MISO s Analysis of EPA s Final Clean Power Plan Study Report

Transcription

1 MISO s Analysis of EPA s Final Clean Power Plan Study Report July 2016 MISO Policy & Economic Studies Department i

2 Contents Executive Summary Introduction Near-Term Analysis Sensitivity Analysis Model Design Implementation of the CPP Scenario Analysis Scenario Descriptions Additional Modeling Assumptions Modeling Rate- and Mass-Based Compliance Modeling State, Regional and Interconnect Compliance Mid-Term Analysis CO 2 Reduction Levels Coal Retirement Analysis Implementation of CO 2 Constraints Study Findings Near-Term Sensitivity Analysis Near-Term Scenario Analysis Scope of Compliance Region Compliance Costs Generation Impacts Rate vs. Mass Compliance Additional Sensitivities Mid-Term Analysis Base Datasets Baseline Generation Expansion Baseline Resource Mix Regional Demand and Energy Forecasts Fuel Forecasts Study Areas Capacity Types Firm Interchange Hurdle Rates Planning Reserve Margin Targets Wind Hourly Profile and Capacity Credits MISO ii

3 5.11 Reserve Contribution Financial Variables Load Shapes Alternative Generator Categories Alternative Generator Data Baseline Renewable Portfolio Standards (RPS) MTEP15 Futures Matrix MTEP16 Futures Matrix Mid-Term Analysis Dataset Demand & Energy Gas Price Forecast Retirements and Conversions Demand-Side Management Mitigating Unintended Emission Shifting MTEP Unit Siting Methodology General Siting Rules for EGEAS Generator Developmental Statuses Site Selection Priority Order for EGEAS Unit-Specific Greenfield Siting Rules for EGEAS Greenfield Coal Siting Rules Greenfield Combined-Cycle Siting Rules Greenfield Combustion Turbine Siting Rules Greenfield Nuclear Siting Rules Greenfield Wind Siting Rules Greenfield Photovoltaic Siting Rules Demand Response and Energy Efficiency Siting Rules Greenfield Siting Rules for EGEAS Appendix A: Energy Efficiency from EPA s draft CPP Building Block 4 used in EWS scenario Appendix B: MISO Hub LMPs for All Years and Scenarios Appendix C: Emissions for All Years and Scenarios Appendix D: MISO Fuel Mix for All Years and Scenarios MISO iii

4 Executive Summary 1 Purpose of MISO s analysis The U.S. Environmental Protection Agency s Clean Power Plan (CPP) for regulating carbon dioxide emissions from the electric power sector could affect the industry in a number of significant ways. Advised by input from its stakeholders, the Midcontinent Independent System Operator (MISO) analyzed the CPP in order to provide its member-states and asset owners with independently derived technical data and other objective information they may wish to consider in preparing their CPP implementation plans. By design, MISO s stakeholder-informed analysis also examined how industry trends and drivers other than the CPP such as low natural gas prices and the increasing penetration of renewable generation, among other things are causing the region s resource portfolio to evolve. MISO expects that these non-cpp policy and economic drivers will continue to reshape the electricity industry regardless of whether the CPP survives the legal challenges that were lodged against it by various entities (including some states and asset owners in the MISO region). These non-cpp drivers figure prominently in MISO s analysis, effectively making it a study about the broader drivers of the evolving resource portfolio as opposed to a look at just the CPP rule itself. The observations in this executive summary and the main body of this report are not recommendations for complying with the Clean Power Plan or addressing the non-cpp factors that are contributing to the evolution of the region s resource portfolio. Instead, the observations in this document are only intended to help MISO s stakeholders better understand how the CPP and the non-cpp drivers could impact the MISO system. States, utilities and other entities should consider these observations within the broader context of their CPP compliance objectives, policy goals and views about their desired future resource mixes. Study focus and key observations Near-Term Analysis All of the key observations cited below are derived from the Near- Term phase of MISO s analysis, which focused on assessing the impacts of complying with the CO 2 -reduction targets in the CPP rule itself. This phase was conducted using scenario-based evaluation on static resource mixes which allowed for very detailed observations in some areas, but limited the ability to make observations about what the most cost effective resource build-out for compliance implementation may be. One primary focus of this analysis is to compare and contrast potential impacts associated with two different approaches to CPP compliance: (1) rate-based compliance, and (2) mass-based compliance. 2 The analysis compares these potential rate/mass impacts on a regional MISO-wide level, as well as on a state-bystate basis. On the regional level, using one set of common modeling assumptions, the analysis yields the following key observations about the rate/mass compliance options and other related matters: Regionally, mass-based compliance is less expensive than The U.S. Supreme Court stay and ongoing CPP litigation On Feb. 9, 2016, the U.S. Supreme Court stayed the CPP until the litigation challenging it runs its full course. Because of the stay, some MISO-member states scaled back or stopped working on CPP-related matters. At the time of the stay, MISO had largely finished its CPP analysis, although some of the study s findings had not been released. Since then, MISO has worked closely with its stakeholders to determine how potential CO 2 - reduction initiatives will be reflected in MISO s transmissionplanning efforts going forward. 1 This executive summary does not attempt to recap the complex modeling methodologies that are explained in detail in the main body of this report. Instead, this summary provides a high-level overview of the scope and key observations of MISO s analysis of the Clean Power Plan and the related non-cpp factors that are driving the evolution of the region s resource portfolio. 2 Rate-based compliance entails limiting how much CO2 is emitted per every megawatt-hour of energy produced in a given state or region, while mass-based compliance entails limiting how much total CO 2 is emitted in a state or region over a set amount of time, such as a 1-year period. 4 FINAL VERSION

5 rate-based compliance, with the gap between the two approaches increasing over time, unless the construction of non-co 2 emitting resources can keep pace with the demand for emission rate credits (ERCs) needed by existing fossil-fueled resources to continue compliant operation. Early compliance targets can be met through existing renewable portfolio standards and coal-to-gas redispatch, but comprehensive planning would need to start expeditiously to meet increasingly stringent compliance targets in the mid-2020s. Over time, the coal fleet faces increased risks of decreased generation and operating hours, along with increased cycling under the CPP. A robust buildout of new non-co 2 -emitting resources would be needed to mitigate CO 2 price increases under rate-based compliance. System dispatch faces relatively less change under mass-based compliance, and thus may require less capital investment. Regional, trading-ready compliance approaches yield lower cost compliance than state-by-state compliance options. MISO also analyzed how individual states in the MISO region could be affected by choosing either the ratebased or the mass-based compliance option. From this part of the analysis, MISO makes the following key observations: Generation will likely rise/fall in similar locations under both rate and mass compliance approaches, so transmission expansion, if needed, will likely be similar under both. Mass-based compliance produces a more balanced mix of buyers and sellers within MISO. Most states see a mass-based compliance advantage unless a regional heavy penetration of renewables and energy efficiency is achieved. Because MISO assumed a static resource mix, the ERCs created by renewables and energy efficiency are assumed to be available for use in compliance. If, however, these resources fail to generate the ERCs necessary for cost-effective rate-based compliance, there is a risk that costs will increase far beyond costs of mass-based compliance. Rate-based states may consider the need to mitigate this risk should it come to fruition. Under a patchwork mix of both rate and mass compliance, states with a rate advantage may lose that benefit if other states go mass. Overall, MISO s analysis shows that flexibility in compliance options leads to lower compliance costs. Individual states and regions derive different benefits from pursuing different options, but working together as a region allows each state to share this diversity in a way that benefits the entire region. MISO as a regional system operator and transmission planner provides the flexibility needed to integrate these preferred compliance options while maintaining reliability and keeping costs as low as possible. Mid-Term Analysis By design and with input from stakeholders, MISO also conducted a Mid-Term phase of analysis that looked more broadly at how the region s generation assets could be impacted by various CO 2 -reduction scenarios that are not based on the specific parameters of the CPP. This phase looked at the optimal resource expansion under mass-based compliance to make observations on what generation resources may be built and/or retired under CO 2 reduction strategies. As noted above, this phase of MISO s analysis focused more on the non-cpp policy and economic drivers that are causing the resource portfolio to evolve regardless of the requirements of the CPP. This phase of MISO s analysis looked at two primary things: (1) potential coal retirements in the region, and (2) the potential for building out the region s levels of renewable energy. This phase modeled the study region into the year 2035, with emission reduction targets continuing through the entirety of the study period. 1. Coal retirements: MISO analyzed how much of the region s existing coal-fired generation may likely retire for economic reasons under the CPP rule as written, as well as the two following hypothetical CO 2 -reduction targets: a. The Partial CPP Future : Assumed that the region s power-sector CO 2 emissions would decline by 17% by 2030 compared to a 2005 baseline. This future was designed to model how the region could be impacted if states and generators began to comply with the CPP, but full compliance was 5 FINAL VERSION

6 slowed or halted due to legal or political challenges to the rule (Note: This future, like all of the others in MISO s analysis, was developed before the U.S. Supreme Court stayed the CPP). b. The Accelerated CPP Future : Assumed that the region s power-sector CO 2 emissions would decline by 43% by 2030 compared to a 2005 baseline, driven by low natural gas prices and decisions by states and generators to aggressively build out renewables and demand-side resources. The 43% figure was based on more aggressive CO 2 -reduction plans, such as the midterm compliance point of the Waxman-Markey climate change bill that the U.S. House of Representatives passed in This future also helped illustrate how things may change if the EPA tightened the CPP targets during a rule review at some future date. Using these scenarios and MISO s practice of dispatching resources in order of economic merit, MISO modeled how much or how little each of the region s existing coal units would run to help the MISO system as a whole to meet the different CO 2 -reduction targets. MISO s key observations regarding coal retirements: MISO s analysis indicates that retiring coal units from service could cause total system costs 3 to decline at least until a certain point for each future studied. For example, total system costs for the Partial CPP Future reach their lowest range when 8-11 GW of coal is retired, climbing to GW for the Accelerated CPP Future. Under the CPP rule itself, MISO s analysis indicates that total system costs would reach their lowest point with GW of coal retirements. Given the similar costs within this range, 16 GW of coal retirements is seen as a likely outcome. Notably, MISO s analysis also indicates that total system costs in all three scenarios start to climb again when a certain level of coal generation retirements is reached. That point is reached when the costs of building and operating new gas-fired and renewable resources start to exceed the costs of continuing to operate existing coal units. 2. Build-out of renewables: MISO s analysis indicates that a near-equal mix of wind and combined cycle plants would likely replace coal units as they retire under the CPP. When greater CO 2 reductions are examined, the proportion of wind (compared to combined cycle) replacing coal increases, and solar resources become more viable. This renewable generation is in addition to what would be built without a national CO 2 -reduction policy, and leads to the need to understand where this additional renewable generation would likely be sited and constructed. Driven by the projected increase in renewable generation, a separate effort was undertaken to analyze where these new wind and solar resources would likely be sited. This separate analysis looked at levels that are more stringent than those of the CPP, in part because MISO wanted to understand the upper bounds of building out viable levels of wind and solar in the region. MISO s key observations regarding build-out of renewables: MISO s analysis indicates that much of the additional wind would likely first be built in areas that are currently experiencing high levels of wind buildout such as Iowa, Minnesota and Michigan. Significantly larger amounts of new wind and solar capacity would need to be built if far more aggressive CO 2 reductions of 50% or 80% were to be pursued in the region. The analysis indicates that the optimal locations for building new wind power to achieve significant CO 2 reduction would be concentrated in eastern Montana, the Dakotas and the Great Lakes region, given the higher wind potential in those areas. Notably, these areas are not particularly close to MISO s biggest load centers. But if the objective was to aggressively reduce CO 2 emissions in the region, MISO s analysis indicates that it would still be cost-effective to site new wind power in these relatively remote areas and build new transmission to deliver the energy to the rest of the MISO footprint. 4 Conclusion MISO s analysis of the Clean Power Plan, along with other studies, indicates that the future will bring significant change to the power sector. The CPP accelerates this change by driving increased levels of renewable and energy efficiency deployment, and by pushing up the retirement timelines for coal assets. This study looks at a range of compliance options and impacts to the generation and transmission assets 3 As used here, the term total system costs includes the following: (1) generation production costs, (2) generation capital costs and (3) generation annual fixed operations and maintenance (O&M) costs. It does NOT include sunk costs of retired coal units nor electric transmission or natural gas transportation costs. 4 For the full report on MISO s study on the build-out of renewables, see MTEP17 Futures Development Workshop Vibrant Clean Energy Report and Documentation. 6 FINAL VERSION

7 within the MISO footprint. Compliance costs are found to vary greatly with the price of natural gas along with the economic and technical potential of both renewable and energy efficiency deployment throughout the study period. Going forward, analysis (including interregional analysis) is required to assess the transmission and natural gas infrastructure needs associated with this industry-wide shift. Future analysis will also require more consideration of energy efficiency as a compliance mechanism, as it can prove a viable means of ERCcreation under rate-based compliance. The results of this study will be used to inform our strategic transmission assessment starting in the fall of It is crucial that planning efforts continue given the long lead time needed to plan, approve and build the infrastructure necessary to enable the cost-effective and reliable evolution of the electric system. 7 FINAL VERSION

8 1 Introduction This document presents an overview of the study assumptions for and results of MISO s Near-Term and Mid-Term Clean Power Plan (CPP) Analyses. For background on the origins of and drivers for MISO s CO 2 study efforts, as well as a discussion of findings from MISO s 2014 and 2015 CO 2 analyses, see the Phase I/II 5 and Phase III 6 study reports. The Near-Term and Mid-Term analyses are two of three stages of study MISO is undertaking to better understand the impacts of the EPA s Final CPP. The overarching goals of these efforts are: - To inform policymakers as they formulate compliance strategies - To enable the reliable, efficient implementation of CPP-related policy decisions made by MISO member-states and asset-owners - To provide insight into the MTEP17 futures development process In particular, the Near-Term analysis of the CPP is designed to: 1) Investigate CPP compliance under a range of sensitivities around economic, technological and policy conditions 2) Provide insight into various compliance pathways and paradigms given six static resource mix scenarios (with varied levels of coal retirements, gas build-out, renewables build-out and energy efficiency program implementation) 3) Examine variation of CO 2 rate targets (sub-category, average) and CO 2 mass targets (with and without the new source complement 7 ) at different scopes (state-level, pool-level and Eastern Interconnect-wide CO 2 constraints) The Mid-Term analysis of the CPP is designed to: 1) Investigate reasonable amounts of coal retirements for three levels of CO 2 reduction (partially meeting compliance, meeting compliance and exceeding compliance) 2) Provide insight on how to model carbon constraints in MTEP17 futures 3) Get a preliminary understanding of how renewables could be selected in meeting carbon constraints Together, the Near-Term and Mid-Term analyses inform MISO s multi-year effort to identify, quantify and interpret grid reliability and economic ramifications of the CPP. These efforts create a bridge between the uncertainty and complexity that exists due to the final CPP and the modeling certainty needed for effective transmission overlay design. 5 See Phase I and II Analyses of the Draft Clean Power Plan 6 See Phase III Analysis of the Draft Clean Power Plan 7 The new source complement provides additional allowances to mass-based states for emissions from new sources associated with satisfying incremental demand. 8 FINAL VERSION

