Forest Ecology and Management 389 (2017) 374–385

Contents lists available at ScienceDirect

Forest Ecology and Management

journal homepage: www.elsevier .com/ locate/ foreco

Historical and current forest conditions in the range of the NorthernSpotted Owl in south central Oregon, USA

http://dx.doi.org/10.1016/j.foreco.2016.12.0290378-1127/� 2016 Elsevier B.V. All rights reserved.

⇑ Corresponding author.E-mail address: hokulea@uw.edu (R.K. Hagmann).

R. Keala Hagmann a,⇑, Debora L. Johnson a, K. Norman Johnson b

aApplegate Forestry LLC, 28831 Tampico Road, Corvallis, OR 97330, USAbDepartment of Forest Ecosystems and Society, Oregon State University, 321 Richardson Hall, Corvallis, OR 97331, USA

a r t i c l e i n f o a b s t r a c t

Article history:Received 24 August 2016Received in revised form 25 December 2016Accepted 26 December 2016Available online 18 January 2017

Keywords:Dry forest restorationHistorical and current Northern Spotted OwlcoverFrequent-fire/fire-prone forestsHistorical forest conditionsEcosystem management

Restoration to increase resilience to current and projected drought and fire in historically open-canopyforests in fire-prone environments may be constrained by concern for species that favor dense forest con-ditions. To assist the recovery of a threatened species, the Northern Spotted Owl (NSO), in dry forest envi-ronments, the US Fish & Wildlife Service (USFWS) recommends embedding NSO conservation andrecovery in restoration efforts that enable ecosystem recovery from past management actions andincrease resilience to changing climate. In this study, we assessed changes between historical and current(1) forest structure and composition and (2) extent of NSO nesting and roosting (NR) or foraging (F) forestcover on 39,000 ha at the eastern edge of the current designation of the range of the NSO. Historicalrecords depict a predominantly open-canopy landscape dominated by large ponderosa pine. Current con-ditions include more than a 600% increase in trees 15–53 cm dbh, substantial decline in treesP81 cm dbh, loss of the widespread distribution of trees P53 cm dbh, and loss of the dominance of pon-derosa pine on mixed conifer sites. NSO habitat assessment involves a suite of attributes including: land-scape context; species composition; canopy cover; basal area; average tree diameter; diameter diversity;and abundance of large trees, snags, canopy layers, coarse woody debris, and mistletoe. We tested for thepresence of forest that met USFWS threshold values for two of these variables, canopy cover and basalarea. Historically none of the area met the 60% canopy cover threshold for NR or F forest cover and almostnone meets it currently. However, several NSO nesting pairs and individual birds have been observed inthe study area over the last 20 years, and studies in other frequent-fire forests show that canopy cover aslow as 50% may be functional for NSO. To assess the implications of lower threshold values, we tested forNR or F forest cover presence at half the recommended thresholds, considerably below published esti-mates. Only five percent of the area exceeded 30% canopy cover historically; much of the current forestexceeds it today. Increase in canopy cover comes at the expense of increasing vulnerability to fire anddrought and loss of historical functions and processes. Conflicting objectives of forest restoration andmaintenance of spotted owl habitat on this site – isolated habitat in the dry margin of the range of theNSO – raises questions about how to achieve forest restoration in altered landscapes where existing,novel conditions favor at-risk species.

� 2016 Elsevier B.V. All rights reserved.

1. Introduction

Recent reviews of key concepts for enhancing resistance andresilience to current and projected wildfire and other drought-related stressors in fire-prone forests in the Pacific Northwest pro-vide recommendations for integrating the recovery of characteris-tic patterns and processes with current management objectivesgiven altered social and geophysical environments (Franklin andJohnson, 2012; Franklin et al., 2013; Stine et al., 2014; Hessburg

et al., 2015, 2016; Lehmkuhl et al., 2015). Restoration is definedin the 2012 Planning Rule to implement the National Forest Man-agement Act as ‘‘The process of assisting the recovery of an ecosystemthat has been degraded, damaged, or destroyed. Ecological restorationfocuses on reestablishing the composition, structure, pattern, and eco-logical processes necessary to facilitate terrestrial and aquatic ecosys-tems sustainability, resilience, and health under current and futureconditions.” (USDA, 2012, 219.19). Predominantly open-canopy for-ests dominated by fire- and drought-tolerant trees have demon-strated resilience and resistance to fire and drought (Stephensand Gill, 2005; Stephens and Fulé, 2005; Collins and Stephens,2010). Historical records and reconstructions provide additional

R.K. Hagmann et al. / Forest Ecology and Management 389 (2017) 374–385 375

reference conditions to inform restoration of landscapes substan-tially altered by management actions; for example, incorporationof the range of variation in historical spatial patterns as metricsto guide restoration efforts (Larson and Churchill, 2012; Churchillet al., 2013). Comprehensive planning at the spatial scale of theprocesses and functions to be restored or maintained is recom-mended prior to implementation. Other key recommendationsinclude: retain and release old trees; shift composition towardfire- and drought-tolerant species; restore fine- and meso-scalevariation in tree, shrub, and herbaceous cover; reduce live treedensity and increase mean tree diameter; protect and restoreunderstory plant communities; and treat activity fuels. However,restoration of conditions resilient and resistant to fire and otherdrought-related stressors may reduce dense forest cover that hasexpanded under a century of fire exclusion and suppression andthat may currently be utilized by northern spotted owl (NSO, Strixoccidentalis caurina). Populations of this threatened subspecies arein decline across their range due increasingly to competition withbarred owls and habitat loss from fires (Dugger et al., 2015; Daviset al., 2016). Management actions cannot jeopardize survival orimpede recovery of the listed species (Endangered Species Act of1973 (16 USC 1531 et seq.)).

Recommendations in the NSO recovery plan (USFWS, 2011)support active, adaptive management, such as dry forest restora-tion, to embed conservation of occupied sites (currently and histor-ically) and high value NSO habitat within broader efforts to restoreecosystem resiliency. Given the importance of dispersal processes,dispersal habitat may also be a priority for conservation. Restora-tion of ecosystem resilience as described in the recovery planincludes reducing the risk of continued loss of valued ecologicalstructures (e.g., large, fire-tolerant trees) and reversing the on-going decoupling of vegetation patterns from characteristic distur-bance processes. One important question, then, would be whetherthe dense forest conditions that NSO prefer to use for NR were pre-sent in the historical landscape. If so, maintenance of that covertype would be a goal of restoration and would help guide it. How-ever, if it was not present in the historical landscape, then its main-tenance could potentially conflict with forest restoration.

Unfortunately few studies of the historical availability of NSOhabitat in the slopes and foothills of the eastern Cascade Range exist,and those that do differ in their conclusions. A study of 85 nest treesin Washington found that almost all (92%) were Douglas-fir, 26%were older than 200 years, 50% were older than 137 years, and treeson south-facing slopes were substantially older than trees on northfacing slopes (Buchanan et al., 1993). Potential nest trees existed his-torically; however, dendrochronological reconstruction of fire historyand historical forest conditions at a randomly selected subset of thesame nest sites found little overlap between historical and currentconditions around those nest trees (Everett et al., 1997). Kennedyand Wimberly (2009) simulated historical forest conditions for theDeschutes National Forest (NF) in central Oregon and estimated thatclosed canopy, older forest conditions potentially suitable as NSO NRcover may have existed on 9% of the area. In contrast to the relativelylow estimates of potential NSO forest cover in the previous studies,an estimate based on a reconstruction of historical forest densityfrom GLO survey data suggests that NSO forest cover was not rarein the eastern Cascade Range of Oregon; potential nest stands existedon 22–39% of the study area, and potential foraging and roostingcover existed on a minimum of 11% and maximum of 68% of thestudy area (Baker, 2015b). Thus, results of these studies, all of whichare based on reconstructed or modeled conditions, conflict in theirconclusions about the historical presence of NSO forest cover.

To further investigate the historical occurrence of potential NRcover in fire-prone portions of the current range of the NSO, weused a data-rich, historical record of trees P15 cm dbh collectedas a 10–20% sample of more than 39,000 ha in the study area. This

1914–1924 timber inventory provides a more substantial record ofhistorical forest conditions than was available to previous studiesof the historical extent of NSO forest cover. Using these historicalrecords and 2014 United States Forest Service (USFS) stand examdata, we explored changes in forest conditions and availability ofdense forest cover commonly used by NSO for NR. The datasetdescribing current forest conditions enables spatially explicit com-parisons of change between historical and current conditions at ascale not previously available (Hagmann et al., 2013). Additionally,we explore changes in the very large (P81 cm dbh), old tree pop-ulation and distribution. Results from this study provide informa-tion about whether restoration is justified on this landscape andalso contribute to the dialogue on the tradeoffs between manage-ment focused on conservation of an at-risk species and restorationof a broader array of ecosystem functions and processes occurringin fire-prone environments across western North America.

