To say the least, the picture that emerges is sobering.
If there were any prior doubt, after this past winter heating season it should be crystal clear that the U.S. natural gas market is experiencing a severe structural deficit. As explained below, supplies of natural gas available to the U.S. market in 2003 are certain to fall at least 1.0 – 1.5 TCf below the Energy Information Agency’s (“EIA’s’) most recent forecast of expected U.S. demand for the year.
This is a staggering shortfall, with profound implications for energy companies and for the health of the U.S. economy. Yet, the potential for a deficit of this magnitude is not yet widely recognized. Nor are most end users or state regulators prepared for the profound dislocations that are likely to occur in both the natural gas and power markets over the next 90 to 120 days.
F urther, this growing imbalance between available supplies of natural gas and expected demand is not likely to be short-lived. Instead, it reflects the early stages of a long-term structural imbalance, in which supplies of natural gas available to the U.S. market are likely to consistently fall 10% or more below the levels achieved during the 1990’s, at the same time that the underlying rate of demand is likely to continue to increase every year -- at least at prices anywhere near current levels. These continuing increases in the amount of natural gas needed to supply the U.S. market are due primarily to increased demand in the power sector -- which is expected to increase by at least 2.5 – 3.0 TCf between now and the end of the decade. See earlier articles on The Coming Natural Gas Crisis and A Cautionary Tale.
Absent adequate supplies of natural gas to fuel the 200,000 MW + of new gas-fired capacity built over the past 4 years (at a cost of more than $ 100 billion), there is no readily apparent means to meet the incremental electricity needs of the U.S. economy over the next 5 – 7 years – raising serious question as to how the growth of the U.S. economy will be sustained during the remainder of this decade, while new, longer-term source of natural gas supply are being developed.
Urgent National Concern
As serious a threat as this long-term structural imbalance is likely to pose during the remainder of the decade, however, the energy industry and the economy face a particularly daunting challenge right now.
This challenge arises from the urgent need to inject unprecedented amounts of natural gas into underground storage during the remainder of this year’s Refill Season and the lack of adequate supplies to meet minimum storage targets required to protect public safety and reduce the risk of runaway natural gas prices this coming winter.
The need to inject record amounts of natural gas into storage could not come at a more difficult time – i.e., just as U.S. production is about to hit its lowest level in 16 years. During the remainder of this year, imports of natural gas from Canada also are likely to decline sharply. Exports of natural gas to Mexico are likely to rapidly increase.
To make matters worse, inventories of distillate also are near record low levels -- limiting the potential for fuel switching.
As this article is being written, only 24-25 weeks remain in the Refill Season (i.e., less than 175 days). The amount of working gas in underground storage remains only modestly above the all-time record lows reached in mid-April.
Despite the short time remaining, however, there does not yet appear to be any clear recognition of how steep a price increase is likely to be necessary in order to free-up the additional supplies of natural gas needed to rebuild storage or the extent of the potential run-up that may occur in the spot market price for electricity before the end of this summer.
Comparison to Injection Levels Required in Earlier Years
In the aftermath of this past winter’s record draw downs from storage, most Local Distribution Companies (LDC’s) east of the Rocky Mountains are keenly aware that they will need to begin injecting natural gas into underground storage earlier than has been necessary in the past and to inject significantly larger quantities into storage than in prior years.
The potential cumulative impact, however, of many LDC’s simultaneously stepping-up their purchases during a single, compressed time period in a chronically under-supplied market is not yet well understood.
This past winter, end-of-withdrawal season storage in the U.S. reached an all-time record low, with the amount of working gas in underground storage bottoming out at 623 BCf on April 11th. Just as significantly, underground storage in Canada was depleted even more severely than in the U.S., with the amount of natural gas in underground storage in Canada falling to a level almost 70% below end-of-season storage levels last year.
To restore the amount of natural gas in storage to year-ago levels, all-time record injections will be required in both the U.S. and Canada.
This should give rise to grave concern – even before taking into account factors that may cause the supply demand balance in the North American market to deteriorate further between now and the end of this year.
As explained in prior articles, relative to historical norms on a weather-adjusted basis the amount of natural gas in storage in the U.S. has been declining for fifteen consecutive months. Month after month, injections during the Refill Season have been falling significantly below the 5-year average and withdrawals during the winter heating season have been consistently exceeding the same benchmark. The size of the deviation from historical norms (both injections and withdrawals) generally has been increasing over time.
