1739 Snowmass Creek Road Snowmass CO 81654, ablovins@rmi.org 14 January 2013, revised 15 April 2013 Personal and confidential Hon. Mick Ireland, Mayor of Aspen, and Council Members Adam Frisch, Torre, Steve Skadron, and Derek Johnson City Hall, 130 S. Galena, Aspen CO 81611 (sent by personal emails to all addressees) Dear Mick and City Council Members, I commend the City's leadership in many fine efforts to promote energy efficiency, renewable energy, and other sustainable practices. Aspen has a long, proud history and wonderful achievements in this sphere. Your 100%-renewable-electricity goal is based on climate concerns that also drive my life's work. I write this to you five elected officials and no one else. I regard you as neighbors and allies whose energy goals I support with this analysis. Though this letter is long, I hope you will agree that the significance of this issue demands its rigor and detail. That said, you must be very busy. If you don't have time to read the whole letter, you can grasp its main points by reading the following three-paragraph executive summary and the italicized headlines, then skim or skip to the concluding suggestions on pp. 30-33. Executive Summary Though this letter contains many points that are important to Aspen's energy future, its thrust derives from two central points: First, though driven by the most laudable goals, the City's economic and strategic analysis of the Castle Creek Energy Center is flawed. CCEC costs more and has higher risks than the City's analysis found, or than suitable alternatives. This must be hard to hear and I know you've heard it before. But saying it gives me no benefit, except the success of our mutual goals. Second, the good news is there are several attractive opportunities to achieve your energy and environmental goals with less cost, risk, and controversy. These options flow from an overarching and simple principle, long central to my work: climate solutions should be chosen to save the most carbon emissions per dollar and per year. In contrast, the City evaluated CCEC's economics by asking if it would be cheaper than purchased power-- not, at least in publicly available documents, whether other alternatives might be even cheaper, faster, and less risky, hence able to save more carbon per dollar and per year. Fortunately, Aspen has abundant and cost-effective renewable alternatives, and just needs a sound, open, and credible process for choosing from an ample menu. Better, Aspen has special opportunities to build rapidly on its laudable past efforts in energy efficiency-- especially a mass retrofit of commercial lighting, which could probably about complete the journey to renewable electricity while providing many benefits to the local economy. 1 How I came to write this letter Before exploring these issues in detail, I want you to know how I got into this issue and where I stand on it. My attention was drawn to CCEC when I received indirectly a 20 December 2012 email from the City's William Dolan inviting public comments on Aspen's Renewable Energy Future. On reading it, I felt that perhaps my experience (Annex 1) might be helpful. My employer Rocky Mountain Institute's longstanding policy is to take no part in nor position on local disputes. That's why I didn't participate earlier, except in a brief neutral remark at the 16 June 2011 forum at Paepcke, and it's why I write you now, by permission, as a private citizen working on my own time. RMI did not see and is not responsible for the views I express here, and had no part in performing this analysis. I take no position on the environmental, water-rights, land-use, or similar issues around CCEC, but focus on electricity economics and related strategic issues and opportunities. And while as a citizen I am interested in wise public decisions, as an analyst I am disinterested-- completely independent, with no dog in this fight. Some of my conclusions may surprise or discomfit you and some other friends, including former colleagues whom I greatly admire and respect. But as I've reviewed the City's electricity challenges, I've been struck by the depth and breadth of the need to improve the quality and transparency of analytic assumptions and methods as well as public process. There are important lessons to be learned here about how the City reached this situation, how to do better, and how to rebuild public confidence and comity. Aspen's 100%-renewable-electricity goal is one of the next audacious environmental goals I'm confident will be reached. It can also help inspire wider efforts to arrest the climate change that threatens both the planet's health and Aspen's economy. However, recent years' path to the 100%-renewable-power goal seems unlikely to produce a sound outcome. Annex 2 explains how a recent effort to elicit public suggestions applied criteria that are technically unsound, excluded the very options being sought, and couldn't properly engage the public nor yield valid results. The City's policy goal of raising the renewable fraction of its electricity from 89% to 100% by 2015 is based on climate threats that I've been working hard to abate since my first professional paper on climate change in 1968. Climate solutions that aren't chosen to save the most carbon emissions per dollar and per year will buy less climate solution than they could have, and hence reduce and retard climate protection from what more judicious investments could have achieved. This letter's economic emphasis is thus about more than cost-effectiveness, returns on investment, and protecting Aspen's historically low electricity prices; it's mainly about how best to achieve the climate-protection goal we all share. Responding properly to Mr. Dolan's welcome call for public suggestions was harder than I expected because the City's website about CCEC and proposed alternatives turned out 2 to be often sketchy, outdated, or inaccurate (Annex 3 describes 15 of its errors as of 15 January). I also reviewed the Canary Initiative website, many other posted documents, the City's 2008 bond prospectus for CCEC 1, two Canyon Engineering reports2 to the City on potential repairs to increase Ruedi hydro output, the City's NMMP economic analysis3, two consultancy reports given to the City in December by third parties 4, local press coverage, and other relevant papers I found online. If my data were incomplete, whatever I needed but couldn't find represented a gap in publicly available information.5 The original version of this letter, sent to you on 14 January 2013, stimulated extensive discussions with City staff, who kindly provided very helpful additional data. I had two technical meetings with David Hornbacher and Phil Overeynder, and also met with William Dolan, Scott Miller, Don Taylor, and Randy Ready, among others. To avoid further confusion, I undertook to send you this extensive revision reflecting that muchappreciated new information. Unfortunately, my revision was delayed by two major trips abroad after I received the last key data set on 11 February. I hope that in the future such information can be posted and accurately described so the public can easily find it. And if I have still misunderstood anything, please forgive and correct any errors so I can check if they'd alter my conclusions. I come to this issue from having led in 2010-11 perhaps the most rigorous whole-system analysis and "grand synthesis" yet done of a coherent solution to our nation's energy challenge. It was performed by 61 of us at Rocky Mountain Institute over six quarters with $6 million of philanthropic support and much help from industry in both content and peer review. Our study Reinventing Fire6 showed how to run a 2.6x-bigger U.S. economy in 2050 with no oil, no coal, no nuclear energy, one-third less natural gas, 82-86% lower carbon emissions, $5 trillion lower net-present-value cost (assuming that carbon and all other externalities are worth zero), no new inventions, and no Acts of Congress--the transition led by business for profit. Broadly speaking, Aspen offers a microcosm where your leadership could yield very similar outcomes and set a national example. I'll say 1 http://emma.msrb.org/MS274833-MS272376-MD550926.pdf. C. Peterson PE, Canyon Engineering, "Ruedi Hydroelectric Project Backwater Review," 29 May 2007 and 11 August 2008 letters to P.A. Overeynder, City of Aspen. 3 NMMP Services, "Castle Creek Hydroelectric Project Economic Analysis," December 2010. 4 Kennedy-Jenks Consultants, Draft Concept-Level Feasibility Analysis and Economic Evaluation of the City of Aspen's Castle Creek Energy Center Hydroelectric Plant and Potential Options, prepared for Public Counsel of the Rockies, 6 December 2011, and Tier One Capital Management, LLC, Report on CostEffectiveness of the City of Aspen Castle Creek Hydro Project, 8 December 2011. 5 Notably, I searched for any posted City commentary on the Kennedy/Jenks report and any substantive rebuttal of the Tier One Capital report's technical content, but found neither. I was also surprised that Aspen Electric apparently doesn't post regular annual or monthly data on how much electricity it sells at what price to what kinds of customers, and generates or buys from what sources--all basic and normally public information for citizens trying to help. Aspen Electric has since privately given me helpful data. 6 By A.B. Lovins and RMI, with Forewords by Marvin Odum (President, Shell Oil Company) and John W. Rowe (then Chairman and CEO, Exelon Corporation), published October 2011 by Chelsea Green (White River Junction VT). Extensive supporting material is at www.reinventingfire.com, a 27-minute TED talk at www.ted.com/talks/amory_lovins_a_50_year_plan_for_energy.html , and a 13-page Foreign Affairs summary article at www.rmi.org/Knowledge-Center/Library/2012-01_FarewellToFossilFuels. Of the electronic editions of this layout-intensive and graphics-rich book (now in fourth printing), I suggest the Google Books version at books.google.com/books/about/Reinventing_Fire.html?id=ZQVZxsGFjnA . 2 3 more later about some ways to start augmenting or strengthening Aspen's efforts. My findings and suggestions come in three parts: 1. the City's economic analysis of CCEC is flawed and unreliable; 2. CCEC has higher costs and risks than available, ample, and suitable alternatives, even neglecting its sunk costs and counting only its to-go costs; 3. a new approach to meeting Aspen's electricity policy and to eliciting and benefiting from public participation could pay major dividends to all parties. Now to these points with their supporting analysis and recommendations--after an initial comment about methodology. 0. A preliminary issue: CCEC's computed cost depends on a contentious and unconvincing post-hoc reallocation of cost between dual roles My original letter of 14 January 2013 didn't address the controversy about whether the CCEC penstock's construction was necessary as an emergency drainline for Thomas Reservoir even if there were no hydro project, in order to mitigate significant flood hazard if the dam breached. Views on this topic diverge sharply, and I hadn't intended to get into it. The history is incomplete, tangled, and controversial. The fullest press report I could find7 suggested hazard mitigation was an after-the-fact rationalization rather than an original purpose, but that account of the history was not considered fair or complete by two City officials, who said it omitted significant contrary material they'd provided. Our discussions usefully clarified some issues but also raised others. Here's my current view: o o The disputed hypothetical safety issue could arise from simultaneous failure of both upstream inlet valves to close, filling the reservoir with maximum flows of 25 cfs from Castle Creek plus 27 cfs from Maroon Creek. The existing drainline could handle 25 but not 52 cfs, so if both valves stuck open and the dam therefore failed8, the excess water could run downhill, inundating certain areas with a lowvolume (total 10.3-af) but high-velocity flow for a fraction of an hour. I was told this might occur before the inlet flows could be shut off using the existing upstream manual valves, especially if a storm obstructed access to the intakes. I therefore asked the obvious engineering question: why not add or convert to remote-controlled motorized valves, which could be installed and made highly reliable at modest cost? I was told this option had been rejected because of the risk that remote operation might fail, or debris might block the valves, requiring too-slow manual intervention to clear them. I asked whether those risks couldn't be handled by adding cheap trash racks above the valves and providing prudent engineered redundancy of actuators, battery power, and telecommunications, but I didn't hear a convincing answer. It seemed to me that the chances of two modern, properly maintained, remote-controlled valves' (or their debris-blocking devices') 7 B. Gardner-Smith, "Aspen defends its journey from penstock to drain line," 4 June 2012, http://aspenjournalism.org/2012/06/04/aspen-defends-its-journey-from-penstock-to-drain-line/. 8 The dam doesn't hold enough water to cause complete structural failure, but could develop a hole. 4 o o failing simultaneously in two different watersheds, then triggering dam failure, were probably very small. I didn't formally study that question, but I wasn't told the City had done so either. I came away unconvinced that this simplest, fastest, and cheapest design option, with probably little or no regulatory cost, had been seriously examined. Rather, I gathered that the dual-purpose penstock/drainline approach was preferred because of its hydroelectric potential. I asked why the City hadn't run a drainline back to Castle Creek at near-reservoir grade--about two-thirds shorter than a dual-purpose line all the way downhill. The answer given was that the intervening landowner wouldn't cooperate, that an underground line was too difficult to build and had frost and esthetic issues, and that the City couldn't condemn a right-of-way because it already owned the downhill one. I can't assess that response, but it would be mooted by the previous alternative, or if the flood hazard were not in fact important. That question should really be asked first. The City experts I spoke to expressed their concern that the flood hazard is unacceptable if not mitigated by an emergency drainline. I have not seen evidence supporting their concern, but there are two contrary findings. First, in 2010, Colorado reclassified the dam from "No Public Hazard" to "Significant Hazard" because of employee housing projects built downhill; required a new spillway to be added to Thomas Reservoir to protect the dam from overflow; but found the existing 24-inch pipeline adequate and didn't require a larger one. The City contends that it follows higher safety standards than the state's minimum requirements, but this rationale isn't supported by my reading of the second key source, McLaughlin Water Engineers' hazard report9 to the City. This report found that a dam breach would be unlikely to cause extensive damage to buildings or roads, that loss of life would not be expected, and that the flow wouldn't touch the neighborhoods supposedly of concern but could cause minor damage to three other buildings. I am therefore not persuaded by the City's view that the penstock was required as an emergency drainline, nor, if it were, that upstream valve solutions were thoroughly considered. If the City has a sound engineering basis for judging otherwise, I haven't seen it. This dispute may seem more relevant to political accountability than to current decisions. However, in the past couple of years the City has changed how it calculates and reports CCEC's costs by reallocating some capital, financing, and operating costs to reflect the penstock's assumed dual-purpose public-safety imperative. Citizens who reject that safety argument could calculate costs very differently. The City's posted project budget goes in both directions, causing confusions that I'll try to unravel next. 9 Dated 11 October 2011 and posted at https://www.documentcloud.org/documents/365970-10-11-2011mclaughlin-water-eng-hazards-analysis.html#document/p6/a59066. 5 1. The City's economic analysis of CCEC is flawed and unreliable I've studied the City's financial-study output spreadsheets. 10 The two outside consultancy reports mentioned above, and my discussions with City officials, have clarified the missing or unclear major assumptions, which I find worrisome in six main ways. a. Half the penstock cost has been removed from the CCEC account, and the financing cost of the other half appears to be understated The City's financial analysis included11 the financing cost both of the $5.5-million bond and of borrowing from the Electric Enterprise Fund, but didn't seem to ascribe a cost of capital to any source of funds (beyond the $0.4-million CORE grant received) used to make up the $10.5-million capital budget, which is $4.3 million over the original budget in the bond prospectus. That financing cost could nearly offset the project's Cityprojected savings.12 Don Taylor explained to me that the missing cost of capital was included, in a spreadsheet column cryptically labeled "Annual Cost Change." A revised version he kindly supplied relabeled this cost stream as "Penstock Lease Water Fund." An annual expenditure of $157,254 for each of the years 2014 through 2028--a total of $2.36 million undiscounted--finances just the hydroelectric half of the penstock-and-tailrace cost from the Water Fund. The other half, I was told, has been billed entirely to the Water Fund and is no longer part of the CCEC cost accounting. This "Penstock Lease Water Fund" column was reportedly added to the spreadsheet in the past couple of years (sometime before November 2011) when the City decided to split the penstock-andtailrace capital cost 50/50 between Electric and Water accounts. This important change raises three obvious issues: 1. How was that cost of capital treated previously? (I can't tell without delving into the history of earlier spreadsheet versions.) If it was omitted, CCEC's cost was materially understated. Or if it was all included in the hydroelectricity budget (as the whole penstock cost is still shown in the City's currently posted project budget), why not leave it that way? 2. Allocating half the penstock costs to the Water Fund removes them from the cost of CCEC's electricity production. Citizens who reject the emergency-drainline argument may therefore feel that the electricity cost is being understated. 