Draft - University of Toronto T-Space · Draft 2 16 Abstract 17 We compare early seral development...

Draft

Early seral pathways of vegetation change following repeated, short interval high-severity wildfire in a low

elevation mixed-conifer-hardwood forest landscape of the Klamath Mountains, California

Journal: Canadian Journal of Forest Research

Manuscript ID cjfr-2019-0161.R1

Manuscript Type: Article

Date Submitted by the Author: 12-Aug-2019

Complete List of Authors: McCord, Millen; Humboldt State University, Department of Biological SciencesReilly, Matthew; Humboldt State University, Department of Biological SciencesJules, Erik; Humboldt State University, Biological SciencesButz, Ramona; USDA Forest Service Pacific Southwest Region, ; Humboldt State University, Forestry and Wildland Resources

Keyword: reburn, early seral vegetation, hardwood resilience, Douglas-fir, state change

Is the invited manuscript for consideration in a Special

Issue? :Not applicable (regular submission)

https://mc06.manuscriptcentral.com/cjfr-pubs

Canadian Journal of Forest Research

Draft

1

1 Title: Early seral pathways of vegetation change following repeated, short interval high-severity

2 wildfire in a low elevation mixed-conifer-hardwood forest landscape of the Klamath Mountains,

3 California

4 Running Title: Early seral pathways of vegetation change

5

6 Authors: Millen McCord1, Matthew J. Reilly1*, Ramona J. Butz2, and Erik. S. Jules1

7

8 1Department of Biological Sciences, Humboldt State University, Arcata, CA 95521

9 2USDA Forest Service Pacific Southwest Region, 1330 Bayshore Way, Eureka, CA 95501, and

10 Department of Forestry and Wildland Resources, Humboldt State University, Arcata, CA 95521

11 *corresponding author: Matthew Reilly (email: [email protected])

12

13

14

15

Page 1 of 34



Draft

2

16 Abstract

17 We compare early seral development between stands subject to single and repeated high-severity

18 wildfire in low elevation mixed conifer-hardwood forests in the Klamath Mountains, California,

19 USA. We used a before-after-control-impact (BACI) approach to assess changes in the density of

20 conifer regeneration and the cover of multiple components of vegetation structure (conifers,

21 hardwoods, shrubs, forbs, and graminoids) to compare pathways of seral development between

22 plots that burned once and twice. Fifty-three field plots were established six years following

23 high-severity fire in 2004. Nineteen of these experienced a second high-severity wildfire 11

24 years later (2015), and all plots were remeasured in 2016/2017. Conifer regeneration following

25 the first fire was abundant, but was greatly reduced by the second fire. Plots that did not reburn

26 increased in conifer, hardwood, and shrub cover, while those that reburned increased in forb

27 cover and decreased in cover of woody vegetation. Despite conifer loss, we found little evidence

28 of shifts to non-forested states following repeated fire due to resilience of resprouting

29 hardwoods. Our results indicate that repeated high-severity fire has the potential to protract early

30 seral development and catalyze transitions from mixed-conifer-hardwood forest to hardwood-

31 dominated early seral conditions.

32

33 Key Words: reburn, state-change, Douglas-fir, early seral vegetation, hardwood resilience,

34 BACI, compound disturbance

35

36

37

38

Page 2 of 34



Draft

3

39 Introduction

40 Larger and more frequent wildfires are altering landscapes throughout forested regions of

41 western North America (Westerling et al. 2006, Miller et al. 2012, Dennison et al. 2014).

42 Increased stand density and fuel loading from a century of fire exclusion, coupled with longer

43 fire seasons, more extreme fire weather days, and drier fuels (Liu et al. 2012, Abatzoglou and

44 Williams 2016, Davis et al. 2017) are expected to increase the potential for fire activity in the

45 coming decades. Although studies from several forest types across western North America

46 demonstrate rapid regeneration of conifers following a single stand-replacing fire (Donato et al.

47 2009a, Kemp et al. 2016, Shatford et al. 2007), there is concern about potential transitions to

48 non-forested states following repeated stand-replacing fire at short intervals. Plant communities

49 with species that are adapted to a particular fire regime may experience a shift in species

50 composition with alterations to the frequency and severity of fire (Bond and Keeley 2005, Odion

51 et al. 2010). Resilience (sensu Holling 1973) and recovery back to a pre-disturbance state, as

52 opposed to shifting to an alternative state, following stand-replacing events will be important to

53 maintain the critical structures and functions of forested ecosystems. However, there are

54 relatively few field studies that document changes in the structure and composition of forests

55 following short-interval, repeated high-severity wildfire (Donato et al. 2009b).

56 The potential conversion from forest to alternative non-forest states has received

57 considerable attention and conceptual development in recent years (Odion et al 2010, Enright et

58 al. 2015, Miller et al. 2019). Obligate-seeding species may experience local extirpation,

59 particularly slow-maturing species, as the time between disturbances decreases and the in-situ

60 seedbank is depleted (Enright et al. 2015). Short intervals between high-severity fires may shift

61 from dominance of obligate-seeding species (e.g., conifers) to alternative ecological states, such

Page 3 of 34



Draft

4

62 as meadows, shrublands, or hardwood forests if regeneration of obligate-seeding species is killed

63 before reaching reproductive maturity (Fig. 1). These states may be self-perpetuating provided

64 that fires continue to occur at a frequency high enough to prevent the recruitment of

65 reproductively mature trees, or those large enough to develop traits associated with fire

66 resistance. Dispersal and recruitment limitation may leave a previously forested area in a state of

67 prolonged arrested succession (Acacio et al. 2007, Tepley et al. 2017, Shrive et al. 2018), and

68 persistent large openings in forested landscapes resulting from short-interval repeated, high

69 severity fires have been observed in several forest types around the world (Bowman et al. 2016,

70 Coop et al. 2016, Enright et al. 2015, Fairman et al. 2016).

71 The Klamath Mountains of northern California are renowned for their ecological

72 diversity (Whittaker 1960) and serve as a model system for conceptual development of

73 ecological state changes following repeated high-severity fire (Odion et al. 2010, Miller et al.

74 2019). This region is historically associated with a mixed-severity fire regime characterized by a

75 complex mosaic of patches at all levels of severity (Halofsky et al. 2009). Patches of stand-

76 replacing fire initially convert the mixed-conifer-hardwood forests of the Klamath Mountains

77 into early seral vegetation mostly consisting of basally-resprouting hardwoods and shrubs that

78 regenerate via resprouting or a persistent seed bank (Shatford et al. 2007). Conifers often have

79 strong post-fire recruitment that will eventually overtop other vegetation after a single stand-

80 replacing fire (Donato et al. 2009a, Halofsky et al. 2011). Drought has increased in recent years

81 (Diffenbaugh et al. 2015) and models of future climate and vegetation change project major

82 changes in the 21st century (Barbero et al. 2015, Serra-Diaz et al. 2018), but studies

83 documenting early seral pathways following fire are needed to assess the potential for vegetation

84 change and transitions to non-forest states with repeated wildfire in this region (Prichard et al.

Page 4 of 34



Draft

5

85 2017). Despite this, there are few studies documenting vegetation change following repeated,

86 short-interval wildfire as studies with pre- and post-reburn data for such events are rare.