The U.S. market has never previously experienced a continuing decline in storage relative to historical norms of this magnitude. On an aggregate basis, the total decline in storage during this 15-month period, after adjusting for weather and for seasonal variations in demand, is more than 1.25 TCf.
This is a stunning decline. It only can be explained as a result of a severe mismatch between newly available supplies and seasonally adjusted demand. This mismatch appears to be widening over time.
For many months, the emergence of this deficit was masked by the combination of a severe manufacturing recession in 2001, the further adverse impact on the U.S. economy of the September 11th terrorist attacks and the extraordinarily mild temperatures that occurred during the ‘01/’02 winter heating season. This unique combination of events resulted in a record build-up in storage during 2001 and early 2002, with the amount of natural gas in storage peaking on a seasonally adjusted basis in February of 2002 at almost 700 BCf above the 5-year average. This existence of this surplus in turn allowed the U.S. natural gas market to operate in a deficit condition for more than a year without a sharp run-up in prices.
The storage build up that occurred in 2001 and early 2002, however, has now been entirely eliminated.
In its place, in a remarkably short time span, storage has been reduced to an all-time record low (viz., as of the week ended May 2nd, 821 BCf – i.e., 545 BCf (or 40%) below the 5-year average for this date). This rapid decline in storage has occurred even though, as recently as last October 24th of last year, the amount of natural gas in storage was still 189 BCf above the 5-year average – a drop of more than 725 BCf relative to historical norms in just the past 6 months.
There is no compelling reason to expect this 15-month pattern of continuing declines to suddenly reverse. Unless it does, however, even matching last year’s end-of-Refill Season storage level of 3,172 BCf may prove to be a nearly impossible task, at least without unprecedented increases in the spot market price for natural gas.
Ensuring Adequate Levels of Storage
If the challenge facing the industry weren’t already sufficiently daunting, however, one lesson that is – or at least should be -- crystal clear after last winter is that merely matching last year’s end-of-Refill Season storage level again this October will not be adequate to protect against severe price spikes or even to protect public safety going into next winter.
As discussed in detail later in this article, contrary to what many analysts assume, temperatures last winter were slightly milder than historical norms.
If the weather last winter had matched the last recent colder-than-normal winter (i.e., the winter of ‘00/’01), the total withdrawal from storage could easily have been another 500 -- 600 BCf greater than the withdrawal that actually occurred last winter – when the cash price in the day ahead market at Henry Hub reached all-time record high.
In all likelihood, if temperatures last winter had matched the ‘00/’01 winter heating season, the amount of working gas in storage would have been reduced to near-zero levels and prices would have spiked even higher than they did this past February and March.
If anything, therefore, last winter’s experience strongly suggests that, in a rapidly-changing market, in which residential and power sector demand is growing every year and supplies available to the U.S. market are continuing to decline rapidly, even filling storage to the brim – i.e., in the U.S., a little over 3,450 BCf – might not fully protect the public interest. Instead, there appears to be an urgent need, to expand existing storage capacity – perhaps by as much as 20 -- 25% (i.e., to at least 4,175 BCf).
As a practically matter, given the short time that remains between now and the end of the Refill Season in mid-to-late October, it is not physically possible to expand storage quickly enough to increase the amount of natural gas that can be stored this winter.
At a minimum, however, given the huge withdrawals from storage that occurred this past winter (in a winter that was slightly milder than historical norms), prudent planning requires that the industry attempt to fill existing storage as close as possible to maximum capacity in both the U.S. in Canada before the end of the Refill Season in mid-to-late October.
This in turn requires LDC’s and their suppliers to inject an additional 2.65 TCf into storage in the U.S. and more than 350 BCf in Canada in a span of no more than 24-25 weeks.
This rate of injections has never previously been achieved in the North American market in any prior Refill Season.
These record injections are necessary in order to: (i) protect against the potential for a colder-than-normal winter; (ii) provide adequate reserves to protect against continued declines in production, pipeline failures, potential outages at nuclear plants and other contingencies; and (iii) ensure that, in the words of one LDC, that “no one’s grandmother freezes to death in late March or early April” if a cold snap hits late in the winter (as occurred in early April of this year in the midwest).
The same underlying structural supply deficit, however, that caused withdrawals from storage to far exceed historical norms this winter make this an impossible goal, at least at prices anywhere near current levels (which, in recent weeks have ranged between $ 5.25 and $ 5.75/MMBTU in the day ahead cash market at Henry Hub in the Producing Region in Louisiana).