9 2 Near-Term Analysis This report begins by describing MISO s Near-Term analysis. The Near-Term analysis was designed to provide a robust and reasonably realistic representation of a world in which the EPA s final version of the CPP is fully implemented. Scenarios were derived and modeled to capture a range of compliance strategies without leaving the bounds of the rule. Implementation of the rule in the models took the form of capacity additions and retirements coupled with constraints representing the EPA s CO 2 rate and mass targets, applied at state, pool 8 and interconnection levels. The Near-Term study included two analytical methods and two models: - Resource forecasting - using the Electric Generation Expansion Analysis System (EGEAS), which optimizes the type, timing and amount of resource additions over a 20-year period, given user-defined study footprint, generation fleet (initial fleet), load profile and constraints (including Planning Reserve Margin) 9 - Production cost modeling - using PLEXOS 10, which produces an optimal, chronological, hourly dispatch for a single year given user-defined study footprint, transmission infrastructure, generation fleet (static across the year modeled), load profile and constraints (including CO 2 rate or mass limits) These methods and models were selected given study needs and prior modeling experience using EGEAS and PLEXOS. More specifically, PLEXOS was selected as the production cost model of choice in part because it can model a dynamic CO 2 rate constraint. In order to capture the early, mid- and late-stage impacts of CPP compliance, the years 2022, 2025 and 2030 were modeled 11 in PLEXOS. In EGEAS, a 20-year period was modeled, beginning in To the extent possible, similar representations of the underlying infrastructure and similar regulatory and economic assumptions were applied in both models. Reliability assessment was not included in the scope of the Near-Term analysis given the focus on investigating compliance pathways for various resource mixes. It will be incorporated into future analyses, once there is a better understanding of localized resource impacts due to the CPP. 2.1 Sensitivity Analysis The first portion of MISO s Near-Term analysis was sensitivity analysis. This involved changing relevant input parameters of the EGEAS model to examine high-level impacts on cost, CO 2 emissions production and resource dispatch. Before doing so, MISO validated the EGEAS CPP models. Each of the three Building Blocks in the final CPP was applied in separate model runs, and then all three were applied simultaneously in a single run. The base MTEP15 economic planning dataset underlay the sensitivity analysis; specific parameters were altered as described below. EGEAS modeled the majority the MISO region as a whole and provided specific insights into how specific variables impact compliance and costs in the MISO region. 8 Pools include MISO, MRO, SPP, PJM, TVA, SERC and NYISO. 9 See Appendix A for details on EGEAS 10 See Appendix B for details on PLEXOS 11 In the final version of the EPA s Clean Power Plan, the first compliance period begins in 2022; the second compliance period, in 2025; and the final compliance period, in FINAL VERSION

10 2.1.1 Model Design The Reference Case in Figure 1 shows emissions for an MTEP15 BAU model run, which does not include the application of any CPP Building Blocks or CO 2 constraints. The first Building Block ( heat rate improvement ) was modeled in 2022 as a 4.3% reduction in heat rate for all coal-fired units at a cost of $100/kW to reflect the capital cost expenditure amortized over 10 years book life. The second Building Block ( Re-dispatch Combined Cycle (CCs) up to 75 % ) was captured in EGEAS by calculating the amount of fuel used by CCs at a 75% capacity factor based on their net summer capacity. The total amount of fuel was then set as a minimum burn constraint for all CC units. The third Building Block ( renewable energy ) was modeled by taking the projected level of renewable energy (RE) generation assumed by the EPA in 2021, applying the five year average increase in capacity through 2023, and then applying the maximum annual capacity increase through Next, the amount of RE generation determined for the entire U.S. was broken down by state based on a state s share of affected source generation, then was split further based on the share of the state in MISO. The RE generation was captured in EGEAS by building out the amount of renewable resources that meet the BB3 level. Figure 1 Validation of CPP EGEAS models using final CPP Building Blocks Each Building Block was applied individually to examine its impact on CO 2 reduction then all Building Blocks were modeled together. One of the outputs of the EGEAS model is CO 2 emissions produced by the operation of the generation fleet. These output CO 2 mass emissions from affected EGUs were compared to the EPA s CO 2 emissions mass target. This process confirmed that the CO 2 reduction levels in EGEAS using the EPA s Building Blocks are comparable to the EPA s CO 2 reduction numbers (Figure 1). After validating the EGEAS CPP models, MISO used them to perform sensitivity analysis. To do so, a select economic, regulatory or technological parameter was varied in the EGEAS model. Hundreds of 10 FINAL VERSION

11 models were executed, each with a different sensitivity change, then the overall trends across these sensitivities were analyzed to understand big picture impacts on CPP compliance. Similar to MISO s analysis of the draft CPP, sensitivities around demand and energy growth rates, natural gas prices, Renewable Portfolio Standards, CO 2 costs, coal retirements and energy efficiency implementation were analyzed (Figure 2). Figure 2 Modeling parameters and sensitivities modeled for the Near-Term CPP analysis Sensitivities in the EGEAS runs helped to inform the scenario analysis that uses PLEXOS. The differences in the tools used to do each analysis as well as the objectives of each analysis do not allow for direct comparison between models. EGEAS analysis modeled capacity expansion; PLEXOS analysis performed production cost analysis. The former modeled a 20-year study period without consideration of electric transmission and the latter, a single year of hourly dispatch with consideration of electric transmission. The generation capacity forecasted in EGEAS (i.e. the RRF units) was modeled in PLEXOS as part of the set of static generation capacity, with which a security-constrained economic dispatch was executed Implementation of the CPP As described above, CPP mass emission targets were not directly modeled in the sensitivity analysis portion of MISO s Near-Term analysis. Sensitivities were altered in the EGEAS model to determine their impact on resulting CO 2 emissions. These emissions were compared to the EPA s CO 2 reduction targets to determine whether a particular set of assumptions results in a compliant system. 2.2 Scenario Analysis The next portion of MISO s Near-Term Analysis was scenario analysis. This analysis used a production cost model, PLEXOS, to capture a range of compliance strategies representing different resource mixes. In PLEXOS, the EPA s CO 2 emissions rate and mass targets were input as constraints on generation and emissions and the model minimized production cost given these constraints. The base MTEP15 economic planning dataset underlay each scenario; specific data elements, detailed below, were modified 11 FINAL VERSION

12 per scenario. PLEXOS modeled the majority of the EI, with the exception of ISO-NE and Florida, but only the results from the MISO region are presented in this report Scenario Descriptions Six scenarios were proposed for MISO s draft rule study, and were used again in MISO s final rule analysis: 1. Business-as-Usual (BAU) 2. CPP Constraints (CPP) 3. Coal-to-Gas Conversions (C2G) 4. Gas Build-Out (GBO) 5. Gas, Wind and Solar Build-Out (GWS) 6. Increased Energy Efficiency with Wind and Solar Build-Out (EWS) Business-as-Usual (BAU) In the first scenario, the system was modeled without CPP constraints. All of the assumptions for the MTEP15 BAU model were used, including Mercury Air Toxics Standards (MATS)-related coal capacity retirements. The figures for MATS retirements assumed in the MTEP15 model were based on MISO s 2011 EPA Impact Analysis 12. Additional assumptions underlying the base dataset for the MTEP15 BAU model are covered in Section 5 of this document CPP Constraints (CPP) For the remainder of the scenarios, the CPP constraints were applied, as defined in Sections and of this document, and the system was dispatched to comply. In the CPP scenario, no generator retirements were modeled beyond those assumed for the MTEP15 BAU model. Similarly, no new electric generation capacity was added to the model beyond that which is added in the MTEP15 BAU model Coal-to-Gas Conversions (C2G) In addition to the 12.6 GW of MATS-related coal unit retirements projected in the MISO footprint, 25% of the remaining coal capacity per region in the Eastern Interconnect was incrementally converted to gasfired combined cycle units. The level of retirements was informed by the results of MISO s Phase II CPP analysis, which modeled sensitivities around coal capacity retirements, including low (0 GW), medium (25% or approximately 14 GW in the MISO footprint) and high (50% or approximately 28 GW in the MISO footprint) levels. The results indicated that a medium level of retirements (beyond MATS-related retirements) was the most cost-effective strategy for compliance with the CPP among those studied. The process for selecting and siting coal units to retire in the C2G scenario is: 1) Select oldest coal units in each region (e.g. MISO, PJM, SPP), amounting to approximately 25% of the coal fleet capacity per region. 2) Convert these coal units to gas units in the model. As conversions, units remain at the same site (same bus) and of the same size (same capacity). Five percent of the coal fleet was converted in the 2022 model, with the remainder of the conversions implemented in the 2025 and 2030 models. The assumption underlying the timing of conversions was that capacity will be removed from the system incrementally (i.e. 5% per year from 2020 through 2024, for a total of 25% in 2024), versus all in one year, due to resource adequacy considerations. The map below shows all of the coal to gas conversions implemented by See 12 FINAL VERSION

13 Figure 3 C2G Coal to CC Conversions Gas Build-Out (GBO) In addition to the 12.6 GW of MATS-related coal unit retirements, 25% of the remaining coal capacity per region in the Eastern Interconnect was incrementally replaced by new gas units, with a balance of approximately 80% gas-fired combined cycle and 20% gas-fired combustion turbines. The timeline for retirement implementation was the same as that for the conversions in the C2G scenario. The process for selecting coal units for retirement and for siting replacement gas units in the GBO scenario is: 1) Select oldest coal units in each region (e.g. MISO, PJM, SPP), amounting to approximately 25% of the coal fleet capacity per region. 2) For MISO: a. Replace retired coal units with 600 MW gas units, per company, as forecasted by EGEAS in the Phase II study. 3) For the rest of the footprint modeled: a. Replace retired coal units with 600 MW gas units, per company (replacement gas-fired capacity is approximately commensurate with coal capacity retirement on a company-bycompany basis, i.e., companies with older coal units will see more gas unit build-out). The sum of the gas replacements is approximately equal to the lost coal capacity. Within the company footprint, site new gas units according to the MTEP siting methodology (see Section 6 for further details). 13 FINAL VERSION

14 Below are two maps detailing the locations of the retired units and capacity expansion units. Figure 4 GBO Retirements (left), GBO Capacity Expansion (right) The difference between C2G and GBO is in the siting and size (capacity, MW) of the replacement generation. The underlying assumption is that these two scenarios will bookend infrastructure expansion needs. The C2G scenario placed new (converted) generators where there is already electric infrastructure but where there may not be the necessary gas infrastructure; the GBO scenario placed new generators in the proximity to gas infrastructure but not necessarily at prime locations within the electric system Gas, Wind and Solar Build-Out (GWS) In addition to the 12.6 GW of MATS-related coal unit retirements projected in the MISO footprint, 17% of the remaining coal capacity per region in the Eastern Interconnect was incrementally converted to wind and solar resources while 13% of the coal capacity was converted to new gas units. The process for selecting coal units to retire in the GWS scenario was based on a retirement screening using PLEXOS to identify retirements based on the intensity of EPA state rate targets and the existing generation of the system. Fifty percent of the retirements were implemented in the 2022 model, 90% in the 2025 model and 100% in the 2030 model. The assumption underlying the timing of retirements was that capacity will begin retirement before implementation of the CPP in an effort to ease compliance in the first year. New capacity was sited according to the MTEP siting methodology (see Section 6 for further details). Below are two maps showing the retirements and capacity expansion units for the GWS scenario. 14 FINAL VERSION

15 Figure 5 GWS Retirements (left), GWS Capacity Expansion (right) Increased Energy Efficiency with Wind and Solar Build-Out (EWS) In the EWS scenario, increased penetration of energy efficiency (EE) was modeled at rates representative of EPA s Building Block number four 13 of the draft rule and was assumed to be a bookend for high-level penetration of energy efficiency. EE rates ramped up from their current levels at 0.2% per year beginning in 2017 until they reached the 1.5% annual target. Once the target was reached, a 1.5% EE level was maintained year-over-year. See the Appendix A of this document for the EPA s assumptions on the potential for avoided energy sales per state. Rate targets for each state were adjusted to give credit for the EE implementation. This was done by adding the MWh of EE to the right hand side of the rate constraint equation (see Section for more details). This scenario also included increased build-out of wind and solar resources, driven by a 15% footprintwide Renewable Portfolio Standard (RPS) target. Current RPS targets accounted for about a 10% target during the same time period. Renewables build out included wind and solar (66% wind and 34% solar) with the exception of SERC which only included solar. Additionally, 25% of SERC s renewables requirement was distributed across SPP, and 25% of it was distributed across MISO. The renewables were sited according to the MTEP siting methodology (see Section 6 for further details), and installed in even increments in 2022, 2025, and Below is a map detailing the siting of the solar and wind units across the study footprint. 13 See Technical Support Document: Goal Computation and Data File: Goal Computation - Appendix 1 and 2 at 15 FINAL VERSION

16 Figure 6 EWS Capacity Expansion Additional Modeling Assumptions Typically, MISO economic models assume that large coal units greater than 300 MW are always committed with must run status. Through the results of preliminary analysis and model runs, it was observed that some states could not comply with their CPP targets and experienced unreasonably high CO 2 prices. To mitigate the inability of many states to comply, these must run statuses were relaxed for this study Modeling Rate- and Mass-Based Compliance In the final rule, the EPA calculated sub-category specific emission rates that apply separately to fossil steam-fired units and combined cycle units. For each sub-category, compliance with each target is measured over three interim compliance periods ( , , ) and a final compliance period (2030 onward). The EPA also created statewide average rate goals for states wishing to use an average rate approach instead. In addition to these two options for rate compliance targets, the EPA also included detailed conversions to mass CO 2 emission targets. Should a state choose to implement mass-based compliance, the state must treat the issue of leakage, which is defined as the shift of generation from existing generators covered by the CPP to new generating units that are not covered by the final rule. One option to mitigate this occurrence, according to the EPA, is for states to cover emissions from new units under their suggested new source complement. The new source complement provides additional allowances to mass-based states for emissions from new sources. Thus, states can choose to use sub-category rate compliance, average rate compliance, mass-based compliance, or mass-based compliance with the new source complement. All four of these compliance implementations were modeled as part of MISO s study. Additional compliance options are available but are outside the scope of this study. 16 FINAL VERSION

