在山區(qū)森林中防止沿成熟度梯度落石外文文獻(xiàn)

Protection against rockfall along a maturity gradient in mountain forests M Fuhr F Bourrier T Cordonnier Irstea UR EMGR 2 rue de la Papeterie BP 76 F 38402 St Martin d H res France Univ Grenoble Alpes F 38402 Grenoble France a r t i c l ei n f o Article history Received 27 February 2015 Received in revised form 5 June 2015 Accepted 8 June 2015 Available online 15 June 2015 Keywords Mature forest Biodiversity conservation Protection forest Rockfall a b s t r a c t When harvesting activities stop forest stands become steadily richer in very large trees and deadwood maturity attributes that are crucial for forest dwelling species On the other hand the maturation process associated with large trees and large gaps between trees has traditionally been thought to be detrimental to the protective function of the forest against gravitational hazards such as rockfall However the fi nd ings of recent studies have contested this belief fi rst because they showed that natural dynamics in aging stands is rather gradual in space and time and second because they highlighted that deadwood may play an important role in protection forests In this study we assessed the protection effi ciency of the forest along a maturity gradient in uneven aged stands using a network of permanent sample plots in the French Alps Plots were selected according to management plans to represent four successive stages along the maturity gradient young stands adult stands just after logging post adult stands that escaped one to two rotations and mature stands We checked that stands gradually matured using the total number of very large trees dbh 77 5 cm the total volume of deadwood and the total volume of degraded large logs We then developed a specifi c module to integrate deadwood into the Rockyfor3D rockfall simulation model and assessed the rockfall protection of the plots Mature stands although not having reached the protection effi ciency of young and dense stands can provide adequate protection against rockfall First because mature stands are rather dense more than 500 stems ha 1 Second because large logs increase the surface roughness of the forest fl oor and act as additional obstacles to the propagation of rocks Consequently mature stands originating from aging irregular stands play a fi tting role in protection forests thereby reconciling biodiversity conservation objectives with protective functions When mature stands originate from aging regular stands dominated by a cohort of very large trees a cautious approach to management will better initiate a progressive irregularization process before promoting patches of mature stands 2015 Elsevier B V All rights reserved 1 Introduction When harvesting activities are stopped or delayed for several decades natural forest dynamics steadily lead to maturity or old growth forest attributes Studies comparing unmanaged versus managed forests see for example Bouget et al 2014 Burrascano et al 2013 Marage and Lemp ri re 2005 Pernot et al 2013 Whitman and Hagan 2007 or focusing on time since logging aban donment gradients Lassauce et al 2012 2013 Pernot et al 2013 Sitzia et al 2012 Vanderkhove et al 2009 highlight the following major structural changes during the maturation process stands become richer in very large trees large snags and large pieces of lying deadwood in various degradation stages increase and tree microhabitats develop Some studies also report that canopy gaps become more frequent Rugani et al 2013 Maturity attributes are crucial for species that depend on forest cover continuity deadwood and very large trees Paillet et al 2010 This is the case for bryophytes Boudreault et al 2002 lichens Boudreault et al 2002 Nascimbene et al 2010 Troy McMullin et al 2010 saproxylic fungi Penttil et al 2004 saproxylic beetles Martikainen et al 2000 Simil et al 2002 Stenbacka et al 2010 Lassauce et al 2011 spiders and ground beetles Isaia et al in press and cavity nesting birds B tler et al 2003 Consequently to improve biodiversity conservation multifunctional forest management aims at restoring maturity attributes using various strategies These strategies consist in http dx doi org 10 1016 j foreco 2015 06 012 0378 1127 2015 Elsevier B V All rights reserved Corresponding author at Irstea UR EMGR 2 rue de la Papeterie BP 76 F 38402 St Martin d H res France E mail address marc fuhr irstea fr M Fuhr Forest Ecology and Management 354 2015 224 231 Contents lists available at ScienceDirect