WEATHER RADAR TRAINING 5.

0 / ISTANBUL-2010 TURKISH STATE METEOROLOGICAL SERVICE (TSMS) WORLD METEOROLOGICAL ORGANIZATION (WMO) COMMISSION FOR INSTRUMENTS AND METHODS OF OBSERVATION (CIMO) OPAG ON CAPACITY BUILDING (OPAG-CB/C.2.) TRAINING ACTIVITIES AND TRAINING MATERIALS

TRAINING COURSE ON WEATHER RADAR SYSTEMS

MODULE C: PROCESSING BASICS IN DOPPLER WEATHER RADARS
OUZHAN REC -Electronics Engineer [email protected] ELECTRONIC OBSERVING SYSTEMS DIVISION TURKISH STATE METEOROLOGICAL SERVICE 10-14 MAY 2010 WMO RTC-TURKEY ISTANBUL FACILITIES, ISTANBUL-TURKEY

IMPORTANT NOTE!
The main sources of the information used in those training documents, which have been stated in the reference list, are the guides published by WMO, technical brochure and instruction manuals of the manufacturers, all related documents issued by research institutions, universities and manufacturers, and all the documents available for anyone in the internet. In addition, previous version of those documents have been reviewed and updated based on the comments of the participants of the previous trainings and view of the instrument experts of TSMS. It must be always kept in mind that those document have been prepared just for such training courses but not for any commercial purposes.

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

MODULE A: INTRODUCTION TO RADAR

MODULE B: RADAR HARDWARE

MODULE C: PROCESSING BASICS IN DOPPLER WEATHER RADARS

MODULE D: RADAR PRODUCTS AND OPERATIONAL APPLICATIONS

MODULE E: RADAR MAINTENANCE AND CALIBRATION TECHNIQUES

MODULE F: RADAR INFRASTRUCTURE

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

PROCESSING BASICS IN DOPPLER WEATHER RADARS CONTENTS CONTENTS FIGURE LIST ABBREVIATIONS 1. 2. 2.1. 2.2. 2.3. 2.4. 2.5. 2.6. 3. 4. 4.1. 4.2. 4.3. 4.4. 5. GENERAL OVERVIEW SIGNAL PROCESSING I/Q Demodulation Analogue Receiver Channels Doppler Signal Processing Techniques Pulse Pair Clutter Correction FFT Implementation MAXIMUM UNAMBIGUOUS RANGE VELOCITY DETERMINATION Doppler Shift Total Distance to Target in Radians Pulse Pair Method Maximum Unambiguous Velocity DOPPLER DILEMMA 3 4 5 7 8 11 11 11 12 14 17 20 23 25 29 31 33 37 41 43 45 46 48 51 52

6. RADAR RANGE FOLDING 6.1. Recognizing Range Aliased Echoes 6.2. Elimination of Second Trip Echoes 7. VELOCITY FOLDING 7.1. Staggered PRF for Velocity Unfolding 7.2. Recognizing Velocity Aliasing 8. REFERENCES

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

FIGURE LIST FIGURE 1: FIGURE 2: FIGURE 3: FIGURE 4: FIGURE 5: FIGURE 6: FIGURE 7: FIGURE 8: FIGURE 9: FIGURE 10: FIGURE 11: FIGURE 12: FIGURE 13: FIGURE 14: FIGURE 15: FIGURE 16: FIGURE 17: FIGURE 18: FIGURE 19: FIGURE 20: FIGURE 21: FIGURE 22: FIGURE 23: FIGURE 24: FIGURE 25: FIGURE 26: FIGURE 27: FIGURE 28: FIGURE 29: FIGURE 30: FIGURE 31: FIGURE 32: FIGURE 33: FIGURE 34: FIGURE 35: FIGURE 36: FIGURE 37: FIGURE 38: FIGURE 39: FIGURE 40: General Signal Flow Chart of Radar System I/Q Demodulation Frequency, Time and Amplitude Axis Rain over PE Spectra and IIR Clutter Notch (Idealised) Rain over PE Spectra and Interpolated Spectra Example of Frequency Domain Clutter Filter FFT of a Three-Frequency Signal with N = 16 FFT of a Three-Frequency Signal with N = 32 FFT of a Three-Frequency Signal with N = 64 FFT of a Three-Frequency Signal with N = 128 FFT of a Three-Frequency Signal with N = 256 Finding the rmax Velocity Example PPI Velocity Radial Velocity Frequency of Stationary and Moving Target The Effect of Moving Target on Frequency Wavelength and Amplitude of a Wave Phase of a Wave Sine Wave (Solid Curve) and a Second Signal 30 Out of Phase with the First Wave (Dashed Curve) Wavelengths and Phase Shifts. (A) T-l is Wave Reference Position. (B) T-2 Waves Position Has Changed 90 from Reference Position (T-l) Doppler Velocity Phases at Different Ranges Basic Block Diagram of Radar Analogue to Digital Converter (ADC) A Sine Wave Sampling at 1 Time per Cycle Sampling at 1.5 Times per Cycle Nyquist Frequency Velocity Interval versus Range Interval and PRF at Different Wavelength Range-Height Diagram Second Trip Echo Example Second Trip Echo Example Rmax and Second Trip Echo Relationship Second Trip Echo Real and Second Trip Echo Velocity Folding or Aliasing Folded Velocity Examples Unfolded Velocities for This Storm Dual PRF Technique 4 8 11 15 16 16 17 18 18 18 19 19 21 23 23 24 25 26 26 27 27 28 30 30 31 34 34 35 35 36 38 40 41 42 42 43 44 46 47 48 50

