NATURE CLIMATE CHANGE | VOL 3 | SEPTEMBER 2013 | www.nature.com/natureclimatechange 771

opinion & comment

here, these range from how climate change research is conducted to the process by which climate change assessments are produced.

Whenever feasible, qualitative uncertainty assessments should be replaced with quantitative ones. At a minimum, a standard error should be attached to any estimate, along with a description of how it was calculated. Ideally, an entire probability distribution should be provided — innovative graphical techniques can be used to communicate these uncertainties in a more effective manner. In addition, modern statistical methods for spatio-temporal data, such as hierarchical models, should be used to reduce uncertainties in trend estimates for climate observations and projections of climate change. By taking into account spatial and temporal dependence, these techniques provide a powerful tool for detection of observed and projected changes in climate. To increase the accuracy with which the climate is monitored, various sources of information need to be combined using hierarchical statistical models. Such techniques can take into account differences in the uncertainties of, for instance, in situ and remotely sensed measurements.

Statistical principles of experimental design should be applied to climate change experiments using numerical models of the climate system. Through making more efficient use of computational resources, uncertainties in climate change projections can be reduced. Methods based on the statistical theory of extreme values should be

used to quantify changes in the likelihood of extreme weather events, whether based on climate observations or on projections from climate models. In this way, information more useful to decision-makers about the risk of extreme events (for example, in terms of changing return levels) can be provided. Finally, to improve the quality of the treatment of uncertainty, at least one author with expertise in uncertainty analysis should be included on all chapters of IPCC and US national assessments. These authors could come from the field of statistics, as well as from other related fields including decision analysis and risk analysis.

If these recommendations are adopted, the improvements in uncertainty quantification would thereby help policymakers to better understand the risks of climate change and adopt policies that prepare the world for the future. ❐

Richard W. Katz1*, Peter F. Craigmile2,3, Peter Guttorp4,5, Murali Haran6, Bruno Sansó7 and Michael L. Stein8 are at 1National Center for Atmospheric Research, Boulder, Colorado 80307, USA, 2University of Glasgow, Glasgow G12 8QW, UK, 3Ohio State University, Columbus, Ohio 43210, USA, 4University of Washington, Seattle, Washington 98195, USA, 5Norwegian Computing Center, NO-0314 Oslo, Norway, 6Pennsylvania State University, University Park, Pennsylvania 16802, USA, 7University of California at Santa Cruz, Santa Cruz, California 95064, USA, 8University of Chicago, Chicago, Illinois 60637, USA. *e-mail: rwk@ucar.edu

References1. Knutti, R. & Hegerl, G. C. Nature Geosci.

1, 735–743 (2008).2. Guttorp, P. Stat. Polit. Policy 3 http://dx.doi.org/10.1515/2151-

7509.1055 (2012).3. Mastrandrea, M. D. et al. Guidance Note for Lead Authors of

the IPCC Fifth Assessment Report on Consistent Treatment of Uncertainties (IPCC, 2010); available via http://go.nature.com/PvUJbk

4. Moss, R. H. & Yohe, G. Assessing and Communicating Confidence Levels and Uncertainties in the Main Conclusions of the NCA 2013 Report: Guidance for Authors and Contributors (National Climate Assessment Development and Advisory Committee, 2011); available via http://go.nature.com/AdQeGO

5. Katz, R. W. Stat. Sci. 17, 97–122 (2002).6. Cressie, N. & Wikle, C. K. Statistics for Spatio-Temporal Data

(Wiley, 2011).7. Gelfand, A. E., Zhu, L. & Carlin, B. P. Biostatistics

2, 31–45 (2001).8. Berrocal, V. J., Craigmile, P. F. & Guttorp, P. Environmetrics

23, 482–492 (2012).9. Mearns, L. O. et al. Bull. Am. Meteorol. Soc.

93, 1337–1362 (2012).10. IPCC Managing the Risks of Extreme Events and Disasters to

Advance Climate Change Adaptation (eds Field, C. B. et al.) (Cambridge Univ. Press, 2012); available at http://www.ipcc-wg2.gov/SREX

11. Coles, S. An Introduction to Statistical Modeling of Extreme Values (Springer, 2001).

12. Milly, P. C. D. et al. Science 319, 573–574 (2008).13. Donat, M. G. & Alexander, L. V. Geophys. Res. Lett.

39, L14707 (2012).14. Hansen, J., Sato, M. & Ruedy, R. Proc. Natl Acad. Sci. USA

109, E2415–E2423 (2012).15. Brown, S. J., Caesar, J. & Ferro, C. A. T. J. Geophys.Res.

