The new standard in 2- to 52-week weather analytics

Salient combines novel ocean and land-surface data with machine learning and climate expertise to deliver the world’s most accurate subseasonal-to-seasonal weather forecasts 2 to 52 weeks in advance.

Request a Demo

2X Accuracy improvement
over competitive forecasts
4 billion Machine learning predictors
5 million API data points generated weekly

Gaining new insight from the global water cycle

Developed by scientists from MIT and Woods Hole Oceanographic Institution, Salient’s forecasts and analytics put superior data and state-of-the-art methodologies to work for you.

Inertia: the true driver of seasonal weather

Conventional forecasts are based on numerical models of atmospheric conditions. However, due to the chaos of the atmosphere, these models provide little skill for subseasonal-to-seasonal weather predictions. And while some statistical models of ENSO cycles provide modest skill at seasonal timescales, there is much more to the ocean than just ENSO. Our decades of research reveal that ocean and land-surface conditions—two global features with greater inertia and heat capacity—have the largest influence on seasonal weather patterns.

A global-scale machine learning platform

Salient uses deep neural networks to analyze a wide range of climate data. Our models are statistical in nature, so they don’t get bogged down by the details of atmospheric physics—and they have the scale and complexity to find predictability in all aspects of the climate system.

Industry Analytics

A breakthrough in predictability

Our deep learning tools efficiently comb through large datasets and identify complex climate system relationships to provide accurate forecasts that are used in diverse applications across agriculture, energy, finance, supply chain, and beyond.