RADHYD - CHMI Activities

• RADHYD Objectivities & Background
• Scientific & Technical Description of the Project
• Initial status of RADHYD project in CHMI (X.96)
• CHMI Resources Related to RADHYD (X.96)
• Status of work in CHMI (II./III.97)
• Status of work in CHMI (V.97)
• Abstract for Final Seminar COST75 in Locarno, 23-27.3.1998

RADHYD Objectives & Background

In the Czech Republic, an analogue Russian weather radar MRL-5 (X-band) at Prague was digitized in 1992/3. The second equipment, digital Doppler weather radar Gematronik METEOR 360AC, was put into operation in 1995/6. Data from both radars (in time interval of 10, resp. 20 min.) with spatial resolution of 2x2 km are used for calculating of the national composite weather radar information, which is used in the operational weather forecast mostly qualitatively. There is a strong need for quantitative use of radar derived precipitation intensities in operational meteorology and hydrology.

In Swiss, there is more than twenty years experience with the use of digital data from weather radar network (from 1970's two radars, at present 3 radars). There is a long experience with the comparison of radar rainfall and raingauges.

Several improvements of the radar data were invented (different clutter corrections, usage of vertical reflectivity profiles ...). In 1993-95 , Swiss weather radar network was equipped with three radars Gematronik METEOR 360AC and with a new processing software. In the proposed project we would like to use the Swiss experience for developing of algorithms for estimating of precipitation by radar, usable in Central European conditions. We would like to share also the experience with the calibration and technical maintenance of Gematronik radars. The results of the project should be used in the meteorology and operational hydrology in the Czech Republic.

Scientific & Technical Description of the Project

For quantitative use of weather radar data, radar should be first correctly electrically calibrated. As a part of the project, standardized calibration procedures for Gematronik weather radars should be developed. The next precondition of quantitative use of radar data is the elimination of the clutter and other artifacts from radar imagery. In the frame of project, the sitting (influence of radar horizon on measurements) for present radar sites will be evaluated as well. From the radar climatology (summation of data for the period of a month or longer), the areas of better and worse representativeness of radar data will be found. From the results, corrections needed for operational radar data will be constructed.

In the following step, the software for calculating and displaying of radar estimates of precipitation sums will be developed. This software will be used for comparison of radar precipitation estimates and raingauge measurements from the network of SYNOP and INTER weather station network of the Czech Republic. The results will be evaluated statistically and possibly implemented in the operational radar data processing in the Czech Hydrometeorlogical Institute. In the same time, the intercomparison of overlapping radars (2 in Czech, 1 in Slovak Republic) will be carried out.

Initial status of RADHYD project in CHMI (X.96)

First attempts for quantitative use of operational radar data started in CHMI this summer (1996). Short term (1-24 h) radar rainfall accumulations are calculated using Marshall-Palmer formula with a=200, b=1.6 (first guess). Preliminary results (will be presented) shows, that rainfall sum patterns are affected by the radar horison. It seems, that in the near future we need :

1. make the reflectivity images "clean" - do better clutter elimination
2. test the hardware radar calibration
3. find better estimate of the reflectivity near the ground (max.Z or pseudo-CAPPI, correction for horison effects / use of vertical Z profiles)
4. fitting of optimum (fixed) Z-R relation

Our field of interest is to develop an operational "near ground" reflectivity estimate, from which the "rainfall intensity / sum" product for hydrological use can be derived in the next step.

At present, we can contribute to the project by :

1. maintaining of radar data archive
2. calculation of the horisons for CZ + SK radars, evaluation of present radar sites
3. testing of algorithms paralelly to the operational data processing

We need to learn from especially Swiss experience in quantitative use of radar data.

CHMI Resources Related to RADHYD (X.96)

1. Radar sites :

1. Prague - Libus (non-Doppler, X-band MRL-5), in operation since 1989, digitized in 1992/3 by Hungarian modification kit WRP-32C (PC386, MS DOS)
2. Skalky (Central Moravia), Doppler, C-band, Gematronik METEOR 360AC with Rainbow software (PC486, Linux), in operation since 1995/6 (unmanned site)

site     lat N   long E  H[m]  R   Fi   int.   #elev  Min 
Libus    50.008  14.447  356  256  0.5  20/10   13    0.3 
Skalky   49.501  16.788  767  260  0.8   10     20    0.1 

H = height AMSL [m], R = max.range [km], Fi = beamwidth [deg], 
int = measurement cycle [min], #elev = # of elevations, 
Min = minimum elevation angle used [deg] 