2. Methods

2.1. Study area

The 39,466-hectare study area occupies the southwest corner ofthe former Klamath Reservation, now the Fremont-Winema NF, inthe foothills of the Cascade Range in south central Oregon (Fig. 1),where the USFS is currently developing a restoration project. TheKlamath Tribes are the senior partner in a stewardship agreementthat covers most of the Fremont-Winema NF (USDA Forest Service,2011) and is focused on implementing the Klamath Tribes’ forestrestoration strategy (Hatcher et al., 2016). Historically, forests in thisarea had predominantly low live tree densities dominated by large,fire- and drought-tolerant trees (Hagmann et al., 2013). Most ofthe study area lies within the Northwest Forest Plan (NWFP) bound-ary (Fig. 1) which reflects scientists’ understanding of the range ofthe NSO when the NWFP was written (1993) and implemented(1994). Under the NWFP, most of the forest in the study area isclassified as ‘‘matrix”, the land-use allocation in which regularlyscheduled timber harvests were expected to occur. Outside theNWFP and designated cultural resource areas, most of the area isclassified as ‘‘timber production” under the Fremont-Winema ForestPlan (fs.usda.gov/main/fremont-winema/landmanagement/planning),although that classification is now tempered with commitments toconserve old forests and the species within them. NSO surveyshave detected several nesting pairs and individual birds over thelast 20 years. The area was not designated critical habitat(USFWS, 2012).

Much of the study area falls within the Pumice Plateau Forestsubsection of the Eastern Cascades section (Cleland et al., 2007).At lower elevations, the forest grades from ponderosa pine (Pinusponderosa) to mixed conifer forests consisting primarily of pon-derosa pine, white fir (Abies concolor), Douglas-fir (Pseudotsugamenziesii), sugar pine (Pinus lambertiana), and incense cedar (Calo-cedrus decurrens). At the highest elevations in the study area, theforest is dominated by white fir and lodgepole pine (Pinus con-torta). Annual precipitation ranges from 41–135 cm; meanmonthly temperatures range from -42 �C in January to 29 �C in July(PRISM, 2012). Elevation ranges from 1269 to 2215 m. Soils in themixed conifer area are predominantly well-drained volcanic ashover very gravelly and loamy material, ashy-skeletal typic cryan-depts in the taxonomic order inceptisols.

2.2. Plant association groups (PAGs)

Potential natural vegetation is used to easily communicate a con-cise and validated description of the capability of an area to supportvegetative ecosystems (Winthers et al., 2005). We used plant associ-ations mapped in the USFS Terrestrial Ecological Unit Inventory

Fig. 1. Inset shows the location of the study area (red) in Oregon in the southwest corner of the former Klamath Reservation (grey) relative to the eastern boundary of theNWFP (black line). Map of study area shows the eastern boundary of the NWFP (heavy black line), study area (thin black line), and plant association groups derived from theUSFS Terrestrial Ecological Unit Inventory. For a color version of this graphic, refer to the web version of this article.

376 R.K. Hagmann et al. / Forest Ecology and Management 389 (2017) 374–385

(TEUI) (Dorr, 2007) to parse geophysical and environmental variabil-ity of the study area. The TEUI incorporates distinct combinations oflandform, morphometry, soil, potential natural vegetation, surficialgeology, hydrology, and climate to map the arrangement of landareas with similar potential productivity and responses to manage-ment activities (Winthers et al., 2005). We classified potential natu-ral vegetation types to PAGs using established guides (Appendix ATable 1). We linked this map of PAGs (Fig. 1) with both the historicaland current data sets to summarize data for the PAGs that form thefocus of this study (ponderosa pine and mixed conifer) and comprisethe majority of the study area (88%).

2.3. Historical forest conditions

Between 1914 and 1924, timber inventory data were collectedby the Bureau of Indian Affairs (BIA, known at the time as the USIndian Forest Service) using uniformly distributed, systematically

located strip surveys referenced to documented locations in theBureau of Land Management Public Land Survey System of landdivision. The inventory represents a 10–20% sample of the area.Cruisers tallied conifers >15 cm in diameter at breast height(dbh) by species. The origin of this extensive, spatially explicit his-torical record and the methods used to convert it from an archivedpaper record to digital data have been described previously(Hagmann et al., 2013). Inventory methods changed substantiallyafter 1919 (Hagmann et al., 2013). Of the townships in this studyarea, eight were inventoried between 1914 and 1917, one wasinventoried in 1920, and 4 sections of one township (T34S R7E)were inventoried in 1924. Appendix A Fig. 1 shows: (1) areasinventoried before and after 1919; (2) areas excluded from thisstudy because (a) density was lower than 25 tph P15 cm dbh,(b) evidence of logging was noted by cruisers (Appendix B Table 1),or (c) the inventory record was incomplete; and (3) areas for whichinventory records were not found.

Table 1Metrics derived from the 2012 NSO critical habitat rule (USFWS, 2012, pp. 71903–71907) or provided by the Klamath Falls Fish and Wildlife Office to identify NR or Fforest cover.

Metric Nesting/roosting Foraging

Canopy cover (%) P60 P60Basal area (m2/ha) 55 32a

a Specified by USFWS Klamath Falls Fish &Wildlife Office; the critical habitat ruledoes not specify basal area requirements for foraging habitat.

R.K. Hagmann et al. / Forest Ecology and Management 389 (2017) 374–385 377

From 1914 to 1919, inventories routinely included ponderosapine, sugar pine, Douglas-fir, and white fir; other species wererecorded in some areas. After 1919, all conifer species wereincluded in the inventory. Absence of lodgepole pine and incensecedar in the areas inventoried before 1919 does not lead to anunderrepresentation of historical densities in these ponderosa pineand mixed conifer forests. On the 762 ha of transect area in pon-derosa and mixed conifer plant associations inventoried after1919, an average of 1% of the trees per hectare (tph) were incensecedar or lodgepole pine. Lodgepole pine were recorded only on anortheast facing slope near or above 1890 m elevation, just belowthe only portion of the study area designated as a white fir-lodgepole plant association (ABCO-PICO/ACOC3-CAINI3) in theTEUI (Fig. 1). Low density of lodgepole pine and incense cedar isconsistent with the predominant plant associations (PIPO/PUTRand ABCO/CEVE) for ponderosa pine and mixed conifer forests inthis area (Appendix A Table 1).

Transects entirely contained within mapped TEUI polygonswere assigned that PAG. Summary statistics describing foreststructure and composition for each PAG include only those tran-sects with density higher than 25 tph, completely within a singlePAG polygon, and on which no logging was noted at the time thearea was inventoried. Logging was noted on transects (n = 34)inventoried in 1920 and 1924 (Appendix B Table 1); most of thesetransects (n = 29) were in T34S R07E and included notes fromcruisers identifying the area as an ‘‘old logging unit”. Mean andstandard deviation for each summary statistic were weighted bytransect area to accommodate differences in plot size. We testedthe assumption that the subset of the transects used representsthe range of conditions across the study area by comparing themean and standard deviation weighted by area occupied for eachPAG to the mean and standard deviation for all transects in thestudy area. We excluded transects that fell more than 50% withinlodgepole pine/bitterbrush plant associations from this test as theywere unlikely to be either ponderosa pine or mixed conifer forest.

2.4. Current forest conditions

The data set representing current conditions was provided byand is available from the Fremont-Winema NF. Using a methodol-ogy developed by the USFS for collecting inventory data acrosslarge landscapes, 178 detailed stand exams were completed withinthe study area in 2014 that represented 60 stand ‘‘types” based onslope, aspect, elevation, and vegetation characteristics derivedfrom classified satellite imagery. Within each sample stand, trees>18 cm dbh were measured on variable plots; smaller trees weremeasured on nested fixed plots. The USFS then used the nearestneighbor analysis within FSVeg Spatial Data Analyzer (fs.fed.us/nrm/fsveg/) to extrapolate sampled stands to all stands in thestudy area. The imputation quality was evaluated for nine variatesand 92% of the stands had a canonical R squared of 0.91 (A. Spen-cer, USFS Fremont-Winema NF, personal communication, 2015).

