Abstract:

To elucidate the influence mechanism of meteorological factors on the green-up period of forage grass in alpine grassland and improve prediction accuracy, this study integrated 2005-2023 phenological observations and meteorological data from the Tongde County Meteorological Bureau (Qinghai Province). Using linear trend analysis, Pearson correlation, and stepwise regression, we characterized interannual variation of the green-up date, identified key climatic drivers, and established a multiple linear regression prediction model. The results demonstrated that: (1) the mean green-up date was 1 May (Julian day 121 d), exhibited pronounced interannual variability (SD=5.8 d; range: 26 d), with no significant long-term trend (0.4 d/10 a, P>0.10), indicating high sensitivity to short-term climate fluctuations. (2) Accumulated temperature ≥0℃ (AT) showed a highly significant positive correlation with the green-up date (r=0.83, P<0.001), while sunshine hours also exhibited a significant correlation (r=0.61, P<0.01), precipitation had no discernible effect. (3) The stepwise regression model retained the ordinal date of the first AT≥3℃ and AT as predictors, achieving an R²=0.869 (F=52.92, P<0.001); Backward substitution validation yielded a mean absolute error of 2.0 d (98.6% accuracy, with 95% of predictions within ±2.0 d). This study confirms that that spring heat accumulation is the dominant driver of green-up timing, whereas water availability plays a secondary role. The proposed model provides a robust framework for phonological monitoring and sustainable grazing management in alpine ecosystems.

Key words: alpine steppe, forage grass green-up date, accumulated temperature ≥0℃, climatic response, phenological prediction, stepwise regression