Estimating sugarcane root length density through root mapping and orientation modelling

Root length density (RLD) is a key factor in crop functioning. A field method was developed to quantify RLD of sugarcane from root intersection density (RID) taking root orientations into account. RIDs were observed on three perpendicular soil planes and RLD was measured for the enclosed volume. RID and RLD of thick and fine roots were measured separately. These measurements were replicated at different ages and sites to test models describing RLD according to RID. Fine roots were nearly isotropic and thick roots had a preferential orientation, i.e. horizontal near the surface and becoming progressively vertical in deeper horizons. Relationships in thick roots were modelled according to CO_t: RLD_t = RID_t. CO_t (CO_t: root orientation coefficient, ranged from 1.3 to 4.9 for thick roots). For fine roots (_f), CO_f?=?2. This theoretical model led to differences between measured and calculated RLD. The ratio between measured and calculated RLD_f (CE_f) increased from 1 to 3 with RID_f. CE_f was introduced as an additional coefficient in the model: RLD_f?=?2. NI_f. CE_f. Intermediate results were obtained for all (_a) roots: CO_a and CE_a were both dependent on RID_a, therefore: RLD_a = NI_a. CO_a. CE_a. The models were validated with independent datasets from Brazil and France. These allowed a more robust prediction of RLD than direct regressions between RID and RLD. They may estimate RLD from RID in soil profiles by root mapping while taking RLD spatial variability into account.