The Temperature's Rising, It Isn't Surprising: Building Network Resilience in a Warmer Climate

Meteomatics & Heimdall Power — "The Temperatures Are Rising: Building Network Resilience in a Warming Climate"

Speakers:

• Candice Thompson — Meteorologist & Customer Success Manager, Meteomatics (Moderator)
• Dr. Matthias Piot — Senior Energy Meteorologist, Meteomatics
• Keanu Giebels — Account Executive, Energy & Utilities, Meteomatics
• Brian Berry — Chief Product Officer, Heimdall Power

1. Introduction — Candice Thompson

A warm welcome to the Meteomatics and Heimdall Power webinar. I'm Candice Thompson, Customer Success Manager at Meteomatics, joined by my colleagues Keanu and Matthias, both energy experts, and our special guest Brian Berry, Chief Product Officer at Heimdall Power and one of our longstanding customers.

Today we're talking about one of my favorite seasons — summer. It's a nice change to discuss something warm rather than the usual cold and icy conditions. Today's theme is "The Temperatures Are Rising: Building Network Resilience in a Warming Climate."

We've already seen very high temperatures in Europe, and there's talk of a strong El Niño toward the end of summer — some headlines are even calling it a "Godzilla El Niño." We'll cover summer energy trends and what they mean for supply and demand, hear from Heimdall Power on how grid capacity can be freed up, look at Meteomatics tools in action with some AI examples, and close with Q&A. Please post questions in the chat as we go.

To start, I'll hand over to Matthias to walk us through the recent European heat and El Niño.

2. European Heat Waves & Grid Stress — Dr. Matthias Piot

Thank you, Candice. We have attendees from all continents today, drawn largely by two stories: the European heat waves and the developing Super El Niño in the Pacific. We'll focus on the European heat waves and what they mean for grid operators.

Let's start with an inconvenient truth. The temperatures across Western Europe were the warmest in modern history. Going back to at least 1940, and beyond, you won't find such a warm June. I'm happy to cover the meteorological mechanisms in the Q&A.

These visualizations were made with our tool, MetX. They show the extent of the heat wave and the parts of the European power network most affected. France saw temperatures up to 17°C above normal. A large part of the Central European grid was hit by scorching temperatures, and national weather services issued red alerts.

Is this the new normal? Not yet — but no one should be surprised by the trend. We've experienced it for several decades, and grid operators will face these situations more and more often. That means more stress on the grid, less energy transmitted through the lines — the ampacity — and more stress on aging European infrastructure.

On forecasts: France is about to experience its fourth heat wave in a row. The 14-day forecast, and the long-term outlook, expect above-normal temperatures through July and August.

Heat drives grid stress. Recent examples: Italy had eight power failures within 24 hours, caused by increased air-conditioning load on old infrastructure. In Western France, a transformer exploded from overheating, leaving 100,000 households without power for days. Wildfires in Southern France added further outages. On top of that, data centers are popping up like mushrooms and electricity demand keeps growing, while electrification of infrastructure is slow and extremely costly.

The key question: how climate-resilient is our power network, country by country? The heat waves won't wait, and they'll return more often.

This short video shows intraday prices across Europe during the two hottest days — prices spiked dramatically, a sign the supply side was under heightened stress to keep the network balanced.

In numbers: French demand rose 10–12 GW purely from air conditioning — a huge figure, given a nuclear reactor is about 1 GW. On the supply side, French nuclear was constrained because river temperatures exceeded 28°C, so warm cooling water couldn't be discharged. UK CCGT gas plants faced similar restrictions. Altogether the heat added roughly 20–35 GW of cooling load across Europe. This is the setup for what we call, in German, the Hitzeflaute — lots of heat, almost no wind, and a lot of solar. Over to Keanu.

3. The Hitzeflaute — Keanu Giebels

Thank you, Matthias. We've all sat through plenty of slides about the winter Dunkelflaute, so I want to talk about its summer cousin — the Hitzeflaute: a sustained period of excessive heat, little to no wind, and still-high solar generation during the day. Germany, and much of Europe, was hit by one about three weeks ago.

This isn't a weather lecture — I want to focus on what it means for the grid, for reserves, and for market prices.

The setup is simple: it's hot, and the air isn't moving. Demand climbs as every cooling system kicks in. Solar helps during daylight — we even see negative prices at midday — but the moment the sun goes down, solar disappears, demand stays high, and there's no wind to lean on. You get an evening cliff where operators scramble for capacity, and prices can swing €300–500 in a single day.

