Pinned Tweet
In the past 6 months we’ve produced a cheaper, more efficient seeding material, trained a hyper-resolution AI weather model, and operated 100+ cloud seeding flights for the governments of Texas and New Mexico.
Recast Systems is coming out of stealth. Here’s some of the research (1/n)
Recast Systems is coming out of stealth. Here’s some of the research (1/n)
84
110
1,019
338,225
Recast Systems retweeted
Turns out people really do love the NAM despite its flaws. We @recast have decided to continue to offer our WOOF-NAM 12km for free indefinitely (link below).
It’s also important to understand that we are cold starting from GFS Analysis which naturally does slightly change model outcomes compared to starting from a native NAM analysis.
Because of that, we will be running quite a few hindcasts of both popular and requested dates and model inits so people can compare. If anyone has any examples they’d like to see, I’ll run them and put it on the site.
woof-nam.recastsystems.com/index.html
It’s also important to understand that we are cold starting from GFS Analysis which naturally does slightly change model outcomes compared to starting from a native NAM analysis.
Because of that, we will be running quite a few hindcasts of both popular and requested dates and model inits so people can compare. If anyone has any examples they’d like to see, I’ll run them and put it on the site.
woof-nam.recastsystems.com/index.html
NOAA retired the NAM this week. We ported it to the GPU and it's back, better than ever: WOOF-NAM, open source.
- 12 km CONUS, all NAM physics, 24 h forecast in ~20 min on a single RTX 5090 (32 GB).
- 8 GPUs: 6 h in 92 s. Multi-card is bitwise to single card.
- 4 km CONUS, 24 h s. 3 km nests anywhere on Earth (Europe, Japan, CONUS): 48 h in ~7 min on one card.
SStarts from the NAM analysis or GFS today, IFS open data in the next release. Beside the operational NAM from the same cycle it's hard to tell them apart (pics).
Today it needs one 32 GB card for the full 12 km CONUS; ~15 GB is the target, coming soon.
github.com/recastsystems/…
10
14
110
16,242
NOAA retired the NAM last week. We thought it deserved to live on, so we ported it to the GPU and open-sourced it.
We’re running the 12 km CONUS NAM ourselves, four times a day (00/06/12/18Z, 84 h, started from GFS), with the hourly files free for anyone: same awphys GRIB2 shape as NOAA’s, so an old NAM fetcher just needs a new host.
woof-nam.recastsystems.com/index.html
The 12 km shape is in the WOOF API today, 5 km and 4 km CONUS this week, so you can run your own cycles any time from GFS or IFS.
We’re running the 12 km CONUS NAM ourselves, four times a day (00/06/12/18Z, 84 h, started from GFS), with the hourly files free for anyone: same awphys GRIB2 shape as NOAA’s, so an old NAM fetcher just needs a new host.
woof-nam.recastsystems.com/index.html
The 12 km shape is in the WOOF API today, 5 km and 4 km CONUS this week, so you can run your own cycles any time from GFS or IFS.
NOAA retired the NAM this week. We ported it to the GPU and it's back, better than ever: WOOF-NAM, open source.
- 12 km CONUS, all NAM physics, 24 h forecast in ~20 min on a single RTX 5090 (32 GB).
- 8 GPUs: 6 h in 92 s. Multi-card is bitwise to single card.
- 4 km CONUS, 24 h s. 3 km nests anywhere on Earth (Europe, Japan, CONUS): 48 h in ~7 min on one card.
SStarts from the NAM analysis or GFS today, IFS open data in the next release. Beside the operational NAM from the same cycle it's hard to tell them apart (pics).
Today it needs one 32 GB card for the full 12 km CONUS; ~15 GB is the target, coming soon.
github.com/recastsystems/…
3
27
167
17,357
Code: github.com/recastsystems/….
Files CC BY 4.0, credit “WOOF-NAM by Recast”.
Files CC BY 4.0, credit “WOOF-NAM by Recast”.
0
0
16
1,241
We open-sourced our GPU port of WRF and named it WOOF (Weather Observations and Operational Forecasting). We want everyone to have access to fast and cheap weather simulations.
Run your own simulations at woof.recastsystems.com
Run your own simulations at woof.recastsystems.com
We ported the government’s weather model (WRF) over to GPUs so it runs a magnitude faster and cheaper.
Here's a 250m resolution forecast of the fog is SF.
sf.recastsystems.com
15
66
424
56,656
WOOF was built so we could run cheaper, near real-time simulations for our operational cloud seeding missions.
We realized it could be useful outside of cloud seeding so we packaged WOOF to be shared.
We realized it could be useful outside of cloud seeding so we packaged WOOF to be shared.
1
0
10
990
It’s a bitwise exact GPU port of WRF-ARW (drycore 4.6.1) which makes it physically coherent but a magnitude faster and cheaper
1
0
8
832
Tropical storm Edouard makes landfall in Texas. We’re opening up our radar and storm tracking software to the public so everyone can track Edouard.
2
9
38
2,834
Recast has developed a natural and biodegradable seeding material. We grow it ourselves.
Thanks to the @nytimes for covering some of our progress as we develop a new generation of materials, tools, and use cases for cloud seeding.
Thanks to the @nytimes for covering some of our progress as we develop a new generation of materials, tools, and use cases for cloud seeding.
2
11
110
27,231
extra, extra, read all about it! nytimes.com/2026/08/24/cli…
0
2
18
1,888
If this is research you’d like to work on. We’re hiring at recastsystems.com
Focusing an image from a laser beam diffraction pattern 1/5
0
1
43
4,325
Growing the cloud seeding industry requires measurements to prove ROI. We build sensing hardware in-house to see what’s happening inside clouds and to measure exactly how much precipitation we've caused.
We use in-line holography to captures images of individual aerosols, water als as they move through a cloud. This lets us directly measure cloud microphysics like how particles interact, how droplets freeze, and how ice crystals grow.
We use in-line holography to captures images of individual aerosols, water als as they move through a cloud. This lets us directly measure cloud microphysics like how particles interact, how droplets freeze, and how ice crystals grow.
2
2
89
7,490
Together, these technologies form a feedback loop. We target optimal clouds, seed them with our new material, and collect data with our hardware to validate real world results against our models, recursively improving our operations.
We’re building across biology, software, and hardware. If any of our research is interesting to you, DM, or apply at recastsystems.com.
We’re building across biology, software, and hardware. If any of our research is interesting to you, DM, or apply at recastsystems.com.
1
0
75
5,092