Windthrow is all too often looked at as an exceptional, catastrophic phenomenon rather than a recurrent natural disturbance that falls within the spectrum of chronic and acute effects of wind on forests, and that drives ecosystem patterns and processes. This paper provides an integrative overview of the nature, contributing factors and impacts of wind-caused disturbance in forests, including its effects on trees, stands, landscapes and soils. Windthrow is examined through the integrating concepts of: the capacity of trees for acclimative growth, the limitation of acclimative growth under inter-tree competition, the recurrent nature of severe weather, how terrain and soil conditions affect local stand vulnerability and the effect of recurrent windthrow on stand dynamics and soils. Windthrow management should take place within a framework of general risk management, with evaluation of the likelihood, severity and potential impacts of wind damage considered – with reference to the broad and specific aims of management. There is much to be gained from interdisciplinary communication about the nature and consequences of recurrent wind damage. There are opportunities for climatologists, engineers, ecologists, geomorphologists and others to develop integrative process models at the tree, stand and landscape scales that will improve our collective understanding, and inform management decision-making.
This paper provides an integrative overview of the nature, contributing factors and impacts of wind-caused disturbance in forests, including its effects on trees, stands, landscapes and soils. There is a great deal of literature on each of these topics, including reviews on wind effects on trees,1,2 on factors contributing to stand-level damage and recovery,3 on the mechanics of windthrow,4–6 on windthrow modeling7 and on the ecological impacts of windthrow in boreal, temperate and tropical forests.8–11 Schaetzl12 and Šamonil et al.13 review windthrow effects on soils, and Bouget and Duelli14 review windthrow effects on insect communities.
Windthrow is all too often looked at as an exceptional, catastrophic phenomenon rather than a recurrent driver of ecosystem patterns and processes that falls within the spectrum of chronic and acute effects of wind on forests. Windthrow disturbance patterns are complex and it is often difficult to see order in the aftermath of an individual storm. While individual windthrow events have immediate impacts, the long-term and cumulative impacts of recurrent windthrow are often overlooked by scientists and managers. Furthermore, the study of windthrow and the management of windthrow-prone landscapes encompasses many scientific and professional disciplines. One of my goals in this paper is to bridge some interdisciplinary gaps. A second goal is to reduce the apparent complexity of windthrow by presenting a set of integrating concepts, knowledge of which helps in the interpretation of the apparent inconsistencies in windthrow patterns within and between events. In this paper, the focus is on windthrow as a recurrent natural disturbance process in forests. Windthrow is examined through the integrating concepts of: the capacity of trees for acclimative growth, the limitation of acclimative growth under inter-tree competition, the recurrent nature of severe weather, how terrain and soil conditions affect local stand vulnerability and the effect of recurrent windthrow on stand dynamics and soils. The interactions between tree, weather and site conditions are framed in terms of diagnostic considerations that should promote greater clarity in predicting and interpreting windthrow patterns and consequences. The implications of recurrent windthrow for stand and landscape management are briefly discussed.
Natural disturbance definition
Grime15 defines disturbance as external factors that limit plant biomass by causing partial or total destruction. Pickett and White16 place this in the context of ecosystems by defining disturbance as discrete events that disrupt ecosystem, community or population structure and change resource or substrate availability, or the physical environment. Forman and Godron17 consider these disruptions as departures from normal patterns or ecosystem functioning. In establishing what is normal, it is useful to think of the range of natural variability, and the temporal and spatial scales at which this variability occurs.18 There is typically an inverse relationship between the disturbance intensity (energy released by the physical process of disturbance, or in the case of wind, the peak wind speeds during the storm) and periodicity (interval between disturbances), and a positive relationship between the disturbance intensity and severity (amount of mortality among tree and other plant populations) in the affected area.10
Chronic effects of wind on trees: acclimative responses
The effects of wind on trees range from chronic to acute, incorporating a spectrum of sub-lethal to lethal effects (see reviews by Grace1 and Telewski2,19). Airflow across the surface of leaves facilitates transpiration and the exchange of carbon dioxide and oxygen between the leaf and the atmosphere. Short-duration displacements of branches and leaders by wind leads to thigmomorphogenetic responses such as formation of flexure wood, increased secondary thickening of stems and structural roots, and reduced shoot extension.2,20 These growth responses produce shorter, thicker and, therefore less slender, trunks and branches which better resist deflection,5 and improve root anchorage.21,22
Trees are self-designing structures that experience increasing loads as they grow larger. They also compete for light and soil resources with neighbouring plants, and must allocate photosynthates efficiently.23 This optimization leads to the conformation of trees with some general design principles.24 Bending of a tree under its own weight, or because of lateral loading by wind, produces stress in the outer fibres and the cambium. The portion of photosynthate allocated to structural increment would be most efficiently used if the cambium produced new wood in a manner that equalizes the distribution of stress along the outer surface of the stem, and this is known as the ‘constant’ or ‘uniform stress hypothesis’.25 While there is observational and experimental evidence in support of this hypothesis, there is ongoing debate on the universality of the theory. Niklas and Spatz26 note that while it is logical to assume that natural selection has lead to biomass partitioning patterns that achieve a balance between bending and counter-resisting moments, leaf, stem and root biomass partitioning patterns must also accommodate the simultaneous performance of other functions.