87 We assessed the empirical support for a conceptual model depicting different pathways of

88 early seral development following single and repeated wildfires (Fig. 1). We leveraged a unique

89 data set with observations before and after a high-severity wildfire that occurred eleven years

90 following a previous fire in low elevation mixed conifer-hardwood forests of the Klamath

91 Mountains. Our study objectives were to examine how conifer regeneration was affected by

92 repeated high-severity fire in our study area and to assess whether there is evidence to support an

93 ecological type change from a conifer-dominated forest to a non-forest or hardwood-dominated

94 forest. We compared (1) changes in the density of conifer regeneration; (2) changes in multiple

95 components of vegetation structure including the cover of conifers, hardwoods, shrubs, forbs,

96 and graminoids; and (3) multivariate trajectories of vegetation change. The results of this study

97 will provide important knowledge regarding the potential for transitions to non-forested states

98 following short -interval high-severity fire and the ecological implications of these compound

99 disturbances.

100101 Methods

102 Study Site

103 The Klamath Mountains of northern California and southern Oregon are a hotspot of

104 biological diversity characterized by tremendous environmental and topographic complexity.

105 The region has a Mediterranean climate, with warm, dry summers and cool, wet winters. Fire has

106 long been a driving force in shaping vegetation patterns of the Klamath Mountains (Whitlock et

107 al. 2004, 2008), and has shaped the evolutionary pathways of individual species as well as

108 landscape-level patterns of low density, multi-aged stands (Shatford et al. 2007, Whitlock et al.

Page 5 of 34



Draft

6

109 2004). Summer thunderstorms are a common source of ignition (Odion et al. 2010, Skinner et al.

110 2006) and Native Americans historically managed fire in the area (Whitlock et al. 2015).

111 This region is primarily associated with a mixed-severity fire regime (Halofsky et al.

112 2009). During the presettlement period prior to fire exclusion, the fire return interval for our

113 study vegetation type, Douglas-fir/hardwood ranged from 12-19 years (Taylor and Skinned

114 1998, Taylor and Skinner 2003), though a millennium-scale lake-sediment charcoal chronology

115 nearby showed periods of less frequent fires and fire-free intervals of up to 180 years over the

116 last 2,000 years (Colombaroli and Gavin 2010). Since 1987, wildfires burned approximately

117 30% of the area in northern California occupied by Douglas-fir or mixed conifer forest types,

118 with ~10% of that area having burned at high severity (regionally calibrated threshold RdNBR

119 >574; Miller et al. 2009, Miller at al. 2012).

120 Our study site was located in the southcentral portion of the Klamath Mountains,

121 northwest of Hyampom, California (Fig. 2). The area is comprised of low elevation (500-1000

122 m) mixed-conifer and hardwood forests. Douglas-fir (Pseudotsuga menziesii) is the most

123 common and dominant conifer, but other conifer species including ponderosa pine (Pinus

124 ponderosa), Jeffrey pine (P. jeffreyi), sugar pine (P. lambertiana), and incense-cedar

125 (Calocedrus decurrens) also occur. All of these conifer species are obligate seeders and rely on

126 seed from surviving trees for regeneration. The hardwood component has a relatively high

127 diversity of broadleaf trees and shrubs compared to other conifer-dominated regions. Evergreen

128 and deciduous hardwoods are common in our study forest, especially tanoak (Notholithocarpus

129 densiflorus), Pacific madrone (Arbutus menziesii), canyon live oak (Q. chrysolepis), California

130 black oak (Quercus kelloggii), and Oregon white oak (Q. garryana). All hardwood species

Page 6 of 34



Draft

7

131 resprout vigorously after fire. Common shrub species include California lilac (Ceanothus spp.),

132 manzanita (Artcostaphylos spp.), and poison oak (Toxicodendron diversilobum).

133 The Sims Fire was ignited by a fallen power line in July 2004 and burned approximately

134 1,630 ha during moderate drought conditions according to the Palmer Drought Severity Index

135 (PDSI) (Appendix Fig. 1). This landscape was largely departed from presettlement fire regimes

136 documented in the fire scar and tree ring records with no evidence of fire in at least the last 100

137 years. The fire included an approximately 700 ha patch of high- and very high-severity fire

138 (RdNBR >750) (Fig. 2). The first few years following the Sims Fire were characterized by

139 warmer than normal temperatures. Although the first year post-fire had unusually high

140 precipitation, the third and fourth years post-fire were characterized by moderate drought. The

141 Saddle Fire was ignited by lightning in June 2015 and burned 624 ha. Most of the Saddle Fire

142 burned within the footprint of the previous Sims Fire, and a large portion of the reburned

143 occurred in an approximately 300 ha patch of high- and very high-burn severity (Fig. 2). The fire

144 occurred following three years of moderate drought. Climatic conditions in the two years

145 following the Saddle Fire were warmer and drier than normal and characterized by moderate

146 drought (Appendix Fig. 1).

147 Field methods

148 In 2010, six years after the 2004 Sims Fire, a series of one hundred 60 m² (0.006 ha)

149 circular plots were established on a 200 m grid system to monitor regeneration of the area burned

150 in the Sims Fire. Some plots were not measured or relocated slightly due to inaccessibility. Since

151 we were interested in the effects of repeated, high-severity fire, we limited our analysis to the

152 fifty-three plots that experienced complete mortality of all trees and RdNBR values >574 (Fig. 2)

153 in the 2004 Sims Fire. A total of thirty-four plots within the once-burned Sims Fire were

Page 7 of 34



Draft

8

154 resampled in the summer of 2016. The nineteen plots burned again in the Saddle Fire were

155 resampled again in the summer of 2017 (see Appendix 1 for a timeline of fire occurrence and

156 sampling dates).

157 All plots were sampled according to an established regeneration plot protocol from the

158 US Forest Service Common Stand Exam (USDA Forest Service 2007). Field measurements at

159 both the first (2010) and second (2016/2017) samplings included ocular estimates of the percent

160 cover of multiple components of vegetation structure including conifers, hardwoods, shrubs,

161 forbs, and graminoids. As we only sampled areas with complete overstory mortality, all conifers

162 on our plots established after the Sims Fire. The age for each individual was determined by

163 counting bud scars or branch whorls and recorded. The technique of aging by bud scars or

164 branch nodes is not entirely accurate and tends to underestimate age (Hankin et al. 2018, Urza

165 and Sibold 2013). However, Hankin et al. (2018) found that the most accurate estimates of tree

166 age by this method were in California Douglas-fir, our primary study tree. In Urza and Sibold

167 (2013), the most accurate aging for Douglas-fir was between 8-11 years of age (approaching

168 100% accuracy at 10 years), roughly the age of most of our conifer recruits.

169 Data analysis

170 We used a before-after-control-impact (BACI) study design to investigate the impacts of

171 repeated high-severity wildfire. This approach focuses on comparing the change in a response

172 variable over time. In our case, we compare the change in cover between those samples

173 experiencing the impact of a second, short interval burn with those that experienced an initial

174 burn but not a second burn. Comparing change standardizes response variables, incorporates the

175 multi-temporal aspect of the data, and facilitates interpretation (van Mantgem et al. 2001). By

176 focusing our statistical analysis on the change in individual components of vegetation structure

Page 8 of 34



Draft

9

177 rather than on the cover before and after the second fire we were better able to characterize the

178 impact of the second fire and quantify the dynamic nature of early seral vegetation structure.

