Final Report Summary - CLEOPATRA (CLEaner OPerations Attained Through Radars' Advance)
Executive Summary:
Radar simulators which are able to accurately reproduce the signal coming from a synthetic environment are powerful instruments for the assessment of the performance of new radar systems. In this document we describe a new airborne meteorological radar simulator, namely CleoSim (Cleopatra Simulator), which combines the description of the meteorological scenario at mesoscale level with the capability of generating accurate time series of raw I/Q signals.
CleoSim was developed in the frame of the Cleopatra (CLEaner OPeration Attained Through Radar Advance) project. It is able to simulate the polarimetric I/Q signals for most significant weather phenomena (both threating and non-threating events for aviation safety) as they appear to the airborne X-band weather radars. From such I/Q signal all of the radar observables (absolute and differential reflectivity, Doppler spectrum,…) can be derived through proper post-processing.
The following main topics have been addressed in the frame of the Cleopatra project:
• Meteorological modelling: all meteorological events are described in terms of 3D distributions of mesoscale parameters (temperature, water content, snow content, graupel content,…)
• Electromagnetic modelling: according to both their nature and related mesoscale data, droplets/hydrometeors are characterized in terms of backward and forward propagation complex amplitudes.
• Microphysical modelling: according to both their nature and related mesoscale data, droplets are assigned a PSD (Particle Size Distribution) and terminal fall speed;
• Weather radar modelling: polarimetric I/Q signals are computed according to droplets/hydrometeors electromagnetic and microphysical properties and according to the underlying terrain morphology and nature.
The project team was composed of four partners, providing all required technical expertise:
• IDS, Ingegneria Dei Sistemi S.p.A has been working since 1980 in modelling complex physical phenomena, with particular reference to electromagnetic and radar applications. IDS currently develops and commercializes multi-language, multi-platform CAE tools.
• POLIMI, Politecnico di Milano is a reference in the scientific community for electromagnetic modelling of weather phenomena (backscattering, attenuation, propagation), as seen from weather radar point of view.
• RCF, Rockwell Collins France is one of the international leaders in the field of airborne weather radar equipment. RCF provided deep knowledge of present and future trends about radar architectures and applications.
• KNMI, Royal Netherlands Meteorological Institute, is one of the main European institutes for meteorological modelling and measurements. Therefore KNMI represented the reference point for scenario definition.
Project Context and Objectives:
The Cleopatra project belongs to the frame of the weather radar research and is part of the “Systems for Green Operations” branch of the Clean Sky program, the main driver of which is the eco-compatible design the main goals of which are:
• Reduction of CO2 emissions
• Optimization of aircraft energy consumption
The polarimetric I/Q signals simulated by CleoSim can be exploited in:
• Testing new weather detection algorithms (A-WxR): simulated signals can be used in place of the measured ones;
• The decision support system for trajectory optimization (Q-AI): once processed, the simulated signals do provide meaningful information about the trajectory safety condition;
• In conjunction with flight simulators: the simulated signals can be displayed providing a more complete representation of the scenario surrounding the aircraft.
In order to simulate the polarimetric I/Q signals, several modelling aspects, hereafter described, have been addressed.
Meteorological Model
The meteorological model defines the meteorological environment in terms of a 3D distribution (latitude/longitude/altitude) of mesoscale parameters, by varying the value of which it is possible to represent several meteorological phenomena.
The meteorological environment mainly focused on significant weather hazards for aviation safety (ice crystals, hail, heavy rain, … ) as well as non-threatening weather events which could however impact the airborne weather radar performances (large scale stratiform rain…)
Meteorological data are simulated through Harmonie and are stored on GRIB (Gridded Binary) file
Microphysical and electromagnetic model
The microphysical and electromagnetic model converts the meteorological mesoscale data into microphysical (PSD - Particle Size Distribution) and electromagnetic quantities (transmission and backscattering matrices).
Hydrometeors/droplets are modelled (shape and permittivity) according to their nature. The scattering amplitudes for rain (and for all distorted hydrometeors except ice crystals) are computed using a point-matching technique.
