« Back to Get the Point main page

Maximum Overdrive:

Natural Gas Power Burn Potential in Summer 2016

May 26, 2016 | By Jack Weixel

Many questions have been raised by the natural gas market about the potential for natural gas power burn demand this summer. With high storage inventories and production levels that are slow to decline, many are left wondering what sectors of demand will take advantage of the market’s low-price environment.

This edition of Get the Point will examine what impact power burn might have at reducing the surplus. We will examine historical pipeline flows and modeled data to unlock the maximum potential of natural gas demand from the electric power sector.

This is an important metric at a time when natural gas prices continue to languish in the $2.00 per MMBtu range. The role of natural gas-fired electric generation has increased as a baseload power source for electric utilities due to environmental and competitive hurdles faced by coal. For the summer of 2016, the level of power demand will have a significant impact on the level of natural gas injected into storage and the price of natural gas heading into winter 2016-2017.

Furthermore, the price of natural gas has implications on the future of lower 48 U.S. natural gas production rates and the supply and demand balance of the natural gas market.

PointLogic measures natural gas burned at power plants using a sample of 645 meter points that deliver gas from interstate pipelines at various locations across the lower 48 U.S. Taken together, these meter points deliver gas to electric generators that make up about 60% of total lower 48 natural gas-fired electric generation capacity.

The remainder of installed capacity is either behind local distribution company (LDC) citygates or on intrastate pipeline systems whose demand profile remains invisible due to the lack of nomination data available to the market. In order to account for the natural gas burn that occurs at these plants, PointLogic must model estimated power burn for the lower 48 using a regression formula that relies on Energy Information Administration (EIA) survey data, temperature and population weighting.

To account for state and regional differences in power burn rates in the sample, power plant sample flow is subdivided by state and region, then modeled individually by region to come up with a total lower 48 modeled estimate. Fortunately, the PointLogic sample flow data tracks very well with EIA survey data and modeled estimates by region; our correlation is over 95% for modeled estimate and sample flow in each PointLogic region and the lower 48 as a whole.

The lower 48 power burn model estimate and sample are charted in the graph below, which shows the past 5 years of data:

Modeled Power Burn vs. Sample Flow (Bcf/d)

Figure 1 Source: PointLogic Energy

As evident in the graph, power burn has seasonality. Summer power burn (April to October) is typically higher, driven by higher temperatures as measured in cooling degree days (CDDs), and winter demand for power burn is typically lower. 

Looking across a period of years, the graph shows that power demand is steadily increasing each year. With each successive winter (November to March) since 2011-2012, winter is showing a higher average burn per degree day. The increase in installed generation capacity was the reason a few years ago, and this has been accelerated in the last year and half by the low price of gas, which has encouraged increased use at existing facilities.   

For the purposes of this article, we will focus on summer power burn, as it has become a major determinant of how much gas is injected into storage each summer and, ultimately, the price of natural gas heading into each winter. 

Modeled Power Burn (Bcf/d)

Figure 2 Source: PointLogic Energy

In terms of average power burn witnessed each summer season, the lower 48 U.S. has registered between 24.0 and 27.9 billion cubic feet per day (Bcf/d) of demand since 2012 (see Figure 2). It’s important to note that even though the summer of 2012 showed a higher monthly average of power burn during the peak CDD month of July (as shown in Figure 1), the total power burn in summer 2015 exceeded 2012. This was due to increased power burn demand in August, September and October.

As the price of natural gas stayed weak this year, this trend has continued. April and May 2016 power burn is expected to average nearly 2.0 Bcf/d higher than April and May 2015 levels. 

Average Montly Modeled Power Burn (Bcf/d)

Figure 3 Source: PointLogic Energy

July 2012 power burn averaged 34.6 Bcf/d -- a record monthly average high that still stands today. Also, July 2012 holds six of the 10 highest individual power burn days.

But July 2015 also set some impressive marks. Digging into the daily data from PointLogic Energy’s supply and demand historical database reveals that July 28-30, 2015 power burn is the three highest power burn days ever, with July 29 setting the daily record high at an impressive 38.3 Bcf/d. 

Power Burn History

Figure 4 Source: PointLogic Energy

Summer 2016 Demonstrated Maximum Power Burn Potential

EIA uses the term “demonstrated maximum working gas volume” to reflect the sum of peak natural gas volumes that can be held in storage across the lower 48 U.S. To calculate this number, EIA has looked at peak inventory levels of all 385 storage facilities in the lower 48 from December 2010 to November 2015to deduce a figure that represents the maximum amount of gas that can be stored at any one time.

Through November 2015, this number was 4,343 Bcf. This represents the non-coincidental maximum amount of gas that could ever be held at each individual storage facility at any point during the subject time period.

EIA then further divides this number by region. But this number is not the same as the design or operational capacity of these same storage facilities – the latter are much higher. The higher number tries to measure what has been observed and therefore what could be achieved again, provided each facility in each region is holding its maximum potential amount of gas.

Start a Free Trial of PointLogic Energy's Pipeline Flow Data

The same analysis can be applied to power plant burn by region – it is just one way to measure the potential to burn natural gas in a low-price market. We can use our PointLogic sample of 645 power plant meters by region to examine on a daily or monthly basis how much gas could possibly be burned over the course of summer 2016.

Reviewing our sample of power plants by region on a daily basis going back to the summer of 2012, we can deduce the maximum sample deliveries to each power plant meter on specific days and model this sample number to provide a model estimate. For the purposes of this example, we will focus on the peak power burn months of July and August 2016. Because they are peak months, July and August have significant influence over both the price and the amount of gas injected into storage during September and October.

