Vol. 9, Special Issue 11, Part B (2025)

A field experiment was conducted at the Instructional Farm of the College of Agriculture, Odisha University of Agriculture and Technology (OUAT), Bhubaneswar, to quantify the relationships between environment as influenced by sowing dates and the growth, phenological development, and grain yield of pigeonpea (Cajanus cajan L. Millsp.) varieties. The experiment comprised four sowing dates (21^st October, 3^rd November, 16^th November, and 29^th November) and four pigeonpea varieties (TDRG-59, GRG-152, GT-104, and Pant A-7). Results revealed that grain yield was significantly influenced by sowing dates, with the earliest sowing (21^st October) producing the highest yield (939 kg ha⁻¹), which was statistically at par with the 3^rd November sowing (878 kg ha⁻¹). Delayed sowing on 29^th November resulted in the lowest grain yield (681 kg ha⁻¹), representing a 27.5% reduction compared to the earliest sowing date. As the date of sowing delayed, there was decrease in accumulation of GDD, HTU & PTU values so as to reach different phenological stages across the sowing dates at Bhubaneswar agro-climatic conditions, except for PTI. The multiple linear regression models explained 46-66% of yield variability (R² values), with the mid to late pod formation stage model showing the highest accuracy (R² = 0.66). Predicted yields closely matched observed values, with errors as low as 0.1-0.8% for specific variety-sowing combinations, enabling forecasts one to one and a half months of pre-harvest. The weather-index-based regression models provide a reliable, statistical tool for timely yield prediction, aiding farmers in agronomic decision-making and resource optimization amid climate variability.