In the following, a brief description of the model used for each specific meteorological event is reported:
• Raindrops are modelled as oblate spheroids whereas cloud water droplets are modelled as spheres. Both PSDs are assumed to be of the gamma type, the parameters (N0, ) of which are related to the temperature and the (cloud) water content. Here below an example of the gamma function, which provides an estimation of the number of the particles per m3 falling in the range of diameter [D, D + dD]:
N(D)=N0Dexp(-D) [m-3 cm-1]
• Ice crystals are modelled as either needles or plates with hexagonal basis, depending on the temperature. The PSD is modelled using the gamma function, the parameters of which are related to air temperature and total cloud ice content. Due to their small dimension, the scattering is computed through Rayleigh approximation (valid up to 50 GHz).
• Hailstones are modelled as ice spheres with a water shell (Mie scattering). The PSD is represented by an exponential function, the parameters (N0, ) of which are related to the intensity of hailstorm (derived from the graupel content). Here below an example of the exponential function:
N(D)=N0exp(-D) [m-3 cm-1]
• Graupel is modelled as spherical particles of density of 0.4 gcm-3 and maximum size of 5 mm. The PSD is described as a gamma function, the parameters of which are related to the graupel content.
• Snow, Melting Layer: snowflakes are modelled as three-medium spheroidal particles made of ice, air and water, the effective complex refractive index of which is calculated using the Maxwell-Garnett mixing rule. During the melting process the electromagnetic, physical and dynamic properties of snowflakes change dramatically with altitude and temperature. The particle size distribution is assumed to be of the gamma type. It changes with altitude following the reduction in particle size. When the melting process is complete, snowflakes are completely turned into raindrops.
Radar Model
The radar model computes the polarimetric I/Q signal by coherently integrating all droplets echo. It manages all of the following modelling aspects:
• Meteorological scenario. Droplets population is generated according to both the mesoscale 3D distribution and microphysical data; the droplets position is updated pulse by pulse, according to the wind speed and terminal fall speed.
• Weather & Terrain Observation Region. Pulse by pulse the portion of both the atmosphere and underlying terrain which is illuminated by the radar beam is estimated. The extension of the beam is defined depending on the type of return signal being computed. For the evaluation of the terrain return, it is usually much larger, including the first side lobes.
• Aircraft kinematics. It is defined as a list of position and attitude samples (yaw, pitch and roll).
• Radar scanning & sampling strategy. It includes the observation volume extension, the maximum range and the temporal scan law for each of the three angles: azimuth, elevation and roll.
• Radar properties. They are represented by the maximum gain, noise figure, side lobe level and radome loss
• Waveform. It is a series of pulses, each being defined in terms of carrier frequency, repetition frequency, width, polarization. The waveform is also assigned a modulation technique (chirp with customizable frequency deviation, Baker Code pulse compression, standard).
• Terrain. It is defined in terms of morphology, (i.e. sea, landscape, mountains, hills) from which all meaningful statistical properties (i.e. mean altitude and standard deviation) are derived and in terms of nature (i.e. grass, trees, sand…) from which reflectivity, for both vertical and horizontal polarization, is evaluated.
Project Results:
The main output of the Cleopatra project is represented by the CleoSim tool, a SW able to evaluate the polarimetric weather radar I/Q signals.
CleoSim comes as a client server SW which makes the computation take place on a central server, letting the human interface reside on the user s’ client.
Measured data, provided by ARPAP (Associazione Regionale per la Protezione Ambientale Piemonte), relevant to a scenario with prevailing stratiform rain, have been used in the validation process.
The same case has been simulated with CleoSim, over a given azimuthal sector, and the related output have been compared against the measured ones.
In order to set up the input meteorological scenario for the simulator, the measured reflectivity data have been processed to estimate back the rain mesoscale parameter, i.e. the water content, by exploiting the know Hagen relationship.
Another test was aimed at verifying the management of polarimetry which is one of the main features of CleoSim.
In case of rain with non-canted hydrometeors, small/medium ranges echo on the horizontal polarization is usually higher than that on the vertical one. This is because the rain oblate particles lay with the major axis on the horizontal plane.
For higher ranges the attenuation is no longer negligible and since it mainly affects the horizontal polarization, the resulting echo prevails on the vertical polarization.
This leads to a sign change in the differential reflectivity.
The above reported validation tests, aimed at verifying the reliability of the amplitude of the I/Q signals. To make validation exhaustive, the related phase had to be tested as well.