Indeed, if anything, at current price levels, it is not clear that injections into storage will even match last year’s anemic levels – in which a meager 1,500 BCf was injected into storage during the period between May 2nd and the first withdrawal of the season in the last week of October.
This is particularly true as a result of the likely impact on natural gas supplies of the NOx trading cap that went into effect in the Northeast beginning on May 1st .
Over the course of the summer, this cap is likely to significantly increase use of natural gas in the generation sector -- and therefore to further reduce the amounts of natural gas available to inject into storage.
Under the new cap, utilities in the Northeast will be required to reduce power plant emissions of NOx by approximately 1/3rd during the period between May 1st and September 30th compared to emissions during the same 5-month period last year. (NOx is a precursor of urban smog, which is believed to cause asthma and other serious health problems.)
In order to comply with the cap, as the summer progresses, many coal-burning utilities in New York, along the Mid-Atlantic seaboard and in Pennsylvania and Virginia are likely to find it necessary to ratchet back significantly on their use of coal compared to last summer and to substitute instead what last year would have been out-of-merit order dispatch of gas-fired generating units. (Gas-fired units emit far lower amounts of NOx than units that burn coal or oil and therefore consume a far smaller number of credits for each megawatt hour of electricity produced.)
Depending in part upon the severity of summer temperatures and the availability factor for nuclear units in the region, this could result in a major increase in the use of natural gas compared to last year. These increases are not yet fully reflected in most forecasts of power-industry consumption of natural gas during June, July, August and September of this year.
This potential increase in natural gas consumption as a fuel to generate electricity in turn could make it even more difficult to match last year’s puny injection levels and create significant further upward pressure on the price for natural gas beginning as early as this coming June or July.
Potential National Crisis Looming
Even if the industry still were able to match last year’s anemic injection levels, however, repeating last year’s injections again this year still would leave the total amount of natural gas in storage as of late October at no more than 2,321 BCf (i.e., storage as of May 2nd of 821 BCf + an injection of 1,500 BCf = total late October storage of 2,321 BCf).
This is more than 1.0 TCf below our estimate (discussed in detail below) of the minimum end-of-Refill Season storage level required to protect public safety this coming winter.
This is an alarming deficit, which should give rise to grave concern on the part of the industry and policy-makers at the federal and state level.
Neither the LDC’s (who take their responsibility to protect public safety very seriously) nor state regulators nor state governors can – or will -- allow end of Refill Season storage to remain at such precariously low levels.
Many state regulators and power companies do not yet appear to fully appreciate the gravity of the crisis the current storage deficit is likely to create over the next several months.
As the Refill Season progresses, however, we believe it will become increasingly apparent that absent steep further increases in the price for natural gas to free-up much larger supplies of natural gas for injection into storage this Spring and early Summer (when the largest injections into storage typically occur) the amount of natural gas injected into storage during the Refill Season is likely to fall far short of the minimum levels required to protect public safety, in both the U.S. and Canada.
Indeed, given the short time remaining between now and the end of the Refill Season in mid to late October, even with aggressive actions to replenish storage, the combined storage deficit in the U.S. and Canada as of the end of the Refill Season still is likely to be in the range of at least 500 – 750 BCf.
This is a deficit of staggering proportions that raises public safety issues of great national urgency in both the U.S. and Canada.
As the severity of the crisis we face becomes more apparent, LDC’s and their suppliers on both sides of the U.S./Canadian border are likely to come under the increasing pressure to enter the spot market in order to buy up larger and larger supplies of natural gas for injection into storage in an increasingly constrained market.
The specific timing of when the next run-up in natural gas prices will occur will depend on how quickly the LDC’s and their suppliers begin to enter the spot market aggressively to step up purchases of natural gas.
Many LDC’s have not yet begun their refill purchases, at least at a significant enough level to materially impact the market. This is because to accelerate the timing of their purchases many LDC’s must obtain approval from their state regulatory commissions to implement revised storage refill plans. Many LDC’s already are seeking these approvals, but final orders have often not yet been issued, delaying the start date for refill injections.
Ironically, however, the longer it takes before the LDC’s begin taking aggressive action to refill storage, the more intense the upward pressure is likely to be on price of natural gas once such purchases begin, since the less time will remain to overcome a deficit that already is at record levels.
Once these purchases begin to accelerate, upward pressure on the spot market price for natural gas is inevitable.
