Simulation of dry matter accumulation, partitioning and yield prediction in processing tomato (Lycopersicon esculentum Mill.)

Simulation of dry matter accumulation and the distribution of crop growth is an important means of predicting yield. In this study, a variety of cultivars of the processing tomato (Lycopersicon esculentum Mill.), sowing dates and densities were tested. By analyzing the quantitative relationships between growth and physiological development time (PDT) based on a knowledge model, dry matter partitioning index and harvest index (HI), mathematical models were developed to estimate total dry matter accumulation, shoot dry matter partitioning and yield for processing tomatoes. A sowing date factor (SDF) was introduced to regulate partitioning intensity and the genetic features of the cultivars tested were considered. Validation of the results using trial data gained from studies with different cultivars and from sowing experiments showed that total dry matter weight was predicted correctly (RMSE < 1313 kg ha, RE < 27.70%). Above ground biomass was predicted with an RMSE < 416 kg ha for stems, < 517 kg ha for leaves and < 545 kg ha for the fruit. Yield, defined as ripe fruit dry matter accumulation, was also accurately predicted (RMSE and RE were 7737 kg ha and 15.51%, respectively). This model can be used to predict different production levels and yield objectives for processing tomatoes planted in arid areas of China. Further improvements to the model are discussed.