Fundamentals of Observing Columns Using Remote Sensing

Tim Wagner

University of Wisconsin – Madison

ARM Summer School 2025

1

May 8, 2025

Primary Blue

#0070C1

ARM Blue

#194173

ARM Green

#00BD70

Gray

#828282

Black

#000000

White

#FFFFFF

Light Blue

#80B8E0

Introduction: the PBL

The Planetary Boundary Layer (PBL) is the layer of the atmosphere closest to the surface

• It’s where sensible and latent heat fluxes are exchanged between the atmosphere and ground
• It’s where mechanical and convective turbulence influences atmospheric flow
• It’s where we live!

​

Thickness varies with the diurnal cycle:

• A few meters at night, 1-2 km during the day

​

Surface observations alone are not enough to inform about all these processes

ARM Summer School 2025

2

May 8, 2025

Stull 1988

Primary Blue

#0070C1

ARM Blue

#194173

ARM Green

#00BD70

Gray

#828282

Black

#000000

White

#FFFFFF

Light Blue

#80B8E0

Advantages of Profilers

Why not just use radiosondes?

• Recurring costs (~$300 per observation)
• Significant manual labor
• Takes several minutes just to penetrate the PBL
• Observations of non-thermodynamic variables (e.g. aerosols) are limited.

​

Profilers enable high temporal resolution observations of the PBL

• A few seconds to a few minutes apart
• Enables monitoring the evolution of atmospheric characteristics

ARM Summer School 2025

3

May 8, 2025

Primary Blue

#0070C1

ARM Blue

#194173

ARM Green

#00BD70

Gray

#828282

Black

#000000

White

#FFFFFF

Light Blue

#80B8E0

Some Key Instruments

ARM Summer School 2025

4

May 8, 2025

Univ. of Köln

ARM

Primary Blue

#0070C1

ARM Blue

#194173

ARM Green

#00BD70

Gray

#828282

Black

#000000

White

#FFFFFF

Light Blue

#80B8E0

Passive Remote Sensing Basics

Different gases in the atmosphere cause varying amounts of absorption and emission at different wavelengths?

​

Why? Different molecules have different shapes, and it takes a very specific amount of energy to excite a particular molecular shape (to stretch, bend, or rotate it).

​

Energy is quantized, and energy is directly related to wavelength: λ = c/ν

​

Only specific wavelengths will get absorbed by a particular gas, while a different gas absorbs a different set of wavelengths

ARM Summer School 2025

5

May 8, 2025

Ackerman and Knox 2008

Primary Blue

#0070C1

ARM Blue

#194173

ARM Green

#00BD70

Gray

#828282

Black

#000000

White

#FFFFFF

Light Blue

#80B8E0

Passive Remote Sensing Basics

ARM Summer School 2025

6

May 8, 2025

Primary Blue

#0070C1

ARM Blue

#194173

ARM Green

#00BD70

Gray

#828282

Black

#000000

White

#FFFFFF

Light Blue

#80B8E0

Passive Remote Sensing Basics

When on the ground looking up, opaque regions are warm, because all of the radiation is being absorbed and emitted near the surface

​

As you move away from these absorption bands, you get more information from higher up in the atmosphere.

​

The amount of radiation you view at a surface is a function of the vertical distribution of temperature and water vapor.

​

Given a spectral observation, can you get the structure of the atmosphere that produced it? Yes, but it ain’t easy…

ARM Summer School 2025

7

May 8, 2025

CO₂

H₂O

O₃

Primary Blue

#0070C1

ARM Blue

#194173

ARM Green

#00BD70

Gray

#828282

Black

#000000

White

#FFFFFF

Light Blue

#80B8E0

Thermodynamic Retrievals and the TROPoe Algorithm

A forward model computes what the AERI-observed spectrum should look like for a given atmospheric profile.

​

Compare a modeled spectrum with the observations, adjust your guess of the atmospheric profile with a big nasty equation:

​

​

Iterate until your modeled spectrum matches your observation: the profiles you used to make the model represent your observation of the atmosphere.

​

​

​

​

ARM Summer School 2025

8

May 8, 2025

Primary Blue

#0070C1

ARM Blue

#194173

ARM Green

#00BD70

Gray

#828282

Black

#000000

White

#FFFFFF

Light Blue

#80B8E0

TROPoe Algorithm

Problem is, there’s not a 1:1 correspondence between spectra and atmospheric states:

• Smoothly varying function but a finite number of observations means more unknowns than knowns.
• Climatological values are used to constrain the algorithm to produce the statistically most likely state.
• While there’s 1000+ channels, many of them are correlated with each other, so there are many fewer pieces of information than you might think.

End result: smoothed curves where small scale structures are not resolved (better than satellite, though!)

​

Still, comparing to collocated radiosondes shows very strong agreement on overall atmospheric structure, especially in the PBL.

​

​

​

​

ARM Summer School 2025

9

May 8, 2025

Primary Blue

#0070C1

ARM Blue

#194173

ARM Green

#00BD70

Gray

#828282

Black

#000000

White

#FFFFFF

Light Blue

#80B8E0

Doppler Lidar (DLID) Basics

Doppler lidars emit a beam of near-infrared radiation (1.5 μm), measure the frequency and intensity of backscatter.

• Sensitive to dust, insects, other aerosols
• Relatively insensitive to air molecules
• Very little solar scattering to worry about
• There must be something in the air for the initial beam to bounce off of.

