This fall, as millions of Californians went without power and wildfires claimed homes and livelihoods, the fragile state of California’s electricity grid came center-stage. Our deep reliance on it to power our homes, businesses, schools, healthcare and medical devices, and other critical services also became painfully clear. Understandably, the current focus is on the grid’s resilience to wildfires.
The unfortunate reality is that wildfire risk is just part of the story. Heat waves, droughts, and flooding affect the grid and will become even more frequent and severe — in California and across the nation – if heat-trapping emissions remain unchecked. Heatwaves can also co-occur with wildfires, increasing risks to the grid. This raises important questions about the electricity grid’s readiness for more extremes today and in the future.
At UCS, we wanted to know what risk, if any, intensifying heatwaves and extreme heat would pose to grid reliability in California’s San Joaquin Valley (SJV or Valley) region by midcentury, absent adaptation and global action to reduce heat-trapping emissions.
This blog describes our analysis of the risks from extreme heat to different grid components using information online; our findings for electricity use, transformers, and transmission lines; and outstanding questions about the climate readiness of these various parts of the grid. We also make several recommendations to enhance the electricity grid’s resilience and reliability.
Our analysis shows that, absent adaptation and global action to reduce heat-trapping emissions, higher temperatures would increase electricity use for recently constructed (or older) schools or residences by mid-century while at the same time potentially constraining the grid’s ability to bring enough energy to meet those needs. This means an already stressed grid could have less cushion to absorb disruptions and prevent service interruptions. Estimating the full extent of the threat requires additional analyses and information from utilities, which we describe below. We also have several remaining questions concerning the condition and readiness of the grid for a future with more heat-related climate variability and extremes. Regardless, we know that utilities and grid operators must design, plan, and operate the grid for the impacts of rising temperatures by mid-century and beyond.
The San Joaquin Valley’s Grid is Already Vulnerable
We focus on the Valley because it is one of the fastest growing regions in California, as well as a region challenged by high rates of poverty. It is one of the warmest areas in the state and climate models project even higher daytime and nighttime temperatures and more frequent and intense heatwaves here by midcentury. Many residents are already vulnerable to heat-related illness. Reliable, safe, and affordable electricity will be critical to prevent heat exhaustion, heat stroke, or even death.
The grid’s condition and design impact its ability to withstand and adapt to climate extremes. Utilities do not normally share information about the age or condition of the grid, so we found very little information online. However, a state investigation found that recent events related to the Camp Fire are “indicative of an overall pattern of inadequate inspection and maintenance of PG&E’s [note: Pacific Gas and Electric Co is the largest utility in the region] transmission facilities.” It’s also widely reported by the media that some utilities have underinvested in the electricity system here in California.
We also know that engineers traditionally designed the grid assuming past climate and weather trends were reasonable predictors of future conditions. As a result of global warming, this is an increasingly invalid assumption. Heat-related equipment failures as recent as earlier this year led to Valley power outages. Grid infrastructure like transmission lines is long-lived (average lifetime of 50 years), so when it was built matters. Today’s grid faces other stressors as well for which it wasn’t designed that can compound its vulnerability.
The two main utilities serving the SJV, Pacific Gas and Electric Co (PG&E) and Southern California Edison (SCE), have taken some steps to upgrade their grid assets in response to past extreme events. But as we describe in our solutions section – and the regulatory community and utilities themselves acknowledge — more must be done.
Hotter days and nights in the San Joaquin Valley by midcentury would further stress the infrastructure people rely on to keep cool when it is dangerously hot outside, like the electricity grid. For example, 1.3 million customers lost power during the July 2006 heat wave in California when over 2,000 distribution transformers suffered heat-related failures.
Homes and Schools Would Face Higher Electricity Use for Cooling
Using climate-adjusted electricity-use data provided by Prof. Brian Tarroja, PhD, P.E. of UC-Irvine, we estimated the percent increase in electricity use (annual, monthly, and peak) for space cooling due to rising temperatures by midcentury at recently built K-12 schools and residences in the SJV. Rising electricity use could result in higher energy bills1 and increased load on the grid, unless offset by on-site generation and/or energy efficiency and conservation.
