurrent fires with an interval since the last fire ...of more or less than 15 years) on the structure of fuel beds were studied at stand level in southeastern France. Three types of vegetation that are among the most common in the study area (pure pine stands, oak-pine mixed stands and shrublands) were studied to determine the role of fire recurrence on biodiversity. Fuel bed description was carried out on aboveground vegetation at understorey and overstorey levels, and on litter in 39 homogeneous study sites. The structure of the understorey was categorized following a gradient of fragmentation including: (i) isolated shrubs, (ii) vegetation patches made up of a single species and (iii) homogeneous vegetation layers including different species of the same height.
The decrease of time interval since the last fire caused an opening of the canopy, except for the lowest vegetation stratum (grasses and low shrubs) that did not vary significantly. There was no significant change in horizontal connectivity between fuel beds (% of layer cover) as a function of the fire regime or the type of vegetation. Conversely, the vertical connectivity (% of tree cover, height of the trees and of the layer) was affected by the fire regime, especially in pure pine stands. Fire regime also influenced the litter layer and litter depth decreased with an increase in the opening of the stand that could be due to fire. These results show that fire risk is greater in the oldest stands (long interval since the last fire).
Flow‐like landslides in clay slopes pose major threats to people and infrastructure, which has led to numerous studies in recent decades. However, the mechanisms leading to the solid–fluid transition ...in clay are still poorly understood, despite numerous studies on its rheological evolution. The aim of this study is to contribute to quantify the degradation of clay at the surface of the Harmalière landslide (French Alps) from the analysis of a series of three unmanned aerial vehicle (UAV) acquisitions. Two approaches were combined to process the acquired optical images. First, image classification was performed applying object‐based image analysis (OBIA) to the red, green and blue (RGB) and surface roughness layers. Second, deeper analysis of the surface roughness allows to describe the morphology evolution and to interpret the degradation scheme from undegraded clay to degraded clay.
The study shows that the applied methodology is appropriate to perform a thorough analysis of the material degradation pattern on the surface of a landslide. The temporal analysis shows an average degradation rate leading to the complete degradation of a block in about 2 years. Meanwhile, a spatial analysis shows that non‐degraded clays degrade faster in the lower part of the study area, reactivated 30 years ago, than in the upper part, reactivated only a few years ago. In addition, roughness analyses enabled to highlight the evolution of the morphology during the degradation process of the clay blocks, from angular blocks to mounds.
The study uses unmanned aerial vehicle (UAV) photogrammetry to perform an in‐depth analysis of the degradation pattern leading to the solid–fluid transition in a clay landslide. The method consists of two steps. First, a classification of the surface materials was used to calculate an average degradation rate and analyse it in time and space. Second, a roughness analysis was used to highlight the evolution of the morphology during the degradation process of the clay blocks, from angular blocks to mounds.
Characterizing time intervals between successive fires in the recent history is of main interest for fire hazard prevention and sustainable environmental management as it indicates what the typical ...fire return interval for each type of ecosystem is. We tested the extent to which fire return intervals (FRIs) depend on fuel type and age, and we compared FRI values between two fire-prone areas of south-eastern France (Provence). These areas had similar weather and roughly similar fuel types but fuels occurred in patches with different sizes and shapes in the landscape. We built a fire database (1960–2010) and we fitted Weibull distributions of FRI in order to compute the probability density function and the hazard of burning. Our results indicate maximal probability of burning again for shrublands (garrigues and maquis), and minimal values for mixed broadleaf-conifer forests and broadleaved forests. Most fuel types of Provence showed no effect of fuel age on the probability of burning again. Only the unmanaged maquis showed a linear increase of fire hazard in time due to a rapid postfire fuel build up. Rather long fire-free intervals and low age-dependency for most forest fuels of Provence suggest that reducing their biomass may not be sufficient to reduce fire risk. In contrast, the flammable shrublands have rather short fire return intervals and represent a high fire hazard for the whole study area. The two areas had statistically significant difference of fire return intervals for a same fuel type (e.g. 18–22 years for shrublands, 20–24 years for pine forests, and 24–27 years for oak forests). This suggested that size, shape and connectivity of fuels play a major role in the probability of burning again and should be taken into account for fire management. The present policy of fire prevention puts efforts into public information and prevention, and preferential management of fuels at risk in the vicinity of roads and wildland–urban interfaces where fires occur preferentially. However, fire suppression may also take advantage of favouring low-flammable fuels with low age-dependency on strategic places in the landscape.
► A first georeferenced fire database (1960–2010) is proposed for southeastern France. ► Shrublands have high fire hazard and high probability of burning again. ► Mixed oak forests limit the probability of new fires. ► Size, shape and connectivity of land covers modify the probability of reburning. ► This study supports a policy of preferential management of fuels at risk.
Display omitted
•We modelled rockfall events on 3886 different forests located in all the French Alps.•We proposed two indicators to assess reductions of rockfall frequency and intensity.•We defined ...one indicator to evaluate the overall rockfall protection of each forest.•The indicators are easily and accurately predicted with three forest characteristics.•This approach has direct applications in forest management and rockfall assessment.
Natural hazards are frequent in mountain areas where they regularly cause casualties and damages to human infrastructures. Mountain forests contribute in mitigating these hazards, in particular rockfalls. Assessing the protective effect of a forest against rockfall is a difficult task for both forest managers and rockfall experts. Accurate and simple tools are therefore required to efficiently evaluate the level of protection that results from the presence of forest.
