By similar calculations, the 90% confidence interval for consumption of diesel falls between a 10 percent reduction and 75 percent increase from current levels by 2030. The range of possibilities for diesel consumption is shown in Figure 17b. Lastly, VMT increases above current levels in 90% of the draws as shown in Figure 17c. VMT has been far less volatile than gasoline and diesel consumption in California and therefore we see a tighter range of uncertainty around future VMT projections.Each gallon of CaRFG contains reformulated blend stock for oxygenate blending and ethanol. Due to the “blend wall” for ethanol, CaRFG, as well as all reformulated gasoline in the U.S., is often referred to as E10. The average gallon of ethanol earns LCFS credits since the volume-weighted CI rating of ethanol used in the program falls below the standard. Therefore, each gallon of CaRFG consumed in California will generate both LCFS deficits and credits. We calculate total CARBOB consumption as 90 percent of CaRFG, with the remaining 10 percent being ethanol. Therefore, the BAU projection assumes that the E10 blend wall persists through 2030. Pursuant to our definition of BAU, currently observed BBD blend rate in the liquid diesel pool persists through 2030 as well. That is, we assume that 20 percent of liquid diesel fuel used in the transportation sector is BBD. The BAU projection extends the status quo assumption to fuel CI ratings over the compliance period. In the case of the biofuels,cannabis vertical farming recent average volume weighted CI ratings reported to the LCFS are used.
These assumptions are summarized in Table 10. The one area where BAU assumptions differ from the status quo is electric vehicles. We have to make an assumption regarding the penetration of electric vehicles to forecast credit generation from electricity as well as to forecast the level of fossil fuel displacement. EV penetration is difficult to predict and its trends have been evolving. For this reason, we use EMFAC projections of the share of all vehicles that are light-duty electric and heavy-duty electric.EMFAC projects 1.3 million EVs on California roads by 2030. Table 11 shows how EMFAC projections translate into parameters that measure EV penetration in California now and in 2030; we follow the EMFAC rate of penetration in our BAU. Table 11 also shows our BAU assumptions regarding the CI rating of electricity. We use the current grid average CI rating and EERs reported by CARB. Policies in place to reduce the CI rating of the grid through increased use of renewables or accelerate penetration of EVs are not implemented in the BAU but considered in scenarios on the BAU and sensitivity analyses on results.In addition to the assumptions in Table 11, we assume in the BAU that future EVs will replace the average internal combustion engine vehicle on the light duty side but be driven 30 percent fewer miles, again taking the BAU stance of extending current conditions to 2030 . We have no information on how VMT for the heavy-duty sector may change with increasing EV penetration.
While vehicles deployed may be well used, as currently the case, fleets and loads may also shift in unexpected ways. For this exercise, for simplicity, we apply the “30 percent fewer miles” assumption also to heavy-duty EVs and assess the displaced petroleum fuel all from the gasoline pool. With these assumptions, we can project a quantity of kWhs of electricity that will be charged and project the resulting number of LCFS credits associated with the demand simulations. Since EVs are assumed to replace average fuel economy ICEVS, gasoline demand declines according to the share of EVs in the vehicle pool.Using the parameters from Table 10 and Table 11, we translate the forecasts of fuel demand, after accounting for gasoline displacement from EVs, into forecasts of the deficit/credit balance over the compliance period subject to BAU conditions. Using the predictions of CaRFG and diesel demand, we calculate CARBOB and CARB diesel deficits in each state of the world represented by our simulations. The distributions of deficits from each fuel are plotted in Figure 18. Figure 18a shows a distribution of CARBOB deficits centered around 290 MMT on average, or approximately 26 MMT per year. For context, the average is approximately 150 percent larger than the 10.3 MMT generated in 2018. The increase in deficits reflects the BAU demand projections as well as the increasing stringency of the standard to 2030: a gallon of fuel of a given CI rating generates more deficits in later years because a higher percentage CI reduction is required to meet the standard. The total demand for LCFS credits plotted in Figure 19 is the sum of CARBOB and CARB diesel deficits, less the bank of system-wide credits accumulated since the beginning of the LCFS, reflecting CI rating reductions beyond required annual levels; the bank currently holds approximately 8.5 million metric tons of credits.