113, D05115 (2008).16. Karl, T. R. & Katz, R. W. Proc. Natl Acad. Sci. USA

109, 14720–14721 (2012).17. Spiegelhalter, D., Pearson, M. & Short, I. Science

333, 1393–1400 (2011).18. http://www.climathnet.org19. http://www2.image.ucar.edu/gsp20. http://imsc.pacificclimate.org21. http://www.nr.no/sarma22. http://www.statmos.washington.edu23. http://go.nature.com/tzfO2Z24. http://cmip-pcmdi.llnl.gov

COMMENTARY:

Powering Los Angeles with renewable energyMayor Antonio R. Villaraigosa, Varun Sivaram and Ron Nichols

The City of Los Angeles is nearly two thirds of the way towards its goal of generating a third of its electricity from renewable sources by 2020; cities around the world can glean valuable technical, economic and political lessons from its experience.

On 22 March, 2013, the City of Los Angeles entered into contracts to end its consumption of coal

power; former US Vice President Al Gore asserted that LA had taken its place among the “five greatest cities in the world where

combating global warming is concerned”. This assessment will be tested over the next decade, as the city continues to deploy renewable energy and energy efficiency as the centrepieces of its coal-replacement strategy. By 2030, cities will account for 76%

of global greenhouse gas emissions, under consumption-based accounting models1,2. If megacities such as LA can demonstrate affordable clean energy, they will provide a template for urban sustainability worldwide. This should be especially important for

© 2013 Macmillan Publishers Limited. All rights reserved

772 NATURE CLIMATE CHANGE | VOL 3 | SEPTEMBER 2013 | www.nature.com/natureclimatechange

opinion & comment

rapidly urbanizing countries with high emissions growth, such as China and India3.

Over the past eight years LA has quadrupled its renewable energy generation, which now comprises 20% of the city’s electricity consumption. By 2020, it will achieve 33% renewable energy and reduce its energy use by at least 10% (ref. 4). Critical examination of the city’s continuing experience with renewables is worthwhile for two reasons. First, LA demonstrates how a city can take control of its energy portfolio and harness state and federal policies to best meet local needs. Second, the challenges and lessons of sustainably powering a megacity are neatly packaged for display because of the unique city-owned, vertically integrated

utility — the LA Department of Water and Power (LADWP).

Los Angeles’s retention of a municipally controlled utility, along with powerful public-sector labour unions, may seem anachronistic and inefficient amidst a national electricity deregulation movement. However, on balance, the city has achieved surprising progress at reasonable cost, enabled by the responsiveness of the LADWP to elected officials and the utility customers they represent. In today’s segmented energy sector, where electricity from independent power producers is often traded on wholesale markets and distributed across expansive service territories, cities can have limited control and visibility over

where their power comes from5. However, procurement, generation, transmission, grid integration and distribution of LA’s renewable energy are all under one organization, elucidating the process of building a city’s renewable energy portfolio largely from scratch.

Towards a robust renewable portfolioCalifornia’s aggressive Renewable Portfolio Standard (RPS) mandated that its three investor-owned utilities (IOUs) — which collectively serve 68% of the state — obtain 20% of their electricity from eligible renewable sources by 20106. Whereas none of the IOUs and no other major US city achieved the 20% mark in 2010, LA succeeded, voluntarily adopting the standard years before it became a legal mandate regardless of utility ownership in 2011. Although RPS policies encourage renewable energy development7,8, the RPS alone is clearly insufficient to explain LA’s success. The crucial factor was the commitment of municipal elected officials to oversee the LADWP’s progress towards the state target.