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

ABBREVIATIONS: PRF PRT IF RF I Q FT FFT DSP V Z W IFD RX TX W/G LNA D dB dBm dBZ V Hz KHz MHz GHz LOG LIN IIR PE CPU AC DC Rmax Vr STALO COHO : Pulse Repetition Frequency : Pulse Repetition Time : Intermediate Frequency : radio Frequency : In-phase : Quadrature : Fast Transform : Fast Fourier Transform : Digital Signal Processor : Velocity : Reflectivity : Spectral Width : Intermediate Frequency Digitizer : Receiver : Transmitter : Waveguide : Low Noise Amplifier : Diameter of Particle : Decibel : Decibel Milliwatt : Logarithmic Scale for Measuring Radar Reflectivity Factor : Antenna Speed : Hertz : Kilohertz : Megahertz : Gigahertz : Logarithmic : Linear : Infinite Impulse Response : Photo-Electron : Central Processing Unit : Alternating Current : Direct Current : Maximum Unambiguous Range : Radial Velocity : Stable Oscillator : Coherent Oscillator

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

1.

GENERAL OVERVIEW

Doppler Meteorological Radars are used to detect, to process, to distribute and display meteorological data in a large scale area. Doppler technology opened a new window in the field of the radar meteorology and increased the performance of weather radars significantly. Doppler weather radars are capable of acquiring particle velocity data in addition to range, direction, and reflectivity data. Software processing is used to control the radar operating characteristics to produce the optimum radar volume coverage patterns and to optimize the radar returns. The resulting base weather data is then processed through the application of meteorological algorithms to generate base and derived weather products. These products are further processed using graphics algorithms to produce readily interpretable weather data displays on colour monitors. After receipt of the reflected echo from target, the signal passes some processing stages for product generation. These stages include many complex process and algorithms. These processes and algorithms including basic products will be explained in this module.

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

2.

SIGNAL PROCESSING

Weather Radars employ high dynamic-range linear receiver and DSPs (digital signal processors) to extract information from the received echo power. Linear receiver output in intermediate frequency (IF) and analogue form is converted to digital form in the analogue-to-digital converter and fed to digital filters to split the power into in-phase (I) and quadrature (Q) components. DSPs process the raw I/Q data and perform phase and amplitude correction, clutter filtering, covariance computation and produce normalized results. These normalized results are tagged with angle information, headers and given out as a data set. Covariance computation is based on pulse pair processing. Intensity estimation consists simply of integrating the power in the linear channel (I2 + Q2) over range and azimuth. The resulting power estimate is corrected for system noise, atmospheric attenuation and transmitter power variations. The signal processing of the linear channel ends with the estimation of reflectivity, mean radial velocity and velocity spectrum width.

Figure 1: General Signal Flow Chart of Radar System. Doppler velocity (V): Doppler velocity is reflectivity-weighted average velocity of targets in the pulse volume and determined by phase measurements from a large number of successive pulses. This is also called radial velocity and gives only the radial component of the velocity vector. It is generally assumed that raindrops and other particles are advected with the wind and have no own motion except their falling velocity.

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

Reflectivity factor (Z): This is the integral over the backscatter cross-section of the particles in a pulse volume. For particles small compared to the wavelength the scatter cross-section is D6, where D is the diameter of the particle. Radars are calibrated in the way to give directly (assuming the dielectric constant of water) the reflectivity factor from the received backscattered energy. Units for the reflectivity factor are mm6 m-3 or the logarithmic value of this in dBZ. Spectral width (W): Spectral measure is a measure of the dispersion of velocities within the pulse volume and standard deviation of the velocity spectrum. Spectral width depends among others from the turbulence within the pulse volume. Sampling: Sampling Rate is defined by the velocity of scan, PRF (Pulse Repetition Frequency) and Resolution.

Resolution= 0.7 Min speed= 6 deg/sec PRF= 1500 Hz Sampling=

0.7 * 1500 = 175 6

Sample=?

Beam spacing (Resolution) =

Nsamp * Ant.speed PRF

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

Example:

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

10

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

2.1.

I/Q Demodulation

Splitter

Mixer

Low Pass Filter I

30 MHz Rx IF
0 deg Mixer Low Pass Filter Q 90 deg Phase Shifter

COHO (30 MHz)

Figure 2: I/Q Demodulation. 2.2. Analogue Receiver Channels

Reflectivity is calculated using the digitised LOG channel. Velocity is calculated using the linear channels I/Q signals. An estimate of a PEs power contribution can be calculated using the digitised linear channel (I/Q data) and then applied to correct the LOG channel power estimate.
2.3. Doppler Signal Processing Techniques

Pulse Pair Spectral Processing using Fourier Transform

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

11

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

2.4.

Pulse Pair

Target velocity can be estimated by use of the Pulse-Pair technique. Pulse-Pairing provides an estimate of a targets velocity by determining the average phase shift of a target that has occurred from PRF to PRF, averaged over several PRF. As many PRF as can be obtained between 1 deg azimuth angle Boundaries are used. More PRF equates to less noise, giving more accurate estimates. The Pulse-Pair algorithm was traditionally used as the CPU horsepower required is not excessive. I&Q data samples are considered to describe a complex vector that will rotate at a speed directly related to the targets velocity i.e. sample[i] = I + jQ. An Auto-Correlation algorithm is applied to each individual range bin, across several PRF, and is defined as below:
NumPRFsamples-1

Lag[n] =

S
i=0

conjugate (sample[i]).sample [i+n]

for n = 0 & 1

The conjugate of a complex value is simply flipping the sign of the imaginary component, in this case the Q value i.e. Conjugate (sample[i]) = I - jQ The Auto-Correlation algorithm conveniently provides two valuable answers:
The real component of lag [0] provides Intensity information. The phase of lag [1] provides Velocity information. Pulse Pair: Algorithm Lag Zero

