Overestimated forest growth in China during 1994-2003
Zhen Yu
Division of Forestry and Natural Resources
West VirginiaUniversity
December 10, 2012
Abstract
In this paper, the distribution of forest age and stock volume were modeled and presented. The result showed that the forest age mainly ranged from 20 to 90 years with average age of 42.59 years, while the total forest stock volume was about 99.79×108 m3(average of 76.40 m3/ha) during the Fifth National Forest Inventory (1994-1998). The forest volume growth from the Fifth to the Sixth National Inventory for different forest age showed a sequence of young forest > mid-age forest > near-mature forest > mature forest > over mature forest, accounted for 37.91 %, 32.14 %, 11.07 %, 10.17 %, and 8.7 % of the total stock volume growth, respectively. However, the modifications of forest criteria had greatly contributed to increase of forest area and stock volume in statistics during the Fifth and Sixth National Forest Inventory. The result showed that the criteria changes of survey methods have contributed to the increase of forest stock by 2.76 ×108 m3(accounting for 2.76 % of the total forest stock volume) and 6.09 ×108 m3(accounting for 5.33 % of the total forest stock volume) during the Fifth and Sixth National Forest Inventory, respectively.
Therefore, with limited potential of afforestation and high potential of forest growth in young forest in China, a more sensible approach of increasing carbon sequestration should be emphasizing the future work on forest culture instead of forest coverage.
Keywords
Forest age, Forest stock volume
1Introduction
Large-scale afforestationprojects have been widely implemented since the late 1980s throughout China as a key approachto improving ecological environments while increasing terrestrial carbon sink to mitigate the climate change(State Forestry Administration, 2009).Thesix key afforestation projects include:Three-North Shelter Forest Project, Natural Forest Protection Program, Grain for Green Project, Beijing and Tianjin Sandstorm Control Project and Fast-growth and High-yieldForest Project (Liu et al., 2008; Zhang et al., 1999). By 2003, the preserved forest area has reached 53.26 million hectares in China, accounting for approximately 1/3 of the world's total planted forest area (Department of Forest Resources Management, State Forestry Administration of China, 2005).China's goal is to increase forest cover to 26% by 2050 by means of the largest planted forest program in the world (Wang et al., 2007).
However, the credits of treeplanting in the increase of forest coverin the past twenty years maybe exaggerated. For example, the threshold of tree cover was lowered from 0.3 of Fourth National Forest Inventory (1984-1993) to 0.2 of the Fifth National Forest Inventory (1994-1998) (Department of Forest Resources Management, State Forestry Administration of China, 2000). During the Sixth National Forest Inventory (1999-2003), criteria of forest changed again by including specified shrub forest as a forest type (Department of Forest Resources Management, State Forestry Administration of China, 2005). These criteria modifications greatly increased forest area, hence resulted in a higher forest stock volume.
Therefore, comprehensive assessment of the plant construction programsimplementedin the past is required to provide important experiences, lessons, and guides for afforestation ofChina and many other parts of the world.Based on nationalinventory materials and remote sensing data, we appliedstatistical method and model analysis to examine the plant constructionactivities during the past two decades. This studyaims to(1) model the forest stock volume growth during the period of 1989 to 2004, and (2)quantify the contributions of modifications in survey methods to the statisitical results of forest growth.
2Materialsand methods
2.1 Forest volume modeling.
During the third and Fourth National Forest Inventory (1989-1993), forest was defined as land with tree cover above 0.3. However, the threshold was lowered to 0.2 in the Fifth National Forest Inventory. The criteria change greatly increased forest area, hence resulted in a higher forest stock volume. During the Sixth National Forest Inventory (1999-2003), criteria of forest changed again by including specified shrub forest as a forest type. Shrubs with coverage above 0.3 and distributed above timberline, in arid area with precipitation below 400mm/yr, or in vulnerable areas such as karst region of tropical and subtropical region and dry-hot valley, were classified as specified shrub forest. These modifications of forest criteria had greatly contributed to increase of forest area and stock volume in statistics during the Fifth and Sixth National Forest Inventory. Therefore, we modeled the forest stock volumes of the period of 1994-1998 to determine the natural growth of forest stock.The process is showed in the Figure1.