This happened in June in Germany. German average demand climbed from about 53 GW to 58 GW. France rose almost 20% over the same two weeks, from about 44 GW to 52 GW. Germany peaked above €700/MWh, France around €430, and Spain lower at about €185. On the supply side, EDF had to pull 8% of France's nuclear generation because the rivers cooling the reactors were too warm.

Two weeks before the peak, German onshore wind was producing over 22 GW. By the day load peaked, wind had collapsed to about 1.3 GW — a 90%+ drop — exactly when the grid needed it most. The same heat-and-calm combination also eats into the thermal headroom on lines and transformers.

The real question is how you find flexibility when the weather is working against you on every front at once. That's where high-resolution, real-time weather data comes in — knowing conductor and ambient conditions well enough to push more through a line when you can, and pull back before you have to. Which leads to our partner, Heimdall Power.

4. Heimdall Power: Dynamic Line Rating — Brian Berry

Thank you. You've heard about the risks from weather, but demand is also surging — from electric vehicles and immense data center loads, across both the US and Europe. This strain shows up all over the grid; we've mentioned transformers overheating, but power lines are especially hard to manage, since they span thousands of kilometers.

Utilities typically use static or seasonal line ratings, assuming the worst case for a season or the whole year. But there's extra capacity hiding on the lines — as long as you measure the weather. That's what we do, either with a sensorless or sensor-based approach. In extreme heat waves, the network is pushed well beyond its assumed worst case, so real-time monitoring of the line and the weather really helps.

The analogy I like: it's like putting a reverse camera on your car. With it, you can park much closer to the wall. Same concept — you can safely operate closer to the line's real limit.

What Heimdall does: we install the Neuron, a sensor about the size of a bowling ball, onto the line using drones. It tells you exactly what the line is doing — the reverse camera. We also offer the Virtual Neuron, powered by Meteomatics data that has been tuned to understand the power system based on our experience with the physical Neurons. Our platform brings it together so customers know exactly what's happening on their grid.

Heimdall has been operating for ten years, active in European and American markets. A few examples:

• Bayernwerk (Germany): They were already using weather stations, but that missed how conditions affected the conductors. Adding sensors to the lines revealed more capacity — a roughly 40% capacity increase. They've since purchased more systems. Many German TSOs and DSOs are following their lead. For consumers, better management of congested regions means cheaper electricity.
• Irish TSO: They wanted to connect a wind farm but would have needed an expensive new line. Using dynamic line rating with the Neurons, tied into a curtailment scheme, they connected the capacity without the physical upgrade — major savings.
• US utility: Using an autonomous, self-flying drone, they installed Neurons on several key lines in congested regions with extreme price swings. In one congestion event, the extra capacity — the reverse camera — saved $3.2 million, savings that ultimately reach consumers.

5. MetX in Action: 19 June Netherlands Storms — Keanu Giebels

 I've described a slow-moving problem — the Hitzeflaute — so now the opposite: a day that went from calm to chaotic in hours, where the only way to keep up was one screen showing the whole picture. That's what MetX is for.

The 19th of June was one of the hotter days in the Netherlands, with temperatures well above 30°C in the evening, plus extreme thunderstorms — over 180,000 lightning strikes detected.

In MetX: top-right shows heat warnings; bottom-right, thunderstorm tracking; top-left, lightning strikes; bottom-left, the grid view of Dutch overhead lines with a warning system built on wind gusts, color-coded to thresholds I defined.

In the animation, you can see thunderstorms moving from Belgium over the Netherlands with the lightning strikes, and the overhead lines changing from yellow (lower wind-gust threshold) to red-purple (high risk). It's about acting proactively on weather events instead of reactively. In today's landscape, speed is turning into the key.

6. The Meteomatics MCP Connector — Dr. Matthias Piot & Keanu Giebels

Candice: Meteomatics recently introduced something we're excited about — the MCP connector, which lets you connect the API to your AI agents.

Matthias: MCP means Model Context Protocol. To visualize it: AI is your computer, and the MCP connector is a USB stick full of information — you can plug in a huge amount of Meteomatics data. It's easy to connect, and works with most AI environments — ChatGPT, Claude, Copilot, and more. There's a specific address to connect to, and we can help.

