Snowfall/Snow depth | 降雪量/降雪深度

• MPAS | WRF | precipitation calculation

---

中国气象局

• 降雪量与降雪深度 | 中国气象报社
• 降雪的形成条件及等级划分 | 中国气象报社
• 科普 | 降雪量≠积雪深度？

降雪量与降雪深度是两个不同的概念，降雪量并不是指积雪深度。

• 在气象观测上，降水量是指某一时段内的未经蒸发、渗透、流失的降水，在水平面上积累的深度，以毫米为单位，取一位小数。如果是降雪量，则指溶化后的水，在水平面上积累的深度，对于降雪量的大小需要用统一规定的标准进行界定，它是气象观测员用标准的降水观测容器（俗称雨量器），将收集到的雪水融化后测量出的量度，以毫米为单位，亦称降水量。目前气象部门安装有固态降水自动观测站,可适时自动观测降雪量。
• 而积雪深度是指从积雪表面到地面的积雪垂直深度，是指气象员使用量雪尺在观测场内或附近平坦、开阔的地方，当落地雪尚未融化时，测得的积雪的深度（厚度），以厘米为单位。

当然，降雪量与降雪深度之间存在一定的关系，下雪大，积雪必定深厚，此外积雪厚度还与气温、风等气象要素相连，温度高，积雪很快就会融化，而温度低，积雪就不易融化甚至维持一段较长时间了。

• 按照降水量强度，降雪分为小雪、中雪、大雪和暴雪四个等级。
  • 小雪：是指下雪时水平能见距离等于或大于1000米，地面积雪深度在3厘米以下，降水量级为24小时降雪量在0.1～2.4毫米之间。
  • 中雪：是指下雪时水平能见距离在500～1000米之间，地面积雪深度为3～5厘米，24小时降雪量达2.5～4.9毫米。
  • 大雪：是指下雪时能见度很差，水平能见距离小于500米，地面积雪深度等于或大于5厘米，24小时降雪量达5.0～9.9毫米。
  • 如果有降雪而没有形成积雪，一般称之为“零星小雪”。
  • 当24小时降雪量达到10.0～19.9毫米时为暴雪，20.0～29.9毫米为大暴雪，超过30.0毫米为特大暴雪。

当降雪落地后无融化时，一般而言，在北方地区 1毫米降雪可形成的积雪深度有8～10毫米，在南方地区积雪深度有6～8毫米。

日常生活中流传的“1毫米降雪量对应1厘米积雪深度”，其实是一个非常粗略的经验值，并非科学意义上的固定换算公式。两者的实际比例，核心取决于雪的“蓬松度”，而雪的蓬松度又由雪花的类型和形成环境决定。

雪花的本质是冰晶，不同条件下形成的雪花，密度差异极大。我们可以把雪花分为干雪和湿雪两类：干雪通常形成于气温较低的环境（一般在-10℃以下），这类雪花晶体结构疏松，内部空隙多，含水量低，蓬松度很高。此时，1毫米的降雪量对应的积雪深度可能达到0.9厘米甚至1.2厘米；而湿雪则形成于气温接近0℃的环境，雪花在降落过程中会部分融化，又重新冻结，晶体结构紧密，含水量高，蓬松度低。这种情况下，1毫米的降雪量对应的积雪深度可能只有0.6厘米到0.8厘米左右。

• 干雪（气温低、雪粒松散）：密度小，1毫米降雪量大概能积1-1.2厘米积雪，手感蓬松，适合堆雪人；
• 湿雪（气温接近0℃、含水分多）：密度大，1毫米降雪量可能只积0.5-0.8厘米积雪，手感沉重，容易压断树枝。

Figures

降水量

• mm of water equivalent

https://www.nmc.cn/publish/precipitation/1-day.html

https://www.nmc.cn/publish/precipitation/1-day.html

colormap

#ffffff, #cccccc, #a2a2a2, #707070, #474446, #753eee, #4e006f
0.01, 2.5, 5.0, 10.0, 20.0, 30.0

NWP

In NWP models like WRF and MPAS, these three variables represent different physical aspects of snow. Understanding the difference between a state variable (what is there now/instantaneous) and an accumulated variable (what has fallen over time) is key.