In 2006, that goal seemed implausible. The LADWP needed to quadruple its renewable generation from 5% of total consumption, whereas the California average was 11%; moreover, LA obtained nearly half of its electricity from coal power, similar to the national profile but well above California’s average of 15%. To understand the dramatic acceleration of LA’s renewable energy programme, some background on the LADWP, the largest municipal utility in the US, is warranted9. Los Angeles owes much of its growth to this enterprising utility, founded in 1902, and the city has a strong history of controlling its generation and transmission assets. Although its customers only account for 10% of California’s energy demand, the LADWP owns 27% of California’s transmission lines, affording it access to several western states to tap into a wide diversity of power sources. Although IOUs suffered rolling blackouts and price volatility during California’s energy crisis in 2001, the LADWP emerged unscathed and actually helped to stabilize the state’s electricity market. The City Charter affords the LADWP an exclusive monopoly to sell power in its service territory, which is largely coterminous with city boundaries.

Insulated from the obligation to meet its power supply from California’s at times volatile power market, the LADWP has leveraged its vast transmission systems and assembled a diverse renewable portfolio through competitive bidding and a rapid, real-world learning process. Moving beyond its voluntary 20% renewable contribution,

Existingwind

Existing solar

Existinggeothermal

New windNewbiogas

New solar

Newgeothermal

9,000

8,000

7,000

6,000

5,000

4,000

3,000

2,000

1,000

0

Elec

tric

ity g

ener

atio

n (G

Wh)

Existingsmall hydro

Existingbiogas

Wholesalepurchase

0

5

10

15

20

25

30

1990 2004 2010

Year

Year

LAD

WP

gene

ratio

n (m

illio

n M

Wh)

or

CO

2 em

issi

ons

(mill

ion

met

ric to

nnes

)

Renewable

20032005

20072009

20112013

20152017

20192021

20232025

20272029

2031

Large hydro

Other

Nuclear

Natural gas

Coal

CO2 emissions

20%RPS Goal: 25% 33%a

b

Figure 1 | Energy sources and emissions. a, LADWP renewable energy generation, historical and projected. ‘Existing’ sources refer to those in operation by 2012, whereas ‘New’ sources will be constructed subsequently. RPS milestones are delineated in red and the major portfolio components to meet the 33% RPS target are listed in bold4. b, LADWP total historical electricity generation broken down by source, with CO2 emissions overlaid.

© 2013 Macmillan Publishers Limited. All rights reserved

NATURE CLIMATE CHANGE | VOL 3 | SEPTEMBER 2013 | www.nature.com/natureclimatechange 773

opinion & comment

the LADWP is bound by and plans to meet subsequent state RPS milestones enacted through recent California mandates — 25% in 2016, and 33% by 2020. For each milestone the LADWP must become more sophisticated, to account for the effect of high renewable penetration on the grid and control costs while diversifying energy sources.

Figure 1a shows the historic and future composition of the LADWP’s renewable energy portfolio. The salient characteristic is diversity — in technology, geography and procurement. By 2020, renewable generation will be divided between geothermal, solar, wind and qualifying small hydro power to smooth the overall generation profile. Los Angeles will acquire its renewable energy from seven different western states (see Supplementary Information), helping to insulate the portfolio as a whole from intermittent solar and wind generation caused by local weather shocks. Finally, the LADWP will procure its renewable energy through competetively bid power purchase agreements (PPAs), proprietary installations and wholesale purchases, to determine the optimal mode of procurement.

The first phase of renewable development, meeting the 2010 milestone of 20%, exhibited relatively little technological diversity. Unfamiliar with large-scale renewable generation, the LADWP had long relied on existing small hydro resources along the LA aqueduct, but that resource was insufficient to meet the target. To meet the 20% level, the LADWP invested heavily in wind — and to a lesser extent biogas — displacing some of the natural-gas supply required for gas-fired generation plants. By 2010, wind energy dominated the renewable portfolio; at the time, wind was far more economical than solar power and the industry was more mature.

Figure 1b illustrates the effect of increasing renewable penetration up to 2010 on LADWP CO2 emissions. Using the popular Kyoto Protocol benchmark of 1990 emissions levels, LA succeeded in reducing emissions by 28% in the electricity sector by 2010.

The present and second phase of the renewable development (to meet the 2016 milestone of 25%) is a period of transition for the LADWP. Following the precipitous decline in the prices of solar panels and the upcoming expiration of federal tax incentives in 2016, the LADWP intends to increase penetration largely through utility-scale photovoltaic installations in the desert and a new feed-in tariff for in-city solar generation. To access the high-insolation Mojave Desert, the LADWP is constructing a new transmission line explicitly for renewables and conducting

a comprehensive competitive bidding process to ultimately provide over 1 GW of solar capacity.

