Global Change Biology Global Change Biology (2014) 20, 2019–2030, doi: 10.1111/gcb.12512

Forest biomass carbon sinks in East Asia, with special reference to the relative contributions of forest expansion and forest growth JINGYUN FANG1,2, ZHAODI GUO1,3, HUIFENG HU2, TOMOMICHI KATO4, H I R O Y U K I M U R A O K A 5 and Y O W H A N S O N 6 1 Department of Ecology, College of Urban and Environmental Science, and Key Laboratory for Earth Surface Processes of the Ministry of Education, Peking University, Beijing, 100871, China, 2State Key Laboratory of Vegetation and Environmental Change, Institute of Botany, Chinese Academy of Sciences, Beijing, 100093, China, 3National Satellite Meteorological Center, China Meteorological Administration, Beijing, 100081, China, 4Laboratoire des Sciences du Climat et de l’ Environnement, IPSL, CEA-CNRS-UVSQ, Orme des Merisiers, Gif sur Yvette, 91191, France, 5Institute for Basin Ecosystem Studies, Gifu University, 1-1 Yanagido, Gifu, 501-1193, Japan, 6Division of Environmental Science and Ecological Engineering, Korea University, Anam-dong, Sungbuk-ku, 136-701, Seoul Korea

Abstract Forests play an important role in regional and global carbon (C) cycles. With extensive afforestation and reforestation efforts over the last several decades, forests in East Asia have largely expanded, but the dynamics of their C stocks have not been fully assessed. We estimated biomass C stocks of the forests in all five East Asian countries (China, Japan, North Korea, South Korea, and Mongolia) between the 1970s and the 2000s, using the biomass expansion factor method and forest inventory data. Forest area and biomass C density in the whole region increased from 179.78 9 106 ha and 38.6 Mg C ha1 in the 1970s to 196.65 9 106 ha and 45.5 Mg C ha1 in the 2000s, respectively. The C stock increased from 6.9 Pg C to 8.9 Pg C, with an averaged sequestration rate of 66.9 Tg C yr1. Among the five countries, China and Japan were two major contributors to the total region’s forest C sink, with respective contributions of 71.1% and 32.9%. In China, the areal expansion of forest land was a larger contributor to C sinks than increased biomass density for all forests (60.0% vs. 40.0%) and for planted forests (58.1% vs. 41.9%), while the latter contributed more than the former for natural forests (87.0% vs. 13.0%). In Japan, increased biomass density dominated the C sink for all (101.5%), planted (91.1%), and natural (123.8%) forests. Forests in South Korea also acted as a C sink, contributing 9.4% of the total region’s sink because of increased forest growth (98.6%). Compared to these countries, the reduction in forest land in both North Korea and Mongolia caused a C loss at an average rate of 9.0 Tg C yr1, equal to 13.4% of the total region’s C sink. Over the last four decades, the biomass C sequestration by East Asia’s forests offset 5.8% of its contemporary fossil-fuel CO2 emissions. Keywords: biomass density, biomass expansion factor, carbon sink, China, East Asia, forest area, forest inventory, Japan, Mongolia, North Korea, South Korea Received 14 June 2013 and accepted 15 November 2013

Introduction As the largest part of terrestrial ecosystems, forests occupy around 30% of the global land surface with about 4.2 9 109 ha (Kramer, 1981; Bonan, 2008). It is estimated that over 80% of terrestrial vegetation carbon (C) is stored in forests (Pan et al., 2011), and the annual C flux between forests and the atmosphere through photosynthesis and respiration accounts for 50–90% of the total annual flux of terrestrial ecosystems (Winjum et al., 1993; Malhi et al., 2002). Because of their capacity for C Correspondence: Jingyun Fang, tel. + 86 10 6276 5578, fax +86 10 6275 6560, e-mails: [email protected]; [email protected]

© 2014 John Wiley & Sons Ltd

storage and high productivity, forest ecosystems play a dominant role in the global C cycle (IPCC, 2007; Pan et al., 2011). The Kyoto Protocol, which was approved in the 1997 United Nations (UN) meeting on climate change, clearly suggested increasing C sequestration through afforestation and reforestation to meet greenhouse gas emission targets (Brown et al., 1999). Therefore, estimation of forest biomass C sinks or sources and their spatial and temporal distributions is both of scientific and of political importance (Watson et al., 2000; Fang et al., 2001; Janssens et al., 2003; Nabuurs et al., 2003; Birdsey et al., 2006; McKinley et al., 2011). Since the early 1970s, regional and national forest inventories have been carried out across the world and provide data for estimating forest biomass on a regional 2019

2020 J . F A N G et al. or national scale (e.g., Brown & Lugo, 1984; Kauppi et al., 1992; Turner et al., 1995; Schroeder et al., 1997; Fang et al., 1998, 2001, 2005, 2007; Brown & Schroeder, 1999; Choi et al., 2002; Goodale et al., 2002; Liski et al., 2003; Li et al., 2010; Pan et al., 2011). Using forest inventory data and long-term ecosystem C studies, Pan et al. (2011) recently estimated that the current C stock in the world’s forest ecosystems was 861  66 Pg C (1 Pg = 1015 g), with 363  28 Pg C (42%) in above- and belowground live biomass. Global forests have functioned as a significant C sink over the last two decades, but there exists a large regional and temporal difference in the magnitude of that sink (Pan et al., 2011). Therefore, detailed assessment of regional forest C sinks/ sources and their spatial and temporal distributions is necessary for understanding the dynamics, processes, and mechanisms of the terrestrial C cycle (Goodale et al., 2002; Birdsey et al., 2006; Fang et al., 2010; McKinley et al., 2011). During the past decade, several key regional programs such as the North American Carbon Program, Carbon Europe Integrated Project, and African Carbon Project have been established to clarify regional C budgets and have greatly contributed to knowledge of the C cycle in their respective regions (Fang et al., 2010; http://www.globalcarbonproject. org/carbontrends). East Asia, which includes the five countries of China, Japan, Democratic People’s Republic of Korea (henceforth referred to North Korea), Republic of Korea (South Korea), and Mongolia, is located at the eastern margin of the Eurasian Continent and the western coast of the Pacific Ocean (Fig. 1). This area has a population