​

By measuring the frequency shift of the returned radiation, can determine the 1-D velocity along the beam via the Doppler effect.

​

​

​

​

​

ARM Summer School 2025

10

May 8, 2025

Primary Blue

#0070C1

ARM Blue

#194173

ARM Green

#00BD70

Gray

#828282

Black

#000000

White

#FFFFFF

Light Blue

#80B8E0

Velocity Azimuth Display (VAD) Profile Retrievals

If you take an observation at a non-zero zenith angle, the radial velocity measured by the DLID gives you height-resolved radial velocity measurements.

​

Take observations at multiple azimuths, use relatively simple trigonometry to convert to wind vectors.

​

Key assumption: the wind field at a given height is constant within the circle traced out by the azimuthal scan.

​

Lidars are active instruments, so dead band adjacent to the surface.

​

A shallower elevation angle means more observations closer to the surface, but lower max height + more wind heterogeneity.

​

​

​

​

​

ARM Summer School 2025

11

May 8, 2025

Primary Blue

#0070C1

ARM Blue

#194173

ARM Green

#00BD70

Gray

#828282

Black

#000000

White

#FFFFFF

Light Blue

#80B8E0

Micropulse Lidar (MPL)

MPL is very straightforward:

• Same principles as a radar
• Emit a pulse of radiation
• Wait for it to be scattered back to the instrument
• Elapsed time for the return tells you how far away the target is
• Intensity of the return tells you how much of the target there is
• Biggest difference from radar: wavelength of the pulse!
• S band ~ 10 cm
• C band ~ 5 cm
• X band ~ 3 cm
• K band ~ 1 cm
• MPL: 532 nm = 0.0000532 cm = green light
• MPL is much more sensitive to aerosol-sized particles than radars are
• But that means you can’t penetrate through clouds well.

​

​

​

​

​

ARM Summer School 2025

12

May 8, 2025

UCAR

Primary Blue

#0070C1

ARM Blue

#194173

ARM Green

#00BD70

Gray

#828282

Black

#000000

White

#FFFFFF

Light Blue

#80B8E0

MPL Polarization and Cloud Phase

​

​

​

​

​

ARM Summer School 2025

13

May 8, 2025

ARM

MPL monitors backscattered radiation for both co-polarization and cross polarization.

​

The light is polarized when it is emitted (the planes of the emitted radiation are geometrically oriented).

​

When light is scattered by spherical objects (e.g. cloud droplets) the returned radiation generally has the same polarization.

​

When light is scattered by irregular objects (e.g. ice crystals, soot) much of the returned radiation is of a different polarization.

​

The greater the ratio of cross polarized to co-polarized, the more likely the observed scatterer isn’t a water droplet.

​

​

​

​

​

​

Primary Blue

#0070C1

ARM Blue

#194173

ARM Green

#00BD70

Gray

#828282

Black

#000000

White

#FFFFFF

Light Blue

#80B8E0

Applications: Solar Eclipse

​

​

​

​

​

ARM Summer School 2025

14

May 8, 2025

Solar eclipses are basically a speed run through the nocturnal cycle:

• Super fast sunset
• A “night” that is four minutes long at most
• Super fast sunrise

​

This has profound impacts on the PBL and its processes.

​

​

​

Wikipedia

Scott Collis

Primary Blue

#0070C1

ARM Blue

#194173

ARM Green

#00BD70

Gray

#828282

Black

#000000

White

#FFFFFF

Light Blue

#80B8E0

TROPoe Observations of 2024 Eclipse

​

​

​

​

​

ARM Summer School 2025

15

May 8, 2025

​

T – T_{max eclipse}

​

​

​

​

​

​

Weak “nocturnal” inversion develops with the eclipse.

​

​

​

Primary Blue

#0070C1

ARM Blue

#194173

ARM Green

#00BD70

Gray

#828282

Black

#000000

White

#FFFFFF

Light Blue

#80B8E0

DLID Observations of 2024 Eclipse

​

​

​

​

​

ARM Summer School 2025

16

May 8, 2025

​

Turbulence grows as expected during the pre-eclipse hours.

​

As sun disappears, turbulence vanishes!

​

After sun returns, penetrative updrafts return much faster than during a normal sunrise

​

PBL grows much faster after eclipse as well.

​

​

​

Primary Blue

#0070C1

ARM Blue

#194173

ARM Green

#00BD70

Gray

#828282

Black

#000000

White

#FFFFFF

Light Blue

#80B8E0

Applications: Lake Breeze

​

​

​

​

​

ARM Summer School 2025

17

May 8, 2025

Primary Blue

#0070C1

ARM Blue

#194173

ARM Green

#00BD70

Gray

#828282

Black

#000000

White

#FFFFFF

Light Blue

#80B8E0

Lake Breeze from High Temporal Resolution Profilers

​

​

​

​

​

ARM Summer School 2025

18

May 8, 2025

Potential temperature cross section (from AERI/TROPoe) shows pre-existing nocturnal inversion gets eroded by strong daytime mixing, but lake breeze induces a new low-level inversion.

​

Wind direction (from DLID) shift occurs at higher heights as time goes on.

​

Backscatter profiles (from HSRL, which is similar to MPL here at Bankhead), show initial growth in PBL with daytime heating, plus enhanced aerosol concentrations along the lake breeze front.

Primary Blue

#0070C1

ARM Blue

#194173

ARM Green

#00BD70

Gray

#828282

Black

#000000

White

#FFFFFF

Light Blue

#80B8E0