The European gas network has been stressed on rare occasions; the most recent events include, on the demand side, the weekly increase of 17 percent observed at the end of February 2018 linked to exceptionally cold temperatures and, on the supply side, the interruption of gas flows in December 2017 linked to the explosion at the Baumgarten node.
Given the low frequency of gas alert occurrences, these two events represent a valuable opportunity to analyze the gas network system response to discontinuities from multiple strategic angles, ranging from the security of supply to market behavior.
We collected and analyzed data relevant to the 2 events using McKinsey Energy Insights EU PipeFlow that allows the processing of detailed information of over 900 points in the European Union, providing hourly data on a country-level overview of all entry/exit points, such as:
- Pipeline imports/exports (transits)
- LNG imports, storage, and break-bulk
- Domestic production
- Storage withdrawals/injections
- Industrial and residential consumption
Demand side case: network response to flow interruption
A high-performing and interconnected energy market is crucial to maintain the security of the energy supply. On December 12th, 2017, an explosion occurred at Austria’s main gas pipeline hub in Baumgarten, resulting in a gas flow interruption to Italy and prompting Italy to declare a state of emergency (see Exhibit 1-2).
Observations highlight several counter-measures that were put in place (see Exhibit 3):
Demand side response: A downsize in demand, possibly linked to contractual flexibility in cases of force majeure, helped to mitigate ~21 percent of the gas flow gap challenge in the immediate aftermath of the event. Demand measured in Italy moved from 318 mcm on December 11th (the day before) to 286 mcm on December 13th (the day after).
Storage network response: Despite the accident occurring during the winter season with storage already at a withdrawal rate close to capacity booking, the physical peak capacity of the storage network allowed the immediate mitigation of 60 percent of the gas flow gap, confirming the high suitability of storage as a flexible source of supply as well as the value of having spare capacity for the system. For this specific case, extra volume was mainly provided by SNAM-managed storage facilities, increasing withdrawal rates from 66 mcm on December 11th to 107 mcm on December 12th (see Exhibit 5).
Pipeline network response: Overall pipeline import increases catered for 17 percent of the gas flow gap, mostly driven by gas uptake in the Mazara entry point, which accounted for nearly 90 percent of the total pipeline import increase.
In response to the event, Italian gas prices reflected the gas supply gap generated by the explosion, moving from 22-25 EUR/MWh over 8-11th December to 50 EUR/MWh on December 12th and peaking as high as 55 EUR/MWh on the same day (see Exhibit 7). However, in less than 24 hours after the explosion, gas prices returned to below 30 EUR/MWh as the aforementioned set of measures promptly restored market equilibrium.
For the sake of exhaustiveness, it is important to consider that additional counter-measures exist and might have been deployed; for instance, the compressibility of natural gas allows the use of a line pack to compensate for fluctuations in gas demand.
Powered by Energy Insights
In conclusion, data analyses on this event show how the gas network system and market behavior promptly allowed the handling of the situation. A key success factor appears to be the availability of multiple levers to pull in response and the system capacity size. As stated by network operator SNAM, spare storage capacity was the single fundamental driver, yet data shows that the other counter-measures were also relevant and allowed the situation to quickly be restored.
Supply side case: network response to peak in demand
Between Feb 23rd and March 3rd 2018, the European gas market was significantly impacted by a sudden but sustained temperature drop driven by the Buran, a powerful freezing wind that swept across North and Central Asia, getting as far as Italy –an event that occurs with a 5-7 year frequency. This condition lowered the average daily temperature in Europe from +0-2 °C to roughly -5 °C, spurring gas demand in Europe by more than +20 percent (i.e., 482 mcm/day, from 2,174 mcm/day on February 23rd to 2,666 mcm/day on February 28th – see Exhibit 8-9).
During the occurrence of the Buran event, it is interesting to observe how gas flows reacted to the sudden gas consumption increase (see Exhibit 10).
First, not all available supply sources responded with increased flow to fill the demand rise. In fact, domestic production and pipeline imports from Norway saw their output lowered on March 1st, immediately after the peak in demand on February 28th, further inflating the need for gas supply. According to press news, the flow reduction was driven by operational challenges due to the unusual weather at the Kollsnes plant.
Second, transmission losses and volumes destined for non-EU countries (e.g., Ukraine) brought the total gas demand increase over the period to 632 mcm/day.
On the supply side, underground gas storage and LNG storage withdrawals combined filled ~97 percent of the total demand increase, increasing by 439 mcm/day (69 percent of total demand) and 174 mcm/day (28 percent), respectively. The remaining 3 percent was shared by a modest increase in pipeline imports from N. Africa and Russia (whose volumes were already flowing at a high rate on February 23rd).
Across these various supply flows, it is interesting to shed light on the impact of flow dynamics on gas prices at the hubs. We leveraged McKinsey Energy Insights EU PipeFlow to analyze the daily distribution from the vast Norwegian upstream basin in combination with Italian and TTF spot prices in Italy and the Netherlands, respectively: with some degree of approximation, it is possible to decompose the price effect into 2 distinct phases (see Exhibit 11).
Phase I—Temperature effect: Between February 23rd and February 27th, when Norway’s supply remained relatively constant between 330-340 mcm/day, both the maximum and the average daily spot price increased from ~23-24 EUR/MWh on February 24th-25th to ~40-50 EUR/MWh on February 26th-27th. We assume most of this price effect during phase I is likely linked to increasingly lower temperatures and, in turn, higher demand.
Phase II—Supply flow effect: Between February 28th and March 3rd, Norway’s supply first went down to ~290 mcm/day due to the Kollsnes plant output reduction to then rise back to the normal level of 330 mcm/day after March 3rd. Cumulatively, Norway delivered ~115 mcm/day less over the period vs the average output of 330 mcm/day previously held.
Over the same period, the cold temperature effect actually reduced and went back to normal and other major flows were stable. Maximum traded prices further increased, e.g., in Italy to ~100 EUR/MWh on March 1st. Given these dynamics, we assume most of this price effect during phase II is likely linked to the unexpected flow decline from Norway.
In conclusion, also in this second event, which was actually a double event (temperature and supply discontinuities), the data analyses show how the European gas network system and market behavior promptly allowed the handling of the situation. A key success factor appears to be the availability of storage, both the traditional underground storage and that at LNG terminals.
At the same time, there is clearly a cost connected with these system-adverse events, as prices rapidly reacted upwards. Clearly many different market drivers (such as contractual arrangements between players, capacity constraints, or maintenance schedules, etc.) might have had an impact on the supply and demand flow behaviors experienced over the period, potentially playing a highly relevant role. However we wanted to give a high-level quantification of the cost for the system versus normal operations (see Exhibit 13).
By focusing on Norwegian pipeline imports specifically and by valuing the full volume traded at the average daily TTF price, the cost to the system for importing gas from Norway rose from an average EUR 78 million/day over February 20th -23rd to an average EUR 99 million/day on February 26-27th. Following the outage in Kollsnes, the total average daily cost to the system increased by a further 113 percent reaching an average EUR 211 million/day between February 28th and March 2nd.
To summarize, the two cases considered in this article offer clear evidence of the sheer magnitude of the intrinsic value of infrastructure capacity (e.g., storage, LNG terminals, multiple pipe gas access, etc.) and of the impact of the market participants’ behavior on flow and price, as well as the relevance of applying advanced analytics solutions to monitor, decompose, and understand market dynamics.