Stands entirely contained within the mapped TEUI polygons foreach PAG were assigned that PAG. Summary statistics describingforest structure and composition include only those stands withmore than 25 tph and contained entirely within a single PAG poly-gon. We tested the assumption that this subset of stands repre-sents the range of conditions across the study area by comparingthe mean and standard deviation for each PAG weighted by thearea occupied by that PAG to the mean and standard deviationfor all stands in the study area. We excluded stands that fell morethan 50% within lodgepole pine/bitterbrush plant associationsfrom this test as they were unlikely to be either ponderosa pineor mixed conifer forest. Mean and standard deviation for eachstatistic were weighted by stand area to accommodate differencesin size.

2.5. Potential NSO nesting and roosting (NR) or foraging (F) forestcover

To estimate potential NR or F forest cover, we applied basal areaand canopy cover thresholds (Table 1) from the 2012 critical habi-tat rule for NSO (USFWS, 2012, pp. 71903–71907) to both the1914–1924 BIA inventory (hereafter historical) and the 2014 USFSstand exam (hereafter current) data sets. Vegetation structure andcomposition were among the most useful of a large number of vari-ables which each make relatively modest contributions to predict-ing NSO habitat quality in the East Cascades zone (USFWS, 2012, p.71904). Canopy cover and basal area are considered among thebest information available for characterizing NSO habitat function(EM Glenn, personal communication, January 12, 2016). Addition-ally, they provide a consistent index of forest cover quality basedon readily available information as suggested by their prominencein descriptions of critical habitat for NSO (USFWS, 2012, pp.71903–71907). Note that the USFWS metric refers to canopy cover(also called crown cover), the proportion of the ground covered bythe vertical projection of tree crowns, not canopy closure, the pro-portion of sky hemisphere obscured by vegetation when viewedfrom a single point (Jennings et al., 1999). Variables other thancanopy cover and basal area that are associated with the classifica-tion of NSO habitat include spatial amounts and configurations ofcover types, species composition (e.g., percent basal area pine),average tree diameter, and diameter diversity as well as abundanceof large trees, snags, canopy layers, coarse woody debris, andmistletoe. The approach we used may overestimate the amountof NR or F forest cover as we applied only a subset of the variableslisted in the critical habitat rule.

We also estimated abundance of potential NR or F forest coverusing half of the USFWS recommended values shown in Table 1because (1) NSO have survived and successfully nested in the areaalthough, as will be demonstrated, canopy cover exceeds 60% onless than 1% of the study area and (2) as described in the 2012 crit-ical habitat rule (USFWS, 2012, pp. 71903–71907), NSO in drier for-est environments may use lower density forests. We consideredthresholds low enough to encompass all possibilities. Canopy coveras low as 50% for NR or F has been recorded (Buchanan et al., 1995;Gerdes et al., 1995; USDA Forest Service, 1997; Loehle et al., 2011;Singleton, 2013). Slightly lower values may have been recorded inthe USDA FS 1997 study; however, the lack of reliable methods forconverting canopy closure to canopy cover confounds comparison.Also, canopy cover and basal area values lower than half theUSFWS thresholds were recorded in areas identified as potentialNR or F forest cover in a 2014 USFS study of NSO habitat qualityin the study area (Johnson, 2016; USDA Forest Service, 2016).

The critical habitat rule does not mention basal area require-ments for foraging habitat; therefore, we used the value providedby the Klamath Falls US Fish and Wildlife Office for the field-based identification of NSO foraging cover conducted by the USFSin the project area. We modeled canopy cover using the South Cen-tral Oregon and Northeast California variant Forest VegetationSimulator (http://www.fs.fed.us/fmsc/fvs/index.shtml), which cal-culates crown width using species and region specific equations,

378 R.K. Hagmann et al. / Forest Ecology and Management 389 (2017) 374–385

sums up the crown areas, and then corrects for the overlap(Crookston and Stage, 1999). We modeled canopy cover in FVSfor both historical and current conditions from the tree lists in boththe 1914–1924 BIA inventory and the 2014 USFS stand exams.

3. Results

3.1. Historical forest conditions

Forests on this landscape were predominantly low density anddominated by large (P53 cm dbh) ponderosa pine with localizedpatches dominated by white fir (Table 2). Median values for den-sity were 59 tph and 14 m2/ha of basal area for trees P15 cmdbh. The 95th percentile values were 145 tph and 29 m2/ha ofbasal area. Basal area consisted primarily of large trees (84 ± 11%)and ponderosa pine (70 ± 19%). Dominance by white fir wasrecorded in areas with >100 cm average annual precipitation onnorth to east facing slopes (Appendix A Fig. 2). Across the moistmixed conifer forest, the ratio of ponderosa pine to white fir basalarea was nearly one to one; however, as most of the white fir den-sity was in small trees (<53 cm dbh), white fir outnumbered pon-derosa pine by more than two to one (Table 3). Ponderosa pinedominated large tree basal area even in moist mixed conifer(Table 3). The largest trees inventoried were Douglas-fir and sugarpine (193–234 cm dbh) although neither species was abundant inthe area. The largest ponderosa pine inventoried was 188 cm dbh.Summary statistics for this subset of transects (100% within a sin-gle PAG) are comparable to those for all transects in the study areawhen summary statistics for the subset are weighted by the area ineach PAG in the study area (Table 2).

Table 2Descriptive statistics for historical forest structure and composition from the 1914–1924deviation, and the range recorded for each metric. The column labeled ‘‘Ponderosa & mixed”first three columns (transects entirely within each of the mapped PAGs) weighted by thecolumn, labeled ‘‘All”, to assess the ability of the subset to represent the variability of condthose transects more than 50% in a lodgepole pine PAG.

Ponderosa Dry mixed

# of transects 31 388Plot area (ha) 174 1332

TPHP15 cm dbh 48 ± 19 (25–100) 76 ± 33 (25–296)15–53 cm dbh 26 ± 12 (10–58) 40 ± 25 (2–177)53–81 cm dbh 18 ± 8 (3–38) 25 ± 11 (1–104)P81 cm dbh 4 ± 2 (0–12) 11 ± 7 (0–33)Ponderosa 47 ± 20 (9–100) 40 ± 20 (0–104)White fir 0 ± 1 (0–7) 28 ± 36 (0–286)Sugar pine 0 ± 1 (0–3) 1 ± 2 (0–13)Douglas-fir 0 ± 0 (0–0) 5 ± 8 (0–63)Othera 1 ± 4 (0–30) 1 ± 2 (0–33)

Basal area (m2/ha)P15 cm dbh 10 ± 5 (3–16) 19 ± 8 (4–59)%P53 cm dbh 86 ± 11 (37–96) 83 ± 11 (16–97)% ponderosa 98 ± 8 (68–100) 64 ± 23 (0–100)

a Incense cedar and lodgepole pine in areas inventoried after 1919.

Table 3Relative distribution of density in the 1914–1924 BIA inventory for ponderosa pine and whfrom the difference between 100% and the sum of the values shown.

tph Basal

PAG Ponderosa White fir Ponde

Ponderosa 95% 2% 94%Dry mixed 63% 26% 66%Moist mixed 28% 66% 44%

3.2. Current forest conditions

Currently, these forests are dominated by small trees (15–53 cmdbh), and trees P81 cm dbh are rare (Table 4). Summary statisticsfor the subset of stands that fall entirely within mapped ponderosaand mixed conifer PAGs are comparable to the summary statisticsfor all stands in the study area when summary statistics for thesubset are weighted by the area in each PAG (Table 4).

3.3. Comparison between historical and current forest conditions

Tree density increased substantially (Fig. 2), and dominanceshifted from large (P53 cm dbh) to small (15–53 cm dbh) trees(Fig. 3). Density (tph) of small trees increased by more than600%. The increase in small ponderosa pine is larger than theincrease in small white fir (Fig. 4). Nonetheless, ponderosa pinebasal area has declined across all three PAGs with the largestdecline in moist mixed conifer forest (Fig. 5, Appendix A Fig. 3).This decline is primarily due to the decrease in large tree density(Fig. 2) and is particularly evident for trees P81 cm dbh whichdeclined from 9 ± 6 tph historically to 3 ± 4 tph currently (Tables2 and 4). Ponderosa pine comprised the majority of the treesP81 cm dbh historically; these large trees, once widely dispersedacross the landscape, are now largely absent (Fig. 6). The 25th per-centile value for percentage of basal area in ponderosa pine was46% historically, i.e., on three-quarters of the area inventoried,more than 46% of the basal area was ponderosa pine. Currently,the 25th percentile value for percentage of basal area in ponderosapine is 25%. Decrease in the large ponderosa pine population(Fig. 6) and expansion and increase in the presence and densityof species other than ponderosa pine (Table 4, Fig. 4) contribute

