LANDSCAPE FRAMEWORK TO PREDICT PHENOLOGICAL EVENTS FOR GYPSY-MOTH (LEPIDOPTERA, LYMANTRIIDAE) MANAGEMENT PROGRAMS

Ability to predict the timing of phenological events (herein termed ta rget events) is an important component of gypsy moth, Lymantria dispar (L.) management programs and integrated pest management programs in g eneral. Several simulation models have been developed that, in part, d emonstrate their validity for predicting events at individual location s. The framework described in this article extends the use of these mo dels to be able to make predictions (i.e., create maps) for heterogene ous landscapes. An algorithm is presented that can predict the time th at a target event will occur anywhere in a landscape using temperature , a digital elevation model, linked egg hatch and larval development m odels, and a linear function that relates elevation to the Julian date when a given target event will occur. The algorithm was validated wit h four data sets collected from Virginia, West Virginia/Pennsylvania, and Utah. Model predictions were satisfactory for the Virginia data se ts and differed significantly from those for West Virginia/Pennsylvani a and Utah data sets. Potential sources of error are discussed. Target event maps are presented that demonstrate how this landscape framewor k can be used in gypsy moth management programs.