Results of humidity sensor intercomparisons of new Russian radiosondes
A. Balagourov, N. Krestyanikova,
KOMET, 3/6, Pervomayskaya Str., 141700, Dolgoprudny, Moscow.Reg., Russia.
Tel.: (+7095)4086104, fax: (+7095)4086865, e-mail:
V. Bogov
Join-stock company Meteo, 2, Dolores Ibarruri Str., 620028, Ekaterinburg, Russia.
Tel.: (+7-3432) 427-646, fax: (+7-3432) 464-997, e-mail:
Comparative laboratory tests of performance of new Russian radiosondes (RF95 and MRZ3AM) humidity sensors were carried out under support of Russian Federal Service for Hydrometeorology and Environmental Monitoring (Roshydromet) on the request of the manufacturer of radiosondes MRZ3AM (JSC Метео, Ekaterinburg) at the enterprise KOMET of Roshydromet in the period since August 2000 till April 2001.
Used on the Russian upper-air network radiosondes MARZ-2 and MRZ-3A utilize old temperature and humidity sensors. And, while the temperature sensor, provided the presence of accurate reference and the individual calibration, in principle allows reaching required accuracy of measurement, the humidity sensor has intrinsic shortcomings not permitting to measure humidity with required accuracy. It essentially concedes to the best foreign analogues on accuracy, a temperature range and dynamic characteristics.
Humidity sensors of radiosondes RF95 /1/ - HUMICAP®. New Russian radiosonde RF95 utilizes modern microelectronic temperature and humidity sensors, manufactured by Finnish company Vaisala for the worldwide known radiosonde RS80 (used on the 50% of the global upperair network) that has allowed improving by several times the performance of upper-air data in comparison with radiosondes MRZ3A (MARZ2).
The sensor unit of radiosonde RF95 comprises capacitive microelectronic sensitive elements of meteorological parameters and transducer, providing conversion of temperature, relative humidity and, in general case, pressure of air into frequency of electric pulses repetition. Electronic switch composes the complete telemetry signal, which is taken from the sensor unit output. The humidity sensor presents a thin film capacitor with polymer dielectric. Thickness of polymer is about 1 micron. The sensor is small (4 x 4 x 0.2 mm) and has a time constant no more than 1 s that provides its fast response to ambient air humidity and temperatures variations. Temperature and humidity sensors are placed together on the thin flexible support ensuring their proper exposure in blowing airflow and electrical connections to the transducer.
Radiosonde RF95 has passed the certification tests and is used on the Russian upper-air network since 1997. Nowadays it is launched routinely at 10 upper-air stations mainly in the North-West Region of Russia that, according to ECMWF monitoring statistics, positively influenced corresponding upper-air data quality characteristics for the whole region. In relation to Russian conventional radiosondes MARZ-2 and MRZ-3A performance, the root-mean-square deviations (RMSD) and the maximum errors of RF95 temperature and humidity sensors are 2…3 times smaller. In particular, the maximum permissible error under working conditions in humidity measurements is less of 7%RH. Range of measurement is extended to 0…100%RH in the whole range of working temperatures. The RF95 humidity sensor time constant is 10 times smaller than one of the corresponding MARZ-2 and MRZ-3A sensors. Sampling rate was increased by 7…15 times.
Humidity sensors of radiosondes MRZ-3AM /2/. These radiosondes passed the first stage of tests during 1994-1997 years, were reworked, have stood certification tests in 1998 and at the time being are manufactured by short runs for the conducting of field tests.
Humidity sensor of radiosonde MRZ-3AM (DVR) comprises the sensitive element similar by design to Vaisala humidity sensor and "capacitancetofrequency" transducer. The sensitive element presents a capacitor, formed by three electrodes, two of which (lower) are situated on dielectric base and the third (upper) moist-penetrable one is placed over moist-sensitive coating above lower electrodes and is not connected with them. It is designed as a plate 6 x 4 mm and 0.6mm width, fixed on semi rigid leads soldered to the lower electrodes. A capacitance of the sensitive elements varies within limits of 120…180 pF with average sensitivity of 0.35 pF/%RH. The sensitive element is incorporated into the "capacitancetofrequency" transducer, assembled on 30 x 40 mm PCB.
HUMICAP and DVR sensors took part in the abovementioned Phase I laboratory test of WMO radiosonde humidity sensors intercomparison (Russia, 1995-1997 years) /2/. But afterwards DVR sensor has undergone reworking and passed certification tests as a component of the radiosonde MRZ-3AM.
Below are given some results of investigations of HUMICAP and DVR sensors' static characteristics made using the references, equipment and comparison technique similar to utilized at Phase I laboratory test of WMO radiosonde humidity sensors intercomparison in 1995-1997 years/2/.
The purpose of investigations consisted in a specification of characteristics of new Russian radiosondes humidity sensors, an estimation of their conformity to requirements of the WMO Guide to meteorological instruments and methods of observation and an evaluation of compatibility of sensors of various types.
As the basic reference instrument, which was applied for the investigation of HUMICAP and DVR sensors' static characteristics, it was used the dew-point hygrometer TOROS with the following specifications:
a range of dew (frost) point measurements: -80…29 °C under environmental temperatures 70…30°C and pressures 1100…10 hPa
sensitivity: 0.01 °C
uncertainty: ±0.15 °C (dew point)
±0.3 °C (frost point).
