This article was downloaded by: [Universita degli Studi di Torino] On: 24 September 2012, At: 03:55 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Scandinavian Journal of Forest Research Publication details, including instructions for aut hors and subscription information: http:/ / www.tandfonline.com/ loi/ sfor20 Age and growth patterns of old Norway spruce trees in Trillemarka forest, Norway Daniele Castagneri a , Ken Olaf Storaunet b & Jørund Rolstad b a Department of AgroSelviTer, University of Turin, Grugliasco, Italy b Norwegian Forest Research Institute, Høgskoleveien 12, NO-1432, Ås, Norway

Accepted author version posted online: 10 Sep 2012.Version of record first published: 20 Sep 2012.

To cite this article: Daniele Castagneri, Ken Olaf Storaunet & Jørund Rolstad (2012): Age and growth patterns of old Norway spruce trees in Trillemarka forest, Norway, Scandinavian Journal of Forest Research, DOI:10.1080/ 02827581.2012.724082

To link to this article: http:/ / dx.doi.org/ 10.1080/ 02827581.2012.724082

PLEASE SCROLL DOWN FOR ARTI CLE

Full terms and conditions of use: http://www.tandfonline.com/page/terms-and-conditions This article may be used for research, teaching, and private study purposes. Any substantial or system atic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. The publisher does not give any warranty express or implied or make any representation that the contents will be complete or accurate or up to date. The accuracy of any instructions, formulae, and drug doses should be independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims, proceedings, demand, or costs or damages whatsoever or howsoever caused arising directly or indirectly in connection with or arising out of the use of this material. Scandinavian Journal of Forest Research 2012, 1Á9, iFirst article

ORIGINAL ARTICLE

Age and growth patterns of old Norway spruce trees in Trillemarka forest, Norway

DANIELE CASTAGNERI1, KEN OLAF STORAUNET2 & JØRUND ROLSTAD2

1Department of AgroSelviTer, University of Turin, Grugliasco, Italy, and 2Norwegian Forest Research Institute, Høgskoleveien 12, NO-1432 A˚s, Norway

Abstract Old trees represent key features of old-growth forests and are important elements for maintaining biodiversity. Due to extensive human exploitation of Fennoscandian boreal forests during several centuries, old Norway spruce trees have become exceedingly rare. We analysed 91 spruce trees in Trillemarka Nature Reserve, southern Norway, to investigate (1) the maximum age of living trees, (2) growth rates of different-age trees and (3) growth trends in very old trees. Increment cores were taken from trees in selected old-growth stands located at 700Á850 m a.s.l. Twelve spruce trees had an estimated total age of 400 years, the oldest one being 529 years and presumably the oldest known still living Norway spruce in northern Europe. A negative relationship between growth rate (basal area increment) and total age was observed, being most distinct for growth rates at 126Á275 years and less marked for early stage growth (26Á75 years). Thus, high age apparently was related more to low growth rates at adult and old stages of life rather than at the earlier stage. Among the trees 400 years, many of them did not show growth decrease with advancing age, indicating that ageing did not reduce growth. We conclude that the maximum age of stand-forming Fennoscandian Norway spruce trees would be in the range of 500Á600 years.

Keywords: Age, dendrochronology, longevity, Picea abies, tree growth.

Introduction ever, species-specific maximum ages can be strongly influenced by extreme growth conditions. For exam- Old trees are key components of old-growth forest, ple, Larson (2001) found that Thuja occidentalis L. constituting unique habitats for many rare species. can live for more than 1600 years on steep rocky cliffs, They also represent exceptional archives of past forest dynamics and climate (Arnan et al., 2012; Lie et al., but rarely exceeds 400 years on fertile sites. 2009; Nascimbene et al., 2009; Schweingruber & Recently, there has been increased interest in what Wirth, 2009). However, even in old-growth forests, controls the lifespan of trees, including possible

Downloaded by [Universita degli Studi di Torino] at 03:55 24 September 2012 individuals close to maximum age are extremely rare. effects of ageing (Pen˜uelas, 2005; Pen˜uelas & Such trees have always fascinated researchers, and Munne`-Bosch, 2010). Across taxa, there is evidence several surveys have been conducted to find the of a negative relationship between growth rate and oldest individuals of different tree species. World- lifespan (Metcalfe & Monaghan, 2003), indicating wide, there are about 20 species that exceed 1000 that slow growth prolongs life expectancy. Regarding years of age, with Pinus longaeva Bailey approaching possible age effects, some studies indicate that signs 5000 years in western US timberlines (Brown, 1996; of senescence (e.g. increase of hydraulic resistance, Currey, 1965; Schulman, 1958) as the oldest. In decrease of photosynthetic rate and reduction of Europe, the most widespread species reach xylogenesis duration) may be more related to size maximum ages between 400 and 700 years (Fagus than age (Mencuccini et al., 2005; Noode´n & sylvatica L., Picea abies (L.) Karst., Pinus sylvestris L., Leopold, 1988; Pen˜uelas, 2005; Petit et al., 2008). Quercus spp.) (Schweingruber & Wirth, 2009), Whereas studies on physiological ageing processes of whereas the European larch (Larix decidua Mill.) plant tissues are rather common, relationships may reach 1000 years (Nola & Motta, 1996). How- between age and growth at the level of whole-tree

