One picture of the whole atmosphere above a single point, from the ground to about 40,000 ft. It looks impenetrable for roughly ten minutes, and then it does not.
This site is not an approved source of aeronautical or meteorological information. The soundings here are model forecasts, not observations, and nothing on this page substitutes for an official FAA preflight briefing. The limits section is specific about what this version can and cannot see, and it is worth reading before you trust a picture.
A weather map shows one slice through the atmosphere, laid out flat: the surface, or 18,000 ft, or wherever the chart is cut. A Skew-T does the opposite. It fixes the place and shows you the entire vertical column above it, all at once.
Every question a pilot asks about air rather than about ground is answered somewhere in that column. Where is the cloud, and how thick is it. Where does it stop being rain and start being ice. Is the air going to stay put or go up. What is the wind doing at the level you plan to fly. Those are all vertical questions, and a plan view cannot answer any of them.
The name is a description of the two axes. Log-P: pressure on the vertical axis, plotted on a logarithmic scale. Skew-T: temperature on the horizontal axis, tilted over to the right instead of running straight up the page. Both choices are deliberate and both are explained below, because a chart whose axes you do not understand is a chart you will misread with confidence.
About three and a half minutes, recorded on this site: a sounding at a point, how to read it, and then a route cross-section from Falcon Field (KFFC) down to Winter Haven (KGIF). The presenter is a synthetic avatar. The site is not.
Balloons and weather models both work in pressure, not feet, so pressure is the atmosphere's own vertical coordinate. It falls roughly exponentially with height, which means plotting the logarithm of it makes the axis very nearly linear in altitude: equal distances up the chart are equal distances up in the air. Useful anchors: 1000 hPa is close to sea level, 850 hPa is around 5,000 ft, 700 hPa around 10,000 ft, 500 hPa around 18,000 ft with half the mass of the atmosphere below it, and 200 hPa is up near 39,000 ft.
Lines of equal temperature, the isotherms, do not run vertically. They lean up and to the right at roughly 45 degrees. That tilt is the whole trick of the chart, and it is also the one thing that catches everyone the first time: a vertical line on a Skew-T is not a constant temperature. To read a temperature off the chart you follow the slanted isotherm down to the scale along the bottom. You never drop straight down.
Air cools with height at something like 2 °C per 1,000 ft. Over the depth of a chart that runs to 40,000 ft, that is a swing of 70 or 80 degrees. On a plain temperature-against-height plot, a real sounding therefore leans hard to the left and lies almost flat across the paper, like a slack rope. Both traces end up crushed into one narrow diagonal band, most of the chart is empty, and the detail you actually want is compressed into the part where the two lines overlap.
Skewing the isotherms to the right cancels most of that lean. As you climb the chart, every temperature is shifted a little further right, which offsets the cooling. The trace stands up. Instead of lying flat it runs roughly vertically up the middle of the chart and uses the full width, which is what makes the picture readable at a glance.
Standing the trace up buys two specific things:
A note on the 45 degrees, since this site would rather be precise than tidy: the skew is a display constant, and in this build it is 0.85, which puts the isotherms about 40 degrees off vertical rather than exactly 45. The convention is the tilt, not the number.
Almost everything worth knowing comes from reading the two lines against each other rather than either one on its own.
The distance between them, measured along an isotherm, is the dew point depression, or "the spread". Zero spread means the air is saturated: it cannot hold the water it is carrying, so the water is cloud. A spread of a couple of degrees means it is close. Ten or twenty degrees means dry air.
Read as a shape up the column, the spread tells a story:
Normally the temperature trace leans left as it climbs, because air cools with height. An inversion is where it does the opposite and bends back to the right: warmer air sitting on top of colder air. Inversions matter out of proportion to their size. They put a lid on convection, they trap moisture and haze underneath (which is why the visibility is often poor below one and excellent above), and the wind frequently changes both speed and direction sharply at the top of one, which you feel as a bump on the way through.
The freezing level is where the temperature trace crosses 0 °C. On a Skew-T that is a crossing point, not a horizontal line you can read off directly, which is exactly why the 0 °C isotherm is drawn brighter than its neighbours.
What matters is not the freezing level on its own but the freezing level combined with where the cloud is. Sub-zero temperature in clear, dry air is a comfort. Sub-zero temperature in cloud is where airframe ice comes from, and the band between 0 °C and about minus 20 °C is where supercooled droplets are most likely to still be liquid and waiting to freeze onto something. That is the band this site hatches.
Three cases the chart shows plainly once you know to look:
Clear air can ice too, and a model that fails to see a cloud layer fails to see the ice in it. A recent pilot report from an aircraft that has actually been through the layer beats any of this.
The column on the right carries a barb per level where the levels are far enough apart to draw one. Near the surface they crowd on a log-pressure axis, so some are dropped rather than smeared on top of each other: ten barbs here for thirteen levels. Count altitude off the pressure axis on the left, never off the barbs. Barbs use the same grammar everywhere in aviation and meteorology, so learning them here is not a local skill.
Two things to read in the column as a whole. First, the wind at the level you intend to fly, which decides your groundspeed and your fuel. Second, and less obvious, how much the wind changes between adjacent levels. A large change over a small vertical distance is shear, and shear is turbulence. A layer where the barbs swing round and grow several ticks longer within a couple of thousand feet is a layer to expect a rough ride in, and it very often sits at the top of an inversion.
The direction the wind turns as you climb is a forecaster's tell: veering, meaning turning clockwise with height, is the classic signature of warm air moving in; backing, turning anticlockwise, suggests colder air on the way.
Work from the ground up, and ask five questions in this order:
Then take the answers to a real briefing and see whether the observations agree with the model. When a PIREP and a sounding disagree, the aircraft that flew through the layer is right.
The Skew-T is on this site because of Dave Venable.
The site's author did not come across the chart in a textbook, or decide to build a weather product and go looking for something to draw. He was introduced to the Skew-T by Dave, a Southeast-based pilot, who took the time to explain what the diagram is for and how to read it properly. That is a different thing from having seen one before. Plenty of pilots have seen a Skew-T. Rather fewer have had somebody sit down and show them why the axes are shaped the way they are, what the gap between the two lines is telling them, and which single feature to look for first.
Everything on this page is downstream of that conversation: the sounding tool, the route cross-section, the 3-D column, and the fact that any of it is on a hobby aviation map at all. The explanation above is an attempt to pass on what Dave passed on, which seems like the right way to say thank you for it.
It is a model forecast on a 13-level grid. It is not a balloon sounding. Nobody released a radiosonde where you clicked. A real sounding is an instrument ascending through the air, sampling continuously, seeing every thin layer and every sharp inversion as it passes through them. This is a numerical weather model, sampled at thirteen pressure levels from 1000 hPa up to 200 hPa, for a time that in most cases has not happened yet. Between those levels the chart draws a straight line through air nothing ever measured.
None of that makes the chart useless. It makes it a good tool for understanding the shape of the day and a poor tool for a go or no-go decision on icing. Use it to know what to ask a proper briefing about.
The fastest way to learn this chart is to take soundings at places you already know and see whether the picture matches the weather you remember. Open the map, press Weather in the dock, and either right-click any point and choose Sounding here, or use Pick a point on the map in the panel. Drag the time slider to watch the column change through the forecast, and tilt the map to stand the same sounding up as a 3-D column over the spot it came from.