基于叶绿素荧光的植被总初级生产力估算

发布时间：2018-07-20 19:30

【摘要】：植被总初级生产力(Gross Primary Production,GPP)反映的是生态系统水平上,陆生植物通过光合作用所固定的有机碳量,是全球最大的碳通量,也是多个生态系统功能的驱动力。精确地掌握陆地生态系统GPP的时空动态特征,不仅可以加深对生态系统与气候系统的相互反馈作用的认知,而且对于全面理解碳循环和制定应对气候变化相关政策具有重要意义。日光诱导叶绿素荧光(Suninduced Chlorophyll Fluorescence,SIF)作为光合作用的伴生产物,被认为是比传统植被指数更为合适的光合活动“指示器”。基于卫星遥感数据反演的SIF数据集,已有研究发现不同植物区系的GPP与SIF均存在很强的线性关系,能够为GPP估算提供了一种更为直接的遥感方法。然而,冠层尺度GPP与SIF的相关关系还会受到环境胁迫、冠层结构、植物功能型等多方面因素的影响,GPP-SIF遥感估算模型还存在很大的不确定性。鉴于此,本文针对基于SIF的GPP遥感估算模型与不确定性这一重要的基础科学问题,开展了GPP-SIF相关关系对光合生理响应的敏感分析、不同波段荧光估算GPP的能力比较、以及基于SIF卫星遥感信号的全球GPP模拟与评估三方面工作,主要研究内容与结论如下:1)本文首先探究了SIF探测植物光合固碳的潜力及不同光合功能型GPPSIF耦合关系的差异特征。基于同步观测的GPP和SIF760数据,我们发现在作物冠层GPP、SIF760的日变化规律与入射光合有效辐射一致,均呈午间高、早晚低的趋势,SIF760信号可以追踪植物光合作用的日变化动态。在冬小麦(C3作物)和夏玉米(C4作物)冠层,GPP与SIF760均呈显著的线性相关关系,进一步肯定了SIF遥感信号可以用于直接估算GPP的能力。然而,由于C3、C4作物荧光量子产量ε_F接近,但是光能利用率ε_P差异较大,使得C4作物的GPP-SIF760的线性斜率是C3作物的2.16倍(若去除C3作物光合午休点对,C4斜率是C3的1.82倍)。因此,在应用基于SIF的经验模型(GPP=ε_P/ε_F×SIF)估算作物冠层GPP时,应首先考虑区分植物光合型,分别建立C3、C4作物冠层的GPP估算模型。2)本文通过模拟与观测实验相结合的方法,探究了红光与远红光波段的SIF对GPP的估算潜力。很多理论研究与观测实验发现,红光与远红光波段的叶绿素荧光对光合生理状态的表征存在一定差异,且荧光在冠层内辐射传输的吸收和散射特性存在较大差异,但目前这两个波段的叶绿素荧光对GPP的估算潜力认识不一。我们首先利用SCOPE(Soil-Canopy-Observation of Photosynthesis and the Energy balance)模拟了不同叶绿素密度Cab和叶面积指数LAI情景下冬小麦冠层的GPP、SIF760、SIF688的日变化数据集。统计分析结果表明,随着Cab和LAI的变化,GPP与SIF760的相关关系依然可以用同一的线性模型表示,但是GPP-SIF688线性关系对Cab与LAI的敏感性较高,而呈“展翅状”。植被指数(特别是NDVI_(rededge)、CI_(rededge)和NDVI×WDRVI_(rededge))的变形可以作为红光荧光逃逸比例f_(esc)(688)很好的近似,抑制GPP-SIF688相关关系对Cab与LAI的敏感性,使得冬小麦冠层的GPP-SIF688×f(VI)线性模型的估算能力与GPP-SIF760相当。然而,考虑到红光荧光难以准确反演以及冠层强烈的重吸收作用,远红光荧光可以为监测生态系统光合作用提供更为可靠的方法。3)本文开展了基于SIF卫星遥感的全球GPP估算研究。众多研究表明,叶绿素荧光遥感可作为直接估算GPP的新方法,但是基于SIF卫星遥感的全球尺度GPP估算研究还十分薄弱。本文基于GOME-2 SIF740与FLUXNET共享数据,在不同植被类型建立了GPP-SIF回归模型,并估算了2007-2014年全球GPPSIF数据集。通过与MODIS GPP产品进行对比,发现两个数据集在时序动态、空间分布、年累计总量较为一致,而且GPPSIF产品算法简单且不需要引入辅助的气象数据和植被指数数据。然而,GPPSIF受制于数据质量和通量数据的可用性,而GPPMOD受限于模型输入数据的精度,两个产品在常绿阔叶林、热带稀树草原和灌木林区域存在一定程度的不确定性。此外,我们还比较了景观尺度上C3、C4作物GPPEC与SIF卫星遥感信号的相关关系,发现C4作物线性斜率是C3作物斜率的1.81倍,与地面实验结果基本一致。景观尺度的GPP-SIF线性模型对植物光合功能型也同样敏感,进一步证实C3、C4植物时空动态分布对于准确估算全球GPP是十分重要的。本文的主要创新性贡献包括:1)首次通过理论推导和实验观测,阐明了C3、C4作物GPP-SIF耦合关系的差异特征。针对GPP-SIF相关关系对光合生理响应不确定的研究问题,建立了基于光合电子传递速率(ETR)模型的C3、C4植被GPP-SIF响应差异的理论计算方法,发现相同荧光产量下C4作物GPP比C3作物高1倍左右;并组织系列GPP与SIF日变化同步观测实验,和GOME-2卫星时序遥感实验,进一步定量分析了SIF在C3、C4作物GPP估算中的应用潜力与差异特征,实验测量与理论推导结果一致。2)首次通过模型模拟和野外实验相结合的方法,定量研究了不同波段SIF估算GPP的潜力。由于红光荧光难以准确获取,目前荧光遥感应用主要聚焦于远红光荧光波段。本文基于SCOPE模拟数据及光谱分辨率为0.3 nm的QE Pro光谱仪观测的日变化数据,评价了红光与远红光荧光估算冬小麦冠层GPP的能力,结果表明远红光荧光受冠层结构的影响相对较小,更适合应用于卫星尺度的GPP估算与光合生理监测。3)建立了卫星尺度的不同植被覆盖类型GPP-SIF遥感估算模型,首次研制了SIF卫星遥感数据直接驱动的GPP估算产品。尽管基于SIF直接估算GPP的方法备受关注,但是目前尚未有基于叶绿素荧光的全球陆地生态系统GPP产品。本文估算的2007-2014年全球的陆地生态系统GPPSIF数据集与MOD17 GPP产品的时空格局具有较高的一致性,而且算法简单且机理清晰,不需要借助气象数据和植被指数产品,可以在一定程度上避免辅助数据引入的不确定性。本研究为更好地理解不同时空尺度GPP与SIF的关联机制提供理论依据、实验基础和模型方法,同时提供的基于SIF卫星遥感数据的陆地生态系统GPP产品可用于分析不同类型生态系统的光合功能状态和生态服务能力,将对全球碳循环研究具有重要意义。
[Abstract]:The total primary productivity of vegetation (Gross Primary Production, GPP) reflects the ecological system level. The amount of organic carbon fixed by the terrestrial plants through photosynthesis is the largest carbon flux in the world. It is also the driving force of multiple ecosystem functions. The precise grasp of the spatial and temporal dynamic characteristics of the terrestrial ecosystem GPP can not only deepen the ecology of the ecological system, but also deepen the ecology of the ecosystem. The mutual feedback interaction between the system and the climate system is of great significance for a comprehensive understanding of the carbon cycle and the formulation of climate change related policies. The sunlight induced chlorophyll fluorescence (Suninduced Chlorophyll Fluorescence, SIF), as a companion product of photosynthesis, is considered to be more suitable for photosynthesis than the traditional vegetation index. Active "indicator". Based on the SIF data set retrieved from satellite remote sensing data, it has been found that there is a strong linear relationship between GPP and SIF in different flora. It can provide a more direct remote sensing method for GPP estimation. However, the correlation of canopy scale GPP and SIF will be subjected to environmental stress, canopy structure and plant work. In