Whiffle

The #Whiffle team is going to be at E-world 2025! If you want to learn more about how our LES-powered forecasting technology can enhance your trading results, drop us a message via this form: ✅ https://lnkd.in/eRBYwy9T

Those who follow Whiffle know that we’ve been doing 100m-scale weather modelling for quite some years now (in fact, we celebrate our 10-year anniversary this week 🎉). If you are an optimistic person like myself, you could even say that we are quite successful at it 😊. However, this summer an interesting scientific paper (https://lnkd.in/d4v3vs9W) came out. It describes challenges and ways forward towards the implementation of… (drumroll) …. 100m-scale weather modelling! Jokes aside; this is a really good paper, which was to be expected from the large list of renowned scientists who wrote it. Yet, when reading it, I went through a wide range of emotions like recognition, frustration, pride, hope and fear. A few thoughts: 😦 Overall, the authors seem quite pessimistic about the future of 100m-scale weather forecasting. They highlight the gigantic computational costs involved and outline a long list of ‘problems’ that must be addressed with parameterizations, treatment of the surface, data assimilation, etc. Furthermore, they discuss the limits of predictability, suggesting that ensembles might offer a solution, though these would further escalate computational demands. By the time you reach the end of the paper, it feels that you must be quite out of your mind to even consider doing this. 🙂 I kind of missed the positive side, the great potential that 100m-scale weather modelling has: superior representation of clouds and rain, simplicity (less need for complex parameterizations), the ability to assimilate more observations, representing exchanges with the surface, orography, extremes, better visualizations, wind farm wakes, you name it.   🏙️ The strong emphasis on urban weather applications was a bit surprising. Certainly relevant, but what about mountain weather, islands, coastal/lake regions, etc? These actually match better with 100m resolution than the urban environment, which probably requires even higher resolutions. And how about renewable energy? In Whiffle’s view the most important use-case, but hardly mentioned. Nevertheless, a super interesting read that also provided lots of insights and suggestions for our own R&D efforts. To the authors (some of whom are literally quite close to Whiffle) I extend a warm invitation for a coffee and a good discussion on 100m-scale weather modelling☕ And as an inspiration: check out last week’s forecast with the Whiffle 100m LES model of a cold-air outbreak (https://lnkd.in/dAMAtn9U) over the North Sea and see how beautifully the honeycomb patterns are captured in the vertical velocity field (red-blue colors) and ice clouds (purple). (Run setup here: https://lnkd.in/dX_Q55nf) Thanks to Chiel van Heerwaarden for pointing out this interesting case!   

One last large-eddy simulation to end the year with, merry Christmas! Here, I used Whiffle's large-eddy simulation code ASPIRE to simulate the atmospheric boundary layer over Cabauw, The Netherlands. And when adding some water vapor at the right places, this is what you get!

Why do oversimplistic approaches to wind resource and yield assessments persist in an industry driven by multi-billion-euro investments? High-fidelity, low-uncertainty models seem like a no-brainer to avoid massive losses. ☝️ Maybe it’s due to the advanced modeler’s curse. Imagine this scenario: an advanced wind resource analyst with a strong scientific background is working on a project for a wind farm tender. The analyst runs an advanced model simulation that captures wake losses, global blockage and other subtle atmospheric effects, works in the time domain rather than using a traditional ‘directional’ approach, and forecasts multiple potential scenarios of future wind farm build-out. He/she also performs a sensitivity analysis on model parameters, climate change, electricity prices, interest rates, and so on. The result? A sophisticated time series of thousands of scenarios of wind farm incomes and expenses over time, beautifully condensed into a probability distribution of risk-weighted NPVs advising 🚨 NOT to bid on this tender, or to bid at a high price. Now imagine this advanced modeler’s simplistic counterpart: driven by wishful thinking, blessed by ignorance and pressured by management, this modeller runs a few simple models and quickly produces a superb business case👌 The result? The modeler’s manager is thrilled; the company submits an aggressive bid and wins the tender, 🙌 followed by a burst of positive press. But what happens ten years later? The project suffers huge financial losses, a grim inheritance of the initial modelling flaws. This narrative isn't just hypothetical; it’s a repeated reality in our industry🤦. Come on, folks! You can’t fool the laws of physics. The truth will come out.

