delusional redd baselines
Post on 21-May-2015
227 Views
Preview:
DESCRIPTION
TRANSCRIPT
Delusional REDD baselines
Britaldo Silveira Soares Filho
MEASUREMENT, REPORTING AND VERIFICATION IN LATIN AMERICAN REDD+ PROJECTS
CIFOR WORKSHOP – MARCH 8-9, 2012 – PETRÓPOLIS, BRAZIL
Britaldo S
ilveira S
oares
Filho
Main concepts
• Additionality: Proof that reduction in emissions from
REDD is additional to reductions that would occur if
initiative were not in place.
• Leakage: reduction in carbon emissions in one area
that results in increased emissions in another
• Permanence: long-term reduced emissions from
REDD. Depends on vulnerability to deforestation and/or degradation.
Concepts and methods borrowed from CDM projects
Britaldo S
ilveira S
oares
Filho
Measuring performance: the baseline approach
The CDM baseline
Historical approach: introducing an innovative process to an installed industrial plant.
Forward-looking approach (ex ante): the building of a new factory with a carbon neutral approach.
Depend strongly on the project and less on the context (Lesser risk of hot air).
Britaldo S
ilveira S
oares
Filho
REDD baselines
Forward-looking approach: Brazil overall GHG target of 39% reduction by 2020 (emissions from agriculture can increase from 480 to 627 MCO2)
Historical approach: Brazil´s National Climate Change Plan.
Reduction
6238
REDD6560
3805
18226
25396
27772
19014
14196
Hitorical
12911
7464
19600
Baseline15405
8935
0
5000
10000
15000
20000
25000
30000km2
Britaldo S
ilveira S
oares
Filho
What to do with countries or regions with low deforestation rates and large extent of forests
Acre Madre de Dios Pando
For more information: visit www.csr.ufmg.br/map
Britaldo S
ilveira S
oares
Filho
baseline
with project
-
2,000,000
4,000,000
6,000,000
8,000,000
10,000,000
12,000,000 C stocks
time
Forward-looking baseline
Project starts
Additionality = $$ credits
Britaldo S
ilveira S
oares
Filho
Project starts
Baseline is not solely based on the project itself, but on a reference region. Therefore, there is a need to understand the effect of the project on the deforestation regional trend.
Forward-looking baseline
Britaldo S
ilveira S
oares
Filho
1. understanding historical trends
2. identifying drivers of deforestation
3. modeling future scenarios
Solution: Deforestation model??
BASIC STEPS OF MODELING
A series of what if decisions !!! Brita
ldo Silveira
Soare
s Filho
• Historical fixed rates
• Historical trend
• Business-as-usual simulation (future scenarios)
Understanding historical trends Modeling approaches for BAU baselines
0
20000
40000
60000
80000
100000
120000
140000
160000
ha
time
average
projected
simulated (BAU)
ObservedWhich one ?
Britaldo S
ilveira S
oares
Filho
Modeling deforestation trajectories
Soares-Filho et al. 2010. PNAS.
2010 2007
Britaldo S
ilveira S
oares
Filho
Modeling scenarios
Soares-Filho et al. 2006. Nature
Nesptad, Soares-Filho et al. 2009. Science.
The End of Deforestation in the Brazilian Amazon
Trajectories can change drastically
Britaldo S
ilveira S
oares
Filho
Other aspects
• Delimiting the geographic reference region.
Reference region
Project
Britaldo S
ilveira S
oares
Filho
Other aspects
• Delimiting the geographic reference region.
Reference region
Project
road
Britaldo S
ilveira S
oares
Filho
Other aspects
• What if there are other projects???
Reference region
Project Project
Project
Madre de Dios, Peru, have 12 or more REDD projects
Britaldo S
ilveira S
oares
Filho
identifying drivers of deforestation deforestation up to 1994
deforestation up to 1999
deforestation from 1994 to 1999
(gray tone)
distance to main roads
buffer zone
overlay
cross-tabulation
deforested non-deforested total
----------------------------------------
distance > 10 km | 194187 1005296 | 1199483
distance < 10 km | 84228 199349 | 283577
----------------------------------------
total | 278415 1204645 | 1483060
They are in fact spatial determinants
Most models only incorporate spatial determinants!!!
