Accurate estimation of roof snow load is important for safely ensuring the structural performance of buildings in snowy regions. Ground snow weight is a basic design factor used to calculate the roof snow load. In a previous model, the ground snow weight was estimated based on daily precipitation and average temperature. However, this model has a limitation in terms of its universality because it is necessary to optimize the coefficients used in the model by comparing the calculated result with observation data in advance.
This study aims to develop a versatile and universally app licable method to estimate ground snow weight based on meteorological observation data and a heat balance model. The proposed model is based on the concept that the snow weight can be estimated by subtracting the snowmelt runoff from the total precipitation. The model proposed by Kondo and Yamazaki (1987) was used to calculate the amount of snowmelt. The ground snow weights were estimated through the proposed model using detailed meteorological observation data for Nagaoka, which is representative of a snowy city in Japan.
According to the estimation results, the threshold temperature of rain/snow, TSR , used in the model did not significantly affect the error in the range of 1.1 to 2.0 °C. Additionally, the maximum water content values, CWVMAX, of 5% and 10% did not show large differences in the estimation error. For the eight winter examples examined in this study, the average normalized mean square error (NMSE) was the smallest when TSR was 1.4 °C and the maximum water content was 5%. Although the accuracy varies with the winter, the variations in the daily ground snow weight and 7-day increasing snow weight were correctly estimated with an error of approximately 20%. The maximum values of each winter were predicted within an error of approximately 10%.
By analyzing the heat balance components obtained by the model, it was confirmed that the snowmelt energy due to radiation mainly contributed to the decreasing snow weight. Under the conditions of the present study, the contributions of sensible and latent heat transport were small. An estimation method based on the heat balance model, as developed in this study, can be helpful for understanding the factors contributing to the variation in the snow weight. This is a great advantage of such a model.
The present method has a limitation in its applicability; that is, all meteorological factors required in the model are only available at very limited sites such as meteorological offices. To overcome this limitation, a method was examined through which unavailable meteorological factors were estimated by meteorological factors measured at AMeDAS observation points; this method was similar to extending AMeDAS meteorological data. As a result, it was shown that the present method can be applied with sufficient accuracy even at sites where only AMeDAS meteorological data are available.