Fig.1 Flowchart showing the algorithm used to calculate the forest volume
First, thedigitalized forest map was adjusted using forest inventory data. The errors of manual digitalization and landcover change were reduced by buffering the areas of specific forest types according to the forest inventory data released by the State Forestry Administration (Department of Forest Resources Management, State Forestry Administration of China, 2000; Hou, 2001.).
Second, the Fifth National Forest Inventory and NDVI data were used to calculate forest age at provincial scale and then downscale to 8×8 km2. Provincial scale forest age was calculated based on the Fifth National Forest Inventory data and forest age group classification system (Tab.1). According to Technique Provision of the National Forest Inventory announced by the State Forestry Administration, tree species are divided into five age groups with different ranges as show in Table 1.
Tab.1Age group of different dominant species
Species / Region / Type / Age groupYoung / Mid-age / Near-mature / Mature / Over-mature
Pinus koraiensis, spruce, cypress,
yew, hemlock / north
north
south
south / natural
planted
natural
planted / <60
<40
<40
<20 / 61 - 100
41 - 60
41 - 60
21 - 40 / 101-120
61 - 80
61 - 80
4l - 60 / 121-160
8l - 120
8l - 120
6l - 80 / 16l
12l
>121
8l
Larch, Fir, Scotch pine,
Japanese red pine, Black pine / north
north
south
south / natural
planted
natural
planted / <40
<20
<40
<20 / 41 - 80
21 - 30
41 - 60
21 - 30 / 81 - 100
31 - 40
61 - 80
3l - 40 / 101-140
4l - 60
8l - 120
4l - 60 / 14l
6l
>121
6l
Chinese pine, Masson pine,
Yunnan pine, Pinus kesiya,
Armand pine, Alpine pine / north
north
south
south / natural
planted
natural
planted / <30
<20
<20
<10 / 31 - 50
21 - 30
2l - 30
11 - 20 / 5l - 60
31 - 40
31 - 40
2l - 30 / 61 - 80
41 - 60
4l - 60
31 - 50 / >81
>61
>61
>51
Poplar, Willow, Eucalyptus, Sassafras, Paulownia, Casuarina, Azedarach, Pterocarya, Acacia / north
south / Planted
planted / <10
<5 / 11 - 15
6 - 10 / 16 - 20
11 - 15 / 21 - 30
16 - 25 / >31
>26
Birch, Elm, Schima superba, Maple, Dove tree / north
north
south
south / natural
planted
natural
planted / <30
<20
<20
<10 / 31 - 50
2l - 30
2l - 40
11 - 20 / 51 - 60
3l - 40
4l - 50
2l - 30 / 6l - 80
41 - 60
51 - 70
31 - 50 / >81
>61
>71
>51
Oak, Tussah, Castanopsis Carlesii，Castanopsis fargesii, Camphor tree, Linden / all
all / natural
planted / <40
<20 / 4l - 60
21 - 40 / 61 - 80
41 - 50 / 81 - 120
51 - 70 / 121
>71
Chinese fir, Cryptomeria, Metasequoia / south / planted / <10 / 11 - 20 / 2l - 25 / 26 - 35 / >36
The ceilingages of age groups in a province were calculated by the following equation:
(2)
Where: is the ceiling age of a forest age group in a province; is the percentage of natural forest of the province; is the ceiling age of natural vegetation type i; is the percentage of planted forest of the province; is the ceiling age of planted vegetation type i.
For young age group, mean age was the average of 1 plus ceiling age of young group (youngest trees are defined as 1 year old). For mid-age, near-mature and mature forest group, mean ages were the average of ceiling age of the forest group and the former forest group plus 1. For over-mature forest group, mean age was the ceiling age of mature forest plus 20 (define the mean ages of over-mature forest were 20 years older than mature forest).