of more than 1.5 billion and covers about 1.2 9 109 ha land area, of which 22% (254.6 9 106 ha) is occupied by forests (FAO, 2010a; Table S1). As one of the most active regions in the global economy, it is of great importance both in scientific understanding of the region’s forest C cycle and in the use of appropriate forest management strategies. Characterized by a warmhumid climate that is under the influence of the Asian monsoon, East Asia has abundant forest types that range from tropical rainforests and evergreen broadleaf forests in the south to boreal forests in the north, providing a model for exploring heterogeneity of ecological attributes of C cycle. East Asia has experienced extensive afforestation and reforestation activities over the past several decades, with about 34.2% of the global plantations located in this region (90.2 9 106 ha in East Asia and 264.1 9 106 ha in the globe in 2010) (FAO, 2010a). This abundance of plantations could provide insight into the effect of forest management on the dynamics of forest C stocks and C sinks or sources (Fang et al., 2010). Using a process-based model of the terrestrial C cycle, Ito (2008) conducted the first regional estimation of the C budget of terrestrial ecosystems in East Asia and estimated that 73.1 Pg C was stored in vegetation and soil with an average C sequestration rate of 98.8 Tg C yr1 (1 Tg = 1012 g) in East Asian forests during 2000–2005. Since the mid-1990s, using national forest inventory data, several studies have estimated forest biomass C stocks in China (e.g., Fang et al., 1998, 2001, 2007; Guo et al., 2010), Japan (Fang et al., 2005) and South Korea (Choi et al., 2002; Li et al., 2010), and these studies

Fig. 1 Locations of the five East Asian countries in this study. © 2014 John Wiley & Sons Ltd, Global Change Biology, 20, 2019–2030

F O R E S T B I O M A S S C S I N K S I N E A S T A S I A 2021 conclude that forests in all these countries have functioned as C sinks since the late 1970s. However, a long-term study that uses consistent methods and wellqualified data sets to evaluate changes in biomass C stocks and the size of C sinks or sources for East Asian forests is lacking. This study uses statistically sound inventory data from the 1970s to the 2000s to estimate forest biomass C stocks and C sinks or sources for the entire East Asia region using the biomass expansion factor (BEF) method and forest inventory data (for China, Japan, and South Korea) and FAO (Food and Agriculture Organization of the United Nations) data (for North Korea and Mongolia). Furthermore, we discuss the relative contributions of forest areal expansion and increased biomass C density to explore the possible mechanisms for forest C dynamics in this region over the last four decades.

Materials and methods

Forest inventory data In recent decades, China, Japan, and South Korea have periodically conducted national-level forest resource inventories. These inventories provide information on forest area and timber volume for each forest type for each administrative unit (e.g., Fang et al., 2005, 2007; Li et al., 2010). China’s forest inventory data used in this study included seven periods: 1973–1976, 1977–1981, 1984–1988, 1989–1993, 1994–1998, 1999–2003, and 2004–2008 (Chinese Ministry of Forestry, 1977, 1983, 1989, 1994, 2000, 2005, 2010). In the inventories, China’s forests were classified into three categories: stands (including natural and planted forests), economic forests (woods with the primary objective of production of fruits, edible oils, drinks, flavorings, industrial raw materials, and medicinal materials), and bamboo forests. In this study, ‘forest’ only refers to ‘forest stands’ with canopy coverage ≥ 20% and therefore excludes economic and bamboo forests (Fang et al., 2007). For each forest stand, the inventories documented detailed information on forest type, age class, area, and volume at the provincial level. Japan’s forest inventory data included seven periods: 1966– 1975, 1976–1980, 1981–1985, 1986–1990, 1991–1995, 1996–2000, and 2001–2005, and were compiled from Japan’s Forest Resources Statistics for 1975, 1980, 1985, 1990, 1995, 2000, and 2005 (Japan Agency of Forestry, 1978, 1982, 1987, 1992, 2000, 2003, 2008). In the inventories, forest is defined as land with ≥20% canopy coverage for government-owned lands and

