Real time hydrologic forecasting requires near accurate initial condition of soil moisture; however, continuous monitoring of soil moisture is not operational in many regions, such as, in Ganga basin, extended in Nepal, India and Bangladesh. Here, we examine the impacts of perturbation/ error in the initial soil moisture conditions on simulated soil moisture and streamflow in Ganga basin and its propagation, during the summer monsoon season (June to September). This provides information regarding the required minimum duration of model simulation for attaining the model stability. We use the Variable Infiltration Capacity model for hydrological simulations after validation. Multiple hydrologic simulations are performed, each of 21 days, initialized on every 5^th day of the monsoon season for deficit, surplus and normal monsoon years. Each of these simulations is performed with the initial soil moisture condition obtained from long term runs along with positive and negative perturbations. The time required for the convergence of initial errors is obtained for all the cases. We find a quick convergence for the year with high rainfall as well as for the wet spells within a season. We further find high spatial variations in the time required for convergence; the region with high precipitation such as Lower Ganga basin attains convergence at a faster rate. Furthermore, deeper soil layers need more time for convergence. Our analysis is the first attempt on understanding the sensitivity of hydrological simulations of Ganga basin on initial soil moisture conditions. The results obtained here may be useful in understanding the spin-up requirements for operational hydrologic forecasts.

实时水文预报需要接近准确的土壤水分初始条件；但是，在许多地区，例如在甘加盆地，尼泊尔，印度和孟加拉国，都没有对土壤水分进行连续监测的工作。在这里，我们研究了夏季季风季节（6月至9月）初始土壤水分条件的扰动/误差对恒河盆地模拟土壤水分和水流及其传播的影响。这提供了有关达到模型稳定性所需的最小模型仿真持续时间的信息。验证后，我们使用可变渗透能力模型进行水文模拟。每21天进行一次多次水文模拟，每5^日初始化一次赤字，盈余和正常季风年份的季风季节的一天。这些模拟中的每一个都是从长期运行获得的初始土壤湿度条件以及正负扰动下进行的。对于所有情况，都获得了收敛初始误差所需的时间。我们发现高降雨量的一年以及一个季节内的潮湿季节迅速收敛。我们还发现收敛所需的时间有很大的空间变化；下甘加盆地等降水高的地区收敛速度更快。此外，更深的土壤层需要更多的时间来收敛。我们的分析是首次尝试了解恒河盆地水文模拟对初始土壤湿度条件的敏感性。