Correspondence: Daniele Castagneri, Department of AgroSelviTer, University of Turin, I-10095 Grugliasco (TO), Italy, E-mail: [email protected]

(Received 29 March 2012; accepted 20 August 2012) ISSN 0282-7581 print/ISSN 1651-1891 online # 2012 Taylor & Francis http://dx.doi.org/10.1080/02827581.2012.724082 2 D. Castagneri et al.

have received less attention (Issartel & Coiffard, pine (P. sylvestris L.) dominates on low productive 2011; Matthes et al., 2008; Schweingruber & Wirth, sites, with a field layer of the Calluna vulgaris Á 2009). A reduced growth in gross volume as trees Vaccinium uliginosum type, often with some Norway approach old age is generally assumed (Duchesne spruce (P. abies) admixed in the understory. Norway et al., 2003; Pen˜uelas, 2005). However, as the spruce dominates on more productive slopes with a meristem regenerates every vegetative season, some field layer of Vaccinium myrtillus (L.). Deciduous tree authors point out that the concept of senescence as species (Alnus incana (L.) Moench, Betula pubescens an endogenously controlled degenerative process Ehrh., Populus tremula L., Sorbus aucuparia L.) occur does not apply to trees (Connor & Lanner, 1990; sporadically. Larson, 2001; Noode´n & Leopold, 1988; Pen˜uelas & Trillemarka forest represents one of a few areas Munne`-Bosch, 2010; Rossi et al., 2008). that host a certain number of very old Norway To elaborate on the age and growth patterns of spruce trees in southern Fennoscandia. Historically, Norway spruce (P.abies (L.) Karst.), one of the most the area has been utilised by human for centuries, widespread tree species in Europe, we performed a including summer dairy farming, grazing by domes- dendrochronological analysis in Trillemarka Nature tic animals, slash-and-burn cultivation and high- Reserve, southern Norway. Trillemarka is one of few grading timber harvesting (Mørch, 1954; Storaunet areas in northern Europe that are little influenced by et al., 2005a, 2005b; Toeneiet et al., 2007). The fire industrial forestry, and where a fair number of very history in Trillemarka shows an increased fire old spruce trees still persist. Our first goal was to frequency during the 1600s and 1700s due to assess the maximum age of Norway spruce trees in anthropogenic influence (Blanck, Storaunet & the Trillemarka forest. Establishing the longevity of Rolstad, unpublished data). However, far less rem- one of the pivotal tree species of the Eurasian boreal nant fire scarred pine stumps are found in the forest is important in understanding forest dynamics, spruce-dominated stands compared to the pine and as a reference for natural forest management forest. Thus, quite a few core areas of the Reserve and conservation. Second, we wanted to see if slow have been identified as little influenced by humans, growth is a prerequisite for getting very old. This has some of which showing high structural diversity already been observed in a few forest stands regarding tree composition, age and size distribu- (Di Filippo et al., 2012; Ro¨theli et al., 2012), and tion, and dead wood characteristics (Bendiksen, can have important repercussions in understanding 2004; Hofton, 2003, 2004). This illustrates the growth and mortality processes. Finally, we analysed gradient in historic human influence, where less trends in growth rate of trees approaching their utilised areas generally are found at higher altitudes maximum age, to see if ageing negatively affected and farther away from main valleys and historic growth. summer dairy farms. In such core areas we expected to find very old trees. Materials and methods Trillemarka Nature Reserve Field sampling and tree-ring analysis The study was conducted in Trillemarka Nature In order to sample the oldest Norway spruce trees in Reserve (608 05? N, 98 10? E), in Buskerud county, the forest, during 2003Á2009 we surveyed an area  Downloaded by [Universita degli Studi di Torino] at 03:55 24 September 2012 southern Norway. The Reserve was established in extending approximately 6 15 km, across a 150 m 2002 and enlarged in 2008, today covering 147 km2 elevation range (700Á850 m a.s.l.) close to the of forest and mountainous terrain between the altitudinal forest limit. Promising stands of spruce valleys of Numedal and Sigdal. Mean annual pre- forest with potential old trees were identified based cipitation is 750 mm, with snow covering the ground on air photos, a general forest inventory, and for 5Á6 months (Tessungdalen meteorological biologically important ‘‘core areas’’ previously station, 775 m a.s.l.). Mean annual temperature is mapped by Hofton (2003, 2004) and Bendiksen 0.98C, with a maximum in July (11.28C) and a (2004). These pre-selected stands (10Á50 ha) were minimum in January (7.98C) (Dagali meteorolo- visited and searched systematically for old trees. We gical station, 828 m a.s.l.). Topography is undulating identified and sampled supposedly old trees (10 with a Precambrian basement rock consisting of cm of diameter) based on external characteristics quartzite and granite, and with elements of gneisses (bark roughness, irregular crown, tree tapering, in the southern parts. The thickness of the morainic etc.). In the vicinity (Â0.1 ha) of these, we also material varies a lot. Locally, rich morainic deposits sampled supposedly younger individuals (also 10 give rise to luxuriant vegetation, but at higher cm). All stands were relatively undisturbed with an altitudes large areas have only thin depositions. Scots open-canopy structure (100Á300 trees ha1). Age and growth of old spruce trees 3