By the end of the summer, we expect natural gas prices to return to the $ 8.00 – 10.00/MMBTU range experienced this past winter – an unprecedented level for summer months. By the fall, prices well above $ 10.00/MMBTU may be inevitable. As the severity of the storage deficit becomes increasingly apparent, the potential for even higher prices cannot be ruled out.
Further, as the run-up in natural gas prices begins to occur, it could have an even more severe impact on power prices, especially as we enter into the heart of the summer in mid-July and early August, when demand for electricity typically is at its peak.
During the peak weeks of the summer, if natural gas prices are in the $ 8.00 – 10.00/MMBTU range (as we believe is likely to occur), the impact on the spot market price of electricity could be dramatic in any region of the country in which gas-fired capacity is the marginal source of supply (i.e., at this juncture, most of the U.S.).
Further, even with natural gas prices at these levels, we are likely to head into the winter heating season with amounts of natural gas in underground storage at precariously low levels – i.e., certainly well below last year’s 3,172 BCf, and quite possibly below the 2,549 BCf peak-to-trough withdrawal that occurred during this past winter’s (milder than normal) withdrawal season.
Thus, unless temperatures next winter prove to be extraordinarily mild (which, as discussed below, is relatively unlikely), the price spikes next winter could make the run-ups that occurred this past February look tame.
As 2003 progresses, we believe that the impact of tighter natural gas supplies on both the natural gas and power markets is likely to become the dominant issue facing the energy industry – and, to a significant degree, the economy as a whole.
The long term impact on the E&P and power industries may well depend to no small degree upon whether the energy industry is able to explain more convincingly than in the past why these price increases are occurring, what price levels are reasonable to expect longer term and what steps, if any, can be responsibly taken to minimize further price increases, to the extent it is realistic to limit the extent of further price run-ups, both for electricity and for natural gas.
Lessons Learned from This Past Winter
The national media focused considerable attention this past winter on the record prices for natural gas reached during the second half of the winter.
This focus on “sticker shock” is hardly surprising – and not inappropriate. For the second time in the past three winters, natural gas prices rose to levels far higher than most analysts predicted before the winter heating season began.
In late February and early March, spot market prices in the day ahead market at Henry Hub reached an all-time high – with intra-day prices on the Intercontinental Energy Exchange reportedly peaking at levels above $ 27.00/MMBTU on February 25th, and closing the day at $ 18.85 (the highest close ever). The delivered price at City-gate locations in Chicago, New York, Boston and other major delivery points in the eastern U.S. frequently traded well above $ 20.00/MMBTU on a number of separate occasions in February and early March, and in some instances at some locations traded well above $ 30.00/MMBTU.
Further, the near-month contract on the New York Mercantile Exchange (NYMEX) briefly shot above $ 11.00/MMBTU (another all-time high) and for several weeks traded in the $ 8.00 – 9.00/MMBTU range.
But for a much earlier-than normal arrival of Spring-like conditions in much of the U.S. at the end of the second week of March that persisted for most of the remainder of the month, prices might have risen to even higher levels later in March, since continued cold weather could have driven storage in the East Consuming Region and the Producing Region to near zero levels.
Record Decline in Storage
These soaring prices, however – as painful as they may have been to end use customers – at best are only half of the story.
Instead, at least as significant is the astonishing decline that occurred in the amount of natural gas in underground storage over the winter months and what it indicates about the magnitude of the structural deficit that exists in the U.S. market.
From peak to trough, the total withdrawal from storage during this year’s withdrawal season was a remarkable 2,549 BCf (i.e., specifically, from 3,172 BCf as of the week ended 10/24/02 to 623 BCf as 4/11/03).
This is by far (i.e., specifically, by 250 BCf) the largest withdrawal ever recorded since EIA began tracking withdrawals in the early 1970’s (i.e., a period of almost 30 years):
All-Time Record Withdrawal From Storage
*If climatologically normal weather between November 1st and March 31st
** If temperatures identical to ’00/’01 winter heating season
The amount of natural gas in underground storage deteriorated at a particularly rapid rate during the first 10 weeks of 2003 – with a net decline in storage of 1,695 BCf during the period between January 3rd and March 14th.
This 10-week withdrawal was almost as large as the withdrawals that typically during the course of an entire winter heating season. This past winter, however, 1,695 was withdrawn in a period of just 70 days (i.e., less than ½ of the winter heating season). This is a sure sign that the supply/demand balance was badly out of whack (since temperatures in the aggregate during this 10-week period were close to historical norms).
