January 10, 2018 | Jack Weixel & Sadie Fulton
In the middle of November, PointLogic published a look ahead at this winter’s gas demand picture, “Addressing the Elephant in the Room: Residential and Commercial Winter Demand.”
In that “Get the Point” article, we looked at what would happen to demand during a “normal” weather scenario and compared that to a 90-day demand forecast derived from population-weighted weather data provided by our weather partner StatWeather. We also took a look back at the past five winters to draw historical comparisons to overall levels of residential and commercial demand (res/com) and its impact on domestic demand. We then layered rising export demand into the equation and came to the undeniable conclusion that this winter would be one of the highest total demand winters on record. We are now nearly midway through winter 2017/18, and within the next 90 days, the winter withdrawal season will come to a close. It’s time to revisit that analysis.
A lot has happened since mid-November. Ten days ago, the U.S. natural gas market witnessed its single-highest domestic demand day, its highest res/com demand day since January 2014, and its highest total demand day, each occurring on Jan. 1, 2018. In addition, the week ending Jan. 4, 2018 set records for the highest weekly average demand periods for res/com, industrial, domestic and total demand.
Inevitably, comparisons were made to the Polar Vortex winter period in early January 2014, when all previous single-day demand records were smashed. The “bomb cycle” storm that blanketed the Northeast with cold, spread snow to parts of the southeast U.S., and stunned many a Florida iguana, only added to the significance of the past two week’s weather pattern.
In this edition of Get the Point, we’ll measure the season’s res/com demand to date and look forward to the end of the season to provide estimates of end-of-winter storage inventories.
Figure 1: Winter Res/Com Demand, Source: PointLogic Energy
The graph shown above makes one thing clear: this winter’s res/com demand through the end of December was not unusual. However, the differential between early winter (November and December) and late winter (January, February and March) forecasts for winter 2017/18 is the largest seen in the last six years (these forecasts account for true January weather through January 8, and either forecasted weather or normal weather through the remainder of winter). While November and December averaged 29.7 Bcf/d, the StatWeather 90-day forecast projects that January through March res/com demand will average 42.1 Bcf/d. For comparison, if weather for those months is the eight-year average, res/com demand would be 40.5 Bcf/d.
As we know, January 2018 came with a bang. For the first week of 2018, res/com demand came in just under 65.0 Bcf/d – nearly 5.0 Bcf/d higher than the next-highest first week of the year this decade, which was seen in 2014. Further, 2018 has already logged two of the top ten highest res/com demand days in the last six years. (Eight of those top ten res/com demand days came in January in one year or another.)
But keep some perspective in mind: even with this strong start, January 2018’s res/com demand is expected to fall short of 2014 and 2015 January levels.
Figure 2: January Res/Com Demand, Source: PointLogic Energy
What Has Occurred and What Lies Ahead
The next 80 days through March 31, 2018 will determine whether winter 2017/18 was a demand dud, like the past two winters, or a shock to the system like winters 2013/14 and 2014/15.
Res/com demand is averaging 37.1 Bcf/d this winter to-date, compared to 40.3 Bcf/d and 37.9 Bcf/d in winters 2013/14 and 2014/15 to-date, respectively. However, due to strength in the power and industrial sector this winter to-date, domestic demand has come in at 84.3 Bcf/d, which is 0.4 Bcf/d higher than winter 2013/14 to-date and 3.1 Bcf/d higher than winter 2014/15 to-date.
Figure 3: Domestic Demand Winter to-date, Source: PointLogic Energy
Plotting out the rest of winter by month using the StatWeather-derived 90-day forecast data through March 31 yields the graph shown below. Winter 2017/18 should average 38.7Bcf/d of res/com demand, good enough for third place behind both winters 2013/14 and 2014/15.
Figure 4: Winter Res/Com Demand Fcst, Source: PointLogic Energy
So what gives? In our 90-day forecast demand scenario, the duration of January res/com demand lags behind the winters of 2013/14 and 2014/15. February res/com demand is also behind these two dominant winters, while March res/com demand will come in above winter 2014/15, but just below winter 2013/14.
