Word Document

2006 Resource Adequacy Report

Produced by

the Staff of the

California Public Utilities Commission

March 16th, 2007

This Report was first issued to the Service List for R.05-12-013 on February 2nd, 2007. During the February 8th RA workshop, parties were given the opportunity to comment and discuss the findings of the report, and to suggest edits in writing on February 20th. This Final Report takes into consideration the comments from those assembled at the workshop, and includes some edits. This constitutes the Final 2006 Resource Adequacy Report.2006 Resource Adequacy Report

Table of Contents

1. Executive Summary 5

2. Goals of the Resource Adequacy Program 7

3. Compliance with RAR in 2006 12

4. 2006 Load Forecast and Resource Adequacy Program Requirements 16

5. Counting Resource Adequacy Resources 30

6. Use of RA resources by CAISO in 2006 50

7. Changes to the RA Program for 2007 53

Index of Tables

Table 1 2006 Aggregated Load Forecast Data (MW) 18

Table 2 Summary of Load Forecast Adjustments in 2006 19

Table 3 2006 RA Filing Summary for CPUC Jurisdictional Entities (MWs) 21

Table 4 Total CAISO Procurement as Percent of Total CAISO Obligation and Peak Demand 23

Table 5 Comparison of Performance of Wind Units on Five Peak Days vs. QC (3-4PM) for 18 of 20 Largest Wind Units 35

Table 6 Comparison of Performance of all Solar Units on Five Peak Days vs. QC (3-4 PM) 37

Table 7 Demand Response Program Allocations vs. Used in 2006 (MW) 39

Table 8 Comparison of DR Program Allocation with DR Load Drop in July 2006 40

Table 9 Enrolled MWs in DR Program in California by IOU and August DR Allocation 41

Table 10 LD Phase-Out Schedule Adopted in D.05-10-042, Section 7.4 41

Table 11 LSE use of RMR Allocations in 2006 RA Filings (MW) 43

Table 12 Import Allocations vs. Used in 2006 (MW) 44

Table 13 Comparison of Import Allocation with Average Imports during Peak in July 2006 45

Table 14 Summary of Resource Categories (Excerpt from the 2006 RA Filing Guide) 46

Table 15 RA Capacity by Resource Category 47

Table 16 RA Capacity in Percentage by Resource Category 47

Table 17 Number of Units with Partial RA Contracts to CPUC jurisdictional LSEs 48

Table 18 Resources Available for 2006 49

Table 19 Outages during summer months in MW 52

Table 20 RMR in 2007 54

Index of Figures

Figure 1 2006 Aggregate Load Forecast Adjustments Reported by LSEs, by Month Showing Load Gained or Lost 20

Figure 2 Total CAISO Summer RA Obligation and Procurement vs. Actual Monthly Peak (MW) 22

Figure 3 CAISO July 2006 Loads and Temperatures 24

Figure 4 Historic Extreme Temperature Events by Service Area 25

Figure 5 CAISO 2006 Predicted versus Actual Peaks 27

Figure 6 CAISO 2006 Predicted versus Actual Peaks Scatter Plot 27

Figure 7 Weather-Normalized CAISO Loads and CEC Forecasts 28

Figure 8 SCE Predicted Loads Assuming Historic Temperatures 29

Figure 9 PG&E Area Predicted Loads Assuming Historic Temperatures 30

Figure 10 Performance as a percentage of NQC for 20 Largest Wind Units during Peak (3-4 pm) on CAISO's Five Peak Days of 2006 36

Figure 11 Performance as a percentage of NQC for Solar Units during Peak (3-4 pm) on CAISO's Five Peak Days of 2006 37

Figure 12 Liquidated Damages Contract Summary for 2006-2012 42

Figure 13 ISO 2006 Summer Peak Loads and Imports and Time of Peak 45

Figure 14 FERC MOO Waiver Denials, Summer 2006 51

Figure 15 LARS Required Capacity Trend (1998-2007) 54

1. Executive Summary

This Report provides background and analysis of the launch of the California Public Utilities Commission's (CPUC's) Resource Adequacy (RA) program. Although RA program policies have been under development for several years, the first active compliance period commenced in June 2006, and the program has been ongoing thereafter. This report summarizes the program's experience to-date, with particular emphasis on the summer period June through September 2006. The report examines how the program is actually working and provides a significant quantity of publicly available information. The report identifies some of the key programmatic changes and expansions that are already in place for compliance year 2007. While the report does not make explicit policy recommendations, it is expected to provide factual input into the policy refinement discussions under consideration in Phase 2 of the CPUC's ongoing RA rulemaking, R.05-12-013.

The Resource Adequacy program is supplying resources to meet CAISO needs.

The CPUC's RA program was launched in 2006, and it provided the California Independent System Operator (CAISO) with access to significant quantities of capacity resources. The basic characteristic of an RA resource is the ability to provide power when needed and be available to the CAISO using the RA must-offer obligation (RA MOO). CPUC-jurisdictional Load Serving Entities (LSEs) procured resources to meet load in all summer months, with total RA procurement ranging from 118 percent to 136 percent of California Energy Commission (CEC) Load Forecast, which was 3 to 21 percent above the Resource Adequacy Requirement (RAR). An LSE's RAR is a monthly obligation equal to its forecast peak load for the month plus a 15 percent planning reserve margin. The availability of RA resources did not meet all CAISO needs and the CAISO procured resources using its Reliability Must Run (RMR) authority and the Federal Energy Regulatory Commission must-offer obligation (FERC MOO).

Across the CAISO, LSEs (both CPUC- jurisdictional and non-CPUC jurisdictional) procured resources sufficient to meet the actual peak loads during the summer of 2006. In July 2006, capacity resources procured by all LSEs (CPUC jurisdictional and non-CPUC jurisdictional), if called and operating, could have provided the CAISO a six percent margin above the actual peak load during the heat storm. This figure includes Demand Response (DR) resources in addition to other resources. The six percent margin was below the seven percent operating margin typically used for CAISO operations.

The load forecast methodology operated effectively, but some parts may need review.

As part of developing the RAR, the CEC established load forecasts for each LSE. Initial forecasts were adjusted, in January 2006, to response to LSE comments. The resulting final forecasts and RARs, for each month, were 287-436 MW less than the benchmark CEC forecast. The year-ahead forecasts and RAR may be adjusted in monthly filings to account for direct access load migration. LSEs adjusted their loads significantly between the January CEC forecast and the RA filing month, increasing forecast 584-1,051 MW. The adjustments were largely concentrated in five Electric Service Providers (ESPs), who reported growth due to load migration that accounted for 97 to 99 percent of all ESP load growth reported during 2006. The monthly increase due to load migration more than made up for decrease in the January forecast adjustments and was not accompanied by a corresponding decrease in other ESPs' loads.