Lag Zero = Signal Power Assume a Unit Vector - i.e. a vector length of 1 @ 45 deg, i.e. 0.707 + j0.707
Lag [0] = (I + jQ) x conj (I + jQ)

= (I + jQ) x (I jQ) = (0.707 + j0.707) x (0.707 - j0.707) = 0.707x0.707 - j(0.707x0.707) + j(0.707x0.707) - j2(0.707x0.707)
WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

12

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

= 0.5 j0.5 + j0.5 -j2(0.5) = 0.5 + j0 +0.5 = 1 + j0

The real component provides the magnitude; the phase will always be zero.

x
sample[1] conjugate(sample[1]

=
lag[0]

Pulse Pair: Algorithm First Lag = Velocity Stationary target:

I&Q samples for PRF2 will ideally have exactly the same phase and magnitude as I&Q samples taken at PRF1.eg. PRF1 = 0.707 +j0.707 & PRF2 = 0.707 + j0.707 Calculating the auto-correlation of the first lag will provide the phase difference between PRF1 and PRF2. The phase of the resultant vector provides the phase difference between PRF1 & PRF2. Lag [1]= (I2 + jQ2) x conj(I1 + jQ1) = (I2 + jQ2) x (I1 jQ1) = (0.707 + j0.707) x (0.707 - j0.707) = 0.707x0.707 - j(0.707x0.707) + j(0.707x0.707) - j2(0.707x0.707) = 0.5 + j0 -j2(0.5) = 0.5 + j0 +0.5 = 1 + j0 => Phase difference = 0 deg

x
sample[2] Conjugate(samp.[1])

=
Lag[1]Result

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

13

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

Pulse Pair: Algorithm First Lag = Velocity Moving target: PRF1 = 0.707 + j0.707

PRF2 = 1 + j0 Lag [1] = (I2 + jQ2) x conj (I1 + jQ1)

= (I2 + jQ2) x (I1 jQ1) = (1 + j0) x (0.707 - j0.707) = (1x0.707) - j(1x0.707) + j(0x0.707) - j2(0 x0.707) = 0.707 - j0.707 + j0 -j2(0) = 0.707 - j0.707 Phase = -45deg. or -/4

x
Sample [1] Conjugate (sample [1])

=
Lag [0]

X
Sample [2] Conjugate (sample [1])

=
Lag [1]

2.5.

Clutter Correction

The major difference between FFT and pulse pair processing is the way in which clutter filtering is performed. The pulse pair processing uses a time domain IIR filter while the FFT mode uses a frequency domain filter. Advantage of the FFT approach is that it is less destructive to overlapped weather than the IIR filter since the clutter filter algorithm attempts to interpolate over the weather. This results in more accurate estimates of velocity, width and clutter correction. Because the clutter correction is more accurate, the resulting reflectivity estimates are more accurate.

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

14

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

Figure 3: Frequency, Time and Amplitude Axis. Pulse Pair: Clutter Correction

Clutter Correction is achieved by passing the incoming I & Q samples through a Time Domain high pass IIR filter. The high pass filter removes any signal power in the 0Hz region i.e. that of apparently non moving targets e.g. clutter. Various cut off frequencies can be chosen to determine how selective we are about removing signals near 0Hz. 1) The unfiltered I&Q samples are auto-correlated and the lag [0] result is known as Uncorrected Reflectivity. 2) The filtered I & Q samples are also auto-correlated but the lag [0] result will now contain the signal power devoid of signals near 0Hz (DC). The difference between the lag [0] results of the filtered and unfiltered data provides an estimate of signal power contributed by any clutter. This estimated value is then subtracted away from the averaged LOG channel data to give
Corrected Reflectivity i.e. Reflectivity corrected for power contributed by a permanent echo.

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

15

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

Pulse Pair: Disadvantages

Pulse Pair clutter filtering is highly invasive to any signal near 0Hz. Recall that the radar can only measure radial velocity. Tangential rain targets will appear as non-moving targets (minimal radial component). Targets with a radial velocity that has folded back into an apparent 0 ms-1 situation will also be filtered by the IIR filter.
Rain rates will be underestimated in either situation.

Figure 4: Rain Over PE Spectra and IIR Clutter Notch (Idealised). Spectral Clutter Processing

I&Q data is transformed from the Time Domain to the Frequency Domain using Fourier Transform methods. Clutter rejection is performed by interpolating across the 0Hz region of the resultant Power Spectra. The reflectivity estimate is not as severely compromised as occurs when using a Time Domain IIR filter.

Figure 5: Rain over PE Spectra and Interpolated Spectra.

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

16

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

Figure 6: Example of Frequency Domain Clutter Filter 2.6. FFT Implementation

If we decide to use only 16 of the collected samples, Fs/N = 1, this means that each frequency varies by 1Hz. We see seven points. The values at frequencies 1, 3 and 5 are correct but in between this is a terrible looking representation of the real FT. All spectrums shown are onesided and were done in SPW.

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

17

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

Case 1: Here we used 16 samples. Now we have a resolution of fs/N = 16/16 = 1. Now we get 7

points, each 1 Hz away.

Figure 7: FFT of a Three-Frequency Signal with N = 16. Case 2: Here we used 32 samples instead of 16. Now we have increased the resolution to fs/N =

16/32 = .5. Now we get 14 points, each .5 Hz away.

Figure 8: FFT of a Three-Frequency Signal with N = 32. Case 3: N = 64 samples. Now fs/N = 16/64 = .25. This Fourier Transform is looking much

better.

Figure 9: FFT of a Three-Frequency Signal with N = 64.

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

18

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

Case 4: N = 128, The FFT looks nearly like the theory.