In combination, the growth responses in roots, stems and branches to deflection and bending enable tree crowns to maintain what Sinnott27 refers to as a genetically determined equilibrium shape – where wind exposure is not excessive. Under higher, directional, sustained winds, the spectrum of responses extends to the formation of reaction wood, asymmetric growth of stems and root systems and flagged (windswept) crowns.22,28,29 In examining the literature concerning windswept crowns, Telewski19 distinguishes between redirection of plant form under constant unidirectional wind (a biophysical response) and mechanoperception coupled with thigmotropism (a growth response). He concludes that redirection of branches is the more probable mechanism for the windswept crown form.
The terms ‘acclimative growth’ and ‘adaptive growth’ are used by windthrow researchers in referring to the process by which acclimation to wind and other mechanical loading develops in trees, and the result of this process.30,31 Acclimation to wind loads includes thigmotropism, deformation and sub-lethal mechanical impacts (abrasion, failure) and their collective effects on the form and biomechanical properties of roots, stems and branches. Since deformation and loss of tissues via impact do not constitute growth, it would be more accurate to use the broader term ‘acclimative responses’ to describe the full spectrum of effects.
Acute effects of wind on trees and stands
If wind loads exceed the resistance of stem or root/soil systems, trees break (stem snap, windbreak) or uproot. These modes of failure are often lethal, but trees may survive, particularly where they are capable of resprouting.32 In either case, this disturbance affects stand and soil conditions. Uprooting and stem breakage are incorporated in the term ‘windthrow’ or more colloquially ‘blowdown’ and ‘windfall’.4,6 In the case of hurricanes, there are gradients of wind speed across the hurricane track, with resulting gradients of damage to trees and stands.33 These gradients provide insights into the hierarchy of tree and stand-level responses to increasing acute wind speeds. Stem or root system failure may be preceded during extreme wind events by partial defoliation or branch failure.11 The loss of foliage and branch shedding may be an adaptive strategy in storm-exposed landscapes, since it reduces drag and the potential for stem breakage and uprooting.34,35 Partial or complete failure of branches also occurs as crowns of neighbouring trees collide during more routine wind events. Periodic pruning of branches or branch tips via collision can lead to gaps between adjacent tree crowns, termed crown shyness, and has implications for tree growth and understory development.29,36,37 At higher wind speeds, as trees within stands begin to uproot or experience stem breakage, they may knock off branches, break or uproot adjacent trees as they fall through the canopy. The extreme winds and flying debris during hurricanes and tornadoes can strip the bark from stems.11
Stand-level damage ranges from the creation of small canopy gaps via failure of individual or small groups of trees, through intermediate levels of failure of overstory trees, to complete and extensive failure of the overstory canopy. Damage patterns within stands can be variable or uniform. The propagation of damage through a stand during a high wind event, and the uniformity of this damage, depends in part on the uniformity of trees within the stand, along with uniformity in terrain and soil conditions that affect wind exposure and root anchorage.6,38
Windthrow mechanics, stand-level instability
Windthrow results when the wind-induced drag force on the tree crown, multiplied over the lever-arm of the stem, results in a turning moment that exceeds the bending resistance of the stem, or the root anchorage. Tree design appears to balance core rigidity with peripheral flexibility.39 The bending and reconfiguration of branches, and consequently tree crowns, during high winds streamlines foliage and reduces the frontal area. Over the range of wind speeds at which this reconfiguration occurs, drag on tree crowns is directly proportional to wind speed rather than to the square of wind speed, as is the case for solid bodies.35,40 The resistance of the stem to failure is proportional to the cube of the stem radius, so small increases in radial growth add substantially to stem strength.5,41 Byrne and Mitchell42 found that the critical moments for trees that fail via stem breakage were similar to the critical moments for trees that failed via uprooting for trees of similar size and form. Root anchorage is a function of root architecture and soil properties, and is empirically very tightly correlated with stem mass.43 The effects of acclimative growth on tree stability in Sitka spruce plantations in Britain were examined by Nicoll et al.30 who found that anchorage increased as wind exposure increased. The stability of trees is affected by precipitation and temperature. Soil moisture affects root anchorage, and uprooting often occurs when high winds coincide with saturated soils.44 Peltola et al.45 found that all of their winched trees failed via stem breakage when soils were frozen, and with higher critical moments than for similar sized trees winched during non-frozen conditions.