Energy bills are the second highest housing cost after rent or mortgage. Affordability concerns may mean low-income households, who already spend a larger portion of their income on energy, face a difficult choice between turning on air conditioning or paying for something else when temperatures skyrocket. This is especially true for low-income renters who tend to live in older, less efficient housing in the hottest parts of cities.
Similarly, older schools often have “outdated, inefficient, and … malfunctioning” HVAC whose use can make up a significant portion of school district electricity bills. While very little information exists on the characteristics of public K-12 school facilities statewide, we do know that in 2012, 30 percent were over 50 years old, which is before the state’s energy efficiency codes took effect.2
When compared to historical levels (2000-2010), annual electricity use for space cooling is projected to increase 20-29 percent by midcentury for residences. Schools would see a 30-39 percent increase. (The ranges represent Southern SJV and Northern SJV, respectively.) These changes are based on temperature data from high-resolution climate models assuming no global action to reduce heat-trapping emissions (or the RCP 8.5 scenario).3
We also looked at monthly total electricity use for September, which is the first full month of school during the warm season. Percent increases due to the influence of rising temperatures on space cooling by midcentury are shown in the graphic below. Not surprisingly, May and October electricity bills would rise as well but to a lesser extent. Monthly peak load would increase as well, with potential implications for electricity bills. Given its assumptions, the analysis may underestimate electricity use increases for older homes and schools.
If no global action is taken to reduce heat-trapping emissions (or RCP 8.5 scenario), hotter temperatures in the month of September by midcentury would lead to increases in total electricity use (as a result of more space cooling) of 10-13 percent for schools, and 9-11 percent for residences in the SJV compared to historical levels. More electricity use could translate to higher energy bills and less resources available for other key educational and household needs and increased load on the grid.
Transformers and Transmission Lines Could Have Less Capacity to Bring Power to Customers When They Need It Most
Increased electricity needs are just one part of the problem facing the Valley as climate change raises temperatures. Very hot days (and nights) combined with high loads to meet air-conditioning needs can push transmission lines and transformers up to and beyond their design limits. Heatwaves can lead to reduced capacities (or power transfer capabilities), shorter lifetimes, and even failures. They are especially problematic if electricity demand for cooling skyrockets while equipment is already operating at or near its rated capacity, and there are no fans or little to no wind to cool them down. (Here is more information on how extreme temperatures affect transformers and transmission lines.) For example, during the July 2006 heat wave in California, over 2,000 distribution line transformers failed, resulting in about 1.3 million customers losing power.4
Utilities and grid operators work to build redundancy into the electricity system so that failures like these do not lead to service outages. As emissions rise, this means adequately planning and operating the grid for longer and more frequent, severe, and widespread extreme heat events.
We wanted to know how rising temperatures would change capacities of transformers and transmission lines in the Valley and increase the risk of equipment damage or failure, absent adaptation. We estimated how often ambient temperatures would exceed the temperature used to rate equipment by midcentury. A piece of equipment’s rating, or the maximum current and voltage that it can safely deliver, depends on several assumptions, including maximum normal operating temperature,5 cooling, ambient temperature, and others. The number of days with temperatures exceeding the rating temperature can therefore serve as a proxy for days with increased risk of fully loaded or nearly fully loaded equipment exceeding their normal thermal limits, assuming little to no wind. We used ambient temperature data from high-resolution climate models developed by Pierce, Cayan, and Dehann for a historical period (1975-2005) and for a midcentury (2035-2064) period under two emissions scenarios: no global action to reduce heat-trapping emissions (RCP 8.5) and slow action (RCP 4.5). More information on specific caveats can be found here.
If transformers and transmission lines operate too long beyond their design limits, they can overheat, risking damage or even failure. Available capacities of this equipment are reduced to avoid this result, as laid out in asset management strategies. Other impacts of hotter days include increased line losses, and higher electricity costs due to constraints on transmission lines and generators.
Transformers
The rating temperature for transformers is a 24-hour average of 86°F (with an absolute temperature not to exceed 104°F at any point) based on an industry standard in lieu of missing utility-specific information. Days above this temperature would become more frequent in the SJV by midcentury. The cumulative deterioration of insulation in transformers leads to failure, so the severity and frequency of these extreme temperature days matter. We had access to data for substations, which house transformers.