We extracted up to three increment cores per tree Regression analysis was used to investigate the at 40Á130 cm above ground level. Diameter at relationships between growth rate and total age. sampling height and at breast height (dbh) was Similar to other works (e.g. Black et al., 2008), recorded. Cores were shaved with a scalpel and BAI was calculated over 50-year periods. Shorter treated with zinc paste to enhance tree-ring borders. periods can be influenced by short-term variability Ring widths were measured with a micrometer to the due to inter-annual climatic fluctuations or distur- nearest 0.01 mm. Ring width series were cross-dated bances. On the other hand, longer periods may be against an available site chronology (Storaunet, too long to capture growth variations along the tree unpublished data) using TSAP package (Rinntech, lifespan. We also used regressions to infer possible Heidelberg, Germany), and also checked with the ageing effect on growth patterns, focusing on growth programme COFECHA (Holmes, 1983). Cores not trends during the last century in trees 400 years reliably cross-dated were excluded from the analyses. old. All analyses were done in SPSS 16.0 (SPSS Inc., For each tree, we used the core with most rings to Chicago, IL). estimate age and the core taken at breast height to analyse growth (Bigler & Veblen, 2009; Ro¨theli et al., 2012). For cores not intersecting the pith, its Results position was estimated by means of a graphical pith We analysed growth pattern of 91 trees, of which 12 locator (Applequist, 1958). Mean ring width of the had an estimated total age 400 years (Figure 1). 10 innermost rings was used to estimate the number The oldest tree had an estimated total age of 529 of missing rings. Cores estimated 30 mm from the years (501 rings counted) in 2009 (Figure 2). pith were discarded (Bigler & Veblen, 2009). To Multiple regression analysis showed that growth estimate total age of trees we conservatively added 2 rates and age distribution were negligibly influenced years per 10 cm growth in height (Granhus et al., by topographic site factors (all p values 0.10 2007; Kuuluvainen et al., 2002). except for age-altitude, p0.04). Expectedly, tree diameter correlated positively with total age, but the relationship was rather weak (R2 0.18; Figure 3). Growth pattern analysis Reconstruction of diameters (at breast height) along For the analysis of growth patterns, basal area the lifespan of trees illustrated that old trees had increment (BAI, mm2 year1) is generally preferred slower growth than younger ones (Figure 4). As an to ring width, since it is a two-dimensional measure example, mean dbh of trees 5300 years old was providing a better proxy for the three-dimensional 29.0 cm at the age of 200 years, whereas that of  mass increment (Visser, 1995). Therefore, growth trees 300 years old was 21.5 cm at the same age. pattern analyses were conducted with BAI calculated To estimate how the growth rate during certain using ring distance to pith, assuming that growth periods of the trees’ lifespan may influence total age, occurs proportionally around the stem. To achieve we regressed the estimated total tree age against normality, BAI was log transformed. Two growth BAI during successive 50-year growth periods. The variables were extracted from the material: growth rate and growth trend (Bigler & Bugmann, 2003). We defined growth rate as the mean of the annual BAIs

Downloaded by [Universita degli Studi di Torino] at 03:55 24 September 2012 within a certain period. Trends were calculated as the slope (beta coefficients) of local linear regres- sions fitted to BAI. Thus, positive coefficients indicate increasing growth rates, whereas negative coefficients indicate growth decrease. Sampled trees were located in relatively homoge- neous sites, but topographic conditions (slope, aspect and elevation) of the individual trees varied. To see if site factors influenced the age distribution, we performed a stepwise regression between estimated total tree age (dependent variable) and slope aspect

(cosine-transformed), slope steepness, and elevation 1450 1500 1550 1600 1650 1700 1750 1800 1850 1900 1950 2000 (independent variables) of each individual location. Calendar year To see if topographic characteristics influenced tree Figure 1. Sample distribution sorted by germination year growth, the same analysis was performed using (n91). Each line ranges from the estimated establishment year growth rate as dependent variable. to the year at which the core was extracted. 4 D. Castagneri et al.

Notably, the negative influence of growth rate appeared more distinct for the older age periods (126Á275 years) than for the younger ones (26Á125 years). Finally, we looked at possible ageing effects on the growth patterns, focusing on the 12 trees that were older than 400 years (Figure 6). None of the trees experienced critical low growth rates during the last decades. Concerning growth trends, we did not find a common pattern. During the last century, four trees experienced a growth decline, two had a constant growth, whereas six trees increased in growth. Considering the last 50 years, six trees declined whereas five increased in growth (Table I).