Above-canopy wind speeds attenuate rapidly within dense stand canopies.6 While open-grown trees are exposed to and acclimate to routine wind events, stand-grown trees are partially sheltered by neighbouring trees. Stand grown trees are also in competition with these neighbours for growing space, and photosynthate is allocated preferentially to shoot growth and fine root production, and in diminishing priority to storage, radial increment and protective chemicals.23,46 Shade-grown trees allocate fewer resources to radial increment than trees grown in full sunlight,47 even when mechanical loads are artificially increased.48 In stands grown at high densities, stems become increasingly slender, crown centre of gravities shift higher and higher and individual trees become increasingly unstable.41,47,49 Crop planning tools such as stand growth models or stand density management diagrams enable projection of height and diameter growth, slenderness, and in some cases crown length. These can be coupled with windthrow models to enable prediction of stand stability for different planting density or thinning regimes.7 At some critical stand height, there is the potential for uniform stands to become mechanically unstable. This likely reflects both the increased leverage for tall, high crowned trees, but also the potential for trees to domino fall during damage events rather than simply falling into and being supported by their neighbours, as happens in short, dense stands. The mechanics of this are demonstrated in mechanistic windthrow models,7,50 albeit without representation of tree collisions. The phenomenon of critical stand height has been observed in planted Sitka spruce (Picea sitchensis (Bong.) Carr.) in the UK,51 and can be observed in natural stands of black spruce (Picea mariana (Mill.) BSP) in eastern Canada,52 mountain beech (Nothofagus solandri var cliffortioides (Hook.) f.) in New Zealand,53 and likely underlies the recurrent stand-replacing windthrow observed in even-aged stands of southern beech (Nothofagus pumilio Poeppig and Endl.) in Tierra del Fuego54 and western hemlock (Tsuga heterophylla (Raf.) Sarg.) and associated species in coastal Alaska.55,56
Factors contributing to windthrow
Recurrent extreme winds
The severity of damage to forests during extreme weather events depends on the duration of the event, the maximum sustained wind speed and precipitation immediately prior to and during the event – in other words, on the intensity of the event.3 The role of wind as a driver of ecosystem processes also depends on how frequently intense weather events recur within a region. Extreme winds are traditionally ranked according to increasing intensity as gales (62–88 km h−1), storms (89–117 km h−1) and hurricanes (≥118 km h−1) according to the Beaufort scale.57 Hurricanes/typhoons and tornadoes are associated with specific synoptic conditions and are ranked according to the Saffir–Simpson and Fujita scales, respectively.58,59 Gale-force and storm-force winds can be produced by a variety of synoptic conditions. In mid-latitude temperate zones, extra-tropical cyclones are a major source of recurrent gale and storm force winds, and consequently forest disturbance. Extra-tropical cyclones develop over the Pacific and Atlantic oceans in response to temperature gradients between tropical and polar climates. These are very large-scale systems that are most intense during the winter months, and are often accompanied by high rainfall. These systems rotate counter-clockwise in the northern hemisphere, and clockwise in the southern hemisphere. They are embedded in mid-latitude airflow and typically track from west to east, in both hemispheres. These systems routinely produce winds up to 120 km h−1 as they make landfall and cause a substantial damage to forests over wide areas in coastal regions in the mid-latitudes. These systems occasionally bring winds in excess of 120 km h−1 well into continental land masses, as occurred in Europe during storms Vivian in 1990 and Lothar in 1999.60 Tropical cyclones develop in response to temperature differentials between the sea-surface and the atmosphere, and are most intense during the late summer and fall months. These relatively compact systems are accompanied by high rainfall and by definition produce sustained winds in excess of 117 km h−1. Tropical cyclones deteriorate rapidly over land and cause the greatest damage in coastal regions, but can bring storm force winds well inland. These tropical systems are initially embedded in tropical airflow, tracking from east to west in both hemispheres, but often veer into the mid-latitudes.58 Convective storms develop over land in tropical and temperate regions on days with high surface heating. Localized storm cells produce vertical down drafts and track with regional airflow.59 Derechos, or straight-line windstorms can develop in association with fast moving bands of intense thunderstorms and can cause bands of damage that extend for long distances.61,62 Under specific atmospheric conditions, convective supercells produce tornadoes, with very localized winds that can exceed 300 km h−1.59,63 Localized temperature and pressure gradients that develop daily or seasonally in coastal, mountainous or glaciated regions under specific atmospheric conditions can result in high winds, such as bora winds. These often recur with sufficient regularity to attract local names,64 and may produce chronic or acute effects on trees and stands. Regional recurrence of damaging storms can be reconstructed from weather records and forest damage reports,65 stand history studies and dendrochronology,66 and analysis of remote-sensing images.56,67
Diagnostic considerations – In evaluating regional climate, consider the return period of high wind events, their coincidence with high rainfall or wet soils, dominant wind directions and the differential between severe weather conditions and routine wind conditions.
