Carbon sinks and sources of China

1
Carbon sinks and sources of Chinas forests from
1901 to 2001

• Shaoqiang Wang1, Jingming Chen2, Weiming Ju2,
  Jiyuan Liu1, Xianfeng Feng1, Mingzhen Chen2,
  Panqin Chen1, Guirui Yu1
• 1.Institute of Geographical Sciences and Natural
  Resources Research, Beijing, Chinese Academy of
  Sciences
• 2. Department of Geography, University of
  Toronto, Canada

2
Contents

• INTEC model
• Materials and method
• Results
• Comparison
• Discussion

3

• The InTEC C model is a process-based
  biogeochemical model designed to simulate annual
  C and N fluxes and pool sizes in forested
  ecosystems.
• It is driven by spatial datasets, including
  climate, soil texture, vegetation parameters (LAI
  and land cover type), forest stand age, and N
  deposition datasets.
• It progressively simulates historical annual NPP
  for each pixel in terms of the initial value of
  NPP in the starting simulation year,
  age-dependent productivity of forest, and the
  combined effects on NPP of climate, CO2
  concentration, and N deposition.

4

• The interannual variability of NPP is separated
  into contributions from changes in CO2
  concentration, growing season temperature, N
  content of foliage, growing season length, and
  LAI
• The model assumes that all disturbances cause
  complete stand mortality and that the disturbed
  forest regenerates without cover type change in
  the second year after a disturbance.
• Like other terrestrial biogeochemical models,
  InTEC uses a fixed land cover map in the whole
  simulation period.
• The incorporation of land use change and
  vegetation dynamics are tasks for further model
  improvement.

5

• For the application of this model to Chinas
  forests, its soil carbon module was calibrated
  using measured SOC density and the decomposition
  rates of different soil C pools measured at
  several forested sites in China.

6
INTEC model

• Integrated Terrestrial Ecosystem C-budget model
  (InTEC)

Management
Disturbance
Spatial and temporal up-scaling algorithm Combine
the upscaling results with age-NPP relationship
Farquhars leaf photosynthesis model
Century C cycle model
N mineralization model
Age-NPP relationship
7
Fig. 1. The major components of the Integrated
Terrestrial Ecosystem Carbon Cycle (InTEC) model
applied to each 1 km pixel. The historical
variation in NPP is central for estimating the
amounts of dead organic matter in nine pools
(Table 1) (Cite from Chen et al., 2003)
8
(Black Spruce in Ontario)
Site index
Role of Forest Age
(Chen et al. 2002)
9
Reference paper

• Chen J., Chen, W., Liu J., and J. Cihlar. 2000.
  Annual carbon balance of Canadas forests during
  1895-1996. Global Biogeochemical Cycles 14(3)
  839-849.
• Chen, W., J. Chen, J. Liu, and J. Cihlar. 2000.
  Approaches for reducing uncertainties in regional
  forest carbon balance. Global Biogeochemical
  Cycles 14(3) 827-838.
• Chen W., J. Chen, and J. Cihlar, 2000. An
  integrated terrestrial ecosystem carbon-budget
  model based on changes in disturbance, climate,
  and atmospheric chemistry. Ecological Modelling
  135 55-79.
• Chen W., Chen M., Price D.T., Cihlar J.. 2002.
  Effects of stand age on net primary productivity
  of boreal black spruce forests in Ontario,
  Canada. Can. J. For. Res. 32 833842 .
• Chen, J., W. Ju, J. Cihlar, et al. 2003. Spatial
  distribution of carbon sources and sinks in
  Canadas forests. Tellus 55B 622641.