Discussion Maximum age Due to increased forest exploitation during the last centuries, very old trees are rare in Europe today and they are mainly located in old-growth forests in remote areas or at high elevation. In Fennoscandia, especially in the southern parts, natural forests harbouring old trees have become very rare (An- dersson & O¨ stlund, 2004; Jonsson et al., 2011). Still, Norway spruce trees exceeding 400 years have been recorded in a few regions of eastern Europe and the Alps (Janda et al., in prep.; Macar, 1973; Motta et al., 1999, 2011; Schweingruber & Wirth, 2009). In Figure 2. The oldest living Norway spruce sampled in the Fennoscandia, about a dozen spruce trees have been Trillemarka Reserve had an estimated total age of 529 years in reported to exceed 400 years of age (e.g. Andersson 2009. &O¨ stlund, 2004; Eidem, 1943; Lie et al., 2009; Rolstad et al., 1996; Wallenius et al., 2002) and two relationship between total age and growth rate was trees have exceeded 500 years (Kullman, 2000, negative for all age periods considered (Figure 5), 2001; Niklasson & Zielonka, 1999). Most of these although not statistically significant for the oldest are not still living, including the two oldest ones. The class (276Á325 years) due to low sample size. oldest tree in our sample (estimated total age of 529 years in 2009) is thus among the oldest Norway spruce trees recorded in northern Europe, and 65 Downloaded by [Universita degli Studi di Torino] at 03:55 24 September 2012 60 possibly the oldest known living today. Although 55 the ages reported in these studies are generally 50 comparable, they may however be in the lower end 45 of total tree ages since it is known that the oldest pith 40 dates of suppressed Norway spruces may be found 35 30 below the root collar (Niklasson, 2002). In contrast 25 to the 608-year-old tree dealt in the study of Kull-

Diameter (cm) 20 man (2000), which was a mountain krummholz, the 15 oldest individuals in our study were not located in 10 particularly harsh environments. Neither were they 5 characterised by unusual morphological characters, 0

0 such as multiple stems, miniature trees, clonal 50

100 150 200 250 300 350 400 450 500 550 individuals, krummholz, etc., even if a few of them, Estimated total age (years) including the oldest one, did have broken tops with Figure 3. Relationship between estimated total tree age and one or multiple secondary tops. On this basis, we diameter at breast height. conclude that the maximum age obtained by Age and growth of old spruce trees 5

65 60 55 50 45 40 35 30 25 20 15

Diameter at breast height (cm) 10 5 0 0 50 100 150 200 250 300 350 400 450 500 550 Estimated total age (years)

Figure 4. Development of individual tree diameter at breast height with an estimated age at germination point. Colours indicate gradients from lowest to highest age: yellow, 5200 years; orange, 201Á250 years; red, 251Á300 years; sky blue, 301Á350 years; blue, 351Á400 years; black, 400 years.

stand-forming Norway spruce trees in Fennoscan- mortality mainly depends on extrinsic causes dian boreal forest would be in the range of 500Á600 (disturbances). Slow-growing individuals tend to years. accumulate more chemical defences and have a higher wood density. Thus they are probably more tolerant to some pathogens and environmental Growth rate and age stresses, and are more likely to reach high ages In different animal and plant organisms, reduced (Larson, 2001; Loehle, 1996). Furthermore, slow- metabolic rates increase life expectancy (McCoy & growing trees reach large size later, thereby reducing Gillooly, 2008). Also among tree species, longevity size-selective mortality that involves large trees in old has been found to be negatively related to growth stands (Fraver et al., 2008; Lorimer et al., 2001). rates (Bigler & Veblen, 2009; Issartel & Coiffard, As we sampled only living trees, we do not know 2011; Kaufmann, 1996; Larson, 2001; Ro¨theli et al., their final ages (age of death). Thus, some of the 2012). However, besides physiological factors, other younger trees could reach high ages. Nevertheless, aspects should be taken into account, as tree our findings still indicate that older trees were

(a) y = -0.70x + 1.62 (b) y = -0.81x + 2.30 (c) y = -1.12x + 3.28 n = 78 n = 91 n = 81 R² = 0.062 R² = 0.098 R² = 0.188 p = 0.027 p = 0.002 p < 0.001 Downloaded by [Universita degli Studi di Torino] at 03:55 24 September 2012

(d) y = -0.91x + 2.85 (e) y = -1.47x + 4.31 (f) y = -0.58x + 2.02 n = 76 n = 53 n = 39 R² = 0.166 R² = 0.218 R² = 0.021 p < 0.001 p < 0.001 p = 0.384 Log estimated total age (years) totalestimated Log age (years) total estimated Log

Log mean basal area increment (cm2) Log mean basal area increment (cm2) Log mean basal area increment (cm2)

Figure 5. Relationships between growth rate (expressed as mean annual BAI of 50-year periods) at 26Á75 (a), 76Á125 (b), 126Á175 (c), 176Á225 (d), 226Á275 (e) and 276Á325 (f) years, and estimated total tree age. n is the number of trees included in analyses, R2 and p are the coefficient of determination and level of significance. Note that y-axes in upper and lower panels differ, and that both axes are log scaled. 6 D. Castagneri et al.