Movement of large-scale air masses generated by extra-tropical and tropical cyclones is modified by regional and local topography. Wind speeds generally increase with altitude, and winds can increase rapidly in higher terrain, particularly in coastal regions. Winds funnel through and over terrain constrictions, leading to areas of higher or lower mean flow and turbulence. Ridgelines, valley gaps and shoulders where large valleys change direction experience higher than average wind speeds.68 Numerical airflow modelling is becoming increasingly sophisticated at capturing the effects of terrain, but is computationally intensive and depends on accurate characterization of surface and atmospheric conditions.69,70 In the absence of airflow models, topographic indices such as topex, distance-limited topex68,71 and expose72 have been found to be useful predictors of wind exposure.
The differing topographic exposure of terrain units can lead to distinctive local disturbance regimes within a landscape.56,72 Trees and stands in windward locations acclimate via crown shaping and height growth suppression,19,28 and local tree-lines may form.53,73 Conversely, shelter from routine winds may lead to increased vulnerability of lee-slope stands to storm winds.73
Diagnostic considerations – In evaluating wind exposure in complex terrain, consider whether a given terrain unit is exposed or sheltered from routine winds and how this will affect the acclimation of trees and stands to wind. Then consider how exposure may change during extreme wind conditions. Finally, consider the effects of slope angle and ground concavity on drainage, since this will affect soil moisture during extreme weather events.
Soil conditions affect site fertility and the capacity of trees to develop effective anchorage. Windthrow is often observed within stands in localized areas of shallow or poorly drained soil,75 and in their analysis of UK tree-winching experiments, Nicoll et al.43 found anchorage to be stronger on deeper soils, for trees of equivalent mass. However, sites with deep soil are often found to be more windthrow-prone than sites with shallow soil.54,55,67,75 This apparent inconsistency is likely due to the effect of soil fertility on tree growth and inter-tree competition, and the limiting effects of inter-tree competition and mutual shelter on acclimative growth.23,47,48 Mitchell et al.75 and Bouchard et al.67 observed that windthrow occurred more frequently on fertile sites than on low-fertility sites, and in dense stands rather than open-grown stands. Deep soils can support tall, dense stands. Sites with fertile but shallow soils, such as seepage sites, would be expected to produce the least stable stands, since height growth and canopy density is promoted, without concomitant root anchorage.
Recent nitrogen fertilization is associated with higher windthrow rates in Scandinavia.77–79 The mechanism for this is not clear. Fertilization can affect stem allometry and above-ground biomass distribution,80,81 however it may not affect coarse root to above-ground biomass ratios.82,83
Diagnostic considerations – In evaluating soil effects on stand stability, consider – is the soil fertile enough to support a closed canopied stand in which trees grow tall and compete for growing space? Does the soil restrict anchorage, or become saturated during severe weather events?
Given the variation in climate, soil, terrain and stand conditions within and between forested regions, it is not surprising that observations on the relationship between tree and stand factors and vulnerability to wind damage vary between studies. Everham and Brokaw3 reviewed 119 reports on wind storm damage from around the world. Although results varied from report to report, they came to the following general conclusions regarding tree size and vulnerability to wind damage: most studies found a unimodal distribution of damage relative to stem size – suggesting that the smallest trees are sheltered from wind while the tallest trees are better acclimated; the proportion of broken (vs uprooted) trees is higher in the smallest and largest tree size classes – smaller trees are often broken when large trees fall on them, whereas large trees may be relatively well anchored and fail via crown or stem break. Everham and Brokaw3 found a considerable variation in the effects of stand structure, composition and age. They found that even-aged stands were generally reported to have greater damage than uneven-aged stands, but point out that the uneven-aged stands are often older, and comprised of species mixes, and often of natural rather than planted origin.