Once connected to our weather API, you can, for example, ask Claude to build an HTML dashboard on the impact of the French heat waves, specifying the Meteomatics MCP connector so the data comes from us, using the best model available — for Europe, that's EURO1k. You can add weather maps, polygons, and even read in a CSV, such as the French electrical grid network. Importantly, you instruct it to use our weather API and not fabricate or use open-source data, ensuring quality.

The result: one user-friendly dashboard where a TSO, DSO, or grid operator can quickly understand where the heat is, what to watch, and the expected ampacity reduction by region — within a few clicks. Extremely powerful, and your creativity is the limit. I strongly suggest testing it.

Keanu: I share the optimism. In the same summer-situational-awareness theme, the prompt on screen is exactly the one I used — written in seconds, in plain natural language. The point is that anyone with access to the Meteomatics API can replicate this. I was skeptical at first about what I'd do with an MCP server, but the possibilities are endless. The dashboard came out in minutes — most of the time was just waiting for Claude to finish. It's limited to a few locations here, but a grid or renewable-asset operator can add any set of locations, assets, and parameters to understand the risks they face at different times of year.

7. Panel Discussion & Q&A

Q1 (from audience): Normally DLR is used to gain capacity when a line is cooler than the static model assumes. In a heat wave, isn't it the opposite — you learn you have less capacity, because the real temperature is higher than the value used in the static calculation?

Brian Berry: Fair question. The short answer: DLR still provides value — because of wind. Of the three factors affecting a line — solar, ambient temperature, and wind — wind has by far the greatest impact, but it's also the hardest to measure accurately. So even in a heat wave, wind can provide relief a static rating would miss. You're right that the temperature side is often strained and can lower your rating; if there's no wind and high ambient temperature, you'll see a loss of capacity, which isn't congestion relief. But it's far better to understand that risk than to be blind to it — it lets operators take the right action, including clamping a line down. Better to know it than not.

Candice / Matthias (on wildfire risk): Hot, dry, windy conditions — especially with strong El Niño events — raise wildfire risk. Matthias noted the Balkans in June 2024, where heat made cables expand and sag; if cables sag too low, there can be a flashover with vegetation, potentially starting wildfires — which contributed to a huge Balkans blackout.

Brian Berry (on wildfire): Wildfire is a huge topic, especially in the US, where utilities have gone bankrupt over liability claims for fires started by their assets. There's no silver bullet — it spans prevention (understanding dry, hot weather and rising risk), infrastructure (underground cables vs. overhead lines, tree trimming), and operations (running the grid even when you must switch off power to mitigate risk). Meteomatics data helps assess and predict risk areas, and the Neuron helps operators run strained grids more effectively when fires do occur.

Brian also noted: even a small, controlled fire passing under a line is disruptive, because the smoke is full of conductive carbon and can cause a flashover. Operators often disable auto-reclose (which normally switches a line back on after a fault) and preemptively manage the line. He and Candice, both from South Africa, recalled tracking fire and lightning events via satellite in a control room back in 2004 — both very disruptive to the grid.

Q2 (from audience): Are there parameters in the API or MetX to track wildfires and their potential spread?

Keanu Giebels: We have wildfire-specific indices like the Fosberg Index, which I used in the MCP example. In MetX you can build custom parameters and situational-awareness dashboards for your use case. But tracking a wildfire's spread isn't really our domain — we identify wildfire risk potential.

Matthias Piot: We have many relevant parameters — wind, vegetation, soil dryness, and thresholds. But dryness alone doesn't cause fire; you need an igniter (human or natural, such as lightning), and there are suppressors (precipitation, low wind) that limit spread. You can combine lightning, wind, suppressors, and precipitation in one MetX or MCP dashboard to identify where risk is highest — but not to track spread or consequences.

Brian Berry: To add: there's no silver bullet — the power is in combining data sources. There are excellent wildfire-fuel APIs and satellite APIs; bringing in the Meteomatics MCP alongside them is straightforward and gives you the full picture from the coincidence of those data points.

8. Closing — Candice Thompson

Thanks to Brian, Keanu, and Matthias — great examples across the heat and summer theme. It's a reminder that weather data has many real-world applications, and you don't need to be a data scientist to use it. (An audience poll on AI adoption followed.)

Thank you for joining. Please reach out to Meteomatics at [email protected] to walk through our products or MCP, and our Customer Success team supports existing customers. We run this webinar once a quarter — look out for the next invitation. A recording will be in your inbox in the next couple of days. Have a great summer.