The on-going RA program creates significant staff work, which is exacerbated by the poor quality of some LSE filings.

On a monthly basis, the Staffs of the California Public Utilities Commission (CPUC Staff), CEC, and CAISO review 15 LSE advice letters, 15 LSE load forecasts, and supply plans from all RA generators. The CAISO also reviews filings from non-CPUC jurisdictional LSEs that demonstrate procurement of capacity to meet their peak loads. This is in addition to the larger year-ahead load forecasts and compliance filings.

Recurrent minor errors consume staff time and delay the processing of filings. In total, staff has required the filing of over 20 Supplemental Advice Letters and over 45 sets of correction sheets to remedy minor errors over the course of 2006.

Resource counting conventions are not consistently accurate and revisions may be needed.

CPUC Staff analyzed the performance of wind and solar resources during the five peak days in the summer of 2006 (July 21st - 26th). Through our analysis of the performance of these resource types, CPUC Staff found that the counting conventions for measuring the generation of solar units during times of system need is more accurate than same method used for measuring generation of wind units. Wind units performed at 12 to 76 percent below NQC, and solar units, when discounting one of the five peak days, performed 12 percent below to 8 percent above NQC over the same period.

CPUC staff also analyzed the performance of Demand Response programs relative to the size of the DR allocation used in the RA program. Demand Response programs generally produced load reductions equal to 80 percent of the values counted for RA.

The CAISO allocated available import capacity to CPUC jurisdictional and non-CPUC jurisdiction LSEs to ensure the State was not relying on more imports than could be accommodated by the existing transmission system. Throughout the summer of 2006, the CAISO allocated 8,410 MW out of 14,941 MW of import capacity to CPUC jurisdictional LSEs, while 5,502 MW was allocated to existing transmission contracts (ETCs). In their monthly RA filings, all LSEs reported between 5,325 and 5,636 MW of import capacity. During peak (3-4 pm) on the five peak days of 2006, there were from 12,153 to 14,461 MW of imported energy delivered into the CAISO.

2. Goals of the Resource Adequacy Program

2.1. Resource Adequacy Policy Framework

The CPUC adopted a RA policy framework in 2004 in order to ensure the reliability of electric service in California. The CPUC established RARs applicable to all LSEs within the CPUC's jurisdiction, including investor owned utilities (IOUs), energy service providers (ESPs), and community choice aggregators (CCAs) within its jurisdiction.1

The Commission's RA policy framework - implemented as the RA program -- guides resource procurement and promotes infrastructure investment by requiring that LSEs procure capacity so that capacity is available to the CAISO when and where needed. The Commission adopted RAR policies that are applicable to all LSEs starting in Decision 04-01-050, with further elaboration in subsequent decisions, especially D.04-10-035, D.05-10-042, and D.06-06-064.

The CPUC's RA program now contains two distinct requirements: System RARs (effective June 1, 2006) and Local RARs (effective January 1, 2007). The majority of this report focuses on System RARs, since the report is primarily looking back at experience to date.

2.2. Overview of Regulatory Decisions

Several decisions summarized below provide the regulatory background for the development of System RAR and Local RAR. Further development of the CPUC's RA program is ongoing in Rulemaking (R.) 05-12-013.

Summary of Key Aspects of D.04-01-050 on January 22, 2004

Summary of Key Aspects of D.04-10-035 on October 28, 2004

Summary of Key Aspects of D.05-10-042 on October 27, 2005

Summary of Key Aspects of Decision 05-12-017 on December 12, 2005

Summary of Key Aspects of D.06-02-007 on February 16, 2006

Summary of Key Aspects of D.06-04-040 on April 13, 2006

Summary of Key Aspects of D.06-06-064 on June 29, 2006

Summary of Key Aspects of D.06-07-031 on July 20, 2006

Summary of Key Aspects of Decision 06-12-037 on December 14, 2006

2.3. Legislature Codifies the Resource Adequacy Requirement in Public Utilities Code Section 380

In 2005, the California legislature codified resource adequacy in Public Utilities Code Section 380. An overview of the some key sections of the law is provided below.

3. Compliance with RAR in 2006

The RA program in 2006 included several components. These components included: RA filing process, compliance review process, identifying compliance issues, and compliance enforcement. Each component is discussed in more detail below. CPUC Staff has implemented Commission Decisions, and overall compliance has been acceptable, although minor filing errors continue to consume staff time.

3.1. Overview of the RA Filing Process

After D.05-10-042 was adopted in October 2005, the CPUC Staff endeavored to quickly develop an RA Filing Template and Guide to facilitate the implementation of the RA program. CPUC Staff issued a Draft Year-Ahead Filing Template and Guide on December 7th, 2005, a final Year-Ahead RA Filing Template on January 9, 2006, and a final Year-Ahead Guide on January 31, 2006. Subsequently, CPUC Staff issued a Monthly RA Filing Template and Guide, largely based on the format of the Year-Ahead format.

The 2006 Year-Ahead and Monthly RA Filing Templates and Guides were attached to the RA Phase 1 Staff Report in April 2006 in R.05-12-012.2 D.06-07-031, Conclusion of Law 8, affirmed that the RA filing guides and templates are approved and CPUC Staff is authorized to modify them as necessary. The 2007 Year-Ahead RA Filing Guide and Template were issued by CPUC Staff in August 2006.3

LSEs are responsible for making Year-Ahead and Monthly RA filings, as well as Year-Ahead and Monthly load forecasts. The RA compliance filings were submitted simultaneously to the CPUC, CEC, and the CAISO. The load forecast information is submitted to the CPUC and the CEC.

LSEs must demonstrate compliance with the 90 percent requirement by showing 90 percent of its RAR is under contract well in advance. RAR is calculated as 115 percent of an LSE's peak load forecast. The filing covers only the summer peak period. For the 2006 peak period, the 2006 Year-Ahead RA filing was due February 27, 2006, and it covered June 2006 through September 2006. In future, the Year-Ahead RA filing is due in October, and covers the period of the following May through September. (The filing due in October 2006 covered May 2007 through September 2007.)