Examining trends within forest types, and differentiating natural from managed stands, may provide greater clarity. In European managed forest landscapes, dominated by even-aged stands, storm damage typically increases with stand height and stand age, and with percentage of conifers, particularly of Norway spruce (Picea abies (L.) Karst), and recent thinning or edge exposure by harvesting is an additional risk factor in these forests.75,84
Additional insights come from patterns of wind-driven disturbance in moist, windy landscapes dominated by natural forests. Rebertus et al.54 in Tierra del Fuego and Nowacki and Kramer55 in coastal Alaska identify a pattern of recurrent windthrow of even-aged stands, within landscapes that include older stands with complex structure. While the even-aged, windthrow-origin stands are more common in wind-exposed positions, the presence of uneven-aged stands in wind-exposed locations indicates that they can persist once established. Although trees in dense, uniform canopied stands may experience relatively less wind loading while the canopy is intact, the high degree of uniformity in crown size and stem form can lead to a substantial propagation of damage from newly exposed stand edges during extreme wind events. Conversely, the heterogeneity in tree form in irregular structured or mixed species stands leads to a wider range of stability and the potential for partial stand survival at all but the highest wind speeds.
Reports on the relative vulnerability of different tree species vary widely. Everham and Brokaw3 arrive at the following conclusions: differences in morphology, wood properties and disease resistance do result in some differences in species vulnerability within and between stands; these differences can be obscured by differences in site and stand conditions, and silvicultural history for different populations of the same species. Vulnerability to different modes of damage (crown, stem, roots) also varies between species, and this along with the post-damage response of damaged individuals is a key determinant in subsequent successional dynamics.
Diagnostic considerations – In evaluating stand-level windfirmness, or conducting storm post-mortems, consider whether individual trees are well acclimated to above-canopy winds, and whether if destabilized during a storm they would fall and destabilize neighbouring trees, enabling propagation of damage through the stand during a storm.
Consequences of windthrow
Windthrow affects vegetation and soil components of forest ecosystems at the landscape, stand and microsite scale (see reviews by Webb,8 Ulanova9 and Lugo11). Individual windthrow events affect stand and soil developmental trajectories and leave legacies that last for centuries or millenia. At the landscape level, wind affects stand condition, patch size and distribution of stand types across the landscape. At the stand-level, windthrow damages the overstory and makes light and soil resources available to subcanopy and understory plants, deposits pulses of foliage, branch and stem material on or near the ground, and inverts and exposes soil profiles. At the microsite scale, downed woody material and upturned rootwads create distinctive substrates, and microtopographic features that persist long after the woody material have decayed. This combination of effects makes windthrow different from other natural disturbance processes in forests. Where windthrow recurrence intervals are shorter than the potential lifespan of local tree species, windthrow becomes a major driver of ecosystem dynamics, particularly in ecosystems where fire is rare.85
Effects on stand dynamics and composition
Windthrow patches vary in size depending on storm intensity and duration, and heterogeneity in site and stand conditions. In natural forest landscapes, individual windthrow gaps can exceed 100 ha, but most are smaller than 8 ha and many are very small, on the scale of a few trees.54,55 Similarly, overstory loss rates within gaps vary, to the point that it is difficult to distinguish between a large gap with a high proportion of surviving trees, and a mosaic of small gaps. The recovery of forest ecosystems following extreme weather events depends on the severity and extent of the initial damage, subsequent damage by other disturbance agents, the growth responses of surviving overstory trees, release of understory trees and colonization of available growing space by new regeneration, or expansion of understory shrubs or lianas at the expense of trees. These alternate mechanisms of recovery from wind damage are described by Everham and Brokaw3 as regrowth, release, recruitment and repression. If windthrow preferentially removes shade-intolerant early seral species from the overstory, and these trees are replaced by regrowth or release of more shade-tolerant late-seral species, then windthrow can be viewed as accelerating succession.86,87
Stand-level disturbances can be broadly categorized as resulting in whole-stand, cohort- and gap replacement (gap-phase) of canopy trees, and these become disturbance regimes when they recur over multiple stand developmental cycles. Over very long time frames, the three regimes may interact, and of course wind is just one of many natural disturbance agents in forest landscapes.10 In landscapes with wind-driven disturbance dynamics, mixed stand/cohort/gap-replacement regimes occur where disturbances such as hurricanes produce a gradient of damage across their trackways,33 or where convective storms produce swathes of damage in landscapes that are also affected by lower intensity extra-tropical cyclones.67,88 Local terrain and soil conditions, combined with the response of local tree species, may make a given regime more likely within particular landscape units, especially in topographically complex terrain where severe weather is associated with large-scale but moderate intensity weather systems such as extra-tropical cyclones.54,56
