1-2 of 2 results
-
Using Machine Learning to Improve Forecasting of Deep Convection
PI Christopher Hennon
CO-I Ronny Schroeder
CO-I Curtis James
CO-I Abd AlRahman AlMomani
We are working to train a neural network to forecast the initiation time, location, and intensity of thunderstorms. These results will support operations during the proposed CONVECT project and could ultimately aid operational forecasting during the North American Monsoon (NAM).
This research, funded through an NSF EAGER grant, seeks to improve forecast accuracy of monsoon thunderstorm activity and precipitation amounts in the Southwest. The project creates an innovative machine learning tool trained using regional numerical weather model output and satellite remote sensing data (the predictors) with respect to known thunderstorm cell locations and intensities detected by radar (the targets). The tool will be designed to extract important fundamental relationships between the predictors and targets that help explain the development and evolution of thunderstorms. After an intense training, validation and testing phase, the relationships will then be leveraged to generate better forecasts of the timing, severity and location of future thunderstorm events in the Southwest. The tool will be shared with the National Weather Service tto help forecasters predict thunderstorm-related hazards such as large hail, flash flooding or wildfire ignition. This innovative approach will also provide a framework for improving operational meteorological and geophysical prediction systems and for guiding scientific field studies.
The project develops a probabilistic model to predict convective initiation, rain rates, and convective cell tracks during the wet phase of the North American Monsoon (NAM). Predictors of convection (e.g., relative humidity, convective available potential energy, precipitable water) will be collected from dynamic mesoscale model (High Resolution Rapid Refresh, University of Arizona-Weather Research Forecast model) analyses and forecasts and combined with new satellite-derived observations of soil moisture and surface temperature to produce a unique prediction tool. A novel machine learning approach – causality informed learning – will be applied to identify the most suitable predictors for further training in a neural network and to gain insight into the processes governing convective initiation and evolution. Hourly forecasts of precipitation occurrence, nature, and categorical rain rates will be produced operationally to guide forecasters and field research.
Categories: Faculty-Staff
-
CONVECT (Convective Organization aNd Venting Experiment in Complex Terrain)
PI Curtis James
CO-I Ronny Schroeder
CONVECT is a major meteorological field research project being proposed for July - August 2025 in north-central Arizona. The project is aimed at improving our understanding and ability to predict the convective development and organization of boundary layer thermals, thunderstorms, and mesoscale convective systems during the North American Monsoon (NAM).
The proposed field project is the Convective Organization aNd Venting Experiment in Complex Terrain (CONVECT), focused in north-central Arizona near the city of Prescott. This targeted region, encompassing the Black Hills, Verde and Prescott Valleys, and Mogollon Rim, provides an ideal laboratory for investigating processes connecting complex terrain to boundary-layer and convective processes. During the summer monsoon season, this region experiences frequent deep, precipitating convection. These storms typically initiate over the most prominent terrain features in this region and then may propagate into the populated lower lying areas or send out density currents or buoyancy bores that subsequently initiate new convection. The thunderstorms are generally spatially localized, forming over a deep convective boundary layer, but are often associated with pulse severe conditions (damaging wind gusts or large hail). Some cells may become terrain-locked or exhibit back-building behavior, leading to intense rainfall and flash flooding.
The multi-scale approach proposed for CONVECT will, for the first time, capture the complete physical chain of land-atmosphere processes that drive water vapor transport and monsoonal precipitation over complex terrain at meso- to micro-scales. This diurnally cycling chain includes energy and moisture exchange over a heterogeneous, sloping surface, thermally-driven planetary boundary layer (PBL) circulations, the venting of PBL air into the free troposphere, and the initiation, upscale growth, and propagation of deep convection. The proposed deployment includes a dense network of surface flux and energy balance probes, lower-tropospheric thermodynamic and kinematic profiling systems, mobile radars, and crewed and uncrewed aircraft with in-situ and remote sensors. The campaign will be carefully guided by multi-scale modeling, and in turn, experimental observations will be assimilated to evaluate their impact on multi-scale predictability and the validity of surface layer and PBL parameterizations in complex terrain. The CONVECT science team of instrument scientists and numerical modelers contains the necessary, complementary expertise in the surface layer, the boundary layer, and deep convection to substantially advance understanding of mountain exchange between the surface and free troposphere, as well as extreme precipitation, through a multi-scale lens.
Categories: Faculty-Staff
1-2 of 2 results