>30% canopy coverage for community- and privately owned lands (Fang et al., 2005; Wang et al., 2011). South Korea’s forest inventory data included five periods: 1972–1974, 1975–1982, 1983–1992, 1993–2000, and 2001–2007, and were compiled from the Agriculture and Forestry Statistical Yearbooks for 1974, 1982, 1992, 2000, and 2007 (Korean Ministry of Agriculture & Forestry, 1975, 1983, 1993, 2001, 2008). In the inventories, forest was defined as land with ≥30% canopy coverage for government-, community-, and privately owned lands (Li et al., 2010). Due to the lack of forest inventory data in North Korea and Mongolia, FAO statistics were used to estimate total forest area and timber volume for these two countries in 1990, 2000, 2005, and 2010 (FAO, 2006, 2010a). According to the FAO report on Global Forest Resource Assessment 2010, forest was defined as land spanning more than 0.5 hectares with trees taller than 5 meters and a canopy cover >10%, or trees able to reach these thresholds in situ. The thresholds for tree height and the areal extent for a forest are similar in the FAO criteria to that of inventories in China, Japan, and South Korea, but that for canopy cover is smaller than the other inventories and may result in relatively higher estimates of forest area for North Korea and Mongolia. We obtained the information on the forest area and timber volume for North Korea in 1970 from Lee (2006), and for Mongolia in 1974 from Persson (1974) and the Mongolian Ministry of Environment, Nature & Tourism (2009). Forest area and timber volume in 1980 for these two countries were estimated by assuming linear relationships of each variable between the 1970s and 1990. Table 1 summarizes the periods of inventory data or FAO reports that were used to estimate the forest biomass C stocks for the five countries from the 1970s to the 2000s.

Estimation of forest biomass C stocks Because forest inventories only report forest area and timber volume, but do not provide detailed information on forest biomass, it is necessary to develop allometric relationships between forest biomass and forest timber volume for each forest type (Fang et al., 1998, 2001; Brown & Schroeder, 1999). The BEF, which is defined as the ratio of stand biomass to timber volume (Mg m3) (1 Mg = 106 g), is used to convert timber volume from forest inventory to biomass (e.g., Fang et al., 2001, 2005; Guo et al., 2010; Li et al., 2010). Previous studies have suggested that BEF is not constant, but varies with forest age, site class, stand density, and site quality (e.g., Brown & Lugo, 1992; Schroeder et al., 1997; Fang et al., 1998, 2001, 2005; Brown & Schroeder, 1999; Brown et al., 1999). Fang et al. (1996, 1998, 2001, 2005), Fang and Wang (2001) derived a simple reciprocal equation from direct field

Table 1 Periods of inventory data or FAO reports used for estimating biomass C stocks for the five countries Decade

China

Japan

North Korea

South Korea

Mongolia

1970s 1980s 1990s 2000s

1973–1976; 1977–1981 1984–1988 1989–1993; 1994–1998 1999–2003; 2004–2008

1966–1975; 1976–1980 1981–1985; 1986–1990 1991–1995; 1996–2000 2001–2005

1970; 1980 1980; 1990 1990; 2000 2000; 2005; 2010

1972–1974; 1975–1982 1983–1992 1993–2000 2001–2007

1974; 1980 1980; 1990 1990; 2000 2000; 2005; 2010

© 2014 John Wiley & Sons Ltd, Global Change Biology, 20, 2019–2030

2022 J . F A N G et al. measurements to express the BEF-timber volume relationship by forest type in China and Japan: BEF ¼ a þ b=x

ð1Þ

where x is the timber volume per unit area (m3 ha1), and a and b are constants for each specific forest type. In Eqn (1), when timber volume is very large (such as mature forest), BEF will approach to a constant parameter, a, while it will be extremely large when timber volume is quite small (such as young forest). This simple mathematic relationship fits for almost all forest types. With this simple BEF approach, one can easily calculate regional or national forest biomass based on direct field measurements and forest inventory data. For detailed information about the BEF method, see Fang et al. (1998, 2001, 2007) and Guo et al. (2010) for China, Fang et al. (2005) for Japan, and Li et al. (2010) for South Korea. Parameters of the BEF equations for major forest types in China, Japan, and South Korea are listed in Table S2. In this study, we used the BEF method with parameters for each forest type from Guo et al. (2010) and Fang et al. (2005) to calculate forest biomass in China and Japan, respectively, from the 1970s to the 2000s. It should be mentioned that the forest inventory data from 1973 to 1976 for China included only total forest area and timber volume at the provincial level. Fang & Chen (2001) established an empirical, linear relationship between mean biomass and mean volume at the provincial level using China’s forest inventory data between 1977 and 1998. We used recent forest inventory data and an updated robust linear relationship to calculate provincial forest biomass in China from 1973 to 1976: BD ¼ 0:704VD þ 19:953ðR2 ¼ 0:968; n ¼ 211Þ

ð2Þ

where BD and VD are biomass density (Mg ha1) and volume density (m3 ha1), respectively. Before 1994, forest was defined as land with >30% canopy coverage in China (Fang et al., 2001). The 1994–1998 inventory data provided both criteria (20% and 30% canopy coverage), and Fang et al. (2007) found that there existed robust linear relationships for the forest area and biomass C stock between the two criteria at the provincial level. In this study, we modified their equations with power function relationships to obtain more accurate conversions: 2 AREA0:2 ¼ 1:290AREA0:995 0:3 ðR ¼ 0:996; n ¼ 30Þ

