We pride ourselves on processing a wide variety of observational and model data. Years ago, we started with the standard National Weather Service models and NEXRAD radar you’ll see in weather apps. Since then, we’ve added a wide variety of specialized data and weather models for our customers.

Machine learning-based models have become a larger part of that. One we’d like to highlight today is from Meandair, a company based in the Netherlands that provides a variety of machine learning model blends for aviation and other uses.

Several of our customers use their data, and we’re adding tighter integration ourselves. Let’s look at one of their datasets and how it can be used.

Weather Models for Cloud Cover

Cloud cover is typically measured as a percentage from the surface to the edge of the atmosphere. 0% means it’s completely clear, and 100% means completely covered.

Most models include a cloud cover variable, often broken into levels such as low, mid, and high. In general, these are going to be pretty useful, but not all that precise. For example, here’s GFS over Europe.

Over the course of the day, you’ll get a pretty good sense of what might happen, but it’s not very precise in the next few hours.

Some products do better in that NOWCAST range, like Meandair’s cloud cover.

These weather models help provide more precise short-term forecasts of cloud cover by combining satellite imagery with multiple model inputs. They describe it better since it’s their process. We just deliver it quickly and efficiently.

Further out, you’ll get something more like a blended model.

Since it’s one data set, we can query it too. Here’s our WCS query along with one of the return values.

For a whole forecast run, this is pretty useful.

Meandair Data Access

The Meandair datasets, including cloud cover, turbulence, precipitation, thunderstorms, and more, are a separate cost option. We pay them, so you have to pay us.

As weather models continue to improve, datasets like Meandair’s are becoming increasingly useful for aviation and operational forecasting. Meandair data is available for display in Terrier, for querying in WCS, and for transformation in Labrador. The latter provides Zarr files that are easy to access in Python, making it simple to work with these advanced datasets in your own applications and workflows. See what you can do with these awesome data sets!