比较了汊点水位预测-校正法（junction-point water stage prediction and correction, JPWSPC）和经典的分级解法，在处理缓流河网汊点处回流效应时的不同，并对比了它们的计算效率。应用这2类方法时，Saint-Venant方程组都采用Preissmann格式离散，生成的非线性离散方程用Newton-Raphson方法求解。比较表明：这2类方法都能处理普适河网，JPWSPC法无需求解整体连接矩阵，同时不会增加每一时间步的迭代次数，因而节约了系统内存，提高了计算效率；河网中河段数目越多，JPWSPC法的效率优势越明显。

This paper compares a newly proposed method, namely the junction-point water stage prediction and correction (JPWSPC) method, for gradually varied subcritical flow simulation in general open-channel networks with conventional three-phase algorithms. Emphasis is on their difference in backwater effect treatment and efficiency. When the two classes of algorithms are implemented, the Saint-Venant equations are discretized using the Preissmann scheme and the resulting nonlinear equation system is solved using the Newton-Raphson method. The comparison shows that the two classes of methods are applicable in general networks, that, furthermore, the JPWSPC method saves computer memory and computation time by avoiding to establish and solve the global matrix without increasing the required number of iterations per time step, and that the advantage of the JPWSPC method over the three-phase algorithms increases with increase of the number of junction points.