Home My Dashboard About Us News Grower's Guide Plant Pest Guide Forecasting Models Expert System Weather IPM Dosage Optimization Agricultural Researchers Farmers Home Gardeners Policies Support Donate Contact Us

Forecasting Models

Accurate forecasting of the arrival of pests has a unrealized potential both economically by reducing pesticide usage and socially by reducing crop losses in situations where pesticides are not economically feasible or practical (Esker, Sparks et al. 2008). There are economic costs associated with false positives, disease prevention measures when unnecessary, and false negatives, disease prevention measures not applied when needed (Dewdney, Biggs et al. 2007). Existing mathematical and statistical modeling techniques have their advantages and disadvantages (Contreras-Medina, Torres-Pacheco et al. 2009) such that different methodologies need to be tailored suited to a particular plant pest. They range from scoring techniques, to single variable thresholds, to simple mathematical curves, to area under the curve techniques, to differential mathematical equations, to multivariate statistical analyses, to complex computer simulations (Fry and Fohner 1985). In some cases, a simple plant pest model does the job perfectly fine. However, most plant pest management problems are rather complicated and a more complex approach is better suited. Models are based upon an understanding of the relationships between a plant pest agent, its plant host, the environment, and in some cases a transmission vector (Agrios 2005).

While there has been some success in reducing pesticide use by a number of specific plant pest models, not enough of these models have been developed and even fewer have been implemented in the field. Furthermore, the delivery of these methods have primarily consisted of expensive single model weather station programmed systems, costly single model desktop software solutions, and time consuming manually entered spreadsheet based models. Finally, the only such solutions available to home gardeners, a population that is responsible for an estimated 20% of overall pesticide (Environmental Protection Agency 2007), are rudimentary web based pest warning systems that provide very little practical usage. Virtually all of the plant disease forecast models in existence were developed in isolation for a specific plant disease on a single host crop. It is essential that any unified computer system of plant disease forecast models deliver their results in a consistent manner that is readily understandable to the user. Otherwise, it is likely that farmers, home gardeners, and even agricultural researchers might misinterpret forecasts from different models and make poor pest management decisions.

The most comprehensive system for plant pest model information to date is Pest Cast developed by the State of California (University of California Statewide IPM Program and California Environmental Protection Agency Department of Pesticide Regulation 2013). While this project provided an excellent starting database for environmental based forecast models of plant pests, it was restricted geographically to California plant pests. The project has severely languished since its initial funding grant expired and only the weather data collection component of the project continues. For most of the models, only the mathematical formulas and links to published research are available. For the models where there is a software implementation, the delivery mechanism is not consistent and instead consists a confusing mishmash of web entry forms, spreadsheets, and desktop database solutions. Furthermore, the only weather data integrated with the system is for the State of California and the system of weather stations is not as comprehensive as those used for NOAA National Weather Service forecasts. Clearly their vision was never fully realized.

DeMilia Research is taking the initiative to rejuvenate research on plant pest data collection and modeling. We are scouring the agricultural literature for plant disease forecast models to be included in DemiAg then organizing it to create a searchable database of Forecasting Models. This information will also be organized into a categorized index here that can be referenced from within the DemiAg expert system, as well as our Grower’s Guide and a Plant Pest Guide. Please let us know if you uncover other information for addition to the Plant Pests Models Database. Eventually, forecasting models for other plant pest models will be added. Details about the implementation and usage of the models into the DemiAg expert system will also be given.

Public domain and freely distributable copyrighted content from other sources will be added manually to the Plant Pests Models Database with appropriate source credits. Restricted use copyrighted and non-distributable content from other websites cannot be reproduced, though, links to them will be provided. A vast collection of agricultural resources is available from the U.S. Cooperative Extension Network with information specific to each state based upon research conducted by each Land Grant University. Direct links to relevant Cooperative Extension publications will be customized based upon your location when you are logged into your personal dashboard. DeMilia Research also will be writing original material for the Plant Pests Models Database. We will be supplementing the content and our original material with photographs, graphics, and video taken on the DeMilia Research Farm.

References Cited:

Coming soon. We anticipate having the database operational during 2017.

 
Home My Dashboard About Us News Grower's Guide Plant Pest Guide Forecasting Models Expert System Weather IPM Dosage Optimization Agricultural Researchers Farmers Home Gardeners Policies Support Donate Contact Us
 
Please contribute to the agricultural research conducted at DeMilia Research LLC by making a donation. If you notice any improperly used copyrighted information then contact us immediately!
 
Copyright © 2017 DeMilia Research LLC All Rights Reserved