3. The $3.761 million of "Penstock Lease Water Fund" financing is priced at 3%/y. 10 Apparently based on the 1 November 2011 revised FERC application, with 2014 start date, and posted "raw" (without assumptions and conventions) as the project's sole online economic justification, making them hard for most citizens to understand and impossible to scrutinize. They're at www.aspenpitkin.com/Portals/0/docs/City/GreenInitiatives/Renewables/Hydro/GI_renewable_hydro_CCE C_FinancialScenarios.pdf. 11 However, the Project Budget at www.aspenpitkin.com/Living-in-the-Valley/GreenInitiatives/Renewable-Energy/Hydroelectric/Finances/ includes no debt service, only unfinanced capital cost. The City says this is because it conventionally treats interest as an operating cost, not a component of capital cost. Fair enough for cashflow and return analysis--but inappropriate for informing public choices. 12 Those savings also include avoided wheeling costs that NMPP has warned may not actually be avoidable, since they're normally based on peak electric loads on the line, not kWh transmitted. 6 Is that the correct valuation for inter-Fund lending? Such internal inter-Fund lending, like any use of City operations' available cashflow, is not free money. It could otherwise be used to reduce indebtedness or otherwise yield earnings or benefits13, so investing it in the project incurs an opportunity cost at least equal to the market marginal cost of that capital. 14 The City reportedly assumed a 3%/y cost of capital because that's the yield of a Treasury-bond proxy over the same [15-year] amortization life. But this is methodologically incorrect for two reasons. First, Treasury bonds are financially riskless, but this project (like any municipal bond) is not. Second, the capital is worth what the City could save by using it to prepay its highest-coupon debt as soon as the call provisions permit, just as using your spare cash to prepay a home mortgage is financially equivalent to a riskless return at your aftertax mortgage rate. That avoidable cost to the City must be at least as high as the 4.58%/y bond cost.15 In fact, economic theory would apply not the 4.58%/y average cost but the 4.85%/y marginal cost of the bond series.16 Moreover, even that 2008 marginal bond cost omits the substantial risk premium that the project's recent history would probably cause bond buyers in 2013, not 2008, to charge: the ballot rejected completion of the project whose revenues were to pay the bonds, so buyers' risk perception would increase significantly. Any capital cost financed from the Water Fund or other City balances or cashflows should thus be booked (at a minimum) just as the Electric Fund borrowing was correctly booked--with principal and market interest at the current risk-adjusted rate. Not doing so overstates CCEC's net benefit. By how much? Well, $3.761 million priced not at the Water Fund's 3%/y but at the latest 4.58% average bond rate and over 28 years like the Electric Fund borrowing, using a standard loan calculation, costs $1.12 million more. That's about $0.63 million in present value, or 22% of the project's City-projected $2.9million total net present value benefit. And if you don't accept the emergency-drainline assumption, then the half of the principal and interest recently reallocated to the Water Fund (another $3.34 million undiscounted or ~$1.87 million in present value) must be 13 The City does invest "temporarily idle cash...in various securities to maximize interest income": http://emma.msrb.org/EP300337-EP14212-EP636311.pdf. 14 Kennedy/Jenks Consultants, at p. 2-2, concur: "The [City's] economic analysis does not consider the forgone investment income from the $3.6 million to be provided by Aspen. Similar to the cost of debt service, this additional cash investment could yield at least 4.5 percent (the cost of capital) [or, I'd add, could be used to the extent permitted by bond and accounting rules prepay or defease the bonds, saving future interest costs]. At this rate of return, the internal rate of return estimated by Aspen would be significantly reduced." 15 To be sure, the City would need to review its whole bond portfolio to check the call provisions and interFund flexibility rules, which may be restrictive. The City might also earn a lower market interest rate for some years until it became permitted to prepay a given tranche of higher-interest debt. But there may well be some older, higher-rate bonds currently prepayable instead on some other fungible Fund account. If the City doesn't fully exploit a maximally fungible approach to managing its cash balances and debt obligations, it should--showing as much ingenuity in the fungibility of sources of repayment as it has in uses and transfers of funds. 16 The 4.58% "true" bond rate (all interest, issuance costs, and premium divided by bond years) is an average cost, not the marginal cost, so it is correct in accounting terms (which are about what you did) but not in economic terms (which are about what you should do). The laddered bonds range from 3.0 to 4.85%/y interest rate depending on their maturity, and NMPP's spreadsheet uses the maximum bond rate of 4.85% as its nominal "interest factor." That's economically correct if an updated risk premium is added. 7 added to the hydroelectric project's cost too. That's 64% of the City's projected net benefit. Thus about 86% of the calculated present-value benefit comes from underpricing half the penstock's cost of capital and transferring the other half from the Electricity to the Water account. b. On the City's reckoning, CCEC would at best be economically marginal Despite that ~$2.5-million advantage and assumed low operating costs (whose correction, as I'll show below, would more than wipe out any remaining net benefit), the City found a very low Internal Rate of Return (IRR), ranging from 1.0 to 7.3%/y, with breakeven times of 14-30 years. Private investors wouldn't touch such a proposition. Kennedy/Jenks Consultants confirmed that the project's cost-effectiveness "is questionable and would rely on a coal power [price] escalation rate of greater than 3 percent [per year]" (discussed below), so they called for the "economic assumptions [to be] thoroughly reviewed before long term economic commitments become irreversible." Tier One Capital, counting all the project's costs of capital, deemed CCEC "not cost effective" with IRRs of 0.08-2.06%, far below the cost of capital 17, and thus costing the City $4-6 million more in net present value than it returns.18 Though I don't agree with every statement in either of those reports, I'll show below why the project's performance will probably be several million dollars worse than these experts have projected. I'll also elaborate why I believe the project would be uneconomic to complete even if its already-sunk costs were ignored. Even the highest IRR for CCEC calculated by the City, 7.3%, is far below that of strong alternatives I'll describe below. For example and comparison, Reinventing Fire found average IRRs of 33% for tripling or quadrupling the total energy productivity of U.S. buildings. Aspen's building stock uses nearly all of the electricity in town, in devices and systems that--despite past efforts at codes, above-code new-build improvements, and retrofits--retain most of U.S. buildings' typical opportunities for improvement. c. The project's economics depend sensitively on its how much water the project can divert without harming streams, on the price of saved coal-fired electricity, and on operating costs. The project loses money unless all three of these unknown values turn out favorably, but in fact they're causally linked, and they're much likelier to get worse all together than better all together. The City's analysis doesn't illuminate that risk The next big issue is that the project's financial performance is extremely sensitive to two 17 Stated in the City's November 2011 spreadsheet to be 4.58% for the bonds ($5.583 million), plus 3.5% for the Electric Enterprise Fund ($0.756 million) and 3% for the Water Fund ($3.761 million)--implying a weighted average cost of project capital around 3.9%/y. However, as stated above, I think the lower figures understate the City's 2008, let alone current, marginal cost of capital. 18 They found the Internal Rate of Return was negative in 50% of the more than one thousand combinations of variables tested, and net present value was negative in 99%, so "there are no scenarios that produce a positive return on investment." The City's former Finance Director concurred that "Aspen's current electric rates will never cover all of Castle Creek's debt service, operations, maintenance and environmental mitigation costs" (www.aspentimes.com/article/20110308/COLUMN/110309866)--requiring higher electric rates or taxes or both to cover the losses. 8 variables that the City has partly analyzed--hydrology, and the future price of avoided coal-fired electricity from the City's wholesale supplier, the Municipal Energy Agency of Nebraska (MEAN)--plus a third variable (operating cost) that the City didn't carefully analyze and didn't sensitivity-test, so I'll analyze it below. The City's posted financial scenarios19 show Internal Rates of Return that I've tabulated: Sensitivity of updated City IRRs to coal-el price and curtailment (CCECIRR6YR, 2014 startup, MNPP costs)) permanent flow reduction (%) 0 17.5 35 %/y real coal-el price escalation 1 4.5 3 1 2 5.9 4.5 2.6 3 7.3 6 4.2 or in graphical form, with % permanent flow reduction along the bottom, %/y real coalelectricity price escalation on the right, and IRR (%) on the vertical axis: I consider the City's preferred combination of assumptions unrealistically sanguine. To get the City's highest modeled IRR, 7.3%, one must assume both the most favorable hydrology scenario (34% or 35% curtailment 2014-16, 17.5% 2017-19, and zero 2020- 85) and that the nominal price of MEAN's coal-fired electricity sales to Aspen will rise 26% during 2012-15, then by an average of 1%/y through 2085. (However, the corresponding real (inflation-adjusted) escalation rates for electricity prices cannot be inferred or analyzed because the rate of monetary inflation is not stated or known.) 19 City's updated CCECIRR6YR financials (2014 startup, MNPP costs, slow start), accessed 12 Jan 2013, www.aspenpitkin.com/Portals/0/docs/City/GreenInitiatives/Renewables/Hydro/GI_renewable_hydro_CCE C_FinancialScenarios.pdf. 9 A City spokesperson reportedly described 20 the 0.3-0.6%/y escalation in real coal21 prices forecast in October 2011 by regional utility giant Xcel Energy as "ridiculously low." Actually, she was apparently confusing real with nominal escalation, Xcel's projection may well prove too high, and it's of dubious relevance anyway: the projection was inexplicably about the price of coal, not of coal-fired electricity. But those two variables are related, and the U.S. Energy Information Administration (EIA) just forecast 22 2011- 40 escalation in the real price of coal to U.S. utilities will average 1.0%/y, but that U.S. electricity's real price escalation will average only 0.3%/y. Both these projections may be too high because EIA projects 0.7%/y growth in U.S. electricity use 2012-40, but by law cannot include the kinds of technical, business, and policy innovations that led our Reinventing Fire team to roadmap roughly 1%/y demand shrinkage for 2010-50. This realization is spreading in the industry23: just the new building codes entering force 2011- 12 in about half of the United States could wipe out nationwide electricity demand growth. Such a demand-shrinkage trajectory is already starting to appear, with an unprecedented (though not yet weather-adjusted) 2011-2012 one-year reduction of 3.7% just reported by EIA in electricity consumption per dollar of real GDP. The key point here is that whatever future U.S. electricity use may be, coal's share of its generation is plummeting. My analysis from EIA data found that just in the past two years, coal lost 19% of its U.S. market share to three competitors--natural gas, efficiency, and renewables. (Over the seven years 2005-12, coal's market share lost 28%.) Coal-fired electricity's prospects are dismal in the U.S. and dwindling abroad-- hardly a recipe for higher prices. Of all U.S. coal-fired power plants, over 70% and half their capacity are over 30 years old, hence fully amortized, and one-third are over 40 years old. Over the next decade, a substantial fraction, perhaps a third but by some recent midrange estimates even two-thirds24, will be shut down rather than paying to comply 20 www.aspentimes.com/article/2011111210/NEWS/111209838 I'd originally assumed that the escalation rates referred to are in real (constant-dollar) terms, like the disputed Xcel Energy coal-price forecasts (Tier One Capital's report, Fig. 4, marked in 2011 $). But I now understand from Don Taylor that the City actually used nominal escalation rates (not adjusted for monetary inflation) and had no explicit assumption about inflation rates. He considers 1.5-2.0%/y a reasonable inflation-rate assumption, bracketing EIA's latest 2011-40 assumption (next footnote) of 1.7%/y. Assuming 1.5-2.0%/y inflation would imply that the City's assumed 2-3%/y nominal escalation rate for coal-fired electricity is equivalent to roughly 0.5-1%/y in real (inflation-adjusted terms). But without knowing assumed inflation rates, nominal escalation rates become unanalyzable, especially over this project's very long lifetime. That's a basic analytic weakness of using nominal rather than real dollars. I can't tell whether the City's assumed average nominal escalation of nearly 1%/y in its purchased electricity prices (and all their components including capacity and wheeling charges) starting in 2016 means that real prices are expected to rise, stay constant, or fall, because I don't know for how long Mr. Taylor's ~1.5- 2.0%/y monetary-inflation assumption is supposed to hold, and neither does he: he simply deferred to the City's Nebraska utility experts and adopted their nominal inflation-rate assumptions. 22 Annual Energy Outlook, www.eia.gov/forecasts/aeo/er/pdf/tbla3.pdf, January 2013. This EIA Early Release forecast doesn't detail all the fine print that's in its July 2012 full forecast, which projected that during 2010-35, utilities' real price for a BTU of coal will grow by 0.9%/y (or 0.7%/y per ton), ranging from -1.0 to 3.0%/y (Table D14, p. 212, www.eia.gov/forecasts/aeo/pdf/0383(2012).pdf), while their consumption of coal will rise only 0.1%/y to 2035 (ranging from -0.8 to +0.4%/y). However, now much more than in July 2012, coal's domestic demand is collapsing and its global demand is softening. 23 E.g., A. Faruqui & Eric Shultz, "Demand Growth And The New Normal," Public Utilities Fortnightly, December 2012, www.fortnightly.com/fortnightly/2012/12/demand-growth-and-new-normal. 24 E.g., Pratson et al., "Fuel Prices, Emission Standards, and Generation Costs for Coal vs Natural Gas 21 10 with long-ignored or -deferred Nixon-era basic environmental laws.25 Yet almost no net new coal-fired capacity will be built in the U.S. (or in the EU or other market economies) because it has no business case. Moreover, U.S. coal's export prospects are increasingly limited by serious competitive, logistical, and demand constraints: in 2012, policy in China and India (together over half the future world coal market) shifted markedly against coal, with India shelving plans for 42 GW of coal plants just since March 2012.26 The frightened U.S. coal industry is striving mightily to reduce its prices to retain fleeing customers, which plan to retire all but the younger and most efficiently operable power stations. In these circumstances, familiar to any energy expert, I think the burden is on the City to show why its assumed coal-power price escalation sustained over the next 75 years is reasonable. I understand that MEAN's bet on robust demand for electricity in general and coal-fired electricity in particular gives it a strategically challenging cost structure that may make its coal-fired output increasingly uncompetitive, threatening a spiral of falling demand, idled capacity, and rising rates over the next decade or two. But over the assumed 75 years, just 1.0%/y nominal escalation means a 111% rise (corresponding to an unknowable rate of real escalation) while coal's main competitors, already robustly beating coal power, get ever cheaper. C'mon, guys. Basic supply and demand will clear markets. The real power prices CCEC would avoid are likelier to fall than to rise. Coal-power price risks really do matter: Holy Cross's original wholesale supplier Colorado-Ute was destroyed chiefly by its general manager Girts Krumins's bad bet on high coal prices. I hope Aspen doesn't make the same mistake. From four decades' utility experience, I do not agree that, as one City official put it, "scarce resources plus increasing demand will result in some level [of] increasing real prices" for U.S. electricity. That is not the sort of future any smart utility investor bets on today as efficiency and renewables get cheaper, electricity demand stagnates or shrinks, and coal's market prospects rapidly dim. Here's a simple way to think about costs. The case for CCEC counts on half-milliondollar-a-year average savings (in mixed future current dollars, inflating for some unknown period by perhaps 1.5-2%/y short-term and perhaps less long-term) from generating CCEC power rather than buying MEAN power. That average works out to nearly 10?/kWh--in nominal dollars of unknown value, compared with average MEAN wholesale rates of 16?/kWh expressed in the same terms. The other 6.2?