高级检索+

新疆南部苹果花期复合冻害大风事件时空变化及风险评估

Spatiotemporal variation and risk assessment of compound frost and wind events during apple flowering in southern Xinjiang

  • 摘要: 为量化新疆南部苹果花期复合冻害大风事件风险,基于1990—2023年3—5月逐日最低温(Tmin)与风速(W)数据构建标准化温度指数(standardized temperature index,STI)及风速阈值,识别四类复合冻害大风事件(S1:STI≤−0.9,WW90;S2:STI≤−0.9,WW95;S3:STI≤−1.3,WW90;S4:STI≤−1.3,WW95)。并采用Vine Copula刻画STI、W与归一化植被指数(normalized difference vegetation index,NDVI)的非线性依赖关系,估算不同复合冻害大风事件下NDVI条件损失概率与重现期。结果表明:研究区复合冻害大风事件总体呈下降趋势。四类事件多年平均发生频次S1最高,为0.339;S2、S3事件次之,分别为0.158、0.126;S4事件最低,为0.059。NDVI条件损失概率高值区主要分布于和田南部和东南部、喀什中部及阿克苏北部,与苹果种植集中区域高度重叠。重现期以0~10与10~20a为主,整体呈现“和田高频、喀什次之、阿克苏低频但极端事件局地突出”的空间格局。该研究可为新疆南部苹果花期防寒防风重点区识别与风险评估提供依据。

     

    Abstract: Apple is one of the most important economic fruit trees in the characteristic forestry and fruit industry of southern Xinjiang. Stable apple production is essential for maintaining fruit supply, supporting regional agricultural development, and increasing farmers’ income. However, apple trees are highly sensitive to adverse meteorological conditions during the flowering period, especially low temperature and strong wind events. In southern Xinjiang, spring cold-air activities are often accompanied by strong winds, which may intensify frost injury through enhanced canopy heat loss, mechanical damage to flowers, and unfavorable pollination conditions. Therefore, identifying compound frost and wind events and quantifying their potential impacts are important for improving regional disaster prevention and orchard management. This study focused on the main apple-producing areas of Hotan, Kashgar, and Aksu in southern Xinjiang. Based on daily minimum temperature, wind speed, and normalized difference vegetation index (NDVI) data from March to May during 1990–2023, a standardized temperature index (STI) and wind-speed percentile thresholds were constructed to identify four types of compound frost and wind events during the apple flowering period: S1 (STI≤−0.9, WW90), S2 (STI≤−0.9, WW95), S3 (STI≤−1.3, WW90), and S4 (STI≤−1.3, WW95). Marginal distribution fitting and a Vine Copula model were further used to characterize the nonlinear dependence among STI, wind speed, and NDVI. On this basis, the conditional loss probability of NDVI under different compound frost and wind events and the return periods of these events were estimated to reveal their spatiotemporal variation and risk patterns. The results showed that the occurrence of compound frost and wind events during the apple flowering period in southern Xinjiang generally decreased from 1990 to 2023. The decadal slopes of S1, S2, S3, and S4 were −2.86, −1.33, −2.22, and −1.20 events per decade, respectively, indicating a weakening tendency of compound frost and wind events over the study period. Nevertheless, obvious interannual fluctuations remained, suggesting that compound hazards may still occur in specific years. The multi-year mean occurrence frequencies of S1, S2, S3, and S4 were 0.339, 0.158, 0.126, and 0.059, respectively, with S1 being the most frequent and S4 the least frequent. The spatial distribution of the four event types showed clear heterogeneity. High-frequency areas of S1 were mainly located in eastern Hotan and northern Kashgar. High-frequency areas of S2 and S3 were concentrated in southeastern Hotan and northwestern Aksu. Although S4 had a relatively low overall frequency, northwestern Aksu still showed locally prominent high-frequency characteristics, indicating that rare but severe compound events may occur in specific areas. The conditional loss probability derived from the Vine Copula model indicated that areas with high NDVI loss probability were spatially consistent under different NDVI quantile thresholds. These high-risk areas were mainly distributed in southern and southeastern Hotan, central Kashgar, and northern Aksu, which largely overlapped with the concentrated apple-growing regions. As the NDVI quantile threshold decreased from 0.50 to 0.10, the spatial extent of high conditional loss probability areas contracted markedly, but the locations of sensitive regions remained generally stable. This suggests that the identified high-risk areas were robust and may represent key zones where apple flowering is more vulnerable to compound frost and wind stress. Return-period analysis further showed that compound frost and wind events in southern Xinjiang were mainly characterized by return periods of 0–10 years and 10–20 years. For S1, return periods of 0–5 years and 5–10 years dominated, indicating frequent mild compound events. S2 and S3 were mainly characterized by return periods of 5–10 years, whereas the area proportions of 10–20 years and ≥20 years increased markedly for S4. This indicates that as the frost and wind thresholds became more extreme, event frequency decreased but the upper bound of potential risk increased. Overall, the compound frost and wind risk during the apple flowering period showed a spatial pattern of frequent occurrence in Hotan, moderate occurrence in Kashgar, and relatively low frequency but locally prominent extreme events in Aksu. These findings provide a scientific basis for identifying key frost and wind prevention areas, improving compound agrometeorological disaster risk assessment, and supporting refined orchard management in southern Xinjiang.

     

/

返回文章
返回