2 CARBON0:2 ¼ 1:147CARBON0:996 0:3 ðR ¼ 0:998; n ¼ 30Þ

ð3Þ

ð4Þ

where AREA and CARBON are forest area (104 ha) and biomass C stock (Tg C) in a province, respectively; subscripts 0.3 and 0.2 stand for the criterion of >30% and ≥20% canopy coverage, respectively. As a result, the provincial forest areas and biomass C stocks with the new criterion in China in 1973–1976, 1977–1981, 1984–1988, and 1989–1993 were calculated based on Eqns (3) and (4), and corresponding forest C densities were hence obtained. For South Korea, we adopted the results of forest biomass from the 1970s to the 2000s estimated by Li et al. (2010), who

used the same BEF method as Fang et al. (1998, 2001) for three major forest types (coniferous, deciduous, and mixed forests) in South Korea. According to the recent report on Global Forest Resources Assessment 2010 from FAO, major forest types in North Korea were oak (Quercus), pine (Pinus), and larch (Larix) (FAO, 2010b), and major forest types in Mongolia were Siberian larch (Larix sibirica), Siberian pine (Pinus sibirica), Scots pine (Pinus sylvestris), and Betula (Betula platyphylla) (FAO, 2010c). There were no direct field measurement data for these two countries and therefore we could not establish BEF functions, but the forest types in these two countries are very similar to those in the northeast and north China. Therefore, similar to Eqn (2), we used the data in the northeast and north China and established empirical relationships between provincial mean biomass and mean volume for North Korea [Eqn (5)] and Mongolia [Eqn (6)], to estimate forest biomass for these two countries from the 1970s to the 2000s: BD ¼ 0:969VD þ 12:800ðR2 ¼ 0:953; n ¼ 178Þ

ð5Þ

BD ¼ 0:898VD þ 19:311ðR2 ¼ 0:965; n ¼ 122Þ

ð6Þ

where BD and VD are the same as Eqn (2). We used a factor of 0.5 to convert biomass to C stock in this study (Fang et al., 2001, 2005).

Data of NDVI and forested area Remotely sensed normalized difference vegetation index (NDVI) data are not only an indicator of land cover and vegetation growth (e.g., leaf area index and net primary production) but also used as a surrogate of growing conditions for vegetation (i.e., the physical environment for plant growth, such as soil conditions, moisture, temperature, and light availability) (Potter et al., 1993; Field et al., 1995; Fang et al., 2003). Because we do not have direct measures of growing conditions at a national scale, we use NDVI and its trend over time to compare the growing conditions of forested areas among East Asian countries. In general, a large NDVI value and a positive NDVI trend imply good site quality and favorable growing conditions for vegetation. The NDVI data used in this study come from the third version of the AVHRR NDVI archive, provided by the Global Inventory Monitoring and Modeling Studies group at a spatial resolution of 8 9 8 km2 over 15 day intervals for the period of 1982 to 2011 (Beck et al., 2011). This data set is one of the most accurate products with which to assess the change in vegetation growth over time and is used widely to depict long-term change in global and regional vegetation cover (Fensholt & Proud, 2012). To eliminate spurious NDVI trends caused by winter snow, our analysis focused on the interannual changes in the forested areas during the growing season (May to September) for East Asia. In addition, we used the global land cover data set with a resolution of 8 9 8 km2, generated by University of Maryland (http://glcf.umiacs.umd.edu/data/landcover/; Hansen et al., 1998, 2000), to document forest area for each East Asian country, except Mongolia because of data unavailability. © 2014 John Wiley & Sons Ltd, Global Change Biology, 20, 2019–2030

F O R E S T B I O M A S S C S I N K S I N E A S T A S I A 2023

Calculation of change rates of forest area, C density, and C stock We used the concept of Forest Identity, proposed by Kauppi et al. (2006) and Waggoner (2008), to explore the relative contribution of changes in forest area and biomass C density (biomass C stock per area) to the C sink for China, Japan, and South Korea. According to the Forest Identity concept, forest area (A, ha), biomass C density (D, Mg C ha1), and total biomass C stock (M, Tg C) can be linked using Eqn (7), and their change rates (a, d, and m) over time (t) can be linked with Eqns (8) and (9). By calculating these attributes, we can assess the relative contribution of changes in forest area and biomass C density to the change in total biomass C stock (i.e., C sink): M¼AD

ð7Þ

Because ln (M) = ln (A) + ln (D), the change rates (m, a, and d) of M, A and D should be: 1 dM 1 dA 1 dD d lnðMÞ d lnðAÞ d lnðDÞ ¼ þ ; or ¼ þ M dt A dt D dt dt dt dt

ð8Þ

Let m

d lnðMÞ d lnðAÞ d lnðDÞ ;a  ;d  dt dt dt

Then, m ¼ a þ d

(9)

where M, A, and D represent total biomass C stock (Tg C, or Pg C), forest area (ha), and biomass C density (Mg C ha1) at the national level, respectively; and m, a, and d depict the

corresponding derivatives (or change rate) of these attributes over time. This identity combines the values of forest area with the biomass density into the changing biomass C stock (i.e., C sink).

Results

Changes in forest area In East Asia, forest area increased by 9.4% from 179.78 9 106 ha in the 1970s to 196.65 9 106 ha in the 2000s, with most of this increase in China (Table 2). The forest area in China increased by 21.88 9 106 ha, from an initial area of 127.31 9 106 ha to 149.19 9 106 ha by the 2000s, which accounted for 129.7% of the total area increment in the East Asian region. Forest area in North Korea and Mongolia decreased by 3.08 9 106 and 1.86 9 106 ha over the study period, respectively, representing 32.8 and 14.1% of the respective country’s forest area in the 1970s. There was a slight decrease of 0.18 9 106 ha in Japan and a slight increase of 0.11 9 106 ha in South Korea (Table 2).

