Enter Note Done
Go to previous page in this tab
Session
  • Presentation | H31I: Utilizing Precipitation Datasets and Quantifying Associated Uncertainties in Hydrometeorological and Climate Impact Applications I Oral
  • Oral
  • Share Icon
    Bookmark Icon
  • [ONLINE] H31I-09: Translating AI-enhanced forecasts and automatic weather station observations into trusted, accurate, and actionable predictions of rainy season onsets in Kenya
    • Schedule
    Notes
  • Online
    Online
    Set Timezone

Generic 'disconnected' Message
Author(s):
Chris Funk, University of California Santa Barbara (First Author, Presenting Author)
Mary Kilavi, Kenya Meteorological Department
Fredrick Sedah, Kenya Meteorological Department
Genevieve Flaspohler, Rhiza Research
Philip Ochieng, Kenya Meteorological Department
Andreas Fink, Karlsruhe Institute of Technology
William Turner, University of California Santa Barbara
Joshua Adkins, Rhiza Research
Paul Kucera, NCAR
Chris Shitote, FEWS NET
Shrad Shukla, University of California Santa Barbara
Johnson Sirengo, Climate Hazards Center
Laura Harrison, University of California Santa Barbara
Frank Davenport, University of California Santa Barbara
Barnali Das, University of California Santa Barbara
Daniella Alaso, The University of Alabama
Gregory Husak, University of California Santa Barbara
Alexander Jong, Karlsruhe Institute of Technology


Plain Language Summary:Rainfall gauges in countries like Ethiopia and Kenya can be blended with satellite gridded precipitation forecasts to provide timely high-resolution estimates of recent past precipitation (Fig. 1). These enhanced gridded data can also be used as the foundation for downscaled tailored national sub-seasonal forecasts (B in Figure 1), which leverage and contextualize information from global models (A in Figure 1). In addition to high spatial resolution, these forecasts align with the statistical distributions of the enhanced rainfall products. Thus, these forecasts and observations can be combined to monitor crop growing conditions in East Africa. Such combinations are already being used to improve food security-related crop projections for the GEOGLAM Crop Monitor and FEWS NET,. More recent research has used quantile-matched downscaled rainfall forecasts for a suite of six SubC dynamic models to improve crop water requirement predictions in East Africa. Downscaled forecasts and observations lead to local onset forecasts (C), which can be used by KMD to support improved advisories (D).



Scientific Discipline
Suggested Itineraries
Neighborhood
Type
Where to Watch
Main Session
Discussion