Regeneration Dynamics of coast redwood, a sprouting conifer species: A review with implications for management and restoration

: Coast redwood ( Sequoia sempervirens (Lamb. ex. D. Don) Endl.) is unique among conifer species because of its longevity, the great sizes of individual trees, and its propensity to reproduce through sprouts. Timber harvesting in the native redwood range along the coast of the western United States has necessitated restoration aimed to promote old forest structures to increase the total amount of old forest, the connectivity between old forests, and to enhance the resiliency of these ecosystems. After disturbance or harvest, healthy redwood stumps sprout vigorously, often producing dozens of sprouts within two years of disturbance. These sprouts form highly aggregated spatial patterns because they are clustered around stumps that may number less than 50 ha − 1 . Thinning of sprouts can accelerate individual tree growth, providing an effective restoration strategy to accelerate formation of large trees and old forest structures or increase stand growth for timber production. However, management, including restoration activities, is a contentious issue throughout the native range of redwood because of the history of overexploitation of this resource and perceptions that overexploitation is continuing. This paper reviews the science of early stand dynamics in coast redwood and their implications for restoration and other silvicultural strategies. ﬁnal stand volume but did have a major effect on the volume increment for both overstory and understory components. Greater overstory density increased the volume increment of the overstory component while decreasing the volume increment of the understory, reﬂecting a reduction in growing space available to the understory. These modeling results indicated that variable retention systems can provide old forest features while providing timber production. Variable retention therefore offers a restoration option for linking old forests in managed landscapes.


Introduction
Restoration is an increasingly common management objective in coast redwood (Sequoia sempervirens (Lamb.ex.D. Don) Endl.)) forests.Decades of over-harvesting in the redwood range have left a small percentage of old presettlement forests [1].Much of this is already preserved in national, state, or local park systems.Other areas that have been harvested exist in a wide range of stand structures resulting from variation in time since harvest, harvest methods, site qualities, and subsequent management activities.Many public agencies and conservation organizations have concluded that restoration is needed to direct the developmental trajectory of some previously harvested stands towards old forest structures to increase the total amount of old forest, connectivity between old forests, and the overall resiliency of these redwood ecosystems [2,3].
Coast redwood (also known as "redwood") is renowned for its magnificent groves of large trees.Redwood trees reach greater heights (>110 m; [4]) than any other species and redwood forests can store more biomass than any other ecosystem [5,6].Redwood stand production rates are also exceptional: Jones and O'Hara [7] reported values for mean annual increment and periodic annual increment of 33.2 and 75.6 m 3 ha −1 year −1 in a young redwood plantation.Similarly, periodic carbon increment can exceed 16 tonnes ha −1 year −1 [7].The natural range of the species is limited to the Pacific coast of central and northern California, and the southwest corner of Oregon over a latitudinal range from 35.68 to 42.15 N, from sea level to 750 m.Coast redwood grows in a strip that is 12-32 km wide and characterized by frequent fog.Old forest groves are found interspersed throughout much of this range.Redwood is also an increasingly popular plantation species in New Zealand and China.
Redwood is an ancient conifer species in the Taxodiaceae family dating to the Mesozoic era [8].It is also a hexaploid species; this means that the same gene can be represented by up to three different alleles.This allows for an enormous number of combinations of those alleles in each generation of trees.The greatest genetic diversity is observed within redwood families [9].Redwood is also an unusual conifer species because of its ability to sprout prolifically after disturbances, including harvest treatments.Trees live in excess of 2000 years and, in old stands, they may host arboreal plant and animal communities in the upper canopy [10].The heartwood of redwood is decay-resistant and large logs may last centuries on the forest floor.Common associated species include tanoak (Notholithocarpus densiflorus (Hook.& Arn.) P.S. Manos, C.H. Cannon, & S.H. Oh), California bay (Umbellularia californica (Hook.& Arn.) Nutt.), red alder (Alnus rubra Bong.), coast Douglas-fir (Pseudotsuga menziesii (Mirb.)Franco var.menziesii), grand fir (Abies grandis (Douglas ex D. Don) Lindl.),Sitka spruce (Picea sitchensis (Bong.)Carrière), western hemlock (Tsuga heterophylla (Raf.)Sarg.), western redcedar (Thuja plicata Donn ex D. Don), knobcone pine (Pinus attenuata Lemmon), and others [8].Conifer species such as western hemlock and coast Douglas-fir have lifespans of up to 500 years or longer, whereas associated broadleaf species have much shorter longevity [8].
A common misconception is that redwood has few stand-level management challenges because of the ease of regeneration and its rapid growth rates [11].However, management challenges do exist and are often unique for this species, particularly with regard to density management and restoration.In this review, we address the regeneration dynamics of this unusual species and their implications for stand-level management and restoration.We also address the contention that management treatments pursuing restoration objectives are ineffective or unnecessary because of redwood's unique silvical characteristics.Our specific objectives were to: (1) review the regeneration dynamics of coast redwood after a variety of disturbance events; (2) review the dynamics of stand development and density management in even-aged and multiaged stands; (3) discuss the implications for management of stands of different structures and managing for different objectives including restoration; and (4) discuss the implications for management with climate change and the presence of invasive organisms.

Disturbance History
Disturbance regimes in coast redwood forests are dominated by fire, which historically occurred throughout the entire redwood range [12].Ignition sources were primarily lightning and Native American burning [13][14][15][16].Redwood is very resistant to fire primarily because of its thick, insulating bark.It also has the ability to sprout prolifically at the base after disturbance.Fires ranged from low to high intensity but were commonly surface fires that meandered on the forest floor and occasionally into the crowns [12].This provides a competitive advantage for existing redwood stems against species that are less resistant to fire, sprout less prolifically, or grow more slowly [16].On alluvial flats where redwood ecosystems are typically most productive [17], flooding and sediment deposition represent another disturbance type [12].Redwood withstands severe flooding, waterlogged soils, and heavy sediment deposition [12,18,19], thereby providing another competitive advantage to redwood on sites where flooding occurs.Mass movement, wind damage, insects, and pathogens are generally not major threats to redwood [12].Redwood's adaptations to disturbance make it competitive throughout its range.Typical disturbance regimes can result in a variety of multiaged stand structures ranging from pure stands to complex mixtures with many different conifer and broadleaved trees.Even-aged structures are also possible but not common.Combined with redwood's longevity and great stature, disturbance regimes also contribute to the dominance of redwood on suitable sites across its range.