/kWh is the City's forecast of what CCEC power will cost, again in nominal dollars of unknown value averaged over 75 years. If MEAN power costs escalate slower or CCEC power proves costlier than the City forecasts, or both, then that spread can dwindle, vanish, or reverse. Both outcomes seem likely. Indeed, MEAN's forecast long-term price escalation Power Plants," Environmental Science & Technology, DOI: 10.1021/es4001642, on Web 15 March 2013. 25 City staff assume that this enforcement will require surviving coal plants to raise their prices. This isn't obvious. It seems more likely that market aversion to coal power will force discounting--pricing-to-sell. 26 A.B. Lovins, "Asia's Accelerating Energy Revolution," 26 March 2013, http://blog.rmi.org/blog_2013_03_26_2013_Asias_Accelerating_Energy_Revolution. A follow-up blog in press for April 2013 on the same site explains corresponding and surprising trends in Germany. 11 could greatly intensify MEAN's financial distress and risk. As I'll describe, the long-run price of US electricity is increasingly being set not by coal but by renewables and by equilibration with efficient use. MEAN's portfolio looks ill-prepared for that future. Reductions in water availability As for hydrology, informed opinions differ sharply about how much diverted flow and hence output curtailment will be needed to protect the streams and surrounding forests and ecosystems. Ultimately this will be an empirical question--hence a risk to the project's economics. This graph (with broken scales on both axes) shows how a 7-cfs difference in stream release levels causes a $6-million difference in the City's projection of undiscounted net revenue--twice the project's projected net-present-value benefit:27 The hydrological and coal futures will both unfold simultaneously, making it harder to interpret the hydrologic risk intelligently, but clearly it's large. Worse, the hydrologic and coal-power-price risks are linked, so both worsen in tandem. Here's why: high coalelectricity prices require rapid growth in U.S. and global coal and electricity demand (otherwise producers' prices soften). But burning that coal intensifies climate change, decreasing snowpack and runoff while condensing snowmelt into fewer months, exacerbating streamflow constraints.28 27 Using all the City's assumptions: 28-y bond life, cost of capital apparently counted only for the $5.5M bond issue and the Electric Fund out of the total $10.5M budgeted expenditure, 2011 Baseline conditions, commissioning 2011, and operational 2012 (later updated elsewhere to 2014). The City's adopted hydrological assumption, used also in the NMPP study, is the 13.3-cfs release, the second point from the left. NMPP says it assumed "projected wholesale power cost increases...over the life of the project" but doesn't say what they were in real terms. The spreadsheets don't show net present value or IRR, but the graphed scenarios span undiscounted-cashflow breakeven dates from 2026 at the left to 2036 at the right. 28 That happens more slowly, but will mostly play out over the project's 75-year life, mainly in the first few decades: SkiCo is already very worried about just the first whiffs of this shift that's already underway and growing. 12 Thus the costly displaced electricity that drives the project's economic benefits also requires more curtailment of electricity production to protect streamflows, reducing those benefits. This synergy between the two variables in the surface graph on p. 10 tends to drive IRRs from the high corner toward the low corner (actually, as I'll show, well below it), worsening the project's already poor economics. And it's even worse than that, because the third sensitive variable, operating and maintenance costs, would also tend to rise in a climate-change scenario. That's because similar costs get spread over lower production, and because increased water stress could require streamflow studies and decision processes, currently assumed to end in 2019, to resume and expand instead. The City has promoted CCEC to help abate climate change, to which Aspen's economy is exquisitely sensitive. So a key strategic question is: If the rest of the country and world doesn't promptly exercise climate leadership as vigorously as Aspen has, so climate change continues and intensifies--surely a big risk--then what will be the highest and best use of streamflow? Some would argue that it's increasing the resilience of the ecosystem by reducing the risk of disastrous drought, hence fires and the like. Husbanding scarcer water throughout that ecosystem is the main way to do that, while there are many ways to provide electrical services, many of which divert no water. A potentially offsetting argument is that the City says its water-saving programs have saved an amount of streamflow diversion roughly comparable to what CCEC would divert.29 I haven't seen an analysis controlling for other key variables, but if true, it could be interesting and relevant. However, in an era of rising climate uncertainty and hence community risk, of which the current drought may be just a foretaste, citizens may consider it a higher priority to reserve any water already saved, and any more that can be saved, for ecosystem "climate insurance." I don't think the available City economic analysis properly presents or analyzes, in a way that's valid and useful for decision-makers like you, the spectrum and linkages of major risks that this project presents--not only the two just discussed, but also seemingly unanalyzed risks related to execution, operation, and long-term reliability and upkeep. One of the execution risks is further capital-cost escalation. Council approved the November 2007 bond election based on a 6.5?/kWh levelized cost (5.2? for construction and bond financing, 1.3? for operating cost) estimated from a $5.1-million 2006 construction-cost estimate that has since doubled. That escalation may not be over yet. And of course the recent construction hiatus further worsen's the project's economics by deferring projected revenues while financing costs remain unchanged. d. Operating and maintenance costs are unreliably analyzed and substantially understated Then there's a seemingly small but actually big issue about the City's assumed CCEC 29 William Dolan wrote me 1 February 2013 that the City's ~56% reduction in potable water production during 1993-2011 leaves ~4.4 more cfs in the stream, and that nonpotable operational improvements save another ~4 cfs, for a total of ~8-10 cfs left instream compared with 1993. 13 running costs--initially 1?/kWh with escalation zero initially and slow later. 30 That assumption was very unconvincingly derived by prorating the average operating and maintenance (O&M) for all U.S. major investor-owned utilities' hydro plants. 31 Those plants average about 77 times higher rated capacity than CCEC 32, so their nearly-all-fixed costs are spread over enormously larger output. The City then spread those costs over more electricity than CCEC would deliver. 33 We need a sounder methodology than that, because each cent per kWh of O&M cost changes the project's net present value by $1.31 million (discounting at the declared ~3.9%/y weighted-average cost of capital). Seeking more relevant data grounded in local experience, I originally analyzed in Annex 434 the Ruedi hydroelectric plant's reported 2003-07 O&M costs in 2011 $ per net kWh delivered to the distribution grid (which would occur directly without transmission loss). These O&M costs would be 3.1?0.88?/kWh if simply renormalized to CCEC's expected delivered output. I then hypothesized that Ruedi may be a conservative proxy for CCEC's O&M costs. CCEC would be rated to produce 3.8x less electricity per year than Ruedi, spreading fixed costs over less output. Though far closer to the City than Ruedi, CCEC also has more farflung and complex physical assets, and would incur monitoring, decisional, and streamflow-adjustment operating costs that Ruedi lacks. I guessed that CCEC's O&M might therefore cost more like 4+?/kWh with little output curtailment-- more with higher curtailment. City experts believe differently. They say that Ruedi's remoteness about doubles its O&M costs because the maintenance contractor spends roughly half his billable time 30 "Castle Creek Hydroelectric Plant Economic Analysis" spreadsheet, July 2010, p. 11, "Base assumptions used for analysis." The result appears to be $41,616 for operation and $28,008 for maintenance and capitaladditions costs, a total of 0.967?/kWh in unstated dollars (perhaps 2011 or 2012 $, since the total is 5.7% more than its cited source reported in 2010 $). This O&M cost is derived from national-average data I'll discuss below, and is stated in the spreadsheet's notes to be allocated to 7.2 GWh/y of gross generation. It should, however, be allocated to CCEC's net delivered output (which produces the economic benefits by displacing MEAN power). It is equivalent to 1.27?/kWh if spread over the 5.5 GWh/y of net delivered output) stated in the City's public information since 2007. However, the NMPP report also moves a hardto-disentangle capital-additions component out of the O&M columns, where the City put it, into capital costs, and I haven't readjusted for that, but just taken the City's spreadsheet assumption for O&M at face value--initially 0.967?/kWh. That's about half the O&M costs for Ruedi graphed--though incorrectly (see Annex 3 below, #4)--at www.aspenpitkin.com/Living-in-the-Valley/Green-Initiatives/RenewableEnergy/Hydroelectric/Finances/, making me wonder why national-average data were substituted. 31 The reference is unclear, but Table 8.2 of EIA's Electric Power Annual has the cited title, and gives 0.915?/kWh as the "average operating expenses" for hydro plants owned by major U.S. investor-owned utilities: www.eia.gov/electricity/annual/html/table8.2.cfm. It is based on FERC-1 reports, which allow each reporting utility to pick its own tacit accounting conventions and cost categories. 32 In the same source, Table 5.1 reports 904 hydro units in 2010 owned by "electric utilities" (not distinguished by ownership), and Table 1.1A reports they have a capacity of 72.974 GW, so the average unit is 80.7 MW--77 times larger than CCEC. The effect of mixing public and private utilities is unknown. 33 The City's spreadsheet assumed 7.2 GWh/y gross production, but that number seems outdated. The Public Works Director's 17 Aug. 2007 decision memo to Council, the bond prospectus, and much currently posted City information give 5.5 GWh/y delivered, probably after some level of curtailment. I'll use Mr. Dolan's more exact 5.476 GWh/y. Any curtailment of this flow would reduce the project's net value. 34 Although Ruedi's annual kWh production doesn't seem to be posted online, its capacity factors for 2003- 07 were in the City's NMMP report. Matching up Ruedi's inferred output with the same years' O&M costs reported in the City's 2008 bond prospectus revealed its 2003-07 O&M costs. City staff later gave me the actual production data and revised costs shown in Annex 5 with my conversion to constant 2011 $. 14 getting to and from the site (a 3.5-hour round trip), and that CCEC's O&M costs are incremental, piggybacking on equipment already bought and staff already on payroll to serve Ruedi and Maroon Creek. The City also argues that CCEC's assets are not "more farflung and complex" than Ruedi's because the penstock would be needed anyhow as an emergency drain and is thus an O&M responsibility of the Water not the Electricity department--an assumption dependent on one's judgment about that supposed dual need. Sure enough, the City, based on its dual-purpose assumption, has assigned the penstock's entire O&M cost to Water (its rare hypothetical use) and none to Electricity (its proposed full-time use)--an unusual accounting convention, since operating costs normally track capital costs, which are currently split 50/50, and the structure's original purpose was for electricity, before the emergency drainline role was added later.35 My initial analysis found that the City's 0.97?/kWh O&M estimate for CCEC looked at least 2?/kWh too low based on Ruedi's O&M costs. The difference had a present value over 75 years of $2.8 million. 36 That nearly wipes out the project's projected $2.9-million net present value, independently of any adjustments for realistic cost (and/or allocation) of capital, hydrology, or coal power price escalation. Adding those would robustly more than eliminate the project's entire City-calculated net benefit. However, I also noticed that the City's published financials (reflected on the second page of my Annex 4 spreadsheet) showed that during 2008-11, Ruedi's operators had made new investments--capital additions--totaling a startling $6.36 million, equivalent to three-fourths of Ruedi's original real construction cost. If this had actually been spent at Ruedi, as stated in the annual financials reported to bondholders37, it would have more than doubled Ruedi's 2003-07 average real O&M cost in 2008 and redoubled it in 2009.38 On inquiring, I was forthrightly told that this money was in fact not spent on the Ruedi plant but on other City renewable energy projects, including CCEC. 39 I was told that the City had renamed and repurposed the Ruedi fund some years ago as a general Renewable Energy Fund, but hadn't updated its financial reporting to show that change. The City's utility and financial staff weren't aware of the resulting reporting anomalies, 35 The City's explanation is that there are no added O&M costs for hydroelectric operation of the drainline/penstock: its operational cost is the same whether it's operated for a single use or as a dual-use facility. In that case it would seem at least equally rational to assign all O&M costs to hydroelectricity. The City has evidently done the opposite in the belief that hydroelectric operation is an option while drainline availability is a public-safety necessity. I don't know how its auditors will view that accounting treatment. 36 Applied to 5.476 GWh/y of net delivery and discounted at CCEC's roughly 3.9%/y weighted-average cost of capital (a common utility-economics convention). 37 The City's 2011 financial report says the Ruedi Hydroelectric Fund's financial statement "accounts for the operations of the Ruedi Hydroelectric Facility" (http://emma.msrb.org/ER611869-ER475142ER878147.pdf, p. D3), so I thought all Net Capital Additions booked to Ruedi had been spent there, but that's not what happened. 38 Ruedi's output spiked to 21 GWh/y in 2008 before falling back to 19.5 in 2009. I couldn't find Ruedi's capacity factors for 2010-11 to see if the new investment were paying off, but I did for 2008-09: my second spreadsheet shows that those major Net Capital Additions raised Ruedi's 2011-$ O&M cost to 7.4?/kWh in 2008 and 14?/kWh in 2009, and raised the 2003-09 average O&M cost to an extraordinary 4.9?4.5?/kWh. An interfering variable on which I have no further information is that apparently BuRec may have meanwhile changed the reservoir's operating mode, reducing the plant's 21 GWh/y design basis. 39 Much of the CCEC-related construction has been funded from the major 2008-11 inter-Fund transfers shown in aggregate on the City's annual accounts but not specified as to purpose. 15 which I've urged them to correct in the interest of accuracy and transparency. Since the published Ruedi O&M expenditures weren't all Ruedi-specific, my initial O&M analysis relying on them was invalid. At my request, therefore, City staff dug down to the four-digit-account-code level to excise from the Ruedi (and Maroon Creek) O&M data any costs incurred elsewhere or otherwise misrecorded. It was also necessary to discard all Ruedi data before 2001 because they included financial distributions to project partners. I have no way to assess the accuracy of the resulting data, but appreciate the City's effort to improve their quality. Taking the results (Annex 5) at face value: o o Ruedi's revised 2002-12 O&M costs averaged 2.09?1.09?/kWh in 2011 $40--onethird less than my original finding, but more than twice the City's financial assumption. However, they're not really comparable because the City seems again to be using mixed current dollars rather than constant dollars. You can't spend mixed current 2002-12 dollars because they no longer exist and have no specific value today. Valid analysis requires specifying costs in real terms. The Maroon Creek hydro plant's revised real O&M during 1995-2012 averaged 6.16?6.03?/kWh (2011 $)--far higher and more variable than Ruedi because, being a run-of-the-river plant, Maroon Creek reportedly needs more disassembly and repair of damage from debris and finer particles. Ruedi Reservoir, in effect, settles and filters the water before it enters the turbine, and at CCEC, Thomas Reservoir would do likewise. The key question is how to use these two disparate data points to estimate CCEC's O&M costs. We need to start by expressing the empirical costs correctly. City Staff assumes CCEC's O&M costs will average one-half of Ruedi's or one-fourth of Maroon Creek's O&M costs, and supposes that these fractions respectively yield 0.85 and 0.72?/kWh in mixed nominal dollars (not a meaningful metric). But in constant 2011 $, which I added in Annex 5 to the City's revised data, half Ruedi's average cost would be 1.05?/kWh and one-fourth Maroon Creek's average cost would be 1.54?/kWh--substantially higher than the City assumed. In fact, the City has used none of these locally-grounded data anyhow; rather, it simply adopted for its CCEC projections the national-average 0.97?/kWh (perhaps in ~2011 $) O&M cost criticized above, then reduced it 24% by spreading it over 31% more kWh than CCEC is expected to deliver. Staff's rationale for applying to CCEC O&M costs one-half to one-fourth those of the City's two existing hydro plants is that CCEC: 1. is not run-of-the-river; that's correct. 2. is conveniently nearby; that's plausible but seems overapplied, because avoided commuting to the remote site applies at most to service-contract and perhaps labor costs totaling 76% of 2002-12 O&M costs, not to capital and equipment costs. 