中文摘要：

对地幔过渡带的研究表明，410km和660km速度间断界面是由于相变引起的不连续界面．穿越过渡带的深俯冲会引起这2个速度不连续面产生起伏．在中国东北等俯冲带地区，地震层析成像观测到了俯冲板片在660km界面上的平卧滞留，地震接收函数也观测到了这种滞留而产生的相变界面加深．然而在最近青藏高原南部展开的几组研究工作表明，地震接收函数并未在该区域发现不连续面存在明显起伏．为了验证该地区是否不存在穿越不连续面的俯冲，还是俯冲引起的不连续面的起伏没有被观测到，进行了该碰撞区域的热模拟，计算不连续面可能的起伏量．研究大角度俯冲、小角度俯冲和印度岩石层使青藏高原整体加厚等模型．在只考虑橄榄石-尖晶石相变作为410km不连续面产生原因、尖晶石-钙钛矿相变作为660km不连续面产生原因的前提下，计算结果表明，以上模型只要存在穿越以上界面的俯冲，就能够产生比较明显的起伏，这样的起伏应该能被地震接收函数观测到．根据以上的研究结果以及接收函数观测，更倾向于支持藏南区域应该不存在到达地幔过渡带的深俯冲，俯冲带可能向北到达更远的地方．

英文摘要：

The study of mantle transition-zone shows that velocity discontinuities of 410km and 660km were caused by phase-change. Subduction through these two discontinuities will change the depth of the discontinuities. A fiat subduction slab lying on the 660km was observed in North-East China subduction-zone by seismic imaging, and a deeper 660km discontinuity was reported by receiver function studies. But the receiver function results in Tibetan collocation zone does not show any significant elevation or depression of the discontinuities. To exam whether there is no subduction-slab through the discontinuities in these areas, or we have not observed the discontinuity caused by subduction-slab, we carry out a thermal simulation in the area, and compute the possible elevation and depression of the discontinuities. We have considered subduction models with different angles and a model the whole Tibetan plateau was thicken by Indian lithosphere. When we only consider 410km velocity discontinuity, and Olivine-Perovskite phase change the Olivine-Spinel phase change as the cause of as the cause of the 660km velocity discontinuity, our result shows that the subduction similar to our models will change the depth of discontinuities significantly, and could be observed by seismic receiver function method. According to our results and the receiver function results, we tend to support the opinion that there does not exist deep subduction slab in Southern Tibet, and the subduction may go further north.