17 Modeling Rate- and Mass-Based Compliance The calculations used to construct the average rate goals and sub-category rate goals within the production cost model are shown in the equations below. EGU Emissions EPA Avg Rate Target EGU Generation + New RE Generation Equation 1 Representation of the average CO 2 rate constraint for a given compliance footprint FS EGU Emissions EPA FS Subcat Rate Target FS EGU Gen + CC: ERCS Gen + CC: GS ERCS Gen + New RE Gen + EE ERCs Equation 2 Representation of the fossil steam CO 2 rate constraint for a given compliance footprint CC EGU Emissions EPA CC Subcat Rate Target CC EGU Gen + New RE Gen + EE ERCs Equation 3 Representation of the combined cycle CO 2 rate constraint for a given compliance footprint The terms used in the equations above are defined here. FS EGU Gen/Emissions = Generation/emissions from affected fossil steam electric generating units (in operation or had commenced construction as of 1/8/2014). Generation/emissions from new fossil steam units are not included in rate equations. CC EGU Gen/Emissions = Generation/emissions from affected combined cycle electric generating units (in operation or had commenced construction as of 1/8/2014). Generation/emissions from new combined cycle units are not included in rate equations. EGU Gen/Emissions = Generation/emissions from affected fossil steam EGUs and combined cycle EGUs (in operation or had commenced construction as of 1/8/2014). CC EGU emissions CC Sub category rate target CC EGU generation CC:ERC Gen = ( CC Sub category rate target ) CC EGU generation Per the EPA s final rule, CCs are eligible to earn ERCs if they are generating at an emission rate lower than their target. If this occurs, this term is positive and represents the number of ERCs the CC is producing. If CCs are instead consuming ERCs, this term is negative in the constraint calculation. CC:GS ERC Gen = ( FS Sub category rate target CC EGU emissions CC EGU generation FS Sub category rate target ) IGF CC EGU generation The first piece of this calculation compares a CC s actual emission rate to the fossil steam rate target. The IGF is the incremental generation factor defined by the EPA in the final rule and is based on the incremental NGCC generation needed to reach 75% NGCC regional capacity, consistent with the EPA s second building block. In general, GS-ERCs are intended to incentivize CCs to run at a higher capacity factor and offset higher emitting fossil steam units. 17 FINAL VERSION

18 New RE Gen = ERCs produced from every MWh generated in 2022 and beyond by zero-co 2 emitting renewable units installed after 1/1/2013 *EE ERCs = ERCs earned by MWh of installed energy efficiency. The EPA specifies what circumstances are required for EE to earn ERCs in the final rule, but for the purposes of this study MISO assumes all EE installed in the EWS scenario 14 generates ERCs. This term is only relevant for the EWS scenario. Modeling of mass-based compliance need only directly concern emissions. The equations for the two types of mass-based compliance are shown below. EGU Emissions EPA Mass Emission Target Equation 4 Representation of the mass emission constraint for a given compliance footprint EGU Emissions + New source emissions EPA Mass Emissions Target + New Source Complement Equation 5 Representation of the mass plus new source complement emission constraint for a given compliance footprint Modeling a Mix of Rate and Mass Compliance In addition to modeling all states under rate-based compliance and all states under mass-based compliance, MISO also modeled states under a mix of rate and mass compliance. To build these mixed models, the relative difference between production-based compliance costs when all states choose sub-category rate compliance and production-based compliance costs when all states choose mass compliance was calculated. Using these differences, the first mixed rate/mass run was created based on which compliance implementation appeared less expensive for each state. These differences were sorted numerically. Using this list, half of the states modeled were assigned rate-based compliance, and half of the states modeled were assigned mass-based compliance, meaning that some states that have slightly lower costs under mass-based compliance were assigned rate by design to create a balance. Next, the relative difference between the costs for the state s choice in the first mixed rate/mass run and its second choice from step 1 was calculated. This determined whether each state s mass/rate choice was still less expensive in the mixed rate/mass run was each state better off implementing the lowercost mechanism they saw when all states implemented the same compliance mechanism? Or would a state have benefited from switching to its second option because of the actions of others? The answers to these questions helped form the second mixed rate/mass run, wherein states that would have benefited from this switch do so in the model. 14 All scenarios are described in Section of this report. 18 FINAL VERSION

19 Figure 7 Creation of mix rate/mass model flow chart Modeling State, Regional and Interconnect Compliance In the final CPP, the EPA offers states the option of constructing a CPP compliance plan that would allow them to either comply alone, or as a part of a larger group of states. As part of MISO s Near-Term analysis, the study footprint was modeled under several geographic footprints to represent this option. PLEXOS was used to model state-level compliance, regional compliance and interconnection-wide compliance. In all cases, to represent the real-world boundaries of different ISO/RTO markets in the model, a hurdle rate (essentially a cost to bias the dispatch of a unit within a pool to that pool s market) was applied. A single security-constrained economic dispatch or SCED was simulated for the Eastern Interconnect. The SCED simultaneously optimized the dispatch, took into account the hurdle rates, and thus, the pool boundaries, and tried to comply with the CO 2 constraints added to the model. This approach modeling hurdle rates to represent pool boundaries, along with the application of constraints in the SCED is common to production cost modeling Modeling State-Level Compliance To represent state-level compliance in PLEXOS, the constraint equation was formulated per the EPA s proposed rate calculations for each state or sub-category. For average rate compliance, affected generators that are physically located within a state must meet the EPA s proposed CO 2 average annual rate target for that state. For sub-category state-level rate compliance, affected fossil steam EGUs physically within a state must satisfy the fossil steam sub-category rate target, and likewise for affected combined cycle units. These affected EGUs are able to use ERCs generated by units physically located within the same state for adjusting their emission rates. Rate constraints on emissions can be represented directly in PLEXOS, as opposed to modeling a CO 2 cost as a proxy for a rate constraint. Furthermore, emission rate constraints at the state-level can be accounted for in the solution algorithm at the same time as dispatch occurs at the market pool level. Mass-based state-level compliance was implemented into PLEXOS as a constraint on CO 2 emissions for each state such that the total sum of CO 2 emissions from affected EGUs within each state must be equal 19 FINAL VERSION

20 to or less than the annual mass emission targets issued by the EPA. For mass-based compliance with the new source complement, the constraint was placed on CO 2 emissions from affected EGUs in addition to CO 2 emissions from new EGUs. State-level compliance was only modeled for the CPP scenario Modeling Regional Compliance Regional (or pool-level) compliance was also modeled for the CPP scenario. This included MISO along with these neighboring external regions: MRO, SPP, PJM, TVA, SERC and NYISO. Modeling regional compliance was meant to simulate region-wide markets for trading ERCs/allowances under different compliance regimes. To model average rate regional compliance, the average rate was calculated per the 2012 affected generation-weighted average of sub-category rates using the total pool-level affected generation; in other words, the combined cycle sub-category rate was multiplied by the % of its contribution to total affected generation in 2012 and likewise for fossil steam. Generation within each pool was dispatched to meet this average target. To model sub-category rate regional compliance, rate constraints were constructed in PLEXOS in the same manner as state-level compliance, but using pool level generation and emissions. Affected EGUs were able to use ERCs generated by units located within the same pool. Mass-based pool-level compliance targets were calculated as the sum of the mass emission targets for each state when a state is located fully within one pool. For a state located in more than one pool, the EPA-calculated mass emission targets were distributed to the relevant pools based on the pro rata share of that state s generation located in a given pool. Mass-based with new source complement compliance targets on the regional level were calculated in the same manner Modeling Interconnection Compliance CPP compliance was also modeled at the interconnection level. All units in MISO s model of the Eastern Interconnect were given one common rate or mass target to simulate an interconnection-wide trading market for ERCs/allowances. Average rate, sub-category rate, mass and mass with new source complement targets were calculated and constructed in the same manner as those for regional compliance, but using interconnection-level generation and emissions values instead of regional values. These compliance methods were modeled for the CPP scenario. A high penetration of new gas units in the states surrounding MISO could cause a larger increase in imports than desired or expected under mass-based compliance. To control for this, MISO assigned the new source complement to the states surrounding MISO that are in PJM and SPP but not MISO. This limited the level of imports from new gas units while not being overly prescriptive as to what method of compliance MISO states might choose. Thus, this model included a mix of mass states and mass with new source complement states. Using this mixed mass/mass with new source complement model and the sub-category rate model, MISO studied the effects of a changing fleet using the C2G, GBO, GWS and EWS scenarios under interconnection-wide compliance. 20 FINAL VERSION

21 3 Mid-Term Analysis In contrast to the Near-Term analysis, the Mid-Term analysis shifted away from studying the specific parameters of the CPP and focused instead on understanding impacts on generation and transmission under several alternative CO 2 emissions reduction scenarios. In MISO s draft rule analysis and the Near- Term analysis of the final rule, it was shown that the level of coal retirements in a region could greatly affect the way in which that region could comply with the CPP. As such, in the Mid-Term analysis, CO 2 reduction levels were analyzed under varying levels of coal retirements using EGEAS. The base MTEP16 economic planning dataset underlay this analysis; specific parameters were altered as described below. EGEAS modeled the MISO region as a whole and provided specific insights into how specific variables impact compliance and costs in the MISO region. 3.1 CO 2 Reduction Levels Three levels of CO 2 mass-based emission reduction were modeled in MISO s Mid-Term analysis. The first was a scenario where the EPA s mass emission targets for affected units were enforced throughout the compliance period (Final CPP). Next, the 2022 reduction targets were applied across all compliance years to represent the effect of states taking initial action towards meeting CPP compliance, but a legal challenge prevents the rule from being put in place (Partial CPP). For both Final CPP and Partial CPP, the 2030 mass emission targets were held constant past The last scenario represented an accelerated economic maturity of renewables and demand-side resources driven by technological advancements and public policy, along with sustained competitive gas prices. Here, CO 2 emission reduction targets were far exceeded, and continued to reduce past 2030 (Accelerated CPP). The emission limits for each scenario are shown in Figure 8. Figure 8 Mid-Term analysis CO2 constraints for affected EGUs Partial CPP modeled a 17% reduction in CO 2 emissions from 2005 levels, the Final CPP modeled a 34% from 2005 levels, and the Accelerated CPP modeled a 43% reduction from 2005 levels. 21 FINAL VERSION

22 3.2 Coal Retirement Analysis As mentioned above, one of the major factors assessed in the Mid-Term analysis was how much coal capacity may economically retire under these carbon constraint scenarios. To prioritize which units to retire, MISO employed an approach similar to the 2011 MATS retirement study. This approach examined an EGEAS model that aligned with the CO 2 limits from the final CPP, without implementing any additional coal retirements. The cumulative net present value of net margin of each coal unit was examined and units were sorted in order from least to most economic. Retirements were then applied accordingly in approximately 1 GW increments. The net margin was determined by deducting the fixed and variable O&M costs along with fuel costs from the revenues of each unit for each year of the study period, then discounted back to the base year dollar value and summed together. The revenues were determined by establishing a dispatch cost for the marginal unit that gets awarded to all dispatched units. Because EGEAS is a transmission-less model, this can be considered a generator-only LMP. Figure 9 Coal capacity retired by cumulative net profitability margin Because there are many other factors to consider when retiring units, and EGEAS excludes transmission limitations and costs, this approach was used to identify quantities of coal retirements, not specific units that could retire. The 1-GW swaths of units identified in Figure 9 were derived from a run with constraints that align with the CPP targets. The same methodology was applied in the scenario that partially meets CO 2 targets and the scenario that exceeds CO 2 targets. 3.3 Implementation of CO 2 Constraints For the Mid-Term EGEAS modeling, the EGU CO 2 constraints were applied based on percent reductions from the 2005 baseline. Aggregating all states MISO portions of the EPA s mass emissions targets yielded a 2022 starting value for all constraint cases of million tons. Prior to 2022, CO 2 was unconstrained. In the Partial CPP case, this initial ton value was used throughout the study period. For the Final CPP case, emissions were constrained to meet the targets laid out in the CPP ending with FINAL VERSION

23 million tons in The Accelerated CPP case followed the initial trajectory in a straight line through 2033, reaching a target of 256 million tons. This presented a spread of final CO 2 outputs of approximately +/- 85 million tons from the final CPP MISO target. All values were entered into EGEAS and the dispatch of EGUs was not allowed to exceed these targets. 23 FINAL VERSION

24 4 Study Findings This section details the results of MISO s Near- and Mid-Term CPP analysis. Note that the results of MISO s analysis are not recommendations. Instead, they are intended to help policymakers understand impacts of the CPP on the MISO system. As a reminder, the scope of this analysis was developed with stakeholder input before the U.S. Supreme Court stayed the CPP during litigation. MISO respects that some states have scaled back or halted work on CPP-related matters in light of the court s decision. 4.1 Near-Term Sensitivity Analysis The results of the Near-Term sensitivity analysis are illustrated in Figure 10. Each diamond in the graphic represents a single sensitivity. The % reduction on the horizontal axis in the figure represents how far above or below the EPA s CO 2 emissions target a given compliance strategy is, according to the results of the analysis. Zero percent equates to exactly meeting the target, negative equates to failing to meet the target by the % indicated and positive exceeds the target by the % indicated. The target CO 2 level for this analysis was based on the EPA's mass emission targets for affected units (existing fossil steam and combined cycle units). Recall that sensitivity analysis was performed using the EGEAS software. Figure 10 Results of Near-Term CPP analysis in EGEAS For the sensitivity analysis, resource capital costs and production-based compliance costs, shown on the vertical axis of Figure 10, were calculated as the difference between production and supply/demand side resource costs from reference case costs. There were three reference cases that are unique to each gas price assumption, but have no incremental retirements, energy efficiency or RPS levels. Three different reference cases were used here because natural gas price is not a sensitivity that represents a policy decision, but rather it responds to a dynamic set of uncertainties outside the control of utilities and regulators. These costs did not include electric and gas infrastructure costs as EGEAS is a transmission- 24 FINAL VERSION