Magnitude of ‘02/’03 Withdrawals Not Attributable to Unusual Weather
Many analysts have attributed the huge withdrawals that occurred this past winter to colder than normal weather.
Any notion that this past winter was colder than historical norms, however, is plainly false -- as can be readily confirmed from publicly available data. (See, in particular, the data on Heating Degree Days available on the web site for the National Weather Service’s Climate Prediction Center, which can be found at www.cpc.ncep.noaa.gov.)
This past winter, unlike a number of winters in the late 1990’s, there were several very cold weeks, especially in the northeast. Further, most of these weeks occurred relatively late in the winter heating season (i.e., specifically, in January and February). As a result, the end-of-season cold spell tends to stand out in many recaps of last winter’s events.
It is important to keep in mind however, that while certain weeks this past January and February were particularly cold, January and February typically are the coldest months of the year. As compared to historical norms, measured in terms of gas-weighted Heating Degree Days ((HDD’s), even the coldest week this past winter (i.e., the week ending January 24th), was only 23 HDD’s (or a little over 10%) above climatologically normal weather for the week (i.e., specifically, 244 gas-weighted HDD’s vs. a norm for the same week of 221 HDD’s).
While this severe cold undoubtedly added to gas consumption during the week ending January 24th (and therefore to the told withdrawal from storage during that week), even in the most extreme week of the winter the increase in consumption attributable to the weather in all likelihood was less than 30 BCf.
Further, while there were several bitter cold weeks in January and February of this year, there also were a number of reasonably mild weeks even in January and February which partially counter-balanced the effect of the colder-than-normal weeks.
During the remainder of the winter heating season, temperatures on average were reasonably mild.
For a stretch of nearly five weeks, for example, from December 10th through January 10th, temperatures in much of the country were unseasonably warm (i.e., more like early to mid November than late December or early January, which in some recent years have been the coldest part of the winter). During this period, the total number of gas-weighted HDD’s fell 166 HDD’s below the historical norm for these same weeks – exceeding by more than 30 HDD’s the total number of excess HDD’s for the 5 weeks this winter that most exceeded historical norms.
This streak of milder-than-normal weather immediately before the coldest weather hit in mid-January significantly reduced natural gas consumption for several weeks and allowed the amount of natural gas in storage to regain gain ground relative to the 5-year average in December and early January despite the overall undersupply condition in the market.
If this warm spell had not occurred, storage in the East Consuming Region and the Producing Region might have come perilously close to being drawn down to zero before the end of the winter heating season and problems in maintain pressures in the pipeline system could have become widespread and severe.
Beginning around March 14th, the weather again turned suddenly Spring-like over much of the country (i.e., more like late-April or early-May) and remained that way, with only brief interludes of colder weather, for most of the remainder of the month.
For the winter heating season as a whole, the data couldn’t be clearer: as measured in terms of gas-weighted Heating Degree Days (the most objective measure available), the weather was 3% milder than historical norms. Only two months (i.e., November and February) were colder than historical norms. Even in those months the exceedances were surprisingly small:
‘02/’03 Winter Heating Season vs. Historical Norm
(Gas-weighted Heating Degree Days)
Further, while temperatures for the winter season as a whole were colder than historical norms in some regions (including, in particular, New York, the Mid-Atlantic Region and New England, where much of the national media live), temperatures in the Midwest were milder than historical norms. This is important, since the Midwest has a large population and very cold winters, with the highest penetration rate for natural gas heating of any region in the country.
As a result, it accounts for more than 40% of heating-related load in winter months.
The milder than normal temperatures in the Midwest more than offset the impact of colder-than-normal temperatures in the east, where the Atlantic Ocean tends to moderate temperatures and the penetration rate for natural gas heating is significantly lower.
Unmistakable Evidence of Large Structural Deficit
Quite clearly, therefore, the huge withdrawals from storage that occurred this past winter can not be explained based upon colder-than-normal weather. To the contrary, total demand for natural gas would have been significantly larger if temperatures had more closely approximated historical norms.
If weather-induced consumption is not the explanation, however, the conclusion becomes inescapable: the huge withdrawals from storage this past winter could only have occurred as a result of newly available supplies of natural gas falling massively short of weather-normalized demand.
We estimate that, during the period from November 1st through March 31st, total withdrawals from storage were approximately 843 BCf greater than should have been expected after fully normalizing for weather:
Excess Withdrawals Not Related to Weather
This should be seen as an alarming figure – raises issues of urgent national concern.
