Early hyperspectral diagnosis of canopy water content for productivity-risk and water-management screening in rainfed winter oilseed rape

Rainfed winter oilseed rape is vulnerable to temporal mismatches between precipitation and crop water demand, creating a need for early indicators of water-related productivity risk. This study developed a hyperspectral framework to estimate canopy water-related traits and assess the management relevance of canopy water content (CWC) under mulching and nitrogen treatments. A three-year field experiment (2020–2023) combined three mulching patterns with five nitrogen rates. Equivalent water thickness (EWT), leaf area index (LAI), CWC and canopy hyperspectral reflectance were measured at 135, 160 and 175 days after sowing (DAS), together with final seed yield (SY) and crop water productivity (WP). Ridge–furrow film mulching increased CWC relative to no mulching at all stages. CWC showed the strongest pooled spectral response (|r| = 0.728), and Gaussian process regression (GPR) showed the most consistent evaluation-subset performance among the compared models. At DAS 160, GPR-derived CWC was strongly associated with measured EWT (r = 0.888, P < 0.001), and the relationship remained significant after accounting for LAI (partial r = 0.671, P < 0.001). The derived CWC was also associated with final SY and WP (r = 0.840; R² = 0.706 for both). Breakpoints of 0.113 and 0.131 kg m⁻² separated high-, moderate- and low-risk classes, with lower measured EWT in the high-risk than in the low-risk class. DAS 160 CWC therefore represents a candidate canopy-scale proxy for productivity-risk screening and prioritization of field water-status verification; the empirical thresholds may support subsequent assessment of supplemental-irrigation need but require independent validation and do not prescribe irrigation timing or amount.