The principle of operation of TOROS is based on the measurement of temperature of chilled mirror at the time of appearance of dew drops or frost. Hygrometer TOROS was certified as a reference instrument by the Russian State Committee on Standardization and Metrology in 1984.
Comparative laboratory tests were carried out under various regimes of temperature and humidity. The common range of temperatures has been from plus 20 up to a minus 57°C along with humidity variations from 10 till 92 %RH. On the Fig. 1 are presented the results of HUMICAP and DVR humidity sensors comparative tests as deviations of their readings from ones of the reference hygrometer. Fig. 2 presents results of processing of comparative tests data as mean (bias) and standard (RMSD) deviations for each sensor type according to temperature. As well, for contrast on the Fig. 1 and 2 are presented in similar way the comparison results for goldbeater skin (GBS) humidity sensors used presently in radiosondes MARZ-2 and MRZ-3A.
From results of tests it is received that radiosondes MRZ-3AM with capacitive humidity sensor had 3…5 %RH bias regarding to the reference hygrometer and standard deviation reaching 5%RH. The maximum registered value of deviation was 14 %RH at temperature 0°C. Nevertheless, approximately in 90 % of cases there were registered the deviations not exceeding 5%RH, that meets to WMO requirements for the accuracy of upper-air measurements of humidity. At the same time the bias of DVR humidity sensor has an obvious trend, which, apparently, is caused by dependence of sensor's readings from temperature. It is necessary to note, that the further work on improvement of DVR characteristics, concerning manufacturing techniques and calibration, should be carried out towards both the accounting of temperature influence and improving of reproducibility.
Radiosondes RF95 with capacitive humidity sensor had rather small values of an error under all investigated temperatures. Bias has not exceeded 0.3 %RH with standard deviation reaching 1.5 %RH. The maximum registered value of deviation was 2.7 %RH at temperature 20°C that is quite compliant to WMO requirements for the accuracy of upper-air measurements of humidity.
Owing to parallel registration of GBS sensors readings during considered comparative tests, it is of interest to compare their performance with one of new Russian radiosondes humidity sensors. From the received results it is visible that GBS humidity sensors' bias exceeds an acceptable level in 10%RH already at environmental temperatures about 0°C. Besides, the GBS deviations from the reference hygrometer has a significant trend which has been noticed also from the results of Phase I laboratory test of WMO radiosonde humidity sensors intercomparison in 1995-1997 years /2/. The maximum registered value of deviation was 37.5 %RH at temperature 57°C that disobey to WMO requirements for the accuracy of upper-air measurements of humidity. It is necessary to note also the value of standard deviation in 7.5 %RH observed as well under the minimal temperature of the tests.
Thus, from results of tests it follows that the performance of DVR and HUMICAP humidity sensors is much better than one of GBS humidity sensors of conventional Russian radiosondes MARZ-2 and MRZ-3A. Accuracy of radiosonde RF95 HUMICAP humidity sensors meets with reserve WMO requirements for the accuracy of upper-air measurements of humidity. Accuracy of radiosonde MRZ3AM DVR humidity sensors basically complies with the requirements; however, it is necessary to carry out work on improvement of production technology and calibration procedure. Results of static measurements of DVR and HUMICAP humidity sensors have shown quite satisfactory compatibility and it is possible to hope that, after improving production process and calibration of DVR humidity sensors, results of humidity measurements of new Russian radiosondes will be quite compatible.
The described comparative tests didn't include investigation of dynamic characteristics as they were quite in detail investigated in Phase I laboratory test of WMO radiosonde humidity sensors intercomparison in 1995-1997 years /2/ and for the past period have not changed. Nevertheless, it is necessary to mention that the dynamic, or lag, error can inroduce the significant contribution to a total error of upper-air humidity measurements. At the same time obtainment of so low values of the response time of HUMICAP sensor allows to realize opportunities of modern means of receiving and processing of upper-air data and, in result, to provide the high vertical resolution and accuracy of the results to meet any requirements of customers concerning content and formats of outgoing upper-air messages /1/.
References:
1. A. Balagurov, V. Grinchenko, A. Kats, 2000, New Instruments for Russian Upper-air Network. Papers Presented at the WMO Technical Conference on Meteorological and Environmental Instruments and Methods of Observation (TECO-2000). Beijing, China, 23-27 October 2000. WMO. INSTRUMENTS AND OBSERVING METHODS REPORT No. 74 (WMO/TD - No. 1028), pp.199-202.
2. A. Balagourov, A. Kats, N. Krestyanikova, 1998, Implementation and Results of the WMO radiosonde humidity sensors intercomparison (Phase I laboratory test). Papers presented at the WMO Technical Conference on Meteorological and Environmental Instruments and Methods of Observsation (TECO-98). Casablanca, Morocco, 13-15 May 1998, WMO. ISTRUMENTS AND OBSERVING METHODS. REPORT No.70. WMO/TD - No.877
Fig. 1. Individual deviations of humidity sensors from the reference hygrometer
Fig. 2. Comparison statistics (regarding the reference hygrometer)