Changes in forest biomass C density, total C stock, and C sink The biomass C density in the region had dramatically increased from 38.6 Mg C ha1 in the 1970s to 45.5 Mg

Table 2 Forest area, C stocks, and C sinks for each country in East Asia from the 1970s to the 2000s Period

East Asia

China

Japan

North Korea

South Korea

Mongolia

179.78 183.14 185.62 196.65 16.87

127.31 131.69 136.06 149.19 21.88

23.82 23.79 23.60 23.64 0.18

9.38 8.59 7.57 6.30 3.08

6.10 6.29 6.26 6.21 0.11

13.17 12.78 12.13 11.31 1.86

6937 7301 7989 8944 2007

4719 4885 5395 6145 1426

954 1173 1395 1615 661

339 311 274 232 107

51 106 157 240 189

874 826 768 712 162

38.6 39.9 43.0 45.5 6.9

37.1 37.1 39.7 41.2 4.1

40.1 49.3 59.1 68.3 28.3

36.1 36.2 36.2 36.8 0.7

8.4 16.9 25.1 38.6 30.3

66.4 64.6 63.3 63.0 3.4

36.4 68.8 95.5 66.9

16.6 51.0 75.0 47.5

21.9 22.2 22.0 22.0

2.8 3.7 4.2 3.6

5.5 5.1 8.3 6.3

4.8 5.8 5.6 5.4

6

Area (10 ha) 1970s 1980s 1990s 2000s Net change C stock (Tg C) 1970s 1980s 1990s 2000s Net change C density (Mg C ha1) 1970s 1980s 1990s 2000s Net change C sink (Tg C yr1) 1970s–1980s 1980s–1990s 1990s–2000s 1970s–2000s

© 2014 John Wiley & Sons Ltd, Global Change Biology, 20, 2019–2030

2024 J . F A N G et al. C ha1 in the 2000s (Table 2). Among the five countries, the largest increase in biomass C density (30.3 Mg C ha1) was in South Korea, with an initial biomass C density of 8.4 Mg C ha1 and a biomass C density of 38.6 Mg C ha1 by the 2000s. Biomass C density increased in Japan and China from 40.1 and 37.1 Mg C ha1 in the 1970s to 68.3 and 41.2 Mg C ha1 in the 2000s, with a net increase of 28.3 and 4.1 Mg C ha1, respectively. It did not change much (ranging from 36.1 to 36.8 Mg C ha1 over the four decades) in North Korea and had a slight decrease of 3.4 Mg C ha1 in Mongolia (66.4 Mg C ha1 in the beginning to 63.0 Mg C ha1 in the 2000s). Total forest biomass C stocks in the region increased by 28.9%, from 6937 Tg C in the 1970s to 8944 Tg C in the 2000s, resulting in a net accumulation of 2007 Tg C. The increase was attributed to increases in China, Japan, and South Korea: over the four decades, the C stocks increased by 1426, 661, and 189 Tg C in China, Japan, and South Korea, respectively, from initial stocks of 4719, 954, and 51 Tg C to the stocks of 6145, 1615, and 240 Tg C by the 2000s, which accounted for 71.1, 32.9, and 9.4% of the total accumulation. On the other hand, the C stocks in North Korea and Mongolia decreased from initial stocks of 339 and 874 Tg C to 232 and 712 Tg C by the 2000s, with an accumulated C loss of 107 and 162 Tg, respectively. Across the entire East Asian region, biomass C sinks increased from 36.4 Tg C yr1 during the 1970s–1980s to 95.5 Tg C yr1 during the 1990s–2000s, at an average rate of 66.9 Tg C yr1 over the four decades. Not surprisingly, among the five countries, the largest C sink occurred in China. Over the study period, China’s forests contributed 47.5 Tg C yr1 (71.1%) to the total region’s C sink and increased from an initial value of 16.6 Tg C yr1 to 75.0 Tg C yr1 during the 1990s– 2000s. Forests in Japan and South Korea also functioned as C sinks, with an average C gain of 22.0 Tg yr1 in Japan and 6.3 Tg yr1 in South Korea, accounting for 32.9% and 9.4% of the total region’s C sink, respectively. North Korea and Mongolia showed C losses of 3.6 and 5.4 Tg yr1 averaged over the last 40 years, respectively.

Relative contributions of forest area and biomass density to C sinks To quantitate the relative contribution of areal expansion and increased regrowth (biomass density) to the total C sink of forests in the three countries (China, Japan, and South Korea) with C sinks, we calculated the relative change rates of forest area (a) and biomass C density (d) for all, planted, and natural forests in China and Japan, and for all forests in South Korea in

each period and over the four decades, using the concept of Forest Identity (Kauppi et al., 2006; Waggoner, 2008) (Fig. 2; Table S3). For all forests, the mean change rates of forest area and biomass density were 0.264% yr1 and 0.175% yr1 in China, respectively, with a larger contribution of the former than that of the latter (60.0% vs. 40.0%) to the C sink over the last 40 years (Fig. 2a). These rates and relative contributions were very much different from those in Japan and South Korea. In those two countries, forest area either decreased slightly (for Japan, with a = 0.013% yr1) or did not change much (for South Korea, with a = 0.030% yr1), and thus their contribution of areal expansion to the C sink was very small or negative. However, the biomass density of (a)

(b)