This has been done by processing the simulated I/Q signals to derive the corresponding Doppler frequency and velocity.
The scenario foresees a radar scanning azimuthally from one direction, orthogonal to the wind speed up to another being perfectly aligned with it. The Doppler quantities, estimated from the phase of the simulated I/Q signals, do accordingly follow the projection of the wind speed onto the radar bore sight.
In the following, further simulations results, including several meteorological events, are reported. No smoothing process has been applied to the shown plots.
The first scenario foresees rain over a urban terrain. The maximum range has been set to 200km with a clockwise azimuthal scan of 90°.
From the test output, the following can be pointed out:
• On the horizontal polarization, the terrain clutter, which is present for ranges higher than 3km, is much stronger on the horizontal polarization, as reported in literature, and completely masks the rain return. Masking effects on the clutter, caused by the terrain roughness, can be also revealed directly from the clutter distribution.
• On both polarizations, the thermal noise, which is included in the model, is always visible in the background.
The second scenario foresees rain over a sandy terrain. The maximum range has been set to 200km with a clockwise azimuthal scan of 90°. The return on both polarizations is much smaller if compared to the urban soil case. This is due to the lower refractive index of the sand. However, the clutter on the horizontal polarizations is still high enough to mask the return echo from the hydrometeors.
The third scenario foresees the only presence of rain and clutter is not included.
The maximum range has been set to 40km with a clockwise azimuthal scan of 180°.
• The return from the rain is in good accordance with the water content distribution. It is also noticeable that the rainstorm front masks all the meteorological events behind it because of the very high attenuation.
The fourth scenario foresees the only presence of ice and clutter is not included. The maximum range has been set to 40km with a clockwise azimuthal scan of 180°.
• Due to the shape of the ice particles (generally oblates for temperature over -25°, needles below), the echo on the horizontal polarization is about 10dB higher than that on the horizontal one (for an incident angle equal to 90° with respect the local vertical axis).
• The typical rain effects, such as masking and the sign change in the differential reflectivity, are very less likely to occur in the presence of ice because of the smaller attenuation.
The fifth scenario foresees the only presence of hail and clutter is not included. The maximum range has been set to 40km with a clockwise azimuthal scan of 180°.
• Hail model bases on the graupel distribution, exploiting the correlation between the distribution of the two kind of hydrometeors.
• Due to the spherical shape of the hail particles the echo on the horizontal polarization equals that on the vertical one.
Potential Impact:
As already reported, CleoSim has been developed in the frame of the Cleopatra project which is located in the domain of the weather radar research and is part of the “Systems for Green Operations” branch of the Clean Sky program, the main driver of which is the eco-compatible design.
The exploitation of the simulated I/Q data will provide benefits to several fields, environmental ones and radar research related.
In the former case the usage of the simulated data in place of measured one, which require a proper flight planning with related cost, will contribute to the:
• Reduction of CO2 emissions
• Optimization of aircraft energy consumption
• Reduction of noise
Regarding to the research field, the usage of a flexible tool like CleoSim, characterized by many input parameters Users can easily customize according to their needs, will allow for relatively fast parametrical analyses. In this regard, CleoSim may represent a meaningful aid to the radar design, speeding up the definition of new radars requirements.
As far as dissemination concern, a number of actions were implemented.
A project public website was set up an kept updated during the project implementation.
With regard to the production of scientific papers a contribution was submitted to the RADAR 2012 conference held in Glasgow (http://conferences.theiet.org/radar/) and accepted for oral presentation. The presentation was given by prof. D’Amico (POLIMI) in October 2012 to an interested audience.
A full paper to be submitted to “Journal of Atmospheric and Oceanic Technology (JTECH)” (of the American Meteorological Society) was finalized.
The final workshop of the project was held in Paris, during the 50th International Paris Air Show (Le Bourget), on 19 June 2013. The workshop lasted from 09 to 11 am, with the objective to present the work done in the framework of the CLEOPATRA project and the results obtained, as well as to do some brainstorming for future expansions. All the partners described in details their contribution to the development of the CLEOPATRA simulator. Interesting comments were received from the audience.
List of Websites:
In the frame of the Cleopatra project a public website has been developed. It can be reached at the following address: http://cleopatra.dei.polimi.it/.
