# How does MSLP change with elevation? trying to interpolate MSLP

My understanding of MSLP is that it is a normalized value for pressure at some location as if that location altitude was zero.

"Thus, MSLP is not a function of elevation"

However, after inspecting few cases, I am beginning to question the previous statement.

Below is a screenshot from windy.tv for MSLP over the Himalayas from ECMWF with a high resolution of 9km:

http://imgur.com/a/IWXxz

Two main characteristics are obvious:

1- MSLP changes steeply between two adjacent points just due to the difference in elevation.

2- MSLP over mountains seems to be a strong function of temperature (since it varies a lot over night and morning)

1) Are these two previous statement correct?

2) Can I normalize MSLP in terms of elevation?

To get more insight at what I am trying to figure out: I am trying to interpolate observational MSLP data; however, since MSLP is very high on mountains I am getting wrong results around those points. In other words my algorithm would assume that a large area around the station that is on top of a mountain has an MSLP of 1020 while in fact, just as you move away few kilometers MSLP changes very quickly to 1010 because of the elevation drop, so how can I fix that?

Important note:

Theses "anomalies" in MSLP can only be seen with high resolution models for example they can be seen in ECMWF and not in GFS

http://imgur.com/a/vDkNh

• – Fred
May 22 '17 at 15:14
• BarocliniCplusplus's answer is quite solid. You can see it in actual data (as opposed to model forecasts) as well. Imagine attempting to find low/high pressure centers when they are working through the mountains! It's a real challenge. But then, almost all things about mountain meteorology are complicated! There's no consistency in mountainous terrain... not MSLP, not station pressure, not temp, nothing. In the end, it's not MSLP that matters as much as trends and centers. I'm not sure you'll have much luck getting anything too useful interpolating :-/ May 23 '17 at 6:34
• A related useful question might be: what is your goal/need from such interpolation? May 23 '17 at 6:34