view of the influence of energy type and other factors, there are still great uncertainties in the GPP-SIF remote sensing estimation model. In view of this, this paper has carried out the sensitivity analysis of the GPP-SIF correlation to the photosynthetic physiological response to the important basic scientific problems of the SIF based GPP remote sensing estimation model and uncertainty, and the capacity ratio of the different band fluorescence estimation of GPP. Compared with three aspects of global GPP simulation and evaluation based on SIF satellite remote sensing signals, the main research contents and conclusions are as follows: 1) this paper first explored the potential of SIF detection of photosynthetic carbon sequestration and the difference characteristics of the coupling relationship between different photosynthetic functional GPPSIF. Based on the GPP and SIF760 data of synchronous observation, we found that in the crop canopy G The daily variation of PP, SIF760 is in accordance with the effective radiation of incident light, which shows a trend of high noon, early and late, and SIF760 signals can track the diurnal variation of plant photosynthesis. In winter wheat (C3 crop) and summer corn (C4 crop) canopy, GPP and SIF760 have a significant linear correlation, and further affirms that SIF remote sensing signals can be used. The ability of GPP is estimated directly. However, because of C3, the fluorescence quantum yield of C4 crops is close to _F, but the light energy utilization ratio is quite different, which makes the linear slope of GPP-SIF760 in the C4 crop is 2.16 times that of C3 crops (if the C3 crop photosynthesis noon break is removed, the C4 slope is 1.82 times C3). Therefore, the empirical model based on SIF is applied. F) when estimating crop canopy GPP, we should first consider the differentiation of plant photosynthesis, establish the GPP estimation model of C3, C4 crop canopy,.2). The estimation potential of SIF to GPP in red and far red light bands is explored by combining simulation and observation experiments. Many theoretical research and observation experiments have found that red and far red light bands are found. There is a certain difference in the characterization of photosynthetic physiological state by green fluorescence, and there is a great difference in the absorption and scattering characteristics of the radiation transmission in the canopy, but at present, the chlorophyll fluorescence of these two bands is different for the estimation potential of GPP. We first use SCOPE (Soil-Canopy-Observation of Photosynthesis and the Energy balan). CE) simulated diurnal variation data sets of GPP, SIF760, SIF688 of winter wheat canopy with different chlorophyll density Cab and leaf area index LAI. The statistical analysis shows that the correlation between GPP and SIF760 can still be expressed with the same linear model with the changes of Cab and LAI, but the GPP-SIF688 linear relationship is more sensitive to Cab and SIF760. The deformation of the vegetation index (especially NDVI_ (rededge), CI_ (rededge) and NDVI * WDRVI_ (rededge)) can be used as a good approximation of the red light escape ratio f_ (ESC) (688), which inhibits the sensitivity of GPP-SIF688 correlation to Cab and LAI, and makes the estimation ability of the linear model of winter wheat canopy. P-SIF760 is equivalent. However, considering the inaccurate inversion of red light fluorescence and the strong reabsorption of the canopy, far red fluorescence can provide a more reliable method for monitoring ecosystem photosynthesis. This paper has carried out a global GPP estimation study based on SIF satellite remote sensing. Many studies have shown that chlorophyll fluorescence remote sensing can be used directly as a direct method. A new method for estimating GPP, but the global scale GPP estimation based on SIF satellite remote sensing is still very weak. Based on the shared data of GOME-2 SIF740 and FLUXNET, this paper establishes a GPP-SIF regression model in different vegetation types, and estimates the global GPPSIF data set for 2007-2014 years. By comparing with MODIS GPP