new temperature based method to separate rain and snow

This paper presents the development and testing of a new method to estimate daily snowfall from precipitation and associated temperature records. The new method requires two variables; the threshold mean daily air temperature at which 50% of precipitation is considered snow, and the temperature rang...

Full description

Saved in:
Bibliographic Details
Published in:Hydrological processes 2008-12, Vol.22 (26), p.5067-5085
Main Author: Kienzle, Stefan W
Format: Article
Language:English
Subjects:
Earth sciences
Earth, ocean, space
Exact sciences and technology
External geophysics
Hydrology
Hydrology. Hydrogeology
mean annual snowfall
model intercomparison analysis
precipitation separation
rain
sensitivity analysis
snow
Snow. Ice. Glaciers
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:This paper presents the development and testing of a new method to estimate daily snowfall from precipitation and associated temperature records. The new method requires two variables; the threshold mean daily air temperature at which 50% of precipitation is considered snow, and the temperature range within which mixed precipitation can occur. Sensitivity analyses using 15 climate stations across south-western Alberta, Canada, and ranging from prairie to alpine regions investigates the sensitivity of those two variables on mean annual snowfall (MAS), the coefficient of determination, and the MAS-weighted coefficient of determination. Existing methods, including the static threshold method, one linear transition method used by Quick and Pipes, and the Leavesley method employed in the PRMS hydrological modelling system are compared with the new method, using a total of 963 years of daily data from the 15 climate stations used for the sensitivity analyses. Four different approaches to using the two input variables (threshold temperature and range) were tested and statistically compared: mean annual variables based on the 15 stations, mean annual variables for each station, mean monthly variables for each station, and a sine curve representing seasonal variation of the variables. In almost all cases the proposed new method resulted in higher MAS-weighted coefficients of determination, and, on average, they were significantly different from those of other methods. The paper concludes with a decision tree to help decide which method and approach to apply under a variety of data availabilities. Copyright © 2008 John Wiley & Sons, Ltd.
ISSN:0885-6087
1099-1085
DOI:10.1002/hyp.7131