25 less model. CO 2 costs were used solely as dispatch modifiers and were not included here. Costs were based on a 20-year Net Present Value (NPV) calculation using a 2.5% discount rate. Results indicated that flexibility in compliance strategies allows for lower compliance costs, and that lower compliance costs could result from strategies outside of the EPA s building blocks. Figure 11 Cost of compliance captured in EGEAS under varying natural gas prices Sensitivities under low, medium, and high gas prices are shown in Figure 11. While the highest priced diamonds represent high gas price sensitivities and the lowest priced diamonds represent low gas price sensitivities, the center swath of sensitivities overlap indicating other variables were at play. Low gas prices did not always equate to low cost impacts. Conversely, high gas prices did not always equate to high cost impacts. 25 FINAL VERSION

26 Figure 12 Cost of compliance captured in EGEAS under varying levels of coal retirements Sensitivities are colored based on retirement levels in Figure 12. Although large overlaps between retirement level sensitivities exist, the patterns in Figure 12 showed a general trend coupling increased coal retirements with increased compliance under mass-based compliance. The range showed that some of the base case sensitivities can comply but the bulk did not. On the other hand, if the highest level of coal retirements was realized, there were almost no sensitivities assessed that did not comply. As an example of how retirement levels interact with sensitivities, Table 1 compares the average resource capital and production-based compliance cost under several levels of coal retirements and several natural gas prices. Average Compliance Cost ($B) GH GM GL Base GW GW GW GW Table 1 Average resource capital and production-based compliance cost per coal retirement level and natural gas price These results indicated that at around 14 GW of coal retirements, the model produced minimum compliance costs. Results also indicated that lower gas prices may lead to lower coal retirements. This counter intuitive result is due to the fact that high gas prices were interfering with the need to switch from coal to gas to aid in compliance. 26 FINAL VERSION

27 4.2 Near-Term Scenario Analysis These results should be considered in the context of the following assumptions: All models assume reliability is maintained through the addition of new resources. Models reflect current generation, assumed retirements and resource expansion, including o Units with signed Generation Interconnection Agreements (GIA) o Resources forecasted part of the MTEP15 7-step process to meet planning reserve margins and renewable portfolio standards Additional scenarios look at other possible resource changes beyond current trends with the assumption that the changes would occur regardless of the CPP. Resource expansion was not optimized for CPP compliance as part of this study, but rather was input into each scenario. The benefits of CO 2 allowances under mass-based compliance are assumed to go to load. Generators are counted for compliance in the state in which they are physically located. Unless otherwise noted, results in this section are for MISO states from interconnection-wide compliance with sub-category rate compliance and mixed mass/mass plus new source complement compliance Scope of Compliance Region These results were produced by the production cost analysis portion of MISO s Near-Term CPP study. The EPA s CPP using rate- and mass-based compliance was modeled using the PLEXOS software under three different geographic scopes: state-level, pool-level and interconnect-level. State-level compliance, as MISO has modeled it, would imply that each state would create individual compliance plans that include individual emissions trading markets for each state. This was considered an unlikely occurrence and thus the results from state-level analysis are not included in this report. Pool-level (MISO-wide) compliance and interconnect-level (EI-wide) compliance could be more likely options for CPP compliance. Figure 13 compares the production-based compliance costs 15 between pool level and EI level compliance Pool - Mass EI - Mass Pool - Rate EI - Rate Figure 13 MISO states production-based compliance costs under pool and EI compliance 15 Production-based compliance costs = production costs + emissions costs + interchange costs 27 FINAL VERSION

28 Costs under pool-level compliance and costs under EI-level compliance were comparable. Because of these cost similarities, the results presented in this report are outputs of interconnect-wide compliance that assumes an available liquid carbon market Compliance Costs Compliance costs in the near term analysis were looked at primarily from the perspective of the costs to operate the electric system. Note that this study did not optimize resource expansion for each scenario, but rather used the results of previous study to inform scenario development. Additionally, resource expansion was performed to meet the Planning Reserve Margin on a regional level, and was unchanged between mass and rate compliance. Optimizing resource expansion, and subsequently calculating resource capital costs, was not the intent of this part of the study; however, the costs are included here to provide a view of resource construction costs to get from the Business-As-Usual scenario to each one of the alternative scenarios studied in the Near-Term analysis. The intention of this section is to provide a more comprehensive view of what compliance costs could look like using the assumptions from MTEP15 and MTEP16, but the key takeaways presented in this report are drawn from the assumption that these scenarios were achieved in a non-cpp future. Rate and mass costs for the scenario analysis 16 are presented in Table 2, separated by a /. Resource capital costs are calculated from cost assumptions in MTEP15 and MTEP16, separated by a. Cost for Rate / Mass Compliance Approaches (20-yr NPV in $B) Incremental Resource Capital Costs for MTEP15 MTEP16 (difference from BAU) CPP C2G GBO GWS EWS Production Costs 88 / / / / 8-69 / -53 Emissions Costs 31 / / / / 1 2 / -2 Resource Capital Costs Total System Compliance Cost Range Table 2 Resource capital and production-based compliance costs for MISO states Production and Emission Costs For the scenario analysis 17, production-based compliance costs for each scenario were calculated by summing production costs (operating and maintenance, fuel, changes in interchange costs) and emissions costs. Costs were based on a 20-year Net Present Value (NPV) calculation using a 2.5% inflation rate. Production costs for 2015 to 2021 were calculated from the BAU case for all scenarios. For 2022, 2025 and 2030, production costs were developed from the results of the PLEXOS simulations; interim year costs were interpolated from these same results. For rate-based compliance, emissions were calculated as the difference between the actual emission rate and the emission rate target multiplied by the generation from affected EGUs. For mass-based compliance, emissions were calculated as the difference between the actual tons of CO 2 emissions and the mass emissions target. For both pathways, 16 For a description of the assumptions underlying each scenario, see Section All rate compliance results are produced using EI-wide sub-category rate-based compliance. Average rate compliance across the MISO region or the EI would require significant coordination and planning, and has a low likelihood of occurring. Mass compliance results are from an EI-wide mixed mass/mass plus new source complement CPP implementation (see section for full description). 28 FINAL VERSION

29 the emission costs were calculated as tons of CO 2 emissions multiplied by the shadow price 18 on the emission constraint output from the PLEXOS model. Interchange was valued at the MISO hub LMP and the overage/underage of emissions was valued at the shadow price on the rate/mass constraint. A negative production cost value indicates that the change case production costs are less than those of the BAU. A negative emissions cost value indicates that MISO is a net seller of ERCs/allowances. CPP C2G GBO GWS EWS Production Costs 88 / / / / 8-69 / -53 Emissions Costs 31 / / / / 1 2 / -2 Table 3 Production and emission costs for MISO states Emissions and production costs across the CPP, C2G, GBO and GWS scenarios were higher under ratebased compliance than under mass-based compliance. A heavy penetration of both renewables and EE in the EWS scenario created significantly lower production costs compared to other scenarios. The trend of heavy renewable penetration also lead to a general rate-based compliance advantage in EWS, as the EPA s mass target only provides additional benefits to RE up to a certain level which the EWS scenario surpasses Resource Expansion Costs Resource expansion costs were calculated based on the number of units installed and the year of their installation for each compliance scenario. A range of costs for each resource is given based on MTEP15 and MTEP16 values, detailed in Sections 5.17 and The costs associated with implementation of energy efficiency in the EWS scenario were calculated from the figures issued by the EPA in their supporting documents for the draft rule. CPP C2G GBO GWS EWS Resource Capital Costs Table 4 Resource capital costs for MISO region under rate and mass compliance These costs were significantly higher in the GWS and EWS scenarios because of the high amount of added renewables. Costs for the C2G and GBO were comparable with one another because similar amounts of gas capacity were built in these scenarios. As seen in Table 3, the EWS scenario had lower production costs than all other scenarios, but required high capital cost investments. Thus, on the whole, EWS was comparable to C2G, GBO and GWS Marginal Cost of Compliance (Implied CO 2 prices) This CPP study used both rate and mass constraints on CO 2 emissions as methods to enforce compliance with the CPP. When a physical constraint is used, PLEXOS calculated a shadow price for that constraint as an output to the simulation. The shadow price on the emissions constraint was considered the marginal cost of CO 2 emissions. 18 The shadow price on the emission constraint represents the marginal cost of reducing one additional unit of emissions under optimal dispatch. 29 FINAL VERSION

30 Figure 14 shows the shadow price for each scenario over the three study years for both rate and mass compliance. A zero price indicates that the cost of economic dispatch for the given resource mix would not increase under CPP compliance. A higher price indicates the system has a more difficult time complying. Note that prices are shown in $/short ton for both rate and mass compliance. This is a result of the way PLEXOS treats a rate constraint, and allows for a more direct comparison between rate and mass compliance. Figure 14 MISO CO 2 prices across scenarios The less stringent CO 2 reduction targets in the early years of compliance lead to lower CO 2 prices under both rate- and mass-based compliance. Under rate-based compliance, an early deployment of renewables drove down CO 2 prices (GWS, EWS), but a continued deployment of renewables would have been needed to sustain these lower prices. Coal retirements also drove down CO 2 prices, but did so with more impact under mass-based compliance (C2G, GBO, GWS). This is due to the fact that under massbased compliance, coal retirements affected the overall efficiency of the coal fleet, which was not the case under rate-based compliance. Note that MISO s scenario analysis assumed a fixed resource expansion implemented over time and scenarios were not optimized for rate and mass compliance. Thus, they may not capture the potential for the system to respond to the high prices of ERCs and allowances Generation Impacts Changes in Dispatch Figure 15 represents the generation mix in MISO in 2030 under rate- and mass-based compliance. Interchange and energy storage accounted for the changing amount of total generation in MISO (thus the height of each bar differs). The region maintained compliance, but energy transfer continued to be optimally scheduled. 30 FINAL VERSION

31 Figure Generation in MISO by fuel type across scenarios Under rate-based compliance, the coal fleet was dispatched down dramatically unless there was a high buildout of renewable generation (EWS), a phenomenon that did not occur under mass-based compliance. Increasing energy efficiency (EE) and renewable energy (RE) allowed for increased CO 2 production because of their ability to produce ERCs under rate-based compliance. EE and RE displaced gas generators, as they are typically the marginal unit, thus coal generation increased. In the other category, steam turbine (ST) gas units also increased production under rate-based compliance. ST gas units are regulated under the same emission standard as coal units which allows them to produce ERCs. These ERCs were high-priced when the ERC market was tight, so the dispatch of ST gas units increased (CPP). In the absence of coal retirements, generation from CTs increased under both rate- and massbased compliance (CPP), but more dramatically under rate-based compliance. Because generation dispatch faced relatively less change under mass-based compliance, it may require less capital investment. MISO s analysis also showed that under both rate- and mass-based compliance in the CPP scenario, generation rose and fell in similar locations. This is shown in Figure FINAL VERSION

32 Figure 16 Generation changes in the CPP scenario Although the magnitude and location of the impacts on generation changed with varying capacity expansion scenarios, within each scenario the impacts of rate and mass compliance were similar. 32 FINAL VERSION

33 Figure 17 Generation changes in alternate capacity expansion scenarios Coal & Gas Unit Performance The coal fleet faces increased risks under CPP compliance. The performance of the existing coal fleet was examined under the CPP scenario, which modeled the grid under current resource trends constrained under CPP compliance. 33 FINAL VERSION

34 Figure 18 MISO coal unit capacity factors under rate and mass compliance As can be seen in Figure 18, coal unit capacity factors decreased under both rate and mass compliance. In 2022, only a small change in the coal fleet was needed under rate-based compliance, so existing units ran with higher capacity factors. Mass-based compliance had a slightly more significant impact on coal capacity factors in However, as the reduction targets became increasingly stringent over time, ratebased compliance became more difficult than mass and thus the rate-based coal unit capacity factors were lower than mass-based coal unit capacity factors, barring additional resource changes. Coal unit performance also changed dramatically in the CPP scenario. In 2030, both rate and mass compliance caused an increase coal cycling, ramping, hours offline and units idled compared to the BAU, detailed in Table 5. Because coal units ran less often under rate-based compliance, they cycled and ramped less frequently. As the stringency of compliance increased, coal units moved from dispatching as baseload to intermediate to peaking units. This may impact a coal unit s ability to continue to economically remain online. Definition 2022 Mixed Mass/NSC 2022 Subcategory Rate 2030 Mixed Mass/NSC 2030 Subcategory Rate Cycling* Number of unit starts 58% -29% 71% 55% Ramping* Total MW traveled (ramp up + ramp down) 11% 2% 30% 7% Hours offline* # of hours of zero generation 68% 3% 157% 246% Total MWh* Total generation -10% -2% -36% -68% Units idled # of units offline all year *Percent change from BAU Table 5 Coal unit operating characteristics under CPP compliance In addition to changes in coal unit dispatch, combined cycle gas units also experienced changes in dispatch. One of the EPA s concerns with the CPP deals with the issue of leakage, defined as the shift of generation from existing generators covered by the CPP to new generating units that are not covered by the final rule. The EPA contends that leakage may occur under mass-based compliance and thus must be addressed when this compliance pathway is chosen. 34 FINAL VERSION

35 2030 Model Natural Gas Offtake (Bcf) Figure 19 Gas unit generation and emissions under CPP compliance The issue of leakage was examined as it pertains to new combined cycle units in the CPP scenario. As can be seen in Figure 19, the generation and emissions from new CCs are comparable under rate and mass compliance across all three compliance years studied. Reliance on gas-fired generation was expected to increase significantly under the CPP. Average daily offtake and peak day offtake were calculated for each scenario studied. Figure 20 shows that rate-based compliance generally saw a greater demand for natural gas compared to mass corresponding with the decreased dispatch of coal seen under rate-based compliance in most scenarios BAU CPP Mixed Mass CPP Subcategory Rate C2G Mixed Mass C2G Subcategory Rate GBO Mixed Mass GBO Subcategory GWS Mixed Rate Mass GWS Subcategory Rate EWS Mixed Mass EWS Subcategory Rate Average Day Peak Day Figure 20 Natural gas offtake across the scenarios Gas-fired generation decreased significantly in the EWS scenario. This was due to the large amount of load-reducing energy efficiency and the significant portion of renewable energy which displaced marginal gas units. The effects of an increased reliance on gas-fired generation on the gas pipeline system were not examined as part of this study. 35 FINAL VERSION

