MEMORANDUM

To:	Margaret Sheppard , U.S. EPA 

Cc:	Yaidi Cancel, U.S.EPA

From:	Emily Herzog, Ed Carr, Mark Wagner, ICF International 

Date:	May 21, 2010

Re:	Sensitivity Analysis CMAQ results on projected maximum TFA rainwater
concentrations and maximum 8-hr ozone concentrations (Deliverable under
EPA Contract Number EP-W-06-008 Task Order 038, Task 06)



In addition to updating the baseline analysis for the Assessment of the
Potential Impacts of HFO-1234yf and the Associated Production of TFA on
Aquatic Communities and Local Air Quality, ICF conducted a sensitivity
analysis based on the findings of Papasavva et al. as described in their
2009 paper Estimated 2010 Refrigerant Emissions of
2,3,3,3-tetrafluoropropene (HFC-1234yf) in the United Stated Resulting
from Automobile Air Conditioning.  This memorandum summarizes the
results from the sensitivity analysis and provides a comparison of the
results with the baseline analysis as well as with the findings of
Luecken et al. (2010).

Please contact Mark Wagner at 202-862-1155 with any questions or
comments.

Assessment of Potential HFO-1234yf Emissions

The following section describes the assumptions used to determine
potential emissions of HFO-1234yf.

Scenarios of HFO-1234yf Market Penetration

For the Assessment of the Potential Impacts of HFO-1234yf and the
Associated Production of TFA on Aquatic Communities, Soil and Plants,
and Local Air Quality, two scenarios were developed.  Both scenarios
assume that HFO-1234yf is used as a drop-in replacement for HFC-134a,
with a replacement ratio of 1:1 HFC-134a to HFO-1234yf.    

Scenario 1: HFO-1234yf is used in the United States as a new car
refrigerant in the MVAC sector.  HFO-1234yf begins penetrating the
market in 2016 reaching 100 percent market penetration by 2020, as
described in   REF _Ref215975477 \h  Table 1  below. 

Scenario 2: HFO-1234yf is used in the United States as a new car
refrigerant in MVAC systems, refrigeration, and stationary AC systems.
Similar to scenario 1, by 2020, HFO-1234yf is used in all new MVAC
systems. For refrigeration and stationary AC systems, HFO-1234yf enters
the market in 2010, reaches 30 percent penetration of the market in
2015, and 100 percent by 2020. 

The market penetration scenarios are summarized in   REF _Ref215975477
\h  Table 1  below.

Table   SEQ Table \* ARABIC  1 : Summary of Assumed Penetration of
HFO1234yf in Scenarios Used in Analysis*

Sectors	2010	2015	2020

Scenario 1

MVACs	0%	0%	100%

Scenario 2

MVACs	0%	0%	100%

Refrigeration	6%	30%	100%

Stationary Air Conditioning	6%	30%	100%

*The dates of introduction of HFO-1234yf are consistent with the earlier
analysis, conducted in 2009.  

The specific dates of introduction are used as a demonstration of how
HFO-1234yf might be phased in; actual dates may differ by several years.
 Examining the period from initial introduction of the refrigerant to
market saturation allows the analysis to consider potential impacts of
HFO-1234yf in the near-term (i.e., 10 - 15 years), during which time
there are already specific State goals for ground-level ozone impacts. 
In addition, the analysis considers the potential longer term impacts of
HFO-1234yf after the refrigerant has saturated the market, assuming
continued growth in use.

HFO-1234yf Emission Assumptions

For the sensitivity analysis, emission assumptions were based on the
findings of Papasavva et al. (2009) regarding the emission projections
of HFO-1234yf from the MVAC sector.  In contrast, ICF assumed that
emissions of HFO-1234yf will be equivalent to emissions of HFC-134a for
the baseline analysis.  Emission assumptions for both the baseline
analysis and the sensitivity analysis are summarized below in   REF
_Ref258934555 \h  \* MERGEFORMAT  Table 2 .

Table   SEQ Table \* ARABIC  2 : Refrigerant Loss Assumptions for
Baseline and Sensitivity Analysis

Assumption	Baseline Analysisa	Sensitivity Analysis 

Leak Rate	8%	5.6%

Servicing Leaks (annualized)	10%	3.2%

End-of-life Loss Rate	42.5%	42.5%

a HFC-134a MVAC assumptions from the U.S. EPA Vintaging Model, Version
VM IO File v4.4_12-16-09.   

