Effects of land management on large trees and carbon stocks

Introduction Conclusions References

ide in the atmosphere would be much faster than currently observed (Reich, 2011). The large potential of trees for either removing CO 2 from the atmosphere or adding it was discovered in early research about forests, the carbon cycle, and climate (Dyson, 1976;Brown and Lugo, 1982;Cooper, 1983;Woodwell et al., 1983). More recently, research has highlighted other mechanisms about how forest canopies affect the radiative forc- 5 ing of the atmosphere by modifying the albedo (Betts, 2000) and evapotranspiration (Swann et al., 2010). Global forests are extremely diverse and provide a variety of ecosystem services such as carbon sequestration, industrial raw materials, flood and landslide protection, biodiversity preservation, and aesthetic and health benefits (Pan et al., 2013). Forests 10 are usually defined by the presence of trees and absence of non-forest land use, even though trees are also numerous outside forests in savannas, pasture lands, and in suburban areas and green city centers (Nowak and Greenfield, 2012). Large and old trees are exceptional entities in most tree populations and they have unique and special qualities beyond their climate mitigation function, yet there are concerns about human- 15 caused losses of old and large trees of the world (Lindenmayer et al., 2012).
While covering only about one quarter of the global land surface, forests dominate the net removal of CO 2 from the atmosphere into land ecosystems (Pan et al., 2011a). Live vegetation, mostly trees, accounts for three quarters of the large and persistent sink of the global forests. The remaining one quarter is shared by dead wood, litter, 20 soils and harvested wood products (Pan et al., 2011a).
In mix-aged forests, large trees are often a significant proportion of aboveground biomass and the carbon density of the site although only a few may be present (Slik et al., 2013;Luz et al., 2011;Martínez and Alvarez, 1998). Large trees also have statistically lower mortality rates compared to small sized trees (Coomes and Allen, Introduction Conclusions References Tables  Figures   Back  Close Full Screen / Esc Printer-friendly Version

Interactive Discussion
Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | biodiversity and integrity (Lindenmayer et al., 2012). They also provide other additional benefits, such as bringing recreational opportunities and engaging symbolic values in many cultural heritages (Blicharska and Mikusinski, 2013). The size distribution of large trees and their long-term dynamics can be detected by surveying the demography of tree populations at annual or multi-annual time steps 5 using forest inventory methodology (Lawrence et al., 2010). Observations on trees size distributions based on forest inventory measurements have not been compiled at the global level. However, regional data and time series are available from Finland and the United States, which provide multi-decadal statistics on the demography of tree populations and changes of the size-distribution of trees. The evolution of forest 10 carbon for Finland and the United States has been described more broadly in Liski et al. (2006) and Birdsey et al. (2006), respectively. The longest time series to our knowledge on statistically representative measurements of timber resources is from a sub region in Finland, where the fieldwork was initiated in 1912 (Kauppi et al., 2010). The first national forest inventories from Finland were carried out in the 1920s and 15 1930s (Ilvessalo, 1927Ilvessalo, 1942), and the national forest inventory in the United States was begun in the 1930s (LaBau et al., 2007).
We focus on large trees, broadly defined as the upper end of the size distribution of live trees. The objective of this research is to analyze the role of large trees in the evolution of the growing stock in regions within Finland and the United States rep-20 resenting different land management histories using data from statistically designed sample surveys. We discuss the impact of large trees on the carbon budget, albedo and evapotranspiration of forests, and the effect of land management on the stock of large trees. 25 Forest inventory is based on measurements taken from a statistically representative sample of all trees within a forest region -for details of this approach, see Tomppo  LaBau et al. (2007). Historical inventory data are available at approximately decadal time intervals from Finland (Kuusela, 1972(Kuusela, , 1978Kuusela and Salminen, 1991;Tomppo et al., 2011;Ylitalo, 2011Ylitalo, , 2012Korhonen et al., 2013) and from the United States (Smith et al., 2009). We extracted data from these published inventories specifically by five sub regions (Fig. 1). The two regions of Finland combined 5 equal all Finland, whereas for the United States we selected three diverse regions. We also looked at nation-wide forests of the United States compiling statistics of tree size distributions divided as hardwoods (deciduous trees) and softwoods (conifers). We prepared time series estimates of the growing-stock volume of large trees and the distribution of growing-stock volume by tree-size classes. Growing stock-volume (in cubic meters, m 3 ) refers to the volume of the tree stem as defined by common merchantability standards. The historical inventories report estimates of growing-stock volume based on consistent definitions.