Thirdly,provincial forest age was downscaled to 8×8 km2 resolution using NDVI data by equation 3(Dai et al., 2011).The change in the physiology and structure of plant canopies can be directly viewed from NDVI (Wang et al., 2004), therefore, NDVI is a useful indicator of forest growth, such as tree-ring width, height increment, and diameter increase (Wang et al., 2004;Lopatinet al., 2006; Di et al., 1994). Consequently, for each forest types, a five-year mean NDVI during 1994-1998 could be used as a reliable proxy for identification of forest age on a large regional scale.
(3)
Where: is the age of pixel i; is the mean age of the type and group which the pixel belongs to;, and are the NDVI of the pixel, maximum NDVI of the group, and minimum NDVI of the group; is the age range of the group.
Fourthly, regression models of each forest type were constructed using NPP, LAI, and forest age based on database compiled by Luo (Luo, 1996). The database contains 1266 samples gathered from field works, various literatures and the continuous forest inventory plots of the forest departments all over the country. The forest volume stock of 1994-1998 was modeled using the regression models of Table 2.
Table 2 Regression models of different forest types
ForestType / Regression model / Significant level / R2 / Sample numberBoreal/temperate Larix forest (BTLF) / / < 0.001 / 0.91 / 42
Boreal/alpine Picea-Abies forest (BAPF) / / < 0.001 / 0.43 / 162
Boreal Pinus sylvestris var. mongolica forest (BPSM) / / < 0.001 / 0.81 / 10
Temperate mixed coniferous-broadleaved forest (TMCB) / / < 0.001 / 0.61 / 17
Temperate typical deciduous broadleaved forest (TTDB) / / < 0.001 / 0.44 / 160
Temperate/subtropical montane Populus-Betula deciduous forest (TSMP) / / < 0.001 / 0.86 / 121
Temperate Pinus tabulaeformis forest (TPTF) / / < 0.001 / 0.83 / 150
Subtropical evergreen broadleaved forest (SEBF) / / < 0.001 / 0.73 / 228
Subtropical mixed evergreen-deciduous broadleaved forest (SMED) / / < 0.001 / 0.66 / 27
Subtropical Pinus massoniana forest (SPMF) / / < 0.001 / 0.88 / 53
Subtropical montane Cupressus and Sabina forest (SMCS) / / < 0.001 / 0.92 / 12
Subtropical montane Pinus armandii, P. taiwanensis and P. densada forest (SMPF) / / < 0.001 / 0.76 / 44
Subtropical Cunninghamia lanceolata forest (SCLF) / / < 0.001 / 0.85 / 74
Subtropical montane Pinus yunnanensis and P. khasya forest (SMPY) / / < 0.001 / 0.97 / 44
Sclerophyllous evergreen Quercus forest (SEQF) / / < 0.001 / 0.75 / 9
Finally, the forest stock volume of the period of 1994-1998 was estimated using the models specific to forest types. For the period of Fourth National Inventory to the Fifth National Inventory, the volume change was calculated by extracting total volume of area with coverage between 0.2 and 0.3 (Hansen et al., 2003). In contrast, criteria change’s contribution of including stock volumes of shrub forestduring the period of 1999 to 2003 was calculated by deducting total increase of stock volume from model simulated total stock volume. The total increase of stock volume is divided into two components: increase of natural growth of forest and new plantings during 1999-2003. Because of the ages of new plantings were below 5 years, the total stock volume of these youngforests was calculated by multiplying the area and the average stock volume of the young forest.
(4)
(5)
(6)
Where is the stand stock volume increase induced by the second criteria change; is the annual total increase of stock volume; is the modelled forest stock volume; is the total increase of stock volumefrom young forest; is the mean annual increase of stock volume; is the mean stock volume of young forest; is the increase of new young forest’s area from the period of 1994-1998 to the period of 1999-2003.