For both analyses, emissions of refrigerant due to servicing motor
vehicle air conditioning systems were expected to be higher during the
spring and summer, with up to 80% of emissions due to repairs occurring
from April through August.  Furthermore, refrigerant emissions from
servicing motor vehicle air conditioning systems were estimated at 47%
of annual refrigerant emissions from motor vehicles.  As a result, the
analysis assumes that 59.5% of annual emissions of HFO-1234yf would
occur from April through August, with 11.9% of annual emissions
occurring during each of those five months.  The remaining 40.5% of
annual emissions was modeled to occur from September through March, with
5.8% of annual emissions occurring during each of those seven months. 
This assumption was also applied to stationary AC end uses based on the
assumption that similar patterns of seasonal emissions would be
attributable to this end-use.

Estimated Emissions of HFO-1234yf

Using EPA’s Vintaging Model, ICF estimated emissions of HFO-1234yf for
both scenarios using both sets of emission assumptions.  These estimates
are summarized in   REF _Ref215975548 \h  Table 3  while estimated
monthly emission estimates of HFO-1234yf for the year 2050 are provided
in   REF _Ref258940654 \h  Table 4 .  Complete tables with estimated
annual emissions can be found in Annex A and Annex B.  

Table   SEQ Table \* ARABIC  3 : Summary of Estimated Emissions of
HFO-1234yf (MT)*

Scenarios	2020	2030	2040	2050

Baseline Analysis 

Scenario 1 (MVAC)	6,244	28,327	33,093	35,851

Scenario 2 (All Ref and AC)	10,501	43,296	55,070	60,616

Sensitivity Analysis 

Scenario 1 (MVAC)	3,053	15,349	18,841	20,411

Scenario 2 (All Ref and AC)	7,309	30,318	40,817	45,175

* ICF projected emissions of HFO-1234yf using special scenarios designed
specifically for this analysis and run on the U.S. EPA Vintaging Model,
Version VM IO File v4.4_12-16-09.   

Table   SEQ Table \* ARABIC  4 : Summary of Estimated Monthly Emissions
of HFO-1234yf in 2050 (MT)

Scenarios	Jan	Feb	Mar	Apr	May	Jun	Jul	Aug	Sep	Oct	Nov	Dec

Baseline Analysis 

Scenario 1 (MVAC)	2,074	2,074	2,074	4,266	4,266	4,266	4,266	4,266	2,074
2,074	2,074	2,074

Scenario 2 (All Ref and AC)	4,025	4,025	4,025	6,489	6,489	6,489	6,489
6,489	4,025	4,025	4,025	4,025

Sensitivity Analysis 

Scenario 1 (MVAC)	1,181	1,181	1,181	2,429	2,429	2,429	2,429	2,429	1,181
1,181	1,181	1,181

Scenario 2 (All Ref and AC)	3,131	3,131	3,131	4,651	4,651	4,651	4,651
4,651	3,131	3,131	3,131	3,131

Figure 1, Figure 2, and Figure 3 graphically depict the estimated annual
HFO-1234yf emissions for each scenario.  Figure 4 graphically depicts
the estimated 2050 monthly HFO-1234yf emissions for each scenario. 

Figure   SEQ Figure \* ARABIC  1 : Estimated Emissions of HFO-1234yf by
End-Use Sector under the Baseline Analysis  (MT)

Figure   SEQ Figure \* ARABIC  2 : Estimated Emissions of HFO-1234yf by
End-Use Sector under the Sensitivity Analysis (MT)

Figure   SEQ Figure \* ARABIC  3 : Estimated Annual Emissions of
HFO-1234yf under All Scenarios (MT)

Figure   SEQ Figure \* ARABIC  4 : Estimated Monthly Emissions of
HFO-1234yf under All Scenarios in Year 2050 (MT)

Results of the Sensitivity Analysis

Using EPA’s CMAQ model version 4.7, a one-year simulation was
performed using 2001 meteorological data to project maximum monthly TFA
rainwater concentrations, monthly maximum dry deposition of TFA, and
maximum 8-hr ozone concentrations for the Los Angeles and Houston
regions resulting from emissions of HFO-1234yf from those regions only. 
These regions were chosen based on the potential presence of vernal
pools, their susceptibility to high TFA deposition concentrations, and
their continuing and chronic problem with high levels of ground level
ozone.  The model used the extended version of the carbon bond five
chemical mechanism as developed by Luecken et al. (2010), which
incorporates the gaseous and aqueous chemistry of HFO-1234yf and its
products.  In ICF’s Assessment of the Potential Impacts of HFO-1234yf
and the Associated Production of TFA on Aquatic Communities, Soil and
Plants, and Local Air Quality (2010) results from the baseline analysis
are presented.  The results for the sensitivity analysis are summarized
below.  