Materials and methods
The volume of growing stock is correlated with the vegetation carbon stock. The ratio of carbon stock/growing stock decreases with tree size; in other words, the contribu-15 tion of stem biomass becomes increasingly large as trees grow in size (e.g. Lehtonen et al., 2004;Jalkanen et al., 2005;Kauppi et al., 2006). Tree size distributions were constructed based on Diameter at Breast Height (DBH) to separate cohorts of trees representing different size classes. The total stem volume (in millions of m 3 ) was estimated for trees within each size cohort and each region. We also included estimates 20 of growing-stock volume of all size classes for both countries to highlight the general trends. Data from Finland referred to nine inventory cycles as follows : 1951-1953 => 1960-1963 => 1964-1970 => 1971-1976 => 1977-1984 => 1986-1994 => 1996-2003 => 2004-2008 => 2009-2012. Finnish data were analysed separately for 25 two regions (southern Finland -about 11.3 million forest ha; and northern Finland -about 11.5 million forest ha; Fig. 1). Measurement teams travel within and across regions. During some years measurements are taken in southern but not in northern Finland, and vice versa. The main part of rural lands in southern Finland has been in Introduction Three sub regions within the United States were selected for more detailed analysis covering a total of 91 million ha ( Fig. 1; Table 1). Regions represent the diverse history of land management and impacts on large trees: the northeast which is largely composed of forests that are re-growing on agricultural land that was abandoned over the last century; the southeast where much of the forest land is intensively managed on 10 short rotations for timber products; and the Pacific Northwest where old-growth forests were still being cleared and regenerated through the mid-1980s but have since been preserved.
Detailed analyses have been published elsewhere on the conversion from growing stock to biomass and carbon stock. A key concept in this conversion is Biomass Ex- 15 pansion Factor (BEF), which has been empirically determined for many tree species and for many regions of the world. An analysis for Finland is available in (Eerikäinen 2009) and in Härkönen et al. (2011).). The biomass for each tree component is modeled by the main tree species (Repola, 2008(Repola, , 2009. Thereafter, BEFs were calculated by dividing total tree biomass by stem volume within tree diameter classes (Lehtonen  11,2014 Effects of land management on large trees and carbon stocks

Results for Finland
The growing stock of Finland's forests accumulated from 1400 million m 3 in [1960][1961][1962][1963] to the latest estimate of about 2300 million m 3 in 2009-2012 (= +60 %). Even though the rate of accumulation was almost the same in southern Finland as in northern Fin-5 land, there were interesting differences between the two regions in the development of the tree size distributions. The stock of large trees hardly changed in northern Finland, where the accumulation of biomass and carbon was concentrated in small trees less than 30 cm in DBH. In southern Finland the growing stock of large trees increased nearly five-fold from about 70 to 340 million m 3 between 1951-1953 and 1996-2003, 10 respectively (Table 2). Even though the volume accumulated in southern Finland, the share of trees > 30 cm DBH was still less than one quarter of the growing stock in 2009-2012 (Fig. 2). The growing stock of Finland's forests consisted predominantly of small and medium sized trees (≤ 30 cm DBH). The stock of such small and medium sized trees increased 15 from about 1230 in the 1960s to 1840 million m 3 in 2009-2012. The relative contribution of large trees to the total growing stock first increased from 10 to 21 % between 1951-1953 and 1996-2003; then declined to 19 % in 2009-2012.

Results for the United States
Trees in the United States are larger in general than those in Finland. In the United 20 States, trees larger than 33 cm in DBH account for more than half of the growing stock. The growing-stock volume of the largest trees in the United States (DBH > 53 cm) declined from 5.9 to 4.6 billion m 3 between 1953 and 1987 and then recovered to 5.9 billion m 3 by 2007 (Fig. 3). Large softwood trees in old-growth forests were still being harvested until the late 1980's especially in the Pacific Northwest region, after which 25 most harvesting of old-growth was stopped to preserve their remaining areas for con- Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | servation goals. In contrast, the volume of the largest hardwood trees, which are more common in the eastern United States, has increased steadily since 1953 on lands that were abandoned from agricultural use and now have forests that are maturing. In the Pacific Northwest region which is dominated by softwood species and has the largest population of larger trees in the United States, there is a contrasting pattern of 5 change over time in the growing stock of trees greater than 33 cm compared with trees less than 33 cm (Tables 3 and 4). In a pattern similar to the national totals, the growing stock of large trees in the Pacific Northwest declined from 1953 to 1987 and then nearly recovered to their prior stocking by 2007, reaching 3.4 billion m 3 . In contrast, the growing stock of trees less than 33 cm increased from 1953 to 1977 and then 10 stabilized at about 1.1 billion m 3 . In the Southeast United States, the growing stock of trees greater than 33 cm doubled between 1953 and 2007, while that of trees less than 33 cm increased only until 1977 then was relatively stable. These changes reflect the increasing influence of industrial plantation forestry over the period. In the Northeast United States where hardwoods predominate, the pattern is similar to the southeast 15 except that the growing stock of trees greater than 33 cm more than tripled between 1953 and 2007, indicating forests that are increasing in age coupled with the absence of significant harvesting or stand-replacing natural disturbances.

