Different crop models including MAIZE Ceres, STICS and other approaches have been used to simulate leaf area index (LAI) in maize (Zea mays L.). These modeling tools require genotype-specific calibration procedures. Studies on modeling LAI dynamics under optimal growth conditions with yields close to the yield potential have remained scarce. In the present study, logistic and exponential approaches have been developed and evaluated for the simulation of LAI in maize in a savannah region of the DR-Congo. Data for the development and the evaluation of the model were collected manually by non-destructive method from small farmers' field. The rate of expansion of the leaf surface and the rate of change of leaf senescence were also simulated. There were measurable variations among sites and varieties for the simulated height of maize plants. At all sites, the varieties with short plants were associated with expected superior performance based on simulation data. In general, the model underestimates the LAI based on observed values. LAI values for the genetically improved maize varieties (Salongo 2, MUS and AK) were greater than those of the unimproved local variety (Local). There were significant differences for K, b, T i , LAI, T f , and parameters among models and varieties. In all sites and for all varieties, the growth rate (b) was higher, while the rate of senescence (a) was lower compared to STICS estimates.

No abstract