10

• Simulation period 101 yr
• Spatial resolution 1 km
• Time resolution 1 yr
• Biotic and abiotic factors
• Soil C pool SOC(0-30cm) litter
• Vege C pool above-vegetation roots
• model integrates
• remote sensing images, gridded climate, soil and
  forest inventory data

11
Materials and method

• Forest stand

• Forth national forest survey 1989 1993
• Provincial statistic data for forest species
• Published materials
• Forest wet N deposition

12
Mean Forest Age Distribution in 2001
13
Percentage histogram of forest age in China in
2001
14

• Soil profiles

• Second soil survey 5405 profiles
• Texture clay, silt and sand (0-30 cm)
• Organic matter and total N (0-30 cm)

15
The spatial distribution of soil profiles in
China (2473) Red color new adding profiles(2932)
16
(No Transcript)
17

• Leaf Area Index (LAI)
• China-wide LAI map in 2001 were provided by
  Dr.Xianfeng Feng. The LAI map was produced from
  250 meter resolution cloud-free composite images
  of MODIS sensor.
• Land cover
• The land cover map was produced from 560 30-m TM
  images during 1999 and 2000 for whole China.
  Forest area 229.01?104km2
• Conifer, deciduous and mixed 167.26 ?104km2 .
• Shrubland area 61.75 ?104km2 .

18
TM Data for Renewing Our Land-use/land-cover
Database and LUCC Analysis
TM Imaging data of 19992000
19
10km10km forest in LU2000 (Units hectare)
20

• NPP and Evapotranspiration (ET)
• Annual average NPP and ET maps in 2001 were
  produced at 1 km resolution using BEPS
  model.(Feng)
• Climate
• Monthly data of mean air temperature and
  precipitation for China in the period 1901-2001
  were obtained from the 0.5º global data set
  interpolated by the UK Climate Research Unit.
  Dr.Weiming Ju calculated the annual mean air
  temperature, precipitation, growing season length
  and mean growing season temperature.

21
Database

• Climate data
• Annual Mean T and P
• Annual growing season T (May to September)
• Annual growing season length
• Total 21.4 Gb
• Reference data
• Total 399 Mb
• Results files
• 5.7 Gb for China.

22
Run InTEC model
23
(No Transcript)
24
(No Transcript)
25
(No Transcript)
26
(No Transcript)
27
(No Transcript)
28
(No Transcript)
29
(No Transcript)
30
Results

• Carbon storage in Chinas forest ecosystems
• Changes in Carbon pools in Chinas forest
  ecosystems
• Spatial patterns of carbon pools and changes

31
NPP of Chinas forests from 1901 to 2001 (Units
g Cm-2)
32
Carbon storages (from 1901 to 2001)

• Ecosystem 28.8-34.6 Pg C (1Pg1015g)
• Vegetation 8.5-13.5 Pg C (above-ground)
• Root 3.0-4.2 Pg C
• SOC 13.3-14.3 Pg C (0-30cm)
• Litter 2.5-3.3 Pg C
• NPP 0.93-1.22 PgC/yr

33
C pools of Chinas forests from 1901 to 2001
(Units Pg yr-1)
34
Compare with inventory research

• Forest ecosystem 12.27 Pg C (Wang 1999)
• Vegetation 5.41 Pg C (volume)
• SOC 6.04 Pg C (0-20cm)
• Litter 0.82 Pg C ( 129 ?104km2 )
• SOC 44.8?7.8 Pg C (0-30cm for whole China)
• Second national soil survey (Wang et al., 2004)
• Simulated forest (conifer, deciduous and mixed,
  shrubland) C pool

• Ecosystem 28.8-34.6 Pg C
  Vegetation 9.5-13.5 Pg C
• SOC 13.3-14.3 Pg C (0-30cm) Litter 2.5-3.3
  Pg C Root 3.0-4.2

