Comparison of WOFOST, CropSyst and WARM for Simulating Rice Growth (Japonica Type - Short Cycle Varieties)

Rice is one of the most important crops worldwide and during the last 15 years crop modellers have pointed their attention on this crop. In this paper, three models are compared for the simulation of rice (Japonica type - short cycle varieties) growth under potential conditions for water, nutrients, pests and diseases. The three models simulate crop growth using different approaches: WOFOST is based on the photosynthesis approach, CropSyst and WARM respectively on the concepts of transpiration and radiation use efficiency. Each of these approaches represents crop growth with a different level of detail and has different requirements in terms of data needs. Data used for models calibration and validation were collected in northern Italy between 1990 and 1996. The three models showed similar accuracy in reproducing crop behavior, both during calibration and validation. Relative root mean square errors ranged from 6.9% to 22.6%, from 9.9% to 18.0%, and from 10.6% to 17.5% respectively for WOFOST, CropSyst and WARM. Similar levels of reliability were reached in spite of the different modelling approaches and levels of detail in the description of the processes related to biomass accumulation; partitioning and leaf area index simulation. The WARM model appeared the easiest to be used because of the lower number of parameters describing cultivars morphology and physiology and because of the absence of parameters purely empirical.

CONFALONIERI Roberto;  GUSBERTI Davide;  ACUTIS Marco; 

2007-05-24

Associazione Italiana di Agrometeorologia

JRC35590