Comparisons of tropopause derived from COSMIC measurements at Nanjing since August 2006

Highlights

• Tropopause properties at Nanjing are compared among the COSMIC, ESRL and NNR data sets since 2006.

• Tropopause temperature between the NNR and ESRL shows a warm bias in summer and a cold bias in winter.

• The tropopause properties of the NNR are significantly different from that of the COSMIC data.

• The ESRL and NNR $p_t$ and $T_t$ show a two-year period in terms of their correlation coefficients.

Abstract

Tropopause temperature ($T_t$) and pressure ($p_t$) at Nanjing are derived from the Constellation Observing System for Meteorology, Ionosphere and Climate (COSMIC) from August 2006 to December 2011. We compare$T_t$and$p_t$among the COSMIC, radiosonde provided by the Earth System Research Laboratory (ESRL) and reanalysis data sets from the National Centers for Environmental Prediction–National Center for Atmospheric Research (NCEP‐NCAR) Reanalysis (NNR) on 10-day, seasonal and annual timescales. Ten-day mean$T_t$ and$p_t$of the COSMIC data are higher than that of radiosonde by 0.2