Forested area excluding shrubland 167.26 ?104km2
Sixth inventory, Forested area in 2003 175
?104km2
35
Table 1. Biomass C storage in Chinas et al.,
regions forest
Author area (106 ha) Biomass (Pg C) Roots (Pg C) Period Region
This paper 167.26 7.310.12 2.780.30 1990s China
Pan et al., (2004) 130.5 4.34 1989-1993 China
Ni (2003) 112 9.11 9.11 1990s China
Goodale et al., (2002) 195.0 4.6 1990s China
Fang et al., (1998) 102.2 4.3 1984-1988 China
Fang et al., (2001) 133.7 4.75 1998 China
Alexey et al., (1995) 771.1 25.60 1990s Russia
Liski and Kauppi (2000) 244.6 11.89 1990s Canada
Birdsey and Heath (1995) 245.9 13.78 1990s U.S.A
36
Table 2. C storage in above-ground biomass in
Chinas forest (excluding shrubland)
Pan et al. (2004) Pan et al. (2004) Pan et al. (2004) InTEC model InTEC model InTEC model InTEC model
Areas C Pool C density Areas C Pool C Pool C density
(106ha) (PgC) (kg m-2) (106ha) (PgC) (kg m-2) (kg m-2)
19731976 123.6 3.836 3.10 167.26 7.310.30 4.370.18 4.370.18
19771981 116.0 3.91 3.37 167.26 6.590.25 3.940.15 3.940.15
19841988 125.4 3.998 3.19 167.26 6.310.07 3.770.04 3.770.04
19891993 130.5 4.339 3.32 167.26 6.740.25 4.030.15 4.030.15
37
Changes in forest C pools
38
Change in C pools of Chinas forests from 1901 to
2001 (Units Tg)
39
NEP of Chinas forests from 1901 to 2001 (Units
g Cm-2)
40

• During 1901-1949, Chinas forests were a source
  of 21.07.8 Tg C yr-1 due to disturbances (human
  activities).
• Its size increased to 122.325.3 Tg C yr-1 during
  1950-1987 due to intensified human activities in
  late 1950s, early 1960s, 1970s and early 1980s.
• The forests became large sinks of 176.744.8 Tg C
  yr-1 during 1988-2001,
• From 1901 to 2001, Chinas forests were a small
  carbon source by 3.32 Pg C, about 32.922.3 Tg C
  yr-1.

41
Table 2. Cumulative stock changes of Chinas
forest above-ground biomass, soil organic carbon,
litter and roots C pools for periods during
1901-1949, 1950-1987 and 1987-2001
1901-1949 (PgC) 1950-1987 (PgC) 1988-2001 (PgC) Total
SOC -0.10 0.29 -0.80 -0.61
Above-ground -0.70 -4.22 2.48 -2.44
Litter -0.10 -0.27 -0.19 -0.56
Roots -0.12 -0.57 0.98 0.29
Total -1.02 -4.77 2.47 -3.32
42
NEP spatial distribution of Chinas forests in
1901 (Units g C m-2)
43
NEP spatial distribution of Chinas forests in
1911 (Units g C m-2)
44
NEP spatial distribution of Chinas forests in
1921 (Units g C m-2)
45
NEP spatial distribution of Chinas forests in
1931 (Units g C m-2)
46
NEP spatial distribution of Chinas forests in
1941 (Units g C m-2)
47
NEP spatial distribution of Chinas forests in
1951 (Units g C m-2)
48
NEP spatial distribution of Chinas forests in
1961 (Units g C m-2)
49
NEP spatial distribution of Chinas forests in
1971 (Units g C m-2)
50
NEP spatial distribution of Chinas forests in
1981 (Units g C m-2)
51
NEP spatial distribution of Chinas forests in
1991 (Units g C m-2)
52
NEP spatial distribution of Chinas forests in
2001 (Units g C m-2)
53
NEP spatial distribution of Chinas forests from
1992 to 2001 (Units g C m-2)
54
NEP spatial distribution of Chinas forests from
1997 to 2001 (Units g C m-2)
55
Comparison

• NEP
• -21.07.8 Tg C yr-1, during 1901-1949
• -122.325.3 Tg C yr-1, during 1950-1987
• 176.744.8 Tg C yr-1, during 1988-2001
  
• Total loss is about -3.32 Pg C from 1901 to 2001
• NEP is around 0.21 Pg C/yr in 1990s.
• (Including shrubland)

56

• 0.04 Pg C /yr in early 1990s (Wang 1999
  biomass)
• Average net forest vegetation sequestration is
  about 0.021 Pg C/yr from 1979 to 1998 (Fang et
  al., 2001 biomass).
• 1984-1988 about 0.011 Pg C
• 1989-1993 about 0.035 Pg C
• 1994-1998 about 0.026 Pg C.
• 0.04 PgC/yr in 1990s (Goodale et al., 2002)
• 0.07 PgC/yr for 1981-1998 in whole China (Cao et
  al., 2003)
• 0.068 PgC/yr from late 1980s to 1990s (Pan et
  al., 2004)
• Without considering SOC sequestration and litter
  C pool.