Radar simulators which are able to accurately reproduce the signal coming from a synthetic environment are powerful instruments for the assessment of the performance of new radar systems. In this document we describe a new airborne meteorological radar simulator, namely CleoSim (Cleopatra Simulator), which combines the description of the meteorological scenario at mesoscale level with the capability of generating accurate time series of raw I/Q signals.
CleoSim was developed in the frame of the Cleopatra (CLEaner OPeration Attained Through Radar Advance) project. It is able to simulate the polarimetric I/Q signals for most significant weather phenomena (both threating and non-threating events for aviation safety) as they appear to the airborne X-band weather radars. From such I/Q signal all of the radar observables (absolute and differential reflectivity, Doppler spectrum,…) can be derived through proper post-processing.
The following main topics have been addressed in the frame of the Cleopatra project:
• Meteorological modelling: all meteorological events are described in terms of 3D distributions of mesoscale parameters (temperature, water content, snow content, graupel content,…)
• Electromagnetic modelling: according to both their nature and related mesoscale data, droplets/hydrometeors are characterized in terms of backward and forward propagation complex amplitudes.
• Microphysical modelling: according to both their nature and related mesoscale data, droplets are assigned a PSD (Particle Size Distribution) and terminal fall speed;
• Weather radar modelling: polarimetric I/Q signals are computed according to droplets/hydrometeors electromagnetic and microphysical properties and according to the underlying terrain morphology and nature.
The project team was composed of four partners, providing all required technical expertise:
• IDS, Ingegneria Dei Sistemi S.p.A has been working since 1980 in modelling complex physical phenomena, with particular reference to electromagnetic and radar applications. IDS currently develops and commercializes multi-language, multi-platform CAE tools.
• POLIMI, Politecnico di Milano is a reference in the scientific community for electromagnetic modelling of weather phenomena (backscattering, attenuation, propagation), as seen from weather radar point of view.
• RCF, Rockwell Collins France is one of the international leaders in the field of airborne weather radar equipment. RCF provided deep knowledge of present and future trends about radar architectures and applications.
• KNMI, Royal Netherlands Meteorological Institute, is one of the main European institutes for meteorological modelling and measurements. Therefore KNMI represented the reference point for scenario definition.
Project Context and Objectives:
The Cleopatra project belongs to the frame of the weather radar research and is part of the “Systems for Green Operations” branch of the Clean Sky program, the main driver of which is the eco-compatible design the main goals of which are:
• Reduction of CO2 emissions
• Optimization of aircraft energy consumption
The polarimetric I/Q signals simulated by CleoSim can be exploited in:
• Testing new weather detection algorithms (A-WxR): simulated signals can be used in place of the measured ones;
• The decision support system for trajectory optimization (Q-AI): once processed, the simulated signals do provide meaningful information about the trajectory safety condition;
• In conjunction with flight simulators: the simulated signals can be displayed providing a more complete representation of the scenario surrounding the aircraft.
In order to simulate the polarimetric I/Q signals, several modelling aspects, hereafter described, have been addressed.
Meteorological Model
The meteorological model defines the meteorological environment in terms of a 3D distribution (latitude/longitude/altitude) of mesoscale parameters, by varying the value of which it is possible to represent several meteorological phenomena.
The meteorological environment mainly focused on significant weather hazards for aviation safety (ice crystals, hail, heavy rain, … ) as well as non-threatening weather events which could however impact the airborne weather radar performances (large scale stratiform rain…)
Meteorological data are simulated through Harmonie and are stored on GRIB (Gridded Binary) file
Microphysical and electromagnetic model
The microphysical and electromagnetic model converts the meteorological mesoscale data into microphysical (PSD - Particle Size Distribution) and electromagnetic quantities (transmission and backscattering matrices).
Hydrometeors/droplets are modelled (shape and permittivity) according to their nature. The scattering amplitudes for rain (and for all distorted hydrometeors except ice crystals) are computed using a point-matching technique.
