4. Resource Adequacy Program Refinements
4.1. Net Qualifying Capacity (NQC) List
The Energy Division made a workshop proposal to post the NQC list that will be used for compliance purposes on the Commission's and/or the California Energy Commission's website. The Energy Division reports that no parties expressed concern about this proposal and that some parties favored it. AReM, DRA, and PG&E supported the proposal in their filed comments. AReM notes that when the CAISO updates the NQC list, it often removes the previous list, making unclear to LSEs which NQC values to use for their RA showings.
Discussion
It should be clear to LSEs what NQC values the Energy Division will apply in its review of LSEs' compliance filings. We therefore endorse this proposal. We concur with the Energy Division that that this is a ministerial action that is within the province of the Energy Division to carry out.
4.2. Local RA Credit for New Resources
In Phase 1 of this proceeding, PG&E proposed that new resources that have not reached commercial operation may be counted toward local RA obligations if the LSE demonstrates local procurement sufficient to cover the obligation in the months preceding the expected commercial operational status of the new resource. PG&E also proposed a stipulation that any LSE relying on a new resource for local RA needs would be responsible for replacing the capacity if the new resource's commercial operational date is delayed. PG&E characterized its proposal as interim, and the Phase 1 decision (D.08-06-031) adopted it for 2009 only. The decision included a provision that an LSE that relies on a new resource that has not become commercially operational as of the date of its final annual local RA compliance showing shall, in such showing, (1) claim the entire new resource and (2) show a single local unit that it will show on every monthly filing to make up the capacity until the new unit has reached commercial operational status.
In the Phase 2 workshops, Energy Division, CAISO, SDG&E, and PG&E proposed to make the interim rule permanent. TURN and other workshop participants suggested that the requirement that a single substitute unit must be shown until the new unit is available is unnecessary and may have an unintended consequence of delaying the retirement of certain older units in some local areas. Instead, these parties believe that LSEs should be allowed to use multiple substitute units in their compliance showings. CAISO, DRA, NRG, PG&E, SDG&E, and TURN filed comments generally supporting continuing the interim rule for counting new resources. Four of these parties (DRA, PG&E, SDG&E, and TURN) proposed that the provisions for claiming the entire new unit and requiring a single substitute unit be eliminated or revised.
Discussion
We will make the interim rule for new resources permanent. As SDG&E points out, the local RA compliance demonstration is made only once a year. In the absence of this program refinement, new capacity could remain uncounted for up to 11 months, unnecessarily driving up costs. We will delete the provision that the LSE must claim the entire new resource and show a single local substitute unit. Allowing only one LSE to claim capacity from a new unit could undermine efficient trading of local resources by inhibiting the owner of the new resource from laying off excess capacity to a capacity-deficient LSE. Requiring that the substitute capacity come from a single resource in the local area would reduce the options available to the LSE for fulfilling its compliance obligation, which would further drive up costs
4.3. Cost-Allocation Methodology (CAM) Credit Allocations
In Phase 1 of this proceeding, AReM sought to change the quarterly allocation of CAM-related RA credits that was ordered by D.07-09-044 to a monthly allocation. During the Phase 1 workshops, most parties agreed that the existence of just one CAM contract in force at that time did not justify the administrative costs of a move to monthly allocations. The Phase 1 workshop discussions then centered on defining the threshold for determining when the change from quarterly to monthly allocations would be justified.
AReM proposed in Phase 1 that the trigger for changing to monthly allocations of RA credits be defined as the date that one additional CAM contract becomes operational. D.08-06-031 noted that this was a straightforward proposal that was not contested by any party, but declined to adopt it in light of unresolved workload issues noted by the Energy Division. The Commission deferred the issue to Phase 2 for resolution.
In the Phase 2 workshops the Energy Division proposed that the switch to a monthly allocation of CAM credit should be triggered when an individual service territory has two or more operational CAM contracts. Service territories with one operational CAM contract would continue to have quarterly reallocations of CAM credits. The Energy Division also proposed that if a reallocation would result in no change greater than 0.5 MW for any LSE, CAM credits would not be reallocated that month. In their filed comments, AReM, DRA, and SCE support the Energy Division proposal, and PG&E does not oppose it.
Discussion
We will adopt the Energy Division proposal for allocating CAM credits, as it strikes a reasonable balance between fairly allocating CAM credits and avoiding additional administrative burdens on the Energy Division.
4.4. Maximum Cumulative Capacity (MCC) Buckets
As explained in the Energy Division's Phase 2 workshop report, MCC "buckets" were established early in the RA program to ensure that LSEs do not over-rely on resources with limited availability to the point that CAISO would not be able to reliably operate the grid with RA resources.1 The buckets represent the maximum cumulative percentage of an LSE's procurement obligation that can be met with use-limited resources (ULRs) and RA contracts that provide less than "7x24" hours per week availability.2
In its workshop proposal, AReM recommended eliminating the requirement to categorize RA resources into MCC buckets. AReM noted several changes to the RA program that have occurred since the MCC bucket approach was established: the ability to count liquidated damages (LD) contracts has been phased out of the RA program, the CAISO tariff requires ULRs to submit monthly use plans, and NQC counting conventions for intermittent resources are under review. AReM contends that these changes justify the elimination of the MCC buckets and associated compliance checks. At the workshop, several parties expressed opposition to AReM's proposal, and no consensus was reached.
In its filed comments, AReM notes that MCC data provided by LSEs have not been used for any purpose except for checking that the data have been provided and comply with the rules. AReM notes in particular that the CAISO had originally pressed for the MCC bucket approach yet it does not use the data. AReM also disputes the view advanced by LSA that MCC buckets are necessary to recognize the capacity value of ULRs. Finally, AReM requests that if the MCC bucket approach is continued, a column be added to the NQC list to identify the MCC category attributable to each resource to ease the filing burden and reduce the potential for filing errors.
Along with AReM, DRA and NRG filed comments in support of eliminating the MCC bucket approach. CAISO, Dynegy, LSA, SCE, and SDG&E opposed the proposed elimination in their filed comments and/or reply comments. PG&E noted the need to coordinate the MCC bucket provision with the RA counting rules. PG&E would not support elimination of the MCC bucket approach if, as suggested by the Energy Division, such elimination would mean that energy contracts that call for delivery or provide dispatch rights during less than all hours of the year could no longer be used to comply with RA procurement obligations. TURN stated no strong preference with respect to this issue but noted that the CAISO's most recent draft Standard Capacity Product (SCP) does not include any reference to MCC buckets.
Discussion
In adopting the MCC approach, D.05-10-042 noted that it was proposed to alleviate over-reliance on ULRs that could not be counted on to serve a large portion of a month outside of the peak period. (D.05-10-042, at 44.) Thus, the MCC bucket approach can be seen as an important reliability measure. As SDG&E notes, if too many ULRs are included in the RA mix, there arises at least a theoretical possibility that the CAISO-operated system could become energy deficient, especially in years when imported hydro generation is low, weather is hotter than normal, and one or more nuclear plants have unexpected outages. The MCC bucket approach can also be seen as a cost-saving measure because it allows for the prudent use of ULRs to make up the RA fleet.