In whole-stand replacement, virtually all of the overstory trees present on a site are damaged during a storm event and new dense cohort of trees is released or rapidly establishes by infill seed. This leads to highly uniform stands with properties that make them vulnerable to recurrent whole-stand replacement as they age.54,55 The dense regeneration that establishes in wind-damaged areas and subsequent stand uniformity predisposes these stands to future instability leading to recurrent episodes of whole-stand replacement. In areas with whole-stand replacement, the potential exists for early seral species to occupy large gaps. However, recruitment of pioneer species depends on their presence within the landscape and their capacity to regenerate on the substrates available within windthrown areas. In coastal Alaska, stand-replacement patches are dominated by western hemlock, a shade-tolerant species that regenerates from seedling banks and infill seed, and by Sitka spruce, a semi-tolerant species that establishes primarily by infill seeding on mineral soil associated with uprooted trees.55,89 In Tierra del Fuego, stand-replacement patches are dominated by southern beech which regenerates en-masse following windthrow events from a bank of short-lived understory seedlings and infill seed.54
In cohort replacement, a substantial proportion of main canopy trees are damaged, often those that have developed large crowns or defective stems with age. For example, in the forests of coastal Alaska, hemlock dwarf mistletoe (Arceuthobium tsugense (Rosendahl) G.N. Jones) along with heart rotting fungi, weaken the crown and bole of older hemlock trees. Recurrent storm damage leads to multi-cohort stands dominated by this shade-tolerant species, with distinct age cohorts.55
In gap-replacement, disturbances lead to canopy gaps from one to several mature tree crowns wide. This is a common mode of wind disturbance in forests and is particularly well studied in the hardwood and mixed forests of the eastern US.10,33,90 For example, the evergreen-deciduous forests of the Great Lakes Region of North America experience recurrent damage from thunderstorm activity, and extra-tropical cyclones.85,88 Partial loss of canopy trees during these storms leads to a mixed regime of cohort replacement and small gap creation. Whole-stand replacement is rare in the Great Lakes Region, leading to forests dominated by shade-tolerant species that regenerate primarily from seedling banks and resprouting of damaged stems. In contrast, Ulanova,9 working in the Russian Boreal forest, reported that light-demanding birch (Betula pendula Roth.) was able to establish on mineral soil seedbeds alongside more shade-tolerant Norway spruce, leading to multi-aged mixed species stands in forests with recurrent wind-driven small gap formation. Gap models such as SORTIE have been used to simulate successional trajectories in mixed species stands for gradients of storm damage.91
In cold climates, stresses induced by low temperatures are compounded by the effects of wind on dessication, wind loading and abrasion by wind-driven snow, freezing rain or ice crystals. In areas with high wind exposure, these combined influences affect the ability of trees to colonize or persist as an erect growth form, and create and modify local tree-lines.73,92 The interaction between wind-induced stresses and tree age, and the consequences for stand development are elegantly illustrated in wave forests. Wave forests are typically mono-specific forests of relatively short-lived trees growing in windy landscapes with strong prevailing wind directions, and have been reported in northeastern US,93 in Newfoundland,29 Japan94 and Tierra del Fuego.95 An exposed front of ageing trees develops perpendicular to the wind direction, with synchronous mortality as trees along this front die from the combined effects of age, exposure and mechanical damage.96 Dying trees are replaced by a new cohort of regeneration of the same species. This front of senescing trees and regeneration gradually moves across the landscape in the direction of the prevailing wind, like a wave, and parallel waves of these bands of progressively old to young trees occur across the landscape. The width of a band reflects the lifespan of the dominant tree species – the time taken for the trees to grow into a condition of fatal wind vulnerability.
Diagnostic considerations – In evaluating the effect of the local severe weather regime on stand dynamics consider whether severe weather recurs at intervals shorter than the potential lifespan of local tree species, where the local disturbance regime falls on the spectrum of whole-stand replacement to small gap creation, or is a mix of regimes, and the manner in which terrain features influence this pattern. Consider whether stands recover via recruitment of shade-intolerant species, release of shade intolerants or regrowth of surviving trees. Consider also the woody debris and soil legacies left by windthrow and the interaction of the windthrow regime with other disturbance processes.