This study defines three novel indicators to quantify the protective effect of forests against rockfalls, regarding (1) the reduction of the frequency of rockfalls, (2) the reduction of their maximum intensity, and (3) the combination of the reduction of the frequency and the energy of the rocks. The first two indicators are relevant for rockfall experts whereas the third is mostly interesting for foresters as it summarizes the protective effect of forest. The Rockyfor3D model was adapted and used to simulate rockfalls propagation on 3886 different forest stands located in all the French Alps. The results of the simulations were used to calculate the three indicators for each forest stand. Finally, the relations between the forest structures and compositions and the indicators values were investigated.
Our principal result shows that only three forest characteristics are required to accurately predict the indicators and evaluate the protective level of a forest against rockfall. The two first variables correspond to the basal area and the mean diameter at breast height (DBH) of the forest stand which are two parameters commonly used by forest managers. The third characteristic is the length of forest in the maximum slope direction which can be computed with a geographic information system (GIS). The method proposed in this study is easily reproducible and is suitable to evaluate the protective effect of European mountain forests at different scales. At local scale, the proposed indicators can enrich rockfall studies in which forests are usually set aside to simplify the evaluation. Moreover, the indicators may find direct applications with foresters by allowing them to identify the protective level of their forest and consequently to adapt their management. Finally, the indicators are convenient to perform spatial analysis and produce maps of the protective effect of mountain forests that could find many applications in land settlement or evaluation of ecosystem services.
•We investigated the protective effect of 3886 French Alpine forests against rockfalls.•The length of forested slope is a predominant variable for rockfalls assessment.•Half of the French Alpine ...forests are very effective in reducing rockfall hazard.•Mix tree species and promote shade-tolerant species is positive for the protection.•High biodiversity and structural heterogeneity in forests improve rockfall protection.
The role of forests in the mitigation of natural hazards has been repeatedly demonstrated. The protective effect of mountain forests against rockfalls has especially been pointed out because it can constitute a natural and cost-effective protection measure in many situations. However, this particular ecosystem service may substantially differ according to the structure and the composition of the forest. Until now, the rockfall protection capability has always been studied at a local scale with only few forest types. Moreover, the comparison of the protective effect of the different forest types studied remains difficult because different methods and indicators were used. For the same reasons, it is not possible to draw conclusions about the influence of biological and structural diversities on the protection capabilities of forests from former works.
The aims of this study were (1) to quantitatively assess the protective effect of forests at the French Alps scale and build a classification based on the protection capability, (2) to compare the protective effect of the different forest types present in the French Alps and (3) to analyze the relations between the protective effect and the forest diversity in terms of stand structure and tree composition. For this purpose, the model Rockyfor3D was used to simulate the propagation of rocks on 3886 different forest plots spread over the whole French Alps. Quantitative indicators characterizing the protective effect of each forest plot were then calculated from the simulation results and used to perform the different analyses.
Our results emphasized the importance of taking into account the length of forest in the maximum slope direction for an accurate assessment of the protective effect. Thus, the minimum length of forest to get a reduction of 99% of the rockfall hazard was chosen as indicator to compare protective effect between forests. Using this indicator, half of the French Alpine forests presented a high level of protection after a short forested slope (190m). A decreasing gradient in the protection capabilities was observed from forest types dominated by broadleaved species to those dominated by conifer species. Moreover, considering an equivalent proportion of conifers, stands dominated by shade-tolerant tree species showed better ability to reduce rockfall hazard. Finally, our study highlighted that a high biodiversity and a structural heterogeneity within the forest have a positive effect on the reduction of rockfalls hazard.
•We established the first georeferenced forest fire database for the New-Caledonia.•Ignitions are mostly driven by human activity and fire weather in New-Caledonia.•Bushfires propagate towards the ...dry and humid forests of high conservation value.•Some areas could become ‘fire traps’ where fire cannot be easily extirpated.•We make recommendations for a sustainable forest management and fire policy.
New Caledonia (NC) is a biodiversity hotspot sheltering terrestrial ecosystems of high ecological and conservation value including tropical dry forests, rainforests, and maquis. However, uncontrolled bushfires threaten this exceptional biodiversity. A science-based fire management policy could reduce the impact of unwanted fires and help facing climate change. However, to date, data on the location, extent, causal factors and spatial patterns of fires had not been collected. We compiled a 13-year-long (1999–2011) spatially-explicit fire database for NC using MODIS and Landsat data. Using boosted regression trees we disentangled the role of anthropogenic factors, physiography, weather and vegetation on fire activity. We also characterized the location of fires and the vegetation composition at the fire edges, in order to determine which ecosystems were especially vulnerable. Fire size distribution was typically asymmetric with many small fires (<10ha) and very few large fires (>500ha). Ignitions were preferentially located close to villages, cities or roads, at low elevation and linked to high values of fire weather index. Fires were larger at the end of the dry season and during El Niño events. Most fires were bushfires burning in savannas, thickets and maquis, while rainforests were rather ‘avoided’ by fire. However, bushfires generally propagated towards forests of high-conservation value, thus increasing the potential for forest edge erosion. As savanna-forest and maquis-forest mosaics are dominant in the landscape, we discuss the extent to which NC could become a ‘fire trap’ where fire cannot be easily extirpated. Based on our spatially-explicit information on fire activity, we make recommendations for a sustainable forest and fire management policy which would balance the traditional use of fire and the conservation of the most valuable ecosystems. In particular, it may help by reducing the damages of large and destructive bushfires ignited during drought peaks.
PDF
