Physiological, vegetation, and biochemical indicators for biomass, nitrogen uptake and lint yield forecast in cotton

Forecasting crop performance through non-destructive tools is crucial for enabling timely, data-driven decisions for sustainable cotton production. This study evaluated physiological, PIs [chlorophyll index (CI), nitrogen balance index (NBI)], vegetation, VIs [normalized difference vegetation index (NDVI), normalized difference red edge (NDRE)], and biochemical, BIs [leaf nitrogen, petiole nitrate-N] indicators to forecast aboveground biomass accumulation (AGB), nitrogen (N) uptake, and lint yield in cotton. We hypothesized that the predictive strength of each indicator would vary by type and growth stage. Field experiments were conducted across five site-years in West and North Florida, USA, using six N rates (0–252 kg N/ha) in four replications. Quadratic regression (QR) model identified PIs at peak flowering and VIs at first flower and cutout growth stages as the most robust predictors of AGB. A random forest regression (RF) model also showed similar results, with PIs at peak flowering and VIs at peak flowering and first flower as the best indicators for AGB prediction. A strong relationship of BIs with N uptake at bloom and post-bloom growth stages evaluated through QR and RF supports their use for early reproductive assessment. Lint yield predicted using QR and RF was best forecasted by VIs at peak flowering, and cutout. Principal component analysis confirmed mid-to-late season VIs as major drivers of AGB and lint yield variability, while as early reproductive BIs as major indicators of N uptake. This research establishes a robust framework for real-time indicator and growth stage-based biomass, N uptake, and yield forecasting in cotton.