As shown in the figure below, more than 40 percent of the region’s substations (230 out of 580) would face at least 30 or more days per year above the rating temperature threshold of 86°F by midcentury, if no emissions reduction action is taken. Historically, the vast majority of substations saw a median of one or no such days annually.
Between 20-40 percent of the SJV’s substations would experience at least 30 or more days per year above 86°F, from the midcentury Slow Action or No Action scenarios, when ambient temperatures would increase transformers’ risk of reduced operating capacity or early failure assuming winds are calm and transformers are fully loaded without sufficient cooling. The presented maps represent the median number of days across climate models for each period and scenario.
Public data and information gaps prevented us from projecting what specific capacity reductions might occur to keep transformers within thermal limits and minimize damage. We have the following questions for utilities and agencies:
- How many transformers are equipped with adequate cooling to avoid overheating? The CEC substation dataset did not have this information.
- What ambient temperature is used by utilities to determine transformer ratings in inland areas? Is it appropriate given that daytime and nighttime temperatures are hotter today than in the past and will become even hotter in the future?
- What is the age and condition of current transformers?
Regardless of these gaps, our analysis makes it crystal clear that regulatory agencies should be asking utilities these questions to better understand how they are planning for future conditions, and in a transparent way.
Transmission lines
For transmission lines, we used a daily maximum temperature of 109°F, based on PG&E’s rating methodology for inland lines.6 Transmission lines would be exposed to ambient temperatures exceeding 109°F much more frequently by midcentury if there is no action to reduce carbon emissions. Line operations are limited by the section of the line with the highest operating temperature, which is influenced by ambient temperature, wind, and other factors.
Whereas no lines experienced more than one week annually at ambient temperatures exceeding 109°F historically, nearly 90 overhead lines in the Valley (totaling 1,700 miles out of 9,500 miles in the SJV) would be exposed to more than two weeks’ worth of days above this temperature. Slow action would result in just two lines, totaling 70 miles, facing more than 14 days, and 170 lines (2,600 miles) exposed for 7-14 days per year by midcentury. Fully loaded lines are particularly at risk of operating with reduced capacity on these oppressively hot days when air conditioning demand rises and if there’s little to no wind. Redundant lines with available capacity to compensate would be critical, yet increasing and more widespread heatwaves (local and regional) would stress many lines at once.
An additional challenge from heat-trapping emissions is presented by more extremely hot days and warm nights in a row. Hotter nighttime temperatures can affect lines’ ability to cool overnight and recover before the next day’s increase in load.
If no global action is taken to reduce heat-trapping emissions, total miles of transmission lines exposed to more frequent and severe ambient temperatures above 109°F will rise dramatically, especially lines serving several major urban areas and several unincorporated communities.
As with transformers, some important public data and information gaps kept us from quantifying the possible magnitude of capacity reductions under no action and slow action scenarios.7 Recent studies help shed some light, with one study estimating a 9°F increase would result in as much as a 7.5 percent reduction or more on a fully loaded transmission line, depending on assumptions. The same study projected an average reduction of 2-3 percent for transformer capacity, while another one found a 10 percent reduction in transformer life for every 1°C increase in ambient temperature. A different study called for more research on climate impacts on the transmission system. We have several questions for which we could not find publicly available data.
- Which conductors are used in specific parts of the SJV? Maximum allowable temperatures vary with the conductor but this information wasn’t available online.
- What is the projected peak load for lines? And to what degree are operators currently using real time weather conditions (dynamic thermal ratings or ambient adjusted ratings) to identify the optimal current rather than a seasonal rating?
- What is the condition of the SJV’s lines? How redundant are they8, and is redundancy planned with these temperatures in mind?
The SJV’s future of more regional heatwaves and hotter days and nights raises an important question: is it still appropriate to use 109°F for rating and transmission planning purposes to minimize risks.
Distribution systems
Historically, most outages are due to events on local distribution systems. They tend to be smaller in scale and less costly than transmission outages, but their impacts are still very real, especially if people have fewer resources and ability to respond. According to several experts, transformers that are under high temperatures and high load tend to be the limiting equipment in distribution systems, rather than lines. We estimated how many days the ambient temperature in each city would exceed a 24-hour average temperature of 86°F in order to understand the risk to urban distribution transformers, as shown below.
