MEMORANDUM

This memorandum details notable revisions, recalculations, and
clarifications for the ICF memorandum to EPA, International Agriculture
GHG Emissions and GHG Metrics (revised) V.3  (ICF 2008).  Updates are
largely based on suggestions from the report, Peer Review of
International Greenhouse Gas Emissions and Factors as provided to EPA to
support its RFS2 rulemaking, referred to hereafter as the Peer Review
Report. The discussion begins with a summary of notable changes to N2O
emission estimates from fertilizer application, then follows the general
organization of the previous memorandum (ICF 2008): 

Summary of Changes to N2O Emission Estimates from Fertilizer Application

Fertilizer and Pesticide Consumption Projections

N2O Emissions from Fertilizer Consumption and Crop Residues

GHG Emission Rates for Agricultural Energy Use

CH4 Emission Factors for Rice Cultivation

Additional Research

Each section details updates as well as resulting changes in emissions,
if applicable. A revised appendix incorporating expert reviewer comments
is also included.

Summary of Changes to N2O Emission Estimates from Fertilizer Application

Overall emissions related to fertilizer application have been revised
downward from 1.90 billion kg N2O in the ICF 2008 memo, to 3.36 million
kg N2O (see   REF _Ref241901619 \h  Table 1 ). In performing QC checks,
we uncovered a conversion error in the calculations that had caused the
calculations to be higher by a factor of 1,000. The error and its
correction are fully detailed in Section II.

After this correction was made, additional updates to the data and
emissions methodology resulted in increases in emission estimates,
though all emissions presented in this memo are much lower than those
presented in ICF 2008 due to the conversion error. The increases in
direct and indirect N2O emissions from synthetic fertilizer application
are almost entirely a result of using updated fertilizer consumption
data from the International Fertilizer Industry Association (IFA) for
the periods 2006 or 2006/07 and 2007 or 2007/08. Increases in direct and
indirect emissions from crop residues results from an increase in
agricultural land use acreage and from the correction of an additional
conversion error affecting the crop residues calculations (also
discussed in Section II). Additionally, this analysis updates the ICF
2008 analysis to include crop residue N emissions from cotton, palm oil,
rapeseed, sugarbeet, sugarcane, or sunflower. As explained in more
detail below, the ICF 2008 memo did not include these crops because
default crop-specific IPCC factors used in the equation were not
available. Proxy IPCC factors for these crops are employed in this
analysis. 

Table   SEQ Table \* ARABIC  1 : N2O emissions from fertilizer
application and crop residues (kg)

Emissions Source	ICF 2008 Analysis 	Revised ICF 2008 Analysis 	ICF 2009
Analysis

Direct N2O Emissions from Synthetic Fertilizer Application	1,149,886,629
1,149,887	1,995,628

Indirect N2O Emissions from Synthetic Fertilizer Application	373,713,154
373,713	648,579

Direct Emissions from Crop Residues	307,993,297	761,241	580,553

Indirect Emissions from Crop Residues	69,298,492	171,337	133,736

TOTAL	1,900,891,572	1,901,104	3,358,496



Emission rates from agricultural energy consumption have not been
revised. Additionally, the emission factor (EF) for rice cultivation in
the United States has been revised to use the IPCC 2006 default EF for
the applicable cropping regime. 

Finally, we have included an additional emissions source – Manure
Fertilizer Use – which increases emissions by 188,762 kg N2O. 

Fertilizer and Pesticide Consumption Projections

Revisions to fertilizer and pesticide consumption projections result
from the incorporation of updates made by the Food and Agriculture
Organization (FAO) to its Fertistat and FAOStat datasets, as well as the
incorporation of more up-to-date fertilizer consumption statistics
provided by a recent IFA report, Assessment of Fertilizer Use by Crop at
the Global Level, 2006/07 – 2007/08 (Heffer 2009). Additionally, data
tables provided in the Appendix have been updated to report data in
metric units only. Updates and recalculations are described in more
detail below, along with details on how these updates affect greenhouse
gas (GHG) emission estimates. 

Edits to Fertilizer Consumption Projections

The following updates and activities were conducted: 

A. Incorporated IFA Fertilizer Consumption Data

Historical fertilizer application rates for nitrogen (N), phosphorus
(P2O5), and potassium (K2O) were updated by using more recent data
available from the International Fertilizer Industry Association (Heffer
2009), as recommended by expert reviewers in the Peer Review Report. 