One alternative to the MCC bucket approach would be adopt the Energy Division's interpretation that elimination of the MCC buckets would mean, for example, that 6x16 energy contracts would no longer count for RA. Another alternative would be to accept RA compliance showings without regard to any use limitations on resources nominated by the LSEs in fulfillment of their capacity procurement obligations. We are concerned that the first alternative would unnecessarily exclude significant amounts of reliable capacity from the RA program and lead to excessive procurement costs, while the second alternative could lead to significant reliability concerns.3 At this time, we are not aware of any additional alternatives, although we note SDG&E's observation that a better tool designed by the CAISO may eventually replace the MCC buckets.
Nothing in the record of this proceeding assuages our concerns about eliminating the MCC bucket approach at this time. In particular, we do not find the phase-out of LD contracts, the fact that the CAISO tariff requires ULRs to submit monthly use plans, or the fact that the counting rule for intermittent resources is under review to be evidence that reliability concerns surrounding ULRs have been adequately resolved. The phase-out of LD contracts has no impact on resources with physical use limitations, the monthly plans submitted by ULRs are informational, and it is at best premature to base elimination of the MCC buckets on the current review of the counting rules for intermittent resources. Finally, the fact that the CAISO has not made direct use of the MCC data contained in the LSE filings may be somewhat puzzling but it does not mean that our decision to limit reliance on ULRs for RA purposes lacks important reliability benefits. Accordingly, we will leave the MCC bucket requirement in place.
We find AReM's request that a column be added to the NQC list to identify the MCC category attributable to each resource to be a potentially worthwhile suggestion that could improve the RA compliance filing process. We therefore urge the Energy Division and the parties to explore with CAISO whether this suggestion is feasible and, if it is, how it might be implemented.
TURN's observation that CAISO's draft SCP omits any reference to MCC buckets raises an issue that needs to be considered in further development of the SCP and/or RA program modifications that may be needed to coordinate the RA program with the SCP tariff.
4.5. Demand Response (DR) Resources
In accordance with earlier RA decisions, dispatchable DR resources are allowed to count for purposes of RA compliance.4 However, DR resources are not treated in the same manner as generation resources. Among other things, they do not currently appear on the NQC list. As part of the annual RA compliance reporting cycle, Energy Division staff performs the ministerial function of calculating DR program impacts and allocating DR capacity credits to all LSEs pursuant to Commission-adopted policy.
D.08-04-050 adopted load impact (LI) protocols for DR resources, and Energy Division staff proposed to use the adopted protocols as RA counting rules for DR capacity. Staff also proposed to list DR programs on the NQC list for information only. This proposal would apply both to DR programs controlled by IOUs and non-IOU controlled DR programs. For those IOU DR programs for which funding has been approved by the Commission, staff would use the underlying load impacts associated with that approval. For non-IOU DR programs that do not involve Commission approval, the LI protocols would serve as NQC counting rules, and the DR program operator along with the LSE that wishes to count the program would provide information comparable to the protocols. Energy Division would then determine the NQC of the program. Finally, staff proposed to use other protocols if appropriate. At the workshops, staff clarified that the intent is to use the adopted LI protocols to the greatest extent possible. Alternative protocols would only be used for new DR programs that do not have Commission-approved LI protocols.
With staff's clarification that it would use the LI protocols to the greatest extent possible, DRA, PG&E, SCE, and SDG&E support the staff's proposal. AReM agrees that those seeking RA credit for DR programs should be obligated to provide information to verify RA capacity, and that there should be a reasonable standard for measurement and verification that all DR programs should meet. Nevertheless, AReM has serious concerns about application of the LI protocols to non-IOU DR programs. As AReM sees it, the LI protocols were developed to determine the cost-effectiveness of IOU DR programs and for long-term resource planning, and are not appropriate for evaluating non-IOU DR programs. AReM contends that cost-effectiveness is unrelated to independently-funded DR programs. Moreover, according to AReM, the LI protocols would be burdensome and expensive obligations that would discourage electric service providers (ESPs) from developing their own programs. AReM also notes that the CAISO has begun a stakeholder process that will address measurement and verification requirements for DR resources, and suggests that those requirements will provide appropriate standards for determining the NQC of RA capacity for non-IOU DR programs. In light of the foregoing, AReM proposes flexible compliance that would allow each non-IOU DR program provider to submit a proposal for counting the RA capacity of its DR program that complies with the measurement and verification requirements of the CAISO. EnerNOC filed reply comments stating concerns similar to those raised by AReM.
Discussion
We are intrigued that the CAISO is developing tariff standards for measurement and verification of DR resource performance, and we look forward to an opportunity to evaluate whether such standards, once adopted, would be appropriate for use in the RA program. However, we cannot commit prospectively to using such standards. At this time, staff's proposal to use the LI protocols for assessing DR impacts provides a means of evaluating DR programs with the use of a defined, uniform standard for all programs, and we therefore approve it. We note that AReM agrees there should be a standard "that all such programs must meet to obtain an RA credit." (AReM Comments, at 3.) Yet, AReM would have the LI protocols apply to IOU DR programs but not other DR programs.
In their discussion of D.08-04-050, AReM and EnerNOC emphasize the role that cost-effectiveness analysis plays with respect to the LI protocols. It is true, as AReM observes, that the Commission found that LI estimates are necessary for analysis of the cost-effectiveness of DR programs and for long-term resource planning. (D.08-04-050, Finding of Fact 1.) However the Commission also found that LI protocols improve consistency and accuracy in the calculation of DR load impact estimates. (Id., Finding of Fact 2.) We believe that improved consistency and accuracy in the calculation of DR load impact estimates would benefit the RA program by giving appropriate weight to the capacity value of these resources. Moreover, D.08-04-050 explicitly stated that evaluations based on the LI protocols may be of use in other proceedings, including RA proceedings. (D.04-08-050, at 28. Further, it directed the IOUs to use the protocols in estimating DR impacts for RA purposes unless directed otherwise by the Administrative Law Judge (ALJ) or assigned Commissioner in the relevant proceeding. (Id., Ordering Paragraph 5.)
We are mindful of AReM's concern that the application of the LI protocols could be burdensome and expensive for DR providers and ESPs, just as we presume it would be for IOUs. As noted above, the CAISO's development of a measurement and verification standard may turn out to be a promising approach that could be used for the RA program. For now, however, the alternative to adopting a clear and defined standard (such as the LI protocols) that can be delegated to the Energy Division for implementation as a ministerial function would be to conduct formal Commission proceedings to review and determine the load impacts of DR programs. This could raise timing problems in the administration of the RA compliance filing cycle.
We do find that it would be appropriate for Energy Division to convene an educational workshop and/or publish guidelines to describe and explain the process and criteria that will be used to apply the adopted LI protocols and we direct that it do so. Finally, we concur with SDG&E's recommendation that the Energy Division should provide notice in the appropriate RA proceeding whenever it seeks to alter application of the LI protocols in determining the NQC for a particular DR resource.
PG&E and TURN jointly submitted a workshop proposal to address what they see as a lack of transparency in the DR allocation process. They initially recommended that the Commission publish an explicit accounting of how the RA megawatts associated with each DR program are allocated to specific LSEs at the same time that LSEs are notified of their respective allocations. The workshop discussions yielded suggestions that DR credit allocations should be provided to LSEs on a program-specific basis and that draft allocations should be shared with LSEs before final allocations are provided. This would allow LSEs to raise any concerns about the allocation with staff. Workshop participants raised a confidentiality concern, and PG&E revised its proposal to provide that the program-specific and total DR allocations to an LSE should only be shared with that LSE.