179 We first checked for spatial autocorrelation in the cover of each component of vegetation

180 structure and the density of conifer and hardwood regeneration. Spatial autocorrelation may

181 result in violation of the assumption of independence among observations, which can increase

182 the rate of Type I errors, or the false rejection of a null hypothesis of no effect (Dormann et al.

183 2007). We used the correlog function in the ncf package in R (R Core Team 2018) to calculate

184 Moran’s I to assess the correlation between observations at spatial lags from 200 to 1000 m in

185 each group of plots and for each time period. The value of Moran’s I approaches 1 when

186 observations close to each other are more similar and exhibit positive spatial autocorrelation,

187 while values approaching -1 are indicative of negative autocorrelation. Values close to zero are

188 spatially independent.

189 We used t-tests to compare changes in the abundance of individual components of

190 vegetation structure (van Mantgem et al. 2001) and test the null hypothesis of no difference

191 between means. These included percent cover of conifers, hardwoods, shrubs, graminoids, and

192 forbs. We also compared the density of conifer seedlings and hardwood resprouts. In cases where

193 the assumption of equal variances between each of the two samples was violated, we used

194 Welch’s unequal variances t-test. Welch’s t-test is considered more reliable when samples have

195 unequal variances and sample sizes (Derrick et al. 2016).

196 Finally, we compared multivariate trajectories of vegetation change using principal

197 components analysis (PCA) to reduce the dimensionality of plant functional groups into

198 gradients of early seral vegetation structure. PCA is based on linear relationships between

199 variables (McCune et al. 2002) and is effective at characterizing the inherently multivariate

Page 9 of 34



Draft

10

200 nature of vegetation structure (Reilly and Spies 2015). We performed the PCA in PC-ORD 7

201 (McCune and Mefford 2018) using the correlation coefficients between structural variables in

202 cross-products matrix. We included observations from all years in a single ordination. Cover

203 estimates were square root transformed to reduce the effect of dominant functional groups

204 (McCune et al. 2002). We used a randomization test to determine the significance of each axis

205 based on the number of randomizations with an eigenvalue for that axis that is equal to or larger

206 than the observed eigenvalue. In order to compare trajectories of change, we then plotted

207 multivariate vectors change between the first and second sampling for each plot with all

208 observations at the first sampling translated to the origin.

209 Results

210 Temporal and Spatial Patterns of Regeneration

211 Conifer recruits mostly consisted of Douglas-fir (>95%) but there were small components

212 of incense cedar, ponderosa pine, and Jeffrey pine. In 2010, six years after the Sims Fire, conifer

213 regeneration was much greater in the area that would subsequently burn in the Saddle Fire (Fig.

214 3). In both areas, conifer regeneration generally increased continuously until four years following

215 the Sims Fire, then decreased. The number of conifer recruits was much lower in 2016/17 in both

216 areas, but the decrease was far greater in the plots that were burned by the Saddle Fire. There

217 was some recruitment of conifers in the area that did not reburn between the first and second

218 sampling, but the total number of seedlings was greatly reduced by the second sampling.

219 Although some regeneration that established after the Sims Fire survived the Saddle Fire, we

220 observed very few seedlings in the plots that reburned by the remeasurement two years after the

221 Saddle Fire.

Page 10 of 34



Draft

11

222 Across both fires and samplings, hardwoods and shrubs generally made up most of the

223 vegetative cover (medians range from ~5 to 20% cover), followed by forbs graminoids, and

224 conifers which were primarily below 10% cover (Fig. 4). However, cover varied between

225 individual fires and sampling times. In plots affected by the Saddle Fire, conifer and hardwood

226 cover were highest at the first sampling, and decreased by the second sampling. Conifer and

227 hardwood cover were generally lower in plots only affected by the Sims Fire where they were

228 higher at the second sampling. Shrub cover was similar in both areas at the first sampling

229 (median ~10%), but was lower at the second sampling following the Saddle Fire and higher at

230 the second sampling in plots that only experienced the Sims Fire. Forb cover was similar in both

231 areas at the first sampling but was higher at the second sampling following the Saddle Fire and

232 lower at the second sampling in plots that only experienced the Sims Fire. Graminoid cover was

233 highest in the Sims Fire at the first sampling and lower in plots that did not experience another

234 fire, but was similar at both sampling times in the Saddle Fire.

235 We found no evidence of spatial autocorrelation at lags between 200 and 1,000 m in

236 either fire at either sampling period with the exception of conifer cover at the first sampling

237 following the Sims Fire in plots that did not reburn. Conifer cover showed strong negative (r = -

238 0.7) spatial autocorrelation at lags of 200 m in the Sims Fire at the first sampling.

239 Changes in vegetation structure

240 We found significant differences in the changes over time of most components of

241 vegetation structure including conifer density and cover, and cover of hardwoods and forbs

242 (Table 1, Fig. 5). Hardwood and shrub cover increased by slightly over 10% over time in plots

243 that did not experience a second fire, while forbs decreased by ~3%. Following the Saddle Fire,

244 the largest decrease occurred in conifer cover (~13%), though hardwood and shrub cover both

Page 11 of 34



Draft

12

245 decreased as well. Forb cover increased by ~7% following the Saddle Fire. Graminoid cover

246 decreased slightly over time in plots that experienced only the Sims fire as well as those burned

247 in the Saddle Fire, but the change was not significantly different between the two.

248 Trajectories of Vegetation Change

249 The PCA identified two major gradients in early seral vegetation structure that were

250 significant in the eigenvalue randomization tests (Table 2). The first significant component

251 accounted for 39% of the variance and had strong (r≥0.5) negative correlations with cover of

252 conifers and hardwoods, and a strong positive correlation with cover of graminoids and forbs.

253 The second significant component accounted for another 23% of the variance had a strong

254 negative correlation with cover of shrubs. A third non-significant axis accounted for an

255 additional 16% of the variance and had a strong negative correlation with conifer cover.

256 Multivariate vectors of change differed between plots burned once in the Sims Fire and

257 those that experienced a second burn by the Saddle Fire (Fig. 6). Direction and magnitude of

258 change was variable within both groups of plots, but change that occurred in the absence of a

259 second fire was primarily towards the negative end of Axis 1, which was strongly associated

260 with greater cover of conifers and hardwoods (Table 2). Plots that burned a second time in the

261 Saddle Fire primarily experienced shifts towards the positive end of Axis 1, which was positively

262 associated with cover of graminoids and forbs. Most of the Saddle plots moved towards the

263 positive end of Axis 2, which was strongly associated with decreased shrub cover and increased

264 cover of forbs and graminoids.

265 Discussion

266 This study provides a unique opportunity to observe the effects of repeated, short interval

267 wildfire on the trajectories of early seral development. We found a major reduction in the conifer

Page 12 of 34



Draft

13

268 component with repeated wildfire, which highlights the vulnerability of young, regenerating

269 post-fire conifer stands to fire. Our results indicate that resprouting hardwoods are a key

270 component of forest resilience to repeated high-severity fire in the Klamath Mountains. Although

271 we found no evidence of a shift towards non-forested states, our findings document a transition

272 away from mixed-conifer-hardwood forests to an alternative early seral trajectory towards

273 hardwood dominated forest. Collectively, our results indicate that repeated high-severity fire

274 prolongs early seral conditions and has the potential to be an important mechanism for future

275 conversion of conifer to hardwood-dominated forests in the Klamath Mountains.