BIA inventory for transects with more than 25 tph. Values indicate mean, standardshows the mean and standard deviation for the subset of transects summarized in thearea in each PAG. These values are provided for comparison with those in the last

itions across the entire study area. ‘‘All” includes all transects in the study area except

Moist mixed Ponderosa & mixed All

208 627 1036411 1917 3596

83 ± 41 (25–245) 70 ± 31 68 ± 25 (25–296)48 ± 34 (2–227) 38 ± 23 36 ± 19 (2–227)24 ± 15 (5–89) 23 ± 11 22 ± 9 (0–104)11 ± 5 (0–24) 9 ± 6 10 ± 4 (0–36)26 ± 14 (0–70) 40 ± 19 40 ± 14 (0–105)56 ± 44 (0–237) 25 ± 29 23 ± 26 (0–286)0 ± 1 (0–17) 1 ± 1 0 ± 1 (0–17)1 ± 3 (0–18) 3 ± 5 4 ± 5 (0–63)0 ± 0 (0–0) 1 ± 2 1 ± 2 (0–79)

20 ± 8 (7–43) 17 ± 7 17 ± 6 (1–59)81 ± 12 (26–97) 84 ± 11 83 ± 8 (0–97)52 ± 20 (0–100) 70 ± 19 72 ± 15 (0–100)

ite fir aggregated at the PAG level. Relative abundance of other species can be inferred

area Basal area P 53 cm dbh

rosa White fir Ponderosa White fir

2% 94% 2%20% 68% 18%49% 52% 41%

Table 4Descriptive statistics for current forest structure and composition from 2014 USFS stand exam data for stands with more than 25 tph. Values indicate mean, standard deviation,and the range recorded for each metric. The column labeled ‘‘Ponderosa & mixed” shows the mean and standard deviation for the subset of stands summarized in the first threecolumns (stands entirely within each of the mapped PAGs) weighted by the area in each PAG. These values are provided for comparison with those in the last column, labeled‘‘All”, to assess the ability of the subset (‘‘Ponderosa & mixed) to represent the variability of conditions across the entire study area. ‘‘All” includes all stands in the study areaexcept stands falling more than 50% in lodgepole pine plant associations.

Ponderosa Dry mixed Moist mixed Ponderosa & mixed All

# of stands 413 1107 265 1785 3654Area represented (ha) 3595 13591 2557 19742 39497

TPHP15 cm dbh 279 ± 143 (75–1124) 306 ± 148 (62–796) 279 ± 166 (75–796) 296 ± 149 293 ± 145 (62–1124)15–53 cm dbh 269 ± 142 (70–1124) 282 ± 151 (10–764) 248 ± 166 (61–764) 274 ± 151 271 ± 146 (10–1124)53–81 cm dbh 9 ± 10 (0–57) 21 ± 21 (0–113) 27 ± 26 (0–88) 19 ± 19 19 ± 19 (0–113)P81 cm dbh 1 ± 2 (0–15) 3 ± 4 (0–21) 4 ± 5 (0–15) 3 ± 4 3 ± 4 (0–21)Ponderosa 239 ± 273 (0–15) 147 ± 205 (0–15) 86 ± 130 (0–7) 161 ± 211 157 ± 207 (0–15)White fir 1 ± 7 (0–1) 107 ± 170 (0–15) 179 ± 232 (0–13) 91 ± 140 85 ± 154 (0–15)Sugar pine 1 ± 9 (0–84) 6 ± 18 (0–132) 1 ± 132 (0–132) 4 ± 31 4 ± 37 (0–132)Douglas-fir 2 ± 20 (0–329) 28 ± 54 (0–329) 7 ± 25 (0–329) 19 ± 42 20 ± 46 (0–329)Othera 22 ± 107 (0–1106) 14 ± 42 (0–672) 4 ± 24 (0–288) 14 ± 55 21 ± 82 (0–1106)

Basal area (m2/ha)P15 cm dbh 22 ± 10 (8–46) 28 ± 11 (7–71) 31 ± 16 (7–71) 27 ± 12 27 ± 12 (7–71)% P53 cm dbh 14 ± 14 (0–64) 29 ± 24 (0–93) 33 ± 24 (0–75) 26 ± 21 26 ± 21 (0–93)% ponderosa 87 ± 23 (0–100) 50 ± 42 (0–100) 33 ± 27 (0–100) 56 ± 35 57 ± 37 (0–100)

a Other includes only lodgepole pine and incense cedar.

Fig. 2. Spatial distribution of density for all trees (top) and for ponderosa pineP53 cm dbh (bottom) historically (left) and currently (right). Breaks for all treesindicate historical quartiles and the 95th percentile value (145 tph); median densitywas 59 tph. Breaks for ponderosa P53 cm dbh indicate historical quartiles andtransects on which no ponderosa P53 cm dbh; median density of ponderosaP53 cm dbh was 17 tph. Historical values are displayed as the area represented bythe sampled area (transects). Blank areas (no squares or rectangles) in the historicalmap indicate areas where inventory records were either not found or wereexcluded because they did not meet study criteria (Appendix A Fig. 1). For a colorversion of this graphic, refer to the web version of this article.

Fig. 3. Comparison of mean density in historical and current forest conditions forall trees (top), for small trees per hectare (bottom left), and for large tree basal area(bottom right). Height of grey bars indicate mean values; error bars indicatestandard deviation. See Appendix A Fig. 3 for comparison of distribution ofhistorical and current forest conditions for each ponderosa and mixed conifer PAG.

Fig. 4. Comparison of historical and current mean abundance of trees 15–53 cm dbh for ponderosa pine (left), Douglas-fir (middle), and white fir (right).Height of grey bars indicate mean values; error bars indicate standard deviation.

R.K. Hagmann et al. / Forest Ecology and Management 389 (2017) 374–385 379

Fig. 5. Comparison of historical and current contribution of ponderosa pine to basalarea by PAG. Height of grey bars indicate mean values; error bars indicate standarddeviation.

Fig. 6. Spatial distribution of ponderosa pine and large trees historically andcurrently. Values from the 1914–1924 BIA timber inventory (left) are displayed asthe area represented by the sampled area (transects) for ease of comparison withthe 2014 USFS stand exam (right). Mixed conifer PAGs are shown in grey. Breaks forponderosa pine P81 cm dbh basal area (m2/ha) are above and below the medianvalue of 4 m2/ha historically. Median basal area for ponderosa P81 cm dbhcurrently was 0. Blank areas (no squares or rectangles) in the historical mapindicate areas where inventory records were either not found or were excludedbecause they did not meet study criteria (Appendix A Fig. 1).

380 R.K. Hagmann et al. / Forest Ecology and Management 389 (2017) 374–385

to the loss of widespread dominance of ponderosa across the studyarea.

Within each measurement period (historical and current), den-sities (tph and basal area) for dry and moist mixed conifer werecomparable; however, between measurement periods, the densi-ties differed substantially (Appendix A Fig. 3). Douglas-fir is listedas a major climax species with 75% constancy and 0–30% cover inthe most wide spread plant association assigned to the study area(Appendix A Table 1), and the largest trees in the historical inven-tory were Douglas-fir. However, it was and remains a minoritycomponent of the forest in this area (Tables 2 and 4, Fig. 4).

3.4. Nesting/roosting (NR) or foraging (F) forest cover

Historically, none of the study area met both the USFWS basalarea and canopy cover thresholds (Table 1) used to characterizeNR or F forest cover (Table 5, Fig. 7). Currently less than 1% ofthe area meets both the USFWS basal area and canopy coverthresholds for NR and F forest cover given a lower diameter limitof 15 cm dbh (Table 5, Fig. 7). To buffer estimates of NR and F coverwe tested for area that met values equal to half the existing thresh-olds (see Methods). Historically, only 4 or 5% of the area invento-ried met even these much lower values used to buffer estimatesof NR or F cover, respectively (Table 5). In areas where less thanhalf the basal area was ponderosa pine, canopy cover higher than30% was rare and canopy cover higher than 50% was almost non-existent at the time of the 1914–1924 timber inventory, (Fig. 8).Currently, 43% and 66% of the study area meet the lower valuesused to buffer estimates of NR and F cover, respectively.

4. Discussion

A historical landscape of predominantly low-density forestsdominated by large ponderosa pines is evident in the 90-year oldtimber inventory. Substantial spatial heterogeneity in densityand species composition is evident in this extensive, detail-rich,systematic sample of historical forests. Current forest conditionsdepart substantially from these historical conditions including:(1) a more than 600% increase in the abundance of small trees(15–53 cm dbh), (2) loss of abundance and widespread distributionof large trees (P53 cm dbh), especially of trees P81 cm dbh, and(3) loss of widespread dominance of ponderosa pine on mixedconifer sites.