Figure 10: FFT of a Three-Frequency Signal with N = 128 Case 5: N = 256, This FFT looks quite satisfactory.

Figure 11: FFT of a Three-Frequency Signal with N = 256.

What conclusion can we draw from these? It is clear that the factor fs/N has the largest impact. So we can always improve the FFT by increasing the size of the FFT.

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

19

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

3.

MAXIMUM UNAMBIGUOUS RANGE

The maximum unambiguous range (Rmax) is the longest range to which a transmitted pulse can travel and return to the radar before the next pulse is transmitted. In other words, Rmax is the maximum distance radar energy can travel round trip between pulses and still produce reliable information. The relationship between the PRF and Rmax determines the unambiguous range of the radar. The greater the PRF (pulses per second), the shorter the maximum unambiguous range (Rmax) of the radar. The maximum unambiguous range of any pulse radar can be computed with the formula:
Rmax = c/(2xPRF), where c equals the speed of light. (3x10e8m/s)

Radar transmits many pulses each second. The rate is given by the PRF. The time T between pulses is thus

T = 1/PRF
The range to a target may be determined by the round-trip time of flight for the echo to return to the radar receiver. The "2" accounts for the distance out and back from the target. We know that electromagnetic radiation travels at the speed of light. t = 2r/c Where c=3x108 m/s t=round trip , time (sec)

Now, given T, we can determine the maximum range a radar signal can travel and return before the next pulse is sent out. This is simply:

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

20

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

rmax = CT/2 rmax = C/(2PRF)

Figure 12: Finding the rmax.

Pulse repetition frequency (PRF) largely determines the maximum range of the radar set. If the period between successive pulses is too short, an echo from a distant target may return after the transmitter has emitted another pulse. This would make it impossible to tell whether the observed pulse is the echo of the pulse just transmitted or the echo of the preceding pulse. This produces a situation referred to as range ambiguity. The radar is unable to distinguish between pulses, and derives range information that is ambiguous (unreliable). In theory, it is best to strike a target with as many pulses of energy as possible during a given scan. Thus, the higher the PRF the better. A high PRF improves resolution and range accuracy by sampling the position of the target more often. Since PRF can limit maximum range, a compromise is reached by selectively increasing the PRF at shorter ranges to obtain the desired accuracy of measurements. In the example above, where we had a pulse repetition time of 1 millisecond (1/1000th of a second), we may calculate how far the beam can travel in that time by multiplying 1 millisecond (0.001 seconds) by the speed of light (300,000km/second) for a result of 300km. However, keep in mind that the beam has to be able to reach its target and reflect back in that time which means the total round trip distance is 300km. That means, with a 1 millisecond pulse repetition time, the total range is half that: 150km.

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

21

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

Determining range of a target

In the graphic example to the left, the radar's beam bounces off a raindrop within the cloud and is detected by the radar 425 microseconds (0.000425 seconds) after it was sent. By multiplying the measured time by the speed of light we know that the beam covered 127.5km and we know that half of that distance was the distance to the cloud and the other half was the distance back. So we know the raindrop we detected is 63.8km away.

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

22

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

4.

VELOCITY DETERMINATION

Doppler technology makes the radars enable to determine the velocity of the targets based on their movement from or towards radar. This is very useful information for meteorologists to be able to predict the direction and future location of the air mass and meteorological systems such as cyclones, tornados, etc. Velocity determination can be managed as described in following part of the document.

Figure 13: Velocity Example.

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

Figure 14: PPI Velocity. 23

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

A Doppler radar can only measure the component of the winds in a direction parallel to the Measured wind speed is called the radial velocity (Vr).

radar beam

Radial velocity is defined simply as the component of target motion parallel to the radar radial (azimuth). It is that component of a target's motion that is either toward or away from the radar site along the radial. Some important principles to remember about Doppler radial velocity are: 1. Radial velocities will always be less than or equal to actual target velocities. 2. Actual velocity is measured by radar only where target motion is directly toward or away from the radar. 3. Zero velocity is measured where target motion is perpendicular to a radial or where the target is stationary.

Figure 15: Radial Velocity.

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

24

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

4.1.

Doppler Shift

Austrian physicist Christian Doppler discovered that a moving object will shift the frequency of sound and light in proportion to the speed of movement in 1842. He then developed mathematical formulas to describe this effect called the Doppler Shift. While not given much thought, you experience Doppler shifts many times each day. The change in pitch of a passing train whistle and a speeding automobile horn demonstrate its effects. When you hear a train or automobile, you can determine its approximate location and movement or you hear the high pitch of the siren of the approaching ambulance, and notice that its pitch drops suddenly as the ambulance passes you. That is called the Doppler Effect.

Figure 16: Frequency of Stationary and Moving Target.

Exactly the same thing happens with electromagnetic radiation as happens with sound. Doppler radar accomplishes much the same thing, but to a higher degree of accuracy. As a target moves toward the radar, frequency is increased; if the target is moving away from the radar, the frequency is reduced. In the case of radar, the usual situation is to have stationary radar observing moving targets.

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

25

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

The radar then compares the received signal with the frequency of the transmitted signal and measures the frequency shift, giving the motion and speed of the target. While frequency of electromagnetic energy is modified by moving targets, the change is usually too slight to measure precisely. Therefore, Doppler radar focuses on the phase of electromagnetic energy. Using phase shifts instead of frequency changes can be compared to viewing an insect under a magnifying glass.

Figure 17: The Effect of Moving Target on Frequency.

Figure 18: Wavelength and Amplitude of a Wave.

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

26

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

Figure 19: Phase of a Wave.

Figure 20: Sine Wave (Solid Curve) and a Second Signal 30 Out of Phase with the First Wave (Dashed Curve).

