中文摘要：

在经典的对流层层析模型基础上提出一种新的层析模型。首先采用非迭代重构算法估算出层析网格顶点处的水汽密度值，然后根据网格顶点处的湿度信息内插出网格中心点处的水汽密度值，并将其作为初始值，通过迭代重构算法对其进行再次修正。组合重构算法能够为迭代重构算法提供高质量、高可靠性的初始值。试算结果表明，组合重构算法相比单一重构算法具有更高的精度。

英文摘要：

We put forward a new tomography model based on the classical tropospheric tomography model. To begin with, the water vapor densities at the vertexes of voxel are estimated using a non-it- erative reconstruction algorithm （NIRA）. Then, the humidity information at the center of the voxel is derived from NIRA-results through interpolation method, and as initial value was amended again by an iterative reconstruction algorithm. The combined reconstruction algorithm has the advantage of providing initial values which are high quality and reliable in the iterative reconstruction algorithm. Test results show that its accuracy is better than a single reconstruction algorithm.