青藏高原蕴含着最典型、最多样的高寒湿地类型。为了探究高寒湿地的退化状况, 首先需要对高寒湿地信息进行精细化提取。选用玛多县GF1-WFV生长季影像, 利用先进的分层分类法, 选择分割尺度为50对影像进行分割, 然后利用单波段阈值法及坡度阈值对湿地进行判识, 结合非生长季影像, 综合利用波谱关系法、混合水体指数法、归一化差异水体指数法及单波段阈值法对玛多县的高寒湿地分类系统的Ⅲ级类别进行逐一提取, 最后得到玛多县高寒湿地类型的地物信息及分布状况。研究表明: (1)随机选取检验样点, 采用混淆矩阵的方法, 对提取的高寒湿地分类影像进行精度评价, 分类精度达到88.59%, Kappa系数为0.8637, 精度通过检验。分类结果显示, 结合了影像的纹理特征和光谱特征的分层分类技术在高寒湿地信息的提取方面更加优于其他传统方法, 能够实现高寒湿地的精细化信息提取与分析。(2)玛多县高寒湿地面积呈东多西少, 北多南少分布, 其中永久性淡水湖的面积最大为1685.58 km2, 占玛多县总高寒湿地面积的69.05%; 其次是草本沼泽和永久性河/溪, 面积分别为495.56 km2和94.81 km2, 占比分别为20.34%和3.88%; 季节性咸水湖、季节性淡水湖、间歇性河/溪、洪泛湿地、泥潭沼泽、灌丛沼泽、内陆盐沼及冰川积雪的湿地类型面积在1.25~73.23 km2, 所占比例不足1%, 其中面积最小的是季节性咸水湖和季节性淡水湖。(3)高分数据在青藏高原高寒湿地信息提取中具有不受高海拔影响分辨率的特点, 因此具有更强的的可行性和应用性, 尤其选择时间分辨率不同的生长季及非生长季的影像, 在获取季节性高寒湿地类型信息方面具有更大的优势。(4)玛多县高寒湿地类型中, 湖泊面积有增加的趋势, 但河流和洪泛湿地面积有减少的趋势, 如果不采取及时、有效的保护措施, 玛多县的湿地有可能继续退化。
The Qinghai-Xizang Plateau contains the most typical and diverse types of alpine wetlands.In order to explore the degradation of alpine wetlands, it is necessary to extract accurate information of alpine wetlands.Research methods Selects growing season images of GF1-WFV in Maduo County and uses hierarchical classification method.Chose the 50-segmentation scale to segment the image, and then the wetlands are identified by means of single band threshold and slope threshold.Comprehensive uses of these methods, including spectral relation method, mixed water index method, normalized differential water index method and single band threshold method.Finally, the information and distribution status of the alpine wetland types in Maduo County, Qinghai Province were obtained, combines with non-growing season images, according to the class III classification standard of alpine wetland remote sensing classification system.Research results Firstly, the sample points were randomly selected and the accuracy of the image was evaluated by the method of confusion matrix.The classification accuracy of this study reaches 88.59%, and the Kappa coefficient is 0.8637, the classification accuracy passes the test.The classification results show that the hierarchical classification technique combining image texture features and spectral features is superior to other traditional methods in the extraction of alpine wetland information, and it is able to achieve Information extraction and analysis of alpine wetland refinement.Secondly, the area of alpine wetland in Maduo County is more in the east than in the west, more in the north and less in the south, and the main types of wetlands are concentrated in the central and northern regions.The largest area of the permanent freshwater lake is 1685.58 km2, accounting for 69.05% of the total alpine wetland area of Maduo County, followed by herbaceous swamps and permanent rivers/streams, with an area of 495.56 km2 and 94.81 km2, accounting for 20.34% and 3.88%, respectively.The wetland area of seasonal saltwater lakes, seasonal freshwater lakes, intermittent rivers/streams, flooded wetlands, mud marshes, scrub swamps, inland salt marshes and glacial snow cover wetlands is within 1.25~73.23 km2, accounting for less than 1%, of which seasonal saltwater lakes and seasonal freshwater lakes are the smallest.Thirdly, the high-score data is not affected by the high-altitude resolution in the information extraction of alpine wetland in Qinghai-Xizang Plateau, so it has more feasibility and application, especially choosing the image of growth season and non-growth season with different time resolution, which has more advantages in obtaining seasonal alpine wetland type information.Lastly, the classification result shows that among the types of alpine wetlands in Maduo County, the area of lakes tends to increase, but the area of rivers and flooded wetlands tends to decrease.If protective measures are not taken timely and effectively, the wetlands in Maduo County may continue to degrade.