Site Productivity
Isolated to a relatively narrow strip on the Pacific coast, the range of coast redwood is indicative of the site-sensitivity of the species.It is a fog-dependent species found in the coastal fog belt [8,20].In addition to being limited to the coastal fog belt, redwood achieves greater productivity on wetter sites including alluvial flats and riparian zones [8,20].Site index, as the traditional measure of site quality in forestry has limitations in redwood because of inconsistencies in annual ring formation [21] and annual height growth patterns [22].Berrill and O'Hara [17,22] explored the complex biophysical factors that affect redwood height and basal area increment revealing different factors affecting each.Redwood volume increment was greater on lower slopes, soils with higher pH, and sites with more Douglas-fir and fewer broadleaved trees.
Floristic-based vegetation classifications have also been used to describe coast redwood forest associations throughout its range [23].Differences in associated vegetation species are closely related to site moisture availability [19].Mahony and Stuart [23] noted that floristic composition and stand structure are highly variable among coast redwood forests.In northern California, Lenihan et al. [24] classified recently harvested coast redwood stands based on understory species and found that associations varied by both level of soil disturbance and site moisture.Sites that were less disturbed were more similar to old-growth redwood stands than highly disturbed sites [25].Similar vegetation associations that vary with topography have been documented in the central portion of the redwood range [8].Indicator species may also be used to help identify old forest status, although designated indicator species may occur in multiple developmental stages or be assigned indicator status in stands outside of their native range [26].Because of this variability, it may be necessary to consider floristic associations in conjunction with site characteristics and disturbance history to evaluate the site productivity and potential for restoration of any redwood stand.
All measures support the view that redwood is highly site-sensitive.This sensitivity is evident across scales from the limits on the redwood range to importance of microsite factors in influencing productivity.This site sensitivity also indicates that comparisons of redwood stand structure and development must be vigilant to assure that site variation is not an unintended variable.

Sprouts
Unlike most conifers, coast redwood primarily relies on vegetative reproduction.New stems sprout from a lignotuber or burl at the base of the original leader [27].These burls eventually form swelling at the base of the tree.Basal burls serve as storage for many suppressed buds that may sprout after damage to the primary leader, contain carbohydrates and nutrients to facilitate growth of sprouts, generate roots for the tree, and may provide increased stability on steep slopes [8].The lignotubers are therefore the source of the sprouting ability that distinguishes redwood from most other conifers.
Redwood trees may produce sprouts within weeks after cutting regardless of age, although older trees have reduced sprouting [28].For example, Barrette [29] reported 95% of young growth stumps produce sprouts and 85% of older stumps produce sprouts.Similarly, others have reported slight declines in redwood sprouting with increasing size and age [30][31][32] and no decline from prescribed burning [28].Redwoods can produce in excess of 100 sprouts per clump [33] that grow exceptionally fast in full sunlight, but also self-thin over time.O'Hara et al. [34] reported sprout growth over 1 m year −1 through five years in light environments where percent above canopy light was about 40% or greater.This process has been described in a sprout clump production model where the development of new photosynthetic capacity in the clump occurs until a physiological equilibrium is reached [31,35,36].After this point, the clump is primarily self-sufficient and no longer relying on stump reserves.Whether the dynamics of clump self-thinning change at this point is not known.Many associated species in the redwood ecosystem resprout after disturbance.In addition to redwood, tanoak, California bay, red alder, among others, are prolific sprouters.

Seedlings
Although redwood trees produce seed each year, seed production varies throughout the range and some areas may experience prolonged periods without significant seed production, though only a proportion of this seed may be sound [20].Compared to other species, redwood seed requires more moisture than most for germination.High levels of leaf litter accumulation in redwood stands, compared to other forest types, result in relatively few sites suitable for seedling establishment [8].Seedlings tend to grow less vigorously than redwood sprouts.This is also true for planted seedlings: Jameson and Robards [37] found that planted seedlings only grew to half the size of naturally regenerated sprouts of the same age.

Spatial and Clonal Patterns
Restoring redwood requires an understanding of spatial patterns.Spatial patterns in redwood stands are highly variable depending on site factors, disturbance history, and time since disturbance.Additionally, the scale of analysis and the sizes of trees included in a spatial analysis may influence the results.For example, Dagley [38] found random patterns of large trees in old forest stands on alluvial sites, but smaller trees formed clumped distributions.Van Mantgem and Stuart [39] also found random patterns of old forest trees on upland sites, which were clumped at smaller spatial scales.In young growth forests, spatial patterns may be highly variable depending on site [40].However, given the propensity for sprouting following disturbance, young stands typically form much more aggregated patterns than the older stands or forests they replace (Figure 1).These results suggest that, following stand replacement disturbance, redwood stands develop from highly aggregated patterns to more random patterns in older stands.This is largely due to the process of self-thinning within clumps from many to few stems over many centuries of development.More complex multiaged stands may express both patterns: clumpiness in younger patches and randomness in older.The implications for restoration treatments are to avoid uniform distributions after the thinning [38,39].
relying on stump reserves.Whether the dynamics of clump self-thinning change at this point is not known.Many associated species in the redwood ecosystem resprout after disturbance.In addition to redwood, tanoak, California bay, red alder, among others, are prolific sprouters.