The month-ahead filings were due on the last day of the month prior to the start of the compliance month.4 These filings were to demonstrate procurement sufficient to meet 100 percent of their RA obligation. The first filing was due May 1st for the month of June 2006, and so on thereafter.

Each LSE's 2006 Year-Ahead RAR was based on its Year-Ahead load forecast submitted to the CPUC and CEC in the summer of 2005. The forecasts were reviewed by the CEC staff and compared to the CEC's aggregate demand forecast (vintage September 2005). If an LSE's forecast was considered unreasonable by the CEC, a plausibility adjustment was made. In consultation with the CPUC, the CEC staff notified all LSEs of their final load forecast for use in their Year-Ahead RA filings. In addition, LSEs were also required to submit monthly load forecast adjustments to account for load migration. The load forecast adjustments were to begin 30 days prior to the first monthly filing deadline, meaning June Load Forecast adjustments were due April 1st. The load forecasts were reviewed by the CEC staff, and the CEC staff notified any LSEs of adjustments to their load forecasts for use in their Monthly RA filings.

The CPUC checked the filings for compliance by verifying that each LSE's submittal was accurate, timely, and satisfied all requirements. The CAISO reviewed the filings to check whether the RA filings submitted by LSEs were consistent with the supply plans submitted by generators.

In 2006, the CPUC Staff worked closely with LSEs to resolve any questions regarding the RA filing process and templates. CPUC Staff has been able to develop answers to numerous questions raised by LSEs that have special or unique circumstances. Working closely with LSEs has contributed significantly to reducing errors or omissions in the filings. Examples of questions brought to CPUC Staff include: treatment of QC for new resources, treatment of QC for resources when initial QC list was inaccurate, and discrepancies between the CEC's and LSE's load forecast. It is the hope of CPUC Staff that this process of working with the LSEs to iron out problems and make revisions will lead to fewer questions in the future and make the RA filing process smoother.

3.2. Compliance Review Process

CPUC Staff, in a coordinated effort with the CEC and CAISO, has reviewed all compliance filings received to date according the process outlined below:

3.3. Compliance Issues

The first year of the implementation of the RA program has been instructive for the CPUC, CAISO, and CEC, as well as all LSEs. CPUC Staff worked with LSEs to correct minor errors in 2006, and saw no pattern of non-compliance issues across LSEs. However, CPUC Staff has identified minor errors on a regular basis, and CPUC Staff has worked with LSEs to provide a timely resolution to these errors. These errors include: filing late, listing units that are within 60 days of commercial operation date, filing information for the incorrect month, inaccurate reporting of demand response allocation, incorrect CAISO resource identification numbers, and failing to send duplicate copies of filing to all three entities, i.e. CPUC, CEC, and CAISO.

Recurrent minor errors consume staff time and delay the processing of filings. For this reason, CPUC Staff is very interested in minimizing the occurrence of errors. In total, staff has required the filing of over 20 Supplemental Advice Letters and over 45 sets of correction sheets to remedy minor errors over the course of 2006. For example, one LSE has been asked to file correction sheets to all but one filing so far, and no LSE has been able to avoid filing corrections. The common LSE has had to file around three sets of correction sheets and one to two Supplemental Advice Letters. This is a significant issue to be remedied as the RA program develops.

In the beginning of 2006, the Commission lacked a simple method for enforcing filing deadlines. The Commission approved Resolution E-4017 on October 5th, 2006 instituting a citation program authorizing Commission Staff to enforce compliance with system and local RA filing requirements.5 The resolution establishes a penalty structure for late, incomplete, or flawed filings. The citation program in Resolution E-4017 is specifically for "failure, absent an approved extension, to submit: (a) any load data, load forecast or other resource adequacy compliance filing in the time and manner required; and (b) other information requested by CPUC Staff or the CEC that is reasonably related to the implementation of resource adequacy requirements6." The procedures identified in the resolution will help CPUC Staff to enforce compliance with the RA program, in particular with filing dates.

3.4. Compliance Enforcement

The essence of the RAR program is mandatory LSE acquisition of capacity to meet load and capacity reserves. Failure of an LSE to file its RAR on a timely basis may result in the issuance of a citation, per Resolution E-4017. Failure of an LSE to meet its RAR overall can result in a penalty for the LSE amounting to a multiple of the cost of new capacity as a reasonable penalty for the deficiency. Commission Decision 05-10-042 established a baseline penalty of 150 percent of the monthly cost of new capacity for 2006 and a baseline penalty of 300 percent of the monthly cost of new capacity for 2007 and beyond. The factors leading to enforcement for RA non-compliance as indicated in Decision 05-10-042 include:

Although CPUC Staff has not initiated enforcement proceedings against any LSEs for failure to follow the requirements of the RA program, CPUC Staff has issued and collected one citation under Resolution E-4017, and it has identified several other potential compliance violations. CPUC Staff is responsible for enforcing the obligations of the RA program for any LSE's failure to comply. If necessary, CPUC Staff will draft an Order Instituting Investigation or other appropriate proceeding to enforce the Commission rules.

4. 2006 Load Forecast and Resource Adequacy Program Requirements

This section describes the 2006 Yearly and Monthly load forecast processes, and the subsequent use of the load forecasts to establish RARs for each LSE. The section also describes the total RA resources procured to meet aggregate System RAR in 2006. From analysis of the RA program throughout the summer of 2006, CPUC Staff found that:

4.1. 2006 Yearly and Monthly Load Forecast Process

The RA program relies on load forecasts supplied and checked by the CEC as the foundation for each LSE's RAR. The load forecast used in the RA program is the most recent CEC "1 in 2" load forecast that is available as of the time the RAR is established for the year. For 2006, the most recent CEC load forecast was compiled by CEC staff in September 2005, and adopted with the IEPR in November 2005. Although the staff revised its forecast for 2006 upwards in June 2006, the revised forecast was not reflected in the CPUC's RA program because the System RARs were already established for the year.

In order to establish the System RAR, CEC reviewed load forecasts submitted by each LSE, reconciled those load forecasts against its own forecast (from September 2005) for the entire IOU service territories, and generated an individual load forecast for each LSE for each month of 2006. For the 2006 Year-Ahead System RA filings due in February 2006, the CEC mailed an individual load forecast to each LSE by certified mail in November 2005. In January, the CEC allowed some LSEs to adjust their forecasts upwards or downwards to account for new estimates of customer retention and migration. This is summarized in Table 1 below.

