Share

Ensemble Forecasting

The aim of this project was to provide modeling support for tracking and forecasting the spread of soybean rust.
The interpretation of model output by the team of PSU meteorology students led by Jeremy Zidek (CEAL MS student) and Dr. Paul Knight (Meteorology)

The interpretation of model output by the team of PSU meteorology students led by Jeremy Zidek (CEAL MS student) and Dr. Paul Knight (Meteorology)

The aim of this cooperative project is to provide state extension specialists, researchers, administrators and indirectly, North American soybean growers, related industry, and the public, with an unprecedented level of modeling support for tracking the spread of soybean rust during the 2006 and 2007 growing seasons. Participants include Scott Isard, Julie Golod, Paul Knight, Jeremy Zidek, and a team of summer intern students (Penn State University), Joe Russo and Joe Petrovich (ZedX Inc), and Ziatao Pan (St. Louis University).

The aerobiology ensemble modeling project, in a manner analogous to the tracking of hurricanes by the National Hurricane Center, follows the movement and development of SBR across the country by integrating and interpreting simulations from the three models: 1) the Integrated Aerobiological Modeling System (IAMS), 2) the HYSPLIT trajectory model (NOAA ARL), and 3) a climatological (historical) based model operated by a St. Louis-Iowa State University team of scientists. A team of PSU meteorology students led by Dr. Paul Knight, PA State Climatologist, with support from Joe Petrovich and Joe Russo of ZedX Inc. run the IAMS and HYSPLIT models on a daily basis and developed a number of unique weather products for PIPE users. Composite Precipitation/Relative Humidity, Solar Radiation/Minimum Temperature, and Wind Speed/Direction maps are created from NWS model output. SBR observations and IAMS model output are interpreted to make SBR transport, deposition, and disease severity “ensemble maps” as well as a product that combined these three maps. The student team integrates the composite weather maps with the “ensemble maps” to create an “SBR activity ensemble” that defined wait, watch and warning zones for field scouting and soybean rust decision support. Three times a week, the student team issues 1-2 day and 3-5 day forecasts in text format on the restricted access IPM PIPE website.

The interpretation of model output by the team of PSU meteorology students led by Jeremy Zidek (CEAL MS student) and Dr. Paul Knight (Meteorology) results in the desired “human-machine mix,” whereby the limitations of individual model simulations are overcome by comparing and integrating output from multiple models with derived weather products and observations. This type of ensemble approach is routinely used for interpreting output of the various meteorological models by weather services throughout the world because it achieves reliable forecasts at a greatly reduced cost in time and effort.

The ensemble modeling project contributes to the continued success of the PIPE during the 2006 and 2007 growing seasons. This project simulates the local development of soybean rust infections based on weather-driven transport of spores from infected geographic regions to downwind areas with potential host vegetation. The maps of spore movement and infection development both in the present and the future are invaluable in guiding extension specialists as to the likelihood, location, and timing of soybean rust in their states. The modeling work of the project PIs was featured in the “The Rust Invasion," part of the USDA CSREES CD titled Feeding America-Partners 17 and at the 2007 National Soybean Rust Symposium (see presentations below)

Isard, S.A. and J. Russo, Application of the Integrated Aerobiology Modeling System to soybean rust forecasting in 2006. Proceedings of the National Soybean Rust Symposium, November 29-December 1, 2006. St. Louis, MO. [Presentation]

Pan, Z., Forecasting the spread of soybean rust using a nested climate model. Proceedings of the National Soybean Rust Symposium, November 29-December 1, 2006. St. Louis, MO. [Presentation]

Petrovich, J., Forecasting the spread of soybean rust using the ensemble HYSPLIT approach. Proceedings of the National Soybean Rust Symposium, November 29-December 1, 2006. St. Louis, MO. [Presentation]

Russo, J., S. Isard, and P. Knight, Ensemble approach to soybean rust forecasting for the 2006 PIPE. Proceedings of the National Soybean Rust Symposium, November 29-December 1, 2006. St. Louis, MO. [Presentation]