36 Curtailment (% of Total Available Energy) Curtailment (% of Total Available Energy) Curtailment of Renewables Because new renewable energy qualifies for ERC production under rate-based compliance, their dispatch was treated as preferential by the model when compared to existing RE. This did not occur under massbased compliance. The chart below shows that system-wide curtailments are not significantly altered by this effect. 14% 12% 10% 8% 6% 4% 2% 0% BAU GWS Mass EWS Mass GWS Rate EWS Rate New RE Existing RE Figure 21 Interconnection-wide curtailments However, the curtailment of individual units could be drastically increased if new RE is sited near existing RE in an area with the potential for congestion. The chart below shows the effects of adding a new RE unit (New_GWS_RE:1) to an area with two existing RE units (Existing_RE:1 and Existing_RE:2) in the GWS scenario. 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% BAU GWS Mass EWS Mass GWS Rate EWS Rate Existing_RE:1 Existing_RE:2 New_GWS_RE:1 Figure 22 Sample area curtailments 36 FINAL VERSION

37 Because no new units were added to this location in the EWS scenario, curtailment of these existing units in EWS remained relatively unchanged from the BAU. The discrepancy in curtailment between new and existing RE under CPP compliance only occurred in the model when a new unit was sited near an existing unit, particularly under rate-based compliance, as seen in GWS Rate. In both GWS Mass and GWS Rate, curtailment of Existing_RE:2 was less than that of Existing_RE:1 because it was interconnected on the 69kV system, closer to load. New RE units were also occasionally dispatched during times of negative local prices in rate-based compliance, when the value of ERC creation outweighs the production cost penalty. This effect did not occur for existing RE, or for any units under mass-based compliance Rate vs. Mass Compliance Comparison of Compliance Costs In addition to examining the generation impacts of rate and mass compliance, the two pathways were also considered from a compliance-based production cost perspective. Each scenario studied represented a unique capacity buildout that was a specific capital investment decision. The model assumed that these capital investments would be made due to economic and/or policy drivers other than the CPP. Capital costs were not included in these cost comparisons. Compliance-based production costs for the scenarios studied are compared in Figure 23. Figure 23 Rate and mass compliance cost comparisons *Michigan includes the modeling of Fermi 3. In 2022, the relatively low stringency of the reduction targets under both rate and mass compliance showed that most MISO states do not have a strong advantage for either compliance pathway. As targets became more stringent, most MISO states saw lower production costs under mass-based compliance in most scenarios studied. In the CPP scenario, a state that had a heavy build-out of non-co 2 emitting resources could see a rate advantage because its build-out pattern is unique and it has a large market of 37 FINAL VERSION

38 ERC buyers. However, in the EWS scenario, the model included a region-wide heavy build-out of renewables and EE, leading to a rate-based compliance advantage for many states. Thus, the massbased compliance advantage existed for most states unless ERC-generating resources were built out in sufficient quantities in the absence of the CPP Emissions Trading The CPP allows for the opportunity to trade compliance currencies between states. This ability to trade was modeled for each compliance year under each compliance scenario. These models assumed that all states choose rate-based compliance or all states choose mass-based compliance and trading is available across the entire EI. Figure ERC/allowance trading potential In Figure 24, each MISO state has two grey-colored vertical bars in its column: 1. The grey bars affixed with the diamonds represent the range of emission rate credits (ERCs) a given state would need to buy/sell across rate-based compliance scenarios. 2. The grey bars affixed with the circles represent the range of emissions allowances a given state would need to buy/sell across mass-based compliance scenarios. Under a mass-based compliance regime, analysis showed that seven MISO-member states were predominantly buyers of mass-based emissions allowances, five states were predominantly sellers, and the remaining three states straddled the buyer/seller line. Conversely, analysis showed that a rate-based regime produced a less-balanced mix of ERC buyers and sellers. Ten MISO-member states needed to 38 FINAL VERSION

39 buy ERCs to achieve CPP compliance, with two other states straddling the buyer/seller line. Meanwhile, only three states Michigan 19, Louisiana and Illinois were clear sellers of surplus ERCs. From an emissions trading perspective, the chief advantage of a balanced mix of buyers and sellers under mass-based compliance is that it may result in a more liquid market for allowances compared to the less liquid market that may be available for rate-based ERCs. Under rate-based compliance, MISO states may be net importers of ERCs from other markets, whereas under mass-based compliance, the MISO footprint produced a large enough market for emissions trading, for the given set of scenarios Additional Sensitivities Patchwork Compliance Production-based compliance costs were examined for MISO states under a mix of rate/mass compliance between states (see Section for methodology). The differences between rate and mass costs under these circumstances are shown in Figure 25. Figure 25 Production-based compliance cost comparison in mixed rate/mass models As Figure 25 indicates, three MISO states were shown to have lower costs under rate-based compliance when all MISO states chose the same compliance mechanism (CPP column). In the first mixed run, CPP 1 st Mixed, states were assigned their least cost compliance mechanism (input compliance regime is indicated by the letter in each cell). Results of this run indicated that all MISO states had lower costs under mass-based compliance (output compliance regime advantage is indicated by the color of each cell). This was confirmed in the second mixed run. The CPP 2 nd mixed model results showed that all input advantages matched the output advantages (the letter in each cell matched the color of each cell), indicating the system had reached equilibrium. This implies that a state s initial lower cost compliance option may not hold if the trading market is not large enough to provide financial benefit Natural Gas Price Variation Due to the large role natural gas-fired generators could play in CPP compliance, MISO performed EI-wide sub-category rate and mixed mass model runs with several natural gas prices. In addition to models with the assumed gas price of $4.67 in 2015, MISO used $2.67 and $6.67 in These additional 19 Modeling of Michigan includes Fermi FINAL VERSION

40 sensitivities were performed using the CPP scenario. The resulting production-based compliance cost trends are detailed in Figure 26. Figure 26 Production-based compliance cost trends under varying natural gas prices As can be seen in the graph above, the cost trends for both rate and mass were consistent under varying natural gas prices, but the magnitude of that trend varied. For example, for the -$2 gas model ($2.67 in 2015), the difference in rate-based costs and mass-based costs was significantly less than in the $2 gas model ($6.67 in 2015). Recall that this study did not optimize capacity expansion as part of the study. MISO contends that gas price changes are likely to change future capacity trajectories, but would not change compliance trends within the same trajectory Fermi 3 Nuclear Unit In the MTEP15 model, MISO included the build of nuclear unit Fermi 3 in Michigan in mid Should Fermi 3 not be installed, MISO also executed EI-wide sub-category rate and mixed mass runs without the installation of Fermi 3. Results of these runs indicated that the lower costs Michigan saw under ratebased compliance in previous runs were driven mainly by Fermi CPP w/o Fermi 3 CPP w/ Fermi 3 Figure 27 Production-based compliance cost comparison for Michigan Figure 27 shows the cost comparison for Michigan without the installation of Fermi 3. Without it, the strong rate advantage that Michigan was previously shown to experience was no longer present, 40 FINAL VERSION

41 indicating that the ERC-producing ability of Fermi 3 was a source of revenue for Michigan under ratebased compliance. While results for Michigan were affected by this change, the rest of the system modeled was not shown to experience significant change. LMPs under both rate-based and mass-based compliance increased by 1%, on average. The CO 2 price in the rate-based model increased by 6% without Fermi 3, but the CO 2 price in the mass-based model remained constant. 4.3 Mid-Term Analysis After applying a range of coal retirement levels under different requirements for CO 2 reduction (described in Section 3.1) to the EGEAS model used for MISO s Mid-Term analysis, total system costs are compared in Figure 28. Figure 28 Total system costs per retirement level under various constraints *Dollar figures are 2016 USD in billions and include capital and production costs. Total system costs were calculated as the sum of fixed O&M costs, variable O&M costs, fuel costs and capital costs. They were based on a 20-year Net Present Value (NPV) calculation using a 2.5% inflation rate. These costs were compared from one level of retirement to the next for each CO 2 constraint scenario. A range of retirement levels that produced the lowest total system costs were identified for each scenario (indicated by tan boxes in Figure 28). From each range, the lower bound was selected for each scenario to represent a conservative estimate for how much capacity may retire. Figure 29 demonstrates that these retirement levels did achieve the required emission reduction in each scenario. Retirements above these levels achieved emission reductions well beyond the required level, as well as increased total system costs. 41 FINAL VERSION

42 Resource Expansion (GW) Figure 29 Emissions under various constraints with identified retirement levels Using the EGEAS software, capacity expansion analysis was performed for each scenario under the coal retirement levels identified in Figure 29, along with the appropriate mass emission constraints. The resulting resources economically selected by the model are shown in Figure 30 (Solar PV Econ and Wind Econ). This figure also includes resources forced into each case to meet the capacity required by RPS mandates (Solar PV RPS and Wind RPS) CC Solar PV - Econ. Solar PV - RPS Wind - Econ. Wind - RPS Partial Final Accelerated Figure 30 Economic unit selection and RPS mandated capacity 42 FINAL VERSION

43 The EPA identified the potential issue of leakage to new CC units this was not allowed in the model. For a description of how the potential for leakage was prevented, see section These results indicated that fleet mix varied greatly with the level of retirement seen and the CO 2 reduction levels targeted. Additionally, as the CPP targets became more stringent, replacement capacity in the form of zero-emitting wind and solar units became more dominant than CCs. 43 FINAL VERSION

44 5 Base Datasets This section describes the preparation of and assumptions for the model that underlies each of the scenarios in MISO s Near-Term CPP analysis. Some of the data presented is not directly incorporated in the CPP study models but is included for information on the creation of the base dataset. The base dataset for the Near-Term study is the 2015 MISO Transmission Expansion Plan (MTEP15) Business-as-Usual (BAU) economic study model. The BAU future captures current policies and trends and assumes a continuation of the status quo throughout the duration of the study period. All applicable 20 EPA regulations governing electric power generation, transmission and distribution (NAICS 2211) are modeled. Demand and energy growth rates are modeled at a level equivalent to the 50/50 forecasts submitted into MISO s Module E Capacity Tracking (MECT) tool 21. All current state-level Renewable Portfolio Standard (RPS) and Energy Efficiency Resource Standard (EERS) mandates are modeled. To capture the expected effects of environmental regulations (primarily the Mercury and Air Toxics Standards) on the coal fleet, 12.6 GW of coal unit retirements are modeled. These assumptions were developed in consultation with stakeholders via MISO s Planning Advisory Committee and approved by a stakeholder vote in early All other scenarios for this study are variations of this underlying dataset. This section also describes the data used to develop the EGEAS model for the Mid-Term analysis of the CPP, which varies slightly from the models in the Near-Term analysis. 5.1 Baseline Generation Expansion The Generation Interconnection Queue is the primary source for out-year capacity; however, the queue is generally limited to five years out or less for new capacity. For this reason, a capacity expansion tool is used to supplement the out years to maintain the load-to-resource balance and Planning Reserve Margin (PRM) target. The Electric Generation Expansion Analysis System (EGEAS), created by the Electric Power Research Institute (EPRI), is the capacity expansion software tool used for long-term regional resource forecasting. EGEAS performs capacity expansions based on long-term, least-cost optimizations with multiple input variables and alternatives. To use Regional Resource Forecast (RRF) units 22 in a production cost model, they must be sited at buses in the powerflow model. Units are sited based on stakeholder-vetted rules and criteria. The objective function of EGEAS study aims to minimize the 20-year capital and production costs, with a reserve margin requirement indicating when and what type of resources could be added to the system. The following sections focus on data assumptions and methodologies specific to EGEAS applications. This baseline generation forecast is used for the Business-As-Usual (BAU) scenario and all other scenarios are built off of this. Below is a map detailing the location and fuel type for these units. 20 The BAU model excludes the Clean Power Plan. 21 For details on the type of information collected through the MECT, see 00%20Level%20Training/Level%20300%20-%20Module%20E%20Capacity%20Tracking%20Tool.pdf. 22 See Section 6 for additional details on RRFs. 44 FINAL VERSION

45 Figure 31 MTEP15 BAU capacity expansion 5.2 Baseline Resource Mix The baseline resource mix per study region is shown in Table 6 and the pie charts in Figure 16 through Figure 22 that follow. The baseline resource mix as defined for this study is the nameplate capacity (in MW) for all existing, under construction and planned units. Region Coal Nuclear Gas Wind Solar Hydro Pumped Storage Oil Other MRO 4, ,781 1, , MISO 73,537 14,953 72,724 16, ,133 2,518 4,200 1,363 NYISO 1,963 5,289 21,406 1, ,876 1,407 4,717 1,378 PJM 76,011 36,308 70,607 7, ,844 5,610 10,285 2,637 SERC 41,063 21,930 49, ,477 4,626 2, SPP 24,421 2,449 28,777 9, , , TVA* 23,710 8,077 20,556 1, ,371 1, Table 6 Existing, under construction and planned units *For EGEAS analysis, Associated Electric Cooperative Inc. (AECI), Louisville Gas & Electric and Kentucky Utilities are combined with the Tennessee Valley Authority (TVA) 45 FINAL VERSION

46 Figure 32 through Figure 38 show the resource mix breakdowns as a percentage of total generation capacity for each modeled Eastern Interconnect region. MISO Region MTEP15 Resource Mix Wind 9% Oil 2% Nuclear 8% Coal 39% Coal Gas Nuclear Wind Oil Hydro Gas 39% Pumped Storage Solar Other Figure 32 MISO resource mix MRO Region MTEP15 Resource Mix Oil 2% Hydro 27% Wind 12% Coal 36% Coal Gas Nuclear Wind Oil Hydro Pumped Storage Gas 22% Solar Other Figure 33 MRO resource mix 46 FINAL VERSION

47 NYISO Region MTEP15 Resource Mix Pumped Storage 3% Hydro 11% Other 3% Coal 5% Coal Oil 11% Wind 5% Gas 50% Gas Nuclear Wind Oil Hydro Nuclear 12% Pumped Storage Solar Other Figure 34 NYISO resource mix PJM Region MTEP15 Resource Mix Pumped Storage 3% Nuclear 17% Wind 4% Oil 5% Coal 36% Coal Gas Nuclear Wind Oil Hydro Pumped Storage Gas 33% Solar Other Figure 35 PJM resource mix 47 FINAL VERSION