Staff suggested in its post-workshop report that publishing the total qualifying capacity of DR programs on a program-specific basis would increase transparency in a way consistent with the parties' suggestion. Staff also suggested that DR allocations could include the distribution area coincident peak load-share percentage used for the LSE and a list of the DR programs allocated using that load-share. Finally, staff noted that providing draft allocations, conferring with LSEs, and then issuing final allocations could delay the date of final allocations.
Parties generally supported the need for greater transparency in their filed comments. AReM reiterated the workshop discussion regarding the need to respect confidentiality of certain data, and would oppose the public posting of LSE-specific allocations. AReM believes that the total RA credits assigned to each IOU DR program should be publicly available. PG&E proposed that the Commission provide clear justification for any reductions made in the RA capacity associated with any specific DR program
With respect to the proposal that staff provide preliminary allocation information to LSEs and provide opportunity for discussion before final allocations are assigned, AReM notes that specific DR allocations are provided to individual LSEs relatively late in the RA compliance cycle. AReM therefore opposes the preliminary review process if it would cause any additional delay. PG&E supports providing LSEs an opportunity to review and discuss preliminary allocations with the expectation and understanding that any delay in making the final assignment would be modest. SCE would support such a process if it would not delay the DR allocations beyond the historical release date of early July, while SDG&E believes the possible delay may be worth the benefit. TURN also believes that the Commission should err on the side of caution and provide for preliminary review, and notes that this could avoid disputes that could lead to even greater delay.
Discussion
To promote fairness and confidence in the RA program, the DR capacity credit allocation process should be transparent to the maximum extent consistent with Commission policy regarding confidentiality of electric procurement data made in D.06-06-066 and subsequent decisions in the underlying rulemaking (R.05-06-040). This includes making public information about the process and criteria used by the staff to administer the program as well as any actual data used in the allocation process where such disclosure would not reveal market-sensitive information.
The PG&E/TURN transparency proposal, clarified to provide that LSE-specific allocations and supporting information should be provided only to the LSE and not made public, would provide greater understanding and should be adopted. Staff's suggestion to provide individual LSEs with the distribution area coincident peak load-share percentage used for the LSE and a list of the DR programs allocated using that load-share is consistent with this approach and is therefore also approved. Staff's proposal to publish the total RA capacity associated each DR program (which AReM supports) would likewise promote transparency and is therefore approved. LSE-specific allocations and the total for each program type should be disclosed to the LSE at the level of aggregated local areas.
Balancing the need for greater transparency and accuracy in allocations, on the one hand, and the need for timely assignment of final DR capacity credit allocations, on the other hand, we find that limited provision should be made for preliminary notice to LSEs of assignments of credits. As SCE points out, the staff has targeted the assignment date in early July of the applicable compliance year. We believe a modest extension, not to exceed 15 days, could be added to the DR credit assignment process to accommodate an opportunity for expedited review based on preliminary assignment notices. Authority to grant such an extension is appropriately delegated to the Energy Division.
Energy Division staff assigns DR program capacity credits based on the established principle that DR impacts should be allocated to LSEs in proportion to the funding that their respective customers provide toward DR programs. (D.05-10-042 at 38.) PG&E and TURN submitted a workshop proposal that the credits associated with IOU DR programs whose costs are recovered in Energy Resource Recovery Account (ERRA) should be allocated exclusively to the IOUs that administer them. The reasoning is that since DR programs whose costs are recovered through the ERRA are paid by bundled service customers exclusively, the same customers should receive the RA benefits of those programs.
AReM proposed that DR credits should be allocated to LSEs based on the shares of bundled and direct access (DA) participants in an IOU's DR program. The IOU would receive credit for the share of bundled participants and ESPs would receive the DA customer share. Among ESPs, credit would be allocated on a load-share basis. AReM notes that commercial and industrial customers are the predominant participants in DR programs and that DA penetration is higher in these customer classes. AReM argues that these customers are disadvantaged by the current method. Several workshop participants responded that AReM's proposal is inequitable because customers that enroll in DR programs are paid for their participation in the program.
AReM reiterated its proposal in its filed comments. CLECA, DRA, PG&E, SCE, SDG&E, and TURN argued for retention of the "who pays" allocation principle.
Discussion
We affirm the established principle that DR program capacity credits should be allocated to LSEs in proportion to the funding that their respective customers provide toward DR programs. The proposed alternative of basing the allocation on relative participation rates of bundled and DA customers in a DR program fails to account for the fact that customers decide to enroll in DR programs because of the direct benefits of doing so. Since bundled service ratepayers generally provide funding for those DR program benefits, they effectively procure DR capacity. It would be inequitable to bundled service customers to assign DR capacity credits to LSEs on the basis of who participates in the DR program, without regard to how it is funded.
The PG&E/TURN proposal to allocate DR credits associated with IOU DR programs whose costs are recovered in ERRA exclusively to the IOUs that administer them, along with PG&E's clarification that credits for DR programs whose costs are recovered through distribution rates should be allocated on a load share basis, are consistent with our adopted allocation principle, reflect current practice, and are hereby affirmed.
4.6. Qualifying Facility (QF) Outage Counting
D.06-07-031 adopted a protocol for determining how the NQC of resources with scheduled outages should be counted. In Phase 1 of this proceeding, PG&E raised a concern that the protocol results in scheduled outages being counted twice in assessing the RA value of certain resources, such as QFs, that utilize historic performance as the basis for setting their NQC. As PG&E explained in Phase 1, the initial NQC calculation for these resources reflects their reduced generation during scheduled outages taken in the three-year historic averaging period. The scheduled outages of these units are applied a second time to reduce their RA counting value under the protocol. The Phase 1 decision (D.08-06-031) found that the double counting of outages for these resources should be corrected to avoid unnecessary procurement, but did not find that the proposed solution was ready for adoption. Instead, it deferred this topic to Phase 2.
CAISO and the three IOUs (PG&E, SCE, and SDG&E) jointly submitted a workshop proposal with a method to remove scheduled outages from the data used to calculate the NQC of non-dispatchable QF units whose NQC is currently based on a three-year rolling average of energy production. Under this proposal, CAISO would provide data on historical outages subject to the scheduled outage counting criterion to the California Energy Commission (CEC). The CEC would then substitute proxy data for the hours of the scheduled outage. This proxy data would be calculated by averaging the same hours for the other two years of data used in the overall NQC calculation. The NQC calculation would then be completed based on the current counting rules. This proposal would not modify the scheduled outage counting criterion.
At the workshop, staff suggested that it would be reasonable to include all units that are subject to the three-year rolling average counting convention within this proposal. Staff does not believe that the contract type (i.e., QF or renewable) justifies different treatment in this regard. CAISO and IOU representatives tentatively agreed to this modification during the workshop discussions.
The CAISO and the IOUs affirmed their support for the proposal in their filed comments. However, the CAISO conditioned its support on a Commission decision to retain the existing replacement rule for scheduled outages set forth in D.06-07-031, Section 3.1. DRA and TURN also support the proposal. No party objects to the Energy Division proposal to extend the proposed method to any resource type for which a rolling average is used to calculate NQC values, and the proposed extension is supported by DRA, PG&E, SCE, SDG&E, and TURN.
