Abstract
Rice (Oryza sativa L.) is a crucial crop for employment and agricultural output and heavily reliant on family labor. This study evaluated the effects of nitrogen levels (80, 120, and 160 kg·ha(-1)) on weed incidence and key agronomic variables, including vegetative growth, yield, and related traits, in Ecuador's primary rice-growing regions, Guayas and Los Ríos. A split-plot randomized complete block design was implemented using two rice varieties (INIAP-FL-Elite and SFL-11) and three planting densities (20 × 30, 25 × 30, and 30 × 30 cm). Weed incidence was higher in Los Ríos, dominated by grasses (55.28%), while Cyperaceae (46.27%) prevailed in Guayas. The data analysis included non-parametric tests to identify significant treatment effects, debiased sparse partial correlations (DSPCs) to reveal key agronomic interactions, and principal component analysis (PCA) to identify influential traits, ensuring robust and normalized interpretations. Analysis of variance indicated significant effects for all variables, with vegetative growth (VG) most affected (p < 0.001). The number of panicles (NP) and days to flowering (DF) showed significant though less pronounced effects, while the panicle length (LP) and 1000-seed weight (TSB) exhibited moderate responses. The DSPCs highlighted the grains per panicle (GP) and total biomass (SB) as critical variables, with significant correlations between the days to flowering and the tiller count at 55 days (r = 0.750, p < 0.001) and between the vegetative growth and the total biomass (r = 0.678, p < 0.001). PCA explained 58.8% of the total variance, emphasizing the days to flowering, plant height, total biomass, and yield as the most influential traits. These findings underline the importance of integrated nutrient and weed management strategies tailored to Ecuador's agroecological conditions.