48 SERC Region MTEP15 Resource Mix Pumped Storage 4% Nuclear 17% Hydro Oil 5% 2% Coal 32% Coal Gas Nuclear Wind Oil Hydro Pumped Storage Gas 39% Solar Other Figure 36 SERC resource mix SPP Region MTEP15 Resource Mix Nuclear 4% Wind 14% Hydro Oil 3% 2% Coal 35% Coal Gas Nuclear Wind Oil Hydro Pumped Storage Gas 41% Solar Other Figure 37 SPP resource mix 48 FINAL VERSION

49 TVA Region MTEP15 Resource Mix Pumped Storage 3% Nuclear 13% Wind 3% Hydro 9% Gas 33% Coal 39% Coal Gas Nuclear Wind Oil Hydro Pumped Storage Solar Other Figure 38 TVA resource mix 49 FINAL VERSION

50 5.3 Regional Demand and Energy Forecasts In PLEXOS, projected future demand and energy growth rates are input at the company level. The MISO baseline value for the demand growth rate is derived from the Module E 50/50 load forecast growth rate (0.8 percent). Low and high values, for both demand and energy, are established by going 1.3 standard deviations above and below the baseline. By utilizing the Load Forecast Uncertainty (LFU) metric, there is an 80% probability that the demand and energy forecast will fall within the high and low growth rates of 0.14% to 1.5%.The effective demand and energy growth rates for each region are calculated after the EGEAS capacity expansion analysis, taking only state-level Demand Side Management (DSM) mandate and goal projections into consideration. The effective growth rates are ultimately used in PLEXOS (seen below in Table 7). Region Demand (%) BAU Energy (%) MISO MHEB MRO NYISO PJM SERC SPP TVA Table 7 Effective demand and energy growth rates ( ) 5.4 Fuel Forecasts All of the fuel forecasts are developed in PowerBase using a pointer system. A pointer system works by designating one fuel as the fuel index and then all other fuel forecasts are based on this fuel index, with some adjustment (usually due to transportation costs) from the index value. In the MTEP database, all natural gas-fired generators point to the Henry Hub natural gas forecast. Therefore, all references to natural gas in the futures matrix are in terms of the Henry Hub forecast. The baseline natural gas forecast used for MTEP15 is developed by Bentek as part of the MISOcommissioned Phase III Natural Gas Infrastructure Analysis 23. High and low forecasts are developed by adding or subtracting 20 percent from the baseline. Since Bentek assumed an inflation rate of approximately 3.5 percent in their forecast, it is necessary to remove this inflation rate and to use the inflation rates for the BAU future scenario that was selected in the Futures development process. The resulting MTEP15 natural gas forecasts are in nominal dollars per MMBtu (Figure 39). 23 Phase III: Natural Gas-Fired Electric Power Generation Infrastructure Analysis An Analysis of Pipeline Capacity Availability. Table FINAL VERSION

51 Natural Gas Price (Nominal $/MMBtu) $15 $12 $9 $ $ $0 Figure 39 MTEP15 natural gas price 5.5 Study Areas The MTEP15 database comprises all areas in the Eastern Interconnect, with the exception of Florida, ISO New England and Eastern Canada. The eight areas referenced in this appendix are: Midwest Reliability Organization (MRO) Midcontinent Independent System Operator (MISO) New York Independent System Operator (NYISO) PJM Interconnection (PJM) SERC Reliability Corporation (SERC) Southwest Power Pool (SPP) Tennessee Valley Authority (TVA) 51 FINAL VERSION

52 Figure 40 MISO market region Figure 41 Study footprint 52 FINAL VERSION

53 5.6 Capacity Types Generation capacity is categorized into existing, under construction and planned units. Assumptions related to each of these categories include the following: Existing: Operating license extensions are assumed on all nuclear units. Under Construction: Units with steel in the ground, but not yet under commercial operation. Planned: All capacity resources with a signed Generator Interconnection Agreement (GIA). 5.7 Firm Interchange Firm interchange contributes to resource adequacy by reducing a region s overall internal capacity needs over time. It is assumed that each modeled region will build generation capacity to meet its own resource adequacy needs. Based on the 2015 Loss of Load Expectation (LOLE) External Ties Model, MISO assumes a net scheduled interchange of 3,157 MW. This capacity is held constant in all 20 years of the EGEAS modeling and is assumed to be available at the time of MISO peak. This is used for capacity forecasting when defining the base dataset used in PLEXOS. In PLEXOS, net scheduled interchange between MISO and their first tier neighbors is scheduled dynamically using SCED and in consideration of hurdle rates. Scheduled interchange with regions that are not modeled in PLEXOS is an hourly value that represents what the interchange would be with those regions. 5.8 Hurdle Rates Hurdle rates influence the capability of a pool to obtain, support or sell energy to other pools. In order for a sale to occur, the difference in dispatch costs between the buying pool and the selling pool must be greater than the hurdle rate between them. PLEXOS performs security constrained unit commitment and economic dispatch, with user-defined hurdle rates. The hurdle rate for the unit commitment step is called the commitment hurdle rate; likewise, the hurdle rate defined for the economic dispatch step is the dispatch hurdle rate. Normally, users set the commitment hurdle rate to be greater than the dispatch hurdle rate. This causes a pool s units to be dispatched against its own pool load first, then allows pool interchange during the final dispatch via the dispatch hurdle rate. Though there is no standard for defining hurdle rates, they are commonly based on the filed transmission through-and-out rates, plus a market inefficiency adder. The dispatch hurdle rates between pools are shown in Table 8. TO: PJM MISO TVA MRO SPP SERC MHEB NYISO TVAO* FROM: PJM * 1 / / 4.8 N/A N/A 4.8 / 4.8 N/A 10 / / 4.8 MISO 8 / 8 * 7.5 / / / / 8 0 / 0 N/A 7.4 / FINAL VERSION

54 TVA 30 / / 30 * N/A - / - 30 / 30 N/A N/A 30 / 30 MRO N/A 6.3 / 5.7 N/A * 6.9 / 6.9 N/A 6.5 / 4.5 N/A SPP N/A 5.1 / / / 5.1 * N/A N/A N/A 5.1 / 5.1 SERC 6.5 / / / 5.0 N/A N/A * N/A N/A 6.8 / 5.0 MHEB N/A 0 / 0 N/A 11.6 / 7.3 N/A N/A * N/A N/A NYISO 3 / 3 N/A N/A N/A N/A N/A N/A * N/A TVAO* 6.5 / / / 8 N/A 8.3 / / 5.7 N/A N/A * Table 8 Dispatch hurdle rates 5.9 Planning Reserve Margin Targets The Planning Reserve Margin (PRM) is entered into EGEAS for the first year of the simulation, and is assumed to remain constant throughout the entire 20-year study period. PRM targets are based on respective system co-incident peaks (MW), with the exception of SPP s, which is based on its noncoincident peak (MW). Table 9 presents the 2014 reserve margin, as well as the PRM target, for each region. Region 2014 Reserve PRM Margin (%) Target (%) MISO MRO NYISO PJM SERC SPP TVA Table 9 PRM margins and targets 5.10 Wind Hourly Profile and Capacity Credits A majority of the wind in the MISO footprint is registered as Dispatchable Intermittent Resources, or DIR. Generators with this designation are able to bid into the day-ahead market using high-confidence wind forecast data. Given that this information is not available for future years, EGEAS models all wind as a non-dispatchable technology using actual historical wind data developed during the MISO Regional Generator Outlet Study (RGOS) and updated as part of the NREL Eastern Renewable Generation Integration Study (ERGIS). All the RGOS wind zone profiles within MISO are averaged to arrive at a single profile, which is used in the EGEAS capacity expansion analysis. Similarly, a single profile for each of the regions external to MISO is made by averaging all NREL wind sites within each respective region. The wind capacity credit is the maximum capacity credit that a wind resource may receive if it meets all other obligations of Module E to be a capacity resource. This value, which is a percent of the maximum 54 FINAL VERSION

55 nameplate capacity of the unit, reflects the risk associated with reliance upon an intermittent resource, such as wind. The capacity factor is the actual annual energy output of the unit as a percentage of the total potential energy output (based on 8,760 hours in a year). The wind capacity credit is updated annually during the MISO Loss of Load Expectation (LOLE) analysis and, for the 2014 planning year, was calculated to be 14.1 percent. Table 10 shows the capacity factors applied to each region as input to the EGEAS model for MTEP15. In PLEXOS, however, all wind units are modeled individually with their own hourly profiles and are considered dispatchable Reserve Contribution Region Annual Capacity Factor (%) MISO 40 MRO 41 NYISO 40 PJM 37 SERC 43 SPP 43 TVA 36 Table 10 Regional wind modeled capacity factors Three specific assumptions were made with regard to reserve contribution: 14.1 percent of nameplate wind capacity is counted toward its reserve capacity contribution. 25 percent of nameplate solar capacity is counted toward its reserve capacity contribution. The summer de-rated capacity for conventional generation is counted toward its reserve capacity contribution Financial Variables Variables associated with the financing of new generation projects are listed in Table 11. Note that these are average values across the footprint. These financial variables are used in MTEP15 EGEAS simulations. Variable Rate (%) Composite Tax Rate Insurance Rate 0.50 Property Tax Rate 1.50 AFUDC* Rate 7.00 * Allowance for Funds Used During Construction Table 11 Financial variables 55 FINAL VERSION

56 5.13 Load Shapes The load shapes used in PLEXOS and their sources are presented in Table 12. Load Shape System Wind Solar Description and Source 2006 hourly profiles from Ventyx 2006 hourly profiles developed by AWS TrueWind for EWITS, and updated as part of the NREL Eastern Renewable Generation Integration Study (ERGIS) hourly profile developed by NREL for the Eastern Renewable Generation Integration Study (ERGIS) Table 12 Load shape descriptions and sources 5.14 Alternative Generator Categories Table 13 and Table 14 list the generic categories of generators used when forecasting future units to meet the planning reserve margin requirements. Supply Side Options Biomass Coal Combined Cycle - with and without sequestration Combustion Turbine Compressed Air Energy Storage Hydro Integrated Gasification Combined Cycle (IGCC) - with and without sequestration Nuclear Pumped Hydro Storage Solar Wind - on-shore and off-shore Table 13 Alternative generator categories supply-side Demand Side Options 56 FINAL VERSION

57 Commercial & Industrial (C&I) Low Cost Energy Efficiency (EE) program C&I Interruptible Table 14 Demand side management alternatives 5.15 Alternative Generator Data Table 15 shows the fixed operation and maintenance (O&M) cost, variable O&M cost, heat rate, lead time (inclusive timeframe for unit construction), maintenance hours, and forced outage rate (FOR) for the alternative supply-side generator categories used in MTEP15 regional resource forecasting. The capacity of each forecasted generic unit from each category is 1,200 MW, with the exception of wind at 300 MW and solar at 25 MW. Monetary values given in the table are in 2014 dollars. Type Fixed O&M Variable O&M Heat Rate Lead Time Maintenance Duration Forced Outage Rate $/kw-yr $/MWh MMBtu/MWh Years Hours % Biomass Coal CC CCS* CT Hydro IGCC IGCCS** Nuclear PV Wind Table 15 Alternative generator data * Combined-Cycle with Sequestration ** Integrated Gasification Combined-Cycle with Sequestration 5.16 Baseline Renewable Portfolio Standards (RPS) Many MISO states have existing RPS mandates and goals with various starting dates, growth rates and program terms. For the MTEP BAU, only the mandates are captured as the goals are not legally enforceable. These numbers were taken from the DSIRE website: Incremental renewable energy required per the RPS mandates, beyond existing qualifying resources, is forced into the model. Most RPS mandates have an energy requirement and only from certain types of renewable resources (see DSIRE for more). The existing units in MISO have expected capacity factors either based on historical performance or fleet averages. This, along with the nameplate capacities, is used to project the energy produced by existing renewable resources. The difference between the energy 57 FINAL VERSION

58 produced by existing renewable resources in each state and the energy required by state mandates drives the amount of renewable resource capacity to be added to the BAU model. The exact capacity amount is back-calculated using the capacity factors for wind, as this is the most common renewable energy resource built. The capacity expansion wind units are built in 300 MW increments. To the extent states have solar requirements specifically, a 25 MW PV unit is used with the characteristics also outlined in Table 15. Wind is assumed to have a 40.16% capacity factor, and an 18.9% capacity factor for solar. These are not the values used to calculate contribution to peak and for purposes of meeting PRM, but rather are used for an energy-to-capacity conversion. The capacity values chosen are specific to accommodate the capacity required without going over the EGEAS limitation of 99 of any given type of unit. This limitation prevents obtaining the exact MW value mandated by either existing RPS or the CPP. This means that if 299 MW are required, a 300 MW unit will be built, and if 301 MWs are required, two 300 MW units will be built MTEP15 Futures Matrix The MTEP15 futures matrix is included as a reference. The highlighted values (Table 16) are used for building the MTEP15 BAU model, the base model for the Phase III study. The capital costs of expansion units are also calculated using these numbers, along with those listed in the MTEP16 matrix (shown in the following section) to create a range of potential resource capital costs under CPP compliance. 58 FINAL VERSION