Probability map of deforestation
Britaldo S
ilveira S
oares
Filho
modeling future scenarios • Spatially-explicit model of deforestation to quantify
where deforestation is likely to occur.
-09 50O `
-10 00O `
-10 10O `
-10 20O
`
-10 30O
`
-55 00O ` -54 50
O `
-09 50O `
-10 00O `
-10 10O `
-10 20O `
-10 30O
`
-55 00O ` -54 50
O `
-09 50O
`
-10 00O
`
-10 10O
`
-10 20O `
-10 30O
`
-55 00O
` -54 50O
`
Land cover map of 1986 Land cover map of 1994 Simulated land cover map of 1994
project project
deforestation is not random but occurs at locations that have a combination of bio-geophysical and economic attributes that are more attractive to the agents of deforestation
True or false?
Britaldo S
ilveira S
oares
Filho
Stochastic nature of deforestation
Performance x realism
Absolute frequency of transition probabilities of observed deforestation
scene 23167
0
500
1000
1500
2000
2500
3000
3500
0 50 100 150 200 250
probability value
num
be
r of
cells
Dinamica EGO simulates landscape structure
Half true, half false
Increasing spatial performance decreases realism
Britaldo S
ilveira S
oares
Filho
-09 50O `
-10 00O `
-10 10O `
-10 20O
`
-10 30O
`
-55 00O ` -54 50
O `
-09 50O `
-10 00O `
-10 10O `
-10 20O `
-10 30O
`
-55 00O ` -54 50
O `
-09 50O
`
-10 00O
`
-10 10O
`
-10 20O `
-10 30O
`
-55 00O
` -54 50O
`
Land cover map of 1986 Land cover map of 1994 Simulated land cover map of 1994
project project
Model validation
t1 t2 t3
training Validation only regarding location, but not spatial structure
Britaldo S
ilveira S
oares
Filho
Model performance assessment • Map comparison methods (how to lie with validation methods)
Label 1 2 3 4 5 6 7
map \ author
Costanza
(1989)
Power et al.
(2001)
Pontius
(2002)
Hagen
(2003)
Van Vliet
et al.
(2011)
Almeida et
al. (2008)1
Almeida et
al. (2008)2
degraded 5x5 0.98 0.83 0.65 0.70 0.66 0.83 1.00
degraded 9x9 0.97 0.81 0.48 0.47 0.48 0.62 0.75
degraded 15x15 0.96 0.81 0.34 0.24 0.34 0.46 0.55
degraded 31x31 0.95 0.80 0.18 -0.03 0.18 0.26 0.32
degraded 51x51 0.95 0.80 0.12 -0.13 0.12 0.19 0.23
random 0.90 0.84 0.002 1.00 1.00 0.04 0.05
line1 x line2 1.00 0.81 -0.02 0.59 -0.02 1.00 1.00
Soares-Filho et al. in review
And model comparisons are quite subjective (e.g. Vega et al. 2011)
Britaldo S
ilveira S
oares
Filho
Different models have different assumptions, geographic scales, spatial resolutions,
scopes, and, therefore, purposes.
Both simulations developed using Dinamica EGO, but none of them were designed to fix baselines
Thank you Reimer
Britaldo S
ilveira S
oares
Filho
This means trouble
Britaldo S
ilveira S
oares
Filho
What about leakage?
Leakage cannot be simulated, but can be evaluated. See Soares-Filho et al. PNAS. 2010
leakage In-out, out-out, diffuse (Fearnside, 2009)
Is modeling for dummies??? Model wizard for rapid
assessment
Effectiveness = Success rate – Leakage 50% = 70% - 20%
Representation of a fictitious software
Leakage Success Effectiveness
Enter value ------------ ------------
Britaldo S
ilveira S
oares
Filho
Good job &Business
• But!!!!