According to the RA program rules, LSEs can submit monthly load forecasts to the CEC to show any changes in load expected due to load migration. The CEC then checks the revised load forecasts to make sure they remain plausible and are within a tolerance level to the statewide forecast, then supplies each LSE with its adjusted monthly load forecast. Pursuant to the process identified in the Monthly RA Filing Guide, monthly load forecasts were mailed to LSEs by the CEC in April 2006. LSEs used the monthly load forecasts for the Monthly RA Filings that were submitted May 1, 2006 to cover the period of June 1 - 30, 2006. The same process was repeated monthly thereafter. The monthly load forecast adjustments are summarized in Table 2.

4.1.1. Yearly Load Forecast in 2006

The CPUC RA obligation is based on load forecasting done by the CEC. D.05-10-042 requires LSEs to submit historical sales figures and a projected forecast for the following year, based on a reasonable assumption of load growth and customer retention. These forecasts are submitted to the CEC and CPUC for evaluation. The CEC worked to clean the data, adjust for transmission losses, and adjust the IOU load for customers returning from direct access. The CEC developed a trigger for a plausibility adjustment. As specified by D.05-10-042, adjustments were made to account for the impact of energy efficiency and distributed generation (EE/DG) and coincidence peak. Table 1 shows the aggregate LSE submissions for 2006 and any adjustments that were made across all three service areas.

Because the historic and forecast data submitted by participating LSEs contain market sensitive information, results are discussed and presented in aggregate. A more complete description of the methodology, along with more supporting data, went out with the forecasts in November of 2005. This process was repeated for the 2007 load forecasts that were mailed to LSEs in June of 2006.

Table 1 2006 Aggregated Load Forecast Data (MW)

Line

Description

Jun.

Jul.

Aug.

Sep.

Oct.

Nov.

Dec.

1

Submitted LSE Forecasts

36,327

39,057

41,679

37,636

32,856

28,056

29,092

2

Adjustment for Returning Load (IOUs only)

437

440

445

427

406

374

360

3

CEC Adjustment for Plausibility

493

544

568

567

540

413

468

4

EE/DG Adjustment

0

0

0

0

0

0

0

5

Net Adjustment to bring forecasts within 1% of statewide forecast

70

36

224

152

221

38

77

6

Coincidence Adjustment

-2,037

-907

-494

-1,201

-932

-509

-547

7

Total of forecasts mailed to LSEs in November, 2005

35,292

39,171

42,422

37,581

33,091

28,372

29,450

8

Load from Nov. Forecasts unaccounted for in revised Jan. Forecasts

-287

-433

-436

-432

-425

-339

-314

9

Total Forecasts used for compliance, reflecting adjustments made in January, 2006

35,005

38,738

41,986

37,149

32,666

28,033

29,134

Source: CEC staff Load Forecast Methodology Letter mailed to LSEs in November, 2005 and adjusted forecasts for use in 2006 RA compliance filings pursuant to Load Migration Adjustment process in 2006

The total load forecasts used for compliance in Line 9 are more than one percent divergent from the total CEC load forecasts, as the adjustments made in January lowered Load Forecasts for some LSEs while not correspondingly raising any others. Line 8 in the table summarizes the load decrease from the original November load forecast that was left unaccounted for in the revised load forecasts summarized in Line 9. This ranged from 287 to 436 MW depending on the month. This load was substantially accounted for in the Monthly Load Forecast Adjustment process however, as forecasts reflected substantial increase without corresponding decrease.

4.1.2. Monthly Load Migration Adjustments in 2006

D.05-10-042 outlined a process to adjust an LSE's load forecast on a monthly basis. The CEC and CPUC administered the program through 2006. The LSEs were directed to submit revised forecasts two months prior to the filing month, which is one month prior to the RA Monthly filing due date. These load forecast adjustments were to be solely for the purposes of accounting for load migration. Load Forecast Adjustments are summarized in Table 2 below. The table shows that the adjusted forecasts each month consistently represent a one to two percent positive adjustment over the original November forecasts. The decrease in load from earlier adjustments was recaptured here, as well as adding an additional 297-737 MW on top of the November load forecasts.

Table 2 Summary of Load Forecast Adjustments in 2006

Line

Description

Jun.

Jul.

Aug.

Sep.

Oct.

Nov.

Dec.

1

Total Forecasts mailed out in Jan. 2006

35,005

38,738

41,986

37,149

32,666

28,033

29,134

2

Monthly Load Forecast adjustments through 2006

584

808

815

914

936

766

1,051

3

Total forecasts used in monthly RA filings in 2006

35,589

39,546

42,802

38,063

33,602

28,799

30,185

4

Line 3 as percent of Nov. forecasts Line 8 from Table 1

101%

101%

101%

101%

102%

102%

102%

5

RAR based on line 4 as percent of Nov. Forecasts Line 8 from Table 1

117%

117%

117%

118%

118%

118%

119%

Source - Aggregated Load Forecast Adjustments submitted to the CEC and CPUC through 2006

As with many other aspects of RA implementation in 2006, there has been a learning curve on which both the LSEs and CPUC Staff have developed and refined the RA program. First, LSEs were not always clear that they were to use their adjusted monthly load forecast for the compliance filings. Second, LSEs have not always submitted monthly load forecast adjustments in a timely or consistent manner. Thirdly, there has been a significant change for some LSEs between the yearly load forecasts and the adjusted monthly load, which has led to additional filing errors and potential misallocations of RMR and DR impacts.

Throughout the summer of 2006, and particularly in June and July, several LSEs did not use the adjusted load forecast as the basis for their RAR in their RA compliance filings. When LSEs used a yearly forecast that was higher than their monthly load forecast, CPUC Staff accepted that as compliant. When the LSE used a load forecast lower than the monthly load forecast, CPUC Staff contacted the LSE and required them to resubmit the RA filing with the correct minimum monthly load forecast used to establish the RA obligation. These errors continue to occur in the monthly filings.

Figure 1 below depicts the magnitude of monthly load forecast adjustments as reported by the ESPs and IOUs. This is to demonstrate the diversity of load forecast adjustments. Four LSEs (one IOU and three ESPs) reported minimal adjustments of less than .05 MW in size, while five of the twelve ESPs reported 97 percent to 99 percent of the total ESP load forecast increases each month. Load forecast increases were not balanced by decreases, and ESPs were primarily responsible for the increases. As discussed above the load forecasts for some ESPs were reduced during the yearly forecast adjustments without corresponding increases in other LSE load. This unaccounted for load may be responsible for part of the upward adjustments in ESP monthly forecasts, but the magnitude of the adjustments exceeds that of the unaccounted for load, indicating that the yearly load forecasts underestimated ESP load while in general correctly estimating IOU load.

