Doppler Radar

Doppler Radar
From Josh Wurman
NCAR S-POL DOPPLER RADAR
Doppler Shift: A frequency shift that occurs in electromagnetic

waves due to the motion of scatterers toward or away from the
observer.
Analogy: The Doppler shift for sound waves is the frequency shift that occurs as race
cars approach and then recede from a stationary observer
Doppler radar: A radar that can determine the frequency

shift through measurement of the phase change that occurs
in electromagnetic waves during a series of pulses.
The electric field of a transmitted wave
Et t E0 cos 2f t t 0
The returned electric field at some later

time back at the radar
Et t E1 cos 2f t t t 1
The time it took to travel
Substituting:
2r
t
c
2r
Et t E1 cos 2f t t 1
c
The received frequency can be determined by taking the time derivative if the
quantity in parentheses and dividing by 2
fr
1 d
2r
2 f dr
2f v
2f t t 1 f t t
ft t r ft f d
2 dt
c
c dt
c
Sign conventions
The Doppler frequency is negative (lower frequency,
red shift) for objects receding from the radar
The Doppler frequency is positive (higher frequency,
blue shift) for objects approaching the radar
These color shift conventions are typically also used on radar
displays of Doppler velocity
Red: Receding from radar

Blue: Toward radar
Note that Doppler radars are only sensitive to the radial motion of objects
Air motion is a three dimensional vector: A Doppler radar can only measure one
of these three components the motion along the beam toward or away from the radar
Magnitude of the Doppler Shift

Transmitted Frequency
Radial velocity
X band
C band
S band
9.37 GHz
5.62 GHz
3.0 GHz
1 m/s
62.5 Hz
37.5 Hz
20.0 Hz
10 m/s
625 Hz
375 Hz
200 Hz
50 m/s
3125 Hz
1876 Hz
1000 Hz
These frequency shifts are very small: for this reason, Doppler
radars must employ very stable transmitters and receivers
RECALL THE BLOCK DIAGRAM OF A DOPPLER RADAR AND THE

PHASE DETECTOR
A0 A1
cos(d t )
2
A0 A1
sin(d t )
2
Amplitude determination:
Phase determination:
A0 A1
I 2 Q2
2
Q
d tan 1
Why is emphasis placed on phase determination instead of

determination of the Doppler frequency?
Typical period of Doppler frequency
Typical pulse duration
fd
= 0.3 to 50 milliseconds
= 1 microsecond
Only a very small fraction of a complete Doppler frequency cycle is contained within a pulse
Alternate approach: one samples the Doppler-shifted echo

with a train of pulses and tries to reconstruct, or estimate,
the Doppler frequency from the phase change that occurs
between pulses.
We can understand how the phase shift can be related to the radial
velocity by considering a single target moving radially.
Distance target moves radially in one pulse period Tr

The corresponding phase shift of a wave between two
Consecutive pulses (twice (out and back) the fraction
of a wavelength traversed between two consecutive
pulses)
Solving for the radial velocity
d Tr vr
2 1 2Tr vr
2 1
vr
2Tr 2
(1)
In practice, the pulse volume contains billions of targets moving at

different radial speeds and an average phase shift must be
determined from a train of pulses
Illustration of the reconstruction of the Doppler

frequency from sampled phase values
Dots correspond to the measured samples of phase
PROBLEM
More than one Doppler frequency (radial velocity) will always

exist that can fit a finite sample of phase values.
The radial velocity determined from the sampled
phase values is not unique
What is the maximum radial velocity possible before

ambiguity in the measurement of velocity occurs?
We need at least two measurements per wavelength to determine a frequency
The phase change between pulses must therefore be less than half a wavelength
vr
2Tr
From (1)
4vrTr

F
vr
vmax
2Tr
4
vmax is called the Nyquist velocity and represents the maximum (or
minimum) radial velocity a Doppler radar can measure unambiguously true
velocities larger or smaller than this value will be folded back into the
unambiguous range
EXAMPLE VALUES OF THE MAXIMUM

UNAMBIGUOUS DOPPLER VELOCITY
Wavelength
Radar PRF (s-1)
cm
200
500
1000
2000
1.5
3.75
7.5
15
2.5
6.25
12.5
25
10
5.0
12.5
25.0
50
Table shows that Doppler radars capable of measuring a

large range of velocities unambiguously have long
wavelength and operate at high PRF
Folded velocities
Can you find the folded velocities in this image?
Folded velocities in an RHI
Velocities after unfolding
http://apollo.lsc.vsc.edu/classes/remote/graphics/airborne_radar_images/newcastle_folded.gif
But recall that for a large unambiguous RANGE