276 Patterns of Conifer Regeneration

277 Our findings are consistent with several other studies documenting resilience of Douglas-

278 fir to a single high-severity fire in the Klamath Mountains (Shatford et al. 2007, Donato et al.

279 2009a, Tepley et al. 2017, Lopez Ortiz 2019), as well as other conifer species in different regions

280 of the western North America (Kemp et al. 2016). Resilience of Douglas-fir forests to stand-

281 replacing fire is likely related to the ability of this species to disperse relatively long distance.

282 Although some studies suggest that conifer dispersal decreases rapidly at 100 m from a seed

283 source (Kemp et al. 2016, Hermann and Lavender 1990, Rother and Veblen 2016), robust

284 regeneration of Douglas-fir has also been observed over 400 m away from a seed source (Donato

285 et al. 2009a, Donato et al. 2009b).

286 The large decrease in regeneration following the second fire is consistent with the

287 findings of other studies documenting the vulnerability of young stands to fire (Stevens-Rumann

288 and Morgan 2016). Previous studies found that high-severity burn patches tend to reburn at high-

289 severity when subsequent fires occurred (Odion et al. 2010), a pattern related to a combination of

290 low resistance of young conifers with thin bark to fire and higher snag density, woody debris,

Page 13 of 34



Draft

14

291 and pyrogenic shrub and hardwood cover following the first fire (Alexander et al. 2006,

292 Thompson et al. 2007, Thompson and Spies 2010, Coppoletta et al. 2016). Collectively, our

293 results and those of others support conceptual models depicting delayed recovery of obligate

294 seeding species (i.e. conifers) in the absence of a long enough period without fire to allow for

295 trees to establish and reach maturity (Odion et al 2010, Enright et al. 2015, Miller et al. 2019).

296 At two years postfire, the reburned areas had much lower regeneration than was present

297 two years after the first fire. Lack of regeneration following the second fire is likely due to a

298 combination of factors including dispersal and recruitment limitation associated with drought

299 conditions in the years following the second fire. Donato et al. (2009b) found ample Douglas-fir

300 regeneration two years following short interval reburn (15 years), but the patches of high-

301 severity fire were smaller (<100 ha) than in our study (~300 ha). Dispersal limitation may be

302 more of a factor following repeated fire than following a single fire as any in-situ seeds that

303 survived the first fire would no longer be present. In addition, low water availability may also

304 limit conifer regeneration (Hogg and Schwarz 1997, Dunne and Parker 1999, Johnstone et al.

305 2010, Dodson and Root 2013, Rother et al. 2015, Harvey et al. 2016, Tepley et al. 2017, Shive et

306 al. 2018). The second fire in 2015 occurred in the midst of a multiyear drought characterized by

307 lower precipitation and warmer temperatures than normal (Appendix Fig. 1). Although the

308 drought officially ended in 2015, it was still warmer and drier than the first year following the

309 Sims Fire (2005). Given that post-fire recruitment in Douglas-fir forests may continue for a

310 period of up to a decade, it is still too early to conclude after only two years that conifers will

311 never establish. However, it seems unlikely that they will regain dominance immediately given

312 the rapid recovery of hardwoods after repeated fire and slower height growth associated with

313 regeneration established at longer time lags following fire (Tepley et al. 2017).

Page 14 of 34



Draft

15

314 Alternative trajectories of early seral development

315 Our field-based study provides empirical evidence to test conceptual models of early

316 seral development following repeated wildfire in Douglas-fir forests of the Klamath Mountains

317 (Fig. 2). Conifer forests dominated by Douglas-fir appear to be resilient to a single high-severity

318 fire, but are subject to transitions from the conifer to hardwood dominated forests following short

319 interval repeated fire. Even after experiencing two very high-severity events, hardwoods

320 persisted and attained levels of cover that were almost equal to the levels observed before the

321 second fire. Resprouting hardwoods are an important mechanism of forest resilience in other

322 forests of northern California (Cocking et al. 2014) and other mixed conifer-hardwood forests of

323 North America (Coop et al. 2016, Hagan et al. 2016). Although hardwoods maintain high levels

324 of cover following repeated fires, cover alone may not reflect the potential for loss of vigor and

325 death of some resprouting clumps observed in other resprouting species after multiple short

326 interval fires (Fairman et al. 2016, Fairman et al. 2017).

327 Although we found limited conifer recruitment after the second fire, it is possible that the

328 area will eventually be able to return to a conifer-dominated forest in the absence of another

329 wildfire in the next few decades. In the single burn, total regeneration declined but recruitment

330 continued over 12 years at low rates of establishment. Kemp et al. (2016) found higher Douglas-

331 fir abundance on sites with a longer time since fire in the northern Rockies, and Tepley et al.

332 (2017) found that conifers made up a low but increasing percentage of the aboveground biomass

333 7-28 years post-fire in the Klamath Mountains. Thus, conifer recruitment may continue into the

334 future, but if so, then likely at a much greatly reduced rate over a longer period of time. In fire-

335 generated chaparral patches in the Sierra Nevada and southern Cascades, conifers eventually

336 overtopped shrubs after 50-70 years across much of the area in the absence of fire (Russell et al.

Page 15 of 34



Draft

16

337 1998, Lauveux et al. 2016). Likewise, many hardwood forests in this region have been

338 encroached by Douglas-fir during the 20th century due to fire exclusion (Long et al. 2018).

339 Additionally, regional heavy seed crop years for Douglas-fir happen every 3-12 years

340 (USFS 2012), and the years after the second fire could have been low seed-producing years.

341 Although resprouting broadleaf species grow more quickly and may outcompete conifers for

342 light and water in early seral conditions, studies have found that the mycorrhizal connections

343 formed between hardwoods, Arctostaphylos spp., and conifer seedlings after disturbance

344 facilitate seedling establishment (Borchers and Perry 1990, Horton et al. 1999, Simard 2009).

345 Ceanothus spp. fix nitrogen, which could contribute the important macronutrient to the soil and

346 help forest recovery after fire (Busse et al. 1996, Busse 2000). In conifer plantations where the

347 hardwood component has been removed or killed in post-logging treatments, forest resilience

348 may be low as hardwood species will have to reestablish from seed as opposed to resprouting.

349 Conclusions

350 The findings of this study are consistent with projections of future vegetation in the

351 Klamath Mountains. Both Lenihan et al. (2008) and Sheehan et al. (2015) projected shifts

352 towards hardwoods the Klamath Mountain in the coming decades using mechanistic ecosystem

353 models (MC2). Repeated, short-interval fires are likely to be an important pathway to ecosystem

354 type change under future disturbance and climate regimes. Climate change may compound these

355 effects and hasten the loss of conifer-dominated forest ecosystems (Tepley et al. 2017, Serra-

356 Diaz et al. 2018). However, projected losses of conifer forest will likely require drastic increases

357 in fire activity as only 3.5% of the Klamath Mountains burned twice at any severity during 1985-

358 2012 (Grabinski et al. 2017).