Figure 1 2006 Aggregate Load Forecast Adjustments Reported by LSEs, by Month Showing Load Gained or Lost

Source: Monthly load forecast adjustment filings submitted by LSEs to CEC.

4.2. 2006 System Resource Adequacy Requirements

For every month of 2006, CPUC-jurisdictional LSEs have satisfied their individual and collective system RAR. All LSEs satisfied RAR in aggregate for all peak months of 2006. The total MWs of RA resources7 procured exceeded the total System RAR by between 3 percent and 21 percent, depending on the month. Please note that the Total CEC Load Forecast is that applicable to the Monthly Filings, from Line 4 in Table 2.

During the forecasted peak month of August 2006, the CPUC's jurisdictional LSEs were collectively required to procure 46,891 MW of resources. Collectively, the LSEs procured 103 percent of the total System RAR, or 48,355 MW, which represents 1,464 MW in reserves beyond that required by the RA program. For the actual summer peak in July, CPUC jurisdictional LSEs were required to procure 43,168 MW of resources and demonstrated procurement of 48,111 MW. Procurement totaled 116 percent of the monthly RA obligation, and providing 4,943 MW of reserves beyond the RA program.

Table 3 2006 RA Filing Summary for CPUC Jurisdictional Entities (MWs)

Line

Description

Jun-06

Jul-06

Aug-06

Sep-06

1

CPUC jurisdictional load forecasts, after 2006 Monthly Load Forecast Adjustments

35,589

39,546

42,802

38,063

2

Demand Response Reported in RA Filings (See Section 5.5)

1,862

2,009

2,027

2,013

3

RAR applied to CPUC-Jurisdictional LSEs ([(RAR= Load Forecast - Demand Response) x 115%]

38,786

43,168

46,891

41,458

4

Total RA Resources Procured

46,918

48,111

48,355

46,760

5

Reserves procured above RAR

8,132

4,943

1,464

5,303

6

Total RA Resources as percentage of RAR*

121%

111%

103%

113%

7

Total RA Resources (Including DR) as a percentage of CPUC-Jurisdictional load forecast**

136%

126%

118%

128%

Source: Aggregated LSE Monthly RA Filings. Note-DR is part of RAR calculation so it is not included in RA resources in Line 4.

* Line 6 = Line 4 divided by Line 3

** Line 7 = (Line 2 plus Line 4) divided by Line 1

4.3. Total RA Resources Available to CAISO in 2006

The CPUC's RA program was launched in 2006, and it provided the CAISO with access to significant quantities of capacity resources. CPUC-jurisdictional LSEs procured resources to meet load in all summer months, with total RA procurement ranging from 118 percent to 136 percent of CEC Load Forecast, or 3 to 21 percent above the RAR. The CAISO was able to call upon RA resources using the RA must-offer obligation (RA MOO). The availability of RA resources did not meet all CAISO needs and the CAISO procured resources using its Reliability Must Run (RMR) authority and the Federal Energy Regulatory Commission must-offer obligation (FERC MOO).

Across the CAISO, LSEs (both CPUC jurisdictional and non-CPUC jurisdictional) procured resources sufficient to meet the actual peak loads in all months. In July, capacity resources procured by all LSEs (CPUC jurisdictional and non-CPUC jurisdictional) totaled 53,355 MW of resources to meet 50,270 MW of peak load. Total procurement across the CAISO, with DR resources added in, gave the CAISO a six percent operating margin above the actual peak load during the heat storm. This margin was below the seven percent operating requirements of the CAISO on peak hours as established by WECC.

Figure 2 compares the total CEC forecast (1 in 2) for the CAISO, the CAISO actual peak load, and the total CAISO Summer Forward Commitment Obligation (including the obligation upon the CPUC jurisdictional entities) for the summer months of June through September, 2006. In all months, the procurement demonstrated through the CAISO's Forward Commitment Obligation exceeded the load forecast and the actual load.

Figure 2 Total CAISO Summer RA Obligation and Procurement vs. Actual Monthly Peak (MW)

Source: Aggregated data compiled from CAISO RCST Analysis

Table 4 demonstrates total procurement for all LSEs within CAISO as a percent of both the total procurement obligation across the CAISO and the peak load across the CAISO during the summer of 2006. The data represented in Table 4 is similar to the data used in Figure 2. Procurement across the CAISO ranged from 106% and 123% of peak demand, and between 103% and 106% of the total CAISO procurement obligation.

Not all resources can be called upon equally by the CAISO; there are sometimes operational and deliverability concerns regarding some resources that may cause system operating difficulties, even though procured resources may surpass actual load. Resources that may have operational concerns to the CAISO include intermittent resources, slow starting units, some types of Demand Response programs, liquidated damage contracts, and import contracts.

Table 4 Total CAISO Procurement as Percent of Total CAISO Obligation and Peak Demand

Line

Summer 2006

Total Reliability Requirements Capacity Showing

Jun.

Jul.

Aug.

Sep.

1

CPUC Jurisdictional LSEs

46,918

48,111

48,355

46,760

2

Non-CPUC Jurisdictional LSEs

4,863

5,244

5,211

5,093

3

Total CAISO RA Procurement

51,781

53,355

53,566

51,853

4

Procurement as Percent of Total CAISO Procurement Obligation

106%

105%

105%

103%

5

Procurement as Percent of Total CAISO Peak Demand

118%

106%

123%

114%

Source: Aggregated RCST data

4.4. Summer 2006 Heat Storm

The mid-July statewide heat storm of 2006 resulted in record high peak demand throughout the state. The CAISO peaked on Monday, July 24, 2006 at 50,270 MW. High temperatures in both the northern and southern portions of the state characterize this period. The San Diego Gas and Electric (SDG&E) service area peaked on Saturday, July 22 when the temperature at Lindbergh Field reached 99 degrees. Temperatures in the SCE region also reached a peak on that Saturday, but with a large percentage of commercial and industrial customers closed on the weekend, it did not produce a system peak. Likewise, the temperature peak in the Pacific Gas and Electric (PG&E) service area occurred on Sunday, July 23. By Monday, July 24, 2006, temperatures had decreased slightly but not enough to avoid a record CAISO system peak. Figure 3 shows the temperatures and corresponding daily peaks during the heat storm.