Seedlings
Although redwood trees produce seed each year, seed production varies throughout the range and some areas may experience prolonged periods without significant seed production, though only a proportion of this seed may be sound [20].Compared to other species, redwood seed requires more moisture than most for germination.High levels of leaf litter accumulation in redwood stands, compared to other forest types, result in relatively few sites suitable for seedling establishment [8].Seedlings tend to grow less vigorously than redwood sprouts.This is also true for planted seedlings: Jameson and Robards [37] found that planted seedlings only grew to half the size of naturally regenerated sprouts of the same age.

Spatial and Clonal Patterns
Restoring redwood requires an understanding of spatial patterns.Spatial patterns in redwood stands are highly variable depending on site factors, disturbance history, and time since disturbance.Additionally, the scale of analysis and the sizes of trees included in a spatial analysis may influence the results.For example, Dagley [38] found random patterns of large trees in old forest stands on alluvial sites, but smaller trees formed clumped distributions.Van Mantgem and Stuart [39] also found random patterns of old forest trees on upland sites, which were clumped at smaller spatial scales.In young growth forests, spatial patterns may be highly variable depending on site [40].However, given the propensity for sprouting following disturbance, young stands typically form much more aggregated patterns than the older stands or forests they replace (Figure 1).These results suggest that, following stand replacement disturbance, redwood stands develop from highly aggregated patterns to more random patterns in older stands.This is largely due to the process of self-thinning within clumps from many to few stems over many centuries of development.More complex multiaged stands may express both patterns: clumpiness in younger patches and randomness in older.The implications for restoration treatments are to avoid uniform distributions after the thinning [38,39].The sprouting ability in redwood creates a spatial distribution pattern where members of one clone, or a distinguishable genotype, are more likely to be clumped together.Narayan [41] found intraclonal distances were usually small and did not exceed 10 m, which was consistent with earlier studies in which clones were identified using allozyme markers [42].In some cases, however, The sprouting ability in redwood creates a spatial distribution pattern where members of one clone, or a distinguishable genotype, are more likely to be clumped together.Narayan [41] found intraclonal distances were usually small and did not exceed 10 m, which was consistent with earlier studies in which clones were identified using allozyme markers [42].In some cases, however, Narayan [41] found the distance between clones was as great as 60 m.Rogers [42] found even greater distances of up to 340 m and suggested that these disjunct patterns may be the result of redwood's ability to sprout from branches driven into the ground after a tree fall.The individuals found in close proximity or in clonal clumps often belong to more than one genotype [41][42][43].This may be a result of several factors, including various patterns of sprout/seedling recruitment.For example, when seedlings and/or sprouts of different genotypes growing in close proximity, fuse together as the stand develops.The genetic difference could also be a result of somatic mutation in the resprouting clones.
Within-stand genetic diversity of redwood is high due to the hexaploid nature of redwood.Mutations have likely accumulated over time and persisted in individuals that are now the only remaining representatives of much greater lineages.The clonal reproduction, while impeding the development of new genotypes, helps to preserve the historically high allelic richness.Occasionally, large numbers of individuals within a clone can be found.Narayan [41] reported the largest number of stems in one clone was 27 stems in old forest stands.Often, however, numbers of ramets or stems per clone tended to be confined to single individuals [41].Spatial patterns of clones, and variations in clone numbers, are highly variable and will be difficult to emulate with restoration treatments.

Even-Aged Stands
Even-aged stands are not common natural structures in redwood forests.Disturbance regimes that are dominated by fire or flooding and sedimentation [12] do not typically result in redwood-dominated even-aged stands.Instead, relatively uncommon disturbances such as landslides, or more severe fires that might occur on the drier sites at the edge of the redwood range, may result in even-aged stands.However, management activities such as clearcutting or seed tree systems, or the heavy-handed extraction practices of early timber harvests, did result in substantial areas of even-aged stands across the redwood range.
Stand development in even-aged redwood forests is, in many ways, typical of development in other even-aged ecosystems.Stem exclusion [44] is quickly reached as trees compete for dominance and growing space.Redwood approaches a stand-level maximum-density line with a similar slope as other species, although with a higher intercept [45].A notable difference from other conifer species is the aggregated distribution of young redwood forests due to the sprouting from cut stumps (Figure 1).There are few studies of self-thinning of individual sprout-clumps and none of sprout clump dynamics in full sunlight.
Other species, particularly tanoak, are prolific sprouters that quickly reoccupy growing space following a disturbance [16,46].However, faster growing conifers and some broadleaved trees soon emerge as dominants.The rapid initial height growth of redwood sprouts generally allows them to outgrow their competitors.This includes planted conifer seedlings [37].However, seedling-origin Douglas-fir will often outgrow the sprout-origin redwood by about age 100 [17,47].Hence, young forests often include a large Douglas-fir component that may persist through the first century or longer of even-aged development.Redwood and Douglas-fir will eventually outgrow tanoak and other broadleaved species [48].Alternatively, Douglas-fir, and probably redwood, can grow through a broadleaved canopy to achieve dominant status [49].