Fig. 2 Mean change rates of forest area and biomass density and their relative contributions to changes in total biomass C stock (i.e., C sink) for China, Japan, and South Korea over the 40 years. (a) For all forests in the three countries; and (b) for planted and natural forests in China and Japan. Numbers above bars are relative contributions (%) of forest area and biomass density to the total C sink over the four decades. The minus value for the change rate of forest area in Japan shows that forest area has shrunk and made its negative contribution to the C sink. The change rates were only calculated for all forests in South Korea because data were not available for planted and natural forests or a small area of natural forests in the country. For details, see text and Table S3. © 2014 John Wiley & Sons Ltd, Global Change Biology, 20, 2019–2030

F O R E S T B I O M A S S C S I N K S I N E A S T A S I A 2025 these countries increased remarkably, with respective average rates of 0.869 and 2.148% yr1 that contributed almost all of their C sinks (101.5% for Japan and 98.6% for South Korea) (Fig. 2a). The minus value for the change rate of forest area in Japan reveals that the area of forests has shrunk by a rate of 0.013% yr1 and thus exerted a negative influence (1.5%) on C gain, acting as a C source over the four decades. For planted forests (Fig. 2b), areal expansion made a larger contribution to the C sink than did the change in biomass density (58.1% vs. 41.9%) in China, with a change rate of 1.249 and 0.902% yr1, respectively. Compared with those in China, Japan’s forests showed much different patterns: increased biomass density dominated the C sink with a contribution of 91.1% (d = 1.235% yr1) and the areal expansion only contributed 8.9% (a = 0.120% yr1) of the C sink. In contrast to planted forests, increased biomass density of natural forests (Fig. 2b) in China was a greater contributor to the C sink than was areal expansion (87.0% vs. 13.0%), with d and a of 0.217 and 0.032% yr1, respectively. In Japan, increased biomass density was responsible for all the C sink (123.9%), while the area of natural forests has shrunk by 6.4% over the last 40 years (a = 0.110% yr1) (also, see Table 3). We calculated the change rates of forest area and biomass density and their relative contributions to the C sink for each period for the three countries (Table S3). In general, the change rate (m, or C sink) of the total C stock tended to increase in China, but decrease in Japan and South Korea, suggesting that the C sink strength increased in China’s forests, but declined in the other two countries. For example, for all forests in China, the m value increased from 0.173% yr1 during the 1970s– 1980s to 0.650% yr1 during the 1990s–2000s, whereas

it decreased from 1.030% yr1 to 0.731% yr1 in Japan and from 3.521% yr1 to 2.089% yr1 in South Korea, respectively (Table S3). However, this trend was not very evident in the different periods for planted and natural forests because of their contrasting patterns of changes in forest area and biomass density.

Discussion

Forest C sinks in East Asia and the relative contributions of forest area and biomass density Over the last four decades, East Asia’s forests have functioned as a persistent C sink, with a peak of 95.5 Tg C yr1 in the 2000s (Table 2). This sink was attributed to increased C stocks in China, Japan, and South Korea; however, the mechanisms underlying the C sinks in these three countries were quite different. In China, areal expansion and forest regrowth (i.e., increase in biomass density) were the major contributors to this C sequestration and have made a respective contribution of 60.0% and 40.0% to the C sink for all forests (Fig. 2a; Table S3). Among a total of 1426 Tg C sequestration within China’s forests, planted and natural forests have almost equal contributions (703 and 723 Tg C, respectively) (Tables 2 and 3), but were driven by different mechanisms (Fig. 2b). For planted forests, areal expansion and biomass density made a respective contribution of 58.1% and 41.9% to the C sink (Fig. 2b). During the study period, 90% of the forest area increment was from planted forests, mainly because of the implementation of national afforestation and reforestation projects since the 1970s (Fang et al., 2001; Lei, 2005). As a result, the area of planted forests dramatically increased from 16.44 9 106 ha to 36.15 9 106 ha in the 2000s, and its proportion to the total forest area

Table 3 Forest area, C stocks, and C densities for planted and natural forests in China and Japan from the 1970s to the 2000s Planted forests

Period China 1970s 1980s 1990s 2000s Net change Japan 1970s 1980s 1990s 2000s Net change

Natural forests

Area (106 ha)

C stock (Tg C)

C density (Mg C ha1)

Area (106 ha)

C stock (Tg C)

C density (Mg C ha1)

