Sunrise at Autumn’s End

Everyday I am out in the weather for at least part of the day. This December has been weird with heavy snowfalls coupled with spells of above freezing ambient temperatures. The talk on the trail is about how the trail surface varies with snowfall and temperature. We are all of retirement age so our concern is testing new muscles and stress on our ankles and joints. Thus far I have been able to navigate rough surfaces without mishap.

The outdoors temperature was 37 degrees Fahrenheit when I woke. By the time I was ready to walk on the trail it had dropped to below freezing. As I entered the state park, the wind picked up and snow began to fall. The trail surface had frozen unevenly around snowmobile, wildlife, and pedestrian tracks. I had to be careful to preserve my joints and ankles. I knew the weather forecast before I left home.

I’m not the only one who takes accurate weather forecasts for granted. My resource is the Weather Channel with its hourly forecasts and radar images. Mostly I seek short-term weather forecasts in order to do something like mow the lawn, take a walk, or work in the garden. Existing technology gives me accurate forecasts that facilitate what I want to do.

Hannah Ritchie is a Scottish data scientist, senior researcher at the University of Oxford, and deputy editor at Our World in Data, according to Wikipedia. Her work focuses on sustainability, in relation to climate change, energy, food and agriculture, biodiversity, air pollution, deforestation, and public health. She recently wrote the following about weather forecasting:

The potential for AI to improve weather forecasting and climate modelling (which also takes a long time and uses a lot of energy) has been known for several years now. AI models have been tested for one- and two-week forecasts with promising results. Scientists will often need to wait weeks for a complex, high-resolution climate model to run; AI might be able to do this hundreds, if not thousands, of times faster. (Email from Hannah Ritchie on Oct. 27, 2025).

Naturally, I asked ChatGPT about this, “When scientists like Hannah Ritchie say weather forecasting can be much improved with artificial intelligence, what do they mean.” As is usual for artificial intelligence, the answer was very long. Here is the crux:

When scientists say weather forecasting can be dramatically improved with artificial intelligence, they usually mean something more specific than “AI makes everything better.” They’re referring to concrete technical bottlenecks in current forecasting and the ways AI can overcome them. Hannah Ritchie, and researchers at ECMWF, NOAA, Google DeepMind, NVIDIA, the UK Met Office, and others, are all pointing to similar developments. (ChatGPT query result, December 2025).

More specifically, in explanation, the machine listed multiple bullet points:

• Weather prediction is limited by physics-based models.
• AI models can “learn the atmosphere.”
• AI lets scientists blend physics and data.
• AI makes weather forecasting more democratic. By that, it means cheaper and more broadly available than on expensive, physics-based computers.
• AI enables longer-range and global risk forecasting: seasonal climate forecasts, agricultural and drought planning, energy-grid load forecasting, and catastrophe-risk modeling.
• But: Scientists emphasize that AI is not a replacement for physics.

The machine summarized: forecast faster, forecast at higher resolution, run at vastly lower cost, improve extreme weather warning lead times, complement physics with learned patterns, and democratize forecasting globally.

According to the machine, consumer-scaled artificial intelligence models might be available by 2032. In the meanwhile, I’m just glad I didn’t turn an ankle on the trail this morning.