3. lacks physically extensive infrastructure because the penstock is treated as a Water-system asset; this is valid only if one accepts the emergency-drainline 40 The denominator is kWh delivered to the City's distribution system, just as CCEC power would be, net of Ruedi's 7.37% grid loss. The "?" shows one standard deviation. 16 rationale and if all penstock O&M is then booked to Water even though only half the capital cost is (the unusual accounting convention the City has adopted). 4. incurs only incremental costs "since necessary staff, equipment, etc. are already programmed & budgeted for Maroon Creek Plant and continue regardless of outcome of the CCEC project." That's dubious, since it assumes that the payroll hours to be booked to running and maintaining CCEC are currently surplus to operating requirements but that staff cannot be reassigned to do other needed tasks in that spare time (and likewise for equipment, like say a pickup truck, where the argument is somewhat clearer). This seems sensible in neither accounting nor economic terms. Each project incurs and should book its own costs without their depending on the existence of other projects. You can't book, for example, a maintenance worker's payroll and benefits cost hours devoted to CCEC as zero just because s/he is already on the City's payroll. I lack the data to quantify factors #2-4 or to assess their relative importance. There also seem to be two missing terms from CCEC's projected O&M cost. The first is the value of the penstock's projected O&M cost. I have asked for this number but not yet received it. Normal accounting may require at least half of it to be booked to CCEC's costs rather than 100% to Water cost. Citizens who reject the emergency-drainline rationale may expect all of it to be booked to CCEC. Second, CCEC would incur additional unique operating costs for streamflow studies, monitoring, and the Board of Experts. The City has given me its estimates 41 of these costs, shown in a column of its spreadsheet that is not displayed in the posted version but does feed into its financial calculations. These assumed costs for "Studies" total an undiscounted $410,000 during 2014-19, equivalent to an extra 1.32?/kWh (in 2011 $ present-valued to 2013) for those six years. This would be highly material to overall project economics if such costs continued. But they're assumed to be zero throughout 2020-2085, reportedly because the City's agreement with the Colorado Division of Parks and Wildlife doesn't require further studies or consultations. That assumption seems to me highly unrealistic, for two reasons: o o It assumes that which was to be proven, namely that monitoring will prove the streamflow diversion causes no significant environmental harm, so the assumed three-year 34%, then three-year 17%, flow reductions (based on the spring 2011 outcomes of community mediation), followed by zero reductions, will satisfy the Board of Experts and the public that the project fulfills its long-term environmental promises at its original design flow. But if the outcome differed, Aspen citizens would doubtless require further monitoring, meetings, and decision processes in a dynamic and unpredictable scientific and political process. Streamflow issues are unlikely to remain static during the following 65 years, when accelerating climate change is expected to precipitation scarcer and more erratic and spring runoff briefer. If that ococurred, more exacting measurements, 41 These cost estimates are said to be consistent with the contractor's previous charges and to include a series of meetings used a proxy for Board of Experts costs. 17 deeper scientific understanding, and more contentious public discussions would be needed, probably over a long period as conditions evolve, to balance fish and ecosystem health against "nearly free" carbon-free electricity production from a sunk-cost CCEC. I don't understand how this could add no cost. To be sure, discounting makes study costs throughout the life of the project at the currently contemplated rates add only ~0.3?/kWh to O&M costs, but that seemingly tiny difference knocks $0.38 million off the project's present value. On balance, though incorrect City data invalidated my original 3.1?/kWh (2011 $) O&M estimate for CCEC, the City's case for <1?/kWh remains sketchy and unpersuasive. I therefore suggest here a range of two O&M cost cases, tabulated at the end of Annex 5: o o a low case assigning all penstock O&M to the Water system, reducing by 50% the 2002-12 service and labor costs per kWh of the remote-site Ruedi O&M while keeping the same equipment and capital costs per kWh, assuming that costsharing with Ruedi and Maroon Creek O&M reduces all O&M costs by 8% (the amount the City apparently assumed when rounding down from half of Ruedi's real O&M cost to the ~0.97?/kWh [in perhaps ~2011 $] used in the financial analysis), counting no diseconomies of small scale for the 3.8x-lower CCEC output, and assuming that the initial six years' monitoring fully and permanently resolves all streamflow questions for the following 65 years, with no further monitoring or decisions required and no post-2019 curtailment. This yields a 2011-$ O&M cost of 1.14?/kWh, slightly above the City's half-of-Ruedi estimate of 1.05?/kWh real. It's also below the 1.27?/kWh (perhaps in ~2011 $) that national-average costs would yield if spread over CCEC's planned output. a high case assigning all penstock O&M (which, absent data, I guess might perhaps be 0.2?/kWh) to CCEC, reducing by 50% the service and labor costs of the remote-site Ruedi O&M and keeping the same equipment and capital costs per kWh but increasing Ruedi O&M cost/kWh by 100% to account for the 3.8x-lower rated output, and continuing the initial six years' monitoring and decision costs for the life of the project. This yields ~2.87?/kWh, 7% below my initial estimate of 3.1?. However, it may well be conservative because the 100% adder would be about 280% if all O&M costs were fixed rather than varying with the plant's output, and in fact, most O&M costs probably are fixed, not variable. The effort and investment needed to paint a pipe, fix a valve, clean a vent, or refurbish a turbine depends much more on whether it's there than on how big it is or (usually) on how many hours a year it runs. It's hard to consider the low cost case realistic, but not hard to imagine other plausible ways the high case could understate costs. For example, any post-2019 flow reduction would almost proportionately raise O&M costs because they're mostly fixed but would be spread over fewer kWh. Serious scientific or political disputes about safe flows-- plausible if water stress rises--could intensify study and decision costs. Bigger storms could damage infrastructure. In short, it's much easier to think of uncertainties that would raise O&M costs than of uncertainties that would reduce them. 18 This O&M cost range is admittedly a crude sensitivity test for this third variable, to which the project's economics are as sensitive as to water flow and electricity price. But it's much better than no sensitivity test at all, which is where we've been until now. And its basic lesson is clear: the ~1??/kWh range between the low and high O&M cost cases has a present value of $2.3 million--four-fifths of the project's calculated total net benefit, ignoring all its other weaknesses. So for the project to achieve a net benefit at all--albeit one far too small to interest any private investor--would require the City's environmental expectations to be fulfilled without controversy and without being disturbed by climate change; the project to be completed on budget and run smoothly for 75 years as expected, again with no loss of output from climate change or other contingencies; the project to displace relentlessly costlier coal power for at least most of that period (inconsistent with the climate-change assumptions, and implicitly assuming no carbon pricing to displace coal power with carbon-free alternatives that already beat it in the market); that CCEC costs about half as much to run per kW-h as Ruedi even though it's rated to produce 74% fewer kWh, as if three-fourths of its costs scaled with output rather than being fixed per plant; and that most of the City's other sanguine assumptions are all realized--all simultaneously. That's a pretty ambitious bet that I doubt any informed investor would care to take. One final note about Ruedi's empirical O&M cost: it would have been proportionately smaller if that project's output had been 17-25 GWh/y as intended and posted 42, rather than the 10.0-20.9, averaging 15.7 GWh/y, that it actually delivered during 2002-12. This shortfall in output is presumably due to the design problems detailed in the 2007-08 Canyon Engineering reports to the City. 43 I'll suggest below that Ruedi should finally be fixed as intended. But if Ruedi's output were indeed raised to its original design level, not only would Ruedi's calculated O&M cost fall correspondingly, but that adjustment would become irrelevant. That's because Ruedi's output would also rise by up to about 30% or 4.9 GWh/y--roughly displacing CCEC's likely output, probably at below CCEC's to-go cost. City staff tell me more recent data suggest Ruedi's actual potential output recovery now looks considerably smaller, but I believe they are misinterpreting the engineering report. Since this is an empirical question, I was glad to hear that they emphatically intend to pursue the Ruedi improvement project and get all the increased yield from it that proves feasible--the sooner the better. e. The City's only publicly reported methodology for choosing CCEC was unsuitable-- asking if CCEC would be cheaper than purchased power, not whether other alternatives could be even cheaper, faster, and less risky, hence able to save more carbon per dollar 42 Posted as if actual values at www.rwapa.org/facts_figures.html. The 2003-09 capacity factor of 0.368 stated in the NMPP report is 23% below the 0.479 implied by the RWPA website's average expected output: 21 GWh/y, vs. the actual 9-year inferred average of 0.368 * 8,766 h/y * 5 MW = 16.1 GWh/y. If the original design's expected output were correct (subject to analysis recommended in 2008) and a complete fix proved feasible, Ruedi's output could rise by about 30% (implicitly counting both relief of its 250-cfm flow limitation and across-the-board efficiency gains as the turbine ran nearer its optimal design conditions--the City's in-house experts seem to have overlooked the latter opportunity). Any reduction of turbine life from cavitation-induced erosion and vibration could also avoid future repair costs whose present value would help to offset the costs of lowering the tailwater elevation and reducing air entrainment. 43 19 and per year. There's a final serious problem with the City's economic analysis: its basic methodology. The report from Tier One Capital, whose expertise is mainly in financial and economic analysis, presents correct basic criticisms of the City's reliance on an avoided-cost methodology: the method City staff used is unsuited for testing the relative worth of alternative energy investments. Whatever the other merits or defects of the Tier One Capital report may be, I think its "Limitations of Economic Analysis" section is exactly right. It tersely explains why the City's cost-modeling approach based on avoided cost could not possibly support a sound decision about whether to prefer one alternative to another. To understand such choices, you would need a completely different approach that fairly compared the costs, uncertainties, risks, and other attributes of all available demand- and supply-side options, spanning their diversity of technologies and attributes. That's what well-run utilities and perspicuous regulators normally do. It's not at all what your staff did, and if done, it couldn't have supported the choices Council made. Staff respond that they used several analytic methods, not just avoided cost, but I haven't seen any other analyses, and suspect that if they were substantial and compelling, they would have been posted. f. The City's decision to proceed with CCEC did not receive proper independent review I'd hoped and expected to find that the City had commissioned a truly independent, highquality analysis of the project's economics vs. the diverse portfolio of alternatives, which I'll outline below. I was disappointed not to. What I did find, NMPP Services' December 2010 "Castle Creek Hydroelectric Project Economic Analysis," was dismaying. It simply compared the City's financial analysis with its own version that slightly tweaked the City's, reached similar conclusions, and found that the City's "assumptions are reasonable and the savings can justify moving forward with construction of the project." NMPP said its review "is to help Aspen determine if the economic analysis provides a reasonable expectation of the financial value of the project," and that "NMPP has used its experience and research to determine whether all significant factors have been considered and assumptions are within industry expected ranges." But NMPP's 14-page report analyzed no sensitivities to coal-power price or O&M cost, adopted the City's 13.3-cfs flow assumption, and considered no investment alternatives. It usefully fine-tuned some of the City's assumptions, but didn't meet the most basic requirements for the "independent analysis" and "independent review" it purports to be. In fact, it's not even independent. As its logo and website show, NMPP is a sister organization of the City's electricity supplier MEAN, with the same address, under the same NMPP Energy organization, and doubtless with collegial if not overlapping staff. It's no criticism of these two Nebraska organizations, whose useful services are widely used by public power in at least four states, to say that they are not independent of each other. But NMPP neither looks nor is independent of Aspen because of Aspen Electric's long business and personal relationships with MEAN. So it's now clear to me, as it's long seemed to many, that CCEC is economically 20 unsound. Because the City must pay for what it's spent, whether the project proves productive or a dry hole, continuing to build CCEC would end up raising your customers' electricity prices, at least for a decade or two and perhaps permanently. (Long-term trajectories depend on many imponderables including climate change, the plant's costs and reliability, MEAN's economics, and the likely downward cost trend for clean competitors.) To be sure, the overall effect on electric rates should be modest: if 8% of your electricity cost, say, three times the average cost (it nearly would, as we'll see below), it'd raise your average rates only 3%. But the political effects may be disproportionate, especially when Aspen rate hikes get unfavorably compared with those of Holy Cross customers. However, depending on how you classify or salvage sunk costs, stopping now and using savings from cheaper alternatives to prepay (starting 2019) or defease (anytime) CCEC's bonds could greatly reduce and perhaps eliminate its upward pressure on rates. And very fortunately, many attractive solutions, probably more than you realize, are available. I'll turn next to these alternatives within their decision context, then recommend some ways forward. 2. CCEC has higher costs and risks than available, ample, timely, and suitable alternatives, even neglecting its sunk costs and counting only its to-go costs To help you understand the full spectrum of paths open to the City, let me offer a simple little calculation in real terms, factoring out monetary inflation, then distinguish costs already sunk from "to-go" costs not yet sunk. When I first analyzed this project in January, the City projected 44 a total budget-- apparently a bare capital cost and probably an overnight cost 45--of $10.5 million, of which $3.646 million was stated to be not yet spent (the "to-go cost"). As noted in Annex 3, item 5, this posted budget does not include the financing cost of any of the capital-- that's extra--but it should have, and it should now. That has nothing to do with accounting rules; it's about political transparency and honest budget comparisons. CCEC power delivered to Aspen's distribution grid through 2035, in constant 2011 dollars, would cost roughly 15-17?/kW, of which about 6-8?/kWh isn't yet spent. Both figures exceed Aspen's 6.2?/kWh price of wholesale power. I've performed a simple but instructive cost calculation (Annex 6) using a common utility analytic technique to derive the levelized 2011-$ cost of a net kWh delivered from CCEC from now to 2035, assuming the $10.5-million budget, the hoped-for 5.476 GWh/y of delivered output, and no further curtailment due to stream conditions, water rights, or climate change. I found that the full CCEC project would generate power at about 15- 44 www.aspenpitkin.com/Living-in-the-Valley/Green-Initiatives/Renewable-Energy/Hydroelectric/Finances/ That's the conventional term for construction cost if done overnight, i.e., incurring no interest cost and no real cost escalation during construction. 46 Some sources indicate $3.1 million, but knowing that the City tries to keep its citizens up-to-date, I take the previous reference, accessed on your website 13 January 2013, as authoritative. The difference is apparently due to the tailrace that must be built at the downstream end of the penstock, if the hydro plant isn't completed instead, to dissipate the water's energy in case of emergency-drainline use. 45 21 17?/kWh47, while its to-go portion would cost about 6-8?/kWh. So looking as far ahead as 2035, CCEC's to-go cost probably exceeds both the originally budgeted total project cost (6.2?/kWh in 2011 $) and the City's current wholesale price of 6.2?