Local and region-wide declines in large, old fire- and drought-tolerant trees represent the loss of unique structures that devel-oped over centuries and that are only partially replaced by treesthat have established on these sites in recent decades. Older treeshave distinctive ecological characteristics and make unique contri-butions to ecological function (Kolb et al., 2007; Lutz et al., 2012;Lindenmayer and Laurance, 2016). Analyses of tree size and age

correlations show that ponderosa pine P53 cm dbh in this areaare probably, but by no means always, at least 150 years old; trees<53 cm dbh may also be older than 150 years (Bright, 1912;Morrow, 1985; Youngblood et al., 2004; Shuffield, 2011). Meanage at which ponderosa pine trees reached 81 cm dbh exceeded

Table 5Comparison of historical and current extent of potential nesting/roosting (NR) and foraging (F) cover types.

Area (ha) and % of area meeting USFWS thresholds Area (ha) and % of area meeting half of USFWS thresholds

BA CC 1914 2014 BA CC 1914 2014 2014 > 0 dbh

Basal area (m2/ha)NR P55 2 0% 897 2% P28 357 8% 19162 43% 21180 47%F P32 145 3% 13090 29% P16 1705 36% 34504 77% 34929 78%

Percent canopy coverNRF P60% 0 0% 162 0% P30% 227 5% 29975 67% 36040 80%

Basal area & percent canopy coverNR P55 P60% 0 0% 162 0% P28 P30% 194 4% 19162 43% 21180 47%F P32 P60% 0 0% 162 0% P16 P30% 227 5% 29741 66% 32592 72%

Fig. 7. Comparison of distribution of basal area (left) and canopy cover (right)historically (top) and currently (center and bottom). Black lines indicate the USFWSbasal area and canopy cover thresholds: P55 m2/acre of basal area for nesting/roosting (NR), P32 m2/ha of basal area for foraging (F), and P60% canopy cover forNR or F. Grey line indicates P30% canopy cover, used to buffer estimates ofhistorical availability of NSO forest cover to encompass values lower than theUSFWS threshold that may be NSO cover (See Methods: Potential NSO nesting androosting (NR) or foraging (F) forest cover).

Fig. 8. Spatial distribution of percent canopy cover where basal area of ponderosapine was less than 50% historically (left) and currently (right).

R.K. Hagmann et al. / Forest Ecology and Management 389 (2017) 374–385 381

300 years in an analysis of 4997 ponderosa pine stumps in 20 sitesalong the east Cascade Range in 1912 (Bright, 1912). We docu-mented a decrease in the abundance and widespread distributionof large, fire- and drought-tolerant trees (particularly treesP81 cm dbh, Fig. 6) that used to dominate basal area across themixed conifer forests in the study area as well as elsewhere acrossthe Klamath Reservation (Hagmann et al., 2013). These declines inlarge and older tree populations are consistent with other studiesof historical forest conditions in mixed conifer forests in Washing-ton and Oregon (Stine et al., 2014, Table 3). Further, the decline ison-going; the 20-year review of the NWFP documents a gross lossof 7 to 17% in forests with structural characteristics associated withages >200 years in the drier, fire-prone provinces in the range ofthe NSO between 1993 and 2012; this loss is offset by recruitment

of forest moving into this category resulting in a net change of +4.5to -10.1% by province (Davis et al., 2015, Table 8). However, forestsmoving into this category also account for the increase in NSOnesting/roosting cover, probably due to increases in Douglas-firand grand/white fir populations which lower the relative contribu-tion of fire- and drought-tolerant pine cover (Davis et al., 2016).

The record of predominantly low-density, open canopy forestsdominated by ponderosa pine in the 1914–1924 timber inventoryis consistent with records of other early surveys of this area: 1866–1909 General Land Office (GLO, blm.gov/or/landrecords/), 1899–1900 United States Geological Service (USGS) (Walcott, 1900),and 1930 USFS (Harrington, 2003) (Appendix C). Pine was the mostubiquitous and abundant species/group in all four historicalrecords (Fig. 9). The relative abundance and widespread distribu-tion of both large trees and large ponderosa pine trees in theGLO survey was comparable to that in the 1914–1924 BIA timberinventory (Fig. 9, Appendix C). Consistency in the depiction of his-torical forest conditions across these independent records collectedover a period of 64 years, from 1866 to 1930 is evident. Everyrecord has its limitations. Nonetheless, each of the historicalrecords depicts landscape conditions that were more similar toeach other than to current conditions. Whatever the particular lim-itations of each individual record or all four records collectively,the degree of change from historical to current conditions is clear.Dominance by shade-intolerant ponderosa pine, especially largeponderosa pine, supports the inference that historically this land-scape was maintained in a predominantly open-canopy forest con-dition by frequent low- to moderate-severity fire.

Historically, none of the study area met both the basal area andcanopy cover thresholds for NR or F forest cover and only 4 to 5%

Fig. 9. Comparison of historical records and maps. (A) 1866–1909 GLO survey: each dot represents a single tree; dot size is scaled by DBH. (B) 1899–1900 USGS survey: chartsshow relative abundance of trees >10 cm dbh, ‘‘other” species were primarily lodgepole with incense cedar and juniper. (C) 1914–1924 BIA timber inventory: relativeabundance of pine (primarily ponderosa with some sugar pine) and fir (primarily white fir with some Douglas-fir) basal area (Table 3). (D) 1930 USFS Survey of ForestResources. Refer to Appendix C for a detailed description of this comparison. For a color version of this graphic, refer to the web version of this article.

382 R.K. Hagmann et al. / Forest Ecology and Management 389 (2017) 374–385

met the lower values (half of the USFWS thresholds) used to bufferestimates of NR and F cover. The open pine forests that existed onthis site historically have been converted to denser forests increas-ingly dominated by shade-tolerant species like white fir andDouglas-fir, conditions that currently may be more suitable foruse by NSO. Our results are consistent with studies that infer lowhistorical availability of NSO forest cover east of the crest of theCascade Range (MacCracken et al., 1996; Everett et al., 1997;Davis and Lint, 2005; Kennedy and Wimberly, 2009). Increase inNR forest cover on this site is consistent with increases since the1930s across the eastern Cascade Range in Washington and Oregonand the Cascade Range in California (Lint, 2005, Table 3-18). Themost recent NSO habitat monitoring report (Davis et al., 2016,Table 6) shows a net increase of 13.5% in NR forest cover on federallands in the Oregon Eastern Cascades physiographic provincebetween 1993 and 2012.

In addition to the predominantly low canopy cover in the studyarea historically, the dominance of ponderosa pine and low num-bers of Douglas-fir probably also reduced the quality of potentialnesting habitat. Forest within 100 m of NSO nest sites along theeastern Cascade provinces in Washington and Oregon typicallyinclude very little (<1%) ponderosa pine basal area (Bailey et al.,2015, Table 4-1); data was compiled for 446 pair locations (nestand day roosts) between 1990 and 2010 (R. Davis, personal com-munication 8/3/2016). Neutral to negative associations betweenincreasing basal area of ponderosa and NSO nesting have beenobserved in southwestern Oregon and north-central California(Irwin et al., 2012) and across the current range of the NSO inthe Pacific Northwest (Davis et al., 2011, 2016). Douglas-fir wasnot and still is not abundant in the area although many of the lar-gest trees were and are Douglas-fir. Areas dominated by white firhistorically were those on north to east facing slopes that receivemore than 100 cm annual precipitation. However, at the time ofthe 1914–1924 timber inventory, canopy cover higher than 30%was rare and canopy cover higher than 50% was almost non-existent even in areas dominated by fir (Fig. 8). If NSO NR habitatexisted in the study area, it would have been of substantially differ-ent structure and composition and probably marginal, given cur-rent characterizations of suitable forest cover.