Multiaged Stands
Old redwood stands have complex age structures [12,50].However, aging large trees is difficult, and most knowledge of age structures is through analysis of cut stumps in harvested stands [51,52].
Given the longevity of coast redwood and low densities of large trees (Table 1), size frequency distributions are highly variable [12,52].The replacement regime in these old forests is predominantly driven by individual or group tree falls, or trees dying in place.Larger gaps may be successfully colonized by new seedlings or sprouts of redwood or other species.In other cases, but not commonly, disturbances may create patches of even-aged trees.Most notable of old forest dynamics is the complexity in the shapes and arrangement of tree stems and crowns.These trees form multiple stems, or reiterations, that dwarf many other tree species and form complex aerial assemblages of plants and animals many meters off the ground [6,10,53].The relatively small number of old forest trees in these stands, their longevity, and their immense size combine to form a "super structure" that is both immense in size, and slow to change.Indeed, our limited understanding of these old forest ecosystems is in part due to the relatively short period of time humans have been studying them [12,51,52].In these complex multiaged stands, a disturbance which makes growing space available can result in similar processes as in even-aged stands but at a smaller scale.Open growing space is filled by seedlings or sprouting species and the site begins a process of stem exclusion.However, the poorer quality light environment may encourage a different suite of species.For example, redwood clumps require sufficient light to persist and develop [34,35].In poor light environments, numbers of stems in individual sprout clumps declined as the light regime declined due to crown closure (Figure 2).In the poorest light environments, entire sprout clumps can die.However, Figure 1 also demonstrates how self-thinning of sprout clumps in better light regimes may be very slow.Because redwood has a high shade tolerance, attains high levels of leaf area index [6,34,35], and has a very plastic crown form capable of quickly reclosing canopy gaps, multiaged stands risk losing regeneration from excessive shade [35].With sufficient light, understory trees can survive and develop into and through upper canopy layers [35].
Forests 2017, 8, 144 7 of 18 the poorest light environments, entire sprout clumps can die.However, Figure 1 also demonstrates how self-thinning of sprout clumps in better light regimes may be very slow.Because redwood has a high shade tolerance, attains high levels of leaf area index [6,34,35], and has a very plastic crown form capable of quickly reclosing canopy gaps, multiaged stands risk losing regeneration from excessive shade [35].With sufficient light, understory trees can survive and develop into and through upper canopy layers [35].

Management
Forest management may be used to pursue many alternative objectives ranging from managing for timber production to restoration.For managed stands, the combination of different previous harvest methods, site quality variation, time since harvest, and post-harvest management activities creates a vast, multidimensional array of existing or potential stand structures.This array of existing and potential stand structures therefore requires a wide array of silviculture: there is not one approach, one silvicultural system, or one form of management that can work for all situations.This is also true for restoration.However, throughout the natural range of coast redwood, management, including restoration, is a contentious issue due to the history of overexploitation of this resource, perceptions that overexploitation is continuing, and the iconic nature of these forests.Management of redwood forests has been an explosive issue for decades [64][65][66] and will likely continue to be contentious.

Young Redwood Stand Management
Given redwood's shade tolerance and prolific sprouting ability, it is amenable to many different silvicultural systems and stand structures.Hence, in the redwood region, industrial owners will commonly use even-aged systems such as clearcutting, or multiaged systems like single-tree and group selection.One of the variables in existing stand structures is the amount of tanoak present, and this is largely related to site characteristics and management history [67].After clearcutting, tanoak is an effective competitor to redwood or Douglas-fir because it sprouts prolifically and there are often more tanoak than redwood clumps.Hence, herbicides are commonly used to control tanoak and other shrub species.These treatments are most commonly hand applications of chemicals directly to target plant species.
Given the low numbers of overstory trees in old forest stands (Table 1), cutting these stands results in a relatively small number of redwood sprout clumps with large spaces between these clumps (Figure 3).Interplanting, or enrichment planting, can be used to increase the redwood

Management
Forest management may be used to pursue many alternative objectives ranging from managing for timber production to restoration.For managed stands, the combination of different previous harvest methods, site quality variation, time since harvest, and post-harvest management activities creates a vast, multidimensional array of existing or potential stand structures.This array of existing and potential stand structures therefore requires a wide array of silviculture: there is not one approach, one silvicultural system, or one form of management that can work for all situations.This is also true for restoration.However, throughout the natural range of coast redwood, management, including restoration, is a contentious issue due to the history of overexploitation of this resource, perceptions that overexploitation is continuing, and the iconic nature of these forests.Management of redwood forests has been an explosive issue for decades [64][65][66] and will likely continue to be contentious.

