Rethinking a cassava crop model
The week from 7th -10th of August a meeting was held at CIAT headquarters between crop modelers and physiologists from CIAT, and the maintainers and developers of the DSSAT-CROPSIM-CASSAVA crop model with the purpose of discussing the improvement of the current cassava model into DSSAT and set up strategies towards the development of a robust crop model applicable under a variety of scenarios, including climate change.
Great strides were made during the gathering where a new cassava module was created within the DSSAT umbrella out from the original CROPSIM-Cassava module.
Additionally, a comprehensive analysis of cassava growth and development was done, which helped to define which processes need to be included and how, whilst at the same time several knowledge gaps that still exist on cassava physiology were highlighted.
A solid team for development that included the two invited external experts and developers of DSSAT-Cassava (Tony Hunt and Gerrit Hoogenboom) was put together and follow-up tasks were assigned to each of the parties to advance forward on the model improvement, activities of data collection and model evaluation for this year and the next.
A stronger collaboration was thus built and roles were defined for each of the institutions and researchers, with CIAT taking a leading role in the model development and evaluation. Further steps on model development will include (1) the final cleaning of the current cassava module; (2) the analysis of species, ecotype and cultivar coefficients; (3) the incorporation of any new required routines to account for processes not already in DSSAT; and (4) a final phase of model evaluation.
Flexibility and easy adjustment are the interesting issues that the DSSAT cassava model is projected to include. We are trying to bring together other researchers working with cassava modeling, and these include the APSIM modeling group at CSIRO, and the “Thai” modeling group, both of which we expect to take advantage from our efforts and data (while also sharing theirs), and work towards the strengthening of our cassava crop modeling consortium. Most principles and parameters used to develop the model would be made available, so that other modelers can incorporate them in their models if they so wish.
Further, feature an open source scheme model which can be used and sustain in a long term is also expected. Within this context we noted that the cassava model can form the basis for the development of models for other root and tuber crops (e.g. yams and sweet potatoes) and possibly also for bananas and plantains, contributing both the cassava specific module and the general DSSAT modules.