Managing Weather Data
Late blight is responsive to changes in weather so measuring weather can be an important tool in disease assessment. The variables that seem most important are temperature, rain and relative humidity, but solar radiation and wind are also important, although there is not much information on how they affect blight.
There has been a virtual revolution in Automatic weather loggers in the past few years. There are now very small and inexpensive weather loggers that collect temperature, relative humidity and other variables at sub-hour frequencies. We use Hobo data loggers by Onset (not an endorsement, they simply work for us); this is described in more detail below. A number of other options are available including more complete weather stations. Information on these can be found in the Internet.
Hobo data loggers are made by Onset computer corporation http://www.onsetcomp.com . The sensors should be put in a radiation shield (available from Onset @ 65 usd). Shields can also be constructed from local material but they should not restrict air movement. Data can be downloaded directly to a PC or via a "Hobo Shuttle" (@ 159 usd). You can use software provided by the manufacturer, but other logger software may also work. We have not had a lot of luck with the shuttles and tend to use computers. For our purposes, a hobo holds sufficient data for one season, so downloading during the season is not necessary but it is a good security measure.
Most late blight forecasting models use hourly weather data, or daily weather data which can only be derived from hourly observations. This type of weather data is typically obtained using automatic weather loggers. CASTOR allows for importation of this type of data (using text files of different formats). CASTOR stores it in a standard format that can be read by other programs (like spreadsheet programs). CASTOR calculates daily, 10-daily and monthly averages or totals.
The weather data can then be used to run late blight forecast models. CASTOR implements a growing number of published forecast models that predicts the outbreak of late blight, and recommend timing of fungicide applications. The forecast models currently included were all implemented according to their description in the Potato Late Blight Disease Model Database, which was compiled by Dr. Janet C. Broome and coworkers of the U.C. Davis, California Statewide IPM project (http://axp.ipm.ucdavis.edu/DISEASE/DATABASE/potatolateblight.html). Before using a model that was not field tested or validated for a specific location, the model should be tested for one or more seasons under local conditions to verify that it will work in your location.
You can use CASTOR to verify the validity of these models in your location, and apply them if you are confident in their validity for your circumstances. The source code (in DELPHI) is also free If you want to continue developing or adding your own model to CASTOR.
To download the CASTOR please read the Manual (.pdf ) first