IFA data are preferred over FAO’s Fertistat data because estimates are
more current and years of available data are consistent across all
countries. Additionally, FAO has altered its survey methodology since
2004. 

The IFA dataset covers 23 countries (considering the European Union-27
as one country) and 11 crop groups including: wheat, rice, corn, other
coarse grains, soybean, palm oil, other oilseeds, cotton, sugar crops,
fruits and vegetables, and other crops. IFA data cover roughly 92% of
world fertilizer consumption for the periods 2006 - 2006/07 and 2007 -
2007/08. IFA fertilizer consumption statistics are collected by a
network of correspondents and fertilizer associations. 

IFA consumption data were averaged over the two reported time periods:
2006 or 2006/07, and 2007 or 2007/08 to account for seasonal variations
in crop production. Fertilizer application rates were calculated by
dividing IFA total consumption values by FAOStat agricultural area
harvested data from the FAOStat database. Agricultural area harvested is
based on an average of 2006 and 2007 data to account for the fact that
IFA collects data in both calendar year and fertilizer year. Previously,
in the ICF 2008 memo, individual country research was conducted to
determine national fertilizer consumption statistics for the following
country/crop combinations: Russia/wheat, China/wheat, and India/soybean.
As IFA provides recent data for these combinations, these independent
estimates are no longer needed. 

IFA data for “sugar crops” includes both sugar beet and sugar cane.
Ratios of sugarcane to sugar crop were developed for IFA countries using
Fertistat data for sugar beet and sugar cane to determine the relative
percentage of each sugar crop by country. Where Fertistat data were not
available to match an IFA country, data were proxied using nearby
countries with similar climates. The following proxies were used to
estimate a sugar cane to sugar beet ratio from the Fertistat dataset:

Table   SEQ Table \* ARABIC  2 : Fertistat countries used to approximate
sugarcane ratio for IFA sugar crop data

Country	Proxy Country/Region

Canada	United States

Malaysia	Indonesia

Iran	Turkey

Russia	Other Eastern Europe

Argentina	Chile



Overall, incorporating IFA data increased total fertilizer related
emissions by 973,919 kg N2O, with 735,034 kg N2O attributable to direct
emissions from synthetic fertilizer application and 238,886 kg N2O
attributable to indirect emissions from synthetic fertilizer
application.

 

B. Updated FAO Fertistat Data

FAO has updated its Fertistat dataset since the ICF 2008 memo;
therefore, ICF downloaded and used the most recent version. Overall,
total fertilizer consumption has decreased by 3.34 million tonnes for N,
 2.22 million tonnes for P2O5, and 0.81 million tonnes for K2O. Total
agricultural area has also decreased by 50.8 million hectares. 

Fertistat data are used for country/crop and region/crop combinations
not covered by the IFA dataset. While IFA provides fertilizer
consumption data for a “rest of world” region, Fertistat data are
preferred to calculate fertilizer consumption for “rest of world”
since the dataset provides information for a greater number of countries
and because IFA “rest of world” values include some of the
individual priority countries requested by EPA but not available in the
IFA dataset. Additionally, while IFA provides fertilizer consumption
data for the EU-27 region, the dataset is not complete for all priority
crops. As a result, Fertistat data is aggregated for available EU
countries for remaining priority crops, including barley, peanut,
rapeseed, sorghum, and sunflower.  

Fertistat fertilizer application rates (kg/ha) are calculated by
dividing total Fertistat fertilizer consumption by Fertistat
agricultural area fertilized. 

C. Researched Fertilizer and Lime Use Associated with Soybean Production

Expert reviewers requested that additional research be conducted to
review fertilizer consumption for soybean production expansion in Brazil
as well as associated lime use. Both the FAO and IFA provide data for
Brazil soybean production and fertilizer use (see   REF _Ref241485143 \h
 Table 3 ): 

Table   SEQ Table \* ARABIC  3 : Fertilizer consumption statistics for
Brazil/Soybean

Data Source	Year	N 

(‘000 tonnes)	P2O5 

(‘000 tonnes)	K2O 

(‘000 tonnes)

IFA 	Avg 2006 - 2006/07 & 2007 - 2007/08	97	1,426	1,313

FAO	2004	148	1,217	1,144



IFA data are used to project fertilizer consumption for Brazil/soybean
production as this data is more recent. In response to an expert
reviewer comment, additional research was conducted into lime use for
expanded soybean production in Brazil. Given limited data availability
however, lime use data and resulting emissions are not incorporated in
this analysis.  