Layering in the Other Elements of Demand
To get a true sense of where winter 2017/18 will end up, we’ve created the graphic below. The “Winter 2017/18 90 Day Forecast” scenario and the “Winter 2017/18 Normal Forecast” are both inclusive of actual data for Nov. 1, 2017 through Jan. 9, 2018. We then layer on the duration of winter demand as projected using the 90-day forecast and the “normal” winter.
Figure 5: Total Demand by Sector, Source: PointLogic Energy
The chart yields some interesting results. Despite res/com demand and domestic demand setting all-time records over the first third of January 2018, over the near term the real lift to total demand comes from the power sector and the total exports category. Total demand in our 90-day forecast scenario is a full 5.2 Bcf/d above winter 2013/14 and 4.9 Bcf/d over winter 2014/15.
These new developments in the Lower 48 U.S. demand profile provides support to our initial conclusions drawn at the beginning of the winter – regardless of weather, export demand will lift this winter over all prior winters from a demand perspective. The topic of the acceleration of power demand this winter will be explored in a future article, but is directly related to price and the availability of natural gas as both a peaking and baseload fuel source.
Impact on End-of-Season Storage Levels
In the graphic below we show winter-on-winter deltas for each element of supply and demand. In the “Normal” demand scenario, we calculate the total demand delta from winter 2016/17 of 7.85 Bcf/d. This delta is only the difference between the two seasons in power, industrial and res/com demand. The same delta in the StatWeather Forecast scenario is 9.05 Bcf/d. Adding the impact of LNG exports and exports to Mexico increases each delta by 1.6 Bcf/d.
Figure 6: Winter 16/17 vs. 17/18 Deltas, Source: PointLogic Energy
In the Normal Demand scenario, total supply is 3.6 Bcf/d short of covering normal forecast demand inclusive of our estimated deltas for exports to Mexico, LNG exports and pipe loss. Over 151 days of winter, this equates to about 545 Bcf of incremental gas withdrawn compared to last winter. Over the course of winter 2016/17, the U.S. Energy Information Administration (EIA) reported that 1,935 Bcf of gas was withdrawn. So our Normal Demand scenario for winter 2017/18 is for a drawdown of 2,480 Bcf.
In the StatWeather Forecast Demand scenario, the supply gap increases to 4.5 Bcf/d, which yields an incremental 679 Bcf of withdrawals compared to last winter. So the drawdown would be 2,614 Bcf.
As of Nov. 1, 2017, EIA estimated that the starting inventory point for winter 2017/18 was 3,784 Bcf. With the PointLogic Normal Demand Scenario of withdrawals, end-of-season storage inventories would be 1,304 Bcf. With the new StatWeather Forecast Demand scenario, inventories would be further plundered down to 1,170 Bcf.
Figure 7: End of Season Storage, Source: PointLogic Energy
While it does not seem likely that winter 2017/18 res/com demand will match Polar Vortex heights, thanks to year-on-year increases to other demand segments (LNG exports, exports to Mexico and power burn), winter 2017/18 total demand is expected to beat demand from the Polar Vortex winters of yesteryear.
If demand follows the path that is predicted by StatWeather’s weather forecasts, end-of-season inventories would end at 1,170 Bcf, or 616 Bcf below five-year average levels. If demand follows the path dictated by “normal” weather, then inventories will land at 1,304 Bcf, or 482 Bcf below five-year average levels.
These estimates are independent of probable supply side responses, including an increase in both associated and dry gas production or a surge in Canadian net imports. Over the first ten days of January, both Canadian and, more surprisingly, LNG imports (or “sendout”), responded heartily to increased demand brought about by the Polar Vortex part deux.
With both supply sources receding over the past few days, PointLogic does not expect this to have a season long impact. It’s also important to note that the prolonged cold weather prompted some production freeze-offs, which while significant, will not affect total dry production estimates over the course of winter 2017/18.
Stay tuned to PointLogic to follow the delicate balance between supply and demand this winter and to see how the 90-day forecast plays out against normal weather and its total impact on res/com and domestic demand daily averages for the season.