3. Results
3.1Forest stock volume simulation
Based on the Fifth National Forest Inventorymaterials, NDVI data, and digitalized vegetation map, forest age was estimated and downscaled to resolution of 8×8 km2 (Fig.2). The result showed that the forest age mainly ranged from 20 to 90 years with an average age of42.59 years. The highest and lowest forest ages werelocated in southwest and southeast of China, respectively. Forest age showed a spatial pattern of increase from southeast to northeast and southwest.Subtropical Pinus massoniana forest (SPMF, 19.75 years) and Subtropical Cunninghamia lanceolata forest (SCLF, 4.89 years) were two types of youngest forest which mainly distributed in southeast of China(Table 3). In contrast, Boreal/alpine Picea-Abies forest (BAPF, 79.24 years) and Temperate mixed coniferous-broadleaved forest (TMCB, 88.61 years) werethe oldest forests distributed in the southwest and northeast China(Table 3).
Figure 2Forest age distribution
Regression models of each forest type were constructed using database compiled by Luo (Luo, 1996). The result showed that total forest stock volume was 99.79×108 m3(average of 76.40 m3/ha) during the Fifth National Forest Inventory (1994-1998), comparedwith100.86×108 m3(average of 78.06 m3/ha) announced by the State Forestry Administration of China (Department of Forest Resources Management, State Forestry Administration of China, 2000).Forest stock volume also showed a distribution pattern of high in southwest and low in southeast.We found that the lowest forest stock was mainly located in northern China, while the highest was in the southeast of China(mainly insouth of Tibet).
The forest volume growth from the Fifth to the Sixth National Inventory for different forest age showed a sequence of young forest > mid-age forest > near-mature forest > mature forest > over mature forest, accounted for 37.91 %, 32.14 %, 11.07 %, 10.17 %, and 8.7 % of the total stock volume growth, respectively.
Figure 3Distribution pattern of forest stock volume during 1994 to 1998
Table 3 Simulated Forest during the Fifth National Inventory (1994-1998)
Forest type* / Average age(years) / Average stock volume
(m3/ha) / Total area
(×104 ha) / Total stock volume
(×108 m3)
BTLF / 54.53 / 95.43 / 1307.20 / 12.47
BAPF / 79.24 / 136.49 / 768.22 / 10.49
BPSM / 54.63 / 124.72 / 33.91 / 0.42
TMCB / 88.61 / 72.16 / 187.17 / 1.35
TTDB / 37.98 / 37.51 / 2835.70 / 10.64
TSMP / 32.62 / 32.18 / 833.26 / 2.68
TPTF / 37.62 / 55.61 / 363.61 / 2.02
SEBF / 63.83 / 154.27 / 1363.22 / 21.03
SMED / 44.81 / 142.90 / 546.92 / 7.82
SPMF / 19.75 / 69.01 / 1757.10 / 12.13
SMCS / 44.70 / 104.71 / 74.91 / 0.78
SMPF / 29.52 / 118.69 / 598.22 / 7.10
SCLF / 4.89 / 34.90 / 1385.55 / 4.84
SMPY / 57.09 / 49.32 / 856.66 / 4.22
SEQF / 68.61 / 120.50 / 149.39 / 1.80
Total / 42.59 / 76.40 / 13061.05 / 99.79
*Abbreviations are the same as Tab.6; Taiwan, Macao and Hongkong are not included.
3.2Forest stock volume increase caused by forestcriteria change
The threshold of forest cover was lowered to 0.2 from 0.3 in the Fifth National Forest Inventory. We estimated that lowering the criteria had contributed to increase of forest area by 62 471 km2, accounting for 24.75% of the total affoerstation area during the period of 1994 to 1998. For the period of 1998 to 2003, criteria change’s contribution of including stock volumes of shrub forest was calculated by deducting total increase of stock volume from modeled total stock volume. The result showed that total forest stock volume would increase 18.75×108m3 from 1998 to 2003 (including 4.34×108m3 of the new growth of young plantings) if no consumption occurred, compared to 24.84×108m3 announced by the State Forestry Administration of China (Department of Forest Resources Management, State Forestry Administration of China, 2005).