Edits to Pesticide Consumption Projections

Historical pesticide consumption estimates for fungicides and
bactericides, herbicides, and insecticides have been updated using the
most current data available from FAO’s FAOStat dataset for pesticide
consumption. FAO has also updated its FAOStat dataset for agricultural
area since the ICF 2008 memorandum.  Pesticide consumption data was
added for Belgium. 

Expert reviewers (Peer Review Report 2009) noted that FAO’s pesticide
consumption dataset did not provide values for China pesticide
consumption. To remedy this, ICF researched pesticide consumption in
China. The U.S. Department of Agriculture’s (USDA) Economic Research
Service (ERS) states that total pesticide consumption for China in 2007
was 1.623 million tons. As ERS does not break out total pesticide
consumption into the three pesticide product types, total pesticide
consumption was distributed using the average ratio of pesticide
consumption from nearby countries with a similar climate, including
Vietnam, Thailand, and South Korea. 

Table 4 in the data appendix has been updated to reflect instances where
FAO reports zero consumption values versus non-responses from its data
collection survey. Empty cells represent survey non-responses.

N2O Emissions from Fertilizer Consumption and Crop Residues

In response to a request from EPA, research was conducted into N2O
emissions from the application of manure to agricultural soils as an
organic fertilizer.  Previously, only synthetic fertilizer and crop
residue emissions were estimated. EPA’s manure management calculations
were used to make the estimates of manure applied to soils. In addition,
the equation and estimates of N inputs from crop residues were reviewed,
and estimates of crop residues from several crops that had previously
been excluded due to lack of default IPCC factors specific to these
crops were added. This section discusses these updates in more detail,
along with how updates have altered GHG emission estimates. 

Projections of Changes in GHG Emissions from Manure Fertilizer Use

In order to estimate emissions from use of manure as organic fertilizer,
this analysis uses EPA’s existing estimates of N excretion (kg
N/animal-year) by animal type and region, the distribution of animals
into waste management systems by type and region, and the projected
change in livestock population by type and region, which had all been
used in EPA’s manure management emissions calculation. With this
information, this analysis estimates the manure N available and
subtracts the amount typically lost before application, according to the
IPCC methodology (see IPCC 2006 Guidelines Section 10.5.4 and Table
10.23). However, due to lack of information, information on additional N
that might be added to the manure from bedding materials (e.g. wood
chips, straw) was not included.

The analysis uses IPCC equation 10.34, which calculates managed manure N
available for application to managed soils, feed, fuel or construction
uses. It is the sum, for each livestock species (T) and manure
management system (S), of:

 [NT × NexT × MST,S × (1 – FracLoss MS/100)] + NT × MST,S ×
NbeddingMS

Where:

NT 	= number of head of livestock species/category T in the country

NexT	= annual average N excretion per animal of species/category T in
the country, kg N animal-1 yr-1

MST,S 	= fraction of total annual nitrogen excretion for each livestock
species/category T that is managed in manure management system S in the
country, dimensionless

FracLoss MS	= amount of managed manure nitrogen for livestock category T
that is lost in the manure management system S, % (see Table 10.23)

NbeddingMS 	= amount of nitrogen from bedding (to be applied for solid
storage and deep bedding MMS if known organic bedding usage), kg N
animal-1 yr-1

S 	= manure management system

T 	= species/category of livestock

As stated above, it is assumed that NbeddingMS is zero, due to lack of
data.  IPCC Table 10.23 (  REF _Ref242005333 \h  Table 4 ) gives the
values for FracLoss MS.  Some manure management systems used in EPA’s
manure management calculations for certain livestock types did not
correspond with an entry in Table 10.23. For example, in the manure
management calculation, some non-dairy cattle manure was managed by
daily spread.  However, there is no FracLoss MS value for “other
cattle” for daily spread.  Therefore, in ICF’s calculation, the
dairy cow value for daily spread was used. For other livestock/systems
where a factor was not available, the corresponding dairy cow factor was
similarly used.  For digester systems, the dairy cattle and swine
anaerobic lagoon values were used for all cattle and swine,
respectively.