16.44 23.47 27.95 36.15 19.71

249 418 584 952 703

15.1 17.8 20.9 26.3 11.2

110.87 108.22 108.11 113.04 2.17

4470 4467 4811 5193 723

40.3 41.3 44.5 45.9 5.6

9.61 10.23 10.34 10.33 0.72

346 497 668 810 464

36.0 48.6 64.6 78.4 42.4

14.22 13.56 13.26 13.31 0.91

608 676 727 805 197

42.8 49.9 54.8 60.5 17.7

© 2014 John Wiley & Sons Ltd, Global Change Biology, 20, 2019–2030

2026 J . F A N G et al. increased from 12.9% to 24.2% in the 2000s (Table 3). Meanwhile, the regrowth of these young forests also made a significant contribution to the C gain: the biomass C density of planted forests increased by 11.2 Mg C ha1, from an initial density of 15.1 Mg C ha1 to 26.3 Mg C ha1 in the 2000s. For natural forests, however, forest area had a slight increase (2% increment), but the biomass C density increased remarkably, with a net gain of 5.6 Mg C ha1 (13.9%). Therefore, regrowth of existing forests was found to be the dominant factor related to C sequestration (87.0% vs. 13.0% for biomass density vs. area) (Fig. 2b). Japan’s forest is the second largest C sink with a total C sink of 661 Tg C over the four decades. However, forest area was reduced by 0.18 9 106 ha, indicating that increased biomass density contributed the entire C sink for all forests in Japan (Fig. 2a; Tables 2 and S3). Within Japan’s forests, over 70% of the total biomass C accumulation (464 Tg C) was derived from planted forests (Table 3). During the study period, biomass C density of planted forests increased by 42.4 Mg C ha1 and the area of planted forests slightly increased by 0.72 9 106 ha (Table 3). Therefore, the regrowth of planted forests dominated this large C sink with a contribution of 91.1% (Fig. 2b). Despite a decrease (0.91 9 106 ha) in forest area over the four decades, total biomass accumulation of 197 Tg C was found in natural forests. Therefore, the C sink was due entirely to the regrowth of natural forests, with a net increase of 17.7 Mg C ha1 of biomass C density during the study period (Fig. 2b; Table 3). In South Korea, total forest biomass C stocks increased by 189 Tg C over the four decades mainly due to the regrowth of existing forests. There was a net increase in biomass C density of 30.3 Mg C ha1, or 3.6 times, from an initial density of 8.4 Mg C ha1 to a density of 38.6 Mg C ha1 in the 2000s. As a result, increased biomass density was largely responsible for this C sink (98.6%) and areal expansion made a small contribution (1.4%) due to a slight increase in total forest area over the four decades (0.11 9 106 ha or 2% increment) (Fig. 2a; Tables 2 and S3). In contrast, forests in both North Korea and Mongolia acted as C sources because of deforestation, but the mechanisms causing the C loss were different in these two countries. In North Korea, land-use change (converting forests to croplands) and harvest of firewood were major causes of the decrease in forest area (Lee, 2006), while in Mongolia, the combined effects of timber cutting, forest fires, pests, and diseases had resulted in the decrease in forest area and biomass C density (United Nations Environmental Programme and Mongolian Ministry of Nature and Environment, 2002).

Factors affecting C sink strength among the countries As shown in Tables 2 and 3, China’s forest has been the largest C sink (1426 Tg C) over the last four decades, followed by that of Japan (661 Tg C) and South Korea (189 Tg C). Compared with China’s much larger forest area, however, forests in Japan and South Korea showed greater C sink strength (C sink per area), with 0.35, 0.93, and 1.01 Mg C ha1 yr1 in China, Japan, and South Korea, respectively. This may be attributed to two major factors: forest age structure and growth conditions. We discuss these factors below, focusing on the forests of China and Japan because information was not available for South Korea. First of all, the age structures of forests (especially for planted forests) are much different in the two countries. In Japan, large-scale plantation and restoration of natural forests had been conducted since the late 1950s and early 1960s (after World War II) (Japan Agency of Forestry, 2000), while afforestation campaigns have been occurring in China since the end of 1970s (Fang et al., 2001; Xiao, 2005). As a result, Japan’s forests are about 20 years older than China’s forests, suggesting that Japan’s forests are mostly in the middle- to prematureaged growing stage in which trees grow fast, while forests in China are young- to mid-aged stands. Although we do not have detailed information on forest age for each study period for both countries, we have some data to support this. In China, national inventories report timber volume for each age class for some forest types and forests were divided into five age classes: young, middle, premature, mature, and overmature-aged classes (Xiao, 2005). In this classification, the age span changes with forest types, but at the national level, this could be generalized as: young- (60 years) (for practically, we combined mature and overmature-aged classes into a single mature-aged class for comparison of the two countries). In China’s planted forests in 2000, the area of young-, mid-, premature-, and mature-aged forests was 40.2%, 37.2%, 13.7%, and 8.9% of the total forest area, respectively (China State Administration of Forestry, 2005). In comparison, the respective numbers were 12.8%, 35.0%, 43.7%, and 8.6% (Japan Agency of Forestry, 2003), suggesting that a majority of forest stands were in the mid- and premature-aged classes in Japan. The younger age structure of China’s forests likely contributed to lower biomass C densities and smaller C sink strength when compared with Japan’s forests. Table 3 shows this clearly. The area in planted forests in China increased by 19.71 9 106 ha over the study period, from 16.44 9 106 ha in the 1970s to 36.15 9 106 ha in the 2000s; likewise, the area of natural © 2014 John Wiley & Sons Ltd, Global Change Biology, 20, 2019–2030

