Comparison Of Global Snow Density Measurements Using Different Methods

Snow density plays a crucial role in most of the snow properties,such as mechanical?e.g.,snow load estimate?,thermal?e.g.,thermal conductivity?,optical?e.g.,light transmission and the extinction coefficient?,and electromagnetic property.Additionally,assessment of snow hydrology is heavily dependent on snow density.The precise snow density measurements in the field are basis to provide a better understanding of the snow related researches.However,there are few studies have concentrated on the comparison of snow density measurements,especially quantitative assessment for the accuracy and stability of it.Thus,this study selected four snow density measurements,all of which are most widely used in different regions of the world.?1?Model VS-43 snow density gauge?most widely used in China and the Former Soviet Union?,?2?Federal sampler?most widely used in North America?,?3?Snow fork?global?,and?4?1000 cm³ wedge snow density cutter?global?.During the 2015-2017,our study did a lot of field work in the Altay Mountains of Xinjiang province?34 snow profile,1284 snow sampler,90 snow course,and 177 snow depth survey?,which help us?1?to understand the spatial characteristics of snow properties,and based on it,?2?to compare the four snow density measurements.In this study,snow depth?>40 cm?and snow water equivalent?>6 cm?in northwest of the Altay Mountains of Xinjiang province was larger than which of southeast?snow depth:20-40 cm,snow water equivalent<6 cm?and pediment?snow depth<20 cm,snow water equivalent<6 cm?of the Altay Mountains of Xinjiang province.It is revealed that snow depth and snow water equivalent are increasing with elevation?40 cm/km and 10 cm/km,respectively?and slope increasing.Overall,snow depth,snow water equivalent,and snow density in north slopes>south slope,margin of forests>open area in forest>forest.Under the different vegetation conditions,the difference of snow depth and snow water equivalent was 10%-65%,and the difference of snow density was about 5%-30%.Set mean value of the four snow density measurements as reference.Overall,Model VS-43 snow density gauge,Federal sampler,and 1000 cm³ wedge snow density cutter overestimated snow density 1.6%,1.7%,and 8.1%,respectively,with the RMSE of 11 kg m^-3,13 kg m^-3,and 21 kg m^-3.However,Snow fork underestimated snow density 11.4%,with the RMSE of 27 kg m^-3.In grassland,Model VS-43 snow density gauge and Federal sampler overestimated snow density 1.2%and 2.2%,respectively.In forest,however,Model VS-43 snow density gauge and Federal sampler underestimated snow density 0.4%and 1.8%,respectively.There no markedly different between the grassland and forest for Snow fork and 1000 cm³wedge snow density cutter,however,causes larger error in the snow layer of depth hoar than other layers.Our measurements presented remarkable difference of snow density collected by different methods with the RPE range of-35.4%–36.3%.Snow fork showed stable performance for the observed snow density range(40-330 kg m^-3),with the absolute deviation no more than 40 kg m^-3.Though the second sampling was required,which was the main source of error,Model VS-43 snow density gauge still performed slightly stably than Federal sampler based on our measurements.