We are delighted to present our new North America Monthly, which provides a brief outlook for the North American gas market.
This withdrawal season we expect LNG exports out of the US to increase by 2.7 Mt y/y to 10.7 Mt, resulting in a by 1.4 bcf/d (40 mcm/d) y/y rise in feedgas demand. Both numbers have risks to the upside and downside that depend on project ramp-up schedules, particularly at Corpus Christi T1 and Sabine Pass T5. Growth in US feedgas demand and LNG exports will be more perceptible in the 2019 injection season, although ramp-up risks still apply. We forecast the US will export 17.7 Mt of LNG during that season, a hefty 5.4 Mt y/y hike, with feedgas in turn jumping by 2.3 bcf/d (65 mcm/d) y/y. That growth in US LNG feedgas demand will come beside increased demand for gas to feed pipeline exports to Mexico and growing industrial consumption. Coupled with low inventories and higher y/y prices for competing fuels in the power sector, we expect more severe Henry Hub cash price spikes in winter 2018-19. Early cold has already pushed the curve up, and a December cold event could see further gains. For exports, the cash market could see off-takers at Cove Point—where delivered prices during the heating season are typically higher than Henry Hub—sell back into the market on peak demand days rather than lift cargoes.
Production hitting new daily highs
The steep upward trajectory in US production gains has been tempered by recent maintenance, with m/m growth now set to be 0.6-0.7 bcf/d (17-20 mcm/d) in November, down from the summer’s growth rate of 1.5 bcf/d (42 mcm/d) m/m. Shoulder season maintenance has helped lead to the lower rate of growth. While new Appalachian infrastructure has been ramping up, it has yet to run at full capacity and still seems to be taking existing flows from other pipes. Still, production nationally hit new daily highs in the past few days of 85 bcf/d (2.4 bcm/d).
Although the start-up of several new pipelines means that takeaway capacity has now officially outstripped production in Appalachia, this has not significantly boosted production. Maintenance limited pipeline capacity through October, but production data for early November does not yet indicate Appalachia output is strengthening. We only expect significant incremental output once all facilities on the Nexus pipeline enter service, which is expected before year-end. We still anticipate Appalachia production will grow at a healthy 4 bcf/d (0.1 bcm/d) y/y this heating season, dipping toward 3 bcf/d (85 mcm/d) y/y in the 2019 injection season.
Over the 2018-19 heating season, even as the pace of sequential output gains moderates, we see US production growth at an incredible 8.5-9.0 bcf/d (0.24-0.25 bcm/d). Appalachia will lead the way, but we see Haynesville adding a notable 1.3 bcf/d (37 mcm/d) y/y. For injection season 2019, we expect growth in both regions to slow, but still be strong with associated production still playing a meaningful role in that growth.
Structural demand to lead y/y production gains
As ever, the winter wild card is weather, given how much total US demand can swing depending on actual realised GWHDDs. US stocks started the heating season nearly on par with that of the week ending 29 December 2017, a tight situation despite all of the incremental production. In addition, cold weather in November (on track to log a GWHDD count c.20% higher than normal) will drive strong storage withdrawals. Because of this, we expect more severe Henry Hub cash price spikes above the 2017-18 peak in winter 2018-19 on growing structural demand, low inventories and higher y/y prices for competing fuels in the power sector.
Our heating season balances assume 10-year normal temperatures, which leads to an essentially flat y/y reading in res-com demand, while we anticipate y/y growth of 1.3 bcf/d (37 mcm/d) from the power sector. Structural demand—feedgas for LNG exports, pipeline exports to Mexico and industrial gas use—will continue to grow, by 2.6 bcf/d (74 mcm/d) y/y. But there are risks to demand stemming from timeliness of both Mexican pipelines and the Elba Island LNG terminal.
Kinder Morgan noted in its recent earnings call that the start-up of the 10 ‘mini-trains’ at Elba Island have now been pushed back from Q4 18. The company now only expects to bring one train to commercial service in Q1 19, with the other nine coming online through the rest of 2019.
As for US exports to Mexico, these have lagged this year on pipeline delays and any further delays to domestic pipes slated to come online this heating season will reverberate through injection season 2019, making it significantly more difficult to raise current gas flow levels out of the Permian. Following October’s news of fresh delays to pipe starts, we expect Mexico’s net pipeline imports to be limited to near 5.1 bcf/d (0.14 bcm/d) this heating season.
Given pipeline delays and declining Pemex output, Mexico will continue to backfill demand with LNG imports, which we forecast will average 0.7 bcf/d (2.6 Mt) across the withdrawal season and 0.8 bcf/d (3 Mt) in the 2019 injection season (the latter is a 0.3 bcf/d (1.5 Mt) upward revision from last month). We currently expect cross-border flows in the injection season to grow to 5.3-5.4 bcf/d (0.15 bcm/d), but further pipe delays could cut this number and raise LNG imports.
Delay-related risk for US LNG exports is less acute over the winter than it is in summer, but project ramp-up will still affect the pace of growth in LNG feedgas demand. We anticipate feedgas will be 1.4 bcf/d (40 mcm/d) higher y/y this heating season, based on the ramp-up of Cheniere’s Corpus Christi T1 and Sabine Pass T5. There is some minor upside to this figure, and hence US LNG exports, depending on how quick the ramp-up is and how consistent feedgas is through the pre-commissioning and commissioning process, as initial feedgas flows can sometimes be choppy. LNG feedgas has already hit a daily high of 4.3 bcf/d (0.12 bcm/d) as both trains are ramping up. For injection season 2019, feedgas will ramp up even more perceptibly with Corpus Christi T2, Elba Island T1-5, Cameron T1 and Freeport T1 expected to start up, although timeliness risks and ramp schedules provide risks. Based on a standard four/five-month ramp, we expect 2019 injection season feedgas to grow by 2.3 bcf/d (65 mcm/d) y/y.
High US gas prices limit Cove Point exports
US cash prices have risen sharply and been volatile so far in November, reaching a four-year high for the month. We expect further price spikes on days of peak demand given that low inventories increase the risk of deliverability issues. With the Henry Hub forward curve also repricing, this has potential implications for US LNG export economics. A Henry Hub price of 4.3 $/mmbtu, for instance, would put the FOB LNG charge from the Gulf Coast at almost 5 $/mmbtu, and with transport at almost 4 $/mmbtu (using spot freight rates), that takes a delivered cost to NE Asia of around 9 $/mmbtu. With peak winter JKM contracts at 11 $/mmbtu, this should not discourage exports. The profitability of Cove Point exports, though, with Transco Zone 5 up at peak winter prices of 8.65 $/mmbtu, look out of the money as it would require delivered prices to Europe over 11 $/mmbtu and to Asia over 13 $/mmbtu—neither region has prices that high.
While winter pricing this year is still a weather play, the quicker-than-expected fall in inventories―our forecasts for the end of the heating season have already come down by 160 bcf (4.5 bcm) compared to a base case using 10-year normal weather―has already shifted the curve, assuming a similarly striking mild event does not follow in the ensuing weeks.
|Fig 1: Haynesville and Appalachia y/y, bcf/d||Fig 2: Total US LNG feedgas, bcf/d|
|Source: EIA, Energy Aspects||Source: Ventyx, Energy Aspects|