Oregon IPM Center Newsletter - Vol I Iss 2 - Summer 2020 - Phenology Tools

IPM Insider Spotlight: 

USpest.org provides weather-based prediction tools nationwide 

Chris Hedstrom and Len Coop, Oregon IPM Center

For nearly 25 years, Len Coop of the Oregon IPM Center has been developing and managing USPest.org, a multi-species phenology modeling platform designed to aid management decisions of extension and agriculture consultants. What began in 1995 as a CD-based software program for Oregon mint growers has expanded into a nationwide resource featuring weather and climate-based tools including pest-specific degree-day calculators, customizable degree-day maps for all of the United States, and disease risk and alerts for multiple crop diseases. While much has changed since the early days, providing a free, comprehensive website for accessing weather and climate models of all types to support IPM decision-making have always been at the core. Today, USPest.org continues to be a key IPM resource for growers across the nation.

The apps at USPest.org have recently been optimized for mobile devices

Long before Google was a thing

In 1995, after wrapping up a postdoc project working on grasshopper millet damage in Africa, Len contacted Marcos Kogan, then-director of IPPC, and proposed a decision aid software program for Mint growers called IPMP (Integrated Pest Management on Peppermint). After completion of this PC software (distributed by CD) by 1996, Len considered a different outlet. At the time, many people, including tech-savvy growers, were starting to use this new thing called the World Wide Web. It’s worth taking a moment to think back to the internet of 1996. Most people were still using dial-up modems and AOL, the top search engine was Yahoo! (Google wouldn’t be founded until the next year), and a 28.8 kbps modem connection was pretty fast (a basic cable connection today is about 1,500 times faster). Recognizing the potential of the new technology, Len created a web version of IPMP, (which is still online!) and began the prototype for USPEST.org. Back then, he was taking over some of the work others in the Entomology Department had started using dial-up bulletin board access to weather data and degree-day accumulations. This work was started by his previous supervisor, Entomology Professor Brian Croft, and his programmer, Kevin Currans, who had by now been diverted to do the initial networking of all the buildings on OSU campus for the College of Science.

Len had been working with Hood River Entomologist Helmut Riedl, and was interested in creating a tool for pest phenology prediction in the different tree fruit growing regions of the state. For a degree-day based tool, three basic components were needed: access to a robust and local weather network, accurate phenology models of the pests, and a website that users could access to run the models as they wished. Like many projects, there was no initial funding.  Len worked to create a prototype that could later be used to justify the grants necessary to expand. One of the first tools that Len and programmers developed was an online phenology and degree-day calculator that used available weather data to predict specific events in the life of a pest, such as egg hatch or adult metamorphosis. This information could then be used to help inform the timing of management actions such as when to monitor or apply insecticide treatments, potentially increasing efficacy and reducing unnecessary sprays. Since its creation, the online tool has become one of the most important tools on USPest.org, with over 98,000 models runs in 2019 and over 800,000 since 2000. The calculator is accessed by a variety of users, such as OSU Extension, PNW Pest Alert Network, Northwest Berry Foundation, and Oregon Department of Agriculture



Putting the data where they're needed

Though it all seems simple, the science behind the tools is complex. Where do all the different pieces come from? The underlying component on which all the models depend is accurate and timely weather data. The long-standing regional agricultural weather network that the system first began with is Agrimet, run by the U.S. Bureau of Reclamation, which remains perhaps the most important and most reliable network throughout the Pacific Northwest.  Today, the data originating from Agrimet and over two hundred other networks such as METAR, RAWS, and APRSWXNET all come through MesoWest, a publicly accessible weather consolidator, with a current total of over 32,000 weather stations. Data from these networks are used by the National Weather Service to aid in weather forecasting, and by researchers and other agencies for weather research and development of tools. These nationwide data points are typically updated every 15 to 20 minutes. One of the challenges over the years has been getting reliable data from private networks that come and go or are increasingly expensive. Len and his colleagues are working to move growers in the Hood River and The Dalles from private networks to the APRSWXNET network, which will allow more time for model additions and features and requires less time for custom data acquisition programming.

The time it takes to develop new and updated phenology models for different pests also requires a fair amount of analysis. Many studies only follow the development of a certain life stage, so the data have to be combined to be able to fit into an annual prediction pattern. Len analyses data from multiple sources, published and unpublished, to create a model that can cover the entire calendar year. The process and the model parameters are posted for review and feedback. The on-line phenology modeling program initially had just a handful of models, such as codling moth and pine shoot moth, but has grown steadily to the current number of 132 models of IPM and invasive insects, crops, biocontrol agents, plant diseases and weeds. Some of the more recent models include bronze birch borer, honeydew moth, tomato leafminer, and CROPTIME models for sweet corn.

Tools of the future

More tools are nearly ready for public testing. Web-based disease risk apps, developed by Len and Dan Upper, a longtime programmer of USPest.org, have been refined for mobile use. These apps provide a risk level based on location and date for fire blight, apple scab, hop powdery mildew, and more. They have also added a notification option that allows users to customize email alerts at a desired frequency, say, every other morning. Len and Dan are currently looking for users to beta-test these programs and hope to have them widely available by the end of the year.

Screenshots from one of the disease phenology risk prediction apps at USPest.org

Len and Brittany Barker of the Oregon IPM Center have been working with the National Plant Phenology group to further develop the idea of freely available, up-to-the-hour pest prediction tools. A new open-source, temporal, and spatial prediction tool called DDRP (Degree-Day Risk Prediction) is entering its fifth year of development. This new program differs from the tools at USPest.org in that it outputs mapped predictions of pest events using current forecast data. It can also account for generation spread, and up to four life stage events per generation for up to four generations per year. Areas can be shown for the likelihood of a life stage event, such as the likelihood of completion of the first generation of emerald ash borer in a specific county. Brittany’s paper describing DDRP is currently in press and available online now


How-to: Using the online phenology model app

Access the apps: http://uspest.org/dd/model_app

  1. Start in the “Station” tab (there are five tabs along the top of the program). Select a weather station, by searching for a city and state or a ZIP Code. Click on the “Map” link if you would like to see the available stations on a Google map. Either select a station from the list or by clicking on a pin in the google map.
  2. Once a location is chosen, the user taps the “Model” tab to select one of the 130+ available models. By selecting a model, the program automatically fills in the parameters, but users can use the generic calculator option and also enter their start dates and max and min development temperatures.
  3. At that point, it’s as simple as tapping on the “Output” tab to get a prediction of life stage events for that organism. Users can see how the degree-days compare to the previous two years and the 30-year normal. Users can jump back to the station tab to try nearby weather stations and compare results, or in the case of some insects, try a different version of the model.

Because it’s web-based, the calculations are done server-side and can be run anywhere there is internet access. It was recently formatted to run on mobile devices for access in the field and an app version is currently at the Android Play Store and is awaiting conversion to the Apple app format.


Bonus: Want some more fun history? Check out the original Degree-Day Calculator interface and the old home page 1997 


Back to the newsletter


This article appears in Oregon IPM Insider, Vol 1 Iss 2, Summer 2020.