) 20 2 18 16 14 12 10 8 6 4 2

Annual basal area increment (cm 0 1500 1550 1600 1650 1700 1750 1800 1850 1900 1950 2000 Calendar year

Figure 6. Basal area increment at breast height of the 12 oldest sampled trees aligned by calendar year. BAI values are smoothed by 50-year cubic spline functions.

characterised by lower growth rates in comparison black spruce (Picea mariana (Mill.) B.S.P.) with a with younger ones at the same cambial age. Similar to lifespan of about 260 years. Tremblay et al. (2011) what has been observed in other species (Black et al., reported similar findings in uneven-aged stands of 2008; Di Filippo et al., 2012; Johnson & Abrams, black spruce with juvenile stage up to 120 years and 2009; Matthes et al., 2008), we conclude that a maximum age of 340 years. Bigler and Veblen Norway spruce trees need to grow at low to moderate (2009) indicated that high growth rate up to 60Á100 rates to reach high age. Reduced growth rates leading years reduced longevity in Norway spruce with to high ages are generally related to competition or to maximum age of approximately 300 years. In unfavourable microsites (Di Filippo et al., 2012; summary, these studies suggest that the growth Matthes et al., 2008; Robichaud & Methven, 1993). rate during the first third of the lifespan appears We believe that the latter was the main factor limiting influential on longevity. However, Ro¨theli et al. growth of old trees in the open-canopy forest of (2012) found that maximum growth rate along the Trillemarka, at the altitudinal limit of Norway spruce whole-tree lifespan was negatively related to long- in southern Norway. evity in Norway spruce. Our results showed that very Some studies of spruce species indicate that slow old trees were characterised by a low to moderate growth early in life increases the likelihood of growth rate up to 275 years. Although Ro¨theli et al. becoming old. Robichaud and Methven (1993) used maximum growth rate over periods of five indicated that longevity was positively affected by years, we used mean growth rate over periods of 50 suppression in the juvenile phase (up to 78 years) in years; both the results indicate that growth rate

Table I. Growth trends (simple regression analysis between BAI and calendar year) of the 12 oldest sampled trees for the periods 1910Á 2009 and 1960Á2009. BAI values were log transformed. Downloaded by [Universita degli Studi di Torino] at 03:55 24 September 2012

Growth trends 1910Á2009 Growth trends 1960Á2009

Tree id Estimated total age b R2 P Trend b R2 p Trend

O83 529 0.0056 0.51 B0.001  0.0059 0.48 B0.001  MK01 485 0.0021 0.12 0.001 Á0.0047 0.22 0.001  D21 479 0.0017 0.02 0.183 0.0067 0.21 0.002  T3Á24 447 0.0028 0.18 B0.001 Á0.0142 0.74 B0.001  O27 445 0.0033 0.24 B0.001 Á0.0112 0.57 B0.001  MKM01 441 0.0007 0.04 0.056 0.0020 0.09 0.045  T3Á23 425 0.0064 0.45 B0.001  0.0121 0.56 B0.001  O57 420 0.0065 0.49 B0.001 0.0023 0.08 0.052 O50 414 0.0039 0.32 B0.001  0.0117 0.72 B0.001  O36 413 0.0021 0.36 B0.001  0.0027 0.19 0.002  GS100 412 0.0055 0.50 B0.001  0.0091 0.37 B0.001  O61 410 0.0043 0.34 B0.001 0.0052 0.11 0.016 

b, R2, and p are the slope, the coefficient of determination and level of significance. Age and growth of old spruce trees 7

during a major part of tree life can affect likelihood Te¯rauds et al., 2011). Our analyses in Trillemarka to reach high ages. forest conform to previous studies indicating that prolonged slow growth is a prerequisite for trees to become really old, which in Norway spruce seems to Growth trends of trees approaching maximum age be in the range of 400Á600 years. Such a time period Due to the rarity of trees close to maximum species is much longer than ordinary rotation cycle in age, little is known about ageing effects on individual forestry practices, where stands 100Á120 years trees. Among the oldest Norway spruce trees in are considered over-mature (Te¯rauds et al., 2011). Trillemarka (400 years), only 4 out of 12 showed a With advancing age, our oldest spruce trees did not growth decline during the last century, whereas the show a growth decrease, indicating that ageing did others had a constant or increased growth. Addi- not affect growth. Thus, individual trees approach- tionally, none of these trees experienced critically ing their maximum age retain the potential to low growth rates during the last few decades, which accumulate biomass, an aspect to consider both in indicates that they could potentially reach even silviculture and in carbon storage models (Fish et al., higher ages. This pattern appears to be in contrast 2010). Furthermore, investigations on maximum with the theoretical sigmoid model, which assumes age and ageing provide new hints in understanding that growth rate increases in the early age, then mortality patterns and natural forest succession. plateaus, and finally declines during old age Aakala and Keto-Tokoi (2011) reported that pre- (Bollandsa˚s&Næsset, 2009; Duchesne et al., vious theories indicated senescence as an important 2003). However, a few studies concur with our mortality cause in natural stands, while recent findings. Kaufmann (1996) observed that growth studies suggest that allogenic small-scale distur- can either increase or decrease with age in old trees of Pinus ponderosa Dougl. Ex Laws. and Pinus bances primarily drive stand development in old contorta var. latifolia Englem., and Johnson and spruce forests. Our study concurs to such theories, Abrams (2009) and Di Filippo et al. (2012) found as mortality due to ageing per se seems to be an increased growth in the oldest trees of several extremely rare event. Few old trees in our study species. Disagreement between the sigmoid model had a decrease in growth, reflecting a healthy state and empirical observations could be related to at ages 400 years. differences in what regulates stand development In addition to their intrinsic value as rare indivi- and what affects individual growth. The modelled duals, old trees are important elements of old- age-related growth decline in productive forests is growth forests (Wirth et al., 2009), and their mainly related to crown closure, changes in stand identification is required to distinguish such stands. structure and resource allocation between trees of As very old trees are characterised by moderate increasing size, rather than to ageing of individual growth rates, they are not necessarily bigger than trees. Indeed, such decline is more evident in dense younger trees. Therefore, identifying the oldest trees stands than in open forests (Ryan et al., 1997; Smith should involve external characteristics such as bark & Long, 2001). Moreover, at the individual level, roughness, irregular crown, bole tapering, twisted age and size effects are often related, as trees get twigs, broken top and multiple secondary tops, taller and larger with age. Some physiological rather than merely large diameter or tall size. Once mechanisms that reduce tree vitality and growth, identified, stands with very old trees should be

