Root growth of cauliflower (Brassica oleracea L. botrytis) under unstressed conditions: Measurement and modelling

Root observations were carried out on cauliflower using the minirhizotron a nd the soil core method in two years on two locations with different soil t ypes, a loess loam and a humic loamy sand. Total root length (RL) (cm cm(-2 )) of cauliflower was correlated to total shoot dry weight (W-sh) (g m(-2)) RL=0.0124(+/- 0.005)s*W-sh, r(2)=0.76. There was an acceptable correlation (r(2)=0.88) between the minirhizotron and the soil core methods for the su b-soil data, whereas the minirhizotron method underestimated rooting intens ity for the top soil. Changes in rooting depth over time could be described for both soil types using a segmented function of temperature sum, consist ing of an early exponential and a later linear phase. The increase of rooti ng depth during the linear phase was 0.107(+/- 0.01) cm degrees C-1 d(-1). A simple descriptive root growth model based on the assumptions of a negati ve exponential decline of root length density (RLD) with soil depth, of a f ixed ratio of RLD at the top of the soil profile and at rooting depth (r(RL D)) and of a fixed fraction of dry matter increase allocated to fine-roots (f(fR)) was formulated and used to describe the temporal and spatial variat ion of RLD found in the field. Slightly different estimates of f(f)R and of r(R)LD could be found for the different soil types, indicating a higher fr action of fine-root dry matter for the loess loam soil and a somewhat deepe r root system for the humic loamy sand soil. A cross validation using the p arameter values obtained from adjusting to the rooting data of one soil typ e for predicting RLD values of the other soil type, however, indicated that still quite satisfactory estimates (r(2)=0.91 and 0.95) of RLD could be ob tained.