Forest Ecology and Management journal homepage extending rotations maintaining existing maturity attributes at the time of harvest e g retention forestry and in isolated cases man made restoration of maturity attributes Martikainen et al 2000 They also consist in setting aside forest areas from logging operations All around the world areas left for natural dynamics vary in terms of size from small patches of around 0 5 ha patches of mature stands to large forest reserves of more than 1000 ha Schmitt et al 2009 On the other hand when the forest has a protective role against gravitational natural hazards e g avalanches rockfall it is recom mended to short circuit the natural sylvigenetic cycle before the mature stages OFEFP 1996 Gauquelin and Courbaud 2006 Indeed typical non mature forests characterized by dense crown cover 50 high stem density 400 stems ha 1 and an absence of large gaps 15 m width have been shown to adequately pro tect downhill areas from rockfall and avalanches Renaud et al 1994 Cattiau et al 1995 Dorren et al 2004 On the contrary mature forests traditionally associated with large trees and a bru tal senescent phase occasioning large gaps are thought to be detri mentaltotheprotectivefunction Leclercetal 1998 Consequently mountain silvicultural guidelines OFEFP 1996 Gauquelin and Courbaud 2006 promote gap or group selection silvicultures to renew forest stands and maintain patches of very dense young stands along the slopes Cordonnier et al 2008 Rammer et al 2015 However as underlined by Brang et al 2006 natural forest dynamics in aging stands is rather gradual in space and time in most European mountain forests Small scale disturbances due to the death of isolated trees or small clumps of trees are dominant and senescent phases are quite rare Moreover recent studies com paring managed and unmanaged forests report that the total stem number can be signifi cantly higher in unmanaged than in managed forests This is for example the case for beech dominated forests in French mountainous areas Pernot et al 2013 and for Norway spruce dominated forests in the Swiss Alps Krumm et al 2011 In these cases unmanaged stands may provide ade quate protective functions against rockfall and avalanche hazards In addition to date most of the studies that evaluated the pro tective function of a forest did not integrate the role played by deadwoodinprotection effi ciency However recentstudies showed that fallen dead stems increase the surface roughness of the forest fl oor and can thereby impede the release of avalanches and increase the protection function against natural hazards Ammann 2006 This may compensate for a lack of density in the living stand for a certain period estimated around 10 years by Sch nenberger et al 2005 or Bigot 2014 during which the stand is regenerated and gradually recovers its protective ability Moreover ecological engineering has developed methods using deadwood to ensure the protective function of the forest in regen eration patches Frehner and Wasser 2005 Bourrier et al 2012 Finally large pieces of deadwood are known to facilitate seedling establishment and growth in constrained climatic conditions thus improving stand resilience We lack studies integrating the role played by deadwood when assessing the protective function of a forest In this work we aimed to assess the protection effi ciency of the forest against rockfall along a maturity gradient using a network of permanent sample forest plots in the French Alps We fi rst selected plots according to the date of their last logging operations and based on structural indicators we checked that the forest became gradually mature Then we used a process based rockfall simula tion model Rockyfor3D Dorren 2012 to quantify the protective function of the forest integrating deadwood We addressed the following issues 1 Considering living trees is the maturation process detrimental to the protective function 2 Does the integration of deadwood modify the assessment of the protective function in mature and immature stands 3 What are the consequences for forest management are patches of mature stands prohibited in protection forests 2 Materials and methods 2 1 Plot selection and plot data Plots were selected within a network of 80 permanent sample plots covering the northern French Alps from the montane belt to the subalpine belt Plots were established by Irstea and the French Forest Service ONF between 1994 and 2002 to study the growth and the dynamics of managed