• Ray Schmitt

Salient's New Energy Forecasts: Supply and Demand, More Accurate than Ever Before

Energy is undeniably central to a thriving economy, perhaps more than any other factor; modern society simply would not function without the many forms of energy (physicist’s definition: “the ability to do work”) presently available. Thus, it’s not surprising that the cost of energy is a key variable influencing the health of the economy. Availability affects the health of people, too, as we're seeing in stark relief in Texas right now. Energy cost and infrastructure capacity are dependent on the supply/demand balance, so energy companies and consumers have a significant interest in anticipating the futures of both. This is especially true for increasingly important renewable energy sources like wind, solar and hydropower.

We are pleased to announce that Salient Predictions has developed a new suite of products to forecast future renewable energy supply and demand on timescales up to a year ahead – with unprecedented accuracy. (We can also predict, as always, the risk to infrastructure from drought, deluge, fires and storms.)

Salient is leveraging unique new insights into the workings of the climate system and using the latest machine learning tools and techniques to provide its customers with the most accurate temperature and precipitation forecasts on timescales from weeks to a year ahead. We’ve demonstrated superior performance in open competitions and in decades of back-testing against the leading forecast products. We have now developed specific products that will empower energy companies to better anticipate energy demand and renewable energy supply well into the future of our changing climate.

Energy Demand

A cold, snowy winter will increase the demand and pricing for heating oil and natural gas (among other resources); similarly, a mild winter will depress demand and prices. A stretch of hot, humid weather in the summer can send electricity requirements soaring and stretch utility resources to the limit. Salient has developed new tools to forecast such events – and their impact on power markets and resources – from two weeks to 12 months out, anywhere in the world.

Energy customers care about a broad range of forecast stats, but one especially useful metric for the winter is the number of Heating Degree Days (HDD = number of days with temperature less than 65oF). Below we show our recent estimate of heating degree days for the next month in the US, as well as the anomalies from climatology.

Figure 1.a: Total Heating Degree Days (HDD) for CONUS for the next month, February 9 to March 10, 2021.

Figure 1.b: Predicted anomalies from climatology for Heating Degree Days (HDD) for CONUS from February 9 to March 10, 2021.

The figures show generally quite cold conditions for much of the midsection of the country, and modestly increased heating demand for the Northwest and Northeast. The Southwest and Southeast are expected to have relatively mild conditions with decreased heating demand. We nailed pretty well in our Winter outlook this pattern of February cold in the Midwest and North after a mild December and January.

Of course, the summer brings quite different energy challenges, as excess heat and humidity will drive up the usage of air conditioning in offices, factories and homes. This will cause electricity demand to spike on warm afternoons, and send prices soaring as high-cost, peaking power plants are brought online. An electric utility that is forearmed with accurate knowledge of the timing and location of future hot spells can secure its electricity supply at more favorable pricing than a utility that only reacts to conditions as they arise. The need for air conditioning is measured by Cooling Degree Days (CDD = number of days with temperature > 65 degrees F). Figure 2.a below shows our prediction for CDD for May through July 2021 for the US. Figure 2.b shows the predicted deviation from climatology for the same period. No other forecaster can supply such accurate forecasts at such long lead times.

Figure 2.a: Salient’s January forecast for Cooling Degree Days across CONUS for May through July.

Figure 2.b: Our predicted anomalies from climatology in Cooling Degree Days for May through July 2021. Warmer conditions are forecast everywhere, but will be especially high in Texas.

Renewable Energy Supplies

2020 was another record warm year around the globe, with heat waves in Europe, extraordinary wildfires in Australia and the Western US, and an unprecedented number of tropical storms in the Atlantic. The need to reduce our reliance on fossil fuels has never been more apparent, and we’ll require significant growth in wind and solar renewable energy sources to deal with our rapidly changing climate. Both wind turbines and solar farm installations are quickly expanding, and anticipating future supply will be key to the management of these renewable resources. We find that available solar power is inversely related to our high-quality precipitation forecasts, as one would expect; clouds bring rain and block the sun. Wind power, derived from the hourly variations in wind and converted to an actual power availability metric (proportional to the cube of the wind speed) tends to have skewed statistics (like rainfall), though does not display the same patterns. The figures below show our predictions of both solar and wind power for the next month as a percent of normal for all of North America. Such forecasts are available for anywhere on the globe at lead times from 3 weeks to 12 months.