Effects on large woody debris dynamics, insects and fuel loading
Like other forest disturbances such as wildfire, insect or disease, windthrow converts living trees to dead broken or downed large woody debris, and this material persists as a legacy of the previous stand, and the damage event. Dead standing and downed trees are an important component of forest ecosystems, providing habitat and substrate for a succession of microbial, insect, plant and animal communities through progressive stages of decomposition.14,97 Log decay rates depend on local climate, log size, species and whether logs are initially above or in contact with the ground, or saturated.98 Large wood that becomes incorporated in stream channels contributes complexity to channel morphology and in-stream habitat.99 Bahuguna et al.,100 however, noted that it may take decades for spanning logs to decay to the point that they drop into the stream channel, and the in-stream longevity of this already decayed material is likely shorter than that of the material entrained in alluvial sediments while still sound.
The weakened trees and debris created by wind storms can also pre-dispose stands to secondary disturbance. Freshly damaged trees attract a succession of insects, including bark beetles – populations of which can build up in downed and damaged material following major windstorms and spread to live-standing trees.14,101 The branch and other fine fuel deposited on or near the ground following storm events adds to the fuel loading in forests, and storm events may be followed by wildfire, even in forest types where live forests are not typically flammable.10 Large-scale windthrow associated with major storms has the potential to influence atmospheric carbon dynamics. Chambers et al.102 estimated that the forest biomass loss associated with Hurricane Katrina was equivalent to 50–140% of the net annual US carbon sink in forest trees.
Diagnostic considerations – When evaluating the role of windthrow in local large woody debris dynamics, consider the periodicity of windthrow events, the species and sizes of trees affected and the proportion of broken vs uprooted trees. Consider how the arrangement of downed material affects decomposition rates, and the contribution of downed and damaged material to bark beetle population dynamics, wildfire fuel loads, terrestrial habitat and stream morphology.
Effects on soil
Windthrow has profound and long-term effects on soil properties. Unlike wildfire, insect or disease-mediated tree mortality, wind-uprooted trees produce upturned root systems and expose and invert volumes of mineral soil and forest floor. This process has been referred to as an example of bioturbation, floralturbation or floral pedoturbation and appears to be a standard feature of forests around the world, even in fire-dominated landscapes.13,103,104 As the woody components decay, a pit-mound complex remains with the mound forming on the leeward side of the pit from soil that was entrained in the upturned portion of the root system. The dimensions of these pit-mound complexes vary depending on whether the leeward root system acts as a hinge (hinge fall), or the leeward portion of the root system slips back into the pit during failure (rotational fall), root-system volume and dimensions, soil depth and cohesion to the root system during failure.103 Pit-mounds may persist for hundreds to potentially thousands of years after the windthrow event.13 Pit-mounds introduce microtopographic and microclimatic complexity – mounds are warmer and drier than pits and adjacent undisturbed soils during the growing season. This microsite-level heterogeneity promotes diversity in understory flora105 and influences successional pathways in storm prone landscapes.9,63
The distinctive microsites and repeated soil turnover caused by recurrent windthrow affect the nature and rate of pedogenesis, particularly in climates where podzolization and paludification occur. At the microsite scale, pits accumulate organic matter, and this combined with the cooler and moister conditions leads to enhanced illuviation of sesquioxides compared with adjacent mounds.13 At the site level, the pulse of downed branch and stem material, inversion of the soil during recurrent windthrow events, reverses podzolization and affects soil carbon and nutrient dynamics,106 site fertility and microbial communities.107,108 In the absence of periodic soil inversion in cool humid climates, the accumulation of organic matter at the soil surface and the illuviation of iron and aluminium oxides and organic matter impairs soil drainage.104,106 The resulting accumulation of organic matter, decline in forest productivity and ultimately bog formation on initially well-drained sites have been described by Simard et al.109 as successional paludification. Bormann et al.104 suggest that windthrow, or disturbances that mimic windthrow, may be necessary at intervals of 200–400 years in order to maintain soil productive capacity in wet cool temperate ecosystems.
Windthrow of trees in areas of steep terrain can also lead to gradual down slope transport of sediment. The potential for gradual down slope movement results when soil entrained in the upturned rootwad deposits on the downslope side of a developing mound, rather than in the upslope pit from which it came.110,111 In areas with thin soil mantles, root penetration of fractured bedrock, and the entrainment and exposure of bedrock in upturned rootwads, facilitates weathering and soil production.112 The role of windthrow in the production and down slope movement of surface soil mantles has been explicitly included in geomorphological models for hillslope processes in forested landscapes.113
Diagnostic considerations – When evaluating the role of windthrow in local soil processes and site fertility, consider the abundance, size and longevity of pit-mounds and the effects of periodic soil inversion on organic matter accumulation and soil drainage. Consider the interaction between soil fertility and stand stability.