[1]Aguinaga O E, Anna M M, White K N, et al, 2018.Microbial community shifts in response to acid mine drainage pollution within a natural wetland ecosystem[J].Frontiers in Microbiology, 9: 1445.
[2]Bai J H, Ouyang H, Cui B S, et al, 2008.Changes in landscape pattern of alpine wetlands on the Zoige Plateau in the past four decades[J].Acta Ecologica Sinica, 28(5): 2245-2252.
[3]Cui X, Guojun J, Tao W, et al, 2013.Dynamic monitoring of sea areas use based on high-resolution remote sensing: A case study of imen Island[C]//International Conference on Remote Sensing, Environment and Transportation Engineering (RSETE 2013), 702-705.
[4]Evenson G R, Golden H E, Lane C R, et al, 2018.Depressional wetlands affect watershed hydrological, biogeochemical, and ecological functions[J].Ecological Applications, 28 (4), 953–966.
[5]Hubert-Moy L, Micheal K, Corpetti T, et al, 2006.Object-oriented mapping and analysis of wetlands using SPOT 5 data[C]//Object-oriented mapping and analysis of wetlands using SPOT 5 data.Geoscience and Remote Sensing Symposium, 2006 IGARSS 2006 IEEE Internation Conference on IEEE, 3447-3450.
[6]Johnston S E, Henry M C, Gorchov D L, 2012.Using Advanced Land Imager (ALI) and Landsat Thematic Mapper (TM) for the detection of the invasive shrub lonicera maackii in southwestern Ohio Forests[J].GIS Science & Remote Sensing, 49(3): 450-462.
[7]Meena D K, Lianthuamluaia L, Mishal P, et al, 2019.Assemblage patterns and community structure of macro-zoobenthos and temporal dynamics of eco-physiological indices of two wetlands, in lower gangetic plains under varying ecological regimes: A tool for wetland management[J].Ecological Engineering, 130: 1-10.
[8]Munyati C, 2000.Wetland change detection on the Kafue Flats, Zambia, by classification of a multitemporal remote sensing image dataset[J].International Journal of Remote Sensing, 21(9): 1787-1806.
[9]蔡迪花, 郭铌, 韩涛, 2007.1990 -2001年黄河玛曲高寒沼泽湿地遥感动态监测[J].冰川冻土, 29(6): 874-881.
[10]曹生奎, 谭红兵, 王小梅, 等, 2005.青藏高原湿地保护与开发利用模式初探[J].干旱区资源与环境, 19(4): 109-113.
[11]成淑艳, 曹生奎, 曹广超, 等, 2018.基于高分辨率遥感影像的青海湖沙柳河流域土地覆盖监督分类方法对比[J].水土保持通报, 238(5): 267-274.
[12]方朝阳, 邬浩, 陶长华, 等, 2016.鄱阳湖南矶湿地景观信息高分辨率遥感提取[J].地球信息科学学报, 18(6): 847-856.
[13]杜红艳, 张洪岩, 张正祥, 2004.GIS支持下的湿地遥感信息高精度分类方法研究[J].遥感技术与应用, 19(4), 244-248.
[14]郭文婷, 张晓丽, 2019.基于Sentinel-2时序多特征的植被分类[J].浙江农林大学学报, 36(5): 849-856.
[15]何鸿杰, 穆亚超, 魏宝成, 等, 2019.分层分类和多指标结合的西北农牧交错带植被信息提取[J].干旱区地理, 42(2): 112-120.
[16]贾祎琳, 张文, 孟令奎, 2019.面向GF-1影像的NDWI分割阈值选取方法研究[J].国土资源遥感, 31(1): 98-103.
[17]李凤霞, 常国刚, 肖建设, 等, 2009.黄河源区湿地变化与气候变化的关系研究[J].自然资源学报, 4(14): 683-690.
[18]李建平, 张柏, 张泠, 等, 2007.湿地遥感监测研究现状与展望[J].地理科学进展, 26(1): 33-43.
