Reanalysis of High-Resolution Weather Data for Post-Detection Modeling of Anomalous Plant Disease Events

NON-TECHNICAL SUMMARY: We will provide weather-based tools that will give agencies and specialists the ability to retrospectively analyze conditions that prevailed prior to detection for plant pathogens that are biosecurity risks and whose development and spread are influenced by weather (a forensic forecast of sorts). Three plant disease systems will be use: downy mildew on grape, citrus canker, and sheath blight of rice.
<P>
APPROACH: We will be accomplish this work by acquiring high-resolution weather information through mesoscale (medium scale) weather models that ingest refined (cleaned-up) global conditions prepared and archived by the Reanalysis Project, a multi-agency effort to allow retrospective forecasts from the very daily data used to forecast global weather for the past 50 years. We will produce weather information needed to drive disease models at a spatial resolution down to 100 m for the U.S, and 1000 m globally. We can provide high-resolution weather data for two days, two months, or two years ago; whatever is needed to reanalyze an anomalous pest event. We will test this tool on three prototype pathosystems: grapevine downy mildew in New York; Asian citrus canker in Florida; and panicle blight of rice in South Korea. Working closely with the National Pest Diagnostic Network, the Cooperative Agricultural Pest Survey, and two groups in the Animal and Plant Health Inspection Service, we intend to develop the ability to rapidly respond to a request for a reanalysis of weather data related to an anomalous disease event.
<P>PROGRESS: 2004/05 TO 2008/04<BR>
OUTPUTS: This project has not produced any major outputs at this time. Development of the high resolution weather data for post-detection modeling had become a more onerous task than originally planned and consequently the intended product is not available at this time. A full discussion is made within the Outcomes section of this report. A major secondary output of this work was the completion of a Surface Wetness Energy Balance (SWEB) model. This model is one of the core models used in the high resolution weather models. Since there is no meteorological standard for measuring surface wetness, even though it is a critical component of many disease models, we needed a physically-based algorithm to estimate wetness at a high resolution. The algorithm has been developed, tested and published, and has been used by other research groups who need to compute estimates of surface wetness. Although the inputs for the SWEB model are complex, they can be easily derived from mesoscale weather models. <BR> PARTICIPANTS: Robert C. Seem, Cornell University, Geneva, NY Kyu Rang Kim, former postdoctoral fellow, now Korean Meteorological Administration, Seoul, KR John Zack, Meso, Inc., Troy NY Joseph Russo, ZedX, Inc, Bellefonte, NY <BR> TARGET AUDIENCES: The target audiences are researchers and regulatory agency staff who need to rapidly acquire detailed weather information related to an anomalous disease event or outbreak. <BR>
PROJECT MODIFICATIONS: In the final year of the grant the subcontract with Meso, Inc, was deobligated and a new subcontract was established with ZedX, Inc. This action permitted the shift from the MASS mesoscale model to the WRF mesoscale model.
<P>
IMPACT: 2004/05 TO 2008/04<BR>
This project brings together several different resources and models that ultimately will aid decision makers and regulators when they are confronted with disease outbreak that could be due to a newly introduced organism, a new, more virulent strain of an existing pathogen, or an intentionally released pathogen designed to cause grave harm to crops. The system can simulate weather conditions in a grid pattern across any region for any period in the last 50 years. We have not achieved the original goals. The proposed work was risky, but the expected outputs were clearly needed. We needed to push the technology to its limit in order to achieve the goal. From a positive perspective, the technology improvements since the start of this grant have been very helpful; from a negative point of view, the technology is still not at a level where we can achieve all our goals. However, we have come a long way towards achieving the goals. In the course of this project we have achieve partial fulfillment of our goals and we have created the groundwork for the project to continue in a modified form to ultimately achieve the original goal. Intermediary success has been achieved in spite of some fundamental alterations of the original methodologies. These changes include replacement of the original mesoscale weather model (MASS) in favor of the more rapidly developing Weather Research and Forecast (WRF) model, and implementation of a different input database (the North America Regional Reanalysis data set). Also we were not able to achieve the necessary computing power to produce the optimal turnaround from the time a request of reanalysis data is made until the data are delivered. However, these alterations have caused the project to slow and delayed all of the expected outcomes originally proposed. The following intermediate outcomes have been achieved. Proof-of-concept that reanalysis data can be used to produce a reliable and accurate set of weather data which can be used to assess development and spread of pathogen after it has been introduce into a site. However, there are two barriers that must be overcome before full implementation can be achieved. These barriers are first, improved computing speed. The best speed we have been able to achieve to date is 28 hours of CPU time to generate a 24-hour dataset for a moderately sized geographic region. However, the final product must achieve better than parity when there is an emergency need for the data. The level of progress in this project has encouraged a private sector business (a subcontractor to this project) to proceed with the further development of this reanalysis project. Therefore, while the major goals were not met, the underlying ideas and principals were strong enough to undertake the work. This is the scenario originally envisioned by the project leaders as the only way to sustain the long-term availability of reanalysis data for biological event modeling. Although the evolution of the project has not progressed as originally planned, the ultimate goal will likely be achieve in partnership with the private sector.