The LADWP has become increasingly advanced in its process for soliciting and evaluating renewable energy sources as it enters this second phase (to ramp from 20% to 25% RPS). As renewable penetration increases, the impacts of intermittent generation on the grid could affect reliable operation and demand expensive contingent power reserves. Solar power provides a more predictable generation profile that coincides far better with peak load than do wind energy resources, which are strongest during off-peak hours. Moreover, the LADWP observed the solar power procurement processes of IOUs elsewhere in the state and now enjoys a buyer’s market, pitting hundreds of developers against each other to deliver less expensive and more bankable projects; this is in contrast to the LADWP’s earlier experience with wind contracting, where it exerted less leverage over price.

Finally, the third phase of development, to meet the 33% milestone by 2020, will require the LADWP to supplement its intermittent wind and solar power with baseload geothermal energy. Geothermal will be the only significant part of the

portfolio that is comparable to the round-the-clock operational characteristics of the coal-fired power plants that the LADWP is replacing. Geothermal will also enable the utility to adjust to the high intermittent penetration of solar and wind, while continuing to increase its share of renewable energy.

Challenges and lessonsAlthough the previous section may imply a smooth learning curve of increasing sophistication towards a growing renewable penetration, LA has had to contend with setbacks and constraints both technical and political, yielding lessons of both varieties. Securing constituent support and political agreement for rate hikes further complicates the already complex challenge of radically altering the electricity supply; LA’s success is dependent on continued cooperation between the LADWP and City Hall.

Over the next 10–12 years, the LADWP will replace 70% of its generation, an infrastructure built over a century. Renewable generation — along with energy efficiency, coal divestment and requirements to repower 2,800 MW of gas-fired generation with new power plants that do not use ocean water for cooling — is driving this transformation. However, the

10

12

14

16

18

Year

20

22

24

26

28

Reta

il el

ectr

icity

rate

(cen

ts p

er k

Wh)

Energy e�ciency

Coal replacement

Other operations

33% RPS

Power reliability programme

OTC repowering

Fuel costs

20122014

20162018

20202022

20242026

20282030

2032

Figure 2 | Drivers of electricity rate escalation. Impacts on the incremental rate of LADWP costs and initiatives over the next 20 years. Once-through-cooling (OTC) repowering refers to the retrofitting of coastal power plants to eliminate inefficient consumption of ocean water and improve generator performance. The power reliability programme carries out basic infrastructural upgrades, such as pole and transformer replacement4.

© 2013 Macmillan Publishers Limited. All rights reserved

774 NATURE CLIMATE CHANGE | VOL 3 | SEPTEMBER 2013 | www.nature.com/natureclimatechange

opinion & comment

costs are non-trivial, as displayed in Fig. 2. To achieve the initial 20% goal, LA utility customers saw an estimated rate increase of 0.9 cents per kWh, but as the complexity of integrating a high proportion of renewables increases, customers are projected to pay around 2 cents per kWh more, due to RPS fulfilment, over the next two decades.

We have engaged in extensive outreach to educate Angelenos about the merits of renewable energy and the necessity for higher rates. The political process behind these increases is arduous, however, and we have faced setbacks when proposals to fund future renewable projects failed to gain support at the ballot box and in the City Council. We learned that demonstrating a clear vision and implementation plan, including projections for mandate fulfilment such as those shown in Fig. 1a, is crucial for securing political support; this approach led to Council approval for a rate increase in 2012. However as rates escalate further, the subsequent Mayor and City Council will face the need to maintain a clear vision, to prudently compromise and recognize customer and voter concerns to secure RPS fulfilment.

Another challenge is state policy, which may not be as easily focused towards a locally optimal solution. California’s RPS requirements created an ambitious level of renewable penetration with the flexibility for LA to construct an appropriate portfolio. However, the state has also proposed a target (though not yet a mandate) of 12,000 MW of distributed generation by 2020. Proportionally, that would allocate 1,200 MW to LA. In the subsequent section, we discuss LA’s support for distributed generation (photovoltaic cells in particular), which leverages existing infrastructure and serves the local economy. Altogether, the city anticipates that between the LADWP-sponsored programmes and voluntary efforts by utility customers, around 500 MW of distributed solar will be installed in LA.