Management in windy landscapes
Forests are managed for a variety of purposes under widely varying management paradigms. The aims of forest management, even within a given region, may span conservation or restoration of natural areas, extensive forest management where production of goods and services is balanced with maintenance of native ecosystems and biological diversity, or intensive production of timber or other forest crops via plantation or other forms of forest cultivation. In forested regions with recurrent severe weather, managers can embrace windthrow as a natural disturbance process in natural or semi-natural ecosystems while attempting to reduce its impacts in managed forests and built-up areas – but they should not overlook or ignore it. Windthrow management should take place within a framework of general risk management, with the likelihood, severity and potential impacts considered prior to deciding on acceptance, mitigation or cost-sharing responses.114 The impacts of windthrow, and preemptive or reactive management responses, should be considered with reference to the broad and specific aims of stand and landscape-level management.
Recognition of the historical natural disturbance regime and how it affects stand and landscape dynamics is identified as a guiding principle in ecologically sustainable management of native forests.115 The differences between stand-replacing and stand-maintaining fire regimes are reasonably well understood and form the basis for stand- and landscape-level planning and prescriptions in fire-prone landscapes.116 A similar appreciation for the role of wind is developing, and this has been incorporated into the concept of natural disturbance types which forms a framework for ecosystem management in British Columbia.117,118 The challenge is to link local disturbance regimes to the climatic and topographic conditions which affect the recurrence, extent and severity of damaging events, and to relate this to the resulting patterns in vegetation and soil conditions. The link between climate, landscape unit conditions and local disturbance regimes is well developed in the field of geomorphology, and Montgomery119 provides a potentially useful integrating framework in his concept of geomorphic ‘process domains’. Geomorphic process domains are spatially identifiable areas in which the local geology, topography, climate and biota result in distinctive suites of recurring geomorphic disturbances which in turn affect soil and vegetation community development. Montgomery119 focused on the geomorphic consequences of natural disturbances, but the process domain concept incorporates fire, wind and other recurrent natural disturbance processes that are similarly affected by the interaction of topography, soils, climate and biotic communities. Re-examining landscapes through the lens of natural disturbance process domains would improve our understanding of the processes underlying soil fertility and successional dynamics of vegetation communities, and enable managers to make more informed decisions about conservation or land-use activities. One such application would be in the design of salvage operations following major storms. As noted by Lindenmayer and Noss,120 logging in the wake of natural disturbances can remove biological legacies and modify rare post-disturbance habitats. These may be distinctive features in landscapes with intermittent, but recurrent wind damage.
In landscapes where management is focused on cultivation of forests for goods and services via plantation and other forms of intensive forestry, and where tree failure impacts human-built infrastructure, managers need to recognize and consider the implications of recurrent severe weather and windthrow disturbance regimes. At the tree and stand level, silviculture regimes can be designed to promote tree-level acclimation and avoid abrupt changes in exposure of retained trees.121 Schelhaas et al.65 reviewed the literature concerning trends in natural disturbances in European forests from 1850 to 2000. They concluded that while storm intensity does not appear to have increased, the forest damage caused by storms has substantially increased since the 1950s. They attribute this to a number of factors, including the expansion of the forest estate and the establishment and maturation of conifer-dominated forests in storm prone locations that had not supported forests for many human generations.
There are various approaches to predicting windthrow in natural or managed forests including diagnostic techniques, and statistical and mechanistic models. The phenomenon of stand failure at critical height in UK Sitka spruce plantations motivated the development initially of a windthrow hazard classification system51 and then the ForestGALES model,122 to assist with the siting and harvest scheduling of plantations in this windy landscape. Hybrid-mechanistic models such as ForestGALES can be refined and extended to apply to a variety of forest types, disturbance regimes and management scenarios,7 and when coupled with stand growth and regional climate models, can be used to explore implications of changing climates.123 However, there remains a significant amount of work to develop process-based models that integrate disturbance mechanics and their short- and long-term effects on ecosystem dynamics, and this is a fertile area for interdisciplinary collaboration.124
Recurrent windthrow is a feature of many forested landscapes and falls within the spectrum of wind effects on forests. It is a complex disturbance process that results from the interaction of regional weather systems with terrain, stand and soil conditions, and influences stand dynamics and soil processes. The variability in windthrow patterns can be better understood if the nature of the regional wind climate and interaction with terrain is considered, if stand stability is evaluated in terms of acclimative growth, and if soils are examined through their effects on stand productivity and anchorage. There is much to be gained in interdisciplinary communication about the nature and consequences of recurrent wind damage. There are opportunities for climatologists, engineers, ecologists, geomorphologists and others to develop integrative process models at the tree, stand and landscape scales that will improve our collective understanding, and inform management decision-making.
Conflict of interest statement