Discussion
With the clarification that it should apply to all resource types whose NQC is calculated using a rolling average, not just QF resources, the proposed "Historical Output Correction" method for correcting the double counting of outages fairly and adequately resolves our concern that such double counting could lead to unnecessary procurement. We therefore adopt it. We make this determination irrespective of the CAISO's stated condition for its support. As TURN points out, we are revising the counting rule for resources whose NQC is based on a rolling average in order to treat such resources more consistently with resources whose NQC is not de-rated for past scheduled outages. Whether to modify the replacement requirement is a separate issue.
4.7. Load Forecasting Issues
D.05-10-042 confirmed the previously established "best estimate" approach in lieu of the "current customer" approach to year-ahead load forecasting. It also indicated a willingness to revisit the determination at an appropriate time.5 The Phase 2 Scoping Memo noted that parties have continued to express concern that some LSEs may systematically under-forecast their load using the best estimate approach, and invited proposed solutions. It also provided that parties making such proposals should show that the conditions for revisiting the topic set forth in Section 6.1 of D.05-10-042 have been met.
AReM submitted a workshop proposal to continue the best estimate approach unless there is evidence of significant under-forecasting by LSEs. AReM observed that even if there is under-forecasting, staff has enforcement authority to seek penalties for the offending LSE. AReM also noted that the conditions for revisiting the issue set forth in D.05-10-042 have not yet been met.
PG&E and TURN submitted a joint workshop proposal to adopt the current customer method. They argue that there is systematic and significant under-forecasting in the year-ahead LSE forecasts. Further, they contended that a current-customer year-ahead forecast would ease the burden of a monthly local true up (see Section 4.7.2).
In its post-workshop comments, AReM reiterated its contention that the conditions stated in D.05-10-042 for revisiting the best estimate versus current customer issue have not been realized. In particular, AReM submits that there is no liquid capacity market in place, either bilateral or centralized. Moreover, according to AReM, LSEs currently have no reasonable and cost-effective means to adjust their local RA portfolios after the year-ahead RA compliance filing is submitted. AReM believes that as the Standard Capacity Product (SCP; see Section 4.8) is adopted and a more liquid capacity market structure is put in place, a review of the forecasting method may be appropriate.
SCE's filed comments concur with AReM that the stated conditions for revisiting the year-ahead forecasting method have not been met, and NRG's comments in effect do so. However, SCE goes on to urge that the Commission use its discretion to disregard the conditions if new circumstances warrant doing so. TURN on the other hand contends that the conditions have been met because LSEs and suppliers have developed a bilateral market for the exchange of RA capacity, and even the smaller LSEs have been able to fulfill their RA obligations on a consistent basis. Apparently agreeing in part with AReM regarding the importance of the SCP, TURN notes that the CAISO has proposed to implement the SCP for the 2010 compliance year, which TURN believes will greatly facilitate the trading of RA capacity. Finally, TURN its reiterates concern about the under-forecasting issue. TURN notes an estimate by the CEC that the gap between LSEs' forecasts and actual load has been in the range of 500 MW.
Discussion
We find that the conditions specified in D.05-10-042 for reviewing whether to replace the best estimate with the current customer method have not been met. It is true that the Commission did not specify that the capacity market would need to be "centralized," but the context of the passage in D.05-10-042 where the Commission described the conditions makes it apparent that the Commission believed that something more conducive to trading than the current bilateral market environment would be needed.
However, we do not wish to overemphasize the "letter of the law" with respect to the preconditions that D.05-10-042 established for replacing the best estimate with the current customer method. As SCE points out, we could exercise discretion to waive the conditions. More important is the underlying principle. The Commission clearly did not want to place LSEs in a position where they could be saddled with excess capacity, or in need of additional capacity, under market conditions where they would not be able to conduct reasonable and appropriate transactions to acquire or dispose of capacity as needed for load migration.
Despite the evolution of the RA program since D.05-10-042 was issued, we are not persuaded that it is time to change the forecast method. The fact that LSEs have been able to meet their year-ahead RA obligations does not provide assurance that, over the course of the compliance year, the market would be sufficiently liquid to accommodate transactions associated with load migration. Most significantly in this regard, the CAISO's SCP tariff proposal has not yet been implemented. Further, as explained in Section 4.8 of this decision, whether the SCP tariff will be implemented in time for the 2010 compliance year is an open question. Given the importance that most parties ascribe to having the SCP in place to facilitate capacity trading, we view the successful implementation of the CAISO's SCP tariff provisions as a necessary condition for adoption of the current customer approach. Accordingly, we decline to change the year-ahead forecast method to a current customer approach for the 2010 compliance year.
To the extent that under-forecasting by some LSEs continues in practice, we are concerned with the potential for cost-shifting from those LSEs that under-forecast to LSEs who more accurately forecast their loads.6 The current customer method could provide incentives for greater forecasting accuracy by focusing attention on the likely loads of existing customers. In addition, it would facilitate measures to accommodate load migration in connection with the Local RA program component (see Section 4.7.2). However, this focus may be limited to migration between ESPs, and may not apply to utilities. On balance, we do not find that the concerns regarding under-forecasting outweigh concerns about the impact of market illiquidity at this time.
After the market liquidity issue has been satisfactorily mitigated with the implementation of the SCP tariff, along with any RA program refinements that may be necessary to coordinate with the SCP tariff, it will be appropriate to further evaluate whether to convert year-ahead forecasts to the current customer method. Assuming that the SCP tariff has been approved and implemented before the record of the proceeding establishing RA requirements for the 2011 compliance year is closed, the conversion could take place for that year. SCE makes two additional proposals: (1) expected load growth or reduction for current customer load forecasting should be determined on a system-wide basis by the CEC and calculated as a percentage factor (positive or negative as weather, economic, and other relevant conditions warrant) that LSEs must apply to their prior year's peak load forecast, and (2) the detailed rules necessary for implementation of the current customer approach, including, for example, the appropriate date for determining LSEs' current customer counts and the process for determining the annual load growth/reduction percentage factor shall be developed in workshop jointly conducted by the Energy Division and CEC staff in a subsequent proceeding. (SCE reply comments at 3-4.) While we are not inclined to adopt substantive proposals that first appeared in reply comments, we generally concur with the procedural aspects of SCE's recommendation and believe it is necessary to include these elements in our review.
Finally, it is unclear what changes in administrative burdens for Energy Division and CEC staff are implied by this shift in forecast methodology. We need to be sensitive to such impacts, particularly since a substantial portion of the burden will fall upon the CEC. We welcome the advice of the CEC as well as our own staff in future proceedings on this topic so that we can give appropriate consideration to staffing needs.
The RA program provides for monthly true-ups of system RA obligations to account for load migration but does not allow such adjustments for local RA obligations. This has certain adverse effects, as described below. In D.07-06-029, the Commission stated that it remained open to considering a mechanism that would true-up local obligations. D.08-06-031 referred this topic to Phase 2 of this proceeding.
