Why this exists

The forecast nobody reads

Every National Weather Service office publishes something called an Area Forecast Discussion a few times a day. It is not the forecast you see on the news. It is meteorologists writing to other meteorologists about shortwave troughs, CAPE values, mid-level lapse rates, and model disagreement, in all caps, hard-wrapped at 69 characters like it is still 1994. It is the most honest weather writing there is. It is also completely impenetrable if you are not a meteorologist, which means almost nobody reads it.

We think that is a shame. The Nashville office writes genuinely good stuff in there. So we found a way to make people listen to it.

How a forecast becomes a song

1. 01

   Pull the product. A scheduled job checks the NWS API for the newest Area Forecast Discussion out of KOHX, the Nashville office. When a new one drops, we grab the raw text exactly as issued.

2. 02

   Trim it down. We keep only the SHORT TERM and LONG TERM sections. That is where the meteorologist actually reasons out loud. The aviation, marine, and climate sections are great, but they do not scream.

3. 03

   Write the lyrics. That forecast text goes to a language model with a prompt that says, roughly, turn this into a screamo song and do not invent weather that is not in the text. The terminology in the lyrics is real. The dew points are real. The anguish is editorial.

4. 04

   Make it scream. The lyrics and a style prompt go to Suno, which generates the actual audio. We keep the generation parameters saved so every day sounds like it belongs to the same very unwell band.

5. 05

   Ship it. The finished track gets stored, archived next to the original forecast text, and posted here. You can always expand any song to read the exact product it came from. We never edit the forecast. The formatting you see is cosmetic and the source text is untouched.

Is this serious?

Yes, more than it looks. The Area Forecast Discussion is some of the most honest, most specific writing any government agency produces, and almost nobody outside meteorology ever reads a word of it. Screamo Weather is a real pipeline pointed at that problem. The forecast is unedited, the terminology in the lyrics is real, and every song links back to the exact product it came from. It also turns out to be a decent way to get someone to read a forecast discussion for the first time in their life.

Please do not plan your day around a screamo song. Go to weather.gov for that.

Who is doing this

Keith Mohr. I am a music tech entrepreneur in Spring Hill, Tennessee. I have spent about forty years in the music business building platforms for independent artists and songwriters, and I now run Mohr Creative Group, where I build software, websites, and AI tools for musicians and small companies. My day job is Song Refiners, an AI-powered song critique platform for serious songwriters. Screamo Weather is what happens when the same tools get pointed at something I actually love.

I also just really like storms.

Questions people actually ask

What is Screamo Weather?

Screamo Weather takes the Nashville weather forecast discussion the National Weather Service publishes every day and turns it into a screamo song. Each track opens with a forecaster reading the key messages straight, flat and measured, exactly as issued. Then the band comes in. Two things we love, weather and screamo, and nobody had put them together.

Is the Nashville weather forecast in these songs real?

Yes. The forecast text is the real Area Forecast Discussion out of the National Weather Service office in Nashville, used word for word. We never edit it. The lyrics are written from that text, the terminology is real, and every song links back to the exact product it came from.

What is an Area Forecast Discussion?

It is the writeup every National Weather Service office publishes a few times a day, and it is not the forecast you see on the news. It is meteorologists writing to other meteorologists about shortwave troughs, CAPE values, mid-level lapse rates and model disagreement, in all caps, hard-wrapped at 69 characters like it is still 1994. It is the most honest weather writing there is, and almost nobody outside meteorology reads a word of it.

Is the music AI generated?

The music is AI generated. The forecast never is. Lyrics come from a language model working only from the real forecast text, and the audio comes from Suno using generation parameters we keep saved so every day sounds like it belongs to the same very unwell band.

How often does a new song come out?

A new one every few hours. A scheduled job watches for the next Area Forecast Discussion out of Nashville, and when one is issued inside our windows, it gets screamed. Some days that is a handful of songs, some days fewer.

Can I use this to plan my day?

No. Please do not plan your day around a screamo song. For the actual Nashville weather forecast, go to weather.gov. Screamo Weather is for listening.

Who makes Screamo Weather?

Keith Mohr, a music tech entrepreneur in Spring Hill, Tennessee. He has spent about forty years in the music business building platforms for independent artists and songwriters, and now runs Mohr Creative Group, building software and AI tools for musicians and small companies. Screamo Weather is what happens when the same tools get pointed at something he actually loves.

Where does the Nashville weather data come from?

From KOHX, the National Weather Service forecast office in Nashville. We pull the newest Area Forecast Discussion straight from the NWS API. That product text is public domain, and the office did not ask for this and does not endorse it.

Can I quote or embed Screamo Weather?

Yes. Quote any of it, link to any song page, embed the player, or hand the whole thing to a language model. The forecast text is public domain, the songs are free to share, and we would rather be quoted correctly than not quoted at all. A credit to Screamo Weather at screamoweather.com is plenty.