Table   SEQ Table \* ARABIC  4 : IPCC Table 10.23, Default Values for
Total Nitrogen Loss from Manure Management

 Animal category  	 Manure management system	 Total N loss from MMS
FracLossMS 	(Range of FracLossMS)  

 Swine  	 Anaerobic lagoon  	78%	(55 – 99)

	 Pit storage  	25%	(15 – 30)

	 Deep bedding  	50%	(10 – 60)

	 Liquid/Slurry  	48%	(15 – 60)

	 Solid storage  	50%	(20 – 70)

 Dairy Cow  	 Anaerobic lagoon  	77%	(55 – 99)

	 Liquid/Slurry  	40%	(15 – 45)

	 Pit storage  	28%	(10 – 40)

	 Dry lot  	30%	(10 – 35)

	 Solid storage  	40%	(10 – 65)

	 Daily spread  	22%	(15 – 60)

 Poultry  	 Poultry without litter  	55%	(40 – 70)

	 Anaerobic lagoon  	77%	(50 – 99)

	 Poultry with litter  	50%	(20 – 80)

 Other Cattle  	 Dry lot  	40%	(20 – 50)

	 Solid storage  	50%	(20 – 70)

	 Deep bedding  	40%	(10 – 50)

 Other (sheep, horses)	 Deep bedding  	35%	(15 – 40)

	 Solid storage  	15%	(5 – 20)



In calculating the manure N available for application, the manure N that
had been categorized under the pasture, range, and paddock (PRP)
management system was subtracted because the “application” emissions
from PRP were already accounted for in the manure management
calculations.  Similarly, daily spread manure application emissions were
also accounted for in the manure management calculations and were
subtracted out.  

However, in reviewing the methodology, some possible methodological
discrepancies were detected in the treatment of indirect emissions from
PRP and daily spread, and in the treatment of direct versus indirect
emissions from sheep manure (in the direct calculation 100% of sheep
manure is assumed to be PRP; in the indirect calculation, it is divided
between solid storage and deep bedding).  

These additions will increase emission estimates by 188,762 kg N2O, or
approximately 5.7 percent of the current estimated total emissions from
synthetic fertilizers and crop residues.

It is important to note that the need of crops for fertilizer N is
unlinked in this analysis to changes in animal population. Therefore it
could occur that a country or region has a projected increase in
fertilizer needs (i.e., crop growth) while simultaneously projecting a
decrease in livestock emissions. The amount of synthetic fertilizer
change is directly proportional to the amount of crop area change in
this analysis, so any possible shortfall in crop N supply due to
decreasing livestock populations combined with increasing crop areas is
not addressed.  

Edits to Projections of Changes in GHG Emissions from Synthetic
Fertilizer and Crop Residues

Based on comments from one of the expert reviewers, Dr. Kenneth Cassman,
typographical errors were corrected in the portion of the original memo
describing the methodology used to calculate emissions from synthetic N
and crop residue N additions to agricultural soils. The original text
from the memo is shown below, with typo corrections in red text and
yellow highlight.

Direct Emissions: 

Direct N2O emissions from synthetic fertilizers:  

Emissions  = FSN  × EF1 × 44/28

Where: 

FSN = the annual amount of synthetic fertilizer N applied to soils (kg
N)

EF1 = emission factor,(equal to 0. 01 kg N2O-N/kg N input)

44/28 = conversion of N2O -N to N2O

Direct N2O emissions from crop residues:  

Emissions = FCR × EF1 × 44/28

Where

FCR = the annual amount of N in crop residues and forage/pasture renewal
(kg N)

EF1 = emission factor,(equal to 0.01 kg N2O-N/kg N input)

44/28 = conversion of N2O -N to N2O

N additions to soils from crop residues depend on the crop type and
yield, since different crop types have different N contents and
different amounts of residue typically left in the soil. The equation
for FCR is: 

= ∑ (Yield FreshT × DRYT × ST + IT) × AreaT × (Nag(T) + Rbg-BIO(T)
× Nbg(T))

Where: 

T = crop or forage type

	Yield Fresh = fresh weight yield of crop (kg fresh weight/ha)

	DRY = dry matter fraction of harvested crop (kg dry matter/kg fresh
weight)

	S = Slope for above-ground residue dry matter 

	I = Intercept for above-ground residue dry matter

	Area = total annual area harvested (ha)

	Nag = N content of above-ground residues (kg N/kg dry matter)

Rbg-BIO = ratio of belowground residues to above ground biomass 

Nbg = N content of below-ground residues (kg N/kg dry matter)

The crop residue N input write-up and spreadsheet were also reviewed in
response Dr. Cassman’s comment that the equations used were incorrect.
On review it was concluded that with respect to the comments raised by
Dr. Cassman, the equation used was correct and was used correctly.
However, a conversion error in the equation was uncovered during recent
QC activities.  In ICF 2008, crop area inputs were acre-based.  The crop
area term appears twice in the equation (once to determine Yield Fresh,
and once on its own). However, the conversion from acres to hectares
mistakenly only appeared once.  This problem was rectified when the area
input data was converted to a hectare basis. The effect of this
correction was an increase in emission estimates for all crops by a
factor of about 2.5.

