Negative Emissions. Energy and CO2 Levels

1
Negative Emissions. Energy and CO2 Levels
Aroon PARSHOTAM1, Ian ENTING,2 and Peter
READ3 1Mathematics, Massey University,
Palmerston North, New Zealand. 2ARC Centre of
Excellence for Mathematics and Statistics of
Complex Systems (MASCOS), The University of
Melbourne. 3Applied and International Economics,
Massey University, Palmerston North, New Zealand.
2
Result of Read and Lermit (2004)
Modelling using FLAMES (Fuel/Fibre/Food Land
Allocation Model for Energy/Environmental
Sustainability) coupled to a C balance shows
using BECS (Bio-energy with Carbon Storage), and
strong assumptions appropriate to imminent abrupt
climate change (ACC), that pre-industrial CO2
levels on maximal scale can be restored by
mid-century.
Read and Lermit, 2005. Bio-energy with carbon
storage (BECS) a sequential decision approach to
the threat of abrupt climate change, Energy xx
1-18.
3
Figure 1 Carbon in atmosphere profiles under
various reference scenarios and under robust
policy with and without an ACC precursor event in
2020
4
Limitation of the analysis the response of the
carbon cycle to net emissions is represented as
relaxation, with single time constant, to
pre-industrial levels
5
This limitation was noted and it was suggested,
on the basis of comparing a range of earlier
studies of emission reductions, that the
differences compared to a more rigorous 4-pool
model (Joos, et al. 1996) could be significant
(Parshotam and Read, 2005).
Parshotam and Read, 2005. CO2 levels under BECS
(Bio-Energy with Carbon Storage) with improved C
dynamics. Submitted to Mitigation and Adaptation
Strategies for Global Change, Special Issue
Abrupt Climate Change and Greenhouse Gas
Emissions contributions to the Expert Workshop,
Paris, 30.ix.04 1.x.04.
6
The 4 -pool Model (Joos et al., 1996)
These 4-pools are initialised to 0 under
pre-industrial CO2 concentrations. Cc is the
inherent rate of carbon uptake by the natural and
uncontrolled components of the global carbon
cycle. They consist of the oceans and the
terrestrial biosphere.
7
Figure 1 Various reference emissions scenarios
8
Figure 2 Comparison of results from the 1-pool
and 4-pool decay models, using the IS92a
emissions scenario
9
Figure 3 Comparison of results from the 1-pool
and 4-pool decay models, using the PH emissions
scenario
10
Figure 4 Comparison of results from the 1-pool
and 4-pool decay models, using the PN emissions
scenario
11

• Some conclusions
• For positive emissions scenarios, the simple,
  one-pool carbon model used by Read and Lermit
  (2004) does not lead to significantly different
  level profiles from the four-pool model.
• major differences occur if the same model is used
  with unchanged coefficient when emissions
  decrease.

12
However, this does not alter Read and Lermits
broad conclusions that Negative emissions
systems, involving increased photosynthesis and
management of the carbon content of the biomass
produced into long term terrestrial sinks,
provide far more effective control of CO2 in
atmosphere than simply cutting emissions from the
energy sector.
13
Here, we give a more comprehensive analysis of
some of the issues involved with negative
emissions scenarios.
14
The FLAMES Model

• Simulates the interaction of energy, timber, and
  land markets under the impact of use-selected
  (policy driven) land-use change.
• Plantations yield both bio-energy and timber as
  joint products, with the outputs treated as
  perfect substitutes, respectively for fossil fuel
  as energy raw material and for other timber
  supplies in the forest products industry.

15
FLAMES Formulation
1) Fuel Market (producers price p, consumers
price p ? P)
non biofuel supply biofuel from long
rotation biofuel from short rotation
demand, a function of post tax price, per capita
demand shifts, population growth, and the macro
impact of energy tax. Initial conditions are
2/GJ, 300EJ, neglecting traditional
non-commercial bioenergy.
16
2) Land Market (price, r (i.e. rent))
Commercial forestry land sequestration land
biofuel land reference case biofuel land
demand for farmland (a function of rent, per
capita demand, etc) wilderness fixed supply
of non-barren non-forest land. Land left to
wilderness inversely related to rent. Ls, Lb and
Lp all functions of t with Ls and Lb selected by
model user. Lp is depleted in proportion to
population in reference case, modified by
de-sequestration in policy cases. Initial
conditions 1.9bHa farmland, 0.7bHa commercial
forestry 10/100/Ha lo/hi rent cases.
17
3) Forest Product Market (price s)
long rotation wood product short rotation wood
product demand for wood products , initially
1.3bTonnes at 130/tonne
18
4) Dedicated Tax ?
Tax transfers the cost of land use change
policies onto the fossil fuel supplier
rent plus operating cost of biofuel land rent
and establishment cost of sequestration land
producer price on biofuel tax on all carbon
fuel. Initial value 0 with zero initial
policy-land
19
5) Carbon Balance
Carbon to atmosphere carbon fuel emissions -
long rotation absorption (with short rotation
absorption short rotation emissions) carbon
from short rotation land carbon from long
rotation land carbon from farmland (all
relative to carbon in wilderness land) -
absorption in ocean. Initial Cat 760Gt
20
A physical reason why a one-pool model does not
work The ocean surface layers are the quick
response pool.  If you want to get the
atmospheric level down you have to dispose of
what is out-gassed from the ocean as well as what
is in the atmosphere. 
21
Background studies

• Young et al. (1996).
• Enting et al. (1994).
• Joos et al. (1996).
• These results show that the difference in
  formulation of the carbon cycle response greatly
  affects the results under conditions of
  decreasing emissions, even when models are
  calibrated to fit the CO2 growth over the
  industrial period.

22
Linear response functions

• In the linear representation, the amount of CO2
  in the atmosphere is given by

where (E(t) is emissions at time t and R(t) is
the proportion of carbon remaining in atmosphere
at time t after emissions. Various
representations of R(t), usually as sums of
exponentials have been obtained by fits to
behaviour of carbon cycle models in response to
unit inputs.
23
Laplace analysis

• Enting (1990) related this linear response
  function to Laplace transforms for exponentially
  increasing emissions (E(t)).

24

• Changing variables, to tt-t, gives

If we consider exponentially growing emissions,
E(t)Aexp (at), and take the limit t0 ? -8, we
obtain
25

• Laplace transforms is a useful tool for
  conceptual understanding of relations
• These results could be generalised to a response
  represented as a sum of n exponentials.
• The Laplace Transform can also show how to
  transform the Read and Lermit calculations so
  that a correction could be made to their work.
• Furthermore, the Laplace analysis lends itself
  conviently to questions such as the following to
  be answered What linearly decreasing pattern of
  emissions scenarios is needed to get CO2 ppm in
  2050, 2060, etc?

26
Conclusions and recommendations

• There needs to be a re-analysis of BECS with
  improved carbon cycle model in FLAMES.
• The Laplace analysis could prove very useful in
  negative emissions scenario analysis.