黄土高原作为中华民族五千年文明的发祥地，历史上曾经森林繁茂，山清水秀，而后来一度成为世界上水土流失最严重的地区。自1999年实施退耕还林政策之后，黄土丘陵沟壑区的土地利用发生了巨大的变化，黄土丘陵小流域土地利用变化及其生态效应研究具有重要的理论和实践意义。本文以黄土丘陵沟壑区小流域作为研究区域，以传统统计学方法为主要技术手段，结合地理信息系统和遥感方法，对小流域的生物量全碳储量和土壤全碳储量的时空变异进行定量研究，尝试建立其估算模型。主要结论如下：（1）小流域地上生物量的分布格局主要受到土地利用类型、植被层次、地形和土壤理化性质的影响。土地利用类型以林地>灌木地>果园>幼林地>荒草地>农地；植被不同层次的地上生物量，乔木层>草本层>灌木层>半灌木层。从地形因子看，低海拔、陡坡、凹坡的地段，地上生物量相对较高；从土壤因子看，速效氮、速效钾、速效磷、有机质、表层土壤全碳、表层土壤总氮含量越高的地方，植被地上生物量就越高。小流域地下生物量主要受到土地利用类型、地形和土壤养分因子的影响，土地利用类型以林地>幼林地>灌木地>果园>荒草>农地；地形越凹、土壤养分越充足，地下生物量就越高。地上植被含碳率，主要受到物种、地形、土壤含水量的影响，在地上物种含碳率较高（沙棘，柠条等），海拔较低、坡形较凹的地段，水分和养分相对较好，人工植被相对较多，植物生长相对较快，植被含碳率较低。地上生物量回归模型以通用多元线性回归的时空预测模型效果最佳，但参数较多，地下生物量回归模型以通用多元线性回归的空间预测模型效果较好，植被碳储量回归模型以通用多元线性回归的时空估算模型为最佳。（2）土壤表层有机碳含量主要受到土地利用类型、地形、土壤氮含量、地上植被类型以及退耕时间的影响。土地利用类型以灌木地>林地>果园>荒草地>幼林地>农地；垂直坡形越凹、海拔越高、退耕时间越早，越有利于土壤中有机碳的积累；土壤中氮元素含量越高，土壤表层有机碳的含量也越高，土壤中的氮元素含量成为有机碳积累的一个限制因素；地上植被类型对土壤有机碳的含量有显著地影响；随着退耕年份的增加，土壤有机碳的含量呈现增加的趋势。土壤全碳含量主要受到土地利用类型、土壤深度和退耕年份的影响。就土地利用类型来看，土壤全碳含量以灌木地>林地>荒草地>幼林地>果园>农地；随着土壤深度的增加而减少，土壤全碳水平空间变异较小；土壤全碳含量随着退耕年份的增加而增加，最后趋于稳定的转变过程。土壤碳储量主要受到土地利用类型和地形的影响。在不同的土地利用类型上，灌木地>荒草地>幼林地>林地>果园>农地；坡度越大、海拔越高，土壤碳储量的累积量会增高。土壤碳储量的多元回归模型优于其余的回归模型，由于研究区的高度空间异质性，其本身的解释精度并不是很好。（3）基于遥感估算中，分区配准的精度要高于整体配准的精度，通过遥感估算得出，大南沟小流域的地上生物量总量为1.4*10的7次方kg。地上植被碳储量为5.8*10的6次方 kg。

Loess Plateau as the birthplace of five thousand years of Chinese civilization, there were a lot of trees in history, and later became the world's most serious soil erosion areas. Since 1999, by returning farmland to forest policy , the land use of the loess hilly region were changed a lot, the ecological effects of this change is worth to be studied. Combining the method of Remote Sensing , GIS and traditional statistic method. After study the distribution of the biomass and total carbon, and then the storage of the total carbon storage of the small catchment is estimated. There are several important results:(1) The distribution of the above ground biomass varied large on different land use ,different vegetation layers, land form feature and soil properties. The biomass based on different land use, the woodland>shrub’s>orchard’s>young forest’s>grass land>farmland’s ;The biomass based on the different layer, the forest’s layer>grass’s>shrub’s>half shrub’s; Locations that have concave, lower altitude, large slope and higher soil nutrient are with higher above ground biomass. The biomass of the below ground is affected by the land use type, land form feature and soil nutrient. The biomass based on different land use, the woodland> young forest’s>shrub’s>orchard’s> grass land>farmland’s ;Locations that have concave and higher soil nutrient are with higher below ground biomass.The C ration of the vegetation is correlated to the type of land use, the land form feature and soil moisture. The C ration of the grass land >shrub’s> young forest’s> forest’s>farm land’s>orchard’s. At the same time , the C rate of the vegetation is affected by the land form feature and soil moisture. Locations that have concave and higher soil moisture are with lower C rate. In all sorts of the models developed to predict above ground biomass , the spatio-temporal generalized multiple-linear regression models performed best but with more parameters in which most of them has poor significance. In all sorts of the models developed to predict below ground biomass, the spacial generalized multiple-linear regression models performed best.