Discussion and conclusions
In the southern Finnish region during the second half of the 20th century, the change 20 in growing stock was mainly driven by the expansion of the large trees cohort. Likewise regarding hardwoods in the United States, the increase in growing stock of trees larger than 33 cm in DBH accounted for most of the density change since 1977. Northern Finland was an exception among the five study regions in the sense that the growing stock of the large tree cohort did not increase during the last six decades. 25 In the United States there have been regional differences in the role of large trees as well as differences between softwoods and hardwoods. Where hardwoods are pre- dominant in the eastern part of the country, there has been a steady and significant increase in biomass of large trees since 1953. Softwood biomass in the southeast has increased mainly in the middle diameter classes; whereas softwood biomass in large trees of the Pacific Northwest declined and then increased since 1987. Globally, forest vegetation has become increasingly dense (Rautiainen et al., 2011). 5 Forest biomass and the carbon stock have expanded even though the global forest area keeps decreasing (Pan et al., 2011a). Ultimately in the successional process, the biomass of forest ecosystems saturates, if losses in tree mortality match the gains of biomass increment. In central Europe, the forest carbon sink was recently reported to show early signs of saturation, likely because large areas of forests have approached 10 their maturity (Nabuurs et al., 2013). The five study regions of this research in Finland and USA did not show similar signs. Forest vegetation biomass has increased in many parts of the world, removing carbon dioxide from the atmosphere, and accounting for three quarters of the estimated gross forest sink of about 4 Pg C yr −1 (Pan et al., 2011a). Trees contain carbon in stems, 15 branches, foliage and roots and provide carbon for the stocks of forest soils, and downstream water and sediments (Richey et al., 2002). Large trees have deep roots, which transfer carbon into the lowest layers of the forest soil. As trees die, the largest individual trees decay most slowly thus maintaining for some time the carbon stocks in snags or coarse woody debris that are standing or laying in forests (Harmon and Hua, 1991;20 Krankina and Harmon, 1995; Pan et al., 2011a). Trees may be transformed into wood based products, which contain significant amounts of carbon and are widely used as a sustainable raw material (Skog et al., 2004). The interaction of forests with the climate system is complex. In high latitudes as biomass expands in areas with low forest density, the mitigating impact of carbon se-25 questration on the radiative forcing of the atmosphere may be offset by the unwanted effects of decreasing albedo and increasing transpiration on radiative forcing (Betts, 2000;Bonan, 2008;Swann et al., 2010). Promoting the expansion of large trees on existing forest areas would selectively favor carbon sequestration with little or no impact BGD 11,2014 Effects of land management on large trees and carbon stocks on albedo or transpiration. Inside the surface layers of the stem, large trees contain heartwood which is rich in carbon but does not contribute to transpiration nor have an albedo impact. Forests of the world are diverse and the definition of "big tree" is perhaps only relative. The biggest tree measured by the stem volume is the General Sherman growing 5 in Sequoia National Park, California. It contains about 1500 m 3 of stem wood and thus 300-400 t C (Alaback, 1991;Pan et al., 2013). In the boreal forests of Finland, all trees are smaller than 20 m 3 in stem volume, and a tree with a stem of just two m 3 can be considered "large". On average, trees in the United States are larger than those in Finland, and exceptionally large in global comparisons are the conifers of Pacific region of the United States (Waring and Franklin, 1979). Measuring carbon sequestration directly is difficult, and in this research we rely on the relationship between growing-stock volume and tree biomass to estimate the vegetation carbon content (Brown et al., 1989;Lehtonen et al., 2004;Härkönen et al., 2011). Carbon sequestration of forest ecosystems is positively and strongly correlated with the 15 accumulation of biomass into tree stems. As the size distribution of trees shifts to larger diameters and stem volumes, large trees gain relative importance, and the carbon stock of forest vegetation accumulates thus sequestering CO 2 from the atmosphere.
In temperate zones and also in Finland, most natural forests have been greatly altered by human activities (Pan et al., 2013;Fritzbøger and Søndergaard 1995). The 20 current forest demography and tree-size distributions generally reflect such a disturbance legacy (Clawson, 1979;Foster et al., 1998;Pan et al., 2011b). The patterns of land use were driven by a historical switch from subsistence agriculture to modern farming many decades ago. Forest transition subsequently switched the sign of change of forests from sources to sinks of carbon as the timber stock started to in-25 crease (Mather, 1992;Rautiainen et al., 2011). The rate of biomass accumulation as a national average in both countries has been 0.5-1.0 per cent per year consistently for many decades. 11,2014 Effects of land management on large trees and carbon stocks In Finland, the forest resources have been utilized intensively for centuries. Tar production -a heavy consumer of stem wood -had an important role in the foreign trade of Finland until the 1820s. Shifting cultivation, a primitive form of agriculture, had a highly destructive effect on the forests of southern Finland. It was widely practiced until the late 19th century on fertile forest lands, which have the highest potential for producing 5 large trees. In reporting the results of the first national forest inventory, Ilvessalo (1927) emphasized the detrimental effect of cattle grazing on the forests of southern Finland. In Finland's exports, tar was replaced by saw mill products, which continued to be the single-most important product group until the early 20th century, only to be gradually replaced by products of pulp and paper industries. Forests situated near reasonable transport routes were heavily exploited. Cutting methods were based on removing the largest and most valuable trees and the remaining tree populations were usually incapable of fully utilizing the growth potential of the site (Fritzbøger and Søndergaard 1995).