Cities in the SJV would face considerably more days with temperatures above a 24-hour average of 86°F by midcentury, especially in the central and southern areas, if no global action is taken to reduce heat-trapping emissions. Slow action would reduce the frequency in midcentury, and rapid action would decrease the number of days at the end of the century to a similar level. Unless planned for, these changes would place fully loaded urban transformers that have little to no cooling at increased risk of early failure or reduced capacity.
Generators May Experience Reduced Efficiency and Capacity
Hot weather makes the conversion of energy to electricity less efficient with lower output. Many studies quantify potential reductions. One recent study on combined cycle power plants assumed a capacity reduction of less than 0.5% per degree C above the generator’s rating temperature. Another on solar photovoltaic panels found efficiency will decrease slightly as temperatures rise. Small individual changes may become increasingly important as California depends more and more on clean energy to power its grid. Reductions in hydropower generation would be due mainly to impacts of warmer winter and spring temperatures on snowpack and snowmelt rather than summertime extreme temperatures.
Solutions Exist Today
Utilities, grid operators, and planners need to plan for the impacts of ever-increasing climate variability and extremes, especially for long-lived investments like transmission lines. The number and severity of extremely hot days and nights facing the grid and electricity customers would increase considerably if heat-trapping emissions continue to rise. Answers to several outstanding questions will help clarify just how ready the power system is for the Valley’s hotter future.
Fortunately, many technical and operational solutions exist today for a more climate-resilient grid. The following heat-related recommendations can help enhance the grid’s readiness for other climate-impacts as well:
- “Non-wires” solutions that reduce reliance on transmission lines and transformers and diversify energy sources, such as distributed energy resources, storage, and micro-grids; more widespread implementation of energy efficiency; and flexible demand solutions. Clean-energy approaches have the additional benefit of decreasing heat-trapping emissions and protecting against electricity rate shocks. These solutions should be accessible to vulnerable communities in the Valley.
- Updated grid technologies and operational and maintenance practices. If they have not already, utilities should quantify the risk of extreme heat to their transmission and distribution systems by midcentury. Thermal limits should be updated to account for more frequent, severe, and longer-lasting heat events. Utilities should consider smart grid technologies and operational approaches that increase the capacity and efficiency of existing power lines and transformers (e.g., cooling systems, reconductoring with higher capacity lines, dynamic thermal ratings or ambient adjusted ratings). They should also dedicate more resources towards grid monitoring (both components and weather conditions) and maintenance.
- Revised design and equipment standards – for the electricity grid, buildings, and appliances – and planning approaches. They should better account for effects and timing of more intense and longer-lasting heat events over the lifetime of a project, including longer planning timelines, and increase energy efficiency. Specific planning timeframes, criteria, and other approaches for integrating climate impacts into key planning processes and documents (like the Integrated Energy Policy Report, Transmission Plans and Integrated Resource Plans) should undergo external expert review to ensure they are appropriate and rigorous.
- Expanded programs that provide cooling assistance to low- and fixed-income households. Some California programs have millions of unspent dollars. Local and state agencies should improve these program’s outreach efforts, using materials translated into locally relevant languages.
- More inclusive and transparent processes for infrastructure decisions. Private sector, utilities, and government authorities should employ community engagement best practices to encourage participation by underserved and marginalized groups and groups most vulnerable to heat to ensure investments address their needs. Otherwise, decisions risk unintentionally harming these communities and perpetuating or worsening existing inequities.
While utilities in California have made some progress on several of the above recommendations, they also admit more must be done. The California Public Utilities Commission is also holding a proceeding on climate adaptation for the electricity sector. Updates to the grid should not unduly burden the ratepayer; instead they should be supported by robust analyses of their need, to ensure they are good value rather than “gold-plating.”