Figure 3 CAISO July 2006 Loads and Temperatures

Source: CEC Data

The CAISO system-wide temperatures of Monday, July 24, 2006, have only been exceeded in five of the last 56 years, shown in Figure 5. Temperatures on the hottest day (Saturday, July 22, 2006) have only been exceeded in two of the last 56 years. High temperatures in the SCE region drove CAISO system-wide temperatures in each of the other occurrences, whereas high temperatures in the PG&E region drove the 2006 heat event.

Figure 4 Historic Extreme Temperature Events by Service Area

Source: CEC Data

4.4.1. Evaluation of the CEC Peak Load Forecast for the CAISO

The CEC's current forecast of peak demand was adopted in June 2006. This forecast is higher than the 2005 Integrated Energy Policy Report (IEPR) forecast, adopted in September 20058, which served as the reference case for the 2006 resource adequacy year-ahead forecasts. The Energy Commission distributed the adjusted 2006 forecasts to LSEs in November 2005. The 2005 IEPR forecast was based on weather normalized 2004 utility peak information, the most current available at the time. In spring 2006, preparatory to the 2007 resource adequacy forecast adjustment process, staff acquired hourly load data from the investor-owned utilities (IOUs) and concluded that 2005 weather-normalized loads were significantly higher than forecast in all three IOU service areas. (Weather normalization is the process of estimating what loads would have been assuming average, or 1-in-2, weather conditions.) The Energy Commission published and adopted a revised forecast of 2007 peak demand for the IOUs in June 2006. This forecast (the "June 2006 Update") was, therefore, not used for the 2006 resource adequacy process, but is the basis for the 2007 resource adequacy forecasts under which LSEs are now operating.

While the CAISO and the individual utility peaks were record-breaking, Energy Commission staff assessed CAISO daily peaks as consistent with the expected load given the magnitude of the temperature. The Energy Commission forecasts demand by utility planning area, but because utility daily peaks are not available to staff on a daily basis, staff developed a proxy method using CAISO daily peaks to evaluate load trends as the summer progressed.

To track weather-adjusted loads through the summer, Energy Commission staff estimated the relationship between 2005 CAISO loads and temperatures using two weather variables. The first is the weighted average of maximum temperatures on three days. The weighting consists of 60 percent of the current day's maximum temperature, 30 percent of the previous day's maximum and ten percent of the second previous day's maximum. The lag is used to account for heat build-up over a three-day period. The second variable is a weighted average of daily maximum temperatures from nine weather stations in the CA ISO, weighted by the distribution of air conditioners in each climate zone. Staff applied the same methodology in the June 2006 Update to develop IOU-level forecasts. That methodology and the adopted forecast are described in detail in the Energy Commission publication, Staff Forecast Of 2007 Peak Demand.9

Figure 5 presents a comparison of actual 2006 summer daily peaks for the CAISO region and daily peaks predicted using the estimated 2005 CAISO temperature-load relationship and increased by the 2005 IEPR 2005-06 forecast growth rate of 1.55 percent. Figure 6 provides a scatter plot of the same information. While the methodology appears to over predict peak at lower temperatures, overall, the methods used to develop the June 2006 Update appear to predict the CAISO peak reasonably well, given the aggregate nature of the data used. This would suggest the Energy Commission's year-ahead forecast for 2006 is reasonably consistent with observed loads. The mean absolute percentage error of the predicted versus actual peak is 2.3 percent.

Figure 5 CAISO 2006 Predicted versus Actual Peaks

Source: CEC data. Actual Daily Summer Afternoon Peaks versus peaks estimated using 2005 relationship and 1.55% forecast growth

Figure 6 CAISO 2006 Predicted versus Actual Peaks Scatter Plot

Source: CEC Data

The weather-normalized 2006 annual peak that Energy Commission staff estimated from the 2006 CAISO daily peaks also supports the Energy Commission's current forecast. Figure 7 shows estimated CAISO peaks using the 2006 temperature-load equation applied to each historic year's (1950-2006) actual temperature patterns. The predicted annual peak values are rank ordered from highest to lowest. The median value is the 1-in-2 (weather normalized peak) and the fifth highest is the 1-in-10 value. This method differs slightly from the method used to develop the Energy Commission forecasts, which builds the CAISO forecast from the utility-level forecasts, but the weather-adjusted peak from this method (47,301 MW) is less than one percent different from the forecasted peak derived from the June 2006 Update (47,603 MW). The 2006 1-in-2 CAISO peak projected in the 2005 IEPR was 46,107 MW, 1.3 percent less than the current Energy Commission forecast.

Figure 7 also shows the current 1-in-2 forecast plus a 15 percent planning reserve margin. The 15 percent planning reserve margin is higher than predicted loads from any temperature seen in last 56 years for the CAISO region, and 6.5 percent higher than a 1-in-10 event.

Figure 7 Weather-Normalized CAISO Loads and CEC Forecasts

Source: CEC data

4.4.2. CEC Forecasts by Service Area

4.4.3. Southern California Edison Planning Area

Staff estimated the 2006 temperature response using 2006 hourly loads for the SCE transmission system area, the three-day-weighted daily maximum temperature, and a weighted daily temperature spread (daily maximum minus daily minimum). Figure 8 shows the distribution of predicted peaks using the peak temperatures that occurred in each of the last 56 years. The estimated load-temperature response for 2006 indicates lower load response at higher temperatures, but the estimated coefficients are not statistically different from those used to develop the June 2006 Update. The June 2006 Update forecast for 2006, using SCE's 2005 actual loads, was 22,791 MW. Applying the coefficients estimated from 2006 loads and temperatures produces an estimate of 1-in-2 loads of 22,447 MW. This is 1.5 percent lower than the June Update forecast, but again the difference is not significant.

Figure 8 SCE Predicted Loads Assuming Historic Temperatures

Source: CEC Data

The 2006 duration curve shows a predicted 2006 peak of 23,352, about a 1-in-5 event. This is based on the Monday temperatures. If the temperatures of July 22, 2006 had occurred on weekday, the predicted peak would be 24,000 MW, slightly less than a 1-in-10 event.

