Excerpt from Preamble V.A.1.d
When considering the proposed Tier 3 sulfur standards, refineries can be grouped into three general categories based on their current post-Tier 2 refinery configurations: those without an FCC unit, those hydrotreating the gasoline stream coming from their FCC unit (i.e., postreating), and those hydrotreating their FCC feed (i.e., pretreating).  Most refineries without FCC units would not need to do anything to meet the proposed Tier 3 sulfur standards.  Refineries equipped with FCC units that invested in an FCC postreater under Tier 2 would likely just need to revamp (i.e., renovate) their existing unit for a modest cost.  Refineries that only have an FCC pretreater would meet the Tier 3 sulfur standards by either revamping their existing pretreaters (perhaps also cutting the heavy portion of FCC naphtha into the diesel pool) or invest in a grassroots FCC postreater.  Our refinery-by-refinery modeling suggests that 29 refineries would not need to make any capital changes, 66 would need to revamp their existing FCC postreaters, and 16 would need to add grassroots postreaters (we did not model any undercutting of heavy FCC naphtha into the distillate pool).  Refiners that need to install a new postreater would have to make the largest desulfurization investments under Tier 3, typical of many of the refinery investments made under Tier 2.  For more on our estimated sulfur control costs, refer to Section VII.B.  
We believe that the choice of technology for each refinery is fairly insensitive to capital cost assumptions.  Revamping an existing FCC postreater will almost always be the preferred compliance path if it is available.  The 16 refineries we project would add grassroots postreaters do not have existing postreaters that could be revamped.  We believe based on conversations with industry technology vendors and engineering firms that installing a grassroots postreater would be more likely for these refineries than revamping their pretreater, which would still incur a significant capital cost and reduce compliance flexibility.  Thus, the refinery-by-refinery analysis performed by EPA for this proposal, higher capital costs (either directly or thru a higher ROI) would be unlikely to alter the selection of pretreater vs postreater control technology in our analysis.  Higher capital costs would likely impact both technology options proportionally with no overall effect. 
We have built in a number of flexibilities that will reduce the compliance burden for refiners.  In particular, coupling the proposed 10-ppm annual average sulfur standard with refinery gate and downstream per-gallon caps should continue to allow for batch-to-batch variability, refinery upsets, and turnarounds while still maintaining the overall air quality benefits of the program.  For more information on the applicable per-gallon sulfur caps, refer to Section V.A.3. 

Excerpt from Preamble III.B.7
These reductions would be offset to some degree by CO2 emissions associated with higher energy use required in the process of removing sulfur within the refinery.  To assess the potential refinery permitting implications of the Tier 3 proposal, we calculated the CO2 emission impacts on a refinery-by-refinery basis.  We used the projected refinery-specific changes from our refinery-by-refinery modeling (see Chapter 5 of the draft RIA) to estimate changes in process energy and then applied emission factors that correspond to those changes.  The results showed an increase of up to 4.6 MMTCO2e in 2017 with the implementation of the lower sulfur standards.  The actual increase is expected to be considerably lower, since this is a permitting analysis and refineries will not be operating at their permit capacity.  The actual increase will also be a function of several factors, including technology options selected by the refineries and the projected use of averaging, banking and trading in avoiding the need for investments at some refineries. As a result, 4.6 MMTCO2e represents an upper-bound estimate of the possible increase in refinery CO2 emissions due to the need for additional process heat and hydrogen production to enable the additional hydrotreating required.  

