THE ROLE OF DAILY MINIMUM TEMPERATURE IN MODULATING THE DEVELOPMENT RATE TO FLOWERING IN SOYBEAN

Time of flowering for soybean (Glycine max L. Merr.) is cultivar speci fic and depends on temperature and night length. The objective of this research was to develop a simple model where a single set of paramete rs accurately predict flowering date across diverse environments. The hypothesis that low daily minimum screen-air-temperature (T-min) is as sociated with a delay to flowering by lowering the daily potential rat e was tested for the cultivar Forrest with phenology data from South A frica and North America in which average daily-minimum air temperature from sowing to flowering ranged from 10.5 to 25 degrees C. The effect of T-min was also evaluated with 13 cultivars from the South African National Soybean Cultivar Trials. Daily development rate was assumed t o be determined by a multiplicative relationship between air temperatu re, night length and T-min. The date of flowering was determined as th e sum of daily rates of development until a threshold is reached. The night-length function is a linear plateau with a minimum night length, below which the rate is zero, and an optimum night length above which the rate is 1.0. The hourly temperature function is a three segmented spline function where the rate is 0.0 below the base temperature, the rate increases from 0.0 to 1.0 at a lower optimum and decreases from 1.0 at a higher optimum to 0.0 at a maximum temperature. The minimum t emperature function is a linear plateau that consists of an optimum te mperature above which the rate is 1.0 and a linear slope that decrease s with decreasing minimum temperature. An optimization procedure, the downhill simplex method, determined the direction of the search and wa s used to estimate parameters. The function for the effect of T-min on development rate reduced the root mean square error (RMSE) from 9.8 t o 6.4 days for the cultivar Forrest when using parameters reported in the literature with 162 observations from North America and South Afri ca. Daily minimum air temperature was associated with delay in develop ment rate to flowering. The effect of T-min appears likely to be assoc iated with canopy-air temperature differences. When T-min was included as a separate effect in the phenology model, consistency of parameter s for night length was improved and model error was decreased.