EGI OpenIR  > 研究系统  > 荒漠环境研究室
典型荒漠植被同化枝叶绿素含量的定量遥感
张思楠
学位类型硕士
导师王权 ; 李兰海
2015
学位授予单位中国科学院大学
学位授予地点北京
学位专业自然地理学
关键词同化枝 叶绿素含量 辐射传输模型 Prospect 高光谱指数
摘要对叶片叶绿素含量的动态监测是理解荒漠生态系统特定的碳循环过程及其生态功能的一个重要前提。定量遥感的方法能快速无损地获取叶片中的叶绿素含量信息,是植被动态监测的有效手段,而对于荒漠植被叶绿素含量的定量遥感研究还比较少。梭梭(Haloxylon ammodendron)和多枝柽柳(Tamarix ramosissima)作为两种代表性的荒漠植被物种,都具有独特的旱生结构,叶片严重退化并形成独特的“同化器官”,对其叶绿素含量的动态监测具有一定的代表性。本研究采用了辐射传输模型和光谱指数两种定量遥感的手段对这两种代表性的荒漠植被叶绿素含量进行了提取。 对于中亚荒漠生态系统中的优势物种之一的多枝柽柳同化枝叶绿素含量的反演结果表明未经校正的辐射传输模型PROSPECT-4模型在反演多枝柽柳同化枝叶绿素含量上表现较差,精度很低。而相比之下,利用实测数据进行校正后的模型能够较为准确的反演柽柳同化枝叶绿素含量,其决定系数(R2)为0.47,均方根误差(RMSE)为28.79 mg/m2,而实测叶绿素的变化范围在57.37-202.27 mg/m2之间。十折交叉验证的均方根误差为26.19 mg/m2,与校正模型的RMSE相近,因此模型校正并不存在过度拟合现象。本研究进一步测试了参与模型校正的数据比例以及重复次数的不同对结果的影响。除了PROSPECT模型本质上的缺陷以外,其在模型校正后反演多枝柽柳同化枝叶绿素含量的精度还是比较令人满意的,不失为高光谱遥感调查荒漠植被叶绿素含量的可行方法。 本研究进一步收集了86种可用于估算叶绿素的植被指数,并测试了其在梭梭和多枝柽柳数据上的表现,这些指数在梭梭数据中的表现较差,虽然在多枝柽柳数据中表现相对更好,但仍然不足以用来进行同化枝叶绿素含量的估算。对此,本研究分别用梭梭和多枝柽柳的实测光谱对10种类型的光谱指数进行波段组合搜索,找出了能在各自数据库中表现最好的4种指数,分别是适用于梭梭同化枝的指数R720/R1165和(D690-D1475)/(D690+D1475),适用于多枝柽柳的指数(R600-R1630)/(R600+R1630)和D540/D1690。经数据检验发现,基于一阶导数的光谱指数的表现要比原始光谱指数更稳定。
其他摘要The understanding of dynamic change of leaf chlorophyll content is an important approach for better tracing featured carbon cycle as well as other ecological functions in desert ecosystems. Quantitative remote sensing is a valid instrument to dynamically monitor vegetation condition since it can obtain the information of leaf chlorophyll content non-destructively and rapidly, and the study of quantitative remote sensing on chlorophyll content of desert vegetation is comparative little. This study aims to estimate the chlorophyll content of assimilating branches of two typical desert species--Haloxylon ammodendron and Tamarix ramosissima, which both have distinct structural features of eromorphism within their young twigs functioned as "assimilating organisms"--through the radiative transfer model and hyperspectral index remote sensing methods. Application of inverse retrieval of chlorophyll for Tamarix ramosissima, a dominant species in central Asia via the radiative transfer model PROSPECT was examined in this study. The results revealed that effect of the inverse retrieval based on the original version of PROSPECT-4 was poor with low accuracies. However, the calibrated version based on the measured data was capable for the accurate retrieval of chlorophyll content, which R2 (coefficient of determination) was 0.47 with an RMSE (Root Mean Square Error) of 28.79 mg/m2. The measured chlorophyll ranged from 57.37 to 202.27 mg/m2. The model calibration proved to be not overfitting, since the ten-fold cross-validation showed a close RMSE value of 26.19 mg/m2. Our experiment also evaluated the effects of the percentage of training dataset and the numbers of replications for calibration. Despite some inherent defects of applying broadleaf-oriented PROSPECT on assimilating branches, it provides a feasible selection to study the inversely retrieve the chlorophyll content for desert plants from the reflectance by using the hyperspectral remote sensing method. Meanwhile, we collected total 86 published spectral indices for assessing the chlorophyll content within Haloxylon ammodendron and T. ramosissima species. The performance of this approach on H. ammodendron was poor, while was relatively better on T. ramosissima. Even so it was not good enough to estimating chlorophyll content of T. ramosissima assimilating branches accurately via these spectral indices. Therefore, we attempted to find new indices for H. ammodendron or T. ramosissima by searching for 10 types of indices from different band combinations of reflected and derivative spectra. The best indices for H. mmodendron assimilating branches were identified to be R720/R1165 and (D690-D1475)/(D690+D1475), while the best indices for T. ramosissima assimilating branches were (R600-R1630)/(R600+R1630) and D540/D1690, respectively. Furthermore, the indices based on derivative spectra showed better performance than those of original reflectance.
学科领域自然地理学
语种中文
文献类型学位论文
条目标识符http://ir.xjlas.org/handle/365004/14636
专题研究系统_荒漠环境研究室
作者单位中科院新疆生态与地理研究所
推荐引用方式
GB/T 7714
张思楠. 典型荒漠植被同化枝叶绿素含量的定量遥感[D]. 北京. 中国科学院大学,2015.
条目包含的文件
条目无相关文件。
个性服务
推荐该条目
保存到收藏夹
查看访问统计
导出为Endnote文件
谷歌学术
谷歌学术中相似的文章
[张思楠]的文章
百度学术
百度学术中相似的文章
[张思楠]的文章
必应学术
必应学术中相似的文章
[张思楠]的文章
相关权益政策
暂无数据
收藏/分享
所有评论 (0)
暂无评论
 

除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。