| 典型区域土地利用/景观格局对黄河上游水体TN的影响 |
| 摘要点击 1241 全文点击 267 投稿时间：2023-10-08 修订日期：2024-01-07 |
| 查看HTML全文
查看全文 查看/发表评论 下载PDF阅读器 |
| 中文关键词 黄河甘肃段 土地利用 景观格局 随机森林回归 BP神经网络 |
| 英文关键词 Gansu section of the Yellow River land use landscape pattern random forest regression BP neural network |
|
| 中文摘要 |
| 为探究土地利用景观格局与水质的相关关系,以黄河上游甘肃段水源涵养区、水土流失区和生态脆弱区为研究对象,基于2020年土地利用数据与水质监测断面数据,选取200 m、500 m、1 km、2 km、5 km和10 km河岸带缓冲区为研究区域,采用单因子指数评价法、随机森林回归模型和BP神经网络等方法量化多空间尺度下黄河上游甘肃段土地利用及景观格局与水质指标的响应关系,并开展基于土地利用景观指数数据的流域水质预测. 结果表明：①通过单因子指数法得出,影响黄河流域甘肃段的水环境的主要指标为总氮（TN）,7~9月丰水期TN指标相对降低,溶解氧（DO）、高锰酸盐指数、氨氮（NH4+-N）和总磷（TP）等地表常规指标均满足地表水环境Ⅲ类水水质标准. ②采用随机森林回归模型分析土地利用和景观指数对TN的影响程度得出不同典型区域内影响TN的存在差异,对水源涵养区、水土流失区和生态脆弱区TN指标影响程度最高的土地利用类型分别为耕地、草地和建设用地. ③采用BP神经网络开展基于土地利用景观指数不同典型区域的水质指标预测,预测得出水源涵养区的结果较好,预测值与实际值误差率均在10%以下,预测精度较高. 研究表明,基于土地利用/景观指数与水质量化关系模型开展水质预测具备良好的水质预测效果. |
| 英文摘要 |
| This study aimed to explore the relationship between land use landscape pattern and water quality in the upstream of the Gansu water conservation, water and soil erosion, and ecological fragile areas. Based on the land use data and water quality monitoring section in 2020 in the 200 m, 500 m, 1 km, 2 km, 50 km, and 10 km riparian buffer area, the single-factor index evaluation method, random forest regression model, and BP neural network were used to quantify the response relationship between land use and landscape pattern of the upper Yellow River in Gansu province and water quality index and to carry out the basin water quality prediction based on land use landscape index data. The results showed that： ① through the single-factor index method, the major indicators of the total nitrogen （TN） in July and September, dissolved oxygen （DO）, permanganate index, ammonia nitrogen （NH4+ -N）, total phosphorus （TP）, and other surface indexes met the surface water environment class Ⅲ water quality standard. ② The random forest regression model was used to analyze the influence of land use and landscape index on TN, and the difference in TN in different typical areas was obtained. The land use types with the highest influence on the TN index in water conservation areas, soil and soil erosion areas, and ecological fragile areas were cultivated land, grassland, and construction land, respectively. ③ The BP neural network was used to predict the water quality index based on different typical areas of land use landscape index. The result of water conservation areas was good, the error rate between the predicted value and the actual value was below 10%, and the prediction accuracy was high. The study showed that water quality prediction based on land use and landscape index/water quality quantitative relationship model had a good water quality prediction effect. |
|
|
|