TFA Rainwater Concentrations

Monthly maximum average TFA rainwater concentrations for each region
under each scenario are summarized in   REF _Ref260304976 \h  Table 5 .

Table   SEQ Table \* ARABIC  5 : Summary of Estimated Monthly Average
TFA Rainwater Concentrations in 2050 (ng/L)

Region	Jan	Feb	Mar	Apr	May	Jun	Jul	Aug	Sep	Oct	Nov	Dec

Sensitivity Analysis, Scenario 1 

Los Angeles	59	21	109	108	497	572	363	484	107	78	128	61

Houston	12	14	21	17	26	38	54	30	28	12	10	17

 Sensitivity Analysis, Scenario 2

Los Angeles	130	47	240	240	1099	1267	803	1072	237	173	283	136

Houston	26	30	46	38	57	85	120	67	63	27	23	38



Under scenario 1 for the sensitivity analysis, the results indicate that
the maximum monthly average TFA rainwater concentration in the Los
Angeles region in 2050 is 572 ng/L, while the maximum monthly average
TFA rainwater concentration in the Houston region in 2050 is 54 ng/L. 
Under scenario 2 for the sensitivity analysis, the results indicate that
the maximum monthly average TFA rainwater concentration in the Los
Angeles region in 2050 is 1,267 ng/L, while the maximum monthly average
TFA rainwater concentration in the Houston region in 2050 is 120 ng/L.  

Similar to the baseline analysis, for Los Angeles, June represents the
month with the highest projected monthly average TFA rainwater
concentration, while for Houston, July represents the month with the
highest projected monthly average TFA rainwater concentration.  The
monthly average maximum TFA rainwater concentrations for these two peak
months are geographically depicted below in   REF _Ref260395334 \h  \*
MERGEFORMAT  Figure 5  and   REF _Ref260395338 \h  \* MERGEFORMAT 
Figure 6 , for both scenarios, for Los Angeles and Houston,
respectively.  

Figure   SEQ Figure \* ARABIC  5 : June 2050 Los Angeles Sensitivity
Analysis Scenario 1 and Scenario 2 Results 

Figure   SEQ Figure \* ARABIC  6 : July 2050 Houston Sensitivity
Analysis Scenario 1 and Scenario 2 Results 

TFA Dry Deposition

  REF _Ref260387062 \h  Table 6  presents the results of both the
maximum annual and monthly maximum TFA deposition expressed on an annual
basis.  

Table   SEQ Table \* ARABIC  6 : Summary of Estimated TFA Dry Deposition
in 2050 

	TFA Maximum Monthly Averagea 	TFA Maximum Annual Average

	nmole/m2-yr	(g/m2-yr	nmole/m2-yr	(g/m2-yr

Sensitivity Analysis, Scenario 1

Los Angeles	8,698	323	2,188	81

Houston	1,455	54	534	20

Sensitivity Analysis, Scenario 2

Los Angeles	16,686	619	4,745	176

Houston	2,813	104	1,148	43



Notes:  Formula weight of TFA is 114 g/mole. 

a Maximum concentration occurred in May for Los Angeles and in July for
Houston.  Results are expressed on annual basis using the peak monthly
deposition rate times 12. 

Ground Level Ozone Concentrations

Under scenario 1 for the sensitivity analysis, results for the Los
Angeles and Houston regions show a maximum increase in the 8-hr ozone
concentration of 0.030 ppb and 0.009 ppb, respectively.  Under scenario
2 for the sensitivity analysis, results for the Los Angeles and Houston
regions show a maximum increase in the 8-hr ozone concentration of 0.057
ppb and 0.016 ppb, respectively.  Assuming an 8-hr ozone standard of 75
ppb, HFO-1234yf is projected to contribute to a maximum of roughly 0.04
– 0.076% of total ground-level ozone in Los Angeles and 0.012 –
0.021% and of total ground-level ozone in Houston.  Maximum increases in
8-hr ozone concentrations are geographically depicted in  REF
_Ref260308122 \h  \* MERGEFORMAT   Figure 7 ,   REF _Ref260309321 \h  \*
MERGEFORMAT  Figure 8 , and   REF _Ref261261039 \h  \* MERGEFORMAT 
Figure 9 . 