No ready solution for REDD (although vendors say so)
There is no magic solution, via software wizard…
MODEL must be built from the ground to incorporate local
knowledge…
Britaldo S
ilveira S
oares
Filho
In sum • Although various MRV methodologies for mapping carbon stocks and modeling
reductions as well as standards for international certification have been developed, the criteria for establishing baselines for crediting purposes are unsound; they do not take into account that forward-looking baselines are questionable because deforestation trajectories can alter drastically in response to changes in a complex set of circumstances (Nunes et al. 2012; Soares-Filho et al. IADB report, 2012).
• Furthermore, it is very difficult to isolate the local effect of a project from the overall trajectory of deforestation as well as to assess leakage arising from the establishment of projects, and none of those projects have attempted to perform such analyses (Soares-Filho et al. IADB report, 2012).
• As more specific standards and indicators are established, more flawed they become, because they are unfounded anyway (Personal Perception). And for financial crediting, every dollar counts. In addition, conflict of interest might lead to gamming (Fearnside, 2012).
Britaldo S
ilveira S
oares
Filho
So, how can we apply modeling to REDD?
Intervention
Information
Modeling Knowledge
Tool for policy design and planning
(let’s keep us in business)
Britaldo S
ilveira S
oares
Filho
Tool for REDD planning and management. Application to Acre
Dinamica EGO software
Simulation as a tool for regional planning www.csr.ufmg.br/map
Method adopted by SISA
Britaldo S
ilveira S
oares
Filho
Setting priority areas for REDD
• Index of threat (time dependent) (Soares-Filho et al. PNAS, 2010)
Priorização de áreas protegidas
Britaldo S
ilveira S
oares
Filho
Calculating costs of reduction
Amazon opportunity costs
Carbon Cost
0
5
10
15
20
25
30
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
year
U$
Carbon cost along the deforestation frontier
year (Nespdad, Soares-Filho et al. 2009; Soares-Filho et al. 2010)
Britaldo S
ilveira S
oares
Filho
Assessing effectiveness and leakage over time
Only monitoring deforestation is not enough.
Soares-Filho et al. PNAS . 2010.
Odds ratio adjusted
Land use zones 1997 2000 2005 Mean
Urban and periurban areas 6.15 2.17 5.40 4.57
Mining concessions 1.26 1.62 1.51 1.46
Small-scale cattle ranching 1.42 1.20 1.28 1.30
Small landowner lots 1.02 0.99 1.55 1.19
Brazil nut concessions 0.76 0.67 0.49 0.64
Logging concessions 2.84 3.28 4.29 3.47
Native communities 0.77 1.20 1.35 1.11
Reforestation concessions 0.19 0.39 0.93 0.51
Conservation concessions 3.85 2.78 2.06 2.89
Ecotourism concessions 0.08 0.08 0.76 0.31
Protected areas 4.85 8.27 5.55 6.22
Indigenous reserves 0.20 0.02 x 0.11
Unassigned land use 0.96 0.94 1.17 1.02
Nunes, Soares-Filho et al. Env. Con. 2012.
Britaldo S
ilveira S
oares
Filho
Which REDD projects in Madre de Dios can make a difference?
Hajek et al. 2011
Nunes, Soares-Filho et al. Env. Con. 2012.
Britaldo S
ilveira S
oares
Filho
Channeling REDD+ investments
Nunes, Soares-Filho et al. Economic benefits of forest conservation: assessing the potential rents from Brazil nut concessions in Madre de Dios, Peru, to channel REDD+ investments. Environmental Conservation, 2012.
REDD funds invested to upgrade the production chain of non-timber forest products
Britaldo S
ilveira S
oares
Filho
Thank you/Obrigado/Gracias
Modeling in support of sound policy Dinamica EGO freeware can be downloaded at www.csr.ufmg.br/dinamica
For many more examples, please access http://www.csr.ufmg.br/dinamica/publications/publications.php
britaldo@csr.ufmg.br
Britaldo Silveira Soares Filho (the culprit for developing deforestation models that are now being applied to REDD+)
Britaldo S
ilveira S
oares
Filho
top related