Young Redwood Stand Management
Given redwood's shade tolerance and prolific sprouting ability, it is amenable to many different silvicultural systems and stand structures.Hence, in the redwood region, industrial owners will commonly use even-aged systems such as clearcutting, or multiaged systems like single-tree and group selection.One of the variables in existing stand structures is the amount of tanoak present, and this is largely related to site characteristics and management history [67].After clearcutting, tanoak is an effective competitor to redwood or Douglas-fir because it sprouts prolifically and there are often more tanoak than redwood clumps.Hence, herbicides are commonly used to control tanoak and other shrub species.These treatments are most commonly hand applications of chemicals directly to target plant species.
Given the low numbers of overstory trees in old forest stands (Table 1), cutting these stands results in a relatively small number of redwood sprout clumps with large spaces between these clumps (Figure 3).Interplanting, or enrichment planting, can be used to increase the redwood stocking in these stands whether they are being managed with even-aged or multiaged systems.Planting stock is either from seedlings, or, more commonly in recent years, from tissue culture of superior genotypes.Subsequent rotations of stands with enrichment planting will not likely need additional planting because of better spatial arrangements of existing redwood trees.Rotations are approximately 50 years, the lowest rotation length permitted in California [68].Early thinning or precommercial thinning is common to reduce overall stand density and to reduce the number of sprouts in redwood sprout clumps.Sprout clumps are typically thinned to two or three sprouts per clump, but repeated generations of stumps can form complex patterns of stumps that may extend over several meters.In these cases, more sprouts may be left.Thinning redwood sprout clumps reduces aggregated spatial patterns, promoting a more random arrangement of stems found in even-aged or complex, multiaged stands [29,[69][70][71] and may not always increase sprout growth [71].Age of sprouts affects sprout size, which affects equipment options for thinning clumps [72].Precommercial thinning is used between ages 10 and 20 to reduce density and improve average spacing.It has been effective at increasing individual tree sizes.Lindquist [73,74] thinned 19-year-old stands and found increased tree growth, but no effects on stand increment.Thinning in 10to 12year-old stands produced similar results [75].The greatest increment over the 12-year study was at a spacing of approximately 1.2 m.However, thinning heavily at age 9-11 resulted in large amounts of ingrowth, particularly at wider spacings, because of the large amounts of growing space that were left unoccupied (Figure 4; [75]).Hence, it is possible to thin too early in stands of sprouting species like redwood.
Modeling early redwood stand growth confirms results from empirical thinning studies: thinning can substantially increase growth of remaining trees and accelerate process of stand development [76].Berrill et al. [77] modeled young plantations with and without thinning and found thinning put stands on a more rapid trajectory towards the desired old forest reference condition.One justification of restoration of old redwood forests is to reduce the vulnerability of younger stands to fire.Although thinning creates activity fuels, these hazards are short-lived after thinning.For example, O'Hara et al. [78] measured slash depth after restoration thinning in young stands at the northern end of the redwood range and found they were negligible ten years after thinning.In other parts of the range, these fuels might be more persistent, but thinning appears to have beneficial longterm effects as a fuel reduction strategy.Early thinning or precommercial thinning is common to reduce overall stand density and to reduce the number of sprouts in redwood sprout clumps.Sprout clumps are typically thinned to two or three sprouts per clump, but repeated generations of stumps can form complex patterns of stumps that may extend over several meters.In these cases, more sprouts may be left.Thinning redwood sprout clumps reduces aggregated spatial patterns, promoting a more random arrangement of stems found in even-aged or complex, multiaged stands [29,[69][70][71] and may not always increase sprout growth [71].Age of sprouts affects sprout size, which affects equipment options for thinning clumps [72].Precommercial thinning is used between ages 10 and 20 to reduce density and improve average spacing.It has been effective at increasing individual tree sizes.Lindquist [73,74] thinned 19-year-old stands and found increased tree growth, but no effects on stand increment.Thinning in 10-to 12-year-old stands produced similar results [75].The greatest increment over the 12-year study was at a spacing of approximately 1.2 m.However, thinning heavily at age 9-11 resulted in large amounts of ingrowth, particularly at wider spacings, because of the large amounts of growing space that were left unoccupied (Figure 4; [75]).Hence, it is possible to thin too early in stands of sprouting species like redwood.Another silvicultural treatment in young redwood stands is pruning.Artificial pruning can increase clearwood production and possibly increase the proportion of more valuable heartwood [79].A potential drawback is the production of epicormics sprouts with severe pruning [80].There is no significant interest in pruning from industrial owners in the redwood region due to cost and uncertainties over future markets.However, there is international interest from countries where redwood plantations are being established [81].Pruning could be used to remove branches in restoration treatments to accelerate the formation of clear boles and their aesthetic values, although it may be cost prohibitive activity for the agencies typically conducting restoration.
The American black bear (Ursus americanus Pallas) can also be an obstacle to restoration in parts of the redwood range [82,83].The bears scrape off the bark and eat the cambium, thereby injuring or even girdling stems.Coast redwood stems will usually resprout below the girdle.Other conifer species can be killed by this activity.The bears have a preference for more vigorous trees [83,84] and may also prefer redwood to Douglas-fir [83,85].Thinning to achieve restoration objectives of faster tree growth and lower densities may increase bear damage by encouraging the development of more vigorous trees.The resprouting of girdled redwood stems is contrary to restoration objectives, as it results in an increase in tree density and smaller tree sizes.Repeated bear damage can potentially prevent a stand from achieving restoration objectives.

Older Redwood Stand Management
Redwood trees respond to the increased growing space provided by thinning in older stands [86].In Lindquist's Whiskey Springs thinning study at Jackson Demonstration State Forest, evenaged stands exhibited a consistent, strong growth response between the ages of 40 to 75 years to thinning [73,87].Periodic diameter growth increased with severity of basal area reduction (Figure 5).The lowest retention of basal area resulted in the greatest periodic diameter growth and resulted in the biggest trees.The most severe thinning also exhibited a statistically significant loss in volume production primarily caused by the leave trees not fully occupying the site [86].Thinning operations accelerate the development of larger diameter redwood trees [73,86,87], promoting old forest structural features beneficial for creation of wildlife habitat [8].Modeling early redwood stand growth confirms results from empirical thinning studies: thinning can substantially increase growth of remaining trees and accelerate process of stand development [76].Berrill et al. [77] modeled young plantations with and without thinning and found thinning put stands on a more rapid trajectory towards the desired old forest reference condition.One justification of restoration of old redwood forests is to reduce the vulnerability of younger stands to fire.Although thinning creates activity fuels, these hazards are short-lived after thinning.For example, O'Hara et al. [78] measured slash depth after restoration thinning in young stands at the northern end of the redwood range and found they were negligible ten years after thinning.In other parts of the range, these fuels might be more persistent, but thinning appears to have beneficial long-term effects as a fuel reduction strategy.
Another silvicultural treatment in young redwood stands is pruning.Artificial pruning can increase clearwood production and possibly increase the proportion of more valuable heartwood [79].A potential drawback is the production of epicormics sprouts with severe pruning [80].There is no significant interest in pruning from industrial owners in the redwood region due to cost and uncertainties over future markets.However, there is international interest from countries where redwood plantations are being established [81].Pruning could be used to remove branches in restoration treatments to accelerate the formation of clear boles and their aesthetic values, although it may be cost prohibitive activity for the agencies typically conducting restoration.
The American black bear (Ursus americanus Pallas) can also be an obstacle to restoration in parts of the redwood range [82,83].The bears scrape off the bark and eat the cambium, thereby injuring or even girdling stems.Coast redwood stems will usually resprout below the girdle.Other conifer species can be killed by this activity.The bears have a preference for more vigorous trees [83,84] and may also prefer redwood to Douglas-fir [83,85].Thinning to achieve restoration objectives of faster tree growth and lower densities may increase bear damage by encouraging the development of more vigorous trees.The resprouting of girdled redwood stems is contrary to restoration objectives, as it results in an increase in tree density and smaller tree sizes.Repeated bear damage can potentially prevent a stand from achieving restoration objectives.