We estimated that the first and second criteria changes had contributed to the increase of forest stock by 2.76 ×108 m3(accounting for 2.56 % of the total forest stock volume) and 6.09 ×108 m3(accounting for 5.39 % of the total forest stock volume), respectively. After removing the increase of forest stock volume resulted from criteria change, the annual growth rate of forest stock volume would be 3.39 % and 3.32 % during the Fivth National Inventory and the Sixth National Inventory, respectively. Therefore, the annual growth rates were much lower than 3.91 % and 4.39 %, which were announced by the State Forestry Administration.
4. Discussion
During the third and Fourth National Forest Inventory (1989-1993), forest was defined as land with tree cover above 0.3. However, the threshold was lowered to 0.2 in the Fifth National Forest Inventory. During the Sixth National Forest Inventory (1999-2003), criteria of forest changed again by including specified shrub forest as a forest type. Shrubs with coverage above 0.3 and distributed above timberline, or in arid area with precipitation below 400mm/yr, or in vulnerable areas such as the karst region of tropical and subtropical regions and dry-hot valley, were classified as specified shrub forests.
The modifications of forest criteria had greatly contributed to the increase of forest area and stock volume in statistics during the Fifth and Sixth National Forest Inventory. We estimated that the first and second criteria changes had contributed to the increase of forest stock by 2.76 ×108 m3and 6.09 ×108 m3, respectively. The modeled annual growth rate during the Fifth and Sixth National Inventory were 3.39 % and 3.32 %, much smaller than3.91 % and4.39% announced by the government. For example, the total forest stock volume of Tibet province was 12.68 ×108 m3in the Fifth National Inventory. In contrast, the number had soared to 22.94×108 m3 in the Sixth National Inventory in Tibet. This sharp change of forest stock volume can not be explained by natural growth solely (annual growth rate of 12.59%).
Carbon sequestration ability will be greatly overestimated by including the increase of forest stock caused by survey methods’ changes. In addition, an overestimated forest stock will mislead forestry administration to make inappropriate decisions.For example, it was reported that the annual consumption shortage gap would be 1.0-1.5 ×108 m3 from 2007 to 2020. However, the shortage gap data was derived from overestimated forest stock growth rate, while the authentic shortage gap of forest stock will be enlarged in the next decade.
5. Conclusions
The current study assessed the forest stock volume change during the past two decades in China. The results showed that the credits of afforestation have been overestimated by including the increase of forest stock caused by survey methods’ changes during the period of 1994-1998.Therefore, to increase carbon sequestration, a more sensible approach is to emphasize the future work on forest culture instead of forest coverage.
Reference:
Cao, S., 2008. Why large-scale afforestation efforts in China have failed to solve thedesertification problem. Environmental Science & Technology 42, 1826–1831.
Cao,S.X., Chen, L., Yu, X.X., 2009. Impact of China's Grain for Green Project on the landscape of vulnerable arid and semi-arid agricultural regions: a case study in northern ShaanxiProvince, Journal of Applied Ecology, 46, 536-543.
Cao, S.X., Sun, G., Zhang, Z.Q., Chen, L.D., Feng, Q., Fu, B.J., McNulty, S., Shankman, D., Tang, J.W., Wang, Y.H., Wei, X.H., 2011a. Greening China Naturally, AMBIO: A Journal of the Human Environment.
Cao, S.X., Chen, L., Shankman, D., Wang, C.M., Wang, X.B., Zhang, H., 2011b. Excessive reliance on afforestation in China's arid and semi-arid regions: Lessons in ecological restoration. Earth-Science Reviews, 104, 240-245.