57
Discussion

• Three main sources of uncertainty.
• lack of forest inventory data. relationship of
  NPP and age
• applicability of INTEC model. Because relevant
  researches were few in China, some coefficient
  and parameters were not validated according to
  actual conditions of Chinas forests.
• classification of forest types. There are various
  kinds of forest types from cold temperate to
  tropical zones in China.

58

• From 1950 to 1987, forests are carbon source in
  China. However, from 1988, Chinas forest is a
  carbon sink
• INTEC model can present rough trends of
  spatiotemporal patterns of C sinks and sources in
  Chinas forest now.
• Uncertainty
• Data
• Modification, validation and verification

59

• The simulation could underestimate or
  overestimate changes in carbon pools.
• Lack of fire, harvest and insect disturbances
  data in Chinas forests,
• We did not separate the effects of disturbance
  factors on C sinks and sources in Chinas forests
  
• We also did not attempt to partition within these
  natural and anthropogenic influences on C balance
  of China-wide forests in this study.
• Simulations results are just the first step of
  our study and need more work on it.

60
Future carbon balance of Chinas forests under
climate change and increasing CO2

• Weimin Ju1, Jing M.Chen1, Danny Harvey1, Shaoqing
  Wang2
• 1.Department of Geography, University of Toronto,
  Ontario, Canada
• 2. Institute of Geographic Sciences and Natural
  Resources Research, Beijing, China

61

• This study predicts the C balance in Chinas
  forests during the 21th century under climate
  change as simulated by the second version of the
  coupled atmosphere-ocean general circulation
  model (CGCM2) from the Canadian Climate Center,
  using radiative forcing from the SRES A2 B2
  emission scenarios.

62

• The previous versions of the InTEC model do not
  explicitly include the effect on NPP of the
  change in canopy conductivity caused by the
  change in soil moisture availability.
• Canopy conductivity determines intercellular CO2
  concentration and therefore has considerable
  influence on photosynthesis.

63

• Previously, InTEC assumes that the ratio of NPP
  to GPP is constant as climate and N status change
  
• To remove this assumption, the effect of
  temperature on autotrophic respiration is added
  in this study.
• Interannual variability of GPP is first
  calculated using a modified algorithm
  (A.18-A.41).
• The historical values of NPP are progressively
  computed from the interannual variations of GPP
  and autotrophic respiration, and the ratio of
  respiration to GPP

64

• This study uses a control and two projected
  future climates (A2 and B2) from the second
  version of the first generation of the Canadian
  coupled general circulation model (CGCM 2)
• In the A2 run,
• CO2 increases from 350 in 1990 to 850 ppmv in
  2100.
• MAT in Chinas forested areas increases by 5 K
  during 2001-2100,
• Precipitation increases by only about 10

65

• In the B2 run,
• CO2 increases from 350 in 1990 to 600 ppmv in
  2100.
• The warming in this scenario is moderate, with an
  increase in MAT of about 3.0 K during 2001 to
  2100.
• This scenario also projects a smaller increase in
  the total precipitation than does CGCM2-A2

66
Figure 1. Time series of mean annual temperature
(MAT), annual precipitation, annual potential
evopotranspiration (PET) used in simulations.
Values prior to 2002 are interpolated and
calculated from historical observations.
Projected climates from control, A2 and B2 runs
of CGCM2 are used for the period from 2002 to
2100.
67
Figure 2a. Spatial distribution of changes in
mean annual temperature (MAT). Values are the
difference of MAT during 2091 to 2100 compared
with that during 2001 to 2010 (a) CGCM2-A2 (b)
CGCM2-B2.
68
Figure 2b. The change in annual total
precipitation during 2091 to 2100 relative to
that during 2001 to 2010 (a) CGCM2-A2 (b)
CGCM2-B2.
69
Simulations performed