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

27

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

A pulse Doppler radar, in its simplest form, provides a reference signal by which changes in the frequency phase of successively received pulses may be recognized. The known phase of the transmitted signal allows measurement of the phase of the received signal. The Doppler shift associated with the echo from which the return originated is calculated from the time rate of change of phase. The phase of a wave, measured in degrees, where 360 degrees equals one wavelength, indicates the current position of the wave relative to a reference position. For example, look at figure below. At time T1 (fig., view A), the position of the wave along the vertical line was as shown, while at time T2 (fig., view B), the position of the wave along the vertical line was as shown. Notice that the wavelength did not change from T1 to T2. However, the waves position relative to the vertical line changed 1/4 wavelength, or 90 degrees. This change is the phase shift.

Figure 21: Wavelengths and Phase Shifts. (A) T-l is Wave Reference Position. (B) T-2 Waves Position has Changed 90 from Reference Position (T-l).

If the radar observes these changes (phase shifts) it will realize that motion has occurred and can then convert this information into target velocity. Keep in mind that the ability of a Doppler
WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

28

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

radar to detect phase shifts and compute velocity depends upon the system maintaining a consistent transmitter frequency and phase relationship from one pulse to the next.

4.2.

Total Distance to Target in Radians

Consider a single target at distance r from radar. The total distance a radar pulse will have to travel to detect this target is 2r since the wave has to go out to the target and back to the radar. Physical change in target distance is r metres, but the RF path length changes by 2r as signal travels both to and from the radar. Knowing the radars wavelength, 2r (full RF cycle) can be expressed as an observed phase change of target:

The total distance (D) travelled by the wave: 2r This distance can also be measured in terms of the number of wavelengths from the radar to the target: 2r/ We can also measure this distance in radians by using the fact that 1 wavelength = 2 radians. So, D in terms of radians:

Distance in radians = (2r/)/2

If a radar signal is transmitted with an initial phase of <f>0, then the phase of the returned signal will be f0 = phase of pulse sent out by radar f = phase of returning signal then,

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

29

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

= 0 + 4r/

Figure 22: Doppler Velocity.

How does the radar then measure fd?

Figure 23: Phases at Different Ranges.

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

30

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

Figure 24: Basic Block Diagram of Radar.

4.3.

Pulse-Pair Method

1) The transmitter produces a pulse with frequency f0 and duration of t. 2) Some power with frequency f0 is mixed with a signal from STALO and is passed to COHO 3) COHO maintains f0 of transmitted wave 4) Receiver/mixer mixes signal from STALO and received signal 5) Mixed signal is then amplified 6) Phases of original and received signals are differenced, i.e., compute f1 = f0 - f. This is the phase of pulse #1. 7) Repeat 1-6 above for successive pulses. This gives you df/dt.

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

31

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

The change of phase with time from one pulse to the next is given by

d/dt = ( 4/)(dr/dt)
Where dr/dt is the time derivative or time rate of change of the parameter. The radial velocity of an object is given by

v = dr/dt
Angular frequency is the time rate of change of angular velocity (or phase) and is defined by:

= d/dt = 2f
Where f is the frequency shift in cycles per second (Hertz).
Thus, by combining Equations, we get the frequency shift caused by a moving target;

f = 2v/

So a given phase shift in a given interval of time becomes a frequency shift which the radar can measure.

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

32

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

4.4.

Maximum Unambiguous Velocity

What is the maximum Doppler-shifted frequency that can be unambiguously measured?

There are limitations in the velocities and ranges that radar can resolve unambiguously. Let us consider velocity ambiguities first. When a target is not moving toward or away from radar, it will have zero radial velocity. This does not necessarily mean that the target is stationary. It simply means that the target is remaining at a constant distance from the radar. It could be moving quite rapidly, in fact, but any movement it has must be perpendicular to the radar's beam. Since the only velocity a Doppler radar can detect using phase-shift principles is the radial velocity, we usually omit the qualifier "radial" and simply talk about the "velocity". While this is convenient, be careful to recognize that a Doppler radar detects only radial velocities (the velocity with which a target moves toward or away from the radar) The maximum velocity a Doppler radar can detect correctly or unambiguously is given by the velocity which produces a phase shift of TI radians. This is also called the Nyquist frequency or Nyquist velocity7, depending upon whether we are referring to the maximum unambiguous frequency or velocity, respectively. Mathematically, we can express this as:

f = 2v/, vmax =fmax./2

Where the maximum frequency fmax is given by:

fmax = PRF/2
(Nyquist Theorem)

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

33

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

And PRF is the pulse repetition frequency of the radar. Thus, the maximum unambiguous velocity detectable by a Doppler radar is:

v = PRF./4
Example:

If PRF = 1000 Hz and l = 10 cm, then Vmax = 25 ms-1 This is an important result. It says that if we want to be able to detect high velocities, we must use long wavelengths, large PRFs or both.
What is Nyquist theorem?

The Nyquist theorem states that a signal must be sampled at a rate greater than twice the highest frequency component of the signal to accurately reconstruct the waveform

Figure 25: Analogue to Digital Converter (ADC).

Suppose we are sampling a sine wave (How often do we need to sample it to figure out

its frequency?

Figure 26: A Sine Wave.

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

34

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

If we sample at 1 time per cycle, we can think it's a constant

Figure 27: Sampling at 1 Time per Cycle

If we sample at 1.5 times per cycle, we can think it's a lower frequency sine wave

Figure 28: Sampling at 1.5 Times per Cycle

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

35

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

Nyquist rate For lossless digitization, the sampling rate should be at least twice the

maximum frequency responses.

Figure 29: Nyquist Frequency.

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

36

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

5.

