A Comparison of Two Approaches for Modelling Cassava (Manihot esculenta Crantz.) Crop Growth

• Published in: Annals of botany Vol. 85; no. 1; pp. 77 - 90
• Main Author: GRAY, V.M.
• Format: Journal Article
• Language: English
• Published: Oxford Elsevier Science Ltd 01.01.2000; Oxford University Press; Academic Press; Oxford Publishing Limited (England)
• Subjects: Allocation; Allocation, simulation, model, crop growth, cassava, Manihot esculenta Crantz; Biomass production; Branches; Carbon; Cassava; crop growth; Manihot esculenta Crantz; Modeling; Parametric models; Plant roots; Plants; Root growth; simulation; Stems; Allocation, simulation, model, crop growth, cassava, Manihot esculenta Crantz
• ISSN: 0305-7364; 1095-8290
• DOI: 10.1006/anbo.1999.0999

Two approaches for modelling the growth and development of cassava, Manihot esculenta Crantz, are described and evaluated. The two models differ only in the hypotheses accounting for storage root growth. In model 1, assimilate allocation to storage roots is governed by the combined Chanter's (1976: Mathematical models in mushroom research and production. PhD Thesis, University of Sussex, UK) growth equation; and in model 2 the spill-over hypothesis for assimilate allocation to storage root governs storage root growth. In both models, canopy photosynthesis generates the carbon substrate required for all growth processes. The growth rates of leaves, stems and storage roots are defined by growth equations subject to substrate saturation kinetics. A key feature of both models is that the growth demands of the stem, fibrous roots and storage roots are related to leaf demand rates. Allocation to stems and branches was modelled by means of a modified logistic growth equation which includes all the parameters and variables (number of nodes, internode lengths, stem density, stem modulus of elasticity and branch tensile strength) that define the limits of the load bearing capacity of the shoot's supportive structures. The correlation coefficients for determination of yield prediction for the models werer =0.898, P =0.0385 (model 1) and r =0.954, P =0.0117 (model 2). For a growth season of 290 d (after which leaf area index equals zero and crop growth ceases), both models simulate the sigmoidal transition from the lag to exponential phase of crop growth. Both models are equally well corroborated by observed data; however, model 1 has greater explanatory power.