4.4.4. Pacific Gas and Electric Planning Area

Staff estimated the 2006 temperature response using PG&E service area hourly loads and the three-day-weighted daily maximum temperatures. Figure 9 shows the distribution of predicted peaks using the annual peak temperatures that occurred in each of the last 56 years. The estimated load-temperature response for 2006 indicates slightly higher temperature response than forecast. The forecast for 2006 using PG&E's 2005 actual loads was 19,162 MW. Applying the coefficients estimated from 2006 loads and temperatures produces an estimate of 1-in-2 loads of 19,471 MW. This is 1.6 percent higher than the June Update forecast. This indicates an increase in the starting point for the next Energy Commissions forecast for the PG&E service area. Energy Commission staff are preparing revised forecasts for each utility planning area in the state as part of the 2007 IEPR proceeding. The Energy Commission will publish a staff draft forecast in May 2007.

Figure 9 PG&E Area Predicted Loads Assuming Historic Temperatures

Source: CEC Data

5. Counting Resource Adequacy Resources

During the development of the RA program, the Commission established counting conventions for the different resource types. To the extent that the counting conventions `over-count' resources, they detract from the PRM, and to the extent that resources `under-count', they add additional insurance to the PRM. This section introduces QC and RA resources, and explains the process on the development of the CAISO's master NQC list, the CAISO's revisions to the master NQC list for 2007, and issues that have arisen from the development of QC.

CPUC Staff analyzed the performance of wind and solar resources during the five peak days in the summer of 2006 (July 21st - 26th). Through our analysis of the performance of these resource types, CPUC Staff found that the counting conventions for measuring the generation of solar units during times of system need is more reliable than those for the generation of wind units. Wind units performed at 12 to 76 percent below NQC, and solar units, when discounting one of the five peak days, performed 12 percent below to 8 percent above NQC over the same period.

Demand response performance during the summer in July 2006 was also analyzed. Over the summer month of July 2006, the seven reliability "Day-Of" DR programs when called achieved a load drop of approximately 874 MW. This represented 80 percent of its DR allocation in the RA program. Within the same month of July 2006, the 13 price responsive "Day-Ahead" programs when called achieved a load drop of approximately 343 MW. This represented 79 percent of its DR allocation.

5.1. Introduction to Qualifying Capacity

QC is the amount of a resource's capacity that can be counted for resource adequacy compliance filings. QC counting conventions vary by resource type, as described throughout this section, but it is intended to reflect the expected capacity value that will be available to the CAISO during periods of system peak demand.

In 2006, LSEs were required to demonstrate they had contracts for, ownership control of, or allocations of QC in order to satisfy the forward commitment obligations required by the RA program. If an LSE had the rights to a resource's QC, then it could use the QC in its RA compliance filings. If a resource is used as part of the RA filing, it is subject to specific offer obligations under both the RA program and the CAISO's tariff. A combination of the CAISO's tariff provisions and the CPUC's RA product definition, including the must-offer obligation, is the backbone of the RA program.10

Specific formulas and methods for deriving QC values were adopted in D. 04-10-035 and further refined in D. 05-10-042.11 Prior to the first RA compliance filings in February 2006, the CAISO published a QC list for use by LSEs in their RA filings. Specific formulas and methods for deriving QC values, generally referred to as counting conventions, vary by types of resource (e.g. wind, solar, hydro, etc) Some types of resource counting conventions heavily discount the productive capacity of units in order to reflect operational uncertainty. For example, as an intermittent resource, a wind resource's QC is lower than a dispatchable natural gas unit with a similar nameplate rating. The counting conventions are discussed further below.

The commonly used industry performance measurement, Net Dependable Capacity (NDC), represents a unit's ability to produce electric power for specified periods of time. NDC is not used in the RA program because of concerns about generation at peak, especially involving intermittent units. NDC is the maximum capacity a unit can sustain over a specified period of time, modified for seasonal limitations, and less the unit capacity utilized for the unit station service or auxiliaries.

Issues have arisen with respect to QC counting conventions. They include the following:

5.2. Establishment of CAISO'S NQC Values List in 2006

The CPUC establishes the criteria for determining the types of resources that are eligible to provide QC and for calculating QC from eligible resource types. The CAISO publishes resource NQC values on its website.

Using the CPUC's QC rules as a starting point, the CAISO establishes NQC. NQC is the QC value reduced, as applicable, by: (1) testing and verification or (2) deliverability restrictions as determined by the CAISO.12

In December 2005, the CAISO requested that all generation owners within the CAISO control area and other generation owners with supply arrangements with LSEs within the CAISO control area provide the CAISO with QC data. The CAISO reviewed the QC submittals, compiled a list of NQC values, and posted the NQC values on the CAISO's website for public access. The CAISO refers to this list as the "Master NQC list.13 The CAISO has stated it will continue to publish an annual NQC list on or about July 1st each year for the following RA compliance year.

All units are subject to derates based on historical performance. Performance means whether or not the unit was available and performed if called upon under the RA must-offer obligation. CAISO has not yet developed procedures for derating QC values based on performance, so 2006 and 2007 QC values were not adjusted for performance.

5.2.1. Revisions to CAISO'S Master NQC List for 2007

On July 14, 2006, the CAISO updated the NQC list to be used for the compliance year 2007. 14 The update of the NQC list was completed for the following adjustments:

The Commission decided that LSEs need some certainty in resources QC value. Therefore, QC values published on the QC list, updated on an annual basis on or about July 1st, are not to be lowered for the next RA compliance year. For example the values contained in the NQC list published in July 2006 count for RA purposes in all 2007 RA compliance filings even if the unit becomes inoperable. At the same time, the Commission allowed units under construction (i.e. not on the QC list) to be counted in system RA filings, but not in monthly filings. Therefore, the RA program must develop a method of allowing for new units to be added to the QC list and counted during the RA year. The CAISO, for its part, has continued to update, modify, and reissue the NQC list as new information has become available.

5.3. QC for Thermal Generation Units

The counting conventions for thermal generation units are perhaps the most straightforward application of QC. The QC is defined as the maximum dependable capacity available from the unit. The QC identified for most thermal units on the QC list is simply the PMax, or the amount of MWs available when the unit is at its "maximum performance".

The capacity of thermal units is in part dependent on the ambient temperature at the generator site when the unit is in operation. Combustion turbine output is especially sensitive to the ambient temperature and these units have less capacity as the ambient temperature increases during peak summer periods.

5.4. QC for Wind and Solar Resources

Due to the uncertainties involved in forecasting the capacity value for intermittent resources, QC for wind and solar resources is based on historical production.

