The Doam watershed is located at alpine areas and the annual average precipitation, including snow accumulation, is significant higher than other areas. Thus, pollutant laden runoff and sediment discharge from the alpine agricultural fields are causing water quality degradation at the Doam watershed. To estimate soil erosion from the agricultural fields, the Universal Soil Loss Equation (USLE) has been widely used because of its simplicity to use. In the early spring at the Doam watershed, the stream flow increases because of snow melt, which results in erosion of loosened soil experiencing freezing and thaw during the winter. Also, extremely torrential rainfall, such as the typhoons 'RUSA' in 2002 and 'MAEMI' in 2003, caused significant amounts of soil erosion and sediment at the Doam watershed. However, the USLE model cannot simulate impacts on soil erosion of freezing and thaw of the soil. It cannot estimate sediment yield from a single torrential rainfall event. Also, it cannot simulate temporal changes in USLE input parameters. Thus, the Soil and Water Assessment Tool (SWAT) model was investigated for its applicability to estimate soil erosion at the Doam watershed, instead of the widely used USLE model. The SWAT hydrology and erosion/sediment components were validated after calibration of the hydrologic component. The R$^2$ and Nash-Sutcliffe coefficient values are higher enough, thus it is found the SWAT model can be efficiently used to simulate hydrology and sediment yield at the Doam watershed. The effects of snow melt on SWAT estimated stream flow and sediment were investigated using long-term precipitation and temperature data at the Doam watershed. It was found significant amount of flow and sediment in the spring are contributed by melting snow accumulated during the winter. Two typhoons in 2002 and 2003, MAEMI and RUSA, caused 33% and 22% of total sediment yields at the Doam watershed, respectively. Thus, it is recommended that the SWAT model, capable of simulating snow melt, sediment yield from a single storm event, and long-term weather data, needs to be used in estimating soil erosion at alpine agricultural areas to develop successful soil erosion management instead of the USLE.