Older Redwood Stand Management
Redwood trees respond to the increased growing space provided by thinning in older stands [86].In Lindquist's Whiskey Springs thinning study at Jackson Demonstration State Forest, even-aged stands exhibited a consistent, strong growth response between the ages of 40 to 75 years to thinning [73,87].Periodic diameter growth increased with severity of basal area reduction (Figure 5).The lowest retention of basal area resulted in the greatest periodic diameter growth and resulted in the biggest trees.The most severe thinning also exhibited a statistically significant loss in volume production primarily caused by the leave trees not fully occupying the site [86].Thinning operations accelerate the development of larger diameter redwood trees [73,86,87], promoting old forest structural features beneficial for creation of wildlife habitat [8].Variable retention silviculture is an alternative to clearcutting in redwood stands that also imparts some features of old forest stands.Berrill and O'Hara [88] used the Multiaged Stocking Assessment Model (MASAM) [89] to model a range of growing space occupancy by understory and overstory cohorts across dispersed retention management regimes.Manipulating overstory density had little effect on final stand volume but did have a major effect on the volume increment for both overstory and understory components.Greater overstory density increased the volume increment of the overstory component while decreasing the volume increment of the understory, reflecting a reduction in growing space available to the understory.These modeling results indicated that variable retention systems can provide old forest features while providing timber production.Variable retention therefore offers a restoration option for linking old forests in managed landscapes.
Most restoration strategies will attempt to develop or accentuate multiaged structures in redwood forests since natural even-aged stands are rare [12].Regeneration is often more of a consideration in multiaged redwood silviculture than in even-aged stands [35].Aggressive redwood crown development will rapidly occupy new growing space and suppress regeneration if the allotted growing space is not sufficient for regeneration to reach the canopy.To account for the aggressive crown development, it is recommended to substantially reduce density at the end of the multiaged cutting cycle.This maximizes understory vigor and productivity, as redwood stands can quickly return to pre-harvest crown closure after a light harvest.Heavier harvests on longer cutting cycles create stands with lower densities and larger trees, while allowing regeneration to develop [35,36].

Restoration
Restoration objectives in redwood forests are primarily focused on restoring old forest structures.Within this old forest umbrella, other objectives include increasing tree and stand growth to develop large trees and structural complexity, developing complex spatial patterns, maintaining genetic diversity and complex clonal patterns, controlling invasive plants, and others.Prescriptions are therefore highly variable depending on the site and stand conditions, and the specific objectives of the stand.Hence, it is difficult to generalize about restoration strategies across the region.Nevertheless, forest restoration in the redwood region is active with hundreds of hectares undergoing treatment.The primary form of restoration is directing young stands towards old forest structures with early stand management treatments such as thinning.Thinning has a demonstrated effectiveness for increasing tree growth rates (Figure 5; [73-75,86,87]), which is a primary objective of many restoration treatments.For example, the Del Norte Coast Redwood State Park at the northern edge of the redwood range, is a former industrial timberland converted to park status.Young stands Variable retention silviculture is an alternative to clearcutting in redwood stands that also imparts some features of old forest stands.Berrill and O'Hara [88] used the Multiaged Stocking Assessment Model (MASAM) [89] to model a range of growing space occupancy by understory and overstory cohorts across dispersed retention management regimes.Manipulating overstory density had little effect on final stand volume but did have a major effect on the volume increment for both overstory and understory components.Greater overstory density increased the volume increment of the overstory component while decreasing the volume increment of the understory, reflecting a reduction in growing space available to the understory.These modeling results indicated that variable retention systems can provide old forest features while providing timber production.Variable retention therefore offers a restoration option for linking old forests in managed landscapes.
Most restoration strategies will attempt to develop or accentuate multiaged structures in redwood forests since natural even-aged stands are rare [12].Regeneration is often more of a consideration in multiaged redwood silviculture than in even-aged stands [35].Aggressive redwood crown development will rapidly occupy new growing space and suppress regeneration if the allotted growing space is not sufficient for regeneration to reach the canopy.To account for the aggressive crown development, it is recommended to substantially reduce density at the end of the multiaged cutting cycle.This maximizes understory vigor and productivity, as redwood stands can quickly return to pre-harvest crown closure after a light harvest.Heavier harvests on longer cutting cycles create stands with lower densities and larger trees, while allowing regeneration to develop [35,36].