However, a method for treatment of newer wind and solar facilities that lack three years of historical data has not been adopted. Going forward, the CPUC's RA program will need to adopt a uniform and consistent methodology of calculating QC for newer facilities that do not have enough historical data. As more wind and solar facilities come on-line, this will be an emerging need.

5.4.1. Comparison of Performance for Wind Generation Units during Five Peak Days vs. QC in July 2006 and 2007

In order to evaluate whether the QC methodology for wind units is understating or overstating the actual QC available during the peak period, CPUC Staff and CAISO worked together to analyze performance data from the wind and solar units during the five peak days of summer 2006, all of which occurred at the end of July. It is generally acknowledged that looking at only the five peak days is a very rough evaluation tool, but it is intended to be used to quickly identify whether there is major cause for concern. In our analysis, wind units provided significantly less actual generation at peak than the NQC value established by the counting conventions.

Table 5 depicts wind generation from 18 of the 20 largest wind units. Performance is measured by peak wind generation expressed as a percentage of its NQC in July 2006 and 2007. The analysis uses the one-hour average of generation of MW of wind units at peak during 3-4 pm over the peak days in July 2006. Two units were excluded because they were off-line and the NQC was adjusted to reflect that.

The NQC for the top 18 wind generating units was 608 MW for July 2006, while the actual generation of the units ranged from 145 MW to 533 MW. The performance relative to NQC was 32 percent on July 21, 24 percent on July 22, 55 percent on July 24, 33 percent on July 25, and 88 percent on July 26. July 26 was a particularly windy day that produced more wind generation than other days in July. NQC exceeded generation by 75 to 406 MW during the test period. While wind performance was 24-88 percent of NQC, Table 5 shows that NQC has been a more accurate performance metric than NDC.

Table 5 Comparison of Performance of Wind Units on Five Peak Days vs. QC (3-4PM) for 18 of 20 Largest Wind Units

Peak

Day

Total NDC MW Rating July 2006 in MW

NQC July 2006 in

MW

NQC

July

2007 in

MW

Peak Generation in MW

Peak Generation - NQC in MW

Performance as a percent of July 2006 NQC

Performance as percent of July 2006 NDC Rating

7/21/06

2,298

608

590

197

-411

32%

9%

7/22/06

2,298

608

590

145

-463

24%

6 %

7/24/06

2,298

608

590

336

-272

55%

15 %

7/25/06

2,298

608

590

202

-406

33%

9 %

7/26/06

2,298

608

590

533

-75

88%

23 %

Source: Source: Data provided by the CAISO. NDC is taken from CAISO Master List, and NQC is taken from CAISO NQC List.

Figure 10 below depicts trend lines for the same wind units used in the analysis in Table 5 across the same five peak days in July. The trend lines shows performance results that demonstrate some uncertainty in the reliability of wind generation units during times of system need. Figure 10 includes all 20 wind units including two units that were off line.

Figure 10 Performance as a percentage of NQC for 20 Largest Wind Units during Peak (3-4 pm) on CAISO's Five Peak Days of 2006

Source: CAISO data.

5.4.2. Comparison of Performance for Solar Generation Units during 5 Peak Days vs. QC in July 2006

Table 6 depicts the performance of the four solar units currently on the NQC list. Performance is expressed by actual generation of all solar units as a percentage of their July 2006 NQC. We have also used the same peak days in July and provided data on the one-hour average of generation of MW of solar units during 3-4 pm. The NQC of Solar units was reasonably accurate on four out of the five peak days in July, but included one unit that consistently performed at a level higher than its NQC. Solar units performance compared to NQC was 106 percent on July 21, four percent on July 22, 88 percent on July 24, 104 percent on July 25, and 108 percent on July 26. On July 22nd, two units appeared to be off-line and one unit did not perform well. Actual generation compared to the solar units 2006 NQC ranged from -339 MW to +28 MW.

The NDC for the solar units is 466 MW compared with 352 MW of NQC. NQC for solar units is measured by actual historical performance during summer system peak. Generally, for solar units NQC is more accurate than NDC.

Table 6 Comparison of Performance of all Solar Units on Five Peak Days vs. QC (3-4 PM)

Peak

Day

Total NDC MW Rating, July 2006

NQC
July 2006 in MW

Generation of Solar Units
at Peak in MW

Peak Generation - NQC

Performance
(Generation as a percent of NQC)

7/21/06

466

352

373

+21

106%

7/22/06

466

352

13

-339

4%

7/24/06

466

352

310

-42

88%

7/25/06

466

352

366

+14

104%

7/26/06

466

352

380

+28

108%

Source: Data provided by the CAISO. NDC is taken from CAISO Master List, and NQC is taken from CAISO NQC List

Figure 11 depicts trend lines for the same solar units used in analysis in Table 6 across the same five peak days in July. The trend line demonstrates that when solar units are on-line, the units perform close to their NQC values.

Figure 11 Performance as a percentage of NQC for Solar Units during Peak (3-4 pm) on CAISO's Five Peak Days of 2006

Source: Data provided by the CAISO. 1 Hour Average of generation in MW of all solar units at peak for 3-4 pm

5.5. Demand Response Resources

Approximately 2,000 MW of Demand Response (DR) programs were allocated to LSEs for RA purposes in 2006 and 2007 and almost all of the allocations were used in the 2006 monthly RA filings. Of the 2,039 MW of DR allocated for RA in July 2006, 1,089 MW was derived from reliability programs and 950 MW was derived from price responsive programs. This was approximately 53 percent reliable and 47 percent price responsive out of the aggregate DR allocation for July 2006. In addition, the total enrolled DR in July 2006 for RA was approximately 2,322 MW. Of the 2,322 MW enrolled DR, 778 MW derived from day-ahead and price responsive programs and 1,544 MW derived from day-of and interruptible programs. Day-ahead represented 34 percent and day-of represented 66 percent of the total enrolled DR for RA in July 2006. Not all programs were called to drop load in our analysis that follows.

In a selective sample of DR program performance during the month of July 2006, approximately 104 events occurred requesting load drop from seven reliability DR programs and thirteen day-ahead DR programs. Generally, reliability and day-ahead DR programs performed at approximately 80 percent of its DR allocated in RA. DR allocations are up approximately 12 percent from 2006 to 2007, representing expansions to the DR programs. DR allocation in July 2006 was 2,039 MW and in the DR allocation in July 2007 is 2,286 MW.