59 Uncertainty Unit Low (L) Mid (M) High (H) New Generation Capital Costs 1 Coal ($/KW) 2,247 2,996 3,745 CC ($/KW) 783 1,045 1,306 CT ($/KW) Nuclear ($/KW) 4,235 5,647 7,058 Wind-Onshore ($/KW) 1,525 2,034 2,542 IGCC ($/KW) 2,898 3,864 4,830 IGCC w/ CCS ($/KW) 5,054 6,738 8,423 CC w/ CCS ($/KW) 1,604 2,139 2,674 Pumped Storage Hydro ($/KW) 4,050 5,400 6,750 Compressed Air Energy Storage ($/KW) 957 1,276 1,595 Photovoltaic ($/KW) 2,225 2,966 3,708 Biomass ($/KW) 3,151 4,201 5,251 Conventional Hydro ($/KW) 2,248 2,998 3,747 Wind-Offshore ($/KW) 4,771 6,362 7,952 Demand Growth Rate 2 % 0.14% 0.80% 1.50% Energy Growth Rate 3 % 0.14% 0.80% 1.50% Demand Response Level 4 % State mandates only State mandates and goals Energy Efficiency Level 4 % State mandates only State mandates and goals Natural Gas 5 MTEP15 UNCERTAINTY VARIABLES ($/MMBtu) Demand and Energy Natural Gas Bentek -20% Fuel Prices (Starting Values) Bentek forecast from Phase III Gas Study Table 16 MTEP15 Futures matrix Bentek +20% Oil ($/MMBtu) Powerbase default -20% Powerbase default 6 Powerbase default + 20% Coal ($/MMBtu) Powerbase default -20% Powerbase default 7 Powerbase default + 20% Uranium ($/MMBtu) Fuel Prices (Escalation Rates) Oil % Coal % Uranium % Emissions Costs SO2 ($/ton) NO x ($/ton) 0 0 NO x : 500 Seasonal NO x : 1000 CO 2 ($/ton) Other Variables Inflation % Retirements MW 12,600 MW 19,000 MW + 11,600 MW age-related retirements = 30,600 MW 8 23 GW Renewable Portfolio Standards % State mandates only 20% MISO-wide mandate Solar 5% of overall mandate 30% MISO-Wide Mandate Solar 10% of overall mandate Notes on uncertainty variables: 1 All costs are overnight construction costs in 2014 dollars; sourced from EIA and escalated according to the GDP Implicit Price Deflator; H and L values are 20% +/- from the M value 2 Mid value for demand growth rate is the Module-E 50/50 load forecast growth rate 3 Energy values are determined based on historical load factors for each Load-Serving Entity (LSE) 4 MTEP13 modeled state mandates and goals for DR & EE 5 Prices reflect the Henry Hub natural gas price 6 PowerBase default for oil is $19.39/MMBtu 7 PowerBase range for coal is $1 to $4, with an average value of $1.69/MMBtu 8 11,600 MW value is based on MTEP13 database 59 FINAL VERSION

60 5.18 MTEP16 Futures Matrix MTEP16 futures introduced several changes to the capital cost assumptions for new resources. For example, maturity curves were assigned to new wind and new photovoltaic units in MTEP16 s CPP futures. Wind and solar capital costs are assumed to start at $1,750/kW in Solar is decreased 10% each year and wind is decreased by 1% each year until The values for other new units are highlighted in yellow in Table 17 followed by a graph of the maturity curves for wind and solar units (Figure 42). MTEP16 UNCERTAINTY VARIABLES Uncertainty Unit Low (L) Mid (M) High (H) New Generation Capital Costs 1 Coal ($/KW) 2,279 3,039 3,799 CC ($/KW) 795 1,060 1,324 CT ($/KW) Nuclear ($/KW) 4,296 5,728 7,160 Wind-Onshore ($/KW) 1,750 2,063 2,579 IGCC ($/KW) 2,940 3,919 4,899 IGCC w/ CCS ($/KW) 5,126 6,835 8,544 CC w/ CCS ($/KW) 1,627 2,170 2,712 Pumped Storage Hydro ($/KW) 4,108 5,477 6,846 Compressed Air Energy Storage ($/KW) 971 1,295 1,618 Photovoltaic ($/KW) 1,750 3,009 5,014 Biomass ($/KW) 3,196 4,261 5,326 Conventional Hydro ($/KW) 2,281 3,041 3,801 Wind-Offshore ($/KW) 4,840 6,453 8,066 Table 17 MTEP16 Futures matrix Figure 42 Renewable cost maturity curves *Prices are shown in real dollars and do not account for inflation. 60 FINAL VERSION

61 5.19 Mid-Term Analysis Dataset While the Near-Term model uses the MTEP15 BAU, the base model for the Mid-Term analysis uses the base data from MTEP16 model without any MTEP16 assumptions to attempt to get closer to a MTEP17 base model as the results are being used to inform the MTEP17 Futures development process. The base data is updated to reflect known retirements and fuel conversions as a result of MATS and other EPA regulations compliance. These updates are made to reflect feedback received in the quarterly MISO EPA Survey Demand & Energy As Module E data was not set for spring of 2016 at the time this study was performed, the previous MTEP16 demand and energy values are used for (2015 is dropped, and the forecast is extended to 2035) Demand (GW) Energy (TWh) Demand (GW) Energy (TWh) Gas Price Forecast The IHS CERA forecast is used for the Henry Hub price. No transportation adders are adjusted from the base data in PowerBase Henry Hub ($/MMBtu) Henry Hub ($/MMBtu) Retirements and Conversions Beyond publicly known retirements and approved Attachment-Y retirements, additional retirements disclosed in the quarterly MISO EPA survey are captured in the model. Additionally, any disclosed coalto-gas conversions from the quarterly survey are also modeled. For these, the assumed technology is changed to Steam Turbine (ST) Gas from ST Coal, the fuel is changed to reflect costs for the nearest fuel hub to capture appropriate transportation adders from the Henry Hub price. The heat rate, fixed O&M, variable O&M and emissions are also updated to reflect a class-average for the type ST Gas in the given state of the conversion Demand-Side Management In late 2015, Applied Energy Group (AEG) provided updated Demand Side Management programs for use in EGEAS modeling. These 11 AEG programs (five demand response, four energy efficiency and two FINAL VERSION

62 distributed generation) come in a number of variations based on MTEP16 futures. These variations a BAU set, a Low Demand set, a High Demand set, and a carbon reduction set based on the CPP futures. As all Mid-Term runs constrained CO 2 to some degree, the CPP-related DSM set is used in these models. These 11 programs vary by future based on how quickly demand and energy is added to the program and at what cost. In all preliminary test-runs of the updated AEG DSM programs, the low-cost Commercial/Industrial and Residential Energy Efficiency (EE) Programs were always selected, so were assumed as base input assumptions for these runs. Over the entire 20-year study period, the Commercial/Industrial EE program grows to a total of 10.4 GW, and the Residential EE to 0.8 GW Mitigating Unintended Emission Shifting In initial test runs, there was a large amount of unintended emission shifting observed, namely in the form of over-utilization of existing combustion turbines running at 85-95% capacity factors, and underutilization of existing gas units. For the Mid-Term analysis, it is assumed that an average annual dispatch over 10-15% is unrealistic based on typical CT design. To limit CT dispatch, the Forced Outage Rate in EGEAS is increased from the typical 5% range to 85%. This allows full reserve contribution and an 80-90% capacity factor in peak months, but limits annual dispatch to a 15% maximum. In typical MTEP models, coal units are assumed to have a must-run status for their first loading block that can be anywhere from 25-50% of the overall unit nameplate capacity, typically with a higher heat rate than a full-load heat rate. In this way, coal units emit CO 2 inefficiently when only the minimum loading block with the highest heat rate is running. This prevents existing ST gas and combined cycle units from running, and pushes more and more energy to new CC units (as dispatch of CTs is limited). This is how the EPA defines leakage under the CPP, and must be avoided. To do so, the must-run status of coal units is removed and minimum coal loading blocks are allowed to be turned off when not economic or prevent the meeting the carbon constraint. This allows increased dispatch of existing ST gas and CCs, and leakage to new CCs is mitigated. 62 FINAL VERSION

63 6 MTEP Unit Siting Methodology Regional Resource Forecasted (RRF) units are an output of Step 1 of the MTEP 7-step process. Given that the generator interconnection queue is typically only useful for one to five years out for capacity, a capacity expansion tool, such as EGEAS, is used to supplement the out years to maintain the load-toresource balance. These units must be sited within the powerflow model for use in PLEXOS. Beginning with MTEP11, MISO included Demand Response (DR) and Energy Efficiency (EE) units in the EGEAS capacity planning process. While EE is simply netted out of the baseline demand and energy values, DR units also have to be sited into the powerflow models for production cost analysis. Therefore, additional siting methodology for DR has been developed. A Geographic Information System (GIS) software program called MapInfo is used to assist in the generation siting. Siting rules, which are detailed below, are used to develop layers within the mapping software showing the potential locations of the resource forecasted units. The siting can be broken into three main categories. The categories are general siting rules, siting priority order and unit-specific greenfield siting. General siting rules apply to all scenarios. 6.1 General Siting Rules for EGEAS The rules outlined in this section show, at a higher level, many of the underlying assumptions that go into the siting of RRF generation. These criteria could be referred to as the first pass siting criteria. Site by region, with the exception of wind. Share the Pain mentality. Not all generation in a region can be placed in one state and one state cannot be excluded from having generation sited. Avoid greenfield sites for gas units (CTs and CCs) if possible - prefer to use all brownfield sites. Site baseload units in 600 MW increments, except nuclear which is sited at 1,200 MW. Limit the total amount of expansion at an existing site to no more than an additional 2,400 MW. Restrict greenfield sites to a total size of 2,400 MW. Limit using queue generation without a signed Generator Interconnection Agreement (GIA) in multiple futures. Transmission is not an initial siting factor, but may be used as a weighting factor, all things being equal Generator Developmental Statuses A generator s developmental status is required to determine how the unit will be treated in both the EGEAS capacity expansion model and the siting process. Existing and queue generation is given one of the following developmental statuses within the PowerBase database: Active Existing, online generation including committed and uncommitted units. Does not include generation which has been mothballed or decommissioned. Planned - A generator that is not online, has a future in-service date, is not suspended or postponed and has proceeded to a point where construction is almost certain, such as it has a signed Interconnection Agreement (IA), all permits have been approved, all study work has been completed, state or administrative law judge has approved, etc. These units are used in the model to meet future demand requirements prior to the economic expansions. Future Generators with a future online date that do not meet the criteria of the planned status. Generators with a future status are typically under one of the following categories: 63 FINAL VERSION

64 proposed, feasibility studies, permits applied, etc. These generators are not used in the models but are considered in the siting of future generation. Canceled Generators that have been suspended, canceled, retired or mothballed. These units are not included in the EGEAS capacity expansion model, although their sites are often considered for brownfield locations in the siting process. 6.2 Site Selection Priority Order for EGEAS Priority 1: Generators with a future status Queue generators without a signed Interconnection Agreement (IA) The New Entrants Generators defined by Ventyx (noted as EV Gens) Both Queue and EV Gens are under the following statuses: o Permitted o Feasibility o Proposed Priority 2: Brownfield sites (Coal, CT, CC, Nuclear Methodology) Priority 3: Retired/mothballed sites that have not been re-used Priority 4: Greenfield sites Queue and New Entrants in canceled or postponed status Priority 5: Greenfield sites Greenfield siting methodology 6.3 Unit-Specific Greenfield Siting Rules for EGEAS Thermal unit siting uses a specific set of rules for each type of capacity. For instance, a coal unit has a different set of criteria than a combined cycle or combustion turbine. Also, demand-side resources have a completely different set of rules to follow when it comes to siting Greenfield Coal Siting Rules Required Criteria: Within 1 mile of railroad or navigable waterway Outside 20-mile buffer of Class I lands Outside air quality non-attainment region Outside 25-mile buffer surrounding a major urban area (population greater than 50,000 and an area larger than 25 mi 2 ) Within 1/2 mile of a major river or lake Optional Criteria: Within 20 miles of a coal mine or dock capable of producing more than 2 million tons per year Access to gas pipeline Multiple railroad lines These rules for coal are input into MapInfo to show potential greenfield placement Greenfield Combined-Cycle Siting Rules Required Criteria: Within 1 mile of railroad or navigable waterway Within 2 miles of river or a lake (lake has to be larger than 100 mi 2 ) Within 10 miles of a gas pipeline (diameter of 12 inches or greater) Within 25 miles of a major urban area 64 FINAL VERSION

65 6.3.3 Greenfield Combustion Turbine Siting Rules Required Criteria: Within 20 miles of railroad or navigable waterway Within 5 miles of a gas pipeline (diameter 12 inches or greater) Optional Criteria: CTs can be located almost anywhere CTs historically have been located near metro areas, but not required CTs do not need a river for cooling Less likely to build pipeline for CT vs. CC CTs may be the preferred generation for coal retirement sites within metro areas Greenfield Nuclear Siting Rules Required Criteria: Use existing nuclear sites only All states are eligible for siting of future nuclear generation Greenfield Wind Siting Rules Required Criteria: Not in a state or national park Not in metro areas Not on state-managed lands Site wind within a state to meet its mandate, unless potential wind capacity is exceeded, then site in neighboring state(s) Greenfield Photovoltaic Siting Rules Photovoltaic (PV) will be sited using annual Solar Global Horizontal Irradiance (kwh/m 2 ). The Solar Global Horizontal Irradiance Intelligent Map Layer includes monthly and annual solar resource potential for the United States. The insolation values represent the average solar energy available to a horizontal flat plate collector such as a PV panel Demand Response and Energy Efficiency Siting Rules Demand response capacity is sited at the top five load buses in each Load Serving Entity (LSE). If an LSE serves load in more than one state, the top five load buses in each state having a DR mandate or goal are used, with the DR being allocated based upon the percentage required in each state s mandate or goal. The impact of energy efficiency is accounted for in the demand and energy growth rates, as EE is typically available during all 8,760 hours in a year. This methodology applies to all Futures. Transmission is not an initial siting factor, but may be used as a weighting factor, all things being equal. Siting is done by region with the exception of wind units. Generation is distributed throughout the states in the study footprint; no one state will have all units sited within its borders; no one state will have zero units sited. Brownfield sites are preferable to Greenfield sites for gas units (CTs & CCs). Baseload units are sited in 600 MW increments and nuclear units, at 1,200 MW each. 65 FINAL VERSION

66 The total amount of expansion to an existing site is limited to no more than an additional 2,400 MW. Greenfield sites are restricted to a total of 2,400 MW. Use of Queue generation in multiple Futures should be limited. 6.4 Greenfield Siting Rules for EGEAS Fuel Type/ Criteria Railroad/ Navigable Waterway Class lands Nonattainment region Urban Area Major River/ Lake Gas Pipeline Coal Mine/ Dock Coal <1 >20 O >25 <0.5 PA <20 LM Biomass <1 >20 O >25s <0.5 PA - LM CC <1 >20 - <25 <2 <10 - CT <20 > <1-2 L Table 18 Greenfield Siting Rules <5 - L = likes : This feature is strongly preferred for siting a unit of this type. LM = likes multiples : Multiple instances of this feature are strongly preferred for siting a unit of this type. <x = within x miles : The unit should be sited within x miles of this feature. >x = outside of x miles : The unit should be sited outside of x miles of this feature. O = outside : The unit should be sited outside of the range of this feature. PA = prefer access : Access to this feature is preferred, though not required. 66 FINAL VERSION