We can create this analysis because we already know that power burn during the shoulder months in 2016 of April, May, June (to a lesser extent), September and October will be higher on average than those corresponding months in 2015. The higher burn reflects increased utilization of gas-fired capacity witnessed over the past several years.

The figures that follow show our maximum daily sample on non-coincidental days and the corresponding modeled estimate. 

Max Demonstrated Power Burn July (Daily in Bcf/d)


Figure 5 Source: PointLogic Energy

In July 2016, the maximum modeled power burn — estimated using the highest sample burn for an individual day in each region -- is 43.5 Bcf/d. To put this number in perspective, the highest daily demand number on record for a single day in July is the July 29, 2015 record of 38.3 Bcf/d. In other words, for July 2016 there’s the potential for an incremental 5.2 Bcf/d during any given day of the month.   

Max Demonstrated Power Burn August (Daily in Bcf/d)

Figure 6 Source: PointLogic Energy

In August, the maximum modeled power burn estimate using the highest sample burn for an individual day in each region results in a modeled power burn number that is slightly lower than July, equating to 41.4 Bcf/d. To put this number in perspective, the highest daily number on record for a day in August is Aug. 2, 2012 at 36.8 Bcf/d. In this case, theres the potential for an incremental 4.6 Bcf/d of incremental burn for any day in August 2016. 

This analysis can be expanded to the full months of July and August. In this case, the maximum peak volumes are slightly more muted in absolute volume than the daily versions. This occurs because power burn demand is typically lower in the first few weeks of July and the last few weeks of August than during the peak days that overlap the end of July and the beginning of August. The monthly figures that follow demonstrate this effect.

Max Demonstrated Power Burn July (Monthly in Bcf/d)

Figure 7 Source: PointLogic Energy

For July, we see that the maximum monthly average peak modeled demand from power equates to an average of 38.6 Bcf/d, or about 4.0 Bcf/d higher than the record-setting average of 34.6 Bcf/d in July 2012. 

Max Demonstrated Power Burn August (Monthly in Bcf/d)

Figure 8 Source: PointLogic Energy

For August, the maximum monthly average peak modeled demand from power burn equates to an average of 36.0 Bcf/d, or about 3.2 Bcf/d higher than the record-setting average of 32.8 Bcf/d in August 2012.

Let's look at the impact that maximum monthly average peak burn could have on storage inventories. For every 1.0 Bcf/d increase in power burn during July and August, an extra 62 Bcf is not injected into storage (31 days in each month). Revisiting Figure 3, we can plug in these monthly average peak modeled power burn figures to compare July and August 2016 to past years.

Average Monthly Modeled Power Burn (Bcf/d)

Figure 9 Source: PointLogic Energy

Comparing the maximum monthly average to 2015 for the months of July and August 2016 yields an extra 5.2 Bcf/d of demand in July and 3.2 Bcf/d in August. To put this in storage terms, it would reduce storage injections by 260.4 Bcf compared to the 2015 months.

Many in the analyst community are expecting storage injections to top out above 4,000 Bcf at the end of October. If maximum monthly average power burn is realized in July and August, then we could see end of October inventories in the range of 3,740 Bcf. This would be a remarkable turnaround from the current situation, and in fact, it would be 50 Bcf below the 5-year average of 3,790 Bcf.

In Conclusion

It is important to remember that the analysis above is purely hypothetical. PointLogic is not stating that power burn will reach maximum levels in July and August this year.

But the exercise is still very valuable. Despite the hypothetical nature of the values derived using demonstrated maximum power burn potential, looking at these numbers shows just how much potential demand is available from the current natural gas-fired electric generation fleet.

Several factors will influence what actually occurs this summer in regards to power burn. Pros for a higher power burn number include:

  • Henry Hub spot prices remain in a reasonable range between $2.00 and $2.45 per MMBtu, as predicted by EIA’s Short Term Energy Outlook and PointLogic Energy’s forecast. This will encourage natural gas use on the margin, vs. other forms of fossil power.
  • The retirement of 19.0 GW of coal-fired capacity in 2015 and another 11.9 GW of retirements in 2016, according to our colleagues at IHS.
  • 5.8 GW of coal conversions to gas (out of the 11.9 GW of coal retirements according to IHS).       
  • Minimal potential for coal to gas switching, due to capacity constraints of existing coal fleet in the face of summer electricity load.

Factors working against a higher power burn number this summer include:

  • Weaker electric power demand if mild weather occurs.
  • Increased Henry Hub spot prices (due to increased demand from natural gas-fired generation) that pushes certain gas plants off the margin of utility dispatch curves.
  • Growth in generation contribution from the renewables sector, both current capacity and units coming online.

This analysis intuitively takes in to consideration a majority of the "pros" in the high summer power burn equation. Even though maximum power burn cannot be anticipated, PointLogic can still reasonably expect that power burn in the summer of 2016 will be higher than that of pervious summers.

Demonstrated maximum power burn potential is just one way to approximate the net impact on the lower 48 storage inventories and the supply and demand balance. Ultimately, the level of 2016 power burn demand will be decided by the choices that electric utilities make to nominate natural gas to power plant meters as it suits their region’s electric power load.

Stay tuned to PointLogic’s Data Suite and Supply and Demand trackers as we monitor summer power burn and its impact on the pace of storage injections, the price of natural gas and pace of lower 48 dry production over the months to come.    


« Back to Get the Point main page


Sign up here to have Get the Point delivered to you each week!





Share With Your Colleagues






© PointLogic Energy, an OPIS Company | Site Map