Modeling and validation of high-temperature induced spikelet sterility in rice

Abstract

Spikelet sterility due to heat stress during reproductive stage would be one of the prominent factors to reduce rice yield under the projected global warming climate. A simulation model that was comprised of equations to estimate the probability distributions of heading date of panicles in the field, flowering date of spikelets on a panicle, and flowering time of spikelets during the day time and the two sterility response functions to the temperature on the day of meiosis and at the flowering time of a spikelet was constructed, calibrated, and validated to predict high-temperature induced spikelet sterility in rice.The model was calibrated and validated against the data collected from a series of experiments conducted in the plastic houses controlling the temperature to ambient, ambient+1.5°C, ambient+3.0°C, ambient+5.0°C, and ambient+7.0°C in 2009 and 2010. Heading, flowering habit, and spikelet sterility of rice exposed to different temperatures during two periods from 20 days before heading and initial heading to 20 days after heading were recorded on panicle basis for model calibration. For model validation spikelet sterility data were collected on hill (pot) basis from rice plants exposed to different temperatures from transiting to maturity in 2009 and 2010.The heading of panicle reached peak on four to six days after and lasted until 12 days after initial heading and the heading date distribution of panicles was well fitted to Poisson's equation. The flowering peak of spikelets on a panicle occurred at about 5 days after heading and lasted 11 days regardless of temperature and cultivars and the flowering distribution was well fitted to the normal distribution function. The anthesis in a day started from 8:00h, reached peak around 11:00h, and lasted until 15:00h regardless of temperature treatments, the flowering peak being a little earlier in a japonica cultivar “Hwaseongbyeo” than in a Tongil type one “Dasanbyeo”. The flower opening time followed the normal distribution with standard deviation of about one hour. The fertility responses to high temperature on the day of spikelet meiosis and at the time of spikelet anthesis were well fitted to logistic functions of heating degree hour above 31°C on 12 days before spikelet anthesis and air temperature at the time of spikelet flowering, respectively. The heating degree hour causing 50% spikelet sterility at meiosis was higher in a japonica cultivar “Hwaseongbyeo” than in a Tongil type one “Dasanbyeo”, while the air temperature causing 50% spikelet sterility at spikelet flowering time was higher in Tongil type cultivars than in japonica ones.The model integrating the above equations predicted the spikelet sterility/fertility response to air temperature with reasonable precision and accuracy.