Downloaded by [Universita degli Studi di Torino] at 03:55 24 September 2012 such as reduced photosynthesis to respiration ratio preserved by a hands-off reserve procedure. and increased hydraulic resistance, appear to be more related to tree size than to age (Pen˜uelas & Munne`-Bosch, 2010; Petit et al., 2008; Rossi et al., Acknowledgements 2008). Thus, the theoretical age-related growth Thanks to Erlend Rolstad and Ma˚lfrid Toeneiet for decline could actually be a size-related growth assisting with the fieldwork. We also wish to thank decline. However, disentangling these relationships two anonymous reviewers for useful comments on an was beyond the scope of our work. earlier version of the manuscript.

Conclusions References Knowledge about growth patterns and species long- Aakala, T. & Keto-Tokoi, P. (2011). The old Norway spruce evity is important in understanding growth patterns, forests of northern boreal Fennoscandia are alive and well: A review of Sire´n (1955). Scandinavian Journal of Forest mortality processes and natural succession of spruce Research, 26(Suppl. 10),25Á33. stands in boreal forests (Aakala & Keto-Tokoi, 2011; Andersson, R. & O¨ stlund, L. (2004). Spatial patterns, density Fraver et al., 2008; Kuuluvainen et al., 2002; changes and implications on biodiversity for old trees in the 8 D. Castagneri et al.

