What is Nowcasting, actually?
The World Meteorological Organization (WMO) defines nowcasting as a detailed analysis and description of the current weather and then forecasting ahead for 0 to 6 hours.
Nowcasting is specifically about addressing the period between Numerical Weather Predictions (NWP) - the forecasts our National meteorological institutes generate for us - that are typically produced and published every 6 hours. But it's not just filling the gap, it's using current sensor data to fill the gap with near-real-time short-term forecasts that are substantially more accurate and more focused.
In the nowcasting time range it has become possible to forecast small scale features, such as individual storm cells with reasonable accuracy, when utilizing the latest information from surface stations (on land, ocean and ice surfaces), radar, satellite, and lightning observations.
Nowcasting is needed for timely, location-specific predictions of high-impact weather, which are relevant to many sectors, not only for the general public, but also for commercial as well as private industries. Any outdoor activities, such as aviation, outdoor sporting events, the construction industry, power utilities and ground transportation organizations, need frequently updated information of current and upcoming weather situations. Nowcasting plays an increasingly important role in disaster risk management and risk prevention, as well.
Weather phenomena for which nowcasting are highly relevant to various user groups include the following: thunderstorms (including associated phenomena such as lightning, hail, damaging wind, tornadoes), heavy precipitation (flash floods, landslides), severe wind, visibility/fog, winter precipitation type (snow, sleet, freezing rain, drizzle, icing).
What's the difference between Forecasting and Nowcasting, then?
Simply and practically speaking, nowcasting and forecasting only differ in the period of focus for the forecast, the frequency of updates and the sensor latency. A nowcast's performance/accuracy is highly correlated to its sensor latency, the amount of time between the actual measured weather phenomena and the publishing of the nowcasts for consumption. (Learn more about sensor latency here)