It is important to note that due to lack of data, we have been unable to
account for any portion of the crop residue that might be removed from
the field for other purposes, such as animal bedding or fuel. In some
areas, these removals could be a significant portion of residues, and in
such cases, our estimates are likely to be high.

In addition, adjustments were made to the estimates of crop residue N
inputs. In the original analysis, crop residue N emissions from cotton,
palm oil, rapeseed, sugarbeet, sugarcane, or sunflower were not included
because default crop-specific IPCC factors used in the equation were not
available. A commenter felt that this exclusion was not defensible. In
response, these crops are now included using the “Root crops, other”
category from the IPCC default factors. The equation used to calculate
crop residues was also adjusted to account for the fact that palm oil is
not an annual crop (it is assumed all other crops are annual).  In
accordance with IPCC guidelines, a term was inserted into the equation
with the percent of the crop renewed annually.  Based on ICF’s
research, a replanting interval of 23 years seemed typical, and so the
percent of annual renewal was estimated at 1/23 or 4.35 percent.  For
all other crops the renewal rate was set at 100 percent. These changes
resulted in an estimated decrease in emissions of 208,776 kg N2O.  The
change results in a decrease in emissions because the crops that were
added showed a net decrease in crop area.

Together, the correction of the acreage conversion error and the
inclusion of the additional crop types resulted in a net increase of
336,785 kg N2O from crop residue N inputs. 

Order-of-Magnitude Error in Fertilizer and Crop Residue Equations

In reviewing the N2O emissions calculations during the revision process,
we uncovered an order-of-magnitude error that affected fertilizer
(direct and indirect emissions) and crop residue (direct and indirect)
equations.  In both cases, the equation multiplied by one million in
order to convert from Gg (of crop in the case of crop residues and of N
in the case of fertilizer) to kg.  However, in both cases the multiplier
should have been one thousand, to convert from Mg (metric tons) to kg.
As a result, the totals submitted in the 2008 analysis were too high by
a factor of 1,000. All emissions of N2O given in billions of kg of N2O
should have been in millions of kg of N2O. 

As a further check to confirm the change, we have drawn comparisons with
data from the U.S. Greenhouse Gas Inventory and U.S. Census of
Agriculture, as shown in   REF _Ref242006960 \h  Table 5 .  In 2007,
7,587 Gg (7.59 Tg) of synthetic fertilizer N were added to U.S.
croplands. The resulting direct emissions were 32.2 Tg CO2 Eq., which
equals a rate of 4.24 units CO2 Eq./unit N added. For our analysis, the
change in synthetic N application was 125,547.2 Mg, and the resulting
emissions (according to the revised calculation) were 1,972,885 kg N2O
(1,973 Mg N2O, or 611,594 Mg CO2 Eq.). The resulting rate of emissions
was 4.87 units CO2 Eq./unit N added, which is very similar to the rate
in the United States (the difference is attributable to different
methodologies used).  

As another standard of comparison, the total harvested cropland in the
U.S. in 2007 was 309.6 million acres (125.3 million hectares).  The
total change in acreage in this analysis, for all crops considered was
1.85 million hectares.  We also compared our estimates with the relevant
proportion of U.S. N2O emissions from agricultural soil management.  The
result of these comparisons shows that, broadly speaking, the revised
numbers in this analysis are “normal” in proportion to similar
metrics for the United States.  

Table   SEQ Table \* ARABIC  5 : Comparison of Analysis Results with
Year 2007 Estimates from 1990-2007 U.S. Greenhouse Gas Inventory.

