1
EP, 25 January 2012
Luisa Marelli
European Commission – DG Joint Research Centre (JRC)
Institute for Energy and Transport
Existing methodologies and best practices on assessing ILUC
2
Outlines
1. How is ILUC measured and how do models work?
2. GHG emissions from different models- uncertainties- impact on food consumption
3. US legislation
4. Other environmental effects
5. Conclusions
3
How to measure ILUC?
There is only one realitySo you cannot know what would have
happened without biofuels
Agro-economic models are used
ILUC cannot be measured directly…..
2020 “Policy” Scenario with extra biofuels
2020 “Baseline” scenario without extra biofuels
Compared with
Models do not compare differences between NOW and 2020
4
Crops for biofuels come from 3 sources
Increased crop demand due to biofuels
less consumption for food
higher crop yields
crop area expansion
crop price increases
emissions from land use change
Land emissions models
Economic models:
5
0%
25%
50%
75%
100%
LEITAPBiod
FAPRIBiod
AGLINKBiod
GTAPBiodmix
LEITAPWht Eth
AGLINKWht Eth
GTAPWht Eth
IMPACTWht Eth
IMPACTCG Eth
FAPRIWht Eth
Outside the EU
Within the EU
For EU biofuels most land use change is outside EUBiodiesel Ethanol
Where does ILUC occur?
Results from different models (JRC model comparison)
6
Commission’s reference results
ILUC GHG emissions differ across models but are significant in all cases (can negate GHG savings from biofuels)
7
JRC vs. IFPRI land emissions models
IFPRI 2011 economic results with IFPRI land emissions model:
IFPRI 2011 economic results with JRC land emissions model:
• Peat-drainage emissions account for about 50% of total EU biofuels ILUC emissions
similar total results:
8
ILUC depends on feedstock:
Oilseeds GHG > cereals > sugar crops
9
Uncertainty, but certainly above zero
US Corn ethanol [Plevin et al, 2010]
Frequencies distribution on ILUC emissions
IFPRI, 2011
10
Impact of preventing forest conversion
Even if we prevent forest conversion globally, GHG emissions are still significant!
No new cropland on
forest
11
Fixing food demand increases LUCLa
nd a
rea
land
requ
irem
ent t
o m
eet i
ncre
ased
feed
stoc
k de
man
d du
e to
bio
fuel
s
By
prod
ucts
repl
ace
feed
cro
ps
Incr
ease
d yi
elds
lower productivity of new cropland
LUC increase with fixed food demand
Net LUC with reduced food demand
Red
uced
food
dem
and
LUC
How economic models work:
12
So......
• Either remove this “iLUC credit” from a reduction in food demand,
• OR accept that part of the biofuels “benefit” is people eating less
Big ILUC credits from less food
All economic models incorporate iLUC savings from reduced food demand.
Without reduction in food consumption, crop production would increase from 10% to 220%, according to feedstock type/model (see figure in backup slide)
IFPRI: 29% increase, but stopping cereals replacing fruit and vegetables (reduction in food quality) would increase by 300-400%)
13
US legislation – ILUC Emissions
Corn Sugar Cane Soy Palm Oil
14
Impacts on biodiversity1. IFPRI says ILUC happens on these land types:
Other environmental effects of ILUC
2. JRC roughly estimated change in species abundance (an indicator for biodiversity):
Results
On the land converted by ILUC, on average there may be up to ~85% loss of biodiversity