67 Appendix A: Energy Efficiency from EPA s draft CPP Building Block 4 used in EWS scenario Demand-Side EE (% of Avoided MWh Sales) State l 2029 State Generation as % of sales 2012 Total MWh (sales x ) Alabama 1.36% 2.11% 3.00% 4.01% 5.13% 6.19% 7.15% 8.01% 8.79% 9.48% % 92,655 Alaska 1.22% 1.95% 2.82% 3.82% 4.93% 6.02% 7.01% 7.91% 8.72% 9.45% 95.58% 6,898 Arizona 5.24% 6.28% 7.22% 8.07% 8.83% 9.50% 10.10% 10.61% 11.05% 11.42% % 80,701 Arkansas 1.52% 2.31% 3.24% 4.28% 5.42% 6.46% 7.41% 8.26% 9.03% 9.71% % 50,379 California 4.95% 6.04% 7.03% 7.93% 8.74% 9.46% 10.11% 10.67% 11.15% 11.56% 71.07% 279,029 Colorado 3.92% 5.08% 6.14% 7.09% 7.96% 8.73% 9.42% 10.03% 10.55% 11.01% 89.08% 57,717 Connecticut 4.71% 5.86% 6.92% 7.88% 8.76% 9.55% 10.25% 10.87% 11.42% 11.88% % 31,707 Delaware 1.14% 1.86% 2.73% 3.72% 4.83% 5.94% 6.96% 7.89% 8.72% 9.47% 45.09% 12,384 Florida 2.03% 2.91% 3.92% 5.03% 6.08% 7.04% 7.90% 8.68% 9.37% 9.98% 90.20% 237,247 Georgia 1.76% 2.60% 3.56% 4.63% 5.73% 6.74% 7.64% 8.46% 9.19% 9.83% 87.75% 140,815 Hawaii 1.29% 2.04% 2.92% 3.93% 5.05% 6.13% 7.11% 8.00% 8.80% 9.52% 96.24% 10,363 Idaho 3.80% 4.98% 6.06% 7.04% 7.93% 8.73% 9.44% 10.07% 10.62% 11.10% 46.83% 25,493 Illinois 4.36% 5.53% 6.60% 7.57% 8.46% 9.26% 9.97% 10.61% 11.16% 11.63% % 154,320 Indiana 3.20% 4.33% 5.49% 6.56% 7.53% 8.42% 9.22% 9.93% 10.56% 11.11% % 113, FINAL VERSION

68 Iowa 4.65% 5.78% 6.82% 7.77% 8.62% 9.39% 10.08% 10.69% 11.21% 11.66% % 49,142 Kansas 1.22% 1.95% 2.83% 3.83% 4.95% 6.05% 7.05% 7.96% 8.78% 9.52% % 43,320 Kentucky 1.91% 2.78% 3.77% 4.87% 5.96% 6.95% 7.85% 8.66% 9.38% 10.02% 97.18% 95,736 Louisiana 1.14% 1.85% 2.71% 3.69% 4.78% 5.88% 6.88% 7.78% 8.60% 9.33% 90.08% 91,094 Maine 5.37% 6.47% 7.48% 8.39% 9.22% 9.96% 10.62% 11.20% 11.70% 12.13% % 12,429 Maryland 4.21% 5.38% 6.45% 7.44% 8.33% 9.13% 9.85% 10.48% 11.04% 11.51% 60.82% 66,456 Massachusetts 4.43% 5.60% 6.68% 7.66% 8.56% 9.37% 10.09% 10.73% 11.29% 11.77% 74.77% 59,467 Michigan 4.59% 5.74% 6.80% 7.77% 8.64% 9.43% 10.14% 10.76% 11.30% 11.77% % 112,690 Minnesota 4.80% 5.92% 6.95% 7.89% 8.73% 9.49% 10.17% 10.76% 11.28% 11.72% 82.84% 73,094 Mississippi 1.40% 2.17% 3.07% 4.09% 5.22% 6.28% 7.24% 8.11% 8.89% 9.59% 98.63% 52,022 Missouri 1.58% 2.38% 3.33% 4.39% 5.54% 6.60% 7.56% 8.43% 9.22% 9.92% 99.47% 88,626 Montana 3.36% 4.51% 5.63% 6.65% 7.57% 8.41% 9.15% 9.81% 10.39% 10.90% % 14,905 Nebraska 2.20% 3.13% 4.18% 5.34% 6.41% 7.38% 8.27% 9.06% 9.78% 10.40% % 33,143 Nevada 2.95% 4.02% 5.18% 6.24% 7.20% 8.07% 8.85% 9.54% 10.16% 10.69% 96.67% 37,822 New Hampshire 2.84% 3.90% 5.08% 6.19% 7.21% 8.14% 8.98% 9.74% 10.41% 11.00% % 11,687 New Jersey 1.25% 1.99% 2.87% 3.88% 5.00% 6.10% 7.11% 8.02% 8.84% 9.58% 76.19% 80,689 New Mexico 3.10% 4.19% 5.32% 6.35% 7.28% 8.11% 8.86% 9.52% 10.10% 10.60% % 24,919 New York 4.42% 5.59% 6.67% 7.65% 8.55% 9.35% 10.08% 10.72% 11.28% 11.76% 93.05% 153,914 North Carolina 2.37% 3.32% 4.39% 5.51% 6.53% 7.45% 8.28% 9.02% 9.68% 10.26% 86.12% 137,704 North Dakota 1.39% 2.16% 3.07% 4.10% 5.24% 6.32% 7.30% 8.19% 8.99% 9.71% % 15, FINAL VERSION

69 Ohio 4.17% 5.35% 6.43% 7.42% 8.32% 9.13% 9.86% 10.51% 11.07% 11.56% 85.97% 163,906 Oklahoma 1.86% 2.71% 3.69% 4.79% 5.88% 6.88% 7.79% 8.60% 9.33% 9.97% % 63,797 Oregon 4.66% 5.77% 6.78% 7.69% 8.52% 9.26% 9.92% 10.49% 10.99% 11.41% % 50,195 Pennsylvania 4.67% 5.81% 6.85% 7.79% 8.65% 9.42% 10.11% 10.71% 11.24% 11.69% % 155,577 Rhode Island 3.90% 5.11% 6.22% 7.24% 8.17% 9.02% 9.78% 10.45% 11.04% 11.56% 97.63% 8,287 South Carolina 2.32% 3.26% 4.32% 5.45% 6.47% 7.40% 8.23% 8.98% 9.65% 10.23% % 83,622 South Dakota 1.60% 2.41% 3.36% 4.42% 5.57% 6.62% 7.57% 8.44% 9.21% 9.91% 82.32% 12,615 Tennessee 2.21% 3.14% 4.18% 5.33% 6.38% 7.33% 8.19% 8.97% 9.65% 10.26% 71.81% 103,620 Texas 1.78% 2.62% 3.59% 4.68% 5.78% 6.79% 7.70% 8.52% 9.26% 9.91% 98.12% 392,523 Utah 3.62% 4.82% 5.91% 6.91% 7.81% 8.62% 9.34% 9.98% 10.54% 11.03% % 31,956 Virginia 1.23% 1.96% 2.82% 3.81% 4.91% 5.98% 6.95% 7.83% 8.62% 9.33% 58.01% 115,890 Washington 4.24% 5.39% 6.43% 7.38% 8.24% 9.01% 9.69% 10.29% 10.81% 11.26% % 99,271 West Virginia 1.77% 2.62% 3.60% 4.70% 5.83% 6.86% 7.81% 8.66% 9.43% 10.11% % 33,132 Wisconsin 4.68% 5.82% 6.87% 7.83% 8.70% 9.48% 10.17% 10.79% 11.33% 11.79% 83.97% 73,988 Wyoming 1.61% 2.42% 3.35% 4.41% 5.53% 6.55% 7.48% 8.32% 9.07% 9.73% % 18, FINAL VERSION

70 Appendix B: MISO Hub LMPs for All Years and Scenarios MISO Hub BAU - Hub Price ($/MWh) DSG Load $ $ $ MISOARK $ $ $ MISOILL $ $ $ MISOIND $ $ $ MISOLOUS $ $ $ MISOMICH $ $ $ MISOMINN $ $ $ MISOTEX $ $ $ Western Load $ $ $ WOTAB Load $ $ $ MISO Hub CPP Mixed Mass - Hub Price ($/MWh) CPP Subcat Rate - Hub Price ($/MWh) DSG Load $ $ $ $ $ $ MISOARK $ $ $ $ $ $ MISOILL $ $ $ $ $ $ MISOIND $ $ $ $ $ $ MISOLOUS $ $ $ $ $ $ MISOMICH $ $ $ $ $ $ MISOMINN $ $ $ $ $ $ MISOTEX $ $ $ $ $ $ Western Load $ $ $ $ $ $ WOTAB Load $ $ $ $ $ $ MISO Hub C2G Mixed Mass - Hub Price ($/MWh) C2G Subcat Rate - Hub Price ($/MWh) DSG Load $ $ $ $ $ $ MISOARK $ $ $ $ $ $ MISOILL $ $ $ $ $ $ MISOIND $ $ $ $ $ $ MISOLOUS $ $ $ $ $ $ MISOMICH $ $ $ $ $ $ MISOMINN $ $ $ $ $ $ MISOTEX $ $ $ $ $ $ Western Load $ $ $ $ $ $ WOTAB Load $ $ $ $ $ $ FINAL VERSION

71 MISO Hub GBO Mixed Mass - Hub Price ($/MWh) GBO Subcat Rate - Hub Price ($/MWh) DSG Load $ $ $ $ $ $ MISOARK $ $ $ $ $ $ MISOILL $ $ $ $ $ $ MISOIND $ $ $ $ $ $ MISOLOUS $ $ $ $ $ $ MISOMICH $ $ $ $ $ $ MISOMINN $ $ $ $ $ $ MISOTEX $ $ $ $ $ $ Western Load $ $ $ $ $ $ WOTAB Load $ $ $ $ $ $ MISO Hub GWS Mixed Mass - Hub Price ($/MWh) GWS Subcat Rate - Hub Price ($/MWh) DSG Load $ $ $ $ $ $ MISOARK $ $ $ $ $ $ MISOILL $ $ $ $ $ $ MISOIND $ $ $ $ $ $ MISOLOUS $ $ $ $ $ $ MISOMICH $ $ $ $ $ $ MISOMINN $ $ $ $ $ $ MISOTEX $ $ $ $ $ $ Western Load $ $ $ $ $ $ WOTAB Load $ $ $ $ $ $ MISO Hub EWS Mixed Mass - Hub Price ($/MWh) EWS Subcat Rate - Hub Price ($/MWh) DSG Load $ $ $ $ $ $ MISOARK $ $ $ $ $ $ MISOILL $ $ $ $ $ $ MISOIND $ $ $ $ $ $ MISOLOUS $ $ $ $ $ $ MISOMICH $ $ $ $ $ $ MISOMINN $ $ $ $ $ $ MISOTEX $ $ $ $ $ $ Western Load $ $ $ $ $ $ WOTAB Load $ $ $ $ $ $ FINAL VERSION

72 Emissions (short tons) Emissions (short tons) Appendix C: Emissions for All Years and Scenarios 1.E MISO States CO 2 Emissions by Scenario - Mass 1.E+08 8.E+07 6.E+07 4.E+07 CPP C2G GBO GWS EWS EPA 2.E+07 0.E+00 IN IL KY MO MI LA TX AR WI IA MS MN ND SD 2025 MISO States CO 2 Emissions by Scenario - Mass 1.E+08 1.E+08 8.E+07 6.E+07 4.E+07 CPP C2G GBO GWS EWS EPA 2.E+07 0.E+00 IN IL KY MO MI LA TX AR WI IA MS MN ND SD 72 FINAL VERSION

73 FS Emission Rate (lbs/mwh) Emissions (short tons) 2030 MISO States CO 2 Emissions by Scenario - Mass 1.E+08 9.E+07 8.E+07 7.E+07 6.E+07 5.E+07 4.E+07 3.E+07 CPP C2G GBO GWS EWS EPA 2.E+07 1.E+07 0.E+00 IN IL KY MO MI TX LA AR WI IA MS MN ND SD 2022 MISO States CO 2 Emission Rate by Scenario - FS 2,500 2,000 1,500 1,000 CPP C2G GBO GWS EWS EPA ND MO AR WI KY IN MN IL IA MI TX SD LA MS 73 FINAL VERSION

74 FS Emission Rate (lbs/mwh) CC Emission Rate (lbs/mwh) 2022 MISO States CO 2 Emission Rate by Scenario - CC 1, CPP C2G GBO GWS EWS EPA AR WI MS MO MN TX MI IN LA IA IL SD KY ND 2025 MISO States CO 2 Emission Rate by Scenario - FS 2,500 2,000 1,500 1,000 CPP C2G GBO GWS EWS EPA MO AR KY ND IN MN WI IA IL TX SD MI LA MS 74 FINAL VERSION

75 FS Emission Rate (lbs/mwh) CC Emission Rate (lbs/mwh) 2025 MISO States CO 2 Emission Rate by Scenario - CC 1, CPP C2G GBO GWS EWS EPA AR MS WI MO MN TX IN LA SD MI IA IL KY ND 2030 MISO States CO 2 Emission Rate by Scenario - FS 2,500 2,000 1,500 1,000 CPP C2G GBO GWS EWS EPA AR MO KY IN WI MN MS IA TX ND IL LA MI SD 75 FINAL VERSION

76 CC Emission Rate (lbs/mwh) 2030 MISO States CO 2 Emission Rate by Scenario - CC 1, CPP C2G GBO GWS EWS EPA AR MS WI MO MN TX IN LA SD IL IA MI KY ND 76 FINAL VERSION

77 Generation in MISO by Fuel Type (TWh) CO 2 Emissions in MISO (Millions of Short Tons) Generation in MISO by Fuel Type (TWh) CO 2 Emissions in MISO (Millions of Short Tons) Appendix D: MISO Fuel Mix for All Years and Scenarios Rate 2022 Mass BAU CPP C2G GBO GWS EWS CPP C2G GBO GWS EWS Nuclear Other Coal Old CC New CC CT Renewable Emissions Rate 2025 Mass BAU CPP C2G GBO GWS EWS CPP C2G GBO GWS EWS Nuclear Other Coal Old CC New CC CT Renewable Emissions 0 77 FINAL VERSION

78 Generation in MISO by Fuel Type (TWh) CO 2 Emissions in MISO (Millions of Short Tons) Rate 2030 Mass BAU CPP C2G GBO GWS EWS CPP C2G GBO GWS EWS Nuclear Other Coal Old CC New CC CT Renewable Emissions 0 78 FINAL VERSION