Potential sources of error in our estimate of the availability ofNR or F cover include: first, the use of only two of many variables

that influence NSO habitat quality; second, lack of informationabout trees smaller than 15 cm dbh; and third, the reliability ofForest Vegetation Simulator (FVS) estimates of canopy cover. Noneof these sources substantially influence our findings for the reasonspresented here. First, the inclusion of additional variables would bemore likely to further constrain or not affect the estimate of suit-able cover than it would be to increase it. Second, it is not likelythat historical abundance of trees smaller than 15 cm dbh wouldalter our findings. Basal area is not substantially affected by theinclusion of trees smaller than 15 cm dbh even at current densities;canopy cover estimates, on the other hand, are sensitive to theinclusion of small trees (Table 5). However, even at current densi-ties, the addition of trees smaller than 15 cm dbh has minimaleffect on the extent of potential NR or F cover (Table 5). Third, For-est Vegetation Simulator (FVS) estimates for canopy cover in low-density forests have demonstrated consistency with ground-basedmeasurements. Canopy cover less than 40% very closely matchedline-intercept measured canopy for 1982 systematically-placedForest Inventory and Analysis plots across all forest lands in Ore-gon (Gray et al., 2014). FVS did not underestimate canopy coverfor open-canopy, dry hardwood stands although it did so in forestswith >70% canopy cover; the authors conclude that the crownoverlap correction used by FVS overcompensates in forests withvery high canopy cover but not in open forests where crown over-lap is less of an issue (McIntosh et al., 2012). FVS estimates werenot significantly different from estimates derived from line-intercept measurements for Jeffrey pine forests with mean canopycover of approximately 25% in Baja California, Mexico (Stephenset al., 2015). Therefore, FVS appears to adequately estimate canopycover for low-density, open-canopy forests like those in this histor-ical record. Further, whether the absolute values modeled by FVSare correct or not, the relative difference between historical andcurrent canopy cover is clear.

These results describing historical availability of NR or F forestcover contrast substantially with an existing study for this area(Baker, 2015b), which suggests that almost half of the area com-mon to both studies (Appendix D Fig. 1) could have functionedas NSO nest stands and almost the entire area as foraging androosting habitat. Baker’s estimate of historical NSO cover (2015b)does not consider canopy cover, a key metric used by the USFWSto identify potential NSO forest cover. When we apply the metrics

R.K. Hagmann et al. / Forest Ecology and Management 389 (2017) 374–385 383

used in Baker (2015b) to the BIA inventory (Appendix D Table 1)only 19% of the area common to both studies would classify as neststands and 24% as foraging and roosting (Appendix D Table 2), sub-stantially less than estimated by Baker (2015b).

In addition, the 1914–1924 BIA inventory recorded substan-tially lower densities than Baker (2012), the reconstruction of his-torical density on which Baker (2015b) is based. Baker (2012)suggests that approximately half of the area common to both stud-ies (Appendix D Fig. 1) supported forests with a density of >143tph, which is greater than the 95th percentile value of 132 tphrecorded in the BIA timber inventory for the same area (AppendixD Fig. 1).

This difference in historical forest density between the BIA tim-ber inventory record and the reconstruction from GLO bearingtrees used in Baker (2015b) cannot be explained by (1) logging(Appendix B), (2) lower diameter limit, (3) greater diversity of spe-cies, or (4) selective sampling of only merchantable forests andexclusion of area in younger, denser forests or recovering burnareas as demonstrated in this study, Appendix D, and Hagmannet al. 2013. However, these criticisms have been presented as rea-sons (Odion et al., 2014; Baker, 2015a, 2015b) for invalidating thetimber inventories on which this analysis is based. For the areacommon to both studies (Appendix D Fig. 1), the BIA timber inven-tory includes 155,442 trees larger than 15 cm dbh. By contrast,only 491 bearing trees (the data set on which the reconstructionof historical density used in Baker (2015b) is based) were recordedin the original GLO surveys for the same area. Thus, we are compar-ing a timber inventory record, which is a systematic sample of 10–20% of the forest area and is consistent with other early surveys(1866–1909 GLO, 1899–1900 USGS, and 1930 USFS), to a modelof forest density derived from a sample of substantially less than1% of the historical population. From this suite of evidence, we con-clude that Baker (2012, 2015b) overestimated stand density.

Other scientists (Fulé et al., 2013; Merschel et al., 2014; Collinset al., 2015; Steel et al., 2015; Stephens et al., 2015) have also chal-lenged the validity of the reconstruction of historical forest densityin Baker (2012) and similar studies (Williams and Baker, 2012;Baker, 2014) as well as inferences drawn from them for forestsacross the intermountain west. Further, in this study, we havedemonstrated that four independent historical records depict his-torical forest conditions that were predominantly open canopyand dominated by ponderosa pine. These conditions are not con-sidered conducive to NSO nesting and roosting (USFWS, 2011;Davis et al., 2016).

4.1. Management implications

Concern for maintaining current NR or F as well as dispersalhabitat on this and similar sites for NSO, a threatened species, istempered by concern for the ecosystems that have been degradedby management actions (e.g., fire suppression, fire exclusion, andlogging) that resulted in the expansion of dense forest conditions.On the one hand, NSO populations have declined more thanexpected in the last 15 years (Dugger et al., 2015). On the other,current forest conditions are at increasing risk of continued lossof desired structures and functions due to the ongoing degradationof forest conditions that, historically, were more resilient to fireand drought (Franklin and Johnson, 2012; Stine et al., 2014;Hessburg et al., 2015, 2016). In its review of management options,the USFWS considers both the ability of the habitat to assist inrecovery of a threatened species and the risk of loss of desiredstructures, diversity, and functions associated with the creationor perpetuation of existing conditions (USFWS, 2011; Hensonet al., 2013).

Generally mixed conifer forests in this study area are degradedrelative to their historical condition. Site productivity is high

enough to support rapid and abundant increases in density and ashift to dominance by shade-tolerant, drought-intolerant species(Johnson et al., 2008; Merschel et al., 2014). The highest priorityfor restoration are those mixed conifer forests where substantialremnant old tree populations persist in highly altered forests(Johnson et al., 2008). However, many of these stands also containthe forest identified by local biologists as owl sites (current andhistorical) or high value habitat. Other mixed conifer forests inthe study area may also qualify as NR or F habitat and much of itmight function as dispersal habitat. Restoration to conditionsapproximating those that existed historically would reduce thequality of these forests as NSO habitat. Maintaining NSO foresthabitat would impede restoration informed by historical condi-tions – conditions expected to be more resilient in the face of achanging climate.

In the heart of the range of the NSO, we might expect thatmanagers would move very cautiously in degrading habitat,given the precarious current state of NSO populations. Restorationmay be largely restricted to non-habitat, until NSO populationsrecover.

These mixed conifer forests, however, are an isolated island onthe eastern edge of the NSO range bounded on the west by KlamathLake, the south by private timber land, and the north and east byponderosa pine forests. They are neither NSO critical habitat norLate Successional Reserves. They are separated from NSO criticalhabitat on the east slopes of the Cascade Range (USFWS, 2012)by several miles of water, marsh, or ponderosa pine forest. Addi-tionally, forests in the study area were classified as primarily ‘‘mar-ginal” to ‘‘unsuitable” with patches of cover suitable for nesting/roosting in an assessment of the suitability of forest cover acrossthe range of the NSO in the Pacific Northwest (Davis et al., 2016).Recent model output characterizing landscape connectivity andsource-sink relationships suggests that the eastern Cascades regionin southern Oregon is one of the strongest sinks within the bound-ary of the NWFP; and the eastern Oregon Cascade province may bea net liability for NSO populations (Schumaker et al., 2014). In fact,a recent paper published in Conservation Biology (Heinrichs et al.,2015) explored implications for population persistence if sinkhabitats were removed; in their analysis, NSO populations declinedbut the risk of extinction decreased.

Would restoration of fire- and drought-tolerant conditions inthese mixed conifer forests jeopardize survival or impede recoveryof the NSO? This question can undoubtedly be asked for other fed-eral forests along the eastern fringe of the NSO range. These largerlandscape considerations deserve attention by land managementand regulatory agencies and in the political-legal process by whichdecisions are made on our federal forests.

Acknowledgements

We gratefully acknowledge constructive reviews from PeterBrown, Ray Davis, Jerry Franklin, and Peter Singleton as well asRussell Kramer, Linda Winter, and three anonymous reviewers thatimproved the quality of manuscript; financial support for prepara-tion of the BIA timber inventory data from the Klamath Tribes;financial support for assessment of historical NSO forest cover fromthe Fremont-Winema National Forest; and data preparation by LoisTaylor and Amy Stephens.

Appendix A. Supplementary material

Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.foreco.2016.12.029. These data include Google maps of the most important areasdescribed in this article.

384 R.K. Hagmann et al. / Forest Ecology and Management 389 (2017) 374–385

References

Bailey, J.D., Vogler, K., Churchill, D., Youngblood, A., 2015. Chapter 4: Silviculturalapproaches to restoring resilient landscapes for northern spotted owls. In:Lehmkuhl, J., Gaines, W., Peterson, D.W., Bailey, J., (Eds.). Silviculture andmonitoring guidelines for integrating restoration of dry mixed-conifer forestand spotted owl habitat management in the eastern Cascade Range. Gen. Tech.Rep. PNW-GTR-915. Portland, OR: U.S. Department of Agriculture, ForestService, Pacific Northwest Research Station, pp. 63–102.

