Identifying optimal planting dates and empirical validation
To maximize crop productivity by harnessing phenotypic plasticity, we propose a strategic approach for selecting the most advantageous planting dates. Indeed, the timing of planting influences the weather conditions experienced by the crop during crucial growth stages. To identify the optimal planting dates at the three tested sites in this study (Fig. 3a), we predicted SOC for several new planting dates with the regression model of SOC on the identified environmental indices (DTR183–192, PR166–195 and UVB144–186), conditioned on the assumption of a linear reaction norm pattern across the environments (see Methods). To ensure a practical planting guide, we limited the dates to within 10 days before or after the actual planting dates. We then compared the measured SOC at actual planting dates to the predicted SOC values for early or late plantings. We determined that planting 10 days earlier would have resulted in higher SOC, while planting 10 days later would have decreased SOC (Fig. 3b; Supplementary Table S8). Specifically, early planting resulted in an increase of up to 13.80% (HF2018), whereas late planting caused a decrease of up to 8.92% in SOC (WH2018).
To assess the general applicability of our approach for selecting optimal planting dates, we predicted SOC at four additional sites not included in the initial testing. We chose these sites from major oilseed producing areas in China (Fig. 3a; Supplementary Table S9). Using historical weather data from 2011 to 2020, we calculated the predicted SOC for each planting date between September 10th and October 28th at all seven sites (Fig. 3c). Early planting also generally leads to higher SOC compared to late planting, even for the four sites from which we lacked experimental data. The SOC varies among the seven planting sites at the optimal and earliest planting dates, with the highest SOC observed at CS (48.9%) and the lowest at CD (38.65%). In addition, the decrease in SOC is more pronounced at high altitude sites (e.g., ZZ) compared to low altitude sites (e.g., CS) when planting is not conducted at or near the optimal date (Supplementary Table S10).
We conducted empirical validation of the above SOC predictions with three planting dates in two new environments. We chose 50 lines from the diversity panel and planted them in Wuhan (WH2023) and Ezhou (EZ2023) on three different dates separated by 10 days each time. Generally, we observed that late planting results in lower SOC in both environments. We also predicted the SOC performance of all 50 inbred lines during the 2023 season using environmental data from the three identified indices (DTR183–192, PR166–195 and UVB144–186). We obtained high prediction accuracies at both locations, with Pearson’s correlation coefficients of 0.62 and 0.67 in WH2023 and EZ2023, respectively (Fig. 4).