Workshop proposals to address the load migration issue were submitted by SES and jointly by PG&E and TURN. SES describes its view of the problem as follows:
"Under the current Year-Ahead Local RA model, Local RA is procured annually; there is no obligation to procure additional Local RA capacity if, during the course of the compliance year, an LSE acquires load. Nor are there any opportunities to re-sell Local RA capacity if an LSE loses load. As a consequence, there is no `market' for the value of Local RA capacity after the year-ahead showing. This results in Local RA capacity losing its local premium value, because if Local RA capacity is resold after the Year-Ahead Local RA showing, it will be valued by the market at the system price for RA. For many LSEs this phenomenon imperils their future viability. From the beginning of the RA program, this risk was identified as an issue that needs to be addressed, for it creates huge regulatory-imposed financial risks. Moreover, it gives any LSE new to the State which acquires customers from an existing LSE an unfair cost advantage, at least for the first year of operation. Allowing LSEs to true-up their Local RA obligations, as they currently do for System RA capacity, may help minimize this financial exposure." (Phase 2 proposal of SES at 2.)
In its pre-workshop submittal, SES proposed to permit LSEs to assign local RA capacity obligations to each end-use customer that migrates. The obligation would be based on the ratio of the customer's August peak demand to the LSE's total August peak demand for all the LSE's customers in a local area. That ratio would be multiplied by the LSE's local capacity obligation for the local area to determine each end-use customer's obligation. For at least the first year of the proposal, the customer-specific obligation would be waived for customers with less than 1,000 kilowatts (kW) August peak demand. As customers migrate from one LSE to another, the losing LSE would report the migrating customer account and the accompanying capacity obligation to the Energy Division. The Energy Division would confirm the release of the capacity obligation on the part of the losing LSE and impose the corresponding obligation on the gaining LSE. SES believes that the number of migrating customers would be low, and therefore does not expect that this process would impose an undue administrative burden on the Commission. SES proposed that the current $40 per kW-year trigger price for local RA capacity remain in effect, and that the utilities be directed to make excess local RA capacity available for purchase. Finally, SES proposed to implement this process on a pilot basis for 2010, assuming that the SCP is implemented during 2009 in time for its use in 2010 procurement.
Responding to concerns raised in the workshop discussions, SES revised its proposal in several respects.7 First, SES proposes that the assignment of local capacity obligations be limited solely to customers with demand meters. Second, in response to questions about how to manage potential disputes between LSEs about a migrating customer's peak load, SES proposes that each LSE show in its year-ahead local RA compliance submittal the customer-specific August peak load and associated local RA obligation for the lagging year (i.e., August 2009 peak for the 2010 compliance showing submitted in October 2009). Third, utilities would be encouraged rather than directed to offer excess local capacity.
PG&E and TURN submitted a pre-workshop proposal to keep the current year-ahead compliance process and allow trading of the obligation for month-ahead demonstrations based on a migrating customer's August usage. This proposal was made in connection with the PG&E/TURN proposal to convert to the current customer method for determining year-ahead procurement obligations. Like SES's revised proposal, the PG&E/TURN proposal does not require the utilities to offer the sale of excess capacity. In its post-workshop comments, PG&E proposed to require month-ahead compliance filings for local RA, similar to system RA monthly filings.
Some workshop participants raised the possibility of unbundling the "local attribute" from an RA contract, so that only the local attribute would be traded from month to month due to load migration. Energy Division suggested that since local RA obligations are allocated to LSEs based on the share of utility area coincident peak, the customer-specific capacity obligation should likewise be based on the coincident peak.
AReM, PG&E, and TURN support adoption of the revised SES proposal, and DRA and SDG&E state their support for a monthly true-up mechanism. The support is generally conditional, however. AReM believes an additional workshop is needed to explore implementation details. Also, AReM notes that some ESPs are concerned about the risk of taking on an additional procurement obligation before a liquid capacity market is available. AReM's support is contingent upon the implementation of the SCP tariff. DRA believes the true-up mechanism would be facilitated by development of the SCP and by allowing trades between long and short LSEs. PG&E believes that the specific language in the SES proposal that defines the "peak-to-load" ratio requires clarification. PG&E also supports several technical clarifications suggested by TURN at the workshops. SDG&E's support for a true-up mechanism is dependent upon use of the current customer approach to year-ahead forecasts and the institution of the ability to trade the local RA attribute of system RA resources. TURN would modify the SES proposal to base the customer-specific local RA obligation on its forecasted August coincident peak demand for the RA compliance year rather than the recorded peak of the prior year. TURN also believes that the true-up mechanism would function much more smoothly with the current customer approach rather than the best estimate approach.
SCE believes that the proposals for true-ups during the compliance year are inappropriate. Among other things, SCE is concerned that LSEs would have additional procurement requirements imposed on them during the course of the compliance year. SCE believes this would be improper to the extent that the Commission retains the best estimate forecasting method, since that method does not necessarily ensure that all customers are accounted for. SCE is also concerned about having to procure capacity pursuant to a monthly true-up obligation because of the extremely short time frame allowed for such procurement and because there would be an obligation for the gaining LSE to buy but no corresponding obligation for the losing LSE to sell the capacity associated with the migrating customer. Finally, SCE sees the true-up proposals as unnecessary because customer migration is already accounted for in the system RA program, which does allow true-ups for migration. Since a losing LSE can sell excess system RA capacity, the actual financial impact is that the losing LSE bears, for no more than one year, the difference in price between system and local RA capacity acquired to serve the lost customer.
Discussion
Local RA procurement obligations are currently established annually for a 12-month compliance period. Thus, when an LSE loses a customer to another LSE during the compliance period, it temporarily remains saddled with Local RA procurement costs associated with that customer.8 At the same time, the LSE that gains the migrating customer has no obligation to procure capacity on behalf of that customer for the remainder of the compliance year. This has the effect of shifting costs to the losing LSE, which runs counter to our policy, and the requirements of Section 380(b)(2), to equitably allocate the cost of generation and prevent cost shifting.9 Among other things, as SES notes, this could provide an unjustified competitive edge to new LSE entrants. SCE argues that this situation is mitigated by the losing LSE's ability to sell system RA capacity on a month-to-month basis, but that is not a complete solution.
This issue has lingered since the Local RA program began, and finding a solution that fairly and effectively resolves the load migration problem without creating problematic new issues has been elusive. After considerable effort over several RA proceedings, SES has presented a proposed mechanism that goes a long ways towards a full solution. By limiting monthly true-ups to instances of documented load migration from one LSE to another, this approach appears less administratively burdensome for LSEs than the proposed alternative of requiring monthly compliance filings for local RA.
We adopt the core principle of the revised SES proposal as set forth in Appendix A, i.e., limiting local RA adjustments to documented load migration, but go no further at this time. In light of the potential benefits of such a process, as well as the time it has taken to get to this point, it is with considerable reluctance that we defer implementation of such an approach to the 2011 compliance period. We find we must do so for the following reasons:
· Several parties, notably including SES, emphasize the importance of implementing the SCP tariff in connection with the true-up proposal. We concur. We declined to adopt the current customer method for year-ahead load forecasts for several reasons, including the fact that the SCP tariff has not yet been implemented. We did so to provide greater assurance of a more liquid market environment in light of procurement obligations that could unduly impact LSEs. Similarly, since the SES true-up proposal imposes new procurement obligations that must be met in a limited time frame during the compliance year, it is necessary and appropriate to provide that the SCP must be in place before implementing the true-up process.
· We believe it is necessary to explore whether basing the quantity of load that has migrated on the customer's historic peak load for the previous year would be any more accurate than using the forecast of load for the customer. A known value has attractive qualities, but it is not necessarily the most accurate value. A forecast developed closer to the point that migration occurs may be more accurate.