and unmanaged forests Their areas vary from 0 25 to 1 ha The dominant management sys tem is uneven aged management and consists in single tree or small group selection cutting Forests stands were according to ecological conditions or management history pure or mixed stands dominated by European beech Fagus sylvatica Silver fi r Abies alba and Norway spruce Picea abies Living and dead standing trees of more than 7 5 cm dbh are reg ularly measured in diameter once every 5 years on average and in most of the plots spatially located with x y and z coordinates Since 2010 we progressively added the measurement of deadwood components using a standardized protocol dedicated to French natural reserves Bruchiamacchie 2005 This protocol mixed cir cular subplots with a radius of 20 m for large pieces of deadwood snags logs and stumps more than 30 cm in diameter circular subplots with a radius of 10 m for small pieces of standing dead wood snags and stumps from 7 5 to 30 cm in diameter and line intersect sampling for small logs 7 5 30 cm in diameter We assessed the decomposition stage of deadwood using three cate gories fresh partly degraded and very degraded We also mea sured the length of large logs 30 cm in diameter or greater We established one to two subplots three to six transects for line intersect sampling in each plot depending on the plot size According to management plans and interviews with managers we selected 24 plots in order to distinguish four groups of six plots representing four successive stages along a Time Since Logging Abandonment TSLA gradient Fig 1 We then checked our TSLA gradient regarding the density of stumps of more than 30 cm in diameter and their degradation stages Table 1 The young stage comprises plots representing young stands that have never been logged They originate from the colonization of open areas that started 80 100 years ago or from old large clearcuts Stumps may be absent or present in abun dance but if so very degraded The so called adult stage is made of plots that have just been logged There are numerous stumps in various degradation stages two thirds of the stumps are fresh stumps The so called sub adult stage comprises plots that were not logged during the last 20 years they escaped one to two rota tions There are numerous stumps but most of them are very degraded The mature stage is made up of plots that have not been logged for at least 40 years or have never been logged Stumps are absent or rare and very degraded The dendrometric structures of the four groups showed tangible differences Table1 Youngstandsareverydense 831 95 stems ha 1 and dominated by small and medium sized trees theiraveragebasalareaisquitehighreaching 48 6 m2ha 1 Adultstandshavethelowestdensity 302 31 stems ha 1 and lowest basal area 36 4 m2ha 1 The average densities of the sub adult and mature stands do not reach the average density of the young stands but are quite high 521 45 stems ha 1and 527 81 stems ha 1 respectively The density of the mature stands is however highly variable ranging from 292 to 840 stems ha 1 The average basal areas of the M Fuhr et al Forest Ecology and Management 354 2015 224 231225 sub adult and mature stands 47 6 m2ha 1and 51 5 m2ha 1 respectively are similar to the basal area of the young stands Additional data on tree ages were consistent with the dendro metric structures the oldest trees were 80 100 years old in Y4 and Y5 plots young group 220 260 years old in M1 and M2 plots mature group 2 2 Maturity attributes We evaluated the maturity of each plot using three structural attributeswhoselinkstomaturityarewidelyrecognized Burrascano et al 2013 Heiri et al 2009 Nilsson et al 2002 Whitman and Hagan 2007 The total number of very large trees We defi ne a very large tree as a tree more than 77 5 cm in dbh referring to recent local studies in similar site conditions that showed that trees more than 75 cm in diameter were rare in managed forests but quite abundant in unmanaged ones Grosso 2012 The total volume of deadwood including snags small logs from 7 5 cm to 30 cm in diameter and large logs more than 30 cm in diameter The total volume of degraded large logs 2 3 The rockfall simulation model taking into account the deadwood compartment None of the few rockfall simulation models integrating forest effects in rockfall processes Dorren et al 2007 Volkwein et al 2011 allows one to account for the infl uence of lying deadwood and stumps on rock propagation For the purpose of this study a process