[19]李林, 李凤霞, 朱西德, 等, 2006.黄河源区湿地萎缩驱动力的定量辨识[C]// 中国气象学会2006年年会“气候变化及其机理和模拟”分会场论文集.
[20]李英年, 2006.祁连山海北高寒湿地植物群落结构及生态特征[J].冰川冻土, 28(1): 76-84.
[21]刘彩红, 苏文将, 杨延华, 2012.气候变化对黄河源区水资源的影响及未来趋势预估[J].干旱区资源与环境, 26(4): 97-101.
[22]刘志伟, 李胜男, 韦玮, 等, 2019.近三十年青藏高原湿地变化及其驱动力研究进展[J].生态学杂志, 38(3): 241-247.
[23]罗栋梁, 金会军, 2014.黄河源区玛多县1953 -2012年气温和降水特征及突变分析[J].干旱区资源与环境, 28(11): 185-192.
[24]牛振国, 宫鹏, 程晓, 等, 2009.中国湿地初步遥感制图及相关地理特征分析[J].中国科学(地球科学), 39(2): 188-203.
[25]潘竞虎, 王建, 等, 2007.长江、黄河园区高寒湿地动态变化研究[J].湿地科学, 5(4): 298-304.
[26]青海省气象局, 2019.《高寒湿地遥感分类技术指南DB63/T 1746-2019》[S].青海省市场监督管理局.
[27]宋昌素, 肖燚, 博文静, 等, 2019.生态资产评价方法研究——以青海省为例[J].生态学报, 39(1): 13-27.
[28]孙洪烈, 郑度, 1998.青藏高原形成演化和发展[M].广州: 广东科学技术出版社, 155-169.
[29]田坤, 王宁, 彭建生, 2018.青藏高原: 中国最大的水乡[J].森林与人类, 2342(12): 68-77.
[30]王启基, 周兴民, 沈振西, 等, 1995.高寒藏嵩草沼泽化草甸植物群落结构及其利用[C]//高寒草甸生态系统(4).北京: 科学出版社: 91-100.
[31]王晓龙, 张寒, 姚志生, 等, 2016.季节性冻结高寒泥炭湿地非生长季甲烷排放特征初探[J].气候与环境研究, 21(3): 282-292.
[32]吴晗, 董增川, 蒋飞卿, 等, 2018.黄河源区气候变化特性分析[J].水资源与水工程学报, 29(6): 4-10.
[33]吴薇, 张源, 李强子, 等, 2019.基于迭代CART算法分层分类的土地覆盖遥感分类[J].遥感技术与应用, 34(1): 68-78.
[34]武慧智, 姜琦刚, 程彬, 2007.基于RS和GIS技术青藏高原湖泊动态变化研究[J].世界地质, 26(1): 68-72.
[35]薛在坡, 李希来, 张红林, 等, 2015.黄河源玛多县3种类型湿地面积动态研究[J].青海大学学报, 33(3): 45-51.
[36]姚红岩, 刘浦东, 施润和, 等, 2017.基于高分辨率遥感影像的湿地互花米草-芦苇混合交错带提取方法[J].地球信息科学学报, 19(10): 1375-1381.
[37]张天举, 陈永金, 刘加珍, 2019.基于典范对应分析的滨海湿地土壤季节性盐渍化特征[J].生态学报, 39(9): 3322-3332.
[38]张婉平, 2018.季节性受淹芦苇湿地蒸散发的模拟研究[D].南京: 东南大学.
[39]张国庆, 2018.青藏高原湖泊变化遥感监测及其对气候变化的响应研究进展[J].地理科学进展, 37(2): 214-223.
[40]赵串串, 张愉笛, 张藜, 等, 2017.黄河源区玛多县湿地生态健康评价[J].安徽农业大学学报, 44(1): 108-113.
[41]赵志龙, 张镱锂, 刘林山, 等, 2014.青藏高原湿地研究进展[J].地理科学进展, 33(11): 1218-1230.
[42]朱丽, 刘蓉, 王欣, 等, 2019.基于FLEXPART模式对黄河源区盛夏降水异常的水汽源地及输送特征研究[J].高原气象, 38(3): 484-496.DOI: 10.7522/j.issn.1000-0534.2019.00015.