· We concur with AReM that additional workshops are needed to resolve technical issues. Examples of details remaining to be worked out include clarification of the "peak-to-load" ratio, whether to aggregate local areas for simplicity, and whether to use actual or forecast ratios. We note that TURN offered several technical suggestions and commend them to the attention of the workshop participants. Also, the workshops would be needed to address any proposal for unbundling the local attribute of RA capacity to accommodate migration. In this regard, we concur with SCE that it would be improper to adopt a major concept that was first suggested in workshops and not properly proposed according to established procedures.
· We remain concerned that administrative burden of processing LSE's migration claims on our staff may not be as negligible as SES anticipates. Deferring implementation has the advantage of providing our staff an opportunity to develop appropriate procedures and to make necessary staffing assignments.
The SES proposal provides a strong foundation for development of a load migration mechanism that we intend to adopt for 2011. We provide the following guidance. First, we concur with parties who claim it would be improper to require or encourage utilities that are long on local RA capacity to sell that capacity to other LSEs, but not impose the same provision on other LSEs. One of the guiding principles of the RA program, and a requirement of Section 380(e), is the nondiscriminatory imposition of the same requirements on all LSEs. We see no basis for disparate treatment here. Also, in light of our concern about the potential administrative burden of this mechanism, it may be appropriate, at least for the first year, to limit the tracking of migrating customers to those customers with peak demands of, for example, 3 MW.
In comments on the proposed decision, SCE reiterated its opposition to a monthly true-up mechanism for local procurement obligations to account for load migration, and it proposed several issues that it believes must be resolved prior to implementation of such a mechanism. Since we are providing for further proceedings on this issue, SCE may introduce its position on unresolved issues in those proceedings. We do not find it necessary to further specify the scope of those proceedings at this time. We also note, as SES observes, that SCE's concerns about market power may be addressed, in whole or in part, by the market power mitigation measures adopted in D.06-06-064.
4.8. Standard Capacity Product (SCP)
At the request of stakeholders over an extended period, and with the encouragement of this Commission in prior orders, the CAISO staff has developed a draft SCP with availability requirements and incentives for RA resources that would enable more efficient transactions of RA capacity. The CAISO distributed a draft SCP document for discussion in the Phase 2 workshops. CAISO staff indicated that it intended to seek approval of its Board of Governors in February 2009 and file an SCP tariff at the Federal Energy Regulatory Commission (FERC) in March 2009. In its opening comments, the CAISO indicated it planned to file the SCP tariff with the FERC in April 2009. At the time of the workshops, CAISO staff was still developing provisions of the SCP proposal dealing with contract grandfathering, unit substitution, and clarifications to the proposal.
The CAISO staff summarized the draft SCP proposal as follows:
· "Availability Standard. If a resource receives payments for providing RA capacity, there is an expectation that the full RA capacity of that resource will be available to the CAISO, i.e., the resource is not on a forced equipment outage or derate that diminishes its ability to provide the full amount of its RA capacity. Under the SCP, hourly resource availability will be tracked on a monthly basis and compared against a single availability standard or target based on the historic performance of the RA resource fleet during the peak hours of each month of the previous year."
· "Availability Incentives. The SCP proposal will provide incentives for each resource to meet or exceed the target availability standard. On a monthly basis the CAISO will assess financial penalties to resources whose availability falls short of the target, and will provide bonus payments to resources whose availability exceeds the target. Bonus payments will be funded only through available financial penalty revenues. This will ensure that the mechanism is revenue neutral on a monthly basis and does not depend on revenues from other sources."
· "Unit Substitution. A resource owner will be able to substitute a non-RA resource for an RA resource on forced outage in order to avoid the outage being counted against the RA resource's availability. A pre-approval process will be required to ensure that the replacement capacity is comparable to the original RA capacity in an operational sense."
· "Transition to SCP. The SCP has provisions for the grandfathering of existing RA contracts that have availability standards and incentives comparable to those specified in the SCP tariff language. Such grandfathered contracts would be exempt from the CAISO-enforced availability standards and incentives under the SCP. Upon the expiration of such contracts, any grandfathering would cease."
· "Deferment of SCP availability standards and incentives for certain RA resource types. The CAISO proposal would not initially apply the SCP availability provisions to intermittent renewable generation (wind and solar), [QFs], and demand response resources. The CAISO intends to revisit the applicability of the SCP provisions to these resource types at a later date."
Since FERC approval of the SCP tariff was not obtained sufficiently in advance of the close of the Phase 2 record, parties have not had an opportunity to comment on the final SCP provisions and whether and how this Commission's RA program might need to be modified in light of them. At this time, we are hopeful that such FERC approval and opportunity for comment can be realized so that, if found to be appropriate, the SCP can be fully integrated into our RA program for the 2010 compliance year. Accordingly, we will leave this proceeding open for a limited time, and for the limited purpose of addressing SCP implementation issues that include whether and to what extent the final SCP should be required for RA compliance and whether the existing replacement requirement of the scheduled outage counting protocol should be eliminated if the SCP is implemented. We clarify that while we defer action on mandating the SCP for RA compliance, contracts that include the final SCP provisions will be eligible to count for RA compliance in 2010.
We also adopt SCE's proposal to maintain a three-month interval between a Commission decision on SCP issues and the date for LSE compliance showings. As SCE observes, it would be poor policy to require parties to assume the outcome of the SCP process and, in effect, begin implementing a program still being conceptualized. Thus, for example, if FERC approval of the SCP tariff occurs in May 2009 and it is then possible to conduct an expedited comment process that concludes by mid-June 2009, then it might be possible for the Commission to act on SCP issues at its scheduled July 30, 2009 meeting. In the event it is not possible to conclude the process with a final decision by July 30, 2009, then the SCP implementation issues would be addressed in a future RA proceeding, and the SCP would be fully implemented with the 2011 compliance year.
4.9. Ancillary Services (AS) Must Offer Obligation (MOO)
Under the CAISO's Market Redesign and Technology Upgrade (MRTU) tariff, RA resources (except units in an outage and certain ULRs) have an obligation to submit in the Integrated Forward Market (IFM) either self-schedules or economic bids for all their RA capacity. This obligation is referred to as the RA MOO. The CAISO staff presented a workshop proposal explaining its plan to file a proposed tariff with the FERC to add an AS MOO to the existing RA MOO. Although related to the draft SCP tariff, it is a separable proposal. Under CAISO's AS MOO proposal, generators who are certified to provide AS must bid the AS capacity into the CAISO's IFM in addition to making an energy bid. CAISO states that the AS MOO is needed so that the MRTU markets can "co-optimize" energy and AS bids and to prevent withholding of AS capacity. CAISO staff clarified that they do not intend this proposal to change the way that LSEs contract for RA. In particular, CAISO is not proposing that there be a requirement that LSEs, either individually or collectively, procure a certain amount of AS certified resources for RA. The CAISO seeks a statement of support for its AS MOO proposal.
DRA, SCE and TURN support the CAISO's proposed AS MOO. PG&E supports it as long as the exemption for ULRs such as hydro is maintained. NRG is concerned about the equity of the AS MOO because it exempts self-scheduled hydro resources that are otherwise certified to provide AS.
