【作者】 谢恩；

【导师】 严冬；

【作者基本信息】 华中科技大学 ， 水利工程， 2024， 硕士

【摘要】 深刻认识干旱流域生态引水与生态系统固碳量之间的关系,并通过水资源管理提升干旱流域生态系统碳吸收能力,对缓解干旱区社会经济和生态环境用水需求与水量供给间的矛盾等方面具有重要的理论和实践意义。但现有研究未能提供干旱流域水量与碳吸收间关系的精确描述。鉴于地下水数据的缺乏,本文专注于干旱区生态引水与植被总初级生产力(Gross Primary Productivity,GPP)的关联这一核心问题,从宏观流域到微观网格两个层面依次开展了研究,并分析了该影响在这两个尺度上的关联性。为此,以新疆塔里木河干流上中游流域为宏观分析的研究对象,分析了生态引水对GPP的滞后和累积效应;接着,基于遥感信息估算了生态闸门引水的漫溢水深分布,并以此为基础建立了精细空间尺度下的GPP预测模型,研究了生态引水与GPP的动态关联。本文主要研究内容及成果如下:(1)通过分析2005～2018年塔河干流上中游总生态引水对GPP的影响发现,在引水期间,GPP发生了显著的时空变化。流域尺度GPP对生态引水有1个月的滞后响应,二者正相关性约为0.59,累积2个月水量与GPP的相关性高达0.63。这些结果均从宏观角度表明了通过闸门实施生态引水的有效性。(2)针对资料缺乏地区,通过GIS集成技术,提出了一种基于遥感影像和DEM的水深估算方法。将此方法运用于塔河中游左岸一典型漫溢型生态闸门的供水区域,计算出该区域的水深空间分布。通过对比实际供水量和估算水体体积,展示了这种方法在反映实际水深空间分布方面的合理性。(3)以上述典型闸门为例,进一步模拟了在500 m空间和月度时间分辨率下,GPP对漫溢水深、温度等环境变量动态变化的响应。采用了弹性网回归模型分析漫溢水淹区域,回归系数表明前1个月GPP和日长对GPP的影响最显著,而水深变量的影响程度相对较小,这主要是受到引水频率较低的影响。另外针对闸门供水区全域,本文基于随机森林建立了考虑网格空间关系的GPP预测模型。该模型的变量相对重要性显示,大多数空间变量(即周围网格的加权平均)的重要性超过了相应的非空间变量,并且水分的时间累积效应及空间滞后作用对GPP产生了显著影响。上述两个尺度下的研究均凸显了水分累积作用对GPP的重要影响。此外,网格尺度中空间变量的重要性也揭示了在流域尺度下,该影响很大程度上是通过空间作用来实现的。更多还原

【Abstract】 A deep understanding of the relationship between Ecological Water Diversion(EWD)and ecosystem carbon sequestration in arid basins,and the enhancement of the carbon sequestration capacity of arid basin ecosystems through water resource management,are of great theoretical and practical significance.However,existing studies have failed to provide a precise description of the relationship between EWD and ecosystem carbon sequestration status in arid basins.To this end,this paper focuses on the core issue of the impact of EWD on vegetation Gross Primary Productivity(GPP)in arid regions,conducting research sequentially from the macro basin scale to the micro grid scale,and analyzing the correlation of this impact across these two scales.For this purpose,the upper and middle reaches of the Tarim River in Xinjiang is selected as the subjects for macro-scale analysis,examining the lagged and cumulative effects of EWD on GPP.Subsequently,based on remote sensing data,the overflow depth distribution from ecological gate diversions is estimated,and on this foundation,a fine spatial scale GPP prediction model is established to study the dynamic relationship between ecological water diversion and GPP.The main research content and results of this paper are as follows:(1)By analyzing the impact of EWD on GPP in the upper and middle reaches of the mainstem of the Tarim River from 2005 to 2018,it is found that GPP has significant temporal and spatial variations during the period.In addition,GPP shows a one-month lagged response to EWD,with a correlation coefficient of about 0.59.Furthermore,the correlation between water accumulated over two months and GPP reaches as high as 0.63.These results collectively demonstrate the effectiveness of implementing EWD through gates from a macro perspective.(2)A water depth estimation method is proposed for areas lacking information,based on remote sensing images and DEM,through the integration of GIS technology.This method is applied to the water supply area of a typical diffuse ecological gate on the left bank of the middle reaches of the Tarim River,and the spatial distribution of water depth in this area is calculated.The reasonableness of this method in reflecting the spatial distribution of the actual water body is showed by comparing the actual water supply volume and the estimated water volume.(3)Using the typical gate as an example,this study further simulates the dynamic response of GPP to environmental variables such as overflow depth and temperature at a spatial resolution of 500 meters and a monthly temporal resolution.The elastic network regression model is used to analyze the flooded areas,and the coefficients indicate that lastmo＿GPP and daylength have the most significant effect on GPP,while the effects of water depth are relatively minor.This is primarily due to the infrequency of EWD.For the entire study area,this paper establishes a GPP prediction model based on random forests that considers spatial relationships between grids.The results indicate that most spatial variables are more important than their corresponding non-spatial variables,and that the cumulative temporal effects of water and its spatial lag significantly impact GPP.Evidence shows that both models have good predictive accuracy.Research at both scales highlights the significant impact of water accumulation on GPP.Additionally,the importance of spatial variables at the grid scale reveals that this impact is largely realized through spatial effects at the basin scale.更多还原