Below are some publicly available datasets for flood prediction, but newer or new datasets may have been added to these datasets. The latest datasets can be found at relevant research institutions, government agencies or data repositories. Some available public datasets include:
NOAA National Weather Service data: The National Oceanic and Atmospheric Administration (NOAA) provides data on weather, precipitation, and river levels that can be used in flood forecasting studies. You can find this data on NOAA's website.
USGS River Flow Data: The United States Geological Survey (USGS) provides flow data on rivers across the United States. The data can be used to monitor river levels and flows to help predict flooding.
CAMELS dataset: The CAMELS (Catchment Attributes and Meteorology for Large-sample Studies) dataset contains hydrological and meteorological data from multiple watersheds in the United States and can be used for hydrological modeling and flood prediction research.
European Space Agency (ESA) Sentinel satellite data: ESA's Sentinel satellite provides global surface coverage and meteorological data that can be used to monitor flooding events.
International Flood Database: The International Flood Database contains historical data on flood events around the world and can be used to study the frequency and impact of floods.
International Hydro-Meteorological Observatory System: This system provides hydrological and meteorological observation data on a global scale, including rainfall, temperature, humidity, etc., which helps flood prediction and monitoring.
These datasets may require some data processing and cleaning to fit a specific flood prediction model or research question. In addition, the availability of and access to data sets may change over time, please consult the relevant data provider's website or data repository for the most up-to-date information and data.