A model to predict diurnal pollen shed in maize

We have developed a mathematical model to predict the diurnal pattern of maize (Zea mays L.) pollen shed on the basis of local meteorological conditions. Our goal is to improve simulations of maize pollen dispersion that have typically released pollen at a constant rate in contrast with measurements of pollen shed that show diurnal variation in the rate of shed. Measurements coupling pollen shed and local meteorological variables were made during controlled experiments and a 2004 field experiment to examine the influence of meteorological conditions on pollen shed. From these data, a model was developed to predict the diurnal pattern of pollen shed as a function of vapor pressure deficit, solar radiation, temperature, and the amount of pollen remaining to be shed. The model was validated by predicting the rate of pollen shed, normalized by the daily total of pollen shed, that occurred hourly for days during a 2003 field study (RMSE = 0.061 h^-1) and results from van Hout et al. (2008; RMSE = 0.089 h^-1). The model captured the general trend of pollen shed and predicted the time of peak shed within an hour of the measured peak on most days. The model, however, tended to underpredict the magnitude of the normalized peak rate of shed and did not account for secondary peaks in pollen shed that were occasionally observed. Thus, future model refinements will depend on identifying additional biological or environmental factors that impact the instantaneous rate of pollen shed.