15
Conclusions
1. There is no scientific support for believing ILUC = 0
3. For EU biofuels most ILUC occurs outside the EU
5. ILUC is not only GHG emissions: the impact on biodiversity could be potentially high.
4. Without savings in food consumption, models would give higher ILUC emissions
6. California and US already account for ILUC
2. Even with uncertainties, ILUC is above zero for all biofuel feedstocks: from ~10 to ~90 gCO2eq/MJ, according to feedstock type (even more, according to US studies )
16
THANK YOU FOR YOUR ATTENTION
All JRC studies available at:
http://re.jrc.ec.europa.eu/bf-tp/index.htm
17
ADDITIONAL SLIDES
Backup/supporting material
18
Main differences in GHG calculations IFPRI-JRC
Peatland emissionsIFPRI: emission factor of 55 tCO2ha-1yr-1
JRC: updated emission factor 86 tCO2ha-1yr-1 (following JRC expert consultation on ILUC – Nov. 2010, recent literature publications and experimental studies)
Land use factor for soil emissionsIFPRI: all crops are considered as ‘annual crop’JRC: oil palm and sugar cane considered as ‘perennial’ and ‘semi-perennial’ perennial crops bring less disturbance to the soil than annual crops
JRC and IFPRI land use models
19
ILUC increases if food consumption is constant – (results of JRC model comparison)
+48%
+220
%
+76%
+101
%
+78% +3
7%
+59%
+87% +1
5%
+87%
+92%
+12%
+19%
+20%
+29%
+21%
Feedstock requirement reported by models Total feedstock with food constantCereals replacing fruits/vegetables
20
Estimation of potential Impacts on biodiversity
1. Additional cropland from IFPRI “central” scenario (~17,000 km2):
• 42% from pasture• 39% from managed forest• 3% from primary forest• 16% from savannah and
grassland
2. Indicator for biodiversity:Mean Species Abundances (MSA), from Global Biodiversity Model (GLOBIO3)
Mean abundance of original species in undisturbed ecosystems
Biodiversity impact
21
3. IFPRI land use classes adapted to GLOBIO3 classes:
Biodiversity impact
22
4. Estimation of total biodiversity loss:
Where: MSAi = Mean Species Abundance of land use type i
%i = % of land conversion according to IFPRI scenario
MSAca = Mean Species Abundance of cultivated area
RESULTS: This rough estimation shows that the land use change foreseen by IFPRI may lead to a decrease in MSA index of ~85% on the converted land
N.B. this is a preliminary estimation of the potential risks for biodiversity. More work foreseen for 2012.
Results – Biodiversity
23
Models
Consequential approaches are very subjective
Already shown
“Historical” approach is oversimplified but verifiable
24
Alternative ILUC approaches: historical
25
Problems with E4tech’s EU-wheat scenario
1. In the most “ILUC negative” scenario, E4tech assume that EU wheat will come from abandoned land in EU. But all economic models show that most crop area expansion caused by EU ethanol demand would be outside EU + it’s unclear how E4tech concluded that EU cropland would be abandoned in the baseline
Historical data shows yields on abandoned EU cropland are much less than average EU yield
Furthermore E4tech worked out too small an area of EU abandoned land by assuming it has EU-average wheat yield
2. That land would otherwise sequester carbon as it reverts to nature.
3. .
But E4tech underestimate the lost carbon sequestration on this land because of a reporting error by Winrock International
26
But they set no limit to how much yield can increase: if they would double the wheat demand, they would automatically almost double the wheat yield.
For EU wheat scenario E4tech get a 12% higher average wheat yield in EU, compared to baseline in the same year. That would require an incredible price increase due to biofuels, according to all published estimates of yield elasticity.
Looked at another way….E4tech roughly doubles the annual rate of yield increase in the EU ethanol scenario. This would mean at least double the rate of investment in farm improvements and research. That would only follow if the expected financial return would also more-than-double. That financial return is proportional to crop price, so the wheat price would have to more than double (due to EU ethanol) to make this possible.
4. E4tech assume that most of the extra wheat in EU will come from yield increase and not from area increase. The ratio of extra yield to extra area is fixed (by historical precedence).
27
3rd E4tech problem: Marginal yields are much less than average in EU
• the countries which lost most crop area 1997-2007 averaged ~65% of EU-average yield [according to EUROSTAT data]
• National data (UK 2004 farm survey) shows cereals yield on marginal UK farms is < 64% of UK average wheat yield.
• The worst field on a farm has on average 63% of the average farm yield. (English farm survey 2004)
0.65 x 0.64 x 0.63 = 0.18
Including any 2 of these 3 factors would more than double the amount of “abandoned land” required, and reverse the E4tech conclusion for EU-wheat -ethanol.