products, two data sets are found. The GPPSIF product algorithm is simple and does not require the introduction of auxiliary meteorological data and vegetation index data. However, GPPSIF is subject to the availability of data quality and flux data, while GPPMOD is limited to the accuracy of the model input data. Two products are in evergreen broad-leaved forests, tropical savage trees There was a certain degree of uncertainty in the grassland and shrub forest area. In addition, we also compared the correlation between C3, C4 crop GPPEC and SIF satellite remote sensing signals on the landscape scale, and found that the linear slope of the C4 crop is 1.81 times the slope of the C3 crop, which is basically consistent with the ground experimental results. The GPP-SIF linear model of the landscape scale has the function of plant photosynthesis. It is also sensitive and further confirms that the spatial and temporal distribution of C3, C4 plants is very important for accurate estimation of global GPP. The main innovative contributions of this paper include: 1) the difference characteristics of GPP-SIF coupling relationship between C3 and C4 crops are clarified for the first time by theoretical deduction and experimental observation. The response to photosynthetic physiological response to GPP-SIF correlation is uncertain. The theoretical calculation method based on the photosynthetic electron transfer rate (ETR) model for the response difference between C3 and C4 vegetation GPP-SIF was established. It was found that the C4 crop GPP was 1 times higher than that of the C3 crop under the same fluorescence production, and the synchronous observation experiment of the GPP and SIF daily changes and the GOME-2 satellite temporal remote sensing experiment were carried out, and the SIF quantitative analysis of SIF was made. The application potential and difference characteristics in the GPP estimation of C3 and C4 crops, the experimental measurement and the theoretical derivation are in agreement with.2). The potential of estimating GPP in different bands of SIF is studied for the first time through the combination of model simulation and field experiment. Because the red light fluorescence is difficult to obtain, the applications of fluorescence remote sensing mainly focus on the far red fluorescence. Based on the diurnal data of the SCOPE simulated data and the spectral resolution of the QE Pro spectrometer with a spectral resolution of 0.3 nm, this paper evaluates the ability of red and far red fluorescence to estimate winter wheat canopy GPP. The results show that the effect of the far red light on the canopy structure is relatively small, and is more suitable for the estimation of GPP and the physiological monitoring of photosynthetic physiology that should be used for satellite scale. .3) established the GPP-SIF remote sensing estimation model of different vegetation cover types on satellite scale, and developed the GPP estimation product directly driven by SIF satellite remote sensing data for the first time. Although the method of direct estimation of GPP based on SIF has attracted much attention, there is no GPP product based on chlorophyll fluorescence in all sphere terrestrial ecosystem. 2007-2 The 014 year global terrestrial ecosystem GPPSIF data set has a high consistency with the spatio-temporal pattern of the MOD17 GPP products. Moreover, the algorithm is simple and the mechanism is clear. It does not need the aid of meteorological data and vegetation index products. It can avoid the uncertainty of the introduction of auxiliary data to a certain extent. This study is to better understand the different time and space scale G The correlation mechanism between PP and SIF provides a theoretical basis, the experimental basis and model method, and the terrestrial ecosystem GPP products based on SIF satellite remote sensing data can be used to analyze the photosynthetic function state and ecological service ability of different types of ecosystems, which will be of great significance to the global carbon cycle research.
【学位授予单位】：中国科学院大学(中国科学院遥感与数字地球研究所)
【学位级别】：博士
【学位授予年份】：2017
【分类号】：S127

【相似文献】