【作者】 刘慧；

【导师】 胡和平；

【作者基本信息】 清华大学 ， 水利工程， 2012， 博士

【摘要】 趋势性气候变化引起的物候变化和极端气候条件产生的干旱对于陆地植被初级生产力有显著影响，进而影响人类生产生活和地球物质循环。土壤水分是影响物候变化和干旱的重要要素，且是影响植被生长的直接水分要素，然而区域尺度的植被物候模型对水分条件的考虑比较粗略，往往采用降水来表征。论文将土壤水分作为影响植被物候的关键环境要素，结合大尺度水文模型和遥感信息，建立综合考虑土壤水分和气温的草地返青物候模型和植被干旱模型，对区域尺度物候变化和植被干旱分析有重要意义。基于遥感归一化植被指数，应用逻辑斯蒂模型对研究区域草地的主要物候现象进行识别，揭示了内蒙古高原草地和青藏高原草地的物候特征，分析了研究区域物候及其影响要素变化趋势，并探讨了干旱条件下逻辑斯蒂模型的适用性。基于不同环境要素对草地返青的影响分析，提出将水文模拟的土壤水分数据用于物候分析，借鉴热量控制的潜在返青日期指标的计算方法，构建了由土壤水分控制的潜在返青日期指标和由降水控制的潜在返青日期指标，论证分析了在大尺度物候研究中土壤水分指标和降水指标对水分条件的代表性。基于上述潜在返青日期指标体系，建立了草地返青模型和草地返青主导要素分析方法。构建的草地返青模型克服了固定影响要素累加时段的问题，草地返青主导要素分析方法能够描述草地返青主导要素的年际波动和多年主导要素定量分析，并可用于未来气候变化情景下草地返青影响要素的演变分析。将模型应用于内蒙古高原草地和青藏高原草地，应用结果显示两个研究区域的草地返青主导要素多数为水分条件而非热量条件，在选定的未来气候变化情景下，内蒙古高原草地多数站点的主导要素对草地返青的控制作用将增强。提出归一化植被指数百分位数，将其应用于干旱分析的时空强度综合法，重建了科罗拉多流域的植被干旱事件，对比了植被干旱和土壤水分干旱的差异，发现归一化植被指数百分位数是描述植被干旱更直接的指标。分析了不同月份和不同植被覆盖类型下植被指标和土壤水分指标的相关关系，提出了植被对干旱的抵抗力和恢复力的计算方法，并比较分析了多种植被对干旱的响应。更多还原

【Abstract】 As an important factor for human activities and material cycle, Net Primary Productivity (NPP) of terrestrial ecosystem is dramatically influenced by phenology change and drought, which are easily disturbed and induced by tendentious climate change and climate extremes. Soil moisture is an important variable that related to pehnology change and drought. Existing regional phenology models do not provide enough consideration on water conditions. Precipitation is frequently used instead of soil moisture which, however, is the more direct variable that the growth of vegetation depend on. By taking soil moisture as a critical environmental variable, based on regional hydrological model and remote sensing, a regional grassland phenology model and vegetation drought reconstruction model were built up, which will be a meaningful work on regional phenology model and vegetation drought research.A logarithmic model was used to identify the main phenomenon of grassland phenology in the study area. The model was based on the NDVI data retrieved from AVHRR. The main characteristics of phenology over Inner Mongolia and Qinghai-Tibetan Plateau grasslands over China were disclosed. Analyses were carried out on the nonstationarities of phenology and its impact factors. The applicability of the model was also discussed.Based on the analyses of phenology factors, Soil moisture was found to be a better element for phenology analysis. Inspired by the idea of TSO (Thermal Spring Onset), An index was made up for potential vegetation spring-up date controlled by soil moisture, SMSO(Soil Moisture Spring green-up Onset date) and PSO (Precipitation Spring green-up Onset date) to stand for the spring-up date controlled by precipitation. Results proved that the index based on soil moisture is superior to the precipitation index in regional phenology study.Based on the index proposed above, a grassland spring phenology model and the grassland spring phenology dominant factor model were established, and the models were then applied over Inner Mongolia and Qinghai-Tibetan Plateau grasslands. The grassland spring phenology model avoided the drawbacks of fixed accumulation span. The grassland spring phenology dominant factor model could depict the decadal shift of dominant factor and quantitatively describe the dominant role of factors. It could also be used to analyze the evolution of dominant factors with the projection of future climate change. Results showed that the dominant factor in both grasslands (Inner Mongolia and Qinghai-Tibetan Plateau) is water condition instead of thermal condition. The control of the current dominant factors is inclined to be reinforced under the SRES B2scenario.An index called NDVIP (NDVI Percentile) was proposed based on NDVI retrieved from MODIS, and a vegetation drought reconstruction model was established. The index was used to reconstruct the drought events over Colorado River Basin. Comparative studies showed that NDVIP is a more direct index for the description of drought. Resistance and recovery abilities of vegetation under drought were also defined and calculated. Different types of vegetations showed different responses to drought events.更多还原