THE DOPPLER DILEMMA

vmax = PRF./4 rmax = C.T/2 rmax = C/(2PRF)

According to equations above, longer wavelength radars can measure larger radial velocity unambiguously for the given PRF. Larger the PRF, larger the radial velocity measurable unambiguously from a given radar. Unfortunately, the larger the PRF shorter the unambiguously measurable range [Rmax= C/(2*PRF)]. On the other hand Reducing the pulse repetition frequency (PRF) and allowing for a longer listening time will alleviate the problem of range folding. However, as just discussed, low PRFs may then lead to the problem of velocity aliasing. Thus there is an inverse relationship between the unambiguous range and the unambiguous velocity, the product of which is a constant (vmax.rmax=C/8), where C is the velocity of light. This is widely known as Doppler Dilemma.
When PRF is low----unambiguous range is high---but that results in a low velocity range. When PRF is high----unambiguous range is low---but that results in a high velocity range.

The combination of maximum unambiguous velocity and maximum unambiguous range form two constraints which must be considered in choosing the PRF for use with a Doppler radar. Notice that non-Doppler radars are only constrained by the maximum unambiguous range; since they cannot measure velocity, the velocity constraint does not apply. If we want to have a large Vrmx we must have a small rmax since the right side of the equation is a constant for given radar. Conversely, if we want to detect echoes at long ranges, we can only detect small velocities.

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

37

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

vmax.rmax = C./8
For example, in order for a radar(for wavelength=5cm) to detect radial velocities of

12.5m/s(45km/h) without aliasing, the PRF would have to be increased to about 1,000 pulses per

second. (V=PRF./4 However, this would reduce the maximum unambiguous range of the radar to about 150km(r=c/2PRF). To have an unambiguous range of 300km, the PRF would have to be 500Hz. If PRF is 500Hz then V=6, 25m/s
Another Example: Suppose a radar can sense up to 250 miles from the location of the radar

(unambiguous range) and can detect velocities of up to 30 m/s before velocity folding occurs (a.k.a. velocity aliasing). If the PRF was increased, the unambiguous range will drop to say 200 miles but the unambiguous velocity will increase to say 35 m/s.

Figure 30: Velocity Interval versus Range Interval and PRF at Different Wavelength.

Figure (based on Gossard and Strauch, 1983) shows the Doppler dilemma graphically. Note that the ordinate (Y-axis) on this figure gives the maximum velocity interval corresponding to the Nyquist frequency. Normally we divide this interval in half with the maximum unambiguous velocity being divided into plus and minus half of the VmM interval. For example, from the figure
WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

38

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

we can see that for S-band radar, if the PRF is 1000 Hz, the maximum unambiguous range is 150 km while Vmm is 25 m/s. For X band radar using the same PRF, rmax is still 150 km, but Vmax is now only 8 m/s. For meteorological situations, we may want to measure velocities as large as 50 m/s out to ranges beyond 200 km, so neither of the limits calculated above is completely adequate. The S-band system comes much closer to being useful than the X-band system, however. And C-band will be intermediate to these two.

One partial solution to the Doppler Dilemma is in our choice of wavelength. We can increase both Vmax and r max by using longer wavelength radar. Unfortunately, longer wavelength radars are more expensive and bigger, and they don't detect weather targets as well as shorter wavelength radars, so using a longer wavelength is not necessarily a solution to the problem. The result is that most Doppler weather radars usually suffer significant range or velocity ambiguities or both.

Even if there were not limitations on range because of PRF or velocity, in the real world, we do not wait very long before sending out a second pulse. There are a number of reasons for this. One is that we cannot detect targets at extremely long ranges or we are not interested in them. Meteorological targets typically exist only 10 to 15 km above the earth's surface. Even though the radar waves bend downward somewhat in their travel through the atmosphere, the earth's surface curves away even faster, so the radar beam usually gets so high above the earth's surface that storms are not detectable beyond 400 to 500 km from a ground-based radar. Another reason we are not interested in distant targets is that the inverse square law decreases the power received from a meteorological target according to 1/r2. If a target is too far away, the power received from it will be so weak that the radar will be unable to detect it. For these and other reasons, radars are designed to send out subsequent pulses of energy at fairly frequent intervals.

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

39

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

Figure 31: Range-Height Diagram.

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

40

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

6.

RADAR RANGE FOLDING

While its true that only targets within radars normal range are detected, there are exceptions. Since range ambiguities (also called aliasing or folding) are so common with modern Doppler radars, let us examine the causes of this in a little more detail. Range aliasing occurs because we don't wait long enough between transmitted pulses. This happens when the first pulse of energy goes beyond maximum unambiguous range rmax and sometimes gets returned by a weather at a distance say r. The first pulse returns while the radar is expecting the second pulse (during the listening time of the second pulse). In other word, we transmit pulses close together (mostly to make the Doppler side of the radar work better), not giving one pulse enough time to cover the distance between the radar and some storms before the radar sends out the next pulse of energy. In this case echoes are displayed in the wrong range interval. If the PRF is high enough and distant echoes tall enough and strong enough, sometimes third or even fourth trip echoes can be detected. The radar displays it at a distance (rrmax) superposed on the normal display. These are also known as multi-trip or second-trip echoes in Pulsed radars. Range folding may cause operators to base crucial decisions on false echoes. The data received from this stray pulse could be misanalyzed and echoes may be plotted where nothing exists. The data may look reliable and the radar may appear to be functioning properly, adding to the deception of normal operation.

Figure 32: Second Trip Echo Example.

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

41

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

Figure 33: Second Trip Echo Example.

Figure 34: Rmax and Second Trip Echo Relationship.

Figure 34 shows 2nd trip echoes.

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

42

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

6.1.