BGD
The United States saw a general progression of harvesting large trees from the east 15 to the west over the last 150 yr. Beginning 100 yr or so ago, eastern forests began regrowing on abandoned agricultural land and now these forests have reached an age where large trees are becoming common once again (Pan et al., 2011b). In the far western US, large trees were still being harvested in significant quantities until about 25 yr ago so their numbers were declining, but now that trend has been reversed with 20 most remaining old-growth forest areas set aside for other purposes besides timber production. Southern Finland and the hardwood region in eastern United States are similar in their forest history with severe forest degradation especially in the 19th century followed by the recovery in more recent decades. Hence the expansion of large tree biomass in 25 these two regions during the latter part of the 20th century indicates a recovery from man-made degradation and disturbances of the past.
Forest carbon density can increase in two alternative or mutually reinforcing ways: (1) as trees in forests become more numerous; or (2) as the average tree of the forest BGD Introduction

Tables Figures
Back Close

Full Screen / Esc
Printer-friendly Version

Interactive Discussion
Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | becomes larger over time. New research is needed here, since this analysis cannot fully separate the relative contribution of the two mechanisms. Understanding the dynamics of tree populations is important for efforts not only to predict the impact on forest carbon balance, but also for evaluating forest structure and function related to other ecosystem services. For instance, in Pacific Northwest, some at-risk plant and animal 5 species need different habitats that are associated with old forests or more opened forests (Buchanan et al., 2010). Large trees often play key ecological roles, such as characterizing forest growth form and structural complexity, and they are essential for ecosystem stability and biodiversity (Lindenmayer et al., 2012). Because of unique ecological roles and ecosystem services that large and old trees provide, the reported 10 global decline of such trees has drawn many concerns about how to balance the need of conservation and forest management (Aerts, 2013;Lindenmayer et al., 2013).
We have shown in this research that an expansion of the stem biomass of large trees, as measured in m 3 of the growing stock volume, has been an important driver of biomass expansion in some areas over certain periods of time. However, the dynamics Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Fritzbøger, B. and Søndergaard, P.: A short history of forest uses, In: Multiple-Use Forestry in the Nordic Countries, edited by: Hytönen, M., METLA, Finnish Forest Research Institute, Helsinki Research Centre, 11-41, 1995. Härkönen, S., Lehtonen, A., Eerikäinen, K., Peltoniemi, M., and Mäkelä, A.:  Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Nabuurs, G. J., Lindner, M., Verkerk, P. J., Gunia, K., Deda, P., Michalak, R., and Grassi, G.: First signs of carbon sink saturation in European forest biomass, Nature Climate Change, 3, 792-796, 2013. Nowak, D. J. and Greenfield, E. J.: Tree and impervious cover change in US cities, Urban For.