In the following, a brief description of the model used for each specific meteorological event is reported:
• Raindrops are modelled as oblate spheroids whereas cloud water droplets are modelled as spheres. Both PSDs are assumed to be of the gamma type, the parameters (N0, ) of which are related to the temperature and the (cloud) water content. Here below an example of the gamma function, which provides an estimation of the number of the particles per m3 falling in the range of diameter [D, D + dD]:
N(D)=N0Dexp(-D) [m-3 cm-1]
• Ice crystals are modelled as either needles or plates with hexagonal basis, depending on the temperature. The PSD is modelled using the gamma function, the parameters of which are related to air temperature and total cloud ice content. Due to their small dimension, the scattering is computed through Rayleigh approximation (valid up to 50 GHz).
• Hailstones are modelled as ice spheres with a water shell (Mie scattering). The PSD is represented by an exponential function, the parameters (N0, ) of which are related to the intensity of hailstorm (derived from the graupel content). Here below an example of the exponential function:
N(D)=N0exp(-D) [m-3 cm-1]
• Graupel is modelled as spherical particles of density of 0.4 gcm-3 and maximum size of 5 mm. The PSD is described as a gamma function, the parameters of which are related to the graupel content.
• Snow, Melting Layer: snowflakes are modelled as three-medium spheroidal particles made of ice, air and water, the effective complex refractive index of which is calculated using the Maxwell-Garnett mixing rule. During the melting process the electromagnetic, physical and dynamic properties of snowflakes change dramatically with altitude and temperature. The particle size distribution is assumed to be of the gamma type. It changes with altitude following the reduction in particle size. When the melting process is complete, snowflakes are completely turned into raindrops.
Radar Model
The radar model computes the polarimetric I/Q signal by coherently integrating all droplets echo. It manages all of the following modelling aspects:
• Meteorological scenario. Droplets population is generated according to both the mesoscale 3D distribution and microphysical data; the droplets position is updated pulse by pulse, according to the wind speed and terminal fall speed.
• Weather & Terrain Observation Region. Pulse by pulse the portion of both the atmosphere and underlying terrain which is illuminated by the radar beam is estimated. The extension of the beam is defined depending on the type of return signal being computed. For the evaluation of the terrain return, it is usually much larger, including the first side lobes.
• Aircraft kinematics. It is defined as a list of position and attitude samples (yaw, pitch and roll).
• Radar scanning & sampling strategy. It includes the observation volume extension, the maximum range and the temporal scan law for each of the three angles: azimuth, elevation and roll.
• Radar properties. They are represented by the maximum gain, noise figure, side lobe level and radome loss
• Waveform. It is a series of pulses, each being defined in terms of carrier frequency, repetition frequency, width, polarization. The waveform is also assigned a modulation technique (chirp with customizable frequency deviation, Baker Code pulse compression, standard).
• Terrain. It is defined in terms of morphology, (i.e. sea, landscape, mountains, hills) from which all meaningful statistical properties (i.e. mean altitude and standard deviation) are derived and in terms of nature (i.e. grass, trees, sand…) from which reflectivity, for both vertical and horizontal polarization, is evaluated.
Project Results:
The main output of the Cleopatra project is represented by the CleoSim tool, a SW able to evaluate the polarimetric weather radar I/Q signals.
CleoSim comes as a client server SW which makes the computation take place on a central server, letting the human interface reside on the user s’ client.
Measured data, provided by ARPAP (Associazione Regionale per la Protezione Ambientale Piemonte), relevant to a scenario with prevailing stratiform rain, have been used in the validation process.
The same case has been simulated with CleoSim, over a given azimuthal sector, and the related output have been compared against the measured ones.
In order to set up the input meteorological scenario for the simulator, the measured reflectivity data have been processed to estimate back the rain mesoscale parameter, i.e. the water content, by exploiting the know Hagen relationship.
Another test was aimed at verifying the management of polarimetry which is one of the main features of CleoSim.
In case of rain with non-canted hydrometeors, small/medium ranges echo on the horizontal polarization is usually higher than that on the vertical one. This is because the rain oblate particles lay with the major axis on the horizontal plane.
For higher ranges the attenuation is no longer negligible and since it mainly affects the horizontal polarization, the resulting echo prevails on the vertical polarization.
This leads to a sign change in the differential reflectivity.
The above reported validation tests, aimed at verifying the reliability of the amplitude of the I/Q signals. To make validation exhaustive, the related phase had to be tested as well.