Recognizing Range-Aliased Echoes

How are second trip echoes recognized on radar? There are a number of ways multitrip echoes can be recognized. One of the easiest is to simply look outside and see what is going on in the real world. If the radar shows a nearby storm in a particular direction but there is nothing outside, it is probably a multitrip echo.

Figure 35: Second Trip Echo.

Figure Illustration of how a storm beyond rmax can be displayed at the wrong range. Two real echoes exist. The first is less than rmax away and is displayed at the correct range. The second is beyond rmax; it is displayed at a range of (r - rmax). The faint, dashed storm near the radar is where the radar would display the distant storm. A second way to recognize multitrip echoes is by their shapes (see Figure 32 and 33). Real storms are usually somewhat circular, elliptical, or irregular. Storms certainly should not know where the radar is located. Anytime a narrow, wedge-like echo is detected which points toward the radar, second-trip echoes should be suspected. Another clue to the existence of multitrip echoes is height (see Figure 32). Real echoes, especially from convective storms, usually extend up into the atmosphere several kilometres.

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

43

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

Figure 36: Real and Second Trip Echo.

Display showing a real echo located to the northeast. To the east is an echo beyond rmax. It is displayed at a distance of r - rmax from the radar. It also has a very narrow shape. In the real and the aliased positions, but its aliased azimuthal distance is much narrower. Also, its reflectivity will be weaker because of the 1/r2 dependence on received power in the radar equation. Thunderstorms are frequently 8 to 15 km in height. If a convective-like echo appears on the radar display but it has an indicated height which is much less than normal, it may be a second trip echo. For example, a real thunderstorm which is 10 km tall at a range of 200 km would be detectable at an elevation angle of about 2.2 (see Figure 36). If it is a second trip on radar with a PRF of 1000 Hz, it would show up at 200 km - 150 km = 50 km. If the echo from this storm disappears at 2.2, its indicated height would only be 2 km. This is a ridiculously small height for a strong storm, so you should expect range aliasing. Finally, second trip echoes can sometimes be recognized by their reflectivities. The power received from a storm decreases according to 1/r2. If our storm being displayed at 50 km were real, it would have a certain reflectivity. If it is really at 200 km, however, the power returned from it would be (200/50)2 less than if it were at 50 km. So the returned signal would be 16 times less. On a decibel scale this would be 12 dB less than if it were at its indicated range. Unfortunately, since we do not know the true reflectivity of a storm without the radar giving it to us, we cannot be sure that a weak echo is simply a weak storm and not a second-trip storm. Nevertheless, low reflectivity combined with shape and height information can help differentiate real from multitrip echoes.

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

44

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

There is one guaranteed-or-double-your-money-back way to unambiguously determine if echoes are range aliased or not: Change the PRF! If we change the PRF and watch the positions of echoes, all correct echoes will not change their range whereas range-aliased echoes will shift in or out in range, depending upon whether the PRF is increased or decreased. Alternatively, we can avoid range aliased echoes by using a PRF so low that rmax is so large that range aliasing cannot take place.

6.2.

Elimination of Second Trip Echoes (Range Unfolding)

1)

Phase-coding (random phase) of the transmitted signal is employed to filter out

range-overlaid echoes. This phase-coding helps in identifying the second-trip echoes from the first-trip echoes for effectively filtering and displaying them in their appropriate range. 2) 3) Change the PRF Use a different PRF every 2-3 pulses, if echo moves, get rid of it!

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

45

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

7.

VELOCITY FOLDING

If a particle's radial velocity is outside the range of the nyquist interval, then the radial velocity will be aliased or folded. This is called velocity folding/aliasing. Example: if nyquist velocity is 25 m/s and the particle's radial velocity is -30 m/s, then it will fold over and the radar will interpret it as +20 m/s

Figure 37: Velocity Folding or Aliasing.

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

46

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

Figure 38: Folded Velocity Examples.

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

47

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

Figure 39: Unfolded Velocities for This Storm.

7.1.

Staggered PRF for Velocity Unfolding

The maximum radar range is related to the PRF in inverse proportion, while the maximum velocity is related to PRF in direct proportion. Thus for a given range, there is an upper limit for maximum velocity measurable unambiguously. But there are techniques to double or triple the maximum unambiguous velocity by staggering the PRF or using dual PRF. Pulse-transmission rate is toggled from a high value to a low value and vice versa, for every set of fixed number of pulses. The velocity estimates from both sets can be combined suitably to increase the composite unambiguous velocity. Velocity aliasing can cause the two velocity estimates to vary significantly, and these differences can be used to resolve the true velocity. A velocity that has actually exceeded the nyquist velocity can be unfolded to its true velocity. This is achieved by using staggered PRF.

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

48

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

Two different, but related, PRF are used for alternating output rays of data i.e. each 1 deg of azimuth. A 2:3 PRF ratio provides a x2 increase of the apparent nyquist velocity. A 3:4 PRF ratio provides a x3 increase. A 4:5 PRF ratio provides a x4 increase. The technique works by searching for a correlation of the phase shift of the target for the each PRF in use, taking into account that each PRF will produce a different phase shift for the same source velocity. The technique is not without its drawbacks; firstly it relies upon a uniform transition in velocities from ray to ray to allow the correct unfolding estimates to occur. It also introduces several more images of the clutter filter notch previously described, which may result in the elimination of valid rainfall data and produce spoking.

Velocity without unfolding method

PRF 2Vmax = 2

Fdmax =

vmax =

PRF * 4

For PRF = 1200

fdmax = 600 Hz

vmax = 300* =16 m/s

800Hz which corresponds to 21,33m/s in real, but Radar sees this echo as 200Hz which corresponds to 5,33m/s for PRF: 1200Hz (vmax =16m/s) and 350Hz which corresponds to 9,33m/s for PRF: 900Hz (vmax =12m/s)

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

49

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

Velocity with Dual PRF Technique:

Figure 40: Dual PRF Technique.