Page 16 of 34



Draft

17

359 Assessing the trade-offs among the ecological implications of large patches of high-

360 severity reburn is complex and requires consideration of a variety of ecological outcomes. There

361 is little evidence that the large patches of high-severity fire in our study were part of the

362 presettlement range of variability in this region (Spies et al. 2018) and high-severity wildfires

363 have negative effects on species dependent on dense, multi-layered old growth forests [e.g., the

364 Northern Spotted Owl (Strix occidentalis caurina)]. However, such disturbances are relatively

365 rare and may also enhance some aspects of landscape diversity given current rates of reburn in

366 this region. Twentieth century fire exclusion reduced the extent of biodiverse early seral and

367 open vegetation types such as meadows and chaparral in the Klamath Mountains (Skinner 1995).

368 Recent fires have reversed declining trends of early seral habitats (Phalan et al. 2019), enabled

369 the expansion of some declining early seral species (Reilly et al. 2019), and may foster

370 understory diversity. However, almost 20% of the understory cover in our study was composed

371 of non-native annual grasses and forbs which may further threaten conifer regeneration (Reilly et

372 al. In Review).

373 Despite a relatively small sample size and limitations due to differences in time since fire

374 and climatic conditions following the fires, this study provides an empirical evaluation of

375 conceptual models of repeated fire effects on transitions between alternative vegetation states

376 using field observations from pre- and post-reburn. Repeated wildfire in this region has the

377 potential to protract early seral development and catalyze a transition from mixed conifer-

378 hardwood to a hardwood-dominated early seral conditions characterized by greater cover of

379 forbs. Our results demonstrate the resilience of conifers (e.g. Douglas-fir) to a single high-

380 severity wildfire, as well as with the vulnerability of conifer regeneration to short, interval

381 reburn. As in other regions, hardwood species are extremely resilient to both single and repeated

Page 17 of 34



Draft

18

382 wildfire and will likely be a key component to forest resilience in the next century as fire activity

383 increases.

384 Acknowledgements

385 We would like acknowledge the USDA Forest Service Region 5 for funding, as well as

386 Vicki Alla, Addison Gross, Stefani Brandt, Taylor Farnum, Kevin Linowksi, and Kevin Welch

387 for their hard work in the field while enduring hot temperatures, steep terrain, and poison oak.

388

389

Page 18 of 34



Draft

19

390 Literature Cited

391 Abatzoglou, J. T., Williams, A. P. 2016. Impact of anthropogenic climate change on wildfire 392 across western US forests. Proceedings of the National Academy of Sciences, 113, 11770–393 11775.

394 Acacio, V., Holmgren, M, Jansen, P.A., Schrotter, O. 2007. Multiple recruitment limitations 395 causes arrested succession in Mediterranean cork oak systems. Ecosystems 10: 1220-1230.

396 Alexander, J. D., Seavy, N. E., Ralph, J., Hogoboom, B. 2006. Vegetation and topographical 397 correlates of fire severity from two fires in the Klamath-Siskiyou region of Oregon and 398 California. International Journal of Wildland Fire 15:237-245.

399 Barbero, R., Abatzoglou, J. T., Larkin, N. K., Kolden, C. A., & Stocks, B. 2015. Climate change 400 presents increased potential for very large fires in the contiguous United States. International 401 Journal of Wildland Fire, 24, 892–899.

402 Bond, W.J., Keeley, J.E. 2005. Fire as a global ‘herbivore’: the ecology and evolution of 403 flammable ecosystems. Trends in Ecology and Evolution 20:387–394.

404 Borchers, S. L., Perry, D. A. 1990. Growth and ectomycorrhiza formation of Douglas-fir 405 seedlings grown in soils collected at different distances from pioneering hardwoods in southwest 406 Oregon clear-cuts. Canadian Journal of Forest Research 20:712-721.

407 Bowman, D.M.J.S, G.J. Williamson, L.D. Prior, and B.P. Murphy. 2016. The relative importance 408 of intrinsic and extrinsic factors in the decline of obligate seeder forests. Global Ecology and 409 Biogeography 25: 1166-1172.410411 Busse, M. D., Cochran, P. H., Barrett, J. W. 1996. Changes in Ponderosa Pine Site Productivity 412 following Removal of Understory Vegetation. Soil Science Society of America Journal 60:1614-413 1621. 414415 Busse, M. D. 2000. Ecological Significance of nitrogen fixation by actinorhizal shrubs in interior 416 forests of California and Oregon. In Forest soil organisms. Powers, R. F., D. Hauxwell, and G. 417 Nakamura (eds.).USDA For. Serv. Gen. Tech. Rep. PSW. In press.418419 Cocking, M. I., Varner, J. M., Knapp, E. E. 2014. Long-term effects of fire severity on oak-420 conifer dynamics in the southern Cascades. Ecological Applications 24: 94-107.

421 Coop, J. D., Parks, S. A., McClernan, S. R., Holsinger, L. M. 2016. Influences of prior wildfires 422 on vegetation response to subsequent fire in reburned Southwestern landscape. Ecological 423 Applications 26: 346-354.

424 Coppoletta, M., Merriam, K. E., Collins, B. M. 2015. Post-fire vegetation and fuel development 425 influences fire severity patterns in reburns. Ecological Applications 26: 686-699.

426 Davis, R., Yang, Z, Yost, A., Belongie, C, Cohen, W. 2017. The normal fire environment 427 modeling environmental suitability for large forest fires using past, present, and future climate 428 normal. Forest Ecology and Management 390: 173-186.

Page 19 of 34



Draft

20

429 Dennison, P. E., S. C. Brewer, J. D. Arnold, and M. A. Moritz. 2014. Large wildfire trends in the 430 western United States, 1984-2011. Geophysical Research Letters 41:2928–2933.

431 Derrick, B., Toher, D., White, P. 2016. "Why Welchs test is Type I error robust" The 432 Quantitative Methods for Psychology. 12: 30–38.

433 Diffenbaugh, N. S., Swain, D. L., & Touma, D. 2015. Anthropogenic warming has increased 434 drought risk in California. Proceedings of the National Academy of Sciences, 112, 3931–3936.

435 Dodson, E. K., Root, H. T. 2013. Conifer regeneration following stand-replacing wildfire varies 436 along an elevation gradient in a ponderosa pine forest, Oregon, USA. Forest Ecology and 437 Management 302:163-170.

438 Donato, D. C., Fontaine, J. B., Campbell, J. L., Robinson, W. D., Kauffman, J. B., Law, B. E. 439 2009a. Conifer regeneration in stand-replacement portions of a large mixed-severity wildfire in 440 the Klamath–Siskiyou Mountains. Canadian Journal of Forest Research 39:823-838.

441 Donato, D. C., Fontaine, J. B., Robinson, W. D., Kauffman, J. B., & Law, B. E. 2009b. 442 Vegetation response to a short interval between high‐severity wildfires in a mixed‐evergreen 443 forest. Journal of Ecology 97: 142-154.

444 Dormann, C.F., McPherson, J.M., Araujo, M.B., Bivand, R., Bollinger, J., Carl, G, Davies, R.G., 445 Hirzel, A. Jetz, W., Kissling, W.D., Kuhn, I., Ohlemuller, R., Peres-Neto, P.R., Reineking, B., 446 Schroder, B., Schurr, F.M. and Wilson, R. Ecography. 2007. Methods to account for spatial 447 autocorrelation in the analysis of species distributional data: a review. 30: 609-628.

