Quiero saber acerca de...

53164

Federal Register / Vol. 86, No. 183 / Friday, September 24, 2021 / Proposed Rules
ammonium and one nitrate. However, unlike measurements and photochemical modeling, a simple emissions ratio does not account for various processes mentioned above; it assumes all the emitted molecules find one another and react. The State found ammonia to be roughly three times as abundant as NOX in 2013 after accounting for their differing molecular weights, and even more abundant in future years. The EPA repeated the exercise to account for SOX and found that the ratio of total ammonia to the ammonia needed to react with both nitrate and sulfate ranged from 2.7 in 2013 to 5.6 in 2028. These results are approximately the same as the CARB
NOX-only results because SOX
emissions are very small relative to NOX
and ammonia emissions e.g., in 2013, winter daily emissions were 8.4 tpd of SOX versus 300.5 tpd of NOX and 309.8
tpd of ammonia.139 These observations support the States finding that PM2.5 is expected to be relatively insensitive to ammonia reductions, though they are not definitive on their own.
The State also points to large decreases or projected decreases in NOX
emissions in the San Joaquin Valley from 2013 to 2024, including a 36
percent reduction from baseline measures by 2020, and a 53 percent reduction by 2024, while CARB projects that ammonia emissions will remain roughly constant i.e., decreasing 12
percent. In conjunction with the ambient evidence that ammonia is already chemically overabundant relative to NOX in the San Joaquin Valley, this indicates that the overabundance will become even greater in the future, and thus ambient PM2.5 is expected to be even less responsive to ammonia reductions. This adds conservatism to the States conclusions about ammonia sensitivity based on the scientific studies.
While the base year for an attainment plan for a given nonattainment area is generally more representative of current conditions, there can be situations in which is it more appropriate to use future conditions representative of when sources will operate, and the EPA
believes that states may use either a base year or a future year for modeling an ambient PM2.5 response to precursor emissions reductions, provided the state explains how the choice of analysis year and associated assumptions are appropriate.140 The 2013 modeled responses cannot be considered current at the present time, in comparison to the 139 2018 PM
2.5 Plan, Appendix B, tables B2, B
3, and B4.
140 PM
2.5 Precursor Guidance, 3536.

VerDate Sep<11>2014

17:10 Sep 23, 2021

Jkt 253001

2020 results. Large NOX emissions reductions have occurred from 2013
2020 and are projected to continue to occur on through 2024, continuing to decrease the ratio of NOX to ammonia.
In light of this ongoing trend, and the ambient data indicating that models underestimate ammonia, the EPA
believes that future year results, which more accurately reflect the expected NOX to ammonia ratio, will continue to be representative, unlike the 2013 base year. These reductions are the result of regulations put in place by past air quality planning decisions and they will occur regardless of the actions that are being proposed herein. In assessing the effect of potential ammonia reductions, the EPA believes it is reasonable to account for these NOX reductions. In addition, as noted above, the greater abundance of ammonia relative to NOX
in the 2024 year modeling is more consistent with recent ambient measurements, which suggest that the 2024 responses are more representative of current atmospheric conditions than the other model years for assessing sensitivity to ammonia reductions.
Therefore, in consideration of the scientific studies and emissions trends, including the projected large amount of NOX emissions reductions through the attainment period, the EPA agrees that use of a future year is appropriate.
Given the available research and ambient data, we conclude that the modeled 2024 year is the most representative of conditions in the San Joaquin Valley.
Even if we were to set aside the more representative 2024 modeling, in the 2020 modeled responses, only the Bakersfield-Planz site is above the contribution threshold, at 1.9 mg/m3. A
single value above the threshold is not determinative, particularly in light of the additional information provided above, indicating that the modeled values overestimate the contribution of ammonia to ambient PM2.5 levels, and that the trend continues toward less contribution in the future as the ratio of NOX to ammonia continues to drop.
Moreover, the monitored 2020 design value is attaining the 1997 24-hour PM2.5 NAAQS because, as discussed above and in section V of this proposal, at the current time there are not PM2.5
levels above the NAAQS. This is further evidence that the single 2020 modeled response above the contribution threshold is not a significant contribution to PM2.5 levels in excess of the NAAQS, even if the 2020 modeling were considered representative.
In the context of interpreting the full set of modeling results for ammonia emissions reductions, the EPA also
PO 00000

Frm 00016

Fmt 4701

Sfmt 4702

considered the States conclusion that the absence of available ammonia controls for sources in the San Joaquin Valley supports its decision to treat a 30
percent reduction as a reasonable upper bound on the ammonia emissions reductions to model in estimating the precursor contribution. As the State correctly notes, the 30 percent to 70
percent range recommended by the EPA
is based on historical NOX and SOX
emissions reductions, and changes in ammonia emissions levels nationally from 2011 to 2017 ranged from a 9
percent decrease to a 6 percent increase.141 The States descriptions of past research relied upon to develop existing rules that apply to ammonia emissions sources, as well as ongoing research, show that it has considered the availability of ammonia controls both in the past and present context, and that the State has a basis for its conclusion that 30 percent is a reasonable upper bound on achievable reductions for ammonia.
In sum, we find that the State quantified the sensitivity of ambient PM2.5 levels to reductions in ammonia using appropriate modeling techniques that performed well, and that the States analysis and use of future year sensitivity data, both 2020 and 2024, is well-supported. We also find that the State adequately documented its basis for using a 30 percent reduction in ammonia emissions as an upper bound in the modeling to assess ambient sensitivity to ammonia emissions reductions. Based on these considerations, the EPA proposes to approve the States demonstration that ammonia emissions do not contribute significantly to ambient PM2.5 levels that exceed the 1997 24-hour PM2.5 NAAQS
in the San Joaquin Valley.
b. SOX
For SOX, the State found that the ambient PM2.5 responses to SOX
emissions reductions were below the EPAs recommended contribution threshold of 1.3 mg/m3 in the Draft PM2.5
Precursor Guidance and below the EPA
recommended threshold of 1.5 mg/m3 in the final PM2.5 Precursor Guidance, and that for most sites there would be an increase in ambient PM2.5 levels in response to SOX reductions i.e., a disbenefit. The EPA has evaluated the States analysis of this disbenefit and resulting conclusion regarding significance.
Because the results of the sensitivity analysis were all below the EPAs recommended 24-hour contribution thresholds at both the 30 percent and 70
141 Id.

E:FRFM24SEP2.SGM

at 30, Table 2.

24SEP2