Gastronomia

ADVANCES IN ATMOSPHERIC SCIENCES, VOL. 28, NO. 1, 2011, 139–146

Intercomparison of Humidity and Temperature Sensors: GTS1, Vaisala RS80, and CFH
BIAN Jianchun∗1 ( DUAN Yunjun3 (
1

), CHEN Hongbin1 ( ), XUAN Yuejian1 (

¨ ), Holger VOMEL2 , ¨ ), and LU Daren1 ( )

Key Laboratory of Middle Atmosphere and Global Environment Observation,

Institute of Atmospheric Physics, ChineseAcademy of Sciences, Beijing 100029
2

GRUAN Lead Center, Meteorologisches Observatorium Lindenberg, German Weather Service, Germany

3

Kunming National Reference Climatological Station, Kunming 650034 (Received 15 February 2010; revised 21 April 2010) ABSTRACT

GTS1 digital radiosonde, developed by the Shanghai Changwang Meteorological Science and Technology Company in 1998, is now widelyused in operational radiosonde stations in China. A preliminary comparison of simultaneous humidity measurements by the GTS1 radiosonde, the Vaisala RS80 radiosonde, and the Cryogenic Frostpoint Hygrometer (CFH), launched at Kunming in August 2009, reveals a large dry bias produced by the GTS1 humidity sensor. The average relative dry bias is in the order of 10% below 500 hPa, increasing rapidly to30% above 500 hPa, and up to 55% at 310 hPa. A much larger dry bias is observed in the daytime, and this daytime effect increases with altitude. The GTS1 radiosonde fails to respond to humidity changes in the upper troposphere, and sometimes even in the middle troposphere. The failure of GTS1 in the middle and upper troposphere will result in significant artificial humidity shifts in radiosondeclimate records at stations in China where a transition from mechanical to digital radiosondes has occurred. A comparison of simultaneous temperature observations by the GTS1 radiosonde and the Vaisala RS80 radiosonde suggests that these two radiosondes provide highly reproducible temperature measurements in the troposphere, but produce opposite biases for daytime and nighttime measurements in thestratosphere. In the stratosphere, the GTS1 shows a warm bias (60% over 300 hPa (Fig. 2a). The relative differences in RH as a function of temperature were also calculated (Figs. 2c and 2d). When the temperature is above 0◦ C, both radiosondes have a dry bias of less than 15%. When the temperature is below 0◦ C, RS80 also has a dry bias of less than 22% (Fig. 2d), but GTS1 shows a dry bias that quicklyincreases with altitude, and one of >50% when the temperature is below −20◦ C (Fig. 2c). The dry bias in GTS1 observations will lead to a dry bias in precipitable water (PW). In a PW comparison between radiosonde and GPS data carried out by Wang and Zhang (2008) at Beijing, China, a change since 2002 from a positive bias for the old humidity sensor (goldbeater’s skin) to a general negative biasfor GTS1 was found. Li et al. (2009a) also obtained a dry bias for PW from GTS1 soundings relative to ground GPS observations at Beijing. In comparison to the old sensor, GTS1 has been found to yield a dry bias in RH measurement at Jiuquan (Zhang et al., 2007) and Hangzhou (Zhu and Hua, 2007), and one that increases with altitude. GTS1 has been shown to yield a lower value in dewpoint observation,with the dewpoint bias varying from −1 to −6 K below the 850-hPa level to −8 to −15 K at the 250-hPa level at Guiyang (Min, 2007), as well as at Jinan and Qingdao (Zou et al., 2006). 3.3 Comparisons between daytime and nighttime humidity observations

The relative differences in RH between GTS1 and CFH radiosodnes, and between RS80 and CFH radiosodnes, were calculated. Both GTS1 and RS80radiosondes show a dry bias relative to the CFH in the troposphere (Fig. 2), which is in agreement with the results of Miloshevich et al. (2001), Wang et al. (2002),

At nighttime, the dry bias of the GTS1 hygristor is only about 5% below the 500-hPa level, and 20%– 30% at 350 hPa (Fig. 3a). However, in the daytime, GTS1 shows a much larger dry bias, the difference being 9% below 500 hPa, and increasing...