Restoration
Restoration objectives in redwood forests are primarily focused on restoring old forest structures.Within this old forest umbrella, other objectives include increasing tree and stand growth to develop large trees and structural complexity, developing complex spatial patterns, maintaining genetic diversity and complex clonal patterns, controlling invasive plants, and others.Prescriptions are therefore highly variable depending on the site and stand conditions, and the specific objectives of the stand.Hence, it is difficult to generalize about restoration strategies across the region.Nevertheless, forest restoration in the redwood region is active with hundreds of hectares undergoing treatment.The primary form of restoration is directing young stands towards old forest structures with early stand management treatments such as thinning.Thinning has a demonstrated effectiveness for increasing tree growth rates (Figure 5; [73][74][75]86,87]), which is a primary objective of many restoration treatments.For example, the Del Norte Coast Redwood State Park at the northern edge of the redwood range, is a former industrial timberland converted to park status.Young stands were established after clearcutting as high density plantations with little spatial diversity.Some of these were replanted with Douglas-fir.Forms of variable-density thinning (VDT; [90]) are being tested as a means to improve diversity and increase the proportion of redwood to historic level [81,87].VDT attempts to diversify stand structures by thinning stands to have greater horizontal, vertical, or species diversity.Methodologies for VDT are developing and many alternative methods have been tested (e.g., [91][92][93][94][95]).Generally, the more successful methods for creating stand complexity, the more difficult they are to implement [93].
Although redwood stands are highly resilient to fire, disturbances such as fire are generally considered undesirable in redwood forests that are in parks or adjacent to vulnerable natural resources or communities.Reducing fuels and making stands more resistant to catastrophic fire is another restoration objective in redwood.These treatments also involve thinning to increase spacing between trees, decrease continuities of crown fuels, and eliminate overly dense, stagnated areas within stands [85].
Another restoration objective is to reduce the amount of tanoak where previous harvesting practices focused on the removal of redwood and other conifers.These former treatments often resulted in tanoak-dominated stands or stands with increased proportions of tanoak.Restoration treatments may vary depending on the extent of the tanoak.In some cases, stands may be clearcut and replanted to redwood or other conifer species, with herbicides used to control sprouting tanoak [96][97][98].In other cases, herbicides may be injected into tanoak stems that are left to die in place.
Many restoration activities are focused on managing invasive organisms.The presence of sudden oak death (SOD; Phytopthora ramorum S. Werres, A.W.A.M. de Cock), an invasive forest pathogen, has resulted in widespread mortality of tanoak in redwood forests since 2000 [99].In the southern portion of the redwood range, Beh et al. [100] found that higher occurrences of California bay laurel, a non-susceptible host, were associated with higher rates of SOD in areas that had experienced high severity wildfire.SOD affects tanoak in redwood stands and infected stems typically exhibit rapid mortality.Tanoak decline, as a result of SOD, results in aggregated stem spatial patterns [101] and disproportionately affects larger tanoak trees [102].After SOD-induced tanoak mortality, redwood stems are unlikely to establish new stems in these recently formed gaps because of the poor sexual reproduction.Small gaps may be captured by overstory redwood stems through crown expansion, rather than through regeneration of tanoak [48].Larger gaps may be colonized by redwood sprouts or seedlings or regeneration of other species.Residual redwoods and other SOD-resistant species may benefit in increased growth [48].Tanoak mortality also results in higher fuel loads compared to stands unaffected by SOD.This increased fuel level alters wildfire behavior and results in increased redwood mortality-at least at the southern extent of the redwood range-overwhelming the characteristic fire resistance of the species [103].
One of the more prominent groups of invasive plants are various species of eucalyptus (Eucalyptus spp.) that have either spread into redwood forests or were purposely introduced.Although these eucalyptus trees are not competitive with redwood on better sites, on the margins of the redwood range, they can be more difficult.Effective treatments may involve cutting and herbicide use to kill the tree and subsequent sprouts.Disturbances may allow exotic plant species to become established in coast redwood stands.Larger canopy openings and more intense understory light conditions facilitate the establishment of exotic, invasive species [104].Pampas grass (Cortaderia selloana (Schult.& Schult.f.) Asch.& Graebn.),jubata grass (Cortaderia jubata (lem.)Stapf), and Scotch (Cytisus scoparius (L.) Link) and French broom (Genista monspessulana (L.) L.A.S. Johnson) are exotic species that are typically controlled with mechanical removal and herbicide.Such species typically persist along roadways and near stand boundaries, in high light environments.DiTomaso et al. [105] suggested that herbicide alone may be as effective as mechanical removal with herbicide at controlling jubata grass.Given wind blown seed spread of several km, control of pampas and jubata grass will be ongoing.Likewise, roadside and in-stand control of broom species will also be a long-term process as seeds remain viable on the ground for more than a decade.Invasive species, such as pampas grass, jubata grass, and Scotch and French broom, may impede the regeneration and growth of native tree and shrub species, and their removal is a common restoration objective.
An alternative to active redwood restoration is to allow stand development to proceed without silvicultural intervention.For example, Russell et al. [11] advocated a "natural recovery", whereby self-thinning and other related processes are left to occur at their own rates.This "natural recovery" assumes a rate of self-thinning that restores clearcut stands to old forest stands in several centuries, but the old forest stands used for reference conditions [11,63] had unusually high densities (Table 1).Self-thinning is likely to take many centuries to achieve the densities of typical old forest stands.The reliance on simple chronosequences of even-aged stand development following clearcutting presumes an even-aged target old forest structure.However, the overwhelming preponderance of stand reconstruction data for old forest stands indicates these have complex multiaged structures [5,12,51,52].By neglecting to understand old forest age structure and density, Russell et al. [11] and Russell and Michels [63] have essentially created a comparison where the occurrence of a declining density trajectory is used to suggest natural development will achieve the correct target structure.However, they fail to demonstrate how structural complexity will be achieved in the old forest communities that are targeted.This error is further compounded by the exceptionally high densities assumed for old redwood forests.
Results from chronosequence studies (i.e., [11,63]) are most valid when each site in the sequence is identical except for age, and when sample sites have the same history "in both its biotic and abiotic components [106]".Redwood forest sites are highly variable (because of topography, moisture, latitude, soils), and floristic associations are also numerous across site type and disturbance history [19,23].Floristic associations should be compared among similar site types close in proximity if effects of disturbance are being evaluated [24].Potential differences in site-specific history may additionally compound any comparison between redwood stands on dissimilar sites.Hence, the chronosequence approach used by Russell et al. [11] and Russell and Michels [63] may have produced erroneous results because of the wide environmental gradients between distant sample locations.