• In order to investigate the response of the C
  cycle in Chinas forests to climate change and
  increasing CO2, we conducted seven model
  simulations (Table1)

70
Table 1 Description of six simulations performed
Run NO. Run name Future Climate Future CO2 (ppmv)
1 Baseline CGCM2-control 371 during 2001-2100
2 CGCM2-A2 coupled CGCM2-A2 From 371 to 850 during 2001-2100
3 CGCM2-A2 climate CGCM2-A2 371 during 2001-2100
4 CGCM2-A2 CO2 fertilization CGCM2-control From 371 to 850 during 2001-2100
5 CGCM2-B2 coupled CGCM2-B2 From 371 to 600 during 2001-2100
6 CGCM2-B2 climate CGCM2-B2 371 during 2001-2100
7 CGCM2-B2 CO2 fertilization CGCM2-control From 371 to 600 during 2001-2100
71
Response of net primary productivity to climate
change and increasing CO2

• NPP of Chinas forests increases from 951.4 Tg C
  yr-1 in 1901 to 1131.0 Tg C yr-1 in 2001
  according to our simulation. due to the effect of
  climate change and forest regrowth acting
  together
• Under the contemporary CO2 concentration and the
  CGCM2 control climate, the mean NPP increases to
  1252.5 Tg C yr-1 during 2091-2100 due to the
  effects of N deposition and changes in stand age
  structure.
• The increase of CO2 concentration will enhance
  the NPP of Chinas forests while pure climate
  warming has a considerable negative effect on
  NPP.

72
Table2. Simulated mean NPP of Chinas forests
under different climate and CO2 change scenarios
(Tg C yr-1)
2001-2100 2001-2010 2045-2054 2091-2100
A2 Climate CO2 1432.4 1274.2 1472.1 1470.5
A2 Climate 1128.1 1233.9 1190.1 889.4
A2 CO2 1425.0 1240.1 1390.6 1602.1
B2 Climate CO2 1323.0 1255.2 1323.1 1341.3
B2 Climate 1148.2 1233.5 1141.7 1025.4
B2 CO2 1354.3 1219.2 1324.9 1494.2
Baseline 1221.3 1198.8 1197.5 1252.5
Values in this table are the NPP simulated using
different scenarios of climate and CO2
concentration changes in comparison with that
simulated in the baseline run using contemporary
CO2 concentration (371 ppmv) and climate from the
control run of CGCM2.
73

• CO2 concentration increases to 550 ppmv,
  simulated NPP will increase by 13 under the
  CGCM2-A2 increase scenario and by 16 under the
  CGCM2-B2 increase scenario,
• CO2 fertilization enhancement of NPP during
  2001-2100 is 28 and 19 for the CGCM2-A2 and
  CGCM2-B2 CO2 increase scenario, respectively.
• NPP will increase in most forests in China during
  the next 100 years due to increasing CO2 and the
  corresponding climate change

74

• The net C uptake by Chinas forests will peak
  around 2020 and then will gradually decline
  mainly due to the change in forest stand age
  structure, which leads to a gradual decrease in
  NPP (Figure 3) and a continuous increase in
  heterotrophic respiration.

