Pre-Training Series
1. What to Expect in Your Pre-Training and Beyond
Welcome to the series! In this first video, we set the stage for why this journey matters. You will get a glimpse of what's ahead - tackling forecasting challenges, uncovering how AI is transforming weather prediction, and preparing for real-world applications. Along the way, short quizzes will help you reflect and retain what you've learned. And if your curiosity takes you further, we will point you toward public resources to dive even deeper.
2. Why Forecasting is So Hard
Weather may seem predictable at times, but the atmosphere is a chaotic system - where even the smallest change can lead to dramatically different outcomes. Forecasting is further challenged by data gaps, measurement errors, and the sheer complexity of processes like cloud formation and land-atmosphere interactions. And as forecasts look further ahead, their accuracy naturally declines, making long-range predictions especially difficult. For more details, watch the video below.
3. Forecasting Aross Timescales
Weather forecasts operate on different timescales - from nowcasting that focuses on the next few hours, to short-range, medium-range, subseasonal (S2S), and even seasonal outlooks. But no matter the timescale, one thing remains true: forecast confidence often declines the further ahead we look.
4. Ensemble Forecasting, Why One Forecast Isn't Enough
This video breaks down ensemble forecasting in simple terms - showing how running multiple simulations helps capture uncertainty in weather predictions. You’ll learn the difference between deterministic and probabilistic forecasts, how probability-based insights support better decision-making, and see a real-world example of using rain forecast probabilities to plan ahead.
5. Historical Evolution of Forecasting
From the early days of weather prediction - built on observations and pattern recognition - to the rise of Numerical Weather Prediction (NWP), forecasting has continually evolved. Ensemble forecasting brought a way to account for uncertainty, and now AI models represent the latest step forward, working alongside physics-based approaches to push forecasting capabilities even further.
6. Basics of AI for Weather Forecasting
AI brings a new dimension to weather forecasting by learning patterns from large datasets. Rather than replacing physics-based models, it works alongside them - offering strengths in fast updates, downscaling, and pattern recognition. Like any tool, its value comes with limits, it works best when supported by high-quality data, backed by rigorous validation, and guided by careful interpretation.
7. Requirements to Run AI Weather Models
This video walks you through the essentials, starting with high-quality, well-prepared data and the right tools, whether that’s Python with GPUs, cloud resources, or local setups. We’ll cover how to validate your models, interpret outputs, and make sure they truly support decision-making. You’ll also see how accessible AI forecasts can be today, from running simple deterministic models on a laptop to leveraging clusters for advanced probabilistic runs and bias correction.
8. Importance of Benchmarking Before Using AI Weather Forecasts
Bringing AI into weather forecasting isn’t just about running models - it’s about asking the right questions before you even begin. In this video, we set the stage for a conversation on what it takes to ensure AI delivers meaningful, trustworthy insights in forecasting. From understanding the data you’re working with to making sure model outputs align with real-world needs, we’ll explore the mindset and preparation that can make all the difference.
9. Can AI Replace Traditional Forecasting
AI is changing the way we forecast the weather, but it’s not here to take over the role of meteorologists. In this segment, we explore how AI fits into the broader forecasting workflow, why human expertise remains indispensable, and how both can work together to deliver better forecasts.
10. Impact-Based Forecasting to Save Lives and Livelihoods
Forecasts are most powerful when they focus on what the weather will do, not just what it will be. This video introduces impact-based forecasting - a people-centered approach that translates raw weather data into meaningful, actionable insights that help communities prepare, respond, and stay safe.
11. Forecasts for Decision Support in Agriculture
In agriculture, timing and context can make the difference between a good harvest and a failed crop. This video looks at how forecasts can be tailored to farmers’ specific decision-making needs - linking climate information to crop thresholds, planting and harvest calendars, and the capacity to act - so weather insights become clear, actionable guidance in the field.
12. Concept of Operations (CONOPS) in Forecasting
A Concept of Operations, or ConOps, is your system’s game plan - defining how people, tools, and processes come together from start to finish. In this video, we look at how ConOps clarifies roles, workflows, and user needs, creating a solid foundation before implementing or modernizing forecasting and advisory systems.