Climate Change and Genetic Diversity
Two redwood traits may make the species more vulnerable to climate change.One is the uncommonness of sexual reproduction and the other is the long-lived nature of the species: both contribute to slow adaptation to change.The reliance of redwood on fog moisture is high.The species benefits from fog condensation [107], and is also capable of the direct foliage absorption of water and nitrogen [108,109].A 33% reduction in fog frequency in the last century [110] may have already increased the environmental stresses on redwood populations.The necessity of adapting to the limitation in moisture might be especially pronounced in the southern populations that are already experiencing stronger selective pressures.Douhovnikoff and Dodd [111] reported that fragmentation in the redwood populations south of the Sonoma-Mendocino border (south of 36.8 • N) may indicate that these populations will have difficulty adapting to the new climatic conditions and thus are a greater conservation priority.
Restoration strategies in a changing climate should include a broad set of tools, as it is not clear yet how species will adapt to the increase in the temperature [112].Multiple scenarios need to be considered.The already difficult process of migration of trees to new habitats, restricted to maximum distances of seed dispersal, is further complicated in redwood because the species is reliant on a sprouting reproduction strategy.The data from common garden trials might be used to evaluate the potential for the assisted migration in redwood.It may be possible to shift the range of the species further north, along the coastline of Oregon, where moisture conditions will be suitable for species survival and reproduction.Another strategy may be enrichment planting of drier and warmer provenances to diversify the gene pool of existing redwood stands and increase their resilience to climate change.However, with either assisted dispersal or blending of provenances, active management of redwood forests will be necessary.

Conclusions
Redwood is an unusual conifer species, primarily because of its prolific sprouting ability, but also because of its longevity and resilience to disturbance.This sprouting ability provides rapid regeneration after harvest, but is spatially aggregated in cutover stands.Redwood may also be the most productive conifer species given rapid individual tree growth rates and its capacity to tolerate high stocking levels.Challenges to management include the social scrutiny that accompanies managing of such an iconic species, and the difficulties in obtaining full stocking of desirable species.For this latter challenge, interplanting of redwoods and aggressive control of shrubs and competing plants are standard operating procedures.Redwood sprouts can also be persistent in their aggregated spatial patterns, thereby necessitating thinning to increase growth and form more random spatial patterns.
Restoration of old forest stand structures is an increasingly common objective in the redwood region as public agencies and other entities attempt to increase the total amount of these structures and to consolidate blocks of old forest structures to provide connections or linkages.Most redwood restoration approaches recognize the complex age structures and the unique clonal patterns of old redwood forests and are using silvicultural treatments that will accentuate stand complexity.Many of these efforts have used variable-density thinning to promote more variable stand structures, decrease highly aggregated spatial patterns in young stands, and encourage more diversity in age class structure.Restoration treatments in existing stands have recognized the value of reducing density and its effect on increasing tree growth rates and developing structural complexity.Although the process of restoration of old forest redwood structures may take centuries, the value of accelerating this process to achieve regional goals of greater amounts and improved connections between old forests have become well-recognized.Simply stated, an active management approach is generally viewed as the best way to restore old redwood forests and the ecological values they provide.
The complexity of redwood forests creates many challenges for management and restoration and there is much to be learned.Given the longevity of the species, our present understanding of the species is only a snapshot in time.Longer term study with fluctuations in climate will yield new understandings.Additional knowledge of other reactions to a changing climate, particularly on the limits of the redwood range, will be especially helpful.Disturbance dynamics and effects on long-term stand structure are also topics for future exploration.The potential of prescribed fire as a restoration tool is also unexplored.Finally, landscape-level concerns related to connectivity of isolated patches of old redwood forest have been the impetus for much of the restoration activity to date.More research is needed to examine these landscape patterns and the different types of stand structures that can meet these needs.

Figure 1 .
Figure 1.Young even-aged redwood forest demonstrating strongly aggregated spatial pattern due to origin of these trees from stump sprout clumps.Photograph taken in Alameda County, CA, USA.

Figure 1 .
Figure 1.Young even-aged redwood forest demonstrating strongly aggregated spatial pattern due to origin of these trees from stump sprout clumps.Photograph taken in Alameda County, CA, USA.

Figure 2 .
Figure 2. Percent survivorship per redwood clump over a five-year period and the percent above canopy light (PACL).Sprout mortality increases with declining light regime (adapted from [35]).

Figure 2 .
Figure 2. Percent survivorship per redwood clump over a five-year period and the percent above canopy light (PACL).Sprout mortality increases with declining light regime (adapted from [35]).

Forests
of better spatial arrangements of existing redwood trees.Rotations are approximately 50 years, the lowest rotation length permitted in California[68].

Figure 3 .
Figure 3. Two-year-old clearcut showing redwood sprout clumps (some of which are indicated with arrows) and spaces between them.Photograph taken in Humboldt County, CA, USA.

Figure 3 .
Figure 3. Two-year-old clearcut showing redwood sprout clumps (some of which are indicated with arrows) and spaces between them.Photograph taken in Humboldt County, CA, USA.

Figure 4 .
Figure 4. Volume increment from 2001-2012 by treatment (average spacing in meters) showing the added increment from ingrowth (dark blue).The treatments included two types of controls (C = no thinning; CV = no thinning and herbicide treatment of shrubs), and the target spacings from 1.2 to 7.3 m (4 to 24 feet) (adapted from [75]).

Figure 4 .
Figure 4. Volume increment from 2001-2012 by treatment (average spacing in meters) showing the added increment from ingrowth (dark blue).The treatments included two types of controls (C = no thinning; CV = no thinning and herbicide treatment of shrubs), and the target spacings from 1.2 to 7.3 m (4 to 24 feet) (adapted from [75]).

Forests 2017, 8 , 144 10 of 18 Figure 5 .
Figure 5. Quadratic mean diameter development with stand age from the Whiskey Springs thinning trial at Jackson Demonstration State Forest.Percent values refer to the residual basal area as compared to the value before thinning at age 40.Initial basal area was approximately 92 m 2 ha −1 [87].

Figure 5 .
Figure 5. Quadratic mean diameter development with stand age from the Whiskey Springs thinning trial at Jackson Demonstration State Forest.Percent values refer to the residual basal area as compared to the value before thinning at age 40.Initial basal area was approximately 92 m 2 ha −1 [87].

Table 1 .
[54]d density of old redwood forests showing upper canopy redwood trees and total density (adapted from[54]).Studies varied in minimum diameters reported and in definitions of upper canopy trees.