/kWh. The total project cost is approaching three times the current MEAN wholesale price. Hydro's potentially very long operating life, conventionally assumed to more than make up the financial loss to ~2035, has now become less an advantage than a drawback in today's radically changing electricity market, where modular alternative investments in mass-produced efficiency and renewable technologies like wind and solar power can now flexibly exploit their rapid evolution and steeply falling cost, yielding even lower overall cost and risk. Wait a minute, I hear you saying: this now-to-2035 perspective doesn't count the presentvalue benefit another half-century of operation at just the O&M cost plus its real escalation. Well, yes, but. The levelized-cost methodology I illustrated in Annex 6 is normal and useful for power plants that are built fairly quickly with a known construction-cashflow profile and are financed over their operating lives. Unfortunately it becomes much less useful for computing lifetime levelized costs when, as in this case, the plant is financed over 28 years starting at least six years before it enters service, its construction cashflow profile isn't publicly known, once operational it runs for about 75 years, its long-term O&M costs are increasingly speculative (but economically critical), and so are the long-term costs of the competing resources it displaces (even their direction of change is unknown). It's also not possible to compare my real-cost calculation with the City's analysis, which uses nominal costs but with unknown monetary inflation rates. I'm therefore not suggesting you give too much weight to my now-to-2035 cost calculation, even though it's methodologically orthodox and I think it's correct in its own terms. Rather, I'd suggest you note its conclusions because near-term costs and electric rates are important and your political discount rate is probably pretty high. To check my conclusions more rigorously, City staff should update their spreadsheet to count all costs of capital (with and without partial allocation to Water accounts), use a realistic range of O&M costs including climate-change cases, and spread O&M costs over 5.476, not 7.2, GW/y of production. Then they should graph transparent and multivariate sensitivity tests to a wide range of waterflows, O&M costs, and coal-electric real price escalation--the latter including at least -1%/y to reflect the range of EIA forecasts. Your decisions should rest even more on these sensitivities and comparative risks than on point-value estimates--staff's, mine, or anyone else's--and should note the causal links that would probably make unfavorable trends apply multiple key variables 47 The City calculates that the average cost of CCEC over its assumed 75-year operating life would be 6.2?/kWh in unstated roughly-contemporary dollars, "considerably less" than its current 6.7?/kWh MEAN power purchase price (www.aspenpitkin.com/Living-in-the-Valley/Green-Initiatives/RenewableEnergy/Hydroelectric/Why-Hydro-/). Oddly, that's the same as CCEC's average electricity cost calculated when the project was originally budgeted at half its currently projected total capital cost! I've since learned that this impressive consistency was achieved by increasing the originally assumed 30-year amortization life to 75 years. 22 simultaneously. But there's another and much bigger reason why I'm comfortable giving little weight to hydro's potential long-term economic advantage of running for many more decades at just its operating cost after its debt is paid off. If the hydro plant really does run reliably and cheaply for 75 years (which assumes little climate change, inconsistently with the conditions needed for high coal-electricity prices), then the 75-year perspective was valid years ago, when the standard alternative was big thermal power plants lasting 30-40 years. But it's no longer valid. Please stay with me here. This is about fundamental electricity strategy, not numbers. Today's competitive generating options--the huge range of efficiency and modern renewable technologies, chiefly wind and solar--have steep learning curves and extraordinary innovation, with no end in sight. Reinventing Fire conveys the full implications of this revolutionary change, but in brief, the means of generating or saving electricity are no longer like building a cathedral over many years. Today they're a scaleable, mass-produced, manufactured product, more like microchips and computers. I recently visited a single Chinese factory that stamps out 2.5 billion watts of top-quality solar cells every year like cookies, 24/7. That's the new electric age, coming right at us: half the world's new generating capacity ever since 2008 has been renewable, non-hydro renewables are winning a quarter-trillion dollars of global private capital investment every year, ~42% of U.S. capacity additions last year were windpower, and reportedly all utility capacity additions in March 2013 were solar. So the game has utterly changed-- and likewise across the vast range of efficiency technologies. This gamechanger makes the venerable strategy of sinking high capital investments into very durable generating assets into a poor strategy because it locks you into old, static technology. It stops you from riding small, fast, granular technologies down their steeply falling cost curves and exploiting their breakthrough innovations as they occur. And it makes you rely on a few discrete, "lumpy" technologies whose single-unit failure is very costly and disruptive, rather than a diversified portfolio of many modular devices that won't all fail at once or in the same way, giving you safety in numbers, just like a diversified stock portfolio. (I wrote an Economist book of the year on this theme in 2002, and it's now the industry bible.48) For these subtle but powerful reasons alone--let alone long-term climate-change and perhaps water-rights issues, the former intensifying the latter, that could turn long-term hydro into a white elephant 49--I think a sound Aspen energy strategy will not consider new hydro's potentially very long lifetime as an advantage justifying its high up-front 48 A.B. Lovins et al., Small Is Profitable: The Hidden Economic Benefits of Making Electrical Resources the Right Size, Rocky Mountain Institute, 2002, www.smallisprofitable.com. This work is important for your strategy because it documents 207 "distributed benefits" that can often make distributed (decentralized) electric generators, or savings, as much as an order of magnitude (about tenfold) more valuable. 49 CCEC bets not only on static technology and relatively stable market conditions but also on long-term climate stability--whose unlikelihood was originally the project's main justification. 23 cost. On the contrary, it could prove a serious handicap to our grandchildren living in a very different world. It would be like doubling down on copper wires and telephone switching relays just before cellphones hit the market. Competitive alternatives can do the same job with less cost, risk, and probably time So looking at those more-predictable CCEC levelized costs to 2035 when the last bond matures, a diverse portfolio of cheaper and currently available alternatives can collectively displace CCEC many times over. These competitive alternatives offer comparable or greater coal displacement, comparable or faster speed, and different but generally lower risks. Here are the main elements: o o o o Even under conservative assumptions, CCEC's to-go cost--about twice that of building the Ruedi hydro plant--exceeds by severalfold any plausible cost of equivalent energy efficiency additions. Efficient use of electricity is your biggest, fastest, cheapest, and most politically winning option. More below. Among the four hydro alternatives that Kennedy/Jenks Consultants recommended more than a year ago be assessed, three existing-infrastructure microhydropower options are individually small but collectively useful and if feasible, may prove cost-effective. Their fourth hydro alternative--Ruedi repairs--might yield nearly as much additional hydroelectricity as CCEC's nominal output (or comparable to its output under curtailment for healthy streamflow or climate change). Based on Canyon Engineering's clear and capable reports to the City in 2007-08, I expect this option could easily beat CCEC's to-go cost. The needed repairs are much smaller and cheaper than building a small hydro plant in the first place. Ruedi hydro, piggybacking on BuRec's dam, cost Aspen a total of ~$900/kW in 1983-85, equivalent to ~$1,700/kW today--less than a fifth of CCEC's total real cost, and far below its to-go cost of about $3,060/kW.50 Importantly, too, the Ruedi upgrade would not increase and would probably decrease its own future O&M costs, but could entirely avoid CCEC's O&M costs. I infer that gaining comparable power output from Ruedi repairs can probably beat CCEC's to-go. Five years after Canyon Engineering's analysis and recommendations, it's time to dig into it more seriously. City staff tell me they're eager to do it. Let's go find out what it can do. New Windbelt windpower could almost certainly beat CCEC's to-go cost. Longterm Power Purchase Agreements signed in 2011-12 for new windpower in that 13-state region (which includes Nebraska) averaged 3.2?/kWh in levelized 2011 $ and ranged from about 2.5 to 4.0?/kWh 51, so net of the apparently ~0.4?/kWh wheeling charge and 2.5% losses to Aspen, they'd easily beat CCEC's to-go cost. 50 I.e., $3.6M to-go cost (note Error: Reference source not found) divided by the FERC permit's nameplate rating of 1.175 MW. 51 All these prices are from the authoritative Federal annual review of the industry, LBNL-5559e, 2012, p. 52, http://eetd.lbl.gov/ea/emp/reports/lbnl-5559e.pdf. City staff have mentioned challenges for MEAN or the City to capture the tax benefits available to private investors. This puzzles me, since a rather large financial industry is devoted to structuring deals that enable public entities to capture those benefits' value. That levelized value in 2011 $ is ~$18/MWh at a 3%/y or $28/MWh at a private 18%/y real discount rate. 24 o o o These new-wind prices, still dropping, have already fallen so dramatically that MEAN's current wind price to Aspen of 6.7?/kWh 52 and its reported late-2011 forward offer price of 8.9?/kWh have become uncompetitive. Such rapid price shifts underscore the need for an agile, flexible, fast-learning utility strategy. If the City has gotten locked into long-term, high-cost wind or other power contracts, especially with beyond-inflation escalators, Council should find out how that happened so it won't recur, and consider how to fix it (Annex 2, #9b). I'm not clear on whether MEAN offers new windpower at competitive and fixed prices, nor on why it shouldn't. Seasonal balancing of wind with Aspen's loads is a real and more complex issue but should be amenable to creative solutions, perhaps including "wind-banking" swaps with other wholesale customers. The Tier One Capital report said MEAN, more than a year ago, could provide carbon-offset thermal power for 6?/kWh or surplus 100%-green power--wind, WAPA hydro, and landfill biogas--for 8.9?/kWh; it's unclear whether that's FOB Nebraska or wheeled to Aspen, and whether and how it's levelized in what year's dollars, but the differences seem minor. Tier One Capital reported that the City passed both on buying more MEAN renewable power and on local solar power projects with levelized costs probably around 8-9?/kWh. If their costs were instead actually around 20?/kWh as Mr. Overeynder tells me, it was right to pass, but I'm puzzled by the discrepancy, and think the range of such offers now merits prompt and expertly updated review independent of MEAN. Interestingly, when Council sent the bond issue to the voters in 2007, its decision memo found windpower competitive with the then-projected total project cost (and thus below the 24%-higher current to-go cost to 2035), but there wasn't more windpower on offer; now there's plenty and it's even cheaper. Yet Mr. Dolan's 20 December 2012 email shows the City has commissioned 15 consultancy studies on hydropower since 1974, but only one on windpower--in 1998, when windpower cost more and the U.S. had 97% less of it than in 2011. Mr. Dolan's updated February 2013 list shows consideration of 19 hydropower, 5 microhydro, 3 wind (the last in 2007), 3-4 photovoltaic, and two other projects. But project-specific studies aren't good at illuminating fast-moving market conditions, of which I wonder if staff have an adequate overview. My impression is that they don't yet. The City's continued use of a highly unsuitable EIA generic levelized-cost table in its just-published 2013 renewable-energy analysis reinforces that impression. Local generating options, such as solar farms and landfill or digester biogas, may even better fit your criteria and desire for energy security. They merit equal consideration. Apparently a PV project at Burlingame is now being discussed. Of course, significant PV deployment could quickly run into the MEAN contract's obsolete prohibition of "behind-the-meter" generation over 2% of requirements.53 If MEAN was offering more WAPA power to Aspen a year ago, perhaps Aspen 52 www.aspenpitkin.com/Living-in-the-Valley/Green-Initiatives/Renewable-Energy/Hydroelectric/WhyHydro-/; however, NMPP also expects 4%/y contractual escalation, which sounds way above monetary inflation but shouldn't exceed it, since windpower's costs, once a windfarm is built, are essentially constant for its life, just like hydropower. If MEAN is indeed entitled to real price escalation (though I have no idea how it could justify that), MEAN wind's levelized cost to Aspen would be higher than 6.7?/kWh. 53 That might become moot as efficiently financed and installed solar power becomes cheap enough that customers can justify dropping off the grid, removing them from the City's "requirements" total. 25 o Electric could buy it directly from WAPA as "preference power"--as I understand already occurs to some degree--though I don't know if such an offer is still available. I was told that WAPA offers firmed windpower for around $100/MWh, a $20-30 premium, but that premium is about ten times what WAPA charged for firming wind a few years ago, and the price seems far above market. Also increasingly relevant, but not yet on Aspen's agenda, are demand-response opportunities--ways to make demand unobtrusively flexible. For example, my electric car charges at a variable speed controlled by grid frequency, dispatching up to 7 kW of valuable one-second fast-regulation services to the Western Interconnect. Electric water heaters, pool or driveway heaters, and ice-storage air conditioners are among the many end-uses where commercially available hardware, with no inconvenience to the user, can help Aspen to manage loads and integrate variable renewables. I don't think this has yet been analyzed. From now to 2035, these options straddle the best-case to-go levelized cost of the project (nearly 7?/kWh), could undercut its ultimate cost into the 2080s if its completion and operation didn't go well, and most importantly, could avoid its major risks. And this list of options is far from complete. SkiCo just brought in a great project using coal-mine vent gas.54 The big hotels and AVH probably have major co- and trigeneration opportunities that they may not have tapped (the state of the art is 93% system efficiency). New drilling and downhole techniques like those mentioned in Reinventing Fire, p. 192, could make geothermal heating cheap; congratulations for starting to explore that resource. To put it another way, if CCEC's to-go costs to 2035 were a standalone new project, which is how its proponents like to think of it (since sunk costs are forever sunk and you can make decisions only about the future, not the past), its roughly $3,060/kW projected capital cost wouldn't look low for a new U.S. small hydro plant. That to-go cost would also be two or three times the cost per kilowatt of a new windfarm in a good Windbelt site, but with less than twice its annual output per kilowatt (the U.S. national average wind capacity factor in 2011 was 33%). And CCEC's to-go cost to 2035 is many times the levelized cost per kWh of end-use electric efficiency, especially in commercial buildings. Let's dig into that option for a moment, because it's your most important and exciting path to a solution. And the silver lining of cost-effective solutions is that every kWh you can save, or generate by any alternative cheaper than MEAN's 6.2? power (not hard to do), saves money that Aspen Electric can use to pay CCEC's debt and hold down rates. 3. A new approach to meeting Aspen's electricity policy and to eliciting and benefiting from authentic public participation could pay major dividends to all parties 54 An option I urged the coal company to push about a decade ago, then tried in vain throughout two G.W. Bush Administrations to get BLM to figure out how to permit: they couldn't figure out how to add a gas permit to a coal mine. Evidently they now have. 26 Aspen's most attractive electric-services options in cost, scale, speed, and risk profile-- probably rivaled only by the Ruedi repair--are in end-use efficiency. If you haven't recently visited my "passive-solar banana farm" in Old Snowmass--one of the world's most efficient and integratively designed buildings, a net exporter of solar electricity, with no fossil fuel or combustion, yet producing 47 banana crops so far with no furnace --let me invite you all to come for a tour and a private chat. I think you'll be astonished by the new technologies we've lately retrofitted (since the original building was turning into a museum of 1983 state-of-the-art). Lighting retrofit opportunities One of the recent improvements was our fifth lighting (and first daylighting) retrofit, designed and installed by one of the world's best practitioners of superefficient and beautiful lighting--Robert "Sardo" Sardinsky55 of Rising Sun Enterprises in Basalt. One of my wife Judy's big landscape photos in the front hall looks better with 12 watts of modern, color-accurate LEDs than it looked in her old Hill Gallery of Photography with 500 watts of incandescents. This illustrates what may well be Aspen's most juicy, highvisibility, and politically potent electric-services alternative: a mass retrofit switching to modern, beautiful, superefficient lighting, starting with commercial buildings because they tend to have the biggest lighting loads and the longest operating hours. Your programs and allies have of course been doing Aspen lighting retrofits for a long time (some with Sardo and other capable folks). But the technology, and the art of lighting design that should inform and apply it, are moving with extraordinary speed: LED technologies change literally every week. Whatever your or CORE's programs installed a few years ago is probably already worth replacing. It would be interesting to go on a walkabout in the Aspen stores, restaurants, and lodgings with Sardo and me and see what we observe. Some Aspen businesses have pretty good lighting efficiency, a few exemplary, but most do not56; any installation before 2009-10's strict new codes is a likely retrofit candidate. Even in relatively well-managed premises, my casual observations continue to show enormous lighting retrofit opportunities that would repay the owner with excellent financial returns but the same or often better merchandising and guest experience. As you can see at our house or at Sardo's display center at the Basalt business center, superlative color rendering, sparkle, dimmability, extreme durability, and any other attribute you might want (except lots of waste heat!) are all off-the-shelf. Commercial lighting modernization is especially valuable because it could so improve offseason cashflow and hence the survivability of marginal enterprises. It should also help Aspen Electric by unloading aging distribution circuits and by saving kWh in the winter when lighting hours are longer and the local system peak is bigger. It saves more carbon because in the winter, local hydro is more limited and wholesale purchases of coal power are higher. It will build local skills and support good local jobs. And besides making downtown buildings more attractive and their goods likelier to sell, the beauty 55 He launched his career coauthoring my encyclopedic 1988 synthesis The State of the Art: Lighting, showing how to save ~92% of U.S. lighting energy with 1-2-y paybacks. 