If 3:4 PRF Ratio applied, folding intersection of Fd max for these two PRF will be 1800Hz In this case new Fd max will be 1800Hz. This means that, radar can detect up to 48m/s), Velocity can be calculated by using two incorrect velocity(5,33m/s and 9,33m/s) with dual-PRF algorithm as 21.33m/s.
PRF Ratio 2:3 3:4 4:5 Fd max 1200Hz 1800Hz 2400Hz Vmax 32m/s 48m/s 64m/s (if PRF1:1200Hz and PRF2:800Hz) (if PRF1:1200Hz and PRF2:900Hz) (if PRF1:1200Hz and PRF2:960Hz)

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

50

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

7.2.

Recognizing Velocity Aliasing

How do velocity-aliased echoes appear on a radar display? The answer to this depends upon where the aliasing takes place. If a large region of echo is being detected by a Doppler radar and a region within it exceeds Vmax, then there will be an abrupt change in velocities surrounding the aliased region. For example, if the storm is moving away and part of it is moving away faster than Vmaxt then strong receding velocities would surround a region with apparently strong approaching velocities. Such a discontinuity is usually quite visible, and it is obvious that velocity folding is taking place. If the storm causing range folding is completely isolated such that there is no surrounding echo, the velocities from the storm may appear entirely correct even though they have been folded into the wrong velocities. This would make recognizing velocity-folded data much more difficult. Fortunately, such isolated situations are not very common, so this is usually not a major problem. There are almost always several echoes on a display at the same time (perhaps even more so when velocities are so strong as to be folded), so velocities of nearby echoes are often useful to indicate whether folding is taking place or not. A more difficult situation, however, occurs when C- or X-band radars are measuring storm velocities. For these radars Vmax can be moderately small. Thus, it is possible to have velocities which are not just folded once but are folded twice or more. This can make it extremely difficult to tell what the true velocities are from a quick visual inspection of the radar display.

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

51

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

8.
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14.

REFERENCES:
Radar for Meteorologists, Ronald E. Rinehart August 1997 Radar Handbook, Merill I. Skolnik Doppler Radar and Weather Observations, Doviak R.J. & Zrnic D.S. Introduction to Radar System, Merrill I. Skolnik Field and Wave Electromagnetics, David K. Cheng,1983 Weather Radar Calibration, R. Jeffrey Keeler January, 2001 Doppler Weather Radar System- Meteor 1000CUser Manuel and DocumentationGematronik GmbH 12.July.2001

RC-57A Weather Radar Training Document and User Manuel- Mitsubishi

Electric Corp. 2002

Radome Influence on Weather Radar Systems, Principle and Calibration Issues

Gematronik GmbH Alexander Manz

Principles of Radar- Wolfgang Manz 12.March .1999 Radar Meteorology- Jrg Joss July.2004 Technical Description TDR Series-C Band Doppler Radar, Radtec Engineering Radar Range Folding and The Doppler Dilemma, Jeff Haby Doppler Radar, A detecting tool and measuring instrument in meteorology
Current Science, Vol. 85, No. 3, 10 August 2003A.K. Bhatnagar, P. Rajesh Rao, S. Kalyanasundorom, S.B. Thampi, R. Suresh and J.P.Gupta

15. 16. 17. 18. 19. 20. 21. 22. 23.

Doppler Weather Radar System, Enterprise Electric Corp. Industrial Assessment of the Microwave Power Tube Industry, Department of
Defense, U.S.A. April 1997

Weather Watch Radar, BoM, Australia Radar Meteorology Doppler, Heikki Pohjoa, FMI Data Quality Improvements on AP Mitigation, Range Velocity Mitigation,
National Weather Service, U.S.A

Radar Training Information, NOAA Detection of ZDR abnormalities on operational polarimetric radar in Turkish weather radar network, WMO- TECO 2005, TSMS, Oguzhan Sireci, 4th.May.2005 Modernization of Observation Network in Turkey, TECO 2005, TSMS, Ercan
Buyukbas, 4th.May.2005

Radar Basics, Renato Croci

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

52

MODULE C- PROCESSING BASICS IN DOPPLER WEATHER RADARS

24. 25. 26. 27. 28. 29.

Feasibility Report for Turkey Radar Network, BoM, Australia,2000 Weather Radar Principles, Firat Bestepe, TSMS, 2005 Principles of Meteorological Doppler Radar, Distance Learning Operations
Course, Instructional Component 5.3. Ver: 0307

Notes on Radar Basics, Serkan Eminoglu, TSMS,2004 Radar Basics, Radar Training Information,NOAA Turkish Radar Network, Hardware Maintenance of Weather Radars, Training Notes, Ercan Bykbas, Oguzhan Sireci, Aytac Hazer, Ismail Temir, Cihan
Gozubuyuk, Abdurrahman Macit, M. Kemal Aydin, Mustafa Kocaman, 2002

30. 31.

Weather Radar Maintenance Procedures and Measurements, TSMS, Aytac

Hazer, Cihan Gozubuyuk, 2005

Operational Use of Radar for Precipitation Measurements in Switzerland

Jrg Joss(1)Bruno Schdler(2) Gianmario Galli(1) Remo Cavalli(1) Marco Boscacci(1) Edi Held(1) Guido Della runa(1) Giovanni Kappenberger(1) Vladislav Nespor(3) Roman Spiess(3) Locarno, 23.Sep.1997

32. 33. 34.

Radar Lecture Notes and Articles available in internet Booklets, reports and guidelines published by WMO Technical Brochures of Radar Manufacturers

WEATHER RADAR TRAINING 5.0 / ISTANBUL-2010

53