boreal landscape of northern Sweden. Biological Hofton, T. H. (2004). Supplerende biologiske registreringer i Conservation, 118, 443Á453. Trillemarka-Rollagsfjell [Supplementary biological inven- Applequist, M. B. (1958). A simple pith locator for use with off- tories of Trillemarka-Rollagsfjell]. Siste Sjanse-notat center increment cores. Journal of Forestry, 56, 141. 2004-2. Oslo: Siste Sjanse. (In Norwegian). Arnan, X., Lopez, B. C., Martinez-Vilalta, J., Estorach, M. & Holmes, R. L. (1983). Computer-assisted quality control in tree- Poyatos, R. (2012). The age of monumental olive trees (Olea ring dating and measurement. Tree-Ring Bulletin, 43,69Á78. europea) in northeastern Spain. Dendrochronologia, 30, Issartel, J. & Coiffard, C. (2011). Extreme longevity in trees: Live 11Á14. slow, die old? Oecologia, 165,1Á5. Bendiksen, E. (2004). Inventory of biological values in boreal Johnson, S. E. & Abrams, M. C. (2009). Age class, longevity and coniferous forests in connection with a conservation plan for growth rate relationships: Protracted growth increases in old Trillemarka Á Rollag Østfjell. NINA Oppdragsmelding, 830. trees in the eastern United States. Tree Physiology, 29, Trondheim: NINA, 67 pp. (In Norwegian with English 1317Á1328. abstract). Jonsson, B. G., Bru¯melis, G. & Kuuluvainen, T. (2011). Early Bigler, C. & Bugmann, H. (2003). Growth-dependent tree classical studies of forest ecology in Northern Europe. mortality models based on tree rings. Canadian Journal of Scandinavian Journal of Forest Research, 26(Suppl. 10),1Á2. Forest Research, 33, 210Á221. Kaufmann, M. R. (1996). To live fast or not: Growth, vigor and Bigler, C. & Veblen, T. T. (2009). Increased early growth rates longevity of old-growth ponderosa pine and lodgepole pine decrease longevities of conifers in subalpine forests. Oikos, trees. Tree Physiology, 16, 139Á144. 118, 1130Á1138. Kullman, L. (2000). The geoecological history of Picea abies in Black, B. A., Colbert, J. J. & Pederson, N. (2008). Relationship Northern Sweden and adjacent parts of Norway. A contra- between radial growth rates and lifespan within North rian hypothesis of postglacial tree immigration patterns. Á American species. Ecoscience, 15, 349 357. Geoo¨ ko, 21, 141Á172. ˚ æ Bollandsas, O. M. & N sset, E. (2009). Weibull models for Kullman, L. (2001). Postglacial immigration of spruce into single-tree increment of Norway spruce, Scots pine, birch Sweden. An old story in new light. Fauna och Flora, 96, Scandinavian Journal of and other broadleaves in Norway. 117Á128 (In Swedish with English summary). Á Forest Research, 24,54 66. Kuuluvainen, T., Ma¨ki, J., Karjalainen, L. & Lethonen, H. Brown, P. M. (1996). OLDLIST: A database of maximum tree (2002). Tree age distributions in old-growth forest sites in ages. In: J. S. Dean, D. M. Meko & T. W. Swetnam (Eds.), Vienansalo wilderness, Eastern Fennoscandia. Silva Fennica, Tree rings, environment, and humanity, pp. 727Á731, Tucson, 36, 169Á184. AZ: Radiocarbon, Proceedings of the International Larson, D. W. (2001). The paradox of great longevity in a short- Conference. lived tree species. Experimental Gerontology, 36, 651Á673. Connor, K. F. & Lanner, R. M. (1990). Effects of tree age on Lie, M. H., Arup, U., Grytnes, J. A. & Ohlson, M. (2009). The secondary xylem and phloem anatomy in stems of great importance of host tree age, size and growth rate as basin bristlecone pine (Pinus longaeva). American Journal of determinants of epiphytic lichen diversity in boreal spruce Botany, 77, 1070Á1077. forests. Biodiversity and Conservation, 18, 3579Á3596. Currey, D. R. (1965). An ancient bristlecone pine stand in eastern Loehle, C. (1996). Optimal defensive investments in plants. Nevada. Ecology, 46, 564Á566. Oikos, 75, 299Á302. Di Filippo, A., Biondi, F., Maugeri, M., Schirone, B. & Piovesan, Lorimer, C. G., Dahir, S. E. & Nordheim, E. V. (2001). Tree G. (2012). Bioclimate and growth history affect beech mortality rates and longevity in mature and old-growth lifespan in the Italian Alps and Apennines. Global Change hemlock-hardwood forests. Journal of Ecology, 89, 960Á971. Biology, 18, 960Á972. Macar, V. (1973). Studie o nejzna´meˇjsˇı´ch mohutnyˆch smrcı´ch Duchesne, L., Ouimet, R. & Morneau, C. (2003). Assessment of ´ sugar maple health based on basal area growth pattern. boubinskeho komplexu. [A study of the well-known giant Canadian Journal of Forest Research, 33, 2074Á2080. Norway spruce trees in the Boubin virgin forest]. Lasnictvı´, Á Eidem, P. (1943). U¨ ber Schwankungen im Dickenwachstum der 19, 341 354 (In Czech with English summary). Fichte (Picea abies) in Selbu, Norwegen [About variations in Matthes, U., Kelly, P. K. & Douglas, W. L. (2008). Predicting the the diameter growth of Norway spruce in Selbu, Norway]. age of ancient Thuja occidentalis on cliffs. Canadian Journal of Á Nytt Magasin for Naturvidenskapene B, 83, 145Á189 (In Forest Research, 38, 2923 2931. McCoy, M. W. & Gillooly, J. F. (2008). Predicting natural Downloaded by [Universita degli Studi di Torino] at 03:55 24 September 2012 German). Fish, T., Wilson, R., Edwards, C., Mills, C., Crone, A., mortality rates of plants and animals. Ecology Letters, 11, Kirchhefer, A. J., et al. (2010). Exploring for senescence 710Á716. signals in native Scots pine (Pinus sylvestris L.) in the Scottish Mencuccini, M., Martinez-Vilalta, J., Vanderklein, D., Abdul- Highlands. Forest Ecology and Management, 260, 321Á330. Hamid, H., Korakaki, E., Lee, S., et al. (2005). Size- Fraver, S., Jonsson, B. G., Jo¨nsson, M. & Esseen, P. (2008). mediated ageing reduces vigour in tall trees. Ecology Letters, Demographics and disturbance history of a boreal old- 8, 1183Á1190. growth Picea abies forest. Journal of Vegetation Science, 19, Metcalfe, N. B. & Monaghan, P. (2003). Growth versus lifespan: 789Á798. Perspectives from evolutionary ecology. Experimental Granhus, A., Hanssen, K. H. & Andreassen, K. (2007). Naturlig Gerontology, 38, 935Á940. foryngelse 11 a˚r etter lukket hogst i bla˚bærgranskog. Mørch, A. (1954). Sigdal og Eggedal. Bind III. Gard og slekt [Natural regeneration 11 years after partially cut stands in [Sigdal and Eggedal, Vol. 3. Farms and families]. Drammen: bilberry spruce forests]. Forskning fra Skog og landskap, 3/07, Sigdal og Eggedal historielag (In Norwegian). 27Á31 (In Norwegian with English abstract). Motta, R., Berretti, R., Castagneri, D., Dukic, V., Garbarino, M., Hofton, T. H. (2003). Trillemarka-Rollagsfjell: en sammenstilling Govedar, Z., et al. (2011). Toward a definition of the range av registreringer med hovedvekt pa˚ biologiske verdier. of variability of central European mixed FagusÁAbiesÁPicea [Trillemarka-Rollagsfjell: Field inventories with emphasis forests: The nearly steady-state forest of Lom (Bosnia and on biological values]. Siste Sjanse-rapport 2003Á5. Oslo: Herzegovina). Canadian Journal of Forest Research, 41, Siste Sjanse. (In Norwegian). 1871Á1884. Age and growth of old spruce trees 9