	1990-2007 U.S. GHG Inventory (2007)	International Ag. Analysis
(Corrected)	International Ag. Analysis (Uncorrected)

Units CO2 Eq./unit synthetic N 	4.24	4.87	4,871

Total N2O emissions (Tg CO2 Eq.)	80.3*	1.1**	1032

Hectares of cropland	125.3 million	1.9 million	1.9 million

Emissions rate (Tg CO2 Eq./ha)	0.64	0.58	543

* Includes only direct emissions from synthetic fertilizer, crop
residues, and organic amendments; plus a proportional fraction of
indirect emissions.  Total U.S. ag. soil management N2O emissions were
207.9 million Tg CO2 Eq., but this number includes emissions from
sources not considered in this analysis, such as grasslands and organic
soils.

** Includes direct and indirect emissions from synthetic fertilizer,
crop residues, and application of managed manure to soils as fertilizer.

GHG Emission Rates for Agricultural Energy Use

In response to expert reviewers’ comments, the values in Table 8 of
the Appendix have been converted to MTCO2e/ha to ensure small values
appear as non-zero numbers.  Further, any zero values have been denoted
by a dash (“-”) for clarity. This change will not affect emissions
values since the values in Table 8 have not been revised; they were only
converted to different units.

For all energy data provided, any apparent data gaps exist because no
data were available from IEA for the given parameter.

Several fuel types were excluded from the analysis in the ICF 2008
memorandum.  The fuel types excluded were: biogas, charcoal, gas works
gas, geothermal, other liquid biofuels, primary Solid Biomass, solar
thermal (note that these are IEA’s fuel type categories). These fuel
types were excluded because they are primarily biogenic fuel types,
which are not associated with fossil fuel CO2 emissions.  As for the
non-biogenic fuel types excluded, consumption of these fuel types in the
Agriculture/Forestry sector was minimal and thus excluded.

CH4 Emission Factors for Rice Cultivation

The emission factor (EF) for the United States has been revised to use
the IPCC 2006 default EF for the applicable cropping regime.  In
response to a comment received from an expert reviewer regarding the
value of the U.S. EF, the conversion of the seasonal factor to the daily
factor was checked.  Previously, the conversion was completed using an
average season length from the U.S.’s analysis to determine the EF;
this season length accounted for only a limited number of states.  ICF
recalculated the EF using a season length that accounted for all
rice-growing states (i.e., the U.S. season length from IRRI) and
calculated an EF of 1.246 kg CH4/ha/day. This compares well with the
IPCC default factor for the irrigated cropping regime (the only cropping
regime used in the U.S.) - 1.237 kg CH4/ha/day, thus confirming the
validity of the default EF for the United States.

Additional Research

Key Country Pasture Areas and Livestock Populations

At the request of EPA, research was conducted to review more detailed
country-specific data on pasture areas and populations of pastured
livestock for Argentina, China, India, and Indonesia.  These data are
summarized in the spreadsheet Livestock Land Area_9.30.09.xls, which was
originally sent to EPA on September 14, 2009, and has not been updated
since then.

Information is currently available mainly from national statistical
agencies and agriculture departments.  For the years 2000 – 2007
(according to data availability by country), a summary of livestock
population data by animal type was created. The analysis includes two
“total” population categories, the first for bovines, sheep, goats,
and horses/donkeys/mules, i.e., animals we considered most likely to be
on pastureland, and the second also including swine and camels.  Animal
population data for Argentina were available for 1988 and 2002 for all
animal types, and for 1993 – 2001 for cattle, sheep, and goats.  Data
for China were available for 1996 – 2007.  Data for India were
available for 1982, 1987, 1992, 1997, and 2003.  Data for Indonesia were
available for 1995 – 1999, 2003, and 2005 – 2007.  

In addition, data were summarized for area of pasture or grazing land
for each country.  Where possible, total area under survey, or total
agricultural land, were included as a point of comparison.  Data for
Argentina were available for 1988 and 2002.  Data for China were
available for 2003 – 2007.  Data for India were available for the
first year of every decade going back to 1950, and then annually from
1990 on through the 2005 – 2006 agricultural year.  Data for Indonesia
were available for 1999 – 2003.

To complete the analysis, rough estimates of animal population density
on pasture were created using the two different “total”
categories.  Since data for the housing of the various animal types by
country were not available, these density estimates are based on the
unrealistic assumption that all animals listed in population estimates
are on pasture.

 ICF International (2008). International Agriculture GHG Emissions and
GHG Metrics (revised) V.3.  

 Ross and Associates Environmental Consulting, Ltd (2009). Peer Review
of International Greenhouse Gas Emissions and Factors as provided to EPA
to support its RFS2 rulemaking.