75
Figure 3. Time series of NPP output from
different simulation experiments. All simulations
were driven by same spatial and historical
climate data sets and produced same NPP values
for the period from 1901 to 2001. The difference
in NPP during 2002-2100 is due to different
future climate and CO2 concentration data sets
used.
76
Figure 4. Time series of the mean of growing
season relative canopy conductivity to CO2 a,
which is related to soil moisture availability
and calculated using Eq.(2). A small a value
indicates that canopy conductivity to CO2 is
limited by soil moisture.
77
Figure 5. Departure of average NPP during
2091-2100 from values during 1991-2000 under
different climate change and increase in CO2
scenarios. Positive values represent NPP
increase, and vice versa.
78
Figure 6. Distribution of different climate
regions in China
79
Figure 7. Time series of simulated total
vegetation C storage in Chinas forests under
different climate change and increasing CO2
scenarios. All simulations were driven by the
same spatial and historical climate datasets and
produced the same vegetation carbon storage for
the period from 1901 to 2001. The difference in
vegetation C storage during 2002-2100 is due to
different future climate and CO2 concentration
datasets used.
80
Figure 8. Time series of simulated total soil C
(including litter and SOC) storage in Chinas
forests under different climate change and
increasing CO2 scenarios. All simulations were
driven by same spatial and historical climate
data sets and output same soil carbon storage
during1901 - 2001. The difference in soil C
storage during 2002-2100 is due to different
future climate and CO2 concentration data sets
used.
81
Figure 9. Time series of simulated total C stock
in Chinas forest ecosystems under different
climate change and increasing CO2 scenarios. The
difference in total carbon storage during
2002-2100 is due to different future climate and
CO2 concentration datasets used.
82
Figure 10. Spatial distribution of carbon balance
in Chinas forests (a) during 1991-2000 (b),
(C), and (d) during 2091-2100 from the CGCM2-A2
coupled, CO2 and climate impact runs,
respectively (e) during 2091-2100 from the
baseline run (f), (g), and (h) during 2091-2100
from the CGCM2-B2 coupled, CO2, and climate
impact runs, respectively. Negative values are
carbon sources, and vice versa.
83
Figure 11. Time series of simulated C balance in
Chinas forest ecosystems under different climate
change and increasing CO2 scenarios. After 2001,
different future climate and CO2 concentration
datasets were used to drive the model.
84
Table 3. Effects of different climate and CO2
change scenarios on mean NPP of Chinas forests
(Tg C yr-1)
2001-2100 2001-2010 2045-2054 2091-2100
A2 Climate CO2 211.1 75.5 274.6 218.0
A2 Climate -93.2 35.1 -7.4 -363.1
A2 CO2 203.6 41.3 199.1 349.6
B2 Climate CO2 101.7 56.4 125.6 88.8
B2 Climate -73.1 34.8 -55.8 -227.1
B2 CO2 133.0 20.4 126.6 241.7
Values in this table are differences between the
NPP simulated using different scenarios of
climate and CO2 concentration changes and that
simulated using contemporary CO2 concentration
(371 ppmv) and climate from the control run of
CGCM2.
85
Scenario Analysis

• Chinas forests were a C sink in the 1990s,
  averaging 189 Tg C yr -1 (about 13 of the global
  total).
• This sink peaks around 2020 and then gradually
  declines to 33.5 Tg C yr -1 during 2091-2100
  without climate and CO2 changes.
• Effects of pure climate change without allowing
  CO2 effects on C assimilation in plants might
  reduce the average NPP of Chinas forests by 29
  and 18 during 2091-2100, respectively.

86

• An increase in CO2 might broadly increase future
  C sequestration of Chinas forests.
• However, this CO2 fertilization effect might
  decline with time.
• The CO2 fertilization effects on NPP by the end
  of this century are 349.6 and 241.7 Tg C yr-1
  under CGCM2-A2 and B2 increase scenarios,
  respectively.
• Under a CO2 increase without climate change, the
  majority of Chinas forests would be C sinks
  during 2091-2100, ranging from 0 to 100 g C m-2
  yr-1.

87

• Stand age structure plays a more dominant role in
  determining future C sequestration than CO2 and
  climate change.
• The prediction of future C sequestration of
  Chinas forest is very sensitive to the Q10 value
  used to estimate maintenance respiration and to
  soil water availability and less sensitive to N
  deposition scenario.
• The results are not yet comprehensive, as no
  forest disturbance data were available or
  predicted after 2001.
• These results could be useful for assessing
  measures to mitigate climate change through
  reforestation.

88

• The combination of an increase in CO2 and climate
  change always has a larger positive effect on NPP
  and vegetation C stock than the increase in CO2
  alone during the first 50 years.
• After 2050, the negative effect of climate change
  will weaken the enhancement of NPP by the
  increase in CO2.
• The prediction of future C sequestration of
  Chinas is very sensitive to the value used to
  estimate the changing maintenance respiration,
  and to soil water availability, while being less
  sensitive to the N deposition scenario.

89
Thank you!