448 Dunne, J.A., Parker, V.T. 1999. Species-mediated soil moisture availability and patchy 449 establishment of Pseudotsuga menziesii in chaparral. Oecologia 119:36–45.

450 Grabinski, Z. S., R. L. Sherriff, and J. M. Kane. 2017. Controls of reburn severity vary with fire 451 interval in the Klamath Mountains, California, USA. Ecosphere 8(11):e02012. 452 10.1002/ecs2.2012

453 Fairman, T.A., C.R. Nitschke, and L.T. Bennett. 2016. Too much, too soon? A review of the 454 effects of increasing wildfire frequency on tree mortality and regeneration in temperate eucalypt 455 forests. Internationl Journal of Wildland Fire 25: 831-848.

456 Enright, N. J., Fontaine, J. B., Bowman, D. MJS., Bradstock R. A., Williams, R. J. 2015. Interval 457 squeeze: altered fire regimes and demographic responses interact to threaten woody species 458 persistence as climate changes. Frontiers in Ecology and the Environment 13: 265-272.

459 Hagan, D.L., Waldrop, T.A., Reilly, M.J., Shearman, T.M. 2015. Impacts of repeated wildfire in 460 long-unburned plant communities of the southern Appalachian Mountains. International Journal 461 of Wildland Fire 24:911-920.

462 Halofsky, J. E., Donato, D. C., Hibbs, D. E., Campbell, J. L. Cannon, M. D., Fontaine, J. B., … 463 Spies, T. A. 2011. Mixed-severity fire regimes: lessons and hypotheses from the Klamath-464 Siskiyou Ecoregion. Ecosphere 2:art40.

465 Harvey, B. J., Donato, D. C., & Turner, M. G. 2016. High and dry: post-fire tree seedling 466 establishment in subalpine forests decreases with post-fire drought and large stand-replacing 467 burn patches. Global Ecology and Biogeography 25:655–669.

Page 20 of 34



http://eprints.uwe.ac.uk/27232/27/p030.pdf

Draft

21

468 Hermann, R. K., Lavender, D. P. 1990. Pseudotsuga menziesii (Mirb.) Franco, Douglas-Fir, in 469 Burns, R. B. and Honkala, B. H. (tech. coordinators), Silvics of North America: 1. Conifers 470 (Agriculture Handbook 654), U.S. Dept. of Agriculture Forest Service, Washington, D.C.

471 Hogg, E. H., Schwarz, A. G. 1997. Regeneration of planted conifers across climatic moisture 472 gradients on the Canadian prairies: implications for distribution and climate change. Journal of 473 Biogeography 24:527-534.

474 Holling, C.S. 1973. Resilience and stability of ecological systems. Annual Review of Ecology 475 and Systematics 4:1–23.

476 Horton, T.R., Bruns, T.D., Parker, V.T. 1999. Ectomycorrhizal fungi associated with 477 Arctostaphylos contribute to Pseudotsuga menziesii establishment. Canadian Journal of Botany 478 77:93–102.

479 Johnstone, J. F., Hollingsworth, T. N., Chain, F. S., Mack, M. C. 2010. Changes in fire regime 480 break the legacy lock on successional trajectories in Alaskan boreal forest. Global Change 481 Biology 16:1281-1295.

482 Johnstone, J. F., Chapin, F. S. 2006. Fire Interval Effects on Successional Trajectory in Boreal 483 Forests of Northwest Canada. Ecosystems 9:268-277.

484 Kemp, K. B., Higuera, P. E., Morgan, P. 2016. Fire legacies impact conifer regeneration across 485 environmental gradients in the U.S. northern Rockies. Landscape Ecology 31:619–636.

486 Larson, A.J., Franklin, J.F. 2005. Patterns of conifer tree regeneration following an autumn 487 wildfire event in the western Oregon Cascades Range, USA. Forest Ecology and Management 488 218: 25-36.

489 Lauvaux, C. A., Skinner, C. N., Taylor, A. H. 2016. High severity fire and mixed conifer forest-490 chaparral dynamics in the southern Cascade Range, USA. Forest Ecology and Management 363: 491 74–85.

492 Lenihan,J.M., Bachelet, D., Neilson, R.P., and Drapek, R. 2008. Response of vegetation 493 distribution, ecosystems productivity, and fire to climate change scenarios for California. 494 Climatic Change 78: 215-230.

495 Long, J.W., Gray, A., Lake, F.K. 2018. Recent trends in large hardwoods in the Pacific 496 Northwest, USA. Forests 9:561

497 Liu, Y., Stanturf, J., Goodrick, S. 2012. Trends in global wildfire potential in a changing climate. 498 Forest Ecology and Management 259: 685-697.

499 Lopez Ortiz, M.J., Marcey, T., Lucash, M.S., Hibbs, D., Shatford, J.P.A., Thompson, J.R. 2019. 500 Post-fire management affects species composition but not Douglas-fir regeneration in the 501 Klamath Mountains. Forest Ecology and Management 432: 1030-1040.

502 McCune, B., J.B. Grace, Urban, D. 2002. Analysis of ecological communities. Gleneden Beach, 503 OR. MjM Software Design.

504 McCune, B. and M. J. Mefford. 2018. PC-ORD. Multivariate Analysis of Ecological Data. 505 Version 7.07

Page 21 of 34



Draft

22

506 Miller, J. D., Knapp, E. E., Key, C. H., Skinner, C. N., Isbell, C. J., Creasy, R. M., Sherlock, J. 507 W. 2009. Calibration and validation of the relative differenced Normalized Burn Ratio (RdNBR) 508 to three measures of fire severity in the Sierra Nevada and Klamath Mountains, California, USA. 509 Remote Sensing of Environment 113: 645-656.

510 Miller, J. D., C. N. Skinner, H. D. Safford, E. E. Knapp, and C. M. Ramirez. 2012. Trends and 511 causes of severity, size, and number of fires in northwestern California. Ecological Applications 512 22:184–203.513514 Miller, A.D, Thompson, J.R., Tepley, A.J., Anderson-Teixeira, K.J. 2019. Alternative stable 515 equilibria and critical thresholds created by fire regimes and plant responses in a fire-prone 516 community. Ecography 42: 55-66.

517 Odion, D. C., Mortitz, M. A., DellaSala, D. A. 2010. Alternative community states maintained 518 by fire in the Klamath Mountains, USA. Journal of Ecology 98: 96-105.

519 Phalan, B.T., Northrup, J.M., Yang, Z., Deal, R.L., Rousseau, J.S., Spies, T.A., Betts, M.G. 520 2019. Impacts of the Northwest Forest Plan on forest composition and bird populations. 521 Proceedings of the National Academy of Sciences 116: 3322-3327.

522 Pounden, E., Greene, D.F, Michaletz, S.T. 2014. Non-serotinous plants behave as aerial seed 523 bank species when a late-summer wildfire coincides with a mast year. Ecology and Evolution 524 4:10.1002/ece3.1247

525 Prichard, S. J., Stevens-Rumann, C. S., Hessburg, P. F. 2017. Tamm Review: Shifting global fire 526 regimes: Lessons from reburns and research needs. Forest Ecology and Management 396:217-527 233.

