Stochastic dynamic simulation of fruit abortion: a case study of sweet pepper

Abortion of reproductive organs diminishes yields in many crops. In indeterminate greenhouse crops, alternating periods of fruit abortion and fruit set exist, resulting in fluctuations in fruit yield. Factors affecting the level of abortion are e.g., the supply and demand for assimilates (source and sink strength, respectively), temperature and cultivar. However, simulation of fruit abortion is still a weak part of crop simulation models. Variation in fruit abortion exists between plants, which results in differences in the timing and the number of set fruits. Therefore, simulating fruit abortion with variation could give more realistic simulation results. The probability of a fruit to abort should be related to factors like source strength and sink strength. The more favourable the circumstances are for fruit abortion, e.g., low source strength or high sink strength, the more likely it is that the fruit aborts. Survival analysis estimates parameters quantifying the influence of explanatory variables on the abortion rate. Time-varying explanatory variables can be used in the analysis. In a case study, we used survival analysis to analyse a data set with observations on flowering, fruit abortion and fruit harvest for sweet pepper. Source and sink strength were used as explanatory variables. The resulting equation determining the probability of abortion per day was implemented in a simple simulation model to simulate fruit set. The model output, as an average of 100 plants, showed similar timing in the fluctuations in fruit set as the observations, although the amplitude of the fluctuations was in some cases underestimated. The percentage fruit set was simulated correctly