F O R E S T B I O M A S S C S I N K S I N E A S T A S I A 2027 forests increased by 2.17 9 106 ha. In comparison, the area of Japan’s plantations fluctuated between 10.23 9 106 ha and 10.34 9 106 ha during the 1980s– 2000s, and the area of natural forests showed a slight decrease (Table 3). With such changes in forest age structure, China’s forests have much lower biomass C density than that of Japan’s forests. For example, biomass C density of planted forests in China was only 15.1 Mg C ha1 in 1970s and 26.3 Mg C ha1 in the 2000s, while that in Japan was 36.0 and 78.4 Mg C ha1, respectively. Growing conditions (or site quality) differ considerably among East Asian countries. Japan has a typical oceanic climate, with abundant precipitation and warm temperature that favor the relatively fast growth for forests as compared with China and other countries (Kira, 1991; Fang et al., 2005, 2010). Consequently, biomass density of both planted and natural forests in Japan generally increases faster than that in the other countries, even at the same forest ages. Because we do not have direct measurements of growing conditions and because remotely sensed NDVI data can be a surrogate to indicate integrative growing conditions (Potter et al., 1993; Field et al., 1995; Fang et al., 2003), we used the NDVI time series data set to compare integrative growing conditions for forests among the different countries. Specifically, we explored possible causes of the large biomass C density and high productivity of Japan’s forests relative to the other countries. In general, a large NDVI value and a positive NDVI trend over time imply good site quality and favorable growing conditions for forests. Figure 3 shows interannual variation in growing season NDVI (May to September) of forested areas in China, Japan, South Korea, and North Korea (data for Mongolia were not available). Table 4 lists statistics for the NDVI attributes. Japan had the largest 30 years averaged NDVI (0.672 units), followed by South Korea (0.638), China (0.606), and North Korea (0.583) (Table 4). The NDVI trends also showed a similar order as did the averaged NDVI (but Japan and South Korea showed a similar value of the trend, or 9.59 9 104 vs. 9.61 9 104). Together, these suggest that Japan has the best growing conditions for forests, followed by South Korea, China, and North Korea. This order is the same as that of biomass C density and C sink (Table 2), suggesting that NDVI and its change over time can not only indicate the growing conditions for forests but also act as good measures to show biomass C density and vegetation production (or C sink strength). Note that the NDVI in North Korea did not exhibit an increasing trend (R2 = 0.000, P = 0.938) and coincidently did not show an increase in biomass C density (ranging from 36.1 to 36.8 Mg C ha1) over the last 40 years. © 2014 John Wiley & Sons Ltd, Global Change Biology, 20, 2019–2030

Fig. 3 Interannual variation in growing season NDVI of forested area for four countries (China, Japan, North Korea, and South Korea) in East Asia from 1982 to 2011. Other than North Korea, all countries show a significant NDVI increase. Mongolia was not included because data were not available.

Table 4 Statistics for NDVI attributes over 30 years between 1982 and 2011 for forested areas in East Asian countries

Country

Mean  SD

China Japan North Korea South Korea

0.606 0.672 0.583 0.638

   

0.124 0.124 0.065 0.069

Trend (9104)

R2

P value

7.49 9.59 0.26 9.61

0.163 0.164 0.000 0.173

0.027 0.027 0.938 0.022

Implications of forest C sinks in East Asia Although we estimated the changes in forest biomass C accumulation in East Asia over the last four decades, we could not estimate the total C budget for this region because we lacked the estimates for the four other important C pools of the forest system: dead wood, litter, soil organic C, and harvested wood products (Pan et al., 2011). Here, we use the ratios of different C pools in US forests to make an approximate account for the C budget of the whole forest sector in East Asia. Similar to forest of the United States, those of East Asia have experienced a long history of forest clearing, agricultural expansion, and subsequent abandonment; currently, forests are still recovering from such activities (Perlin, 1991; Brown et al., 1997; Fang et al., 2005; Li et al., 2010; Pan et al., 2011). The ratio of net change among different C pools of the forest sector in the United States was 0.49: 0.11: 0.01: 0.02: 0.37 for living vegetation: dead wood: litter: soil organic C: harvested forest products (Woodbury et al., 2007). By applying these ratios to our data, we obtained an average C sink rate of 136.5 Tg C yr1 for the whole forest sector in East Asia over the four decades.

2028 J . F A N G et al. Table 5 The percentage of the national fossil-fuel CO2 emissions offset by East Asia’s living forests during the 1970s– 2000s

Region

C sink (Tg C yr1)

Fossil-fuel emission (Tg C yr1)*

Offset percent (%)

China Japan North Korea South Korea Mongolia East Asia

47.5 22.0 3.6 6.3 5.4 66.9

749 292 39 71 2 1153

6.3 7.5 9.2 8.9 270.0 5.8

*The mean fossil-fuel emission during 1970s–2000s.

We used national fossil-fuel CO2 emissions data provided by the Oak Ridge National Laboratory of the United States Department of Energy (Boden et al., 2012) to estimate average CO2 emissions from fossil fuels in each country and the entire region over the last four decades (Table 5). The C sink rate averaged 47.5, 22.0, and 6.3 Tg C yr1 in living forests of China, Japan, and South Korea over the last four decades, respectively, which offsets 6.3, 7.5, and 8.9% of total fossil-fuel CO2 emissions of the respective countries over the study period (Table 5). However, over the same period, living biomass of forests in North Korea and Mongolia released 3.6 and 5.4 Tg C yr1, equaling to 9.2% and 270.0% of total fossil-fuel CO2 emissions of these two countries, respectively. Overall, the net C sink (66.9 Tg C yr1) in living biomass in East Asian forests has offset 5.8% of the total fossil-fuel CO2 emissions in this region over the last 40 years. If the four other C pools mentioned above were also included in the account, then the forest sector in this region could offset 11.8% of its contemporary fossil-fuel CO2 emissions.

Uncertainty of estimations The most important uncertainty may come from the quality of forest area and timber volume data from the forest inventories and FAO. For China, Japan, and South Korea, forest inventory data used in our study specified the precision requirement in the sampling design: in China, the forest area and timber volume precision were required to be >90% in almost each province (>85% in Beijing, Shanghai, and Tianjin) (Xiao, 2005); in Japan, the error of timber volume in the inventories at the national level was

Forest biomass carbon sinks in East Asia, with special reference to the relative contributions of forest expansion and forest growth.

Forests play an important role in regional and global carbon (C) cycles. With extensive afforestation and reforestation efforts over the last several ...
254KB Sizes 0 Downloads 0 Views