Motta, R., Nola, P. & Piussi, P. (1999). Structure and stand Ryan, M. G., Binkley, D. & Fownes, J. H. (1997). Age-related development in three subalpine Norway spruce (Picea abies decline in forest productivity: Pattern and process. Advances (L.) Karst.) stands in Paneveggio (Trento, Italy). Global in Ecological Research, 27, 213Á262. Ecology and Biogeography, 8, 455Á471. Schulman, E. (1958). Bristlecone pine, oldest known living thing. Nascimbene, J., Marini, L., Motta, R. & Nimis, P. L. (2009). National Geographic Magazine, 113, 355Á372. Influence of tree age, tree size and crown structure on lichen Schweingruber, F. H. & Wirth, C. (2009). Old trees and the communities in mature Alpine spruce forests. Biodiversity meaning of ‘old’. In: C. Wirth, G. Gleixner & M. Heimann and Conservation, 18, 1509Á1522. (Eds.), Ecological studies: Vol. 207. Old-growth forests: Func- Niklasson, M. (2002). A comparison of three age determination tion, fate and value,pp.35Á54, Berlin, Heidelberg: Springer. methods for suppressed Norway spruce: Implications for age Smith, F. W. & Long, J. N. (2001). Age-related decline in forest structure analysis. Forest Ecology and Management, 16, growth: An emergent property. Forest Ecology and 279Á288. Management, 144, 175Á181. ¨ Niklasson, M. & Zielonka, T. (1999). Norra Europas aldsta gran Storaunet, K. O., Rolstad, J., Gjerde, I. & Gundersen, V. S. Picea abies [The oldest Norway spruce in Northern Europe]. (2005a). Historical logging, productivity, and structural Á Svensk Botanisk Tidskrift, 93, 287 293 (In Swedish with characteristics of boreal coniferous forests in Norway. Silva English summary). Fennica, 39, 429Á442. Nola, P. & Motta, R. (1996). Una cronologia plurisecolare di Storaunet, K. O., Rolstad, J. & Toeneiet, M. (2005b). Urskog Larix decidua ’ larice ( Mill.) per l Alta Valmalenco (Sondrio, eller kulturskog Á skogshistorikk i Gamlevolltjønnan i Italia) [A multi-century chronology of larch (Larix decidua Trillemarka [Primeval or cultivated forest Á Forest history Mill.) for the Alta Valmalenco (Sondrio, Italia)]. at Gamlevolltjønnan in Trillemarka]. Aktuelt fra skogforsknin- Dendrocronologia, 14,31Á42 (In Italian with English gen, Norwegian Forest Research Institute, 5/05,27Á32 summary). (In Norwegian). Noode´n, L. D. & Leopold, A. C. (1988). Senescence and aging in Te¯rauds, A., Bru¯melis, G. & Nikodemus, O. (2011). Seventy-year plants. San Diego: Academic Press. changes in tree species composition and tree ages in state- Pen˜uelas, J. (2005). A big issue for trees. Nature, 437, 965Á966. owned forests in Latvia. Scandinavian Journal of Forest Pen˜uelas, J. & Munne`-Bosch, S. (2010). Potentially immortal? Research 26 Á New Phytologist, 187, 564Á567. , , 446 456. ˚ Petit, G., Anfodillo, T. & Mencuccini, M. (2008). Tapering of Toeneiet, M., Rolstad, J. & Storaunet, K. O. (2007). 600 ars fi xylem conduits and hydraulic limitations in sycamore (Acer brannhistorikk i Trillemarka [A 600 years forest re history Á pseudoplatanus) trees. New Phytologist, 177, 653Á664. in Trillemarka]. Norsk Skogbruk, 53(3), 20 22 (In Robichaud, E. & Methven, I. R. (1993). The effect of site quality Norwegian). on the timing of stand breakup, tree longevity, and the Tremblay, M. J., Rossi, S. & Morin, H. (2011). Growth dynamics maximum attainable height of black spruce. Canadian of black spruce in stands located between the 51st and 52nd Journal of Forest Research, 23, 1514Á1519. parallels in the boreal forest of Quebec, Canada. Canadian Rolstad, J., Rolstad, E. & Groven, R. (1996). Nordens eldste Journal of Forest Research, 41, 1769Á1778. gran? [The oldest Norway spruce in the Nordic countries?]. Visser, H. (1995). Note on the relation between ring widths and Blyttia, 54,7Á8 (In Norwegian). basal area increments. Forest Science, 41, 297Á304. Rossi, S., Deslauriers, A., Anfodillo, T. & Carrer, M. (2008). Age- Wallenius, T. H., Kuuluvainen, T., Heikkila¨, R. & Lindholm, T. dependent xylogenesis in timberline conifers. New (2002). Spatial tree age structure and fire history in two old- Phytologist, 177, 199Á208. growth forests in Eastern Fennoscandia. Silva Fennica, 36, Ro¨theli, E., Heiri, C. & Bigler, C. (2012). Effects of growth rates, 185Á199. tree morphology and site conditions on longevity of Norway Wirth, C., Gleixner, G. & Heimann, M. (Eds.). (2009). Old- spruce in the northern Swiss Alps. European Journal of Forest growth forests: Function, fate and value. Ecological studies, Vol. Research, 131, 1117Á1125. 207. Berlin, Heidelberg: Springer. Downloaded by [Universita degli Studi di Torino] at 03:55 24 September 2012