528 R Core Team. 2018. R: A language and environment for statistical computing. R Foundation for 529 Statistical Computing, Vienna, Austria. URL https://www.R-project.org/.530531 Reilly, M.J., and T.A. Spies. 2015. Regional variation in forest structure and development in 532 Oregon, Washington, and inland northern California, USA. Ecosphere 6:art192. 533 http://dx.doi.org/10.1890/ES14-00469.1534535 Reilly, M.J., V. Monleon, E.S. Jules, and R.J. Butz. 2019. Biogeographic patterns of population 536 structure and dynamics of a serotinous conifer, knobcone pine (Pinus attenuata L.). Forest Ecology 537 and Management 441: 182-191.538539 Reilly, M.J., McCord, M., Brandt, S., Linowski, K., Butz, R.J., and E.S. Jules. Invasion of non-540 native plant species following repeated high-severity wildfire in Klamath Mountains, northern 541 California, USA. In Review, Biological Invasions542543 Rother, M. T., Veblen, T. T. 2016. Limited conifer regeneration following wildfires in dry 544 ponderosa pine forests of the Colorado Front Range. Ecosphere 7: e01594.

Page 22 of 34



https://www.R-project.org/

http://dx.doi.org/10.1890/ES14-00469.1

Draft

23

545 Rother, M. T., Veblen, T. T., Furman, L. G. 2015. A field experiment informs expected patterns 546 of conifer regeneration after disturbance under changing climate conditions. Canadian Journal of 547 Forest Research 45:1607–1616.

548 Russell, W. H., McBride, J., & Rowntree, R. 1998. Revegetation after four stand-replacing fires 549 in the Lake Tahoe basin. Madrono~ 45:40–46.

550 Serra-Diaz, J., Maxwell, C., Lucash, M.S., Scheller, R.M., Laflower, Miller, A.D., Tepley, A.J., 551 Epstein, H.E., Andersen-Teixeira, Thompson, J.R. 2018. Disequilibrium of fire-prone forests sets 552 the stage for a rapid decline in conifer dominance during the 21st century. – Sci. Rep. 8: 6749.

553 Shatford, J. P. A., Hibbs, D. E., Puettman, K. J. 2007. Conifer Regeneration after Forest Fire in 554 the Klamath-Siskiyous: How Much, How Soon? Journal of Forestry 105:139-146.

555 Sheehan, T., Bachelet,D., and Ferschweiler, K. 2015. Projected major fire and vegetation 556 changes in the Pacific Northwest of the conterminous United States under selected CMIP5 557 climate futures. Ecological Modeling 317: 16-29.

558 Shive, K., H. Preisler, K.R. Welch, H.D. Safford, R.J. Butz, K. O’Hara, and S.L. Stephens. 2018. 559 From the stand scale to the landscape scale: predicting the spatial patterns of forest regeneration 560 after disturbance. Ecological Applications 28: 1626-1639.

561 Simard, S. W. 2009. The foundational role of mycorrhizal networks in self-organization of 562 interior Douglas-fir forests. Forest Ecology and Management 258:S95-S107.

563 Skinner, C.N. 1995. Change in spatial characteristics of forest openings in the Klamath 564 Mountains of northwestern California, USA. Landscape Ecology 10: 219-228.

565 Skinner, C. N., A. H. Taylor, J. K. Agee. 2006. Klamath Mountains bioregion. Pages 170-194 in 566 N. G. Sugihara, J. W. van Wagtendonk, K. E. Shaffer, J. O. Fites-Kaufman, and A. E. Thode, 567 editors. Fire in California's ecosystems. University of California Press, Berkeley, California, 568 USA.

569 Spies T.A., Hessburg, P.F., Skinner, C.N., Puettmann, K.J., Reilly, M.J., Davis, R.J., Kertis, J.A., 570 Long, J.W., Shaw, D.C. 2018. Chapter 3: Old growth, disturbance, forest succession, and 571 management in the area of the Northwest Forest Plan. In: T.A. Spies, P. Stine, R. Gravenmier, J. 572 Long, M.J. Reilly (Tech. Coords.), Synthesis of Science to Inform Land Management Within the 573 Northwest Forest Plan Area. General Technical Report PNW-GTR-370 Volume 1. USDA 574 Forest Service, Portland, OR, USA.

575 Stevens-Rumann, C., Morgan, P. 2016. Repeated wildfires alter forest recovery of mixed-conifer 576 ecosystems. Ecological Applications 26:1842-1853.

577 Taylor, A.H., Skinner, C.N. 1998. Fire history and landscape dynamics in a late-successional 578 reserve, Klamath Mountains, California, USA. Forest Ecology and Management 111: 285-301.

579 Taylor, A.H., Skinner, C.N. 2003. Spatial patterns and controls on historical fire regimes and 580 forest structure in the Klamath Mountains. Ecological Applications 13:704–719.

Page 23 of 34



Draft

24

581 Tepley, A. J., Thomspson, J. R., Epstein, H. E., Anderson-Teixeira, K. J. 2017. Vulnerability to 582 forest loss through altered postfire recovery dynamics in a warming climate in the Klamath 583 Mountains. Global Change Biology 10:4117-4132.

584 Thompson, J. R., Spies, T. A., Ganio, L. M. 2007. Reburn severity in managed and unmanaged585 vegetation in a large wildfire. Proceedings of the National Academy of Sciences 104, 10743–586 10748.

587 Thompson, J.R., Spies, T.A., 2010. Factors associated with crown damage following recurring 588 mixed-severity wildfires and post-fire management in southwestern Oregon. Landscape Ecology 589 25, 775–789.

590 USDA Forest Service. 2007. Common stand exam field guide. Natural Resource Information 591 System, Washington, DC. Available on-line at: http://www.fs.fed.us/nrm/fsveg/.

592 USDA Forest Service. 2012. Ch. 500: Planning for Cone and Seed Production. Forest Service 593 Handbook Northern Region (R1) Seed Handbook. USDA Forest Service, Missoula, MT, USA. 594 pp 31

595 van Mantgem, P., M. Schwartz, and MaryBeth Keifer. 2001. Monitoring fire effects for managed 596 burns and wildfires: Coming to terms with pseuoreplication. Natural Areas Journal 21:266-273.

597 Westerling, A. L., Hildalgo, H. G., Cayan, D. R., Swetnam, T. W. 2006. Warming and earlier 598 spring increase western U.S. forest wildfire activity. Science 313:940-943.

599 Whittaker, R. H. 1960. Vegetation of the Siskiyou Mountains, Oregon and California. Ecological 600 Monographs 30:279-338.

601 Whitlock, C., Skinner, C. N., Bartlein, P. J., Minckley, T., Mohr, J. A. 2004.Comparison of 602 charcoal and tree-ring records of recent fires in the eastern Klamath Mountains, California, USA. 603 Canadian Journal of Forest Research 34:2110-2121.

604 Whitlock, C., Marlon, J., Briles, C., Brunelle, A., Long, C., Bartlein, P. 2008. Long-term 605 relations among fire, fuel, and climate in the north-western US based on lake-sediment studies. 606 International Journal of Wildland Fire 17:72-83.