It reflects a huge shortfall in supply -- far larger than the largest supply deficit that has occurred in any previous winter in the U.S.
By way of comparison, it is greater than the total shortfall in supply that occurred in all of the year 2000 (i.e., a “current account” deficit of 805 BCf).
The shortfall in supply that occurred in 2000 was sufficient to cause natural gas prices to quadruple in the last 8 months of that year. It also was a major precipitating cause of the meltdown in California – which resulted in $ 14 billion in unanticipated power supply costs in California alone, forced the largest utility in the state to file for bankruptcy and nearly bankrupted the entire state.
It hardly should be surprising, therefore, that as a direct result of an even larger “current account” deficit this past winter, natural gas prices again set an all-time record.
Further, this supply deficit did not emerge suddenly in early November. Instead, as has been discussed in earlier articles, it has been building for more than a year, as reflected in a steady decline in the amount of natural gas in underground storage that has been occurring since early February of 2002.
Even before the winter heating season began, storage already had fallen by more than 500 BCf relative to the 5-year average:
Once this winter heating season began, storage then dropped like a rock – both in absolute terms and relative to the 5-year average:
In total, during the 15-month period between February 1, 2002 and March 31, 2003, supplies of natural gas delivered to the U.S. market have fallen a whopping 1.25 TCf below actual consumption during this period.
The U.S. has never previously experienced a mismatch between supply and demand of this magnitude, persisting for such a sustained period.
As noted earlier, the amount of natural gas in underground storage now stands at 545 BCf below the 5-year average – with little or no prospect for restoring storage to more normal levels any time this year.
The exposure of the U.S. market to a further severe run-up in natural gas prices, therefore, is clearly far greater today than it was at the end of April last year -- when the amount of natural gas in storage was more than twice current levels (i.e., as of May 2nd, 1,645 BCf, 824 BCf above the 5-year average).
Agenda for Secretary Abrams and Other Policy-Makers
Given these circumstances, at least three issues need to be addressed urgently at the national level:
1. Given the experience this past winter, what level of storage is needed to protect public safety and protect against the possibility of extreme price run-ups this coming winter?
2. Is it feasible to inject sufficient amounts of natural gas into storage in the short time period that remains between now and the end of the Refill Season in mid-to-late October to satisfy these targets? If so, what actions need to be taken to ensure that the required level of injections occurs?
3. If it is not feasible to satisfy these targets by mid-to-late October (either because there is not sufficient storage capacity or because supplies are insufficient to achieve the required level of injections or both), what additional steps should be taken to protect public safety and to reduce vulnerability to severe price spikes this coming winter?
Given the urgency of the issues at stake, we urge the Secretary of Energy to immediately convene a senior level task force or blue ribbon panel to comprehensively address all three issues on an urgent, priority basis.
We also urge the major Local Distribution Companies, in concert with the National Association of Regulatory Commissioners (NARUC) to meet on an emergency basis to develop comprehensive plans for addressing these issues.
In the hopes of stimulating further discussion, the remainder of this article will present some initial thoughts regarding the first and second issue.
A subsequent article will address potential solutions, both short and longer-term.
Establishing a New Storage Target
As with any planning decision, there is no, single objective “right” answer to the question of “how much end-of-Refill Season storage is enough?”
During the last half of the 1990’s, the general belief within the industry was that end-of-Refill Season storage would be adequate if the total amount in storage was in the range of 2,800 – 3,200 BCf.
This amount, in turn, consisted of two components:
Based upon the experience this past winter, however, it should be obvious that the question of “how much storage is adequate” needs to be entirely rethought.
Further, this reexamination needs to occur immediately.
Only 24 to 25 weeks remain in the Refill Season. Further, the largest injections occur near the beginning of the season (i.e., between late April and the end of June).
After then, the injection rate tends to taper off fairly rapidly – first due to peak summer demand for electricity and then due to the early stages of the heating season. While injections still occur in late September and early October, they typically are very small. (The last 4 injections last year, for example, averaged only 40 BCf per week.)
The time remaining to refill storage in anticipation of the next withdrawal season, therefore, is dwindling far more rapidly than many observers realize. With each passing week, it becomes increasingly difficult – potentially to the point of becoming physically impossible – to compensate for significant deficits in prior weeks.
Size of the Potential Withdrawal
As in prior years, the first step in the process of establishing a reasonable end-of-season target is to estimate the likely maximum size of the withdrawal for the next winter.