Doppler radars must operate at a low PRF
rmax
2F
vmax
THE DOPPLER DILEMA: A GOOD CHOICE OF PRF TO ACHIEVE A

LARGE UNAMBIGUOUS RANGE WILL BE A POOR CHOICE TO
ACHIEVE A LARGE UNAMBIGUOUS VELOCITY
rmax vmax
The Doppler Dilema
Ways to circumvent the ambiguity dilema

1. Bursts of pulses at alternating low and high pulse repetition frequencies
Measure reflectivity
Measure velocity
Low PRF used to measure to long range, high PRF to measure velocity
2. Use slightly different PRFs in alternating sequence

For 1st PRF
f d
Vr
2nvmax
2
For 2 PRF
Vr
nd
f d
2nvmax
2
Solve simultaneously
f d f d
4
nvmax
nvmax
f d f d nF nF
Example: = 5.33 cm, F = 900 s , F = 1200 s
-1
-1
vmax
F
12 ms 1
4

vmax
MEASURE fd = -150 hz, fd = 450 hz
300 1200n 900n

n n 1
Data is folded once
F
16 ms 1
4
Real characteristics of a returned signal from a distributed target

Velocity of individual targets in contributing volume vary due to:
1) Wind shear (particularly in the vertical)
2) Turbulence
3) Differential fall velocity (particularly at high elevation angles)
4) Antenna rotation
5) Variation in refraction of microwave wavefronts
NET RESULT: A series of pulses will measure a

spectrum of velocities (Doppler frequencies)
Power per unit velocity interval (db)
The moments, or integral properties, of the Doppler Spectrum
S f df
Pr
Average returned power
Mean radial velocity
vr
S v dv
vmax
vmax
vmax
S v dv
vmax
vmax
vS v dv
vmax
vS v dv
r
vmax
Pr
vmax
vmax
Spectral width
v v S v dv
2

2
v
vmax
vmax
vmax
S v dv
r
vmax
v v S v dv
2
vmax
Pr
Example of Doppler spectra

As a function of altitude
measured in a winter
snowband. These spectra
were measured with a
vertically pointing Doppler
profiler with a rather wide (9
degree) beamwidth
Melting level
Note ground clutter
The Doppler spectrum represents the echo from a single contributing region
Mean Doppler frequency (or velocity)
Spectral width
Related to the reflectivity weighted mean

radial motion of the particles
Related to the relative particle motions
RECALL: Fluctuations in mean power from pulse to pulse occur due

to interference effects as the returned EM waves superimpose upon
one another.
Fluctuations are due to the relative motion of the particles

between pulses and therefore to the spectral width
Effects of relative particle motion:

Consider two particles in a pulse volume
Return from 1:
Return from 2:
E1 t E1 cos D1 t 1
E2 t E2 cos D 2 t 2
Where:
2f t
D1, 2
4vr1, 2
With a bit of trigonometry.
E1 t E2 t E1 cos cos D1t sin sin D1t
E2 cos cos D1t sin sin D1t
Where: t 1
t 2
Total Echo power proportional to sum of two fields squared

2
E1
E2
Pr
E1 E2 cos D1 D 2
2
2
Constant term
Term which depends on particles relative velocities and wavelength
For a large ensemble of particles

2
Ei
Pr
2
i
i
E E cos
1
Di
Dj
To determine the echo power, one must average over a

large enough independent samples that the second term
averages to zero
HOWEVER!!
To determine the Doppler frequency (and velocity) from

consecutive measurements of echo phase, the samples must
be DEPENDENT (more frequent) than those required to
obtain the desired resolution in reflectivity
Determining the Doppler Spectrum

1. Doppler spectrum is measured at a particular range gate (e.g. at r
ct
)
2
2. Must process a time series of discrete samples of echo Er(t) at intervals

of the pulse period Tr
3. Analyze the sampled signal using (fast) Fourier Transform methods:
1
E ( mTr )
M
M 1
F kf 0 cos 2kf 0 mTr
m 0
M = # of samples
f0 = frequency resolution
M 1
F (kf 0 ) Er mTr cos 2kf 0 mTr

m 0
4. Frequency components (radial velocities) occur at discrete intervals, with M

intervals separated by intervals of 1/MTr = fD
Discrete Doppler spectra

computed for a point
target, with M = 8. Dots
represent the discrete
frequency components of
the spectra.
Point target, M = 8
fD = 2 f 0
If Doppler frequency is not an

integral multiple of the
frequency resolution (normally
not the case), the discrete
Fourier transform will smear
power into all of the frequencies
across the spectrum.
Point target, M = 8
fD = 2.5 f0
Signal appear in all M lines
of the spectrum
With a distributed target, which has a spectrum of Doppler frequencies, the

