【作者】 魏靖琼；

【导师】 柳小妮；

【作者基本信息】 甘肃农业大学 ， 草业科学， 2012， 硕士

【摘要】 草地是气候变化的承载者，既对生态环境变化反应敏感，又对气候变化产生反馈作用。由于气候变化和社会经济的快速发展，使草地生产力逐渐下降。本文基于CASA模型，利用2001~2010年的遥感数据（NDVI）、气象数据（月降水量、月均温和月太阳辐射），在GIS平台上估算甘肃省草地NPP。旨在为甘肃省草地资源的利用和保护提供基础数据。主要结论如下：1）在气象数据空间插值多元回归+残差分析法（Analytic Method based on MultipleRegression and Residues, AMMRR）的基础上，利用微观地形因子（坡度、坡向）修正的多元回归+残差分析法（Micro-topographic factors and Analytic Method based onMultiple Regression and Residues, M-AMMRR）对甘肃省Σθ和r进行插值，发现考虑微观地形因子的M-AMMRR法插值精度较高，插值效果更高，与实际更接近。2）修正了CASA模型的最大光能利用率（max）和水分胁迫系数，将实测数据和其它研究结果与CASA模型的模拟值进行对比，发现修正后的CASA模型模拟效果较好，能准确估算甘肃省草地NPP。3）2001~2010年，甘肃省草地NPP的年均值不同，年际平均值为133.87gC·m-2·a-1，变化范围在122.67~142.49gC·m-2·a-1之间。NPP总量的年际平均总量为3.62×1013gC·a-1，变化范围为3.32×1013~3.85×1013gC·a-1。甘肃省草地NPP的季相明显，夏季NPP总量最大为2.16×1013gC·a-1，占全年NPP的60.93%；而冬季NPP总量占全年NPP总量的3.15%，仅为1.12×1012gC·a-1。4）高寒草甸和山地草甸的NPP总量最高，为1.09×1013gC·a-1和9.79×1012gC·a-1，分别占甘肃草地面积的29.98%和26.96%。沼泽、暖性草丛、高寒荒漠和高寒草甸草原的NPP总量较低。甘肃省主要草地类型（低地草甸和温性荒漠）的NPP平均值较低，分别占草地总面积的31.96%和23.07%。5）甘肃省草地NPP由西南向东北逐渐减少，空间分布格局与水热条件紧密相关。其中NPP的高值区集中在甘南高原，低值区分布在北山山地。甘肃省草地NPP随年份不同总体呈下降趋势，在空间分布上有很大差异，约占草地总面积74.42%的区域的年际变化不显著；显著降低和增加的面积分别占草地总面积的15.86%和9.74%。6）气候因子对甘肃省草地NPP影响复杂，草地NPP的年均值与年降水量呈极显著正相关，与年均温呈不显著负相关，与太阳辐射呈弱的正相关；草地NPP与降水量的相关性明显高于温度与太阳辐射的。地形因子中，草地NPP与海拔高度呈极显著正相关，坡度和坡向对草地NPP的影响小。除农村人口外，其它人类活动因素与草地NPP呈不显著正相关。随人类活动的频繁，人类活动对草地NPP的影响逐渐加剧。更多还原

【Abstract】 As supporter of climatic change, grasslandgrassland not only is sensitive to entironmentchange, but also can feedback to climatic change. The productivity of grasslandgrassland hasbeen decreased gradually by reason of climatic change and rapid development of communityand economy.In order to supply basic data support for using and protection of grasslandgrasslandresource in Gansu province, we selected CASA model and its driven data of NDVI andmeteorological data contaning monthly precipitation, mean monthly temperature and monthlysolar radiation to estimate grasslandgrassland NPP in Gansu province on Arc GIS platform.The main conclusions were as following:1) It was found that the interpolation of Σθ and r with M-AMMRR method has moreprecision, realistic performance which took micro topographic factors such as slope andaspect into regression formula based on AMMRR method.2) According to the comparision between simulative grasslandgrassland NPP andobservated grasslandgrassland NPP or previous study results, we discovered that theimproved CASA model with perfectly sumilative performance, whose εmaxand water forcingcoefficient were modified can accurately estimate grasslandgrassland NPP in Gansu province.3) From2001to2010, the mean annual NPP of grasslandgrassland fluctuated in yearswith mean value of133.87gC·m-2·a-1, varing from minimum value of122.67gC·m-2·a-1tomaximum value of142.49gC·m-2·a-1, and its corresponding mean annual total NPP was3.62×1013gC·a-1, ranging from3.62×1013gC·a-1to3.85×1013gC·a-1. Seasonly variation ofgrasslandgrassland NPP was significant with total NPP (60.93%) in summer and total NPP(3.15%) in winter, respectively.4) High-cold meadow and Temperate montane meadow had the highest NPP with1.09×1013gC·a-1and9.79×1012gC·a-1, occupying29.98%and26.96%of mean annualgrasslandgrassland NPP in Gansu province, respectively. But, the preponderantgrasslandgrassland classes of Lowland meadow and Temperate desert had the lowest NPP,which accounted for31.96%and23.07%of mean annual grasslandgrassland NPP in Gansuprovince, respectively.5) The spatial distribution pattern of grasslandgrassland NPP represent declining trendfrom southwest to northwest, which was highly agreement with water and heat condition instudy area. There had significant difference of its spatial distribution with high NPP inGannan plateau and low NPP in north mountains. The spatial variation showed74.42%areaof mean annual grasslandgrassland NPP with non-significant changes,15.86%withsignificant decrease and9.74%with significant increase, respectively.6) There had extre-significantly positive, non-significantly negative and weakly positivecorrelations between mean annual grasslandgrassland NPP and annual precipitation, mean annual temperature and annual solar radiation. It showed a extre-significantly positivecorrelation between mean annual grasslandgrassland NPP and altitude, but with weakcorrelations to slope and aspect. Other human activities had non-significantly negativecorrelations with annual grasslandgrassland NPP exception for rural population.更多还原