| 摘要: |
| ［目的］川西横断山河谷是我国乃至世界高山酿酒葡萄的新兴产区，赤霞珠是该区域主栽品种之一。复杂的地理条件使得该区域与其他酿酒葡萄主产区的自然环境差异显著。科学评估川西横断山河谷区赤霞珠适宜空间分布，可促进该区域酿酒葡萄可持续发展。［方法］文章综合运用MaxEnt模型和GIS技术，构建了赤霞珠适宜分布与环境变量的关系模型，筛选出主导环境变量，利用存在概率划分适宜等级，预测了赤霞珠在川西横断山河谷的适宜分布。［结果］赤霞珠适宜分布与环境变量的关系模型准确度非常好（AUC>090）。在川西横断山河谷，主导环境变量依次为最热月平均气温、海拔、pH、秋季降雨量、年日照时数、年温差、全氮。川西横断山河谷大部分地区为不适宜区，低适宜区约为2 2752km2，中适宜区约1 29608km2。高适宜区面积为1 16428km2，且集中分布在丹巴县、小金县、金川县、九寨沟县和巴塘县。［结论］MaxEnt模型通过存在概率进行作物分布模拟和预测，对作物适宜区划具有较强应用价值和指导意义。对于不同区域和作物应注意环境变量，空间尺度，物种采样位置等问题，提高作物适宜区划精度。最后应加快该区域生产空间调整，在丹巴县、九寨沟县、金川县、小金县、巴塘县赤霞珠还有很大潜力空间，在黑水县、马尔康市、松潘县、白玉县、德荣县、康定县需适度发展，在汶川县、理县、茂县、稻城县、乡城县，应加强生产管理，减少不利环境影响。仁和区不宜再进行规模扩张。 |
| 关键词: MaxEnt模型赤霞珠地理分布适宜性区划横断山 |
| DOI： |
| 分类号:F3231S6631 |
| 基金项目:四川省杰出青年科技人才项目“四川经济作物时空格局变化及响应机制研究”（2020JDJQ0073）； 四川省软科学计划项目“四川贫困地区主要作物（经作）产业可持续发展科技问题与对策研究”（2018ZR0196）； 四川省财政创新能力提升工程“四川省特色水果适宜性评价及布局优化研究”（2018QNJJ010） |
|
| ASSESSMENT OF SPATIAL DISTRIBUTION OF CABERNET SAUVIGNON IN HENGDUANSHAN VALLEY OF WESTERN SICHUAN BASED ON MAXENT MODEL |
|
Lin Zhengyu1,2, Chen Qiang2, Chen Chunyan1, Li Xiao1, He Peng1
|
|
1.Agricultural Information and Rural Economy Institute, Sichuan Academy of Agricultural Science, Chengdu 610066, Sichuan, China;2.College of Resources, Sichuan Agriculture University, Chengdu 611130, Sichuan, China
|
| Abstract: |
| The Hengduan Mountain Valley in Western Sichuan is a new high mountain wine grape production area in China and even in the world. Cabernet Sauvignon is one of the main varieties in this region. The complex geographical conditions make the natural environment of this region significantly different from other wine grape producing areas. This research aims to scientifically evaluate the suitable spatial distribution of Cabernet Sauvignon in Hengduanshan valley of Western Sichuan, so as to promote the sustainable development of wine grape in this region. In this research, the MaxEnt model and GIS technology were used to construct the relationship model between the suitable distribution of Cabernet Sauvignon and environmental variables, and screen the dominant environmental variables, and then the existence probability was used to classify the appropriate grade, predict the suitable distribution of Cabernet in the Hengduan Mountain Valley of Western Sichuan. The results were showed as follows. The accuracy of the relationship model between the suitable distribution of Cabernet Sauvignon and environmental variables was very good (AUC > 0.90). In the Hengduan Mountain Valley of Western Sichuan, the dominant environmental variables were the average temperature of the hottest month, altitude, pH, autumn rainfall, annual sunshine hours, annual temperature difference and total nitrogen. Most of the Hengduan Mountain Valley in Western Sichuan was not suitable area, the low suitable area was about 2 275.2 km2, and the medium suitable area was about 1 296.08 km2. The area of high suitability area was 1 164.28km2, and it was mainly distributed in Danba , Xiaojin, Jinchuan, Jiuzhaigou and Batang. In summary, MaxEnt model can simulate and predict crop distribution through the existence probability, which has strong application value and guiding significance for crop suitable zoning. For different regions and crops, attention should be paid to environmental variables, spatial scale, species sampling location and other issues to improve the accuracy of crop suitable zoning. Finally, we should speed up the adjustment of production space in this region. In Danba, Jiuzhaigou, Jinchuan, Xiaojin and Batang, Cabernet Sauvignon still has great potential. In Heishui, Maerkang, Songpan, Baiyu, Derong and Kangding, moderate development is needed. In Wenchuan, Lixian, Maoxian, Daocheng and Xiangcheng, production management should be strengthened to reduce adverse environmental impact, and Renhe is not suitable for scale expansion. |
| Key words: Maximum entropy model(MaxEnt) Cabernet Sauvignon geographical distribution suitability zoning Hengduan mountains |