Figure 3.a: Solar power available over CONUS for February 9 to March 10 as a percent of normal. Output in Texas, the Southwest and Florida will be good, while most of the rest of the country will be near normal, with a slightly below normal stretch from the Mid-Atlantic states to Arkansas.

Figure 3.b: Wind power anomalies for North America and offshore regions for February 9 to March 10, as a percent of normal.

These maps show that there will be good wind power supply for the Northwest, Southwest and across the South to Florida, while the Midwest, Mid Atlantic and Northeast will be below normal. Offshore wind sites on the East Coast should perform well. Though many wind farms there are still in the planning stages, such short- and long-term forecasts of their performance will have great value to those managing them.

Hydropower Resources: Salient is also developing a sophisticated hydrology, snowpack and river flow model in partnership with Climatics. The predictive model utilizes our highly accurate precipitation forecasts in specific watersheds to forecast stream flows and reservoir levels. These outputs let us anticipate hydropower resources up to a year in advance. While hydropower resources have been mostly built out and won’t be expanding like wind and solar, efficient management of the resource will be greatly aided by Salient’s accurate long-range forecasts. More information on our hydropower forecast can be found in our previous blog post by Ken Westrick and myself. Hydropower resources are especially critical because they can also serve as an important storage batteries for the more variable solar and wind resources in our power grids.

Asset management: Whether it’s springtime tornadoes in the Midwest, summer fires in California, fall hurricanes in Louisiana or winter nor’easters in New England, it’s important for utility companies to pre-position field crews in anticipation of outages or to plan shut-downs of power lines that pose fire risk. The ability to look far into the future of weather risk for assets anywhere on the globe will enable pricing and hedging decisions that smoothly anticipate coming risks and opportunities.

Ensemble Probability: For Salient’s customers, it’s important that we make a probabilistic forecast available as an aspect of each forecast we produce. That is, it’s often useful to know not just the most likely forecast conditions, but also the possibility for an unusual extreme. Examples are shown in Figures 4.a and 4.b below. These charts show forecasts for average precipitation anomalies, month by month, for all of 2021 in Minneapolis. Note that the probability distribution information shows that, while the median forecast may be moderate, there are times when there occurs a significant chance for a much larger rainfall excursion, such as in September and October. Figure 4.b shows week-by-week spreads in the ensemble forecasts for temperature anomaly in Minneapolis from now to mid-May. This sort of information helps customers hedge against the possibility of extreme events.

Figure 4.a: A month-by-month time series of the spread in precipitation anomalies for Minneapolis, Minnesota for 2021. The possibility of high rainfall events peaks in March and September.

Figure 4.b: The week-by-week spread in temperature anomalies for Minneapolis from now to mid-May. Cooler conditions are forecast through March, with warmer conditions after April 1, and the spread in likely outcomes is shown by the various color shadings.

Summary Overview

Salient’s new suite of energy demand and renewable energy supply forecasts will enable asset and plant operators, grid managers and power marketers to better manage their resources and price risk exposure. Having reliable long-range forecasts for wind, solar and hydropower resources – as well as for energy demand – will help large grid operators to maximize efficient use of their assets and decrease the cost of electricity for consumers. Assets at risk of heavy rain, drought or extreme temperatures can be better protected. Energy traders will be able to hedge against future price swings and anticipate market volatility. Don’t let your company be left in the dark without these new forecast tools to drive better financial and operational decisions. Our forecasts are significantly better than both climatology and any competing product, and are available via a variety of different formats (including API, flat files, map-based, etc.). Please reach out to us at if you’d like to learn more.

33 views0 comments

© 2020 by Salient Predictions, Inc.