During the second half of the 1990’s, the total withdrawal from underground storage during the ‘95/’96 winter heating season provided a fairly good proxy for the likely maximum withdrawal in future years.
This is because the winter of ‘95/’96 was the coldest winter in many years, with a total withdrawal of 2,300 BCf (i.e., 250 BCf less than the total withdrawal this past winter).
While the winter of ‘95/’96 didn’t set an all-time record for the highest number of Heating Degree Days (HDD’s), it came close.
Further, during 1995-1999 timeframe, total demand for natural gas varied little from year to year. Total supplies also were nearly the same every year.
Given this stable pattern, it was perfectly reasonable to assume that the maximum withdrawal in any winter heating season was not likely to be materially larger than the withdrawal that occurred during the ‘95/’96 season (i.e., the coldest year of the decade).
The experience this past winter, however, clearly blows the ‘95/’96 standard out of the water (even though some analysts still appear to be assuming that there is no need to adjust this year’s refill target to reflect this past winter’s experience).
For openers, as noted earlier, the actual withdrawal during the last withdrawal season, peak to trough, was 2,549 BCf -- i.e., 250 BCf greater than the previous record withdrawal of 2,300 BCf.
Even the most superficial analysis, therefore, suggests that the end-of-season target of 2,800 – 3,200 BCf (including a working reserve) should be increased by at least this amount, to a minimum of 3,050 BCf to 3,350 BCf – i.e., 2,550 BCf actual peak-to-trough withdrawal + 500 – 800 BCf working reserve = minimum acceptable reserve of 3,050 – 3,350 BCf.
Further, even this target almost certainly is too conservative.
Storage last fall started at almost precisely the mid-point of the range just suggested (i.e., specifically, 3,172 BCf as of the week ended October 24, 2002).
Nonetheless, even though winter temperatures this past winter were slightly milder than historical norms, prices spiked to all-time record levels before the end of February. In addition, serious operating problems were experienced in much of the eastern U.S. in late February and the first two weeks of March.
These problems in all likelihood would have become significantly worse if the weather during the second half of March had not suddenly turned unusually mild and remained that way across most of the U.S throughout the remainder of March.
At a bare minimum, therefore, prudence strongly suggests that, going into the current Refill Season, the end-of-Refill Season target ought to be set at the upper end of the range suggested above (i.e., 3,350 BCf). Anything less would leave the U.S. market vulnerable to price spikes and/or operating difficulties at least as severe as those experienced this past winter even if temperatures this coming winter were identical to last winter (i.e., milder than the historical norm).
Even this adjustment, however, understates significantly the size of the withdrawal that could occur next winter – and therefore the appropriate end-of-season target for this year’s Refill Season.
In particular, in estimating the size of the potential withdrawal this coming winter (i.e., the first step in the process, before adding a working reserve), a prudent planner undoubtedly would want to take into account at least three other factors:
1. Likely increase in consumption if weather matched historical norms.
As noted earlier, despite the widespread believe to the contrary, temperatures this past winter, as measured in terms of gas-weighted Heating Degree Days, in fact were milder than historical norms.
Further, this deviation from historical norms was not simply a matter of normal fluctuations from historical norms (although normal variations could easily have produced variations of the magnitude that occurred).
Instead, at least a portion of the variance that occurred this past winter is directly attributable to the El Nino-like conditions that existed in the Pacific from the Spring of last year through at least the end of January.
These conditions did not prove to be nearly as powerful as the National Weather Service predicted throughout much of the winter. See, for example, the Weather Service’s 90-day outlook issued January 9, 2003, which predicted much warmer than normal winter temperatures, particularly in the Midwest and the Northeast, for most of January, February and March of this year. (Even though the Weather Service’s January 9th forecast proved to be far off the mark, it had a significant impact in limiting increases in the price of natural gas, both in the day ahead spot market and in the futures market, in January and early February, setting the stage for particularly steep increases later in February.)
While the Weather Service’s forecast of reasonably strong El Nino affects this past winter proved to be dead wrong, there were mild El Nino-like conditions earlier in the winter (i.e., particularly in November, December and early January). These conditions moderated temperatures during the first half of the winter, but would not be expected to reoccur in a more normal year.
The El Nino-like conditions that existed in the Pacific this past winter, however, have now largely dissipated; unlike conditions a year ago, surface water temperatures in the Pacific conditions are now back within normal ranges, and most forecasters are not currently predicting El Nino conditions for next winter.