based approach was developed to integrate such effects in the simulation code Rockyfor3D The rockfall protection of the stands was assessed using the process based simulation code Rockyfor3D Dorren 2012 It sim ulates the propagation of rocks down a slope through a rasterized digital terrain model by successive sequences of free fl ights through the air rebounds on the slope surface and impact against trees When a rock impacts a tree it loses part of its kinetic energy depending on the tree type broadleaves versus conifers on the vertical and horizontal locations of the impact on the tree and on the rock trajectory before impact Trees with their related diam eters and precise locations x y coordinates are explicitly used as input parameters see Section 2 4 Astheeffectofdeadwoodisnotintegratedintothe Rockyfor3D rockfall simulation model a specifi c module was developed for this study to assess the infl uence of both large logs and large stumps more than 30 cm in diameter in rock Grenoble Fig 1 Location of the plots 226M Fuhr et al Forest Ecology and Management 354 2015 224 231 trajectories The impact of the rocks on logs was modeled as the interaction between a moving sphere with a mean diameter Dr mean the rock and a fi xed cylindrical object with the same diameter as the log s mean diameter Dl mean located in an obli que position in the direction of the steepest slope Both stumps and logs can be considered as obstacles that locally change the slope of the terrain Stumps were modeled with a local slopea set at a nil value a 0 in the cell considered For logs a local slope reductionalwas applied Fig 2 in the cell considered This slope reduction was determined from geometrical consider ations It depends on the mean diameters of the rock Dr mean and the log Dl mean the larger Dl meancompared to Dr mean the larger the slope reduction Energy dissipation for these two impact types was modeled using a classic restitution coeffi cient approach Volkwein et al 2011 Dissipation along the normal resp tangential directions toward the contact surface was quantifi ed using normal resp tan gential restitution coeffi cients called rn or rt defi ned as the ratio between the normal resp tangential direction after and before impact The values of the restitution coeffi cients for fresh deadwood logs and stumps were estimated from rockfall experiments on sites where logs were recently installed Bourrier et al 2012 The restitution coeffi cient for fresh deadwood was set at the val ues measured during the experiments rn 0 8 and rt 1 Decay processes reduce the rigidity and strength of wood and consequently increase energy dissipation during impacts classi cally associated with rnonly Volkwein et al 2011 Decreasing values of the restitution coeffi cient rnwere thus arbitrarily set for increasing decay stages given the lack of quantitative informa tion about the relation between decay rate and energy dissipation Intermediate deadwood and degraded deadwood were associ ated with rn 0 4 and rn 0 2 respectively Preliminary sensitivity analysis of these values of rnwas done to check that the same ten dencies were obtained for different rnand that the model was not highly sensitive to these values Table 1 Characteristics of plots GroupPlotAltitude m asl Composition relative abundance in stem number Total number of stumps N ha 1 Total number of recent stumps N ha 1 Total number of living trees N ha 1 Basal area of living trees m2ha 1 Quadratic mean diameter of living trees cm Picea abiesAbies albaFagus sylvaticaOthers YoungY11150991684115257 725 3 Y2120088317396081629 321 4 Y3120031224161141391235 722 3 Y4155056202324097658 327 6 Y515009640058066 038 1 Y61030336115954855340 830 6 Mean116 5910 8831 9548 627 5 2 5 AdultA11230878529218428032 338 3 A2140099124615728041 343 3 A3105024641111117941934 432 3 A41350495115412036631 333 0 A511001779429621924653 152 4 A6132069221228421925 438 4 Mean203 35140 23302 3131 439 6 3 0 Sub adultSA1118019631351365640634 532 9 SA21150544411754749245 134 2 SA314303971513270038 826 6 SA417009821283243272 846 3 SA5115092351242060439 428 3 SA614306941034749251 536 5 Mean136 1039 5521 4545 634 2 2 8 MatureM1148056321216050860 338 9 M2147054281832062453 933 2 M3120013701430055745 032 1 M4140061668100029231 837 2 M51650991241684065 431 5 M6135086590034451 543 7 Mean12 63 3527 8151 536 1 1 9 Fig 2 Slopeafor large stumps and slope reductionalfor large logs calculated using Dr mean or Dl mean the mean rock or log diameter M Fuhr et al For