A new method of simulating dry matter partitioning in the Canegro sugarcane model

Crop simulation modelling is widely accepted as a useful tool for research and management purposes in agriculture, also in sugar industries worldwide. Successful model application depends on accurate predictions of stalk and sucrose yield. The literature shows that the simulation of dry matter partitioning in sugarcane models such as the Canegro, model could be improved by using source-sink concepts. This should lead to a better understanding of the processes involved and could lead to improved predictions of sucrose yield. With these aims in mind a new dry matter partitioning model was developed and incorporated into the Canegro model, and tested against independent experimental data. In the proposed model, daily partitioning of assimilate between roots and aerial parts is simulated as a non-linear function of total biomass. A constant fraction of aerial dry mass is partitioned to stalk when thermal time since emergence exceeds a given value. The rate of dry matter partitioning to stalk is regarded as the source strength. Partitioning of stalk dry matter is regulated by sink capacity and the source to sink ratio. Sink capacity is dictated by current growing conditions, current stalk mass and varietal characteristics. The sucrose accumulation component of the model is based on a framework of sucrose distribution within stalks as it is affected by temperature and water stress. Values of model parameters for cultivar NCo376 were determined from the literature and by model calibration on data from an experiment conducted at Mount Edgecombe, South Africa. The new and old method of simulating dry matter partitioning was compared to observations from 24 NCo376 crops grown at Umhlanga, La Mercy and Pongola, South Africa. The new method performed better than the old method in all respects. The RMSE of aerial dry mass, stalk dry mass and sucrose mass predictions was 6.94, 5.48 and 2.6 t/ha respectively. These values compare favourably with other models in the literature. The results show that the new method is sufficiently robust to accurately predict aerial dry mass, stalk yield and sucrose yield for a wide range of conditions. It is also capable of simulating subtle changes in partitioning associated with shifts in temperature and water stress conditions. It provides a good platform for further refinement of sugarcane dry matter partitioning such as the inclusion of chemical ripener, nitrogen and variety effects. (C) 2002 Elsevier Science B.V. All rights reserved.