discrete Fourier transform will always produce power in all frequencies.
The power will be relatively uniform at frequencies not associated with the true
Doppler frequencies, and peak across the range of true Doppler frequencies.
Noise
Signal
Noise
In most applications (such as the operational NEXRADs), the Doppler

spectra are not needed.
Recording the entire Doppler spectra at each range gate takes an
enormous amount of data storage capability, quickly exceeding the
capacity of current electronic storage devices.
What are needed are the moments of the spectra the average
returned power, the mean Doppler velocity, and the spectral width
How can the moments be obtained from the series of discrete samples?
1. Record time series at each range gate and Fourier analyze Doppler
Spectra. Calculate the moments. Discard Spectral data.
(Computationally inefficient, given that these calculations must be
done for every range gate on every beam!
or
2. Calculate moments as the time series is recorded using the
Autocorrelation function (see below), and discard data continuously
following the calculation (little data storage required and
computationally efficient)
Problems complicating process:

1. Noise
Tends to bias Vr to 0
and spectral width to vmax/3
2. Folding
3. Clutter
RECALL THE PHASE

DETECTOR IN A
DOPPLER RADAR
SYSTEM
A0 A1
cos(d t )
2
A0 A1
sin(d t )
2
Amplitude determination:
Phase determination:
A0 A1
I 2 Q2
2
Q
d tan 1
Sample of I/Q channel voltage at time 1:
A0 A1
R1
exp i D t 1
2
Sample of I/Q channel voltage at time 2:
A0 A2
R2
exp i D t 2
2
Autocorrelation function:
A02 A1 A2
R1 R
exp i 2 1
4
*
2
1
C
M
R R
n 1
*
n 1
Representation of
I/Q signal on a phase
Diagram in complex
space
d
Amplitude
A0 A1
2
A0 A1
2
Graphical depiction of how average amplitude (returned power)

And phase (radial velocity) are recovered from autocorrelation function
A02 A4 A5
4
A02 A3 A4
4
A02 A2 A3
4
2
0
A A1 A2
4
The spectral width can also be recovered from autocorrelation function

Doppler Radar

Uploaded by

Copyright:

Available Formats

Doppler Radar

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Doppler Radar

Uploaded by

Copyright:

Available Formats

Doppler Radar

From Josh Wurman

NCAR S-POL DOPPLER RADAR

Doppler Shift: A frequency shift that occurs in electromagnetic

Doppler radar: A radar that can determine the frequency

The electric field of a transmitted wave

The returned electric field at some later

The time it took to travel

Red: Receding from radar

Magnitude of the Doppler Shift

RECALL THE BLOCK DIAGRAM OF A DOPPLER RADAR AND THE

Why is emphasis placed on phase determination instead of

Alternate approach: one samples the Doppler-shifted echo

Distance target moves radially in one pulse period Tr

Solving for the radial velocity

In practice, the pulse volume contains billions of targets moving at

Illustration of the reconstruction of the Doppler

Dots correspond to the measured samples of phase

More than one Doppler frequency (radial velocity) will always

What is the maximum radial velocity possible before

EXAMPLE VALUES OF THE MAXIMUM

Radar PRF (s-1)

Table shows that Doppler radars capable of measuring a

Can you find the folded velocities in this image?

Folded velocities in an RHI

Velocities after unfolding

But recall that for a large unambiguous RANGE

THE DOPPLER DILEMA: A GOOD CHOICE OF PRF TO ACHIEVE A

The Doppler Dilema

Ways to circumvent the ambiguity dilema

2. Use slightly different PRFs in alternating sequence

MEASURE fd = -150 hz, fd = 450 hz

300 1200n 900n

Data is folded once

Real characteristics of a returned signal from a distributed target

3) Differential fall velocity (particularly at high elevation angles)

NET RESULT: A series of pulses will measure a

The moments, or integral properties, of the Doppler Spectrum

Average returned power

Mean radial velocity

Example of Doppler spectra

Note ground clutter

Related to the reflectivity weighted mean

Related to the relative particle motions

RECALL: Fluctuations in mean power from pulse to pulse occur due

Fluctuations are due to the relative motion of the particles

Effects of relative particle motion:

With a bit of trigonometry.

E1 t E2 t E1 cos cos D1t sin sin D1t

E2 cos cos D1t sin sin D1t

Total Echo power proportional to sum of two fields squared

Term which depends on particles relative velocities and wavelength

For a large ensemble of particles

To determine the echo power, one must average over a

To determine the Doppler frequency (and velocity) from

Determining the Doppler Spectrum

2. Must process a time series of discrete samples of echo Er(t) at intervals

F kf 0 cos 2kf 0 mTr

F (kf 0 ) Er mTr cos 2kf 0 mTr