56 Aspen buildings aren't greatly different from buildings in other places with decent building codes. Only in the past few years have our local lighting-efficiency code requirements reached sensible stringency. 27 and visual effectiveness of modern superefficient lighting could add the kind of new tourist attraction that Soldier's Grove, Wisconsin, achieved when it used its move out of a floodplain to add daylighting, and unexpectedly found shoppers coming from faraway to enjoy the more attractive shops. Such work must be done right. You may recall that after Sardo's Council Chamber lighting retrofit, ballasts and lamps kept randomly burning out for a couple of years. It took that long to discover, a few months ago, that five different electricians had been consistently miswiring the ballasts by ignoring their labels' wiring diagrams! Problem solved, but such annoyances teach us the importance of knowledge, training, and quality control. Fortunately, many of the biggest opportunities are simple screw-in or plug-in replacements, like LED versions of those power-hungry and very costly MR16 halogen lamps. Much commercial lighting is so standardized and commoditized--especially back-of-house and in service-business clusters--that a mass retrofit, one building complex or City block at a time, can yield huge economies of scale, with many replacements as simple as picking parts out of a shopping-cart of standard plug-in equipment and swapping it for what's there. There are similar economies of scale in parking-lot lighting: the best LED luminaires like Kim Lighting's Warp9 can use just 2% of the normal ASHRAE-standard lighting watts per square foot, or 1% with light-colored paving, but they look better and you see better. In much exterior architectural lighting, savings around 70-90% are not unusual, with better esthetics and durability. At and after the City's evening citizen openhouse, I had valuable discussions with City and CORE efficiency staff and was delighted to hear of their recent progress, including the 13 City buildings retrofitted by McKinstry. Aspen lighting retrofits have generally matched or exceeded my expectations. For example, the parking garage saved over half its lighting energy with a 3-3?-year payback, while Aspen Square saved at least 80% of lighting energy with a 3.4-year payback and high satisfaction. Encouragingly, Kincaid Gallery's lighting retrofit has since spread to two more galleries--often among the most demanding and skeptical customers. Energy Smart Colorado is assembling all these and other Aspen data, aiming to have a concise summary by the end of April for somewhere between a dozen and ~30 downtown projects already undertaken. That's plenty of data to design an aggressive program around. The lodging sector is especially promising, and a hotel makeover contest already underway in Aspen reportedly has 11 signed, 8 participating, and 3 active establishments so far. Your leadership can expand that. Other retrofit opportunities Of course, there are many fat targets besides lighting. Wandering around town, I keep seeing juicy opportunities in supermarkets (which RMI has redesigned for 40-60% savings, many retrofittable, in several major chains--but City Market, for example, has one engineer maintaining 164 stores). There are more in the ratty old fans and pumps and motor systems I keep seeing and hearing, in data centers like Quest's, and in the ubiquitous office equipment that could save most of its electricity just by better-informed shopping, with equal attributes and no greater capital cost. 28 And there are new ways to do old things, like RMI's integrated package retrofit for electric water heaters that can save ~70% with a one-year payback, or the LBL aerosolized-chewing-gum spray-sealant to plug hidden duct leaks from the inside--not to mention all the retrofit techniques, from simple weatherization to superwindows to closing exterior doors, to save electric space-heating. If we do a walkabout, I'll bring my high-resolution infrared camera to spot how some buildings are still space-heating by heating outer space. There are only a couple of air-curtain shops in town. A winter survey of closed-vs.-open-door merchant policies, and perhaps some Breckenridge-like signagepolicy tweaks, could help make closed doors the new normal. It's some years since I had close contact with the practitioners who work with Canary, CORE, Holy Cross, and other Valley efficiency-program leaders. But early discussions confirm that there's an exciting opportunity here to surge new efficiency that works astonishingly better and costs less than traditional methods. I'd be happy to discuss how I, and RMI, might be able to help find and capture these targets. Some such efforts will also save lots of natural gas (hence CO2) and water (hence electricity for treating and pumping). Financing and delivery innovations Aspen has another exciting opportunity to innovate in financing and delivery channels. Commercial-building PACE bonds, or Aspen Electric on-bill financing if you want to keep it on the City's balance sheet, could overcome landlord/tenant split incentives by removing the up-front capital burden and yielding positive cashflow from day one. Savings from these extremely high-return investments could be split with Aspen Electric to help prepay debt. Or Aspen Electric could lease the lighting equipment as a "lending library"--including equipment mandatorily saved when a tenancy rolls over and the space is normally gutted, often sending perfectly good, already-installed, and almost infinitely durable lighting gear to landfill when it could be saved and recycled to the next tenant or other customers. (It'd be even better for landlords to leave lighting equipment in place, potentially avoiding reinstallation cost--especially for controls--until the next tenant has been found and has explicitly rejected it; then it could return to Aspen Electric's inventory. A conversation with landlords, brokers, and merchants seems timely: in such a small town, it shouldn't be so hard to move this durable, fungible stuff around and think of its negawatts as the City's "distributed power plant.") You might be wondering how much total electricity a commercial-lighting mass retrofit can be expected to save. I don't know, because actual analysis would need to be based on estimates of what electricity is used for in Aspen, and I've never seen Aspen end-use data. But U.S. commercial buildings in 2003 used an average of 38% of their electricity for lighting; non-mall retail buildings used 53%.57 Early Aspen projects confirm that average lighting-energy savings around 50+% are realistic without, or 70-80+% with, controls (which add cost, complication, and skill needs), so in round numbers, saving at least half of 38% saves at least 19%. Thus after allowing for the commercial-building 57 EIA CBECS database, 2003 (the latest available--another is currently underway), Table E3A, www.eia.gov/emeu/cbecs/cbecs2003/detailed_tables_2003/2003set19/2003pdf/e03a.pdf. 29 fraction of Aspen's total electricity usage, we're probably pretty close to saving 11% of total consumption and thus making Aspen's electricity 100% renewable in the next few years just from commercial lighting retrofits. An expert meeting, including your staff and others who know Aspen Electric's customers, could refine that rough estimate, but it's probably not far off, and if it were, there'd be plenty of other straightforward efficiency targets that a mass retrofit could catch on the same visit. It's easier to do a pilot experiment quickly than to do a study, because we know this target is plump and hard to miss, whatever we get will be good, and it's better to start now and learn quickly and cheaply in a small, representative group of buildings than to delay and ruminate. In these circumstances, a bias for visible action that clearly benefits everyone will be economically valuable, signal openness and decisiveness, and help rebuild fractured coalitions and regruntle disgruntled constituencies. Besides obvious pilot opportunities in Aspen, some in the Holy Cross territory could be useful too, notably at the AABC, and could help better coordinate the two utilities' efforts. Some unsolicited political advice I've never been an elected government official, but I've worked with them, from 22 heads of state to hundreds of councilors, mayors, governors, and cabinet members all over the world. I have immense respect for the difficulty of their jobs and the dedication of their effort. It may seem presumptuous to offer you political advice on top of all this technical and economic stuff, but let me try anyway and hope it helps. CCEC's total cost may make it the costliest hydro plant ever built--hardly the legacy you strive for. It shouldn't and wouldn't have been started if it had been properly compared with many alternatives then available. Hence the importance of transparent analyses and choices now as you mull over some tough choices. It was an economic and strategic blunder to dismantle Castle Creek hydro in 1958 in pursuit of slightly cheaper grid-delivered coal power, because better options then emerged. (Aspen was in good company, though--many other small-hydro operators made the same mistake.) Paradoxically, in today's profoundly different and climateaware conditions, rebuilding Castle Creek hydro would be wrong too, because better options for displacing coal power are already widespread and getting rapidly even cheaper and better. But these opposite outcomes flow from the same cause. In both cases, inadequate consideration of available alternatives and strategic risk management caused bad decisions; input from Council's technical advisors was either ill-informed or misinterpreted; and input from the public, which in this region includes world-class independent experts, was improperly solicited and inadequately considered. My five broad suggestions for a public process are: 1. Recognize from the outset that this is all about 8% of the City's electricity. That's really not much. Smart firms have cut their energy intensity 6-16% per year, some countries by over 5% per year, chiefly through modern end-use efficiency. You're only 30 seeking 8%, ideally 11%, spread over three years. For those of us who are practitioners, not theorists, and do solutions, not problems, that's about as close as it gets to a piece of cake, and your menu lists lots more than cake. 2. Consider all options impartially--all ways to save or produce electricity to deliver desired services, regardless of their type, technology, size, location, or ownership. Consider their diverse attributes as a richly complex tradespace, not a rigid checklist or screening process. Consider options not just singly (and especially not requiring any one option to do the whole job) but in sensible combinations to capture synergies and maximize societal benefits. 3. Focus first and mainly on the demand side, because there's overwhelming evidence that end-use efficiency and demand response nearly always cost less, usually by severalto manyfold, than incremental supply. Then make your demand-side portfolio systematic, comprehensive, and modern--and preferably informed by integrative design58 (featured in Reinventing Fire), so you can often achieve expanding rather than diminishing returns to investments in end-use efficiency, making very large energy savings cost less than small or no savings. This isn't yet a common practice in the Valley nor adopted in City analytics, but your folks are plenty smart enough to do it. It took me a half-hour to teach our master plumber, 30+ years in the trade, layout tricks that cut our new-piping friction 97%. I'd be happy to offer that class to all Aspen's or the Valley's plumbers, pipefitters, duct-fitters, contractors, architects, and engineers. 4. Consider all realistic supply-side options too, like the half-dozen or so I listed above. Even if the City entered a long-term windpower purchase contract at the top of the market, new windpower should be procurable at far more competitive prices. A truly independent review of MEAN's prices, terms, and conditions may be timely. Though MEAN is in a tough spot, committed to coal power that an increasing number of its members want to use less of, Aspen has been able to work very creatively with MEAN to overcome traditional restrictions and achieve a remarkably high renewable power fraction. A senior RMI colleague and I have offered, and City staff have welcomed, some free consultation about potential ways to increase flexibility in the MEAN relationship and in handling month-to-month imbalances such as more windpower might cause.. 5. Transparently, dispassionately, clearly, and publicly compare economics and risk profiles for all options including CCEC. Otherwise no meaningful, valid, and useful conversation can occur. The public will be exquisitely sensitive to any whiff of bias. I doubt that the public, especially the "no" voters, will believe anything fundamental has changed until there is a clear and unambiguous signal that minds are open and rational exploration of better ways forward can begin. (Sorry, but Mr. Dolan's well-intended email of 20 December 2012 isn't it. Remarks to individuals or reporters aren't either. I mean a clarion signal, not muttering.) 58 My two 2010 short papers at www.rmi.org/Knowledge-Center/Library/2010-09_IntegrativeDesign and http://www.rmi.org/cms/Download.aspx?id=4984&file=201010_10XEPrinciples.pdf&title=Factor+Ten+Engineering+Design+Principles introduce the concept, and Reinventing Fire is full of practical examples. 31 At this stage we need managerial reflection and repair, but what we most need is lucid, inspiring, unmistakable leadership. Some potential elements of the signal might include: - - - since the first rule of holes is, when you're in one, stop digging...explicitly suspend construction (perhaps already done59, but so far only de facto) pending the outcome of the next three steps; engage truly expert advice on quick, cheap alternatives and on how best to work with MEAN to expand the mutual benefits of more diversified portfolios; set up a one-day expert practitioners' workshop to design a fast-learning, fastspreading rollout of an aggressive City-wide "blitz" on commercial lighting mass retrofits (I'd be glad to host and convene such an expert session at my home) building on pilot projects already performed, and in parallel, launch a deftly and fairly facilitated, reasoned, all-cards-on-the-table external dialogue--ideally integrated with the update of the 1990 comprehensive water management plan--whose neutrality, openness, and integrity are above reproach. I would be willing to devote a little of my private time to such a conversation if I felt it met these criteria. The economic comparisons in #5 above will be critical. I suspect that such a process could well identify and produce dollar savings from efficiency alone big enough to cover a lot of the sunk costs 60--an approach we've developed for multi-billion-dollar abandoned power plants elsewhere over the past few decades. Advanced energy efficiency is so much cheaper than buying power from MEAN that arbitraging the spread in financial returns could be structured to return to the City a substantial cashflow from MEAN savings to recoup its project losses, while still stabilizing or cutting customers' rates and markedly reducing participants' bills. There's probably more I've missed, but this will get you started. If you'd like to meet, just let me know. Meanwhile, thank you for your kind attention...and three deep breaths! With best wishes-- 59 www.aspendailynews.com/section/home/155560, 13 Nov. 2012, reporting the City Manager's remarks in the previous day's Council meeting. 