607 Whitlock, C., McWethy, D. B., Tepley, A. J., Veblen, T. T., Holz, A., McGlone, M. S., Perry, 608 G.L.W., Wilmshurst, J.M., Wood, S. W. 2015. Past and present vulnerability of closed-canopy 609 temperate forests to altered fire regimes: a comparison of the Pacific Northwest, New Zealand. 610 Bioscience 65:151-163.

Page 24 of 34



Draft

25

611 Table 1. Results from t-tests comparing the change in the density of conifer regeneration and 612 cover of multiple components of early seral vegetation structure in fifty-three 60 m2 plots subject 613 to single (n=34) and repeated high-severity wildfire (n=19).

Structural Component

Mean Sims (1x burned)

Sims and Saddle Fire (2x burned)

t-statistic Degrees of freedom p-value

Conifer density -3.5 -41.5 -2.29 19.2 0.03Conifer cover 2.1% -13.1% -2.59 19.5 0.018

Hardwood cover 11.1% -4.5% -3.06 47.5 0.004Shrub cover 10.2% -1.2% -2.46 45.6 0.018Forb cover -3.3% 6.8% 2.91 29.1 0.007

Graminoid cover -2.0% -1.6% -0.08 39.6 0.94614

615 Table 2. Percent of variance explained by the first two dimensions of a principal component 616 analysis along with correlation coefficients for cover of multiple components of early seral 617 vegetation structure in plots subject to single and repeated high-severity wildfire. Axes with a * 618 were significant (p=0.001) based on an eigenvector based randomization test.

Component Axis 1* Axis 2* Axis 3% of variance

explained 39% 23% 16%

r r rConifer cover -0.469 0.444 -0.743

Hardwood cover -0.822 0.160 0.152Shrub cover 0.153 -0.811 -0.472Forb cover 0.547 0.456 -0.108

Graminoid cover 0.843 0.253 -0.110619

620

Page 25 of 34



Draft

26

621 Figure 1. A conceptual model of possible vegetation state change after short-interval, high-622 severity fire in mixed conifer-hardwood forests of the Klamath Mountains in northern California.

623 Figure 2. Study location and patterns of burn severity (RdNBR) in the Sims (2004) and Saddle 624 Fire (2015). Inset in the top left shows the location of the Klamath Mountains in northern 625 California and southwestern Oregon, USA. The hatched area overlaying the Sims Fire perimeter 626 is the perimeter of the Saddle Fire which is presented in the inset in the bottom right corner.

627 Figure 3. Timing of conifer establishment for: a) the Sims Fire in 2010, b) after the Sims Fire and 628 prior to the Saddle Fire in 2010, and c) in 2016 following only the Sims Fire, and d) in 2017 629 following both the Sims Fire and the Saddle Fire.

630 Figure 4. Distribution of cover of conifers, hardwoods, shrubs, forbs, and graminoids in early 631 seral plots that experienced a second high-severity fire in 2015 (Saddle 2x burned) and plots that 632 only burned during the first fire in 2004 (Sims 1x burned). Fires and sampling times are 633 abbreviated as follows: SM1=Sims Time 1 (2010), SM2=Sims Time 2 (2016), SD1=Saddle 634 Time 1(2010), SD2=Saddle Time 2 (2017).

635 Figure 5. Comparison of changes in cover of conifers, hardwoods, shrubs, forbs, and graminoids 636 between samplings in early seral plots that burned only in the first fire in 2004 (Sims 1x burned) 637 and plots that experienced a second high-severity fire in 2015 (Saddle 2x burned).

638 Figure 6. Multivariate trajectories of early seral vegetation change based on a principal 639 components analysis of cover of conifers, hardwoods, shrubs, forbs, and graminoids. Plots are 640 translated to the origin, which represents the first sampling in 2010, six years following the Sims 641 Fire. Vectors show contrasting directional change in vegetation structure between those that 642 experienced a second high-severity fire in 2015 (Saddle Fire) and those that did not experience a 643 second fire (Sims Fire).

644

Page 26 of 34



Draft

27

645 Appendix 1 – Recent climate trends near the Sims and Saddles Fire study area outside of 646 Hyampom, California.

647 Figure. 1. Recent trends climate trends (1980-2019) for the Sims and Saddle Fires study area in 648 Trinity County near Hyampom, California: a) the annual temperature (deg. F) departure from 649 normal (2001-2010), b) the growing season (May, June, July, August) precipitation percent of 650 normal (2001-2010) ,and c) the Palmer Drought Severity Index (PDSI). Data accessed from 651 https://wrcc.dri.edu/wwdt/time/ on 1/22/2019.

Page 27 of 34



Draft

Figure 1. A conceptual model of possible vegetation state change after short-interval, high-severity fire in mixed conifer-hardwood forests of the Klamath Mountains in northern California.

254x190mm (300 x 300 DPI)

Page 28 of 34



Draft

Figure 2. Study location and patterns of burn severity (RdNBR) in the Sims (2004) and Saddle Fire (2015). Inset in the top left shows the location of the Klamath Mountains in northern California and southwestern

Oregon, USA. The hatched area overlaying the Sims Fire perimeter is the perimeter of the Saddle Fire which is presented in the inset in the bottom right corner.

215x279mm (300 x 300 DPI)

Page 29 of 34



Draft

Figure 3. Timing of conifer establishment for: a) the Sims Fire in 2010, b) after the Sims Fire and prior to the Saddle Fire in 2010, and c) in 2016 following only the Sims Fire, and d) in 2017 following both the Sims

Fire and the Saddle Fire.

134x175mm (300 x 300 DPI)

Page 30 of 34



Draft

Figure 4. Distribution of cover of conifers, hardwoods, shrubs, forbs, and graminoids in early seral plots that experienced a second high-severity fire in 2015 (Saddle 2x burned) and plots that only burned during the

first fire in 2004 (Sims 1x burned). Fires and sampling times are abbreviated as follows: SM1=Sims Time 1 (2010), SM2=Sims Time 2 (2016), SD1=Saddle Time 1(2010), SD2=Saddle Time 2 (2017).

203x152mm (300 x 300 DPI)

Page 31 of 34



Draft

Figure 5. Comparison of changes in cover of conifers, hardwoods, shrubs, forbs, and graminoids between samplings in early seral plots that burned only in the first fire in 2004 (Sims 1x burned) and plots that

experienced a second high-severity fire in 2015 (Saddle 2x burned).

82x190mm (300 x 300 DPI)

Page 32 of 34



Draft

Figure 6. Multivariate trajectories of early seral vegetation change based on a principal components analysis of cover of conifers, hardwoods, shrubs, forbs, and graminoids. Plots are translated to the origin, which

represents the first sampling in 2010, six years following the Sims Fire. Vectors show contrasting directional change in vegetation structure between those that experienced a second high-severity fire in 2015 (Saddle

Fire) and those that did not experience a second fire (Sims Fire).

194x143mm (300 x 300 DPI)

Page 33 of 34



Draft

152x203mm (300 x 300 DPI)

Page 34 of 34



Draft - University of Toronto T-Space · Draft 2 16 Abstract 17 We compare early seral development...

Documents

Transcript of Draft - University of Toronto T-Space · Draft 2 16 Abstract 17 We compare early seral development...