Baker, W.L., 2012. Implications of spatially extensive historical data from surveysfor restoring dry forests of Oregon’s eastern Cascades. Ecosphere 3, 1–39.

Baker, W.L., 2014. Historical forest structure and fire in Sierran mixed-coniferforests reconstructed from General Land Office survey data. Ecosphere 5. art79.

Baker, W.L., 2015a. Are high-severity fires burning at much higher rates recentlythan historically in dry-forest landscapes of the Western USA? PLoS ONE 10,e0136147.

Baker, W.L., 2015b. Historical Northern spotted owl habitat and old-growth dryforests maintained by mixed-severity wildfires. Landscape Ecol. 30, 655–666.

Bright, G.A., 1912. A study of the growth of yellow pine in Oregon. Unpublishedtypescript report obtained from the National Archives, College Park, MD; recordgroup 95. [Place of publication unknown]: U.S. Department of Agriculture,Forest Service. 106 p.

Buchanan, J.B., Irwin, L.L., McCutchen, E.L., 1993. Characteristics of spotted owl nesttrees in the Wenatchee-national-forest. J. Raptor Res. 27, 1–7.

Buchanan, J.B., Irwin, L.L., McCutchen, E.L., 1995. Within-stand nest site selection byspotted owls in the eastern Washington Cascades. J. Wildl. Manage., 301–310

Churchill, D.J., Larson, A.J., Dahlgreen, M.C., Franklin, J.F., Hessburg, P.F., Lutz, J.A.,2013. Restoring forest resilience: from reference spatial patterns to silviculturalprescriptions and monitoring. For. Ecol. Manage. 291, 442–457.

Cleland, D.T., Freeouf, J., Keys, J., Nowacki, G., Carpenter, C., McNab, W., 2007.Ecological Subregions: Sections and Subsections of the Conterminous UnitedStates [1:3,500,000] [CD-ROM]. Sloan, A.M., cartog. Gen. Tech. Report WO-76.Washington, DC: U.S. Department of Agriculture, Forest Service.

Collins, B.M., Lydersen, J.M., Everett, R.G., Fry, D.L., Stephens, S.L., 2015. Novelcharacterization of landscape-level variability in historical vegetation structure.Ecol. Appl. 25, 1167–1174.

Collins, B.M., Stephens, S.L., 2010. Stand-replacing patches within a ‘mixed severity’fire regime: quantitative characterization using recent fires in a long-established natural fire area. Landscape Ecol. 25, 927–939.

Crookston, N.L., Stage, A.R., 1999. Percent canopy cover and stand structurestatistics from the Forest Vegetation Simulator.

Davis, R.J., Dugger, K.M., Mohoric, S., Evers, L., Aney, W.C., 2011. Northwest ForestPlan—the first 15 years (1994–2008): status and trends of northern spotted owlpopulations and habitats. Gen. Tech. Rep. PNW-GTR-850.

Davis, R.J., Hollen, B., Hobson, J., Gower, J.E., Keenum, D., 2016. Northwest ForestPlan—the first 20 years (1994–2013): status and trends of northern spotted owlhabitats. Gen. Tech. Rep. PNW-GTR-929. Portland, OR: U.S. Department ofAgriculture, Forest Service, Pacific Northwest Research Station. 54 p.

Davis, R.J., Ohmann, J.L., Kennedy, R.E., Cohen, W.B., Gregory, M.J., Yang, Z., Roberts,H.M., Gray, A.N., Spies, T.A., 2015. Northwest Forest Plan–the first 20 years(1994–2013): status and trends of late-successional and old-growth forests.Gen. Tech. Rep. PNW-GTR-911. Portland, OR: U.S. Department of Agriculture,Forest Service, Pacific Northwest Research Station. 112 p.

Davis, R.J., Lint, J., 2005. Chapter 3: Habitat status and trend. In: Lint, J., tech. coord.,(Ed.). Northwest Forest Plan—the first 10 years (1994–2003): status and trendsof northern spotted owl populations and habitat. Gen. Tech. Rep. PNW-GTR-648. Portland, OR: U.S. Department of Agriculture, Forest Service, PacificNorthwest Research Station. pp. 21–82.

Dorr, J., 2007. Terrestrial Ecological Unit Inventory within the Winema NationalForest. Available from fs.fed.us/r6/data-library/gis/frewin/.

Dugger, K.M., Forsman, E.D., Franklin, A.B., Davis, R.J., White, G.C., Schwarz, C.J.,Burnham, K.P., Nichols, J.D., Hines, J.E., Yackulic, C.B., 2015. The effects ofhabitat, climate, and Barred Owls on long-term demography of NorthernSpotted Owls. The Condor 118, 57–116.

Everett, R., Schellhaas, D., Spurbeck, D., Ohlson, P., Keenum, D., Anderson, T., 1997.Structure of northern spotted owl nest stands and their historical conditions onthe eastern slope of the Pacific Northwest Cascades, USA. For. Ecol. Manage. 94,1–14.

Franklin, J.F., Johnson, K.N., Churchill, D.J., Hagmann, K., Johnson, D., and Johnston, J.,2013. Restoration of Dry Forests in Eastern Oregon: A Field Guide. Available atWww. Conservationgateway. Org. The Nature Conservancy, Portland, OR.

Franklin, J.F., Johnson, K.N., 2012. A Restoration Framework for Federal Forests inthe Pacific Northwest. J. For. 110, 429–439.

Fulé, P.Z., Swetnam, T.W., Brown, P.M., Falk, D.A., Peterson, D.L., Allen, C.D., Aplet, G.H., Battaglia, M.A., Binkley, D., Farris, C., Keane, R.E., Margolis, E.Q., Grissino-Mayer, H.D., Miller, C., Sieg, C.H., Skinner, C., Stephens, S., Stephens, S.L., Taylor,A.H., 2013. Unsupported inferences of high severity fire in historical westernUnited States dry forests: response to Williams and Baker. Glob. Ecol. Biogeogr.23, 825–830.

Gerdes, M.G., Carlton, A., O’Neil, C., Andrade, T., Maffei, H., Eglitis, A., Bonacker, R.,Byerly, D., Wesseler, L., 1995. Deschutes National Forest: A Late SuccessionalReserve Overview. US Department of Agriculture, Forest Service, PacificNorthwest Region, Deschutes National Forest. Available from fs.usda.gov.

Gray, A.N., McIntosh, A., Garman, S.L., 2014. Stand canopy cover prediction fromindividual tree measurements: stocking, crown width, and overlap functions.

Poster presented at IUFRO 2014 World Congress, October 5-11, 2014, Salt LakeCity, UT.

Hagmann, R.K., Franklin, J.F., Johnson, K.N., 2013. Historical structure andcomposition of ponderosa pine and mixed-conifer forests in south-centralOregon. For. Ecol. Manage. 304, 492–504.

Harrington, C.A., comp, 2003. The 1930s survey of forest resources in Washingtonand Oregon. Gen. Tech. Rep. PNW-GTR-584. Portland, OR: U.S. Department ofAgriculture, Forest Service, Pacific Northwest Research Station. 123 p.

Hatcher, W., Rondeau, S., Johnson, D.L., Johnson, K.N., Franklin, J.F., 2016. KlamathTribes—Managing their Homeland Forests in Partnership with the USDA ForestService. J. Forestry. Available on line December 15, 2016.

Heinrichs, J.A., Lawler, J.J., Schumaker, N.H., Wilsey, C.B., Bender, D.J., 2015.Divergence in sink contributions to population persistence. Conserv. Biol. 29,1674–1683.

Henson, P., Thrailkill, J., Glenn, B., Woodbridge, B., White, B., 2013. Using ecologicalforestry to reconcile spotted owl conservation and forest management. J. For.111, 433–437.

Hessburg, P.F., Churchill, D.J., Larson, A.J., Haugo, R.D., Miller, C., Spies, T.A., North,M.P., Povak, N.A., Belote, R.T., Singleton, P.H., 2015. Restoring fire-prone InlandPacific landscapes: seven core principles. Landscape Ecol. 30, 1805–1835.

Hessburg, P.F., Spies, T.A., Perry, D.A., Skinner, C.N., Taylor, A.H., Brown, P.M.,Stephens, S.L., Larson, A.J., Churchill, D.J., Povak, N.A., 2016. Tamm review:management of mixed-severity fire regime forests in Oregon, Washington, andNorthern California. For. Ecol. Manage. 366, 221–250.