Discussion
The CAISO states the approval of the AS MOO tariff will not impose any additional burdens or costs on LSEs. By requiring suppliers to bid into the AS market, the CAISO will be able to better optimize the resources its selects, which should result in lower costs for ratepayers. We therefore generally support this proposal. However, we note that the record was not adequately developed to enable us to weigh the substantive merits of NRG's or PG&E's concerns regarding the exemption of certain hydro resources from the AS MOO. We further note that the record does not enable us to weigh the concern of the Independent Energy Producers Association that the proposed tariff could create uninterested incentives and decrease the overall supply of ancillary services.
4.10. Qualifying Capacity (QC) for Intermittent Resources
Qualifying Capacity (QC) represents the gross amount of a resource's capacity, prior to an adjustment for deliverability, that can be counted for meeting the Commission's RA procurement obligation. Net qualifying capacity (NQC) is the amount of a resource's capacity that can actually be counted for RA compliance filings. For intermittent resources, including wind, solar, biomass, and as-available cogeneration, QC values are calculated for each month of the year based on averages of historic production performance data.
In recent years, concerns have arisen that the averaging method may not be appropriate for determining the NQC of wind and solar resources. In its 2007 Resource Adequacy Report (April 15, 2008), the Energy Division provided data showing that the current method overstates the available capacity of wind resources during peak demand periods. Energy Division observed that:
"...[D]aily production deviates broadly, in both directions, from the established NQC." (2007 Resource Adequacy Report, at 20.)
"Wind production is extremely variable." (Id.)
"...[W]ind production is negatively correlated with CAISO system load and prices in both zones (North of Path 15 (NP 15) and South of Path 15 (SP 15)) during the summer months, indicating that wind production is generally lower during the periods of high prices and high demand. (Id. at 23-24.)
"Wind production at super-peak hours very often falls below NQC. Figure 8 shows that in only one of the twenty hours of highest load during the summer of 2007 did the actual hourly wind production [exceed] NQC." (Id. at 24.)
In view of such concerns, a review of the counting rule for intermittent resources was taken up in Phase 1, where the issue was referred to Phase 2. While much of the focus has been on existing wind resources, we also consider the intermittent resource counting rule as it applies to solar resources. As the Energy Division notes, many large new solar resources are expected in California in the next several years, and the counting rules for new solar units warrant discussion.
Workshop proposals regarding the counting rule for intermittent resources were submitted by CAISO, SCE, and SDG&E (Joint Proponents; see Appendix B); CalWEA and AWEA; DRA; Dynegy; LSA; and PG&E. Along with these parties, NRG, SA, Sempra Generation, and TURN addressed this issue in their post-workshop comments.
As the staff workshop report explains, the proposals for counting the QC of intermittent resources can generally be grouped into three categories:
Historical Average (current method): Either a straight or weighted average of historical data typically from a specific set of "important" hours. CalWEA and AWEA proposed to continue this approach. LSA also proposed to maintain the current counting method but suggested another approach for intermittent resources, described below. PG&E proposed averaging the production of intermittent generation during the top ten load hours from each month.
Historical Exceedance: Usually in a set of important hours, this method uses a percentile to estimate how much generation is available for some percent of the time. For example, what quantity of generation was exceeded during 80% of the important hours? Exceedance approaches recognize an important diversity effect; an exceedance of multiple, different resources will generally have a higher exceedance than the sum of the exceedances of the individual resources.
Dynegy presented an exceedance approach with an exceedance factor of one minus the forced outage rate of thermal RA generators. Joint Proponents presented an exceedance proposal for both wind and solar resources that uses three years of data, all days of the month, and five important hours per day. Wind units would receive a "diversity benefit" based on the difference between the exceedance value of a wind area and the sum of the exceedance value of the resources in that wind area. The diversity benefit would be allocated proportionally to the individual exceedances. New units would use proxy data based on the wind area for wind units or Transmission Access Charge (TAC) area for solar resources. Energy Division suggested a variation of the exceedance method intended to capture the benefits of geographic diversity. This proposal would apply to three different classes of intermittent resources: wind, solar, and intermittent cogeneration. An exceedance value would be calculated for each class, statewide. Then the total class exceedance would be allocated to individual units based on energy production.
Effective Load Carrying Capacity (ELCC): A statistical calculation of the amount of "reference" capacity needed to improve reliability by the same amount as the intermittent resource. Reference capacity is often a thermal resource with a zero forced outage rate. An ELCC calculation requires an hourly risk metric such as Loss of Load Probability (LOLP) for the historical hours used in the analysis. ELCC studies are generally performed for all hours of the day at once; no set of "important" hours is defined or used. However, the production of the intermittent resource is most important to the calculated ELCC at high risk hours. CalWEA and AWEA believe that the ELCC approach has yielded results that are consistent with continued use of the averaging method. DRA proposed an ELCC approach based on a technique known as the "Garver Approximation."
LSA took a different approach. Rather than changing the QC counting rule, LSA proposes to use the MCC buckets, revised as needed due to changing grid realities, for recognizing both the capacity value and the limitations of intermittent resources. Sempra Generation proposes that the counting rule be considered in conjunction with our current review of the planning reserve margin (R.08-04-012).
In its Phase 2 workshop report, Energy Division observed that parties have proposed inconsistent objectives for the RA program. According to the Energy Division, some parties argued that QC rules should measure the performance of a unit at and near the time of system peak load, while others argued that the CAISO needs assurance that RA resources are available at all times. This "all hours" versus "peak hours" dichotomy turns out to be the key issue in determining and resolving the appropriate treatment of intermittent resources in the RA program.
Referring to statements in D.04-01-050 and D.04-10-035, CalWEA and AWEA contend that the Commission has determined that the RA program is not focused solely on the peak hours:
"The purpose of the RA program is to provide `reliable service at least cost.' [Footnote reference to D.04-01-050.] The focus of the program is enhancing system reliability. Some parties may assert that the purpose of the RA program is the narrower goal of serving demand during the monthly system peak hour. Although providing capacity during the system peak hour is one aspect of reliability, it does not fully measure a resource's contribution to reliability - indeed, it is a simplified measure, because an electric grid is at a significant risk of failing to meet load in many hours, not just in the peak hour. Indeed, in D.04-10-035, which implemented the RA program, the Commission clarified that the intent of the RA obligation is not limited to serving the single peak hour, but rather the set of hours whose demands are within 10% of the monthly peak. [Footnote reference to D.04-10-035.]" (CalWEA/AWEA proposal, at 2.)
We concur with the proposition that peak system load conditions are not the sole concern of the RA program. We have often acknowledged the importance of maintaining reliability at all times. For example, as we noted earlier in this decision, our MCC bucket approach for determining how to count ULRs provides greater reliability in off-peak periods. The fact remains, however, that providing assurance of dependable resource availability to the CAISO at peak demand periods is and should be the primary focus of the RA program, not just another aspect of it.
After the decisions cited by CalWEA and AWEA were issued, the legislature enacted Section 380, the resource adequacy statute that now constitutes the blueprint for the program.10 Section 380 (c) provides that:
(c) Each load-serving entity shall maintain physical generating capacity adequate to meet its load requirements, including, but not limited to, peak demand and planning and operating reserves. The generating capacity shall be deliverable to locations and at times as may be necessary to provide reliable electric service. (Emphasis added.)