Higher levels are being evaluated, subject to a better understanding of the effect on local system reliability. Recognizing local needs and constraints is essential, and the state target — at over a third of average peak load — could overburden the LADWP’s infrastructure and delay progress toward the RPS.

On the technical side, the LADWP is learning how to integrate intermittent renewable resources. At present, solar and wind come laden with hidden costs, fundamentally because of variable production and inability to predict generation output from these resources confidently — resulting in a lack of dependable capacity10. This refers to the percentage of power generation of a renewable generator, under optimal conditions, that the utility can reliably count on during peak load times. Table 1 illustrates this point — for example, although typical wind turbines under contract to LA have a 24% to perhaps 37% capacity factor, their dependable capacity is only on the order of 10% after averaging geographically diverse wind resources. The difference implies that the LADWP needs to acquire expensive single-cycle gas turbines to provide contingent reserves and rapidly increase and decrease output to compensate for fluctuating wind-energy production. This is an expensive proposition, so the LADWP is now working to improve capabilities in forecasting wind-energy production, by placing meteorological equipment along known upwind corridors, for example. This will help to improve the dependable capacity value of its renewable generators.

Finally, the city is wary of possible ‘curtailment’, under which the LADWP would be forced to sell excess peak power, or even pay to offload that power. This occurs when customer loads are low (especially late at night), and the combined generation of wind turbines and minimum operating levels of gas-fired generation

exceeds the aggregate customer demand. Elsewhere in the world (for example, in Australia11) curtailment has increased the cost of renewable energy, because at high levels of renewable penetration, weather and demand can conspire towards excess supply. As LA advances toward higher renewable targets, its intermittent resources will produce unexpected consequences, rain or shine.

Focus on distributed solar powerHere we have argued that cities can implement state and federal guidelines to suit local needs. With utility-scale solar, it is easy to leverage federal tax credits to meet California’s RPS target, but that singular approach misses valuable opportunities for improving the local economy12. In February 2013 LA introduced the largest urban feed-in tariff (FIT) in the nation, which will deliver 150 MW of distributed solar power capacity in the city.

Distributed generation can benefit the grid — through lower line losses, more robust protection against cascading failure and reactive power support — as well as the local economy, by creating jobs in the city13,14. Los Angeles has supported distributed solar for 15 years, but until recently did so through upfront installation rebates for ‘net-metered solar’, which refers to grid-connected photovoltaic systems that utility customers install to offset their purchases of energy from the LADWP. By switching to a FIT and paying a fixed price per kWh of energy delivered to the LADWP under a PPA, the city can spread out its incentive payments over 20 years, enabling a fourfold increase in the pace of local installations15. Moreover, although net-metered solar has effectively increased market exposure and demand for rooftop photovoltaic panels, its implementation creates some serious inequities; the offset in the energy used by the customer only accounts for about half of the total cost of service, which includes electric distribution, customer service, risk management and storm response. These costs are borne by customers not participating in net-metered solar, so the LADWP is quickly moving forward with an FIT model that avoids such cost-shifting. The first 20 MW offered in 2013 were fully subscribed in a week, attesting to the success of the programme.

The first 100 MW offering of the FIT was designed to incorporate best practices from around the world, such as a declining rate structure over time. However, the final 50 MW exhibits the kind of policy innovation that cities must deploy to leverage their particular strengths. As previously discussed, LA has attracted hundreds of developers

Table 1 | LADWP estimates of renewable energy costs and generation characteristics.

Generation source Levelized cost ($ MWh−1) Capacity factor (%) Dependable capacity (%)Wind 105 24–37 10Solar PV PPA 116 25–32 27Solar PV FIT 152 19 27Geothermal 109 91–95 90Combined-cycle gas 80 59 100Simple-cycle gas 225 9 100

Solar PV FIT refers to the FIT programme for distributed solar (photovoltaic). Levelized cost is the cost of generating electricity at the point of interconnection to the electricity grid, computed by incorporating the initial capital, discount rate and continuous operating costs (such as fuel and maintenance) over the 2012–2032 timeframe. The capacity factor is defined as the ratio of actual output to potential output on the basis of manufacturer-specified peak power ratings. Combined cycle and simple cycle are variants of natural-gas-fired power plants: the former are more efficient and economical to operate continuously, whereas the latter are optimized for quick ramping of power production to balance immediate demand fluctuations4.