60 The politics of those sunk costs could also be interesting. Nearly half the sunk cost is for the "emergency drainline" whose necessity you will doubtless feel bound to defend. Apparently the raw-water improvements (debottlenecking the upstream supply pipes), emergency drainline, and tailrace would all be charged to Water accounts if CCEC weren't finished. You could probably recover something like half the $1.5M cost of the turbine by reselling it, assuming it's too late for any force majeure escape hatch that might be in its procurement contract. Or if you kept the turbine, the raw-water and soft costs could perhaps be treated as an investment in a potential future resource that the City now defers but could still choose later to pursue if and as needed and cost-effective--a sort of "electricity reserve" asset that might then stay on the balance sheet rather than being written off. This approach might also help bolster your case for intended maintenance of the water rights, much as the City has reportedly cited its unadopted 1990 water plan and subsequent policies in support of filings to maintain conditional water rights on Maroon and Castle Creeks (www.aspendailynews.com/section/home/156238, 6 Jan. 2013). 32 Amory B. Lovins 33 Annex 1: Biographical sketch of Amory B. Lovins Recovering physicist and energy innovator Amory B. Lovins has advised the energy and other industries for four decades and the Departments of Energy and Defense, and is a member of the National Petroleum Council and an advisor to the Chief of Naval Operations. He is considered among the world's leading experts and practitioners in energy (particularly its efficient use and sustainable supply) and its links with economy, security, development, and environment. His work in 50+ countries has been recognized by the "Alternative Nobel," Blue Planet, Volvo, Zayed, Onassis, Nissan, Shingo, and Mitchell Prizes, MacArthur and Ashoka Fellowships, 11 honorary doctorates, and the Heinz, Lindbergh, Time Hero of the Planet, National Design, and World Technology Awards (among others). A Harvard and Oxford dropout, former Oxford don, honorary U.S. architect, and Swedish engineering academician, he has briefed 22 heads of state and been sole or main author of over 500 publications including 31 books, such as Natural Capitalism, Small Is Profitable, Winning the Oil Endgame, and Reinventing Fire. The latest of his ten visiting chairs were in Stanford's School of Engineering and (currently) at the Naval Postgraduate School. His senior advisory relationships have lately served the leaders of such firms as Coca-Cola, Deutsche Bank, Ford, Holcim, Interface, and Wal-Mart. Cofounder in 1982 and now Chairman and Chief Scientist of Rocky Mountain Institute--an independent nonprofit think-and-do tank that drives the efficient and restorative use of resources--he has led the superefficient redesigns of scores of buildings, several vehicles, and $30+ billion worth of industrial facilities, and the creation of three of RMI's five for-profit spinoffs, including the leading source of technical and strategic information on advanced energy efficiency and distributed generation. He did the conceptual and energy design for his 1984 owner-built home (and RMI's original headquarters) in Old Snowmass, still one of the world's most efficient and integratively designed buildings. He has advised more than 100 electric utilities worldwide, led perhaps the world's most detailed synthesis of advanced electric efficiency and cofounded and -led PG&E's ACT2 experiment on it, given expert energy testimony in eight countries and over 20 states, and devised many of the fundamental concepts and practices that underlie modern electricity strategy. The Electricity Journal wrote, "No advocate has had a greater influence on the changing orientation of U.S. utilities toward energy efficiency--even the terms we use to talk about it--than Amory Lovins." Dr. Alvin Weinberg, former Director of Oak Ridge National Laboratory, called him "surely the most articulate writer on energy in the whole world today"; Newsweek, "one of the Western world's most influential energy thinkers." Dr. John Ahearne, then Vice President of Resources for the Future, said "Amory Lovins has done more to assemble and advance understanding of [energy] efficiency opportunities than any other single person." The Wall Street Journal's Centennial Issue named him among 39 people in the world "most likely to change the course of business in the 1990s"; Car called him the 22nd most powerful person in the global car industry; and The Economist wrote in 2008 that "history has proved him right." In 2008, he was named one of America's 24 Best Leaders by U.S. News & World Report and Harvard's Kennedy School, and received the first Aspen Institute / National Geographic Energy and Environment Award for Individual Thought Leadership. In 2009, Time named him one of the world's 100 most influential people, and Foreign Policy, one of the 100 top global thinkers. 34 Annex 2: Comments on Mr. William Dolan's proposed "criteria" for solutions Mr. Dolan's 20 December 2012 email to other selected citizens in the Aspen area, inviting them to submit ideas on how to achieve 100% renewable electricity for Aspen, is commendable as an explicit solicitation of the public input that has long been made to feel unwelcome. All praise to him and you for that gesture. But his invitation's last section, providing "some additional information and criteria to help guide you in identifying realistic and practicable projects to meet our goal," undoes the goodwill he'd just created. He first offers eight criteria, then lists six standard talking points for the "proposed Castle Creek Energy Center [as] a useful benchmark for comparison." This framing invites the unhelpful inference that the eight criteria were tailored to restrict valid and admissible choices to CCEC, which Aspen voters just rejected. Moreover, his eight criteria are often unfounded or technically incorrect. Taking them in turn: 1. The initial criterion of providing ~8 GWh/y of "reliable, clean renewable electricity" to reach the 100% goal assumes that this must be done by additional supply rather than by more-efficient use (or a mixture), even though the invitation's previous paragraph rightly identifies both supply- and demand-side options. This criterion seems to me arbitrary, wrong, and important. I'd be astounded if the ~11% "gap" between projected 2014 renewable electricity supply and the 100% goal couldn't be met entirely, with much room to spare, by moreproductive use of electricity at far lower cost than any kind of increased supply, cheaper and and probably faster. In other words, saving 11% of projected electricity consumption (or roughly one and a half average megawatts--one watt for every ~75 square feet of land in the City) would be much cheaper, easier, more benign, and probably quicker than increasing renewable supply by that amount using any supply technology whatsoever, including CCEC's to-go cost. I don't mean by this to imply any criticism of the people or programs currently advancing energy efficiency in the City; rather, as a citizen observing homes, shops, restaurants, hotels, public buildings, and other electricity-users throughout the City and Valley for more than three decades, I've concluded that very, very few customers lack important opportunities for cost-effectively deploying modern technologies that could collectively save much more total electricity than 11% (with bigger savings in some places, smaller in others) while delivering the same or better services. Ruling out enhanced efficiency at the start and focusing solely on supply makes this inquiry futile. That may not be what Mr. Dolan meant, but it's what he wrote. 2. The second criterion seems to call for levelized lifecycle costs of electricity comparable to those of current wholesale supply at 6.2?/kWh currently (but subject to planned escalation and unknown volatility). This criterion too is fallacious in three ways: (a) we should be looking at services like hot showers, cold beer, visibility, and shaftpower--not the kWh used to deliver them; (b) we need to compare those services delivered to the customer, not compare delivered retail energy or services vs. a wholesale price for power not yet delivered (because distribution, even if its capital costs so far are sunk, does have incremental and 35 lifecycle capital and operating costs and losses); and (c) the proper cost comparison is levelized lifecycle cost (not a snapshot of the current price) plus the hedging cost of any price volatility, plus any externalities (hidden social costs, which of course are key issues for both CCEC and coal power). 3. The third criterion, "Completion timeframe of 2015," appears to have no technical need or analytic framework; rather, it expresses a political goal and is therefore arbitrary. A cheaper or otherwise superior solution that takes longer to complete may provide greater net public good. The fraction of Aspen's total carbon emissions at stake here, vs. the 2004 Inventory, is just 0.6%; timing isn't critical. 4. "Directly offsets reliance on coal-fired generation" seems a reasonable general goal--I used it to optimize the design of my own net-solar-power-producing house, whose PV system is tuned to back out as much coal-fired generation as possible--but of course this could also be achieved by any other way to deliver the services now provided by coal-fired electricity. If this is a major goal, it's best achieved by whatever option(s) will displace the most coal per dollar and per year; any costlier option will buy less coal displacement and thus less climate protection than it could have. 5. "Preferably local" has unclear logic and analytic basis. Local supply has many merits described in definitive books I've written, such as the Pentagoncosponsored Brittle Power: Energy Strategy for National Security (1982) and the Economist book of the year Small Is Profitable: The Hidden Economic Benefits of Making Electrical Resources the Right Size (2002). But scale and locality are very complex and nuanced attributes with many other implications. They can't be properly used to compare alternatives without understanding the underlying goals. 6. "Limited transmission requirements" would again apply to any local demand- or supply-side resource, but it's awfully vague and its logic is unclear. Nobody is calling for any option with unlimited transmission requirements, and evidently there is spare transmission capacity, or the City couldn't have agreed to boost its renewable-electricity fraction by 14 percentage points with new Ridgeway hydropower nor suggested it could buy more windpower. All the transmission capacity originally built to serve Aspen's needs before Aspen became 75% renewably powered should still be there, implying a big surplus capacity. MEAN warns that Aspen may have to pay most if not all of the associated wheeling charges regardless of its utilization of that transmission capacity, since those charges are typically based on winter peak loads that would scarcely change with CCEC. (If so, and in contrast, efficiency and winter-peak local production could cut the wheeling charges.) Lacking information on the transmission situation, I can only infer from the City's and NMPP's spreadsheets that wheeling costs are too small (apparently ~0.4?/kWh) to be important in comparing local with remote resources; typically the uncertainties in the cost of the local resources are bigger. On the other side of this argument, transmission systems do fail occasionally, so 36 the modern approach to secure electricity supply--illustrated by my own house, adopted by the Pentagon, piloted in Denmark, highly successful in rural Cuba--is to rearchitect the grid into netted, islandable microgrids supported by local generation, so critical loads can be met even if transmission fails. That would be an excellent idea, merits public discussion, but doesn't yet seem to be on the City's agenda to explore or execute, and it should be. The key to making it work is very high end-use efficiency, which makes distributed renewable supplies very affordable. 7. "Potential for partnership or ownership interest" is vague and peculiar. The City always has make-or-buy decisions, but I don't understand why full or partial municipal ownership should automatically be considered preferable to ownership by private parties under suitable contractual arrangements. This is more a matter of ideology than of economics, and it needs more logical support than just supposing it feels good to own stuff. (This ideological dispute sometimes comes out in odd ways, such as claims that CCEC's critics are motivated by hatred of government. Some might be. Many more don't seem to be. It's a red herring.) 8. "Not based on [REC]...incentives or purchases" is sensible if you want to make sure your electricity's green origins are real and not vaporous claims or multiple counting and if you don't discriminate between low- and high-quality offsets. But tit's not sensible to exclude high-quality, independently certified green power on principle. Like many customers hereabouts, I have an all-requirements windpower contract with Holy Cross for the electricity my house uses at night (generally less than it exports annually from its daytime solar production). It doesn't bother me in the least that there is no direct and unique wire connection between my house and Excel's windfarms built to fulfill contracts like mine, nor that all the electrons I buy from Holy Cross are physically identical no matter how they were produced. (All kinds of generators mix their output indistinguishably in the grid, so the link from wind turbines to my meter is contractual rather than physical.) It shouldn't bother you either, so long as the accounting is honest and auditable and the renewable resources are counted once and only once. Indeed, the City's 2007 Electricity Plan explicitly calls for buying renewable carbon offsets if and as needed.61 That's a valid policy, but it's not consistent with this criterion. 9. In addition, Mr. Dolan poses as "key challenges" two opaque requirements of dubious relevance and validity: - 61 "finding power sources that fit Aspen's consumption curve (to limit the production and/or purchase of unusable excess energy" (e.g., poorly matching local loads' seasonality). That hasn't been the way American utilities run their systems for well over a century, and shouldn't be; it's not a "key challenge" but simply a silly and unnecessary notion. Although avoiding major spillage of renewable generation is always part of the economic calculus for a large power system, such as the Western Interconnect, it is practically never a sound criterion www.nrpa.aspenrecreation.net/2-CEP_Exec_Summary.pdf . 37 for a small municipal system. Utilities don't connect one source to one load--all sources supply the grid, which in turn serves the loads--so load-matching on all timescales is routinely handled by the full spectrum of demand response, efficient end-use, and dynamic operational or contractual adjustments in the diversified supply portfolio. I'm mystified why Mr. Dolan should think that over-/undergeneration is problematic. You never import more power than you need at that moment. The laws of physics and the practice of grid regulation guarantee that balance. If you locally produce more than you need, you reduce your imports accordingly, export more, or clear a market at higher demand. If you ever became a large net producer of local renewable power, you could run your transmission line backwards and sell your surplus. I'm also unclear how Mr. Dolan's mention of seasonality helps the CCEC case, since Aspen's winter-peaking loads poorly match hydropower's spring-peaking, winter-sparse output. - limiting new renewable projects to Aspen's wholesale supply contract with MEAN. As I understand it, that contract makes MEAN the exclusive supplier for 98% of the City's renewable electricity consumption, less any onsite renewable generation by individual customers (limited to 2%): that is, the contract requires Aspen Electric to buy 98% of its incremental supply exclusively from MEAN, other than the three existing hydro projects plus 1.78 MW of new hydro (evidently meant to include chiefly CCEC). But if that's what the contract says, it merits independent review with a view to modest renegotiation, which would be as much an option as any other sort of supply- or demand-side resource. Slightly loosening the straitjacket should expand the portfolio of local options that could nicely meet Mr. Dolan's other criteria. There's no good reason to require that 1.78 MW of new Aspen Electric procurements from sources other than MEAN must be hydropower. That odd requirement, apparently adopted on Jim Markalunas's watch (but he doesn't recall why), limits and prejudges the outcome of any transparent public comparison of options. If that's what really the contract's Exhibit B says, Council should consider working with MEAN to amend it so that equivalent nonhydro resources, local or otherwise, can compete fairly with local hydropower as well as with MEAN's offerings. MEAN shouldn't care whether Aspen's next 1.78 of incremental supply comes from hydropower or from any other resource; kWh are kWh to them.62 However, such renegotiation wouldn't be needed to procure the hydro addition to Homestake plus in-system hydro opportunities (such as pressure-reducing valves) explored by the Kennedy/Jenks Consultants report of 6 December 2011. The existing contract presumably also lets proposed Ruedi repairs restore original design output without counting against the 1.78-MW exception. These hydro options alone could probably make Aspen Electric 100% renewable without further efficiency gains or other renewables. 62 A MEAN official told me he doesn't consider windpower a reliable source over more than a decade, although his wind Power Purchase Agreements are for 20 y and the industry norm is 25 y, often backed by insurance; cf. Annex 3, #9. But he also confirmed that MEAN's goal is to help its members meet their own goals without disadvantaging other members. This should be achievable for both MEAN and Aspen. 38 Annex 3: Fifteen examples of material errors and omissions in the City's online explanations of CCEC and other electricity options 1. www.aspenpitkin.com/Living-in-the-Valley/Green-Initiatives/RenewableEnergy/Hydroelectric/Finances/ gives as its sole evidence that "a new hydroelectric power source is one of the most affordable new sources of power" an undated table purportedly from USEIA. Of course actual costs are highly site-dependent--most of all for hydropower, whose levelized kWh costs commonly vary manyfold with size and site even if the water flow is constant. EIA's latest standard levelized-cost table 63 shows that a new-hydropower cost range of 5.8-14.8?