Thus, while it is certainly possible that temperatures next winter will be as mild as this past winter, at this point in the year, in establishing storage targets for this coming winter, there is no defensible basis for assuming that milder-than-normal conditions are likely to reoccur this coming winter.
Instead, at a bare minimum, in assessing the potential size of the withdrawal from storage next winter, any prudent planner would undoubtedly adjust the size of any projected withdrawal to take into account what the magnitude of the withdrawal would have been last winter if temperatures had matched historical norms.
Even this minimal adjustment, however, increases the size of the potential withdrawal next winter by at least 200 - 225 BCf, from 2,549 BCf to 2,750 – 2,800 BCf.
The effect of this adjustment is to increase the minimum storage target from the 3,050 – 3, 350 BCf range suggested previously to 3,250 – 3,575 BCf (i.e., a potential withdrawal of 2,750 – 2,800 BCf + a working reserve of 500 – 800 BCf = a minimum target of 3,250 – 3,575 BCf).
This in turn immediately creates a problem. As discussed previously, given the price spikes and operating difficulties experienced this past winter, the appropriate target probably should be set at or near the upper end of this range (i.e., 3,575 BCf). The maximum current storage capacity in the U.S., however, currently is only 3,450 BCf (i.e., 125 BCf below the target range).
Just to prepare for the possibility of a statistically normal winter, therefore, would require end-of-Refill Season storage in excess of current total U.S. storage capacity.
2. Allowance to take into account potential for colder-than-normal weather.
Even this adjusted target, however, does not make any allowance for the potential that next winter could prove to be colder than normal – even though statistically there is a 50/50 probability that temperatures this coming winter will be colder than historical norms.
As with any planning decision, there is no clear objective standard as to how cold a winter a prudent planner ought to assume (i.e., 1 year in 5, 1 year in 10 etc.). The issue, however, obviously needs to be taken very seriously; the safety of the public depends upon LDC’s and their suppliers injecting sufficient natural gas into storage during the Refill Season to ensure that, no matter how cold the weather might be this coming winter, adequate natural gas will remain in storage near the end of the season to meet whatever heating demand may still arise if a string of colder-than-normal days should occur in the second half of March or April.
At a minimum, therefore, presumably the amount of natural gas in storage at the end of the current Refill Season should be sufficient to meet the withdrawals that would be likely to occur if the weather this coming winter were to be an exact repeat of the last colder-than-normal winter – i.e., in this instance, the ‘00/’01 winter heating season (just three years ago).
If temperatures this past winter, however, had been identical to temperatures three winters ago, when the number of gas-weighted Heating Degree Days between November 1st and March 31st was 426 gas-weighted HDD’s higher than this year (i.e., 4,156 HDD’s in ‘00/’01 vs.3,730 HDD’s in ‘02/’03), the total withdrawal this past winter in all likelihood would have been at least 500 -- 550 BCf higher than the 2,550 BCf peak-to-trough withdrawal that actually occurred this past winter. This in turn would have resulted in a total withdrawal of approximately 3,050 BCf.
Adding a 500 – 800 BCf working reserve to this total would require a total amount of natural gas in storage as of the end of the Refill Season of 3,550 BCf to 3,850 BCf – i.e., 100 – 400 BCf greater than total current storage capacity in the U.S.
3. Adjustments for expected increases in demand and/or reductions in supply.
Finally, even this estimate of the potential withdrawal for next winter does not yet taken into account the impact of known or expected increases in demand or decreases in supply.
At this juncture, a number of factors already can be identified that are likely to cause the supply/demand balance to deteriorate compared to last winter. These include:
- likelihood of continued declines in production in both the U.S. and western Canada (in part because drilling rates in both countries remain well below the level necessary to offset continued declines in production from existing wells);
- potential need to retain substantially more natural gas in Canada next winter in order to maintain storage in Canada at adequate levels (which in turn could reduce the amount of natural gas available to export to the U.S. by as much as 1.5 BCf/day);
- Likely increases in the amount of natural gas exported to Mexico;
- Likely increases in consumption of natural gas for residential heating due to the continued high rate of construction of new homes and the high penetration rate being achieved byr gas heating; and
- Potential further increases in the use of natural gas to generate electricity
On a net, all-in basis, however, we expect the overall supply/demand balance this coming winter to deteriorate by at least another 1.5– 2.0 BCf/day compared to last winter. This in turn could increase the size of the withdrawal over the ‘03/’04 winter heating season by another 225 -- 300 BCf compared to last year.