Irwin, L.L., Rock, D.F., Rock, S.C., 2012. Habitat selection by northern spotted owls inmixed-coniferous forests. J. Wildl. Manage. 76, 200–213.

Jennings, S., Brown, N., Sheil, D., 1999. Assessing forest canopies and understoreyillumination: canopy closure, canopy cover and other measures. Forestry 72,59–74.

Johnson, D.L., 2016. Summaries of Canopy Cover Based Upon Imputed Stand ExamData in the Lobert Project Area. Unpublished report on file with the Fremont-Winema National Forest.

Johnson, K.N., Franklin, J.F., Johnson, D.L., 2008. A Plan for the Klamath Tribes’management of the Klamath Reservation Forest. Report to the Klamath Tribes.<www.klamathtribes.org>. Accessed August 23, 2012.

Kennedy, R.S.H., Wimberly, M.C., 2009. Historical fire and vegetation dynamics indry forests of the interior Pacific Northwest, USA, and relationships to northernspotted owl (Strix occidentalis caurina) habitat conservation. For. Ecol. Manage.258, 554–566.

Kolb, T., Agee, J., Fule, P., McDowell, N., Pearson, K., Sala, A., Waring, R., 2007.Perpetuating old ponderosa pine. For. Ecol. Manage. 249, 141–157.

Larson, A.J., Churchill, D., 2012. Tree spatial patterns in fire-frequent forests ofwestern North America, including mechanisms of pattern formation andimplications for designing fuel reduction and restoration treatments. For.Ecol. Manage. 267, 74–92.

Lehmkuhl, J., Gaines, W., Peterson, D.W., Bailey, J., 2015. Silviculture and monitoringguidelines for integrating restoration of dry mixed-conifer forest and spottedowl habitat management in the eastern Cascade Range. Gen. Tech. Rep. PNW-GTR-915. Portland, OR: U.S. Department of Agriculture, Forest Service, PacificNorthwest Research Station. 158 p.

Lindenmayer, D.B., Laurance, W.F., 2016. The ecology, distribution, conservation andmanagement of large old trees. Biol. Rev.

Lint, J., tech. coord., 2005. Northwest Forest Plan—the first 10 years (1994–2003):status and trends of northern spotted owl populations and habitat. Gen. Tech.Rep. PNW-GTR-648. Portland, OR: U.S. Department of Agriculture, ForestService, Pacific Northwest Research Station. 176 p.

Loehle, C., Irwin, L., Beebe, J., Fleming, T., 2011. Factors influencing the distributionof northern spotted owls in the eastern Cascades, Washington. Northwest.Natural. 92, 19–36.

Lutz, J.A., Larson, A.J., Swanson, M.E., Freund, J.A., 2012. Ecological importance oflarge-diameter trees in a temperate mixed-conifer forest. PLoS ONE 7, e36131.

MacCracken, J., Boyd, W., Rowe, B., Forest health and spotted owls in the easterncascades of Washington, 1996. In: North American Wildlife and NaturalResources Conference (USA).

McIntosh, A., Gray, A.N., Garman, S.L., 2012. Estimating canopy cover from standardforest inventory measurements in western Oregon. For. Sci. 58, 154–167.

Merschel, A.G., Spies, T.A., Heyerdahl, E.K., 2014. Mixed-conifer forests of centralOregon: effects of logging and fire exclusion vary with environment. Ecol. Appl.24, 1670–1688.

Morrow, R.J., 1985. Age structure and spatial pattern of old-growth ponderosa pinein Pringle Falls Experimental Forest, central Oregon.

Odion, D.C., Hanson, C.T., Arsenault, A., Baker, W.L., DellaSala, D.A., Hutto, R.L.,Klenner, W., Moritz, M.A., Sherriff, R.L., Veblen, T.T., 2014. Examining historicaland current mixed-severity fire regimes in ponderosa pine and mixed-coniferforests of western North America. PLoS ONE 9, e87852.

PRISM, 2012. United States Average Annual Precipitation, 1981–2010. PRISMClimate Group, Oregon State University. Created 10 July 2012. Available fromprism.oregonstate.edu. Last accessed 12/29/2012. The PRISM Group, OregonState University, Corvallis.

Schumaker, N.H., Brookes, A., Dunk, J.R., Woodbridge, B., Heinrichs, J.A., Lawler, J.J.,Carroll, C., LaPlante, D., 2014. Mapping sources, sinks, and connectivity using asimulation model of northern spotted owls. Landscape Ecol. 29, 579–592.

Shuffield, C.D., 2011. Overstory composition and stand structure shifts within inter-mixed ponderosa pine and lodgepole pine stands of the south-central Oregonpumice zone.

R.K. Hagmann et al. / Forest Ecology and Management 389 (2017) 374–385 385

Singleton, P.H., 2013. Barred owls and northern spotted owls in the easternCascade Range, Washington (Doctoral dissertation). Retrieved fromUniversity of Washington Research Works Archive. (http://hdl.handle.net/1773/22911).

Steel, Z.L., Safford, H.D., Viers, J.H., 2015. The fire frequency-severity relationshipand the legacy of fire suppression in California forests. Ecosphere 6, 1–23.

Stephens, S.L., Gill, S.J., 2005. Forest structure and mortality in an old-growth Jeffreypine-mixed conifer forest in north-western Mexico. For. Ecol. Manage. 205,15–28.

Stephens, S.L., Fulé, P.Z., 2005. Western pine forests with continuing frequent fireregimes: possible reference sites for management. J. For. 103, 357–362.

Stephens, S.L., Lydersen, J.M., Collins, B.M., Fry, D.L., Meyer, M.D., 2015. Historicaland current landscape-scale ponderosa pine and mixed conifer forest structurein the Southern Sierra Nevada. Ecosphere 6. art79.

Stine, P.A., Hessburg, P.F., Spies, T.A., Kramer, M.G., Fettig, C.J., Hansen, A.J.,Lehmkuhl, J.F., O’Hara, K.L., Polivka, K.M., Singleton, P.H., Charnley, S.,Merschel, A., White, R., 2014. The ecology and management of moist mixed-conifer forests in eastern Oregon and Washington: a synthesis of the relevantbiophysical science and implications for future land management. Gen. Tech.Rep. PNW-GTR-897. Portland, OR: U.S. Department of Agriculture, ForestService, Pacific Northwest Research Station. 254 p.

USDA, 2012. 36 CFR Part 219. National forest system land management planning.Federal Register, 77(68). 21162–21276.

USDA Forest Service, 1997. Late Seral Reserve Assessment for the Oregon EasternCascades Physiographic Province South of Crater Lake National Park. WinemaNational Forest Supervisor’s Office, Lakeview, OR.

USDA Forest Service, 2011. Master Stewardship Agreement Between KlamathTribes, Lomakatsi Restoration Project, The Nature Conservancy, and the USDAForest Service, Fremont-Winema National Forest. Available from fs.fed.us/spf/tribalrelations/documents/agreements/KlamathTribesMasterStewardship-Agreement20110928.pdf. Last accessed 4/8/2016.

USDA Forest Service, 2016. Analysis of Northern Spotted Owl Habitat Based uponStand Exam Data. On file Fremont-Winema National Forest.

USFWS, 2011. Revised Recovery Plan for the Northern Spotted Owl (Strixoccidentalis caurina). USDI Fish and Wildlife Service, Portland, OR. xvi-258.

USFWS, 2012. Endangered and Threatened Wildlife and Plants; Designation ofRevised Critical Habitat for the Northern Spotted Owl; Final Rule, 50 CFR, Part17, FWS–R1–ES–2011–0112; 4500030114, RIN 1018–AX69; Federal Register77, 233, Tuesday, December 4, 2012. Rules and Regulations; US GovernmentPrinting Office, Washington, DC, USA.

Walcott, C.D., 1900. Twenty-First Annual report of the Director of the United StatesGeological Survey. Part V - Forest Reserves 1899–1900, 209–498.

Williams, M.A., Baker, W.L., 2012. Spatially extensive reconstructions showvariable-severity fire and heterogeneous structure in historical westernUnited States dry forests. Global Ecol. Biogeogr. 21, 1042–1052.

Winthers, E., Fallon, D., Haglund, J., DeMeo, T., Ferwerda, M., Robertson, G., Gallegos,A., Rorick, A., Cleland, D., Robbie, W., 2005. Terrestrial Ecological Unit InventoryTechnical Guide: Landscape and Land Unit Scales. United States Department ofAgriculture Forest Service; Gen. Tech. Report W0-68.

Youngblood, A., Max, T., Coe, K., 2004. Stand structure in eastside old-growthponderosa pine forests of Oregon and northern California. For. Ecol. Manage.199, 191–217.