Increasing the octane rating of the U.S. gasoline pool (currently ∼ 93 Research Octane Number (RON)) would enable higher engine efficiency for light-duty vehicles (e.g., through higher compression ratio), facilitating compliance with federal fuel economy and greenhouse gas (GHG) emissions standards. The federal Renewable Fuels Standard calls for increased renewable fuel use in U.S. gasoline, primarily ethanol, a high-octane gasoline component. Linear programming modeling of the U.S. refining sector was used to assess the effects on refining economics, CO2 emissions, and crude oil use of increasing average octane rating by increasing (i) the octane rating of refinery-produced hydrocarbon blendstocks for oxygenate blending (BOBs) and (ii) the volume fraction (Exx) of ethanol in finished gasoline. The analysis indicated the refining sector could produce BOBs yielding finished E20 and E30 gasolines with higher octane ratings at modest additional refining cost, for example, ∼ 1¢/gal for 95-RON E20 or 97-RON E30, and 3-5¢/gal for 95-RON E10, 98-RON E20, or 100-RON E30. Reduced BOB volume (from displacement by ethanol) and lower BOB octane could (i) lower refinery CO2 emissions (e.g., ∼ 3% for 98-RON E20, ∼ 10% for 100-RON E30) and (ii) reduce crude oil use (e.g., ∼ 3% for 98-RON E20, ∼ 8% for 100-RON E30).

This analysis uses linear programming modeling of the U.S. refining sector to estimate total annual energy consumption and CO(2) emissions in 2025, for four projected U.S. crude oil slates. The baseline is similar to the current U.S. crude slate; the other three contain larger proportions of higher density, higher sulfur crudes than the current or any previous U.S. crude slates. The latter cases reflect aggressive assumptions regarding the volumes of Canadian crudes in the U.S. crude slate in 2025. The analysis projects U.S. refinery energy use 3.7%-6.3% (≈ 0.13-0.22 quads/year) higher and refinery CO(2) emissions 5.4%-9.3% (≈ 0.014-0.024 gigatons/year) higher in the study cases than in the baseline. Refining heavier crude slates would require significant investments in new refinery processing capability, especially coking and hydrotreating units. These findings differ substantially from a recent estimate asserting that processing heavy oil or bitumen blends could increase industry CO(2) emissions by 1.6-3.7 gigatons/year.

Response to Comment by Karras on "Analysis of Refinery Energy Use and CO 2 Emissions in the U.S. Refining Sector, With Projections to 2025"O ur paper 1 dealt with the application of an appropriate analytical technique, refinery LP modeling, to estimate the effects on CO 2 emissions of ambitious, but plausible, increases in the use of Canadian bitumen crudes in U.S. refineries. We neither critiqued nor rejected Karras's model. We mentioned his results in the context of comparing our estimates of increased CO 2 emissions from refining Canadian bitumens with three published estimates.The refinery modeling system we used in our analysis is proprietary for good reason. Developed without public funds, it is the product of thousands of man-hours of our professional effort. It embodies proprietary model design, process representations, data, and correlations, as well as information developed from public sources (some of which is shown in our Supporting Information). The client community vets proprietary refinery LP models because they are used in applications of practical consequence. The vetting considers previously published analyses, the developers' reputation and qualifications, and the endorsement of clients.Neither the editor nor any reviewers found our model's proprietary status an impediment to accepting our paper.Karras's paper 2 states (twice) that "a switch to heavy oil and tars sands could double or triple refinery emissions and add 1.6−3.7 gigatons of CO 2 from [processing] the oil". His paper deals with the U.S. refining sector. His "PD model" was estimated using only U.S. refining data and therefore has dubious applicability to refining operations elsewhere. Hence, we assumed that his estimate of 1.6−3.7 gigatons of increased CO 2 emissions from refining Canadian bitumens applies to U.S. refining (crude run ≈ 15 MM Bbl/day). If his estimate applies to global refining (crude run ≈ 82 MM Bbl/day), it is still 3−4 times higher (per barrel of crude) than our estimate − not within 3−10% of it, as he asserts. Indeed, his estimated increase in ref inery CO 2 emissions is higher than the estimated increases in total life-cycle CO 2 emissions published by the U.S. State Department and CERA (ref 1, Supporting Information Table SI -13).Karras claims that our projections "diverge from observed data as crude density increases." Responding to that requires some background information.Refinery energy use depends on many factors, including ▶ The refined product slate (i.e., the volume shares of products in refinery output) ▶ Refined product specifications (e.g., gasoline sulfur and benzene content) ▶ The crude slate's "hetero-atom" content (especially sulfur) ▶ Other crude slate properties (e.g., distillation curve, density, aromatics content) These factors differ from one U.S. refining region (PADD) to another. For example, PADD 5 (West Coast) refining, dominated by California's refining sector, has the nation's

This paper outlines the nature and scope of a linear-programming model system (LORENDAS) designed for quantitative analysis of world-wide Long-range Energy Developments and Supplies. Basic projections, to the year 1996, of U.S. and European energy resource developments, conversions and imports as indicated by the model solution are reviewed, and the variations of these projections in response to a range of OPEC pricing policies are presented. pricing policies are presented Introduction Until fairly recently, the semblance of resource abundance and freedom of choices made it justifiable to approach individual energy issues separately, on the basis of individual analyses. However, with the rapidly vanishing margin between requirements and reserves practically all the slack may soon be gone from the world energy system. As a result, changes in any one component of the energy system - including constraints imposed by policies - are likely to impinge on everything else. Conversely, whenever one seeks to assess the implications of a given policy or technological alternative concerning energy supplies, he is inevitably faced with a complex set of technical, economic and often political interactions. These interactions stem from the choice among different processes for the development and supply of a given form of energy, from the substitution possibilities among the different energy forms and/or different supply sources and, finally, from the competition between different energy consuming areas and end-use sectors for the supplies of energy. In addition, one has to account for the tight temporal interdependence between the capabilities of energy supply available at any given moment and previous investments in reserves and/or plant capacity, on the one hand, and the decline of the natural resources, on the other. The need for analyzing these interactions between a multitude of variables has motivated the development of LORENDAS, a comprehensive mathematical modeling system, in linear-programming format, of the worldwide Long-Range Energy Development and Supplies. This computer-based system serves essentially to simulate all the different operations and processes concerning the supply of energy, accounting for their technical and geographical interdependence, as well as their evolution through time. At this stage the LORENDAS modeling system is a fully operational tool available to government agencies, universities and industry. The system has been used for a number of preliminary investigations concerning the impacts of taxation on North Sea oil production, the effects of strip mining moratoria on U.S. production, the effects of strip mining moratoria on U.S. energy supplies, the impacts of oil and gas price regulations, and various other sensitivity analyses. In this paper, following a review of the LORENDAS formulation and of the basic model inputs, we present the initial results of a series of studies focusing on the effects of possible OPEC price variations over time on the indigenous energy resource developments in the U.S. and Europe, the volumes of oil imports, and the mix of fuels or energy forms supplied. The term "OPEC" is used as a surrogate for all major oil exporters including Mexico, Canada and other prospective exporters. Thus, the presented results and conclusions deal only with the impacts of possible price paths for world oil, and do not treat the question of how the various exporters arrive at pricing decisions. OUTLINE OF LORENDAS MODELING SYSTEM Overview The LORENDAS modeling system used in the present studies provides an explicit engineering type description, in Linear Programming format, of the complete set of operations and investments in the chain of energy resource development, conversion and distribution.

No abstract