/kWh lies behind its 8.9? average (2011 $), and specifically assumes seasonal water storage to even the flow. Using a meaningless point value for the cost of "a new hydroelectric power source" is especially misleading because CCEC costs several to many times as much per kW or per kWh as typical new small hydro installations in the US. The sweeping implication that generic new hydro "is one of the most affordable new sources of power" is specifically inapplicable to the Aspenspecific, no-seasonal-storage context in which it is presented. 2. Many of the same table's other figures are also seriously outdated. For example, windpower is shown at 9.7?/kWh, but as noted in the text of this letter, new Windbelt power-purchase contracts written in 2011-12 actually averaged 3.2?/kWh and ranged from about 2.5 to 4.0?/kWh.64 Solar photovoltaics are shown at 21?, but utility-scale groundmount PV power cleared the California auction in April 2012 at an average of 8.92? (levelized 2012 $); moreover, installed PV system costs in the first half of 2012 averaged twice as high in the U.S. as in Germany, which buys the same equipment in the global market but has streamlined its installation by scaling additions to ~8 GW/y, so U.S. prices too are continuing to drop rapidly, passing nowadays through ~6-7?. Delivery channels are also evolving rapidly: cloudy Connecticut, for example, just doubled installed PV power in four pilot towns, adding 2.3 MW, in just 20 weeks (more in than in the previous 7 years) via a Clean Energy Finance and Investment Authority. 65 3. Most importantly, the table does not include more-efficient use of electricity, which in 2009 U.S. utility programs had an average levelized cost of about 2.6?/kWh (ranging from about 1.7 to 3.4?)66 and, like modern renewables, continues to get cheaper. 4. The same URL graphs ongoing annual savings from Ruedi hydropower starting a decade ago. There are two issues here: (a) That snapshot says nothing about the levelized cost or the cost-effectiveness of that 1985 project's power over its lifetime--an important point that would help CCEC's case. Ruedi hydro may well be a great deal, and if it runs well for a long time, it should be (at its order-of-magnitude lower-than-CCEC original capital cost), but this graph doesn't go far enough to the right to show that. Interestingly, though, it does show an asymptotic power cost (set at breakeven) of 2?/kWh for this 563 www.eia.gov/forecasts/aeo/electricity_generation.cfm, 12 July 2012, probably a later version of the data the City used for its summary table, which seems to show levelized busbar costs in some unstated year's dollars for various marginal busbar electricity supplies, but not for efficient use. 64 In levelized 2011 $: LBL-5559e, p. 52, Fig. 34, http://eetd.lbl.gov/ea/emp/reports/lbnl-5559e.pdf. 65 See www.ctcleanenergy.com and www.smartpower.org. 66 Friedrich et al., ECEEE, Sept 2009, http://www.aceee.org/research-report/u092. 39 MW project, presumably covering a mixture of long-term debt service and current O&M cost. The graph shows much higher costs in the earlier years during debt service. It would be instructive to know the lifetime levelized cost, which I presume would support the implied claim that the project was advantageous to City customers, on the same lines as the qualitative argument about 28-year bonds vs 75-year life. (b) The green-line Ruedi costs shown appear to be only old projections not corrected to actual values. Reporting projected rather than actual costs is misleading and casts doubt on the other content. 5. The stated CCEC budget seems to be just for overnight capital cost, but doesn't say so. I gather that it includes some normal (15%?) construction contingencies, but I haven't seen any mention of real escalation during construction, which has more than doubled the budgeted capital cost since 2006.67 Most importantly, the budget doesn't include the City's actual or shadow cost of capital, which totals quite a few million dollars. The bond prospectus shows undiscounted totals of $5.5 million principal plus $4.02 million interest, adding 73% to the project's undiscounted cost. The City's website doesn't mention this. Nor does it ascribe any cost to most of the $4+ million of other City capital earmarked for the project. (The nearly four million dollars transferred from other City funds either incurs an actual borrowing cost or could have been used instead to save interest on existing debt. Even CORE's $0.4-million grant has its own opportunity cost that can hardly be less than the City's cost of capital.) Not disclosing these costs of capital as part of the project's budget gives a false impression of the project's actual total cost to electricity customers (and City property-taxpayers if electric revenues, for any reasons, proved insufficient to repay the debt). Presenting only the unfinanced cost is like saying you can afford a million-dollar house if it were financed with free money. The City's financial case for CCEC counts the cost of two-and-a-half of its three sources of non-CORE-grant capital, as discussed in the text of this letter (even if some of the interest rates seem understated), but the City's lay-public budget representations include none. 6. www.aspenpitkin.com/Living-in-the-Valley/Green-Initiatives/RenewableEnergy/Hydroelectric/Why-Hydro-/ interprets "base load" as "the minimum power necessary to meet customer demands at any given time," which is syntactically ambiguous but appears to mean the level of demand below the shoulder of the annual load-duration curve, i.e. load that is or appears to be steady in aggregate (whether or not each of its elements is actually constant). This is one of five meanings of "baseload," and is sometimes used by utility load analysts. However, for a utility resource purchaser, "baseload" means the resource with the lowest levelized lifecycle cost, while for a utility operator, "baseload" means the resource with the lowest dispatch (short-run marginal) cost. Note that both these dominant definitions are economic. New hydroelectricity can seldom meet the first and may or may not meet the second--certainly not if competed against demand-side resources. 7. www.aspenpitkin.com/Living-in-the-Valley/Green-Initiatives/Renewable-Energy/ 67 "Castle Creek Hydroelectric Plant Ballot Questions," memo from Phil Overeynder, Public Works Manager, to Mayor and Council Members, 17 August 2007 for 27 August 2007 Council Meeting, gives a 2006 estimate of $5.1 million and says the recommended (and later approved) $5.5-million bond issue "includes an allowance for cost escalation from the time of the 2006 budget estimate." Evidently that allowance was grossly inadequate. 40 repeats the load-analyst-perspective definition of "base load" just noted, then shifts to a different definition: "It is energy that must be there when a customer turns on a computer, lights, or refrigerator whether or not the sun is shining or the wind is blowing." Actually, that's about reliability, not loadshape. All noncurtailable electricity demand must be met by instantaneously balancing supply and demand at all times, including onpeak, in order to meet the utility's reliability-of-service criteria. Evidently the writer is confused both about what "baseload" means and about how power supply works--my next point. 8. www.aspenpitkin.com/Living-in-theValley/Green-Initiatives/Renewable-Energy/ misrepresents how electricity systems work. While it is correct that CCEC could help protect Aspen from transmission failures--an important argument, though equally applicable to other local efficiency or renewable resources--it also states that "local hydropower from the CCEC is always on." But I thought CCEC is supposed to shut down for about four months a year to protect minimum streamflows (in midwinter-- inconveniently, since that's when Aspen's electricity usage peaks), and the page acknowledges that output may also be lost in severe drought. Either way, neither CCEC nor any other power source will be "always on." The same page further states that "base load energy resources must be constant and reliable--wind and solar are neither." This is highly misleading. Ever since Thomas Edison's time, the electricity system has routinely backed up failed units with working units. That's the main reason the grid exists. But the same grid capability to move power around makes it perfectly feasible and common to provide "virtual baseload" in the City writer's sense by combining a portfolio of variable resources such as solar and wind. (Siemens, for example, aggregates different and differently sited variable renewables in Germany and sells their collective output as steady firm power.) For example, the balancing resources required for a windpower portfolio providing up to half, or even more, of total electricity in diverse U.S. electricity systems are typically less than the reserve margin required to generate the same firm power from large thermal power stations, because the wind portfolio has smaller units and is more diversified. Moreover, its outages are more predictable than those of the large thermal plants, whose intermittence requires those even larger, and quite costly, reserve margins and spinning reserve. It appears that whoever wrote this page is simply unfamiliar with modern grid-integration practice and utility strategy. The misconceptions conveyed are common but regrettable, and the City should not be propagating them. 9. www.aspenpitkin.com/Living-in-the-Valley/Green-Initiatives/RenewableEnergy/Hydroelectric/Why-Hydro-/ claims that "Wind turbines only last between 15 and 20 years on average." That hasn't been true for over a decade. Early turbines were typically rated at 20 years, but the industry norm is now at least 20-25 years, often 30 68 , which is why most Power Purchase Agreements for new windfarms have a 25-year term. Proper maintenance now makes 30+-year operational life realistic. 69 Modern wind-system lifetime ratings and warranties are based on accelerated lifetime testing, sophisticated sensors and monitoring, and preventive maintenance. The life-limiting factor is typically the major moving parts, notably the generator and/or gearbox if any (the latest turbines have none), but even complete replacement of the blades, gearbox, and generator 68 69 E.g., www.elpower.com/wind. See e.g. www.windpowermonthly.com/news/1148629/Breathing-new-life-old-wind-turbines/. 41 typically costs only ~15-20% of the original turbine cost. 70 With such highly costeffective refurbishment, "Some developers take the view that the 'useful life' of a wind turbine may be indefinite if properly operated, maintained and refurbished."71 10-12. The same City webpage says that "Without sufficient levels of locally produced renewable energy, Aspen cannot provide reliable, clean and sustainable base load to the system into the future, and must instead rely upon supplemental purchases of costly, and non-local sources of energy which can be dirty or unreliable." Wrong on three counts besides the odd and irrelevant use of "base load": (a) reliable, clean and sustainable power (steady or variable) can be freely chosen to use any mix of local and wheeled sources; (b) the latter can now be considerably cheaper, e.g. from new wind contracts as cited above; (c) many non-local generating options are neither dirty nor unreliable, but could yield the same carbon savings as local new hydro at comparable or lower marginal cost. (Of course energy efficiency, not mentioned, beats any marginal supply cost.) 13. The same City page describes "uncertainty surrounding long-term maintenance, upkeep, and cost stability associated with wind energy," but these uncertainties are smaller than for traditional generators or for a single local hydropower plant dependent on snowpack that's now at serious risk from climate change. 14. The same City page complains about "Disappearing Federal Subsidies" that "have provided as much as 20% of wind energy project costs over the past 10 years"; true, but the wind Production Tax Credit for all projects under construction by the end of 2013 would apply to any new wind Power Purchase Agreement signed in the next year or two. It partly offsets nonrenewable electricity's permanent and generally larger subsidies. 15. The same City page twice states that 7% is the national-average transmission loss. The correct figure shown by EIA and EEI data is 3%. (A further 4% is lost in the local distribution system, and applies to any resource except those at customers' premises like efficiency and rooftop solar.) The reader could easily get the impression that 7% of Aspen's wheeled power is lost en route. The City's NMMP reports that the actual wheeling loss from Nebraska to Aspen averages 2.5%, so why not simply give the right number rather than repeatedly giving one that is bigger, wrong, and irrelevant? These observations, which are far from exhaustive, suggest a need for improved technical understanding and quality control among those responsible for creating such information, whether for the Mayor and Council or for Aspen's and other citizens. 70 71 www.windmeasurementinternational.com/wind-turbines/om-turbines.php. http://renewableenergylawyer.blogspot.com/2011/12/what-is-useful-life-of-wind-turbine.html . 42 Annex 4: My original analysis of Ruedi hydro's operating and maintenance costs, 2003-07 (including net capital additions), followed by 2003-07 data plus anomalous 2008-09 results 43 Ruedi hydro ABL, 12 Jan 2013, including anomalous 2008-09 data current-$ GDP Implicit year cap. factor (1) out (2) cost (3) cost (4) Price Deflator (5) cost 2003 0.278 12.18 232,987 1.91 1.2042 2.30 2004 0.252 11.07 316,071 2.86 1.1713 3.34 2005 0.362 15.86 290,754 1.83 1.1336 2.08 2006 0.405 17.74 243,701 1.37 1.0981 1.51 2007 0.391 17.13 409,304 2.39 1.0672 2.55 2008 0.477 20.95 1,485,492 7.09 1.044 7.40 2009 0.408 17.87 2,418,769 13.54 1.0331 13.98 mean 0.368 16.11 771,011 4.43 1.11 4.90 st dev 0.078 3.44 852,600 4.45 0.07 4.52 mean renormalized to CCEC's projected net/gross output ratio of 0.805 (6) 6.1 5.6 memo (7): 2010 1,790,194 1.0213 2011 3,285,575 1 notes 1 NMPP Services, "Castle Creek Hydroelectric Project Economic Analysis," Dec 2010, Table 5 2 3 \l Inferred from 5--MW nominal rated capacity, id., uncorrected for gross vs. net output Stifel Nicolaus prospectus for City of Aspen $5.5-million bond, 2008, Table X, 2003-07; 2008: p. F6 is capex); 2009: p. F6 is capex) Using GDP Implicit Price Deflator and 8,760 h/y (8,784 in 2004 2008) USEIA, Annual Energy Review 2011, Table D1 Kennedy-Jenks Consultants, Draft Concept-Level Feasibility Analysis and Economic Evalua- tion of the City of Aspen's Castle Creek Energy Center Hydroelectric Plant and Potential Options," 6 Dec 2011, 2-1 (6.2 net from 7.2 gross GWh/y) 2010 was $789,851 op plus capex ER878147.pdf, p. 2011 was $796,173 op plus capex p. so total nominal--dollar 2008-2011 capex was three--fourths of original constructior in 1984-85, equivalent to in 2011 44 Annex 5: The City's revised (11 February 2013) Ruedi and Maroon Creek hydroelectric operating and maintenance costs, and my CCEC O&M cost range 45 Annex 6: An instructive cost calculation Utility financial analysts commonly use a "fixed charge rate" to convert a capital investment72 into an annual capital charge that must be recovered to cover depreciation plus cost of capital. In this case, that sum looks simplistically like 2.5%/y for 40-year depreciation plus 4.58%/y for the cost of bond capital 73, a total of 7.1%/y. The actual, more complex, standard formula for a tax-exempt municipal utility yields a levelized real fixed charge rate of 7.264%/y for CCEC's financing parameters. 74 Then a $10.5-million project cost would incur an annual real capital charge of $0.76 million. If the project delivered the intended 5.476 GWh/y, with no allowance for further curtailment if stream monitoring or climate changed, then the real (constant-dollar) capital charge would be 13.93?/kWh, of which 34%75 or 4.77?/kWh represents costs (and proportionate financing costs) not yet incurred. I have not adjusted for niceties of timing, such as the deferred output with the project's suspension since the election. Adding the 1.14-2.87?/kWh range of O&M costs derived above, the full CCEC project would thus generate power at ~15-17?/kWh, or just its to-go portion at ~6-8?/kWh, all in constant 2011 $. The former far exceeds, and the latter is comparable to, the City's current calculation (most recently provided to me 11 April 2013) of 6.2?/kWh total cost in unspecified nominal dollars. 72 It's the whole $10.5-million budget, so one needn't worry about whether its whole cost of capital was counted--the fixed charge rate does that automagically. 73 Why not ~3.9%/y for the weighted-average cost of capital? It's economically proper to use the higher marginal rate from the bond sale, because it reflects a minimum estimate of the opportunity cost of using cheaper existing capital for this purpose rather than, for example, to prepay the bonds. However, as current and expected near-term monetary inflation is low, I have not explicitly adjusted for it. All this could be properly and explicitly handled if I, or the City, knew what monetary inflation rate its analysis assumed. 74 I used the formulae on pp. 41-43 of the 2009 Ten Year Site Plan of Jacksonville Electric, (www.psc.state.fl.us/publications/pdf/electricgas/2009JEATYSP.pdf), a far larger municipal utility (#8 in the U.S.) but with financial parameters broadly similar to Aspen's. I then assumed 28-y bond financing at 4.58%/y, the 2.036% issuance costs shown in Aspen's bond prospectus, a 1-y Debt Service Reserve Fund, and a 0.5%/y property insurance cost (from Jacksonville's plan--I didn't see such a cost in Aspen's accounts) to derive a Capital Recovery Factor of 0.0649/y, an Adjusted CRF of 0.06764/y, and a Levelized FCR of 0.07264/y. 75 I.e., the ratio of $3.6M to-go costs to $10.5 total costs shown at www.aspenpitkin.com/Living-in-theValley/Green-Initiatives/Renewable-Energy/Hydroelectric/Finances/ on 12 January 2013. 46