Variable	Full Name	Units	Category	Description
SNOWC	Snow Cover	Fraction (\(0–1\))	State	The horizontal area of the grid cell covered by snow.
SNOWH	Snow Depth	Meters (\(m\))	State	The actual physical height (thickness) of the snowpack on the ground.
ACSNOW	Accumulated Snow	\(kg/m^2\) (or \(mm\))	Accumulated	The total amount of snow that has fallen since the model start (Water Equivalent).

---

1. SNOWC (Snow Cover)

• What it represents: The "horizontal" extent. If SNOWC is \(0.5\) , it means \(50\%\) of that grid box is covered in snow and \(50\%\) is bare soil or vegetation.
• Why it matters: It is used by the Land Surface Model (LSM) to calculate the Grid Albedo. If the ground is only half-covered, the model will blend the high albedo of snow with the lower albedo of the soil.
• Common trap: Even if you have a very deep snowpack (\(1m\)), SNOWC can never exceed \(1\).

2. SNOWH (Snow Depth)

• What it represents: The "vertical" extent. This is what you would measure with a ruler in your backyard.
• Dynamic Nature: Unlike the variables that track water mass, SNOWH changes due to compaction. Over time, the model (especially Noah-MP) simulates the snow settling under its own weight, so the depth may decrease even if no melting occurs.
• Note: If you are comparing model data to weather station observations, this is the variable you want.

3. ACSNOW (Accumulated Snowfall)

• What it represents: The "history" of the run. This is a bucket that collects every snowflake that hits the ground.
• The Key Difference:
• SNOW (SWE) tells you how much snow is on the ground right now. If the snow melts, SNOW goes down.

• ACSNOW tells you how much has fallen in total. If the snow melts, ACSNOW does not go down. It only increases.

• Units: While it's called "snow," it is almost always stored as Liquid Water Equivalent. To get the "accumulated depth" in inches or cm, you would need to divide by a snow-to-liquid ratio (e.g., \(10:1\)).

How they work together

Imagine a storm where \(10\text{ mm}\) of water equivalent falls as snow.

1. ACSNOW jumps from \(0\) to \(10\)
2. SNOWH jumps from \(0\) to \(0.1\text{ m}\) (assuming a \(10:1\) ratio).
3. SNOWC jumps from \(0\) to \(1.0\).

Six hours later, if half of that snow melts:

1. ACSNOW stays at \(10\) (it records the event, not the current state).
2. SNOWH drops to \(0.05\text{ m}\).
3. SNOWC might stay at \(1.0\) (if the ground is still fully covered) or drop slightly if the melting is patchy.

WRF

• How to get hourly snowfall? | 2023
  • If you're looking at a case that only had snowfall (no liquid precipitation), then you can use the value of RAINC + RAINNC because these equal total convective and non-convective precipitation (including snow and ice). Otherwise, yes, I believe ACSNOW would be the way to go. It is initialized to zero at the model initial time and then accumulates over time. Regarding melting, this variable is handled differently in various surface physics schemes. For whichever surface scheme you're using, you can take a look at specific code in the WRF/phys/module_sf_<scheme>.F file (for e.g., if you're using Noah, look in module_sf_noahdrv.F). Another option would be the SNOWH variable, which is the physical snow depth. You would need to subtract the depth at the previous hour so determine the hourly rate.
• How to calculate ACSNOW | 2021
  • This is true in general for WRF. ACSNOW is always initialized to be zero at the beginning of the model run, It is then calculated in various surface schemes . You can find related codes in phys/module_sf_clm.F, phys/module_sf_noahdrv.F. etc.
  • 
• question about snow in WRF | 2018
  • You can probably use the variable ACSNOW, but you will need to turn it on, as the registry file does not have it automatically output. There are 2 ways to turn it on