2. Operational data products (both radars) :

1. Data resolution (pixel size):
   • 2 x 2 km in horizontal
   • 1 km in vertical
2. Picture size :
   • Single radar : 300 x 300 pixels (including sideviews), ground view 256 x 256 pixels
   • Composite : 400 x 300 pixels (including sideviews), ground view 356 x 256 pixels
3. Data values : 4 bits - values 0, 1, ...., 15
4. Level slicing :
   • Reflectivity Z : data threshold 4 dBZ, step 4 dBZ
   • Echo top height H : data threshold 0 km, step 1 km
5. File format : internal CHMI format for displaying purposes (header + run-length encoded data)
6. Data types :
   • "X" (max.reflectivity) - incl. sideviews
   • "H" (pseudo-CAPPI 1.5km ) - only ground view
   • "T" (echo top) - only ground view
7. Geographical projection of composite radar images is the local Prague-Libus radar (= azimuthal gnomonical) one.
8. Operational radar products archive : on MO discs
   • for Prague radar : since May 1993,
   • for Skalky radar : since Dec.1995

3. Raw data available :

1. Prague radar :
   • Volume scan : 13 x PPI in Cartesian coordinates
   • Resolution : 2x2 km horis., 8 bits (0.5 dBZ steps).
   • Archived on MO discs since May 1993.
2. Skalky radar :
   • Volume scan : dataset in spherical coordinates (20 elev's)
   • Resolution : 1.25 km x 1 deg. horis. 8 bits (0.5 dBZ steps)
   • Presently not archived, available only on site about 6 h after measurement.

4. Present operational procedures needed for quantitative use of radar data :

1. Radar calibration : Hardware, off-line (N months period), no online checking
2. Clutter elimination :
   1. Prague radar : static clutter maps averaged from fine weather conditions, used for individual PPI elevations (in Cartesian coords, 2x2 km resol.)
   2. Skalky radar : Doppler clutter filter (built-in RVP-6)
3. Attenuation correction : none
4. Vertical profile correction : none
5. Raingauge adjustment : none
6. Quality checking of radar data : only subjective, time-by-time

5. Software tools available :

1. Displaying software "RADSHOW" (for CHMI formatted data, operational use) - animation, compositing, printouts of radar products
2. Volume processing of reflectivity data "R3D" : for Prague radar in operational use; for Skalky radar under development
3. Rainfall accumulation "RAINCUM" (for CHMI formatted data, preoperational use)
4. Radar horizon + clutter area calculation "HORI1X" calculation according to COST-73 procedure using digital model of relief with 1 km horizontal resolution, CZ + SK area)
5. Scan strategy calculation "ELEV1"

Status of work in CHMI (II./III.97)

1. Radar coverage maps :

2. Computation of "radar rainfall" sums :

3. Comparison radar-raingauges :

Status of work in CHMI (V.97)

1. Radar coverage maps :

2. Computation of "radar rainfall" sums :

3. Comparison radar-raingauges :

4. Project cooperation :

5. Radar calibration :

Czech weather radar network, consisting of 2 radars - digitized X-band MRL-5 in Prague and unmanned Doppler C-band Gematronik Meteor 360AC at Skalky (Central Moravia) - has been used mostly for qualitative purposes. In 1996, first attempts for quantitative use of data started. Estimates of daily precipitation sums calculated from 4-bit maximum reflectivity products using standard Z=200.R^1.6 relation were compared with the measurement of cca 800 raingauge stations. Preliminary results of one-year operationally running tests shows following :

1. The precision and especially the stability of radar hardware has to be ensured. A method of regular quantitative calibration of the both systems which allows to check the receiver and digitisation chain has been evaluated.
2. Radar horizon (screening by topography obstacles) influences strongly the precipitation amount observed by radar, and with it, the ratio (radar/gauge). Minimum visible pixel height calculated using digital model of relief shows the areas of shadowing where radar informatin needs to be corrected.
3. The reflectivity field used for calculation should be clutter-free and close-to-ground. First results suggest better performance of max.Z field than CAPPI 1.5 km. Attempts to find optimum reflectivity product and to take into account vertical reflectivity profiles for improved estimates of precipitation are presented.

Preliminary results of cooperative Swiss-Slovak-Czech project RADHYD (Exchange of ideas and development of algorithm for estimating precipitation by radar for operational applications in meteorology and hydrology) are outlined.