Figure   SEQ Figure \* ARABIC  7 : LA May 22, 2050 Sensitivity Analysis
Scenario 1 and Scenario 2 Results

Figure   SEQ Figure \* ARABIC  8 : Houston June 11, 2050 Sensitivity
Analysis Scenario 1 and Scenario 2 Results

Figure   SEQ Figure \* ARABIC  9 : Houston August 29, 2050 Sensitivity
Analysis Scenario 1 Results

Comparison of Results

For comparison purposes, the following table summarizes results of the
baseline analysis, the sensitivity analysis, and Luecken et al. (2010). 

Table   SEQ Table \* ARABIC  7 : Comparison of Results for Southern
California

Analysis	Maximum Projected TFA Rainwater Concentration (ng/L)	Maximum
Estimated TFA Dry Deposition ((g/m2)*	Maximum Increase in 8-hr Ozone
Formation

Baseline, Scenario 1 	1,116	143	0.07%

Baseline, Scenario 2 	1,700	237	0.11%

Sensitivity, Scenario 1 	572	81	0.04%

Sensitivity, Scenario 2 	1,267	176	0.08%

Luecken et al. (2010)	1,264	240	0.1%

*ICF results represent the annual average maximum for Southern
California, while Luecken’s results represent the 2.5 month spatial
average.  For Southern California, Luecken’s results show TFA dry
deposition to be about 150 (g/m2.

The relationship between HFO-1234yf emissions and projected maximum TFA
rainwater concentrations for both regions are also displayed below in  
REF _Ref261284717 \h  Figure 10  and   REF _Ref261631193 \h  Figure 11 .
 

Figure   SEQ Figure \* ARABIC  10 : Relationship between HFO-1234yf
emissions and maximum TFA rainwater concentrations in LA

Figure   SEQ Figure \* ARABIC  11 : Relationship between HFO-1234yf
emissions and maximum TFA rainwater concentrations in Houston

As shown above, the relationship between HFO-1234yf emissions and TFA
rainwater concentrations appears to be relatively linear.  To support
this relationship, the data for each region were fitted to a linear
regression model.  The equations for each regression model are as
follows:

Los Angeles TFA Rainwater Concentration = 51.55 + (HFO-1234yf Emissions
* 0.027)

Houston TFA Rainwater Concentration = 4.64 + (HFO-1234yf Emissions *
0.0026)

The equations above can be used to estimate TFA rainwater concentration
projections for other HFO-1234yf emission scenarios.  However, it must
be noted that the equations assume that the current chemical mechanism
of gas-phase and aqueous chemistry of HFO-1234yf and its products
remains substantially the same in future analyses.  Also, it is possible
that results may differ if other meteorological datasets (e.g., 2005) or
higher spatial resolutions (e.g., 12-km) were used in the original
analysis.  Finally, due to localized variability in meteorology and
emission distributions, these results cannot be applied to other regions
of the U.S.  Annex A: Annual HFO-1234yf Emissions Estimated under Each
Scenario for the Baseline Analysis for Years 2010 through 2050 (MT)*

