Optimal selection of the phenological models for wine grapevine and the impact of climate change on phenology in northwest China

Grape wine is the largest type of fruit wine industry in the world. Among them, Northwest China is one of the most important regions to produce wine grapes with enormous market potential and high added value, due to the unique climate conditions for the growth and development of grapes. However, the Sixth Assessment Report (AR6) of the Intergovernmental Panel on Climate Change (IPCC) also released that the global climate system has posed a significant impact under industrialization and human activities. Particularly, climate change has threatened the regional development of the wine grape industry in Northwest China. This research aims to conduct on the phenological phases of wine grapes in the northwest, and then their changes were predicted to efficiently utilize the local climate resources. The surface temperature and phenological parameters (budburst, flowering, and veraison phases) of wine grapes (Cabernet Sauvignon, Pinot Noir, Merlot, and Chardonnay) were collected from 2007 to 2022. The effective accumulated temperature (GDD₅, GDD₁₀, and GFV), BRIN, and WE models were optimized during the growth period. The phenological phase of wine grapes was further simulated under the SSP245 and SSP585 scenarios using climate model data. The results showed that the BRIN model exhibited the best simulation for Cabernet Sauvignon, Merlot, and Chardonnay, while the GDD₅ model was most effective for Pinot Noir during the budburst phase. Furthermore, the GDD₁₀ model shared the best performance for Cabernet Sauvignon and Merlot during the flowering phase, whereas, the GFV and WE models were superior for Pinot Noir and Chardonnay, respectively. In the veraison phase, the WE model demonstrated the best simulation for the Cabernet Sauvignon, Pinot Noir, and Merlot, while the GFV model was the most accurate for Chardonnay. Compared with the historical period, the phenological phases of the four wine grape species showed no significant difference under the SSP245 scenario. However, the budburst, flowering, and veraison phases of wine grapes all showed a noticeable trend of advancement under the SSP585 scenario, except for the veraison phase at the northern foot of the Tianshan Mountains. Further analysis revealed that there was no significant advancement in the future frequent occurrence of high-temperature disasters in the northern foothills of the Tianshan region, even delayed in the veraison phase of wine grapes. Meanwhile, the extreme spring temperatures were underestimated to delay the BRIN model, according to the climate model data. Three phenological models shared great variations to simulate the different phenological phases of wine grapes. Overall, the best performance was achieved in the flowering phase, followed by the veraison and budburst phase. According to the phenological phases, the BRIN model demonstrated that there was a significant advantage to simulating the budburst phase, while the WE model was more accurate than that in the veraison phase. Therefore, the effect of high temperatures on the phenological phase was evaluated to predict the veraison phase. Especially, the early effects caused by rising temperatures were partially offset by high temperatures in summer. Therefore, the WE model with high temperatures was usually more accurate to reflect future trends. The phenological phase of wine grapes was also determined under the future climate scenarios. The finding can provide the scientific basis to optimize the wine grape phenology models, in order to promote the development of the local wine grape industry.