If there were any doubt prior to the enactment of Section 380, the statute now makes clear that the adequacy of physical generation capacity to meet peak demand plus reserves is a key objective for the RA program. We also note that several aspects of the broader RA program as administered by this Commission, the CAISO, and the CEC are designed based on peak demand hours and are consistent with a primary focus on meeting the peak demand. These include local RA studies, deliverability, import capacity, load forecasts, transmission system availability, CAM allocations, DR allocations, Path 26 allocations, and import allocations.
By and large, there is little dispute regarding the contention that the current counting rule overstates the availability of wind resources during peak periods.11 Similarly, parties have not contested the findings that there is a negative correlation between wind production and loads on the CAISO controlled grid, and that wind production is extremely variable and difficult to predict in advance of the hour of interest. Instead, proponents of maintaining the status quo emphasize the need to assure reliability during off-peak periods, in effect acknowledging that intermittent QC as now measured might not be dependable during peak hours. For example, in their January 15, 2009 workshop proposal, CalWEA and AWEA state that "[i]mportantly, the ELCC measures the capacity value of a resource across all hours of the year, and does not focus on just a few peak hours." (CalWEA/AWEA proposal at 5.)
We find this emphasis on off-peak hours to be incompatible with the key objective of the RA program to meet peak demand. Given the demonstrated variability in wind production, the current averaging method is inaccurate because it can produce NQC values that overstate, by a significant amount, the actual, dependable capacity available to the CAISO during the conditions in which monthly peaks are experienced. For example, as the Joint Proponents point out (Joint Proposal, Footnote 5, at 11), production by wind resources in the San Gorgonio wind region for 2005, 2006, and 2007 was 4.9%, 2.4%, and 40.4% of nameplate capacity, respectively. The three-year average is 15.9%, which far exceeds the actual output for two of the three years. Such a discrepancy between calculated QC and actual availability demands a correction in how the calculation is made.
Some proponents of continuing the current averaging method argue that the proposed exceedance method could lead to higher procurement costs due to the need to replace the devalued intermittent capacity. While replacement procurement would be required to offset devaluation that results from a more accurate measure of peak availability, that does not constitute a valid argument for continuing a method that overstates the QC of a resource. The goal of resource adequacy is achieve reliability at least cost, not simply to achieve least cost. To the extent we design resource adequacy requirements that fail to provide the resources needed by the CAISO, the CAISO could find it necessary to activate its backstop procurement mechanisms such as the Residual Unit Commitment process, the Exceptional Dispatch process, or the Interim Capacity Procurement Mechanism. The costs of such backstop procurement would ultimately be passed on to ratepayers, offsetting savings that would be realized from a more liberal QC counting rule.
We find that subject to certain clarifications and modifications noted below, the Joint Proposal of the CAISO, SCE, and SDGE best meets our objectives for RA.12 It calculates QC of intermittent resources that the CAISO can reasonably rely on to serve peak load, thereby meeting the RA program's reliability objective, and it will best mitigate backstop procurement. In addition, it addresses solar as well as wind resources. Moreover, apart from the proposals to continue the current averaging method, it is the only comprehensive proposal that is ready for implementation with the 2010 compliance period. We believe implementation of a more accurate counting convention for intermittent wind and solar resources is important for reliability as soon as practicable, and the Joint Proposal provides a means of achieving such implementation.
Accordingly, we will adopt the Joint Proposal with the following clarifications and modifications. First, we note that Joint Proponents suggest that the exceedance level could be set between 70% and 80%, and propose initially setting it at 70%. We adopt a 70% exceedance level and specify that any change would be considered in a future RA proceeding. Second, we adopt the CalWEA/AWEA/SA proposal to aggregate the diversity benefits of solar and wind generation to recognize the complementary profiles of these resources. Conceptually, this is not unlike the Joint Proposal's provision for aggregating wind resources within a defined wind area, and it gives appropriate recognition to the growing importance of both wind and solar generation in California. Finally, we are persuaded that it would be reasonable to recognize and incorporate into the exceedance method the locational diversity benefit of aggregating intermittent resource on a statewide basis. Although the CAISO expressed concern that transmission constraints may limit the practical benefits of geographic diversity because, for instance, Northern California wind resources may not be deliverable to Southern California during a peak event, we are not persuaded that systematic congestion during peak loads prevents intermittent resources from being deliverable. To the extent that the CAISO has concerns about deliverability due to congestion on Path 26, such concerns should be revised in the next RA update process. The adopted rule is set forth in Appendix C.13 Finally, we note our concern that data on nameplate capacity for existing units may contain errors. While we adopt the use of nameplate capacity as described in the Joint Proposal for 2010, we direct Energy Division to investigate the issue and propose a solution, if appropriate, in the successor to this proceeding.
1 D.05-10-042, Section 7.1, at 43-51.
2 Some resources, because of fuel limitations, limits on annual emissions, and similar constraints, are capable of generating only limited amounts of energy during the course of the year.
3 The reliability concern would undoubtedly be translated into a cost concern. The CAISO anticipates that in the event that the MCC bucket requirement is eliminated and not replaced with a comparable mechanism, its use of the Residual Unit Commitment Mechanism, Exceptional Dispatch, and the Interim Capacity Unit Commitment Mechanism would be more frequent.
4 D.04-10-035 at 26-27; D.05-10-042 at 51-54.
5 The best estimate approach, adopted by D.04-10-035, requires LSEs to submit load forecasts using their best estimates of future customers and their loads. The current customer approach would require LSEs to assume that their customer base will remain fixed for the forecast period, i.e., that load migration will not occur. D.05-10-042 denied a petition for modification in which TURN sought to reverse the determination and adopt the current customer approach. D.05-10-042 noted that an organized capacity market might provide LSEs with a means of addressing the impact of load migration on their RA obligations, and stated the conditions for revisiting the topic:
"In particular, if a capacity market is in place and it has been shown that the load migration problem can be readily addressed by the ability of LSEs to acquire and dispose of increments of capacity sufficiently small (and located where needed) to match such migration, then it would be reasonable to revisit this topic." (D.05-10-042 at 35-36.)
6 Mitigating this concern is the Energy Division's findings that for the 2008 compliance year, monthly load migration adjustments were significantly decreased from previous years and plausibility adjustments contributed more significantly to total adjustments made to LSE forecasts. (2008 Resource Adequacy Report, at 9.) This suggests that to a greater extent than in prior years, the CEC load forecast review process is resolving and correcting under-forecasting before final load forecasts are assigned to the LSEs.
7 Excerpts from the revised SES proposal, which was appended to the Energy Division workshop report, are reproduced in Appendix A to this decision.
8 We recognize the concern that this statement holds true only to the extent that the customer's load was reflected in the year-ahead load forecast that underlies the local capacity procurement obligation for that LSE.
9 Section references herein are to the Public Utilities Code.
10 Section 380, enacted by Assembly Bill 380 (Stats. 2005, Chapter 367), became effective January 1, 2006.
11 CEC staff, California Wind Generation on Hot Summer Days, presented at the Phase 1 workshop on March 25, 2008.
12 Excerpts from the Joint Proponents' comments describing the Joint Proposal are copied in Appendix B.
13 To clarify the adopted exceedance methodology, we have made minor changes to the wording of Appendix C as it appeared in the proposed decision. It clarifies that the diversity benefit is allocated based on energy production.
