2. Moderate problems

2.1. Birds

Birds reflect microwaves. As they move, they are not removed by the Doppler filter. Single bird echoes can be removed by statistical (speckle) filters, but bird flocks remind precipitation very much especially during migration. Some algorithms using either pattern recognition or studying the doppler spectra have been developed.

2.2. Overhanging precipitation

Precipitation aloft is measured correctly by the radar, but it evaporates before hitting the ground. This is very common in warm fronts. Data from surface stations and NWP can be used to recognize and correct for this. Sometimes this is made as part of the vertical profile correction.

2.3. Interfering emitters, jamming.

Radar pictures get infected by mictrowaves coming from other sources: other radars and microwave links. The military disturb radars on purpose (jamming). Most of the emitter signatures can be identified by pattern recognition from polar data. Interfering pulse waves can be filtered in the signal processor, and for continuous waves the SQI filter works sometimes. Also national and international laws should demand better control of frequencies.

2.4 Wet radome attenuation

Microwaves wasted on wet radome. The problem gets worse as the radome gets old and dirty. Could be tacked on time series analysis. the Spanish met. service has some articles.

According to A. Manz (Radcal meeting) the rain influence can only be estimated with a high degree of uncertainty. Laboratory results are extensive, but rain attentuation is always depending on the specific rain rate, the wind conditions and the radome surface conditions. Water as a thin film attenuates more than the same amount in the form of drops. Hence the attenuation can be reduced by hydrophobic coating.

The attenuation could be estimated by ground clutter echoes (Spanish method), if such a target is sufficiently close so that the two-way precipitation attenuation is not dominant.

2.5 Attenuation by icy, sleety, salty or dirty radome

Icy or sleet on radome may produce strong attenuation or even block a sector completely. The only way to deal with this is to provide sufficient heating so that ice is not formed on the radome and if ice and sleet appears, they melt and drop away when the weather condition changes.

Dirt on radome does not increase the attenuation as such. It is not probable that it attaches to the surface to in such amounts that it is important. However, dirt prevents the water from running down the radome and thus increases the water film thickness. The drying of the radome also takes longer. Ageing produces effects similar to those of dirt.

The effect of the salt is pretty much unknown. Salt will prevent water from running down the radome just as dirt does. On the other hand, rain may wash the salt away from the radome, although it may not happen once the salt has dried. Dry salt as such increases the attenuation because the salt crystals contain water.

2.6. Infrastructure: electricity, tower structure

We have assumed that the overall quality of the electricity service is good. Hence no alternated power sources (diesel generators) are installed at the radar sites. A UPS is sufficient to deal with short power cuts, but longer power break stops the radar operation. It is best to use underground electricity cabling in spite of the added cost.

Tower structure should be sturdy enough so that the tower stay erect in wind. For radar on moving platforms (ships and oli rigs) there is software solutions so compensate for the anettna movement.

2.7. Water phase

Radar reflectivity depends on diameter and dielectricity (surface) of the particles. Raindrops are small but glossy, while the same amount of water as snowflakes has bigger particle diameter but less glossy surface. Strongest scattering comes from partially melted snowflakes, as they are large and glossy. Ice particles having sizes close to used wavelength are not even Rayleigh scatterers.

Different Z(R) and Z(S) (reflectivity to rainfall) equations have been derived for rain and snow, but the water phase must be known for selection of the right equation. Also attenuation is affected by water phase, and thus information of 3D water phase is needed. Water phase can be determined by dual polarization radars, VPR analysis and external temperature data sources such as NWP, soundings and weather stations.

2.8. Miscalibration

Miscalibration affects compositing and produces biases in the rainfall rate estimates. It also interferes with precipitation attenuation correction. An overestimate of Z produces an exponentially increasing error.

The choice of the radar hardware and software and regular maintenance and calibration of the radar are the basis for good radar reflectivity measurements. Various radar parameters change with time and the real-time calibration (active source, noise measurement) cannot deal with all these effects. Thus the calibration level of the radar does not necessarily stay valid from maintenance to maintenance. Various tools are available to monitor this. These include rainfall accumulations, integrated reflectivity and numerical comparison of the reflectivicty with Nordrad QA tools. These tools do not yet produce estimates which could be used to correct the raw data.

2.9. Pointing error (elevation)

Antenna pointing error in elevation changes the altitude of the measurement. The effect is largest at low elevations. The altitude is important for the vertical profile correction and thus elevation pointing errors produce errors in the correction.

Antenna pointing in azimuth is rather stable but the pointing in elevation is more vulnerable. Regular maintenance of the antenna system is essential for preventing antenna pointing problems. Antenna elevation is checked by mechanical pointing (plumb line) and sun observations. Methods to monitor antenna elevation include the use of composite images and Nordrad QA tools. Operational sun observations are a promising method which will be developed in the near future. One should also be aware of that the nominal and actual angles are not exactly identical. The effect of this is rather small.

2.10. Z(R, S) equation

Conversion from radar reflectivity to rainfall or snowfall rate depends on drop size distribution. Rough assumptions are used.

2.11. Ships

Sidelobes and anomalously propagating slices of mainlobe can hit ships. Doppler filter does not help as the ships move. As the backscattering cross section of ships is billions times more than that of raindrops, ship-echoes are strong even in sidelobes.

Ships are serious problem escpecially when generating accumulation products and using advective nowcasting methods. Both ship and their sidelobe echoes can be reliably detected and removed by applying pattern recognition methods to polar data.

2.12 Scan strategy

There is various needs for scan strategy, and it is always a compromise. For good quality we need many samples to be processed together (that means slow antenna speed), but for nowcasting frequent updates are preferred. For good Doppler winds high PRFs, for big range small PRFs. For precipitation frequent updates, for 3D structure of the atmosphere several elevations.

2.13 Second trips

Radar detects previous pulse which has traveled beyond rmax, and displays it at distande w=r-rmax. As the squared distance in radar equation is taken as the apparent distance to the echo, second trip echoes appear much weaker than real echoes. Often beyond rmax the measurement is made so high in the atmosphere that there is seldom weak echoes. Problem is biggest when PRF is big (and rmax thus small), and when strong reflectivities appear high in the troposphere, e.g. in hail situations.

SQI threshold removes second trip echoes as they are incoherent. However, careless use of SQI threshold can remove valuable information.