This has been done by processing the simulated I/Q signals to derive the corresponding Doppler frequency and velocity.
The scenario foresees a radar scanning azimuthally from one direction, orthogonal to the wind speed up to another being perfectly aligned with it. The Doppler quantities, estimated from the phase of the simulated I/Q signals, do accordingly follow the projection of the wind speed onto the radar bore sight.
In the following, further simulations results, including several meteorological events, are reported. No smoothing process has been applied to the shown plots.
The first scenario foresees rain over a urban terrain. The maximum range has been set to 200km with a clockwise azimuthal scan of 90°.
From the test output, the following can be pointed out:
• On the horizontal polarization, the terrain clutter, which is present for ranges higher than 3km, is much stronger on the horizontal polarization, as reported in literature, and completely masks the rain return. Masking effects on the clutter, caused by the terrain roughness, can be also revealed directly from the clutter distribution.
• On both polarizations, the thermal noise, which is included in the model, is always visible in the background.
The second scenario foresees rain over a sandy terrain. The maximum range has been set to 200km with a clockwise azimuthal scan of 90°. The return on both polarizations is much smaller if compared to the urban soil case. This is due to the lower refractive index of the sand. However, the clutter on the horizontal polarizations is still high enough to mask the return echo from the hydrometeors.
The third scenario foresees the only presence of rain and clutter is not included.
The maximum range has been set to 40km with a clockwise azimuthal scan of 180°.
• The return from the rain is in good accordance with the water content distribution. It is also noticeable that the rainstorm front masks all the meteorological events behind it because of the very high attenuation.
The fourth scenario foresees the only presence of ice and clutter is not included. The maximum range has been set to 40km with a clockwise azimuthal scan of 180°.
• Due to the shape of the ice particles (generally oblates for temperature over -25°, needles below), the echo on the horizontal polarization is about 10dB higher than that on the horizontal one (for an incident angle equal to 90° with respect the local vertical axis).
• The typical rain effects, such as masking and the sign change in the differential reflectivity, are very less likely to occur in the presence of ice because of the smaller attenuation.
The fifth scenario foresees the only presence of hail and clutter is not included. The maximum range has been set to 40km with a clockwise azimuthal scan of 180°.
• Hail model bases on the graupel distribution, exploiting the correlation between the distribution of the two kind of hydrometeors.
• Due to the spherical shape of the hail particles the echo on the horizontal polarization equals that on the vertical one.
Potential Impact:
As already reported, CleoSim has been developed in the frame of the Cleopatra project which is located in the domain of the weather radar research and is part of the “Systems for Green Operations” branch of the Clean Sky program, the main driver of which is the eco-compatible design.
The exploitation of the simulated I/Q data will provide benefits to several fields, environmental ones and radar research related.
In the former case the usage of the simulated data in place of measured one, which require a proper flight planning with related cost, will contribute to the:
• Reduction of CO2 emissions
• Optimization of aircraft energy consumption
• Reduction of noise
Regarding to the research field, the usage of a flexible tool like CleoSim, characterized by many input parameters Users can easily customize according to their needs, will allow for relatively fast parametrical analyses. In this regard, CleoSim may represent a meaningful aid to the radar design, speeding up the definition of new radars requirements.
As far as dissemination concern, a number of actions were implemented.
A project public website was set up an kept updated during the project implementation.
With regard to the production of scientific papers a contribution was submitted to the RADAR 2012 conference held in Glasgow (http://conferences.theiet.org/radar/) and accepted for oral presentation. The presentation was given by prof. D’Amico (POLIMI) in October 2012 to an interested audience.
A full paper to be submitted to “Journal of Atmospheric and Oceanic Technology (JTECH)” (of the American Meteorological Society) was finalized.
The final workshop of the project was held in Paris, during the 50th International Paris Air Show (Le Bourget), on 19 June 2013. The workshop lasted from 09 to 11 am, with the objective to present the work done in the framework of the CLEOPATRA project and the results obtained, as well as to do some brainstorming for future expansions. All the partners described in details their contribution to the development of the CLEOPATRA simulator. Interesting comments were received from the audience.
List of Websites:
In the frame of the Cleopatra project a public website has been developed. It can be reached at the following address: http://cleopatra.dei.polimi.it/.