A more resilient and equitable grid is central to a safer future, but by itself, it will not be enough to protect people from heat over the long-term. There must also be deep and swift reductions in global carbon emissions (including here in the U.S.), consistent with, or ideally going beyond the Paris Agreement goals. California set a goal of achieving carbon neutrality by 2045, which will require a complete transformation of its economy. A cleaner, smarter electricity grid and more efficient buildings will play a key role in meeting this target and increasing resilience.
Energy efficiency strategies and solar power are two strategies that can help increase buildings’ resilience to heat-related impacts on the electricity grid. The California building code requires solar panels on new single family homes and multifamily buildings up to three stories high built starting in 2020.
The Sum is Greater Than the Parts
Continued emissions of heat-trapping gases will turn up the heat on an electricity grid that is already stressed, causing it to operate closer to – and possibly beyond — its limits. Very hot, long, and widespread heat events increase electricity use for cooling and affect multiple grid components simultaneously. While individually each effect on the grid may be manageable, their combined impact on the power system’s ability to absorb disruptions and prevent service interruptions could be significant, with potentially dire consequences, if not addressed in grid planning and operations.
It is not possible to avoid every negative effect of extreme temperatures on the grid but risks can be minimized and ratepayer dollars spent more wisely if electricity infrastructure, planning, and operating practices are climate-resilient, low carbon, and equitable. Projected growth in the SJV over the next few decades provides an important opportunity to reshape the grid to better protect and meet the needs of a growing population during dangerous heat events. As grid operators and planners work to ensure that inevitable equipment failures do not lead to outages, they must plan accordingly for an increasingly variable and extreme climate future.
We did not find enough answers to reassure us that the negative effects of extremely hot days and nights and heatwaves on the condition and operation of the SJV’s electricity grid – though potentially small individually – would not be widespread by midcentury. Given what is at stake, residents of the San Joaquin Valley and the rest of California deserve to know.
The author would like to thank the following people for their review of the analysis: Merwin Brown, PhD (CIEE); Joseph Eto, PhD (LBNL); Guido Franco, P.E. (California Energy Commission); Prof. Brian Tarroja, PhD (UC-Irvine); Prof. Alexandra von Meier, PhD (UC-Berkeley); Laura Wisland (Heising Simons Foundations). The opinions expressed in this blog do not necessarily reflect those of the reviewers. Dr. Owen Doherty (Eagle Rock Analytics) and Dr. Pablo Ortiz (UCS) assisted with the analysis.
1 It is challenging to predict whether reduced natural gas use for heating might offset this increase since electricity and natural gas rates could change between now and mid-century. Currently, natural gas is significantly cheaper per thermal unit than electricity in the PG&E service territory. A back of the envelope analysis using PG&E’s current base rates for annual electricity and natural gas use revealed slight net increases in residential energy bills and much larger increases for schools. One school in the Valley shared that space cooling costs were currently significantly greater than heating costs.
2 Projects funded by Proposition 39 have improved efficiency at more than 4,400 school sites (out of more than 10,000 statewide). More than 20 percent of schools in the Valley had on-site solar PV in 2017 (source: gosolar).
3 While the ‘no action’ scenario assumes no global action, Tarroja does include a high level of renewables in California.
4 The 2006 heatwave had 12 consecutive days above 105°F in Fresno and 5 consecutive nights with temperatures staying above 80°F. Cal-Adapt projects an average of 13 days in a row above 105°F and 2 nights back-to-back above 80°F in Fresno by midcentury.
5 Refers to hot spot temperature for transformers.
6 SCE uses 104°F, but for the sake of consistency we will use 109°F for SCE lines as well.
7 We did not consider the effect of changes in wind speed between now and midcentury on line operating temperatures. Wind has a cooling effect that can lower line temperatures. However, there is a high level of uncertainty in projecting these changes at a local level due to global warming according to climate experts. Regardless, a substantial increase in the number of days with ambient temperatures above 109°F points to the need to revisit line operations and planning to minimize risks. Future analyses should incorporate robust hourly wind speed projections (and hourly temperature projections) as they become available.
8 The degree of redundancy in the transmission system for the SJV was not clear for each line in the CEC dataset. In PG&E service territory, no urban areas in the SJV have distribution systems with secondary networks, which are more reliable than radial networks.
UPDATE 1/3/20: A label in the schematic in this post has been updated to read “conventional power plants”