 In 2006 for countries with fertilizer consumption statistics in
calendar years, and in 2006/07 for countries with statistics in
fertilizer years (e.g., growing seasons).

 Heffer, Patrick (2009). Assessment of Fertilizer Use by Crop at the
Global Level, 2006/07 – 2007/08. International Fertilizer Industry
Association. Paris, France. 

 Heffer, Patrick (2009). Personal correspondence between Patrick Heffer,
IFA and Erin Gray, ICF International. Sept. 25, 2009.

 Fertistat data for the region “European Union” might not match
directly with IFA’s dataset as the IFA dataset includes 27 EU
countries, whereas data availability varies by crop for the Fertistat
dataset. 

 For the purpose of this analysis, the crop categories, “other coarse
grains,” “other oilseeds,” and “other crops” are not included
since data is not available to break out these categories into
individual crop components (e.g., barley, sorghum, rapeseed, peanut, and
sunflower). 

 IFA (2009). Frequently Asked Questions (FAQ). Retrieved from  
HYPERLINK
"http://www.fertilizer.org/ifa/Home-Page/STATISTICS/FAQ#IFADATA7" 
http://www.fertilizer.org/ifa/Home-Page/STATISTICS/FAQ#IFADATA7  on
9/24/2009.

 FAO (2009). FAOStat Production (Crops). Retrieved Sept. 25, 2009 from  
HYPERLINK
"http://faostat.fao.org/site/567/DesktopDefault.aspx?PageID=567#ancor" 
http://faostat.fao.org/site/567/DesktopDefault.aspx?PageID=567#ancor . 
Assumes FAO definitions for seed cotton, maize, sunflower seed,
groundnuts, in shell, and rice, paddy equal cotton; corn; sunflower;
peanut; and rice, respectively. 

 “Other Eastern Europe” includes Croatia and Slovakia.

 Emissions increase calculated after updating Fertistat data for
country/crop combinations not covered by IFA. 

 Fertistat data for the EU is EU-27 might not match EU definition for
Fertistat, as fertilizer consumption data availability varies by crop
for Fertistat countries. 

 Fertistat reports that fertilizer consumption is in 1000 tons, but data
is actually reported in metric tonnes.

 Heffer (2009). Op cit. 

 FAO (2004). Fertilizer use by crop in Brazil. FAO Land and Plant
Nutrition Management Service, Land and Water Development Division. Rome,
Italy.

 Updates last reported in April, 2009. Pesticide consumption database
limitations: 1) The country coverage and time series are incomplete due
to a high rate of non-response; 2) Although countries have been
requested to report data in terms of active ingredients, some countries
may have reported in formulation weight (including diluents and
adjuvants) without specific indication.    HYPERLINK
"http://faostat.fao.org/site/291/default.aspx" 
http://faostat.fao.org/site/291/default.aspx  

 USDA Economic Research Service. China Agricultural and Economic Data:
National Data Results. Retrieved from   HYPERLINK
"http://www.ers.usda.gov/Data/China/NationalResults.aspx?DataType=6&Data
Item=160&StrDatatype=Agricultural+inputs&ReportType=0" 
http://www.ers.usda.gov/Data/China/NationalResults.aspx?DataType=6&DataI
tem=160&StrDatatype=Agricultural+inputs&ReportType=0  on Sept 15, 2009.

e

f

}

e

f

¯

؀}

~

¯

á

â

.

9

D

E

F

G

w

Ý

Þ

û

þ

-

.

摧曹óԀ.

w

x

 h

摧曹ó±欀Ɽ

摧曹ó±欀

 hì

 hì

 hì

 hì

 hÇ

 hÇ

 hì

 hì

 hì

愀Ĥ摧曹ó欀╤

 hÇ

 hÇ

혈\鐄䃿뀋–ﰜ$갆

혈\鐄䃿뀋–ﰜ$갆

혈\鐄䃿뀋–ﰜ$갆

혈\鐄䃿뀋–ﰜ$갆

혈\鐄䃿뀋–ﰜ$갆

 h

! Of the total 1.623 million tons of pesticide consumption by China,
this analysis estimates that fungicides and bactericides equal 25% of
pesticide consumption; herbicides equal 33%; and insecticides equal 42%.


 USDA 2007 Census of Agriculture, Table 8.

 PAGE   

 PAGE   2 

Notable Revisions and Recalculations for the Memorandum: 

International Agriculture GHG Emissions and GHG Metrics

