Saturday, 5 March 2016

Record Total Precipitable Water through January and February

Below is a graph of Total Precipitable Water (TPW) through January and February from NCEP/NCAR.

Now for some very quick and rough back of envelope calculations. This is quick and off the top of my head as I have to get out and will be busy for the rest of this weekend. I never got round to including snow depth into my simple model calculations of winter sea ice growth.

This is a measure of the amount of precipitation at the surface....

EDIT - posting early in the morning yesterday and in a rush I think this is probably wrong. As Ronan points out in the comments TPW should be a measure of the total column water vapour content and the following working out would be wrong. But yesterday when plotting the above graph it only let me do it for the surface, so I think I jumped to the conclusion that it was actually meaning surface precipitation.

So I am crossing out the following and replacing it with the following two implications as I see it.

1) Snow cover may have substantially reduced thickening in Jan and Feb. Snow insulates the ice from the atmosphere and reduces thickening. (PIOMAS models this effect and that data will be out soon).

2) The high TPW supports the idea that the warmth of the last two months is due to influx of warmer air from lower latitudes.

I felt rushed to get implication 1 posted as it could counter my previous post in which I calculated the impact of the Jan/Feb warmth on sea ice and found it relatively small. The effect of temperature and snow might not be small.


In the comments Kevin makes a very good point, that large open water areas (in the Atlantic) might be causing the high precipitable water figure. So I change the region to north of 70degN and only between 90degE and 210degE, which excludes the Atlantic.

The plots look different, so perhaps greater open water in the Atlantic sector or warmer water is causing an increase. But even so in the series since 1948, 2016 is a record high.

Probably meaningless back of envelope comments follow - see comments.

The units on this graph are not stated, but elsewhere in documentation the amount kg/m^2 is stated. When this is at the surface does this mean it's the total in the atmospheric column or is it what has fallen? If we take water density as about 1000kg/m^3, and the 1970 to 1995 period as about 2.6kg/m^2, then this is 2.6cm of water.

If we take a snow density of about 300kg/m^2 then this is about 8cm of snow.

Jan/Feb 2016 has been 30% above the 1970 to 1995 period, if my calculations above are anywhere near correct (and on that I'm not sure), that's a snow depth of 11cm. However Warren et al 1999 find May snow depths of 30cm to 40cm during around that period on multi-year ice. Maybe this 11cm is reasonable when added to snowfall earlier in the winter?

Snow tends to insulate sea ice from the atmosphere, so should reduce the amount of thickening of sea ice. TPW is the highest since 1948, much higher than before 2000, about 20% (1/5) higher than in recent years. This will have the effect of tending to reduce sea ice thickness, PIOMAS and Cryosat will indicate to what degree this has happened (see the end of my last post for comments on Cryosat volume/thickness data). But that is a 20% on top of what is usual for October to December (and last year for that period was pretty usual).

However, the above does support the interpretation that this year's weird Arctic weather in January and February was driven by influx of warmer more moist air from lower latitudes.


Ronan said...

It was my understanding that total precipitable water was simply a measurement of how much water was present in a column of the atmosphere, not how much water has fallen or is likely to fall as snow/rain/what have you. It's correlated with total snowfall, of course (the snow's got to come from somewhere, after all), but it's no more than a correlation; actual precipitation can either be lower than the value implied by the TPW (not all the water falls) or higher (water is drawn in from surrounding regions of the atmosphere, replenishing the water content of a particular column of the atmosphere). Or am I misunderstanding what you're saying, here?

Chris Reynolds said...

Thanks Ronan,

That was my understanding, but when I tried to plot the variable for anything but the surface the webpage spat out an error. I know that either this variable or a similar one is used for snow cover in PIOMAS.

One issue is that if the variable is for the entire column, and if the units are kg/m^2 then there doesn't seem to be enough water to give the snow cover reported in Warren for the past.

This was a very rushed post because in the previous post I calculated the effect of the recent warmth on Jan Feb sea ice thickening and found it was not significant. However this data suggests that snowfall might have been significant and could reduce winter thickening of the ice. So as data contrary to my previous point I felt I had to get it out there as soon as I could.

I think there are two points here:

1) Snow cover may have substantially reduced thickening in Jan and Feb.

2) The high TPW supports the idea that the warmth of the last two months is due to influx of warmer air from lower latitudes.

Kevin O'Neill said...


The graph for TPW includes a lot of open water in the Atlantic sector that was usually ice-covered in previous years. I would assume that the open water also leads to more TPW. I don't know if it requires much in the way of inflow into the Atlantic sector.

Chris Reynolds said...

Thanks Kevin,

Good point. I have regraphed and added a new plot to the main post. The plot changes but Jan/Feb 2016 is only 0.2 below that calculated and including the Atlantic ocean.