Year	Baseline Scenario 1	Baseline Scenario 2

	MVACs	MVACs	Refrigeration	Stationary AC	TOTAL

2010	-	-	39	7	46

2011	-	-	111	20	132

2012	-	-	218	40	258

2013	-	-	361	66	426

2014	-	-	541	98	640

2015	-	-	762	138	900

2016	407	407	1,093	196	1,696

2017	1,229	1,229	1,541	273	3,043

2018	2,472	2,472	2,110	369	4,951

2019	4,143	4,143	2,807	485	7,435

2020	6,244	6,244	3,636	620	10,501

2021	8,332	8,332	4,487	766	13,585

2022	10,413	10,413	5,389	916	16,718

2023	12,513	12,513	6,278	1,066	19,857

2024	14,644	14,644	7,209	1,216	23,069

2025	16,812	16,812	8,128	1,367	26,307

2026	19,003	19,003	9,043	1,516	29,561

2027	21,203	21,203	9,953	1,678	32,834

2028	23,962	23,962	10,914	1,848	36,724

2029	26,335	26,335	11,817	2,017	40,170

2030	28,327	28,327	12,748	2,221	43,296

2031	29,926	29,926	13,591	2,412	45,929

2032	31,118	31,118	14,348	2,577	48,043

2033	31,295	31,295	14,918	2,720	48,932

2034	31,514	31,514	15,479	2,861	49,854

2035	31,791	31,791	15,924	3,010	50,725

2036	32,084	32,084	16,382	3,213	51,678

2037	32,374	32,374	16,824	3,408	52,606

2038	32,627	32,627	17,249	3,592	53,468

2039	32,857	32,857	17,656	3,767	54,280

2040	33,093	33,093	18,046	3,931	55,070

2041	33,345	33,345	18,371	4,008	55,724

2042	33,637	33,637	18,668	4,080	56,385

2043	33,906	33,906	18,921	4,151	56,977

2044	34,177	34,177	19,156	4,217	57,549

2045	34,450	34,450	19,374	4,278	58,102

2046	34,726	34,726	19,566	4,315	58,606

2047	35,004	35,004	19,756	4,351	59,110

2048	35,284	35,284	19,944	4,383	59,611

2049	35,566	35,566	20,132	4,415	60,112

2050	35,851	35,851	20,318	4,447	60,616

* ICF projected emissions of HFO-1234yf using special scenarios designed
specifically for this analysis and run on the U.S. EPA Vintaging Model,
Version VM IO file v4.4_12-16-09. 

Annex B: Annual HFO-1234yf Emissions Estimated under Each Scenario for
the Sensitivity Analysis for Years 2010 through 2050 (MT)*

Year	Sensitivity Scenario 1	Sensitivity Scenario 2

	MVACs	MVACs	Refrigeration	Stationary AC	TOTAL

2010	-	-	39	7	46

2011	-	-	111	20	132

2012	-	-	218	40	258

2013	-	-	361	66	426

2014	-	-	541	98	640

2015	-	-	762	138	900

2016	199	199	1,093	196	1,488

2017	601	601	1,541	273	2,415

2018	1,209	1,209	2,110	369	3,688

2019	2,026	2,026	2,807	485	5,317

2020	3,053	3,053	3,636	620	7,309

2021	4,073	4,073	4,487	766	9,326

2022	5,091	5,091	5,389	916	11,396

2023	6,118	6,118	6,278	1,066	13,461

2024	7,159	7,159	7,209	1,216	15,584

2025	8,219	8,219	8,128	1,367	17,714

2026	9,290	9,290	9,043	1,516	19,849

2027	10,366	10,366	9,953	1,678	21,997

2028	12,206	12,206	10,914	1,848	24,968

2029	13,868	13,868	11,817	2,017	27,702

2030	15,349	15,349	12,748	2,221	30,318

2031	16,647	16,647	13,591	2,412	32,650

2032	17,749	17,749	14,348	2,577	34,674

2033	17,819	17,819	14,918	2,720	35,456

2034	17,918	17,918	15,479	2,861	36,259

2035	18,077	18,077	15,924	3,010	37,011

2036	18,257	18,257	16,382	3,213	37,851

2037	18,444	18,444	16,824	3,408	38,675

2038	18,595	18,595	17,249	3,592	39,436

2039	18,718	18,718	17,656	3,767	40,141

2040	18,841	18,841	18,046	3,931	40,817

2041	18,973	18,973	18,371	4,008	41,352

2042	19,150	19,150	18,668	4,080	41,898

2043	19,303	19,303	18,921	4,151	42,375

2044	19,458	19,458	19,156	4,217	42,830

2045	19,613	19,613	19,374	4,278	43,265

2046	19,770	19,770	19,566	4,315	43,651

2047	19,928	19,928	19,756	4,351	44,035

2048	20,088	20,088	19,944	4,383	44,415

2049	20,249	20,249	20,132	4,415	44,795

2050	20,411	20,411	20,318	4,447	45,175

* ICF projected emissions of HFO-1234yf using special scenarios designed
specifically for this analysis and run on the U.S. EPA Vintaging Model,
Version VM IO file v4.4_12-16-09.   

 In all scenarios, HFO-1234yf replaces only pure HFC-134a, not any
blends with HFC-134a as a component.

.

 Hoffpauir, Elvis.  2009.  Email to Karen Thundiyil, 10 March.

 U.S. EPA Vintaging Model, Version VM IO File v4.2 10-07-08 

  PAGE   \* MERGEFORMAT  8 