At Whiffle we like small-scale weather phenomena 🔎 And we like it even better to simulate small-scale weather phenomena. But we like it best if our own Whiffle LES model accurately simulates a particular small-scale weather phenomenon 🙌 Therefore, this particular movie of a simulation of the famous 'Levanter cloud' over Gibraltar made our day 🤩 It seemed quite accurate if you compare it with this movie https://lnkd.in/dvh_QHR7. A nice description of the phenomenon can be found here: https://lnkd.in/duxGR8TV Oh, and yes, we realize that this is not a high-impact weather event. But exploring the exotic corners of meteorology teaches us about how our model captures certain physics, so therefore we do it. PS: Indeed, it was of course Peter Baas who made the movie again 🙏

The November edition of The Eddy Effect is here! This month’s key update? Version tracking is live in Whiffle Wind! 🛠️ You can now easily see which Whiffle LES model version powers your simulations. Browse the newsletter to find our more (link in comments 👇 ). #WhiffleWind #WhiffleLES #Whiffle

We had the pleasure to welcome Pedro Costa at Whiffle for a presentation about his work on efficient pressure solvers for turbulent flows. And what a wonderful story we were treated on! You could see all (25+) technical folks in our company soaking it up. We look forward to collaborating on exascale simulations of, in Whiffle's case, atmospheric flows. The work Pedro and co-workers did on the CaNS solver (https://lnkd.in/ePdRCfq3) is extremely promising. Below a beautiful image from the CaNS GPU paper https://lnkd.in/eUmEmz4D.

🚀 #Whiffle is growing - we're on the hunt for a Meteorological Data Scientist! Do you thrive at the intersection of data science, software engineering, and meteorology? Dive into a role designed to continually advance of our short and medium-range forecasting products, ensuring our forecasting capabilities remain at the forefront of the industry. What will you do? ➡️ Work closely with meteorologists and scientists to blend their expertise into our forecasting systems. ➡️ Design, test and refine new modelling approaches that drive our business forward. ➡️ Collaborate with ML and DevOps team members to bring machine learning models into production. ➡️ Stay at the forefront of advancements in meteorology and AI, ensuring our forecasting remains cutting-edge. We’re seeking a team member with a solid foundation in physics, coding (Python) and time series forecasting, and of course, a problem-solving drive to convert complex weather data into actionable insights that enhance renewable energy decision-making. If this sounds like a challenge you’re up for, learn more via: https://lnkd.in/etM4VKeh and apply!   (Also, if you know someone who might be a great fit for the role, do share this post!) #Whiffle #WhiffleLES #WhiffleTeam

The October edition of The Eddy Effect is now live! In this issue, we share with you recent #WhiffleWind updates that have made our user-friendly, wind resource and yield modelling tool, all the more intuitive: 👉 Clone turbine models for your simulations, with a single click 👉 Track your simulation spending effortlessly via the new Balance page 👉 Use Color-code maps to easily distinguish between multiple wind farms … and more! Plus, if you missed the recent Whiffle Tech Talks (on Challenging Atmospheric Flows) at the #ACPRT24 Conference you can now explore the full presentation. (Dive in and start reading - newsletter link in comments!) #TheEddyEffect #Whiffle #WhiffleLES

👏 Whiffle has officially achieved the ISO 9001 and ISO 27001 certifications! These standards recognise our unwavering commitment to quality management and information security.   🏆 ISO 9001 certifies our ability to consistently deliver services that meet customer and regulatory standards. 🔐 ISO 27001 validates our efforts to ensure the utmost security of sensitive information. These milestones are not just achievements for our company; they reflect our dedication to excellence and secure practices in every operation.   A heartfelt thank you to our incredible team for making this possible and to all our partners for your ongoing support and trust. #Whiffle #ISO9001 #ISO27001 #QualityManagement #InformationSecurity