This distribution characterizes the number of LCFS credits that would need to be supplied to the market to cover aggregate deficits expected to be generated under BAU conditions for the period 2019-2030. Note that our approach is high-level, examining aggregate net deficits for the compliance period, and abstracting away from annual compliance decisions and situations that could impact Until this point, we have described BAU forecasts for LCFS deficits and for credit generation from BBD, ethanol, and on-road electricity. However, there are other pathways to credit generation that must be considered before estimating a credit/deficit balance. As shown in Figure 20, BBD, ethanol, and on-road electricity make up 90 percent of the credits that were generated in 2018. For the remaining pathways, we simply that credit generation under BAU remains constant at 2018 levels. We will consider alternative credit-generating assumptions regarding these other pathways in the next section. The other pathways include renewable natural gas – including from landfills and dairy, off-road electricity, projects such as carbon capture and sequestration and innovative crude production, alternative jet fuel, and hydrogen. We use measures of different scenarios laid out by CARB in their illustrative compliance scenario calculator to quantify these credits.We can now combine projected distributions of deficits and credits, some of the latter tied to the demand scenarios through blend levels and vehicle penetration rates, and others held constant at levels proscribed by CARB in its LCFS ICS, to illustrate the future compliance outlook through 2030 under BAU uncertainty. We present this in Figure 21 by taking slices of the distribution according to the percentile of net deficits remaining after BAU assumptions are applied. That is, we identify the percentile of each simulation according to the level of net deficits in that simulation and plot credits by each pathway in those simulations. Under the BAU, Figure 21 shows the scope of under-compliance in the LCFS. The under-compliance result under BAU assumptions is not a surprise, since LCFS targets were chosen to mandate substantial change in California’s fuels mix, and the BAU freezes several key elements. It does, however, show magnitude of change required for compliance if past trends in fuel consumption and the state economy. On average across simulations, deficits are 163.MMT greater than credits over the entire compliance period. Figure 19 shows that the average number of deficits is about 360 MMT, indicating that our credit supply assumptions under a BAU would cover less than half the compliance requirements for the period 2019 to 2030.The BAU case depicted in Figure 21 allows infrastructure credits to be generated at their maximum potential rate,drying cannabis which is 5% of the previous year’s CARBOB deficits. It is readily apparent that infrastructure credits are much too small a source to make a meaningful difference in the net deficits, and that the burden of compliance will fall on other sources of credit generation.Our projection of deficits under BAU uncertainty provides a range of possibilities of demand for LCFS credits. Next, we present compliance scenarios in which we overlay a range of possibilities for LCFS credit supply. We begin with a baseline scenario and then consider adjustments to each of the baseline assumptions. Throughout, we make the assumption that biomass-based diesel will be the marginal fuel for compliance under the LCFS. This is the most likely case given past trends, and due to policy and capacity constraints inherent with other regulated pathways. Of current credit generators, the constraint from the ethanol blend wall is notable. Blends of ethanol up to E85 require a specialized vehicle not being prioritized for sales. E15, while allowable nationally, must go through an additional approval process for use within state.Massive growth in newer technologies such as hydrogen, natural gas, or electric vehicles would require those technologies to be lower cost than the already mature renewable diesel. This may be possible, or additional credit generating opportunities may be opened up by regulatory amendments as in the past , but these situations are too unknown or uncertain to be included here.
In the previous section, we presented a distribution of credit shortages assuming that BAU trends continue on both the demand and supply side. In this section we relax assumptions on the supply side and answer the question of how much BBD would be necessary to reach annual compliance under the LCFS. We take this approach to evaluating the difficulty of compliance because BBD is the marginal fuel for compliance. Therefore, we consider different assumptions regarding credit generation and assume the resulting net deficits must be satisfied by BBD credits. We assume a smooth draw down of the existing credit bank going into the study period and require annual compliance through use of additional fuels, neither of which is imposed by the regulation. Our analysis is meant to illustrate difficulty of compliance.Certain elements of credit supply are tied to demand, whereas we assume others are independent of demand. We calculate the factors that depend on demand from output of the VEC model and the simulations. Ethanol volumes in each simulation, for example, are equal to 10 percent of gasoline demand so we calculate the volume of ethanol for each draw of the simulations. For the factors that are separate from demand we run our simulation using different policy and supply scenarios to understand their impact. To characterize the relative influence of different assumptions, we evaluate each scenario against a baseline. In the baseline scenario, we assume all CI ratings remain at 2018 levels, infrastructure credits are maximized, and the other credit generating categories achieve the minimum values in the ICS. In Table 12 we summarize each scenario and its assumptions, relative to the BAU assumptions in the previous section and the baseline compliance scenario. In all scenarios, we assume that infrastructure credits are at the maximum allowable level of 5% of the previous year’s CARBOB deficits.The other credits we use from ARB’s ICS are independent of our model of demand for LCFS credits. They are developed within the ARB modeling system, based on demand scenarios, and policy and credit pricing assumptions out to 2030. To illustrate the magnitude in which these sources could affect BBD demand and LCFS compliance, we consider a scenario in which the maximum of each source across scenarios is realized. Specifically, we take the maximum number of credits across the ICS scenarios in each year for each pathway. This set of assumptions is A1 in Table 12. This characterizes a scenario with a higher credit profile for renewable natural gas and projects. We consider a scenario in which the number of EVs rises sharply over the compliance period. In 2018, California Governor Jerry Brown announced a $2.5 billion plan with the objective of getting 1.5 million zero-emission vehicles on California roads by 2025 and 5 million by 2030.This trajectory would be a stark deviation from any historical trends and would not be captured in our model of BAU fuel demand. Therefore, we consider a scenario in which 1.5 million EVs are on the road by 2025 and 5 million by 2030 at a constant rate. We refer to this set of assumptions as A2 in Table 12.The CI rating of ethanol is also independent of demand. CARB, in the ICS, assume a path for starch, sugar, and cellulosic ethanol in which the volume-weighted average CI rating of ethanol falls to 40 by 2030, a 38.5 percent reduction from the current level.