© 2013 Macmillan Publishers Limited. All rights reserved

NATURE CLIMATE CHANGE | VOL 3 | SEPTEMBER 2013 | www.nature.com/natureclimatechange 775

opinion & comment

keen to secure contracts to develop large utility-scale solar installations outside city limits and the LADWP distribution system. Under a new and innovative programme, the LADWP is requesting bids to develop four parcels in a 200-MW-capacity site in the Mojave Desert, northeast of the city, and committing winning developers to collectively install 50 MW of distributed solar capacity within the LA Basin as a condition of the award. By empowering developers to reduce costs though scale by bundling local projects along with utility-scale installations, LA is supporting local economic development and a healthier electricity grid in an affordable way. This new bundled model for the FIT’s second phase will provide valuable insight into the relative economics and ease of implementation of solar in LA, as compared to the ‘local-only’ FIT. Los Angeles will learn from these different models to determine the best paths for distributed solar implementation for the LADWP in the future.

Los Angeles offers a unique window onto the construction of a robust renewable energy portfolio to power a megacity. Cooperation between the city and the LADWP has allowed a dramatic increase in renewable penetration, despite political and technical obstacles; if LA can successfully implement its 33% target, it will provide insight into sustainably powering cities around the world. ❐

Mayor Antonio R. Villaraigosa1, Varun Sivaram1,2* and Ron Nichols3 are at the 1City of Los Angeles, 200 North Spring Street, Los Angeles, California 90012, USA, 2Department of Physics, Oxford University, Parks Road, Oxford OX1 3LB, UK, 3Los Angeles Department of Water and Power, 111 North Hope Street, Los Angeles, California 90012, USA. *e-mail: varun.sivaram@physics.ox.ac.uk

AcknowledgementsWe acknowledge Romel Pascual, Jonathan Parfrey, James Barner, William Glauz, Cindy Parsons, Mark Sedlacek and the LADWP Integrated Resource Planning team.

References1. Hoornweg, D., Sugar, L. & Trejos Gomez, C. L. Environ. Urban.

23, 207–227 (2011).2. Satterthwaite, D. Environ. Urban. 20, 539–549 (2008).3. International Energy Outlook 2011 (US Energy Information

Administration, 2011).4. Power Integrated Resource Plan (LADWP, 2012); available via

http://go.nature.com/Hd1U7e5. Griffin, J. M. & Puller, S. L. Electricity Deregulation:

Choices and Challenges Vol. 4 (Univ. Chicago Press, 2005).6. Renewables Portfolio Standard Quarterly Report: Q1 2011

(California Public Utilities Commission, 2011). 7. Wiser, R., Namovicz, C., Gielecki, M. & Smith, R. Electricity J.

20, 8–20 (2007).8. Yin, H. & Powers, N. Energy Policy 38, 1140–1149 (2010).9. Baer, W., Ederlman, E., Ingram, J. III & Mahnovski, S.

Governance in a Changing Market: The Los Angeles Department of Water and Power (RAND, 2007); available via http://go.nature.com/oyrlC8

10. Milstein, I. & Tishler, A. Energy Policy 39, 3922–3927 (2011).11. Sayeef, S., Heslop, S., Cornforth, D. & Moore, T.

Solar Intermittency: Australia’s Clean Energy Challenge (Australian Solar Institute, 2012); available via http://go.nature.com/3a54kL

12. Wiser, R., Barbose, G. & Holt, E. Energy Policy 39, 3894–3905 (2011).

13. Andersson, G. et al. IEEE Trans. Power Systems 20, 1922–1928 (2005).

14. Chiradeja, P. & Ramakumar, R. IEEE Trans. Energy Conver. 19, 764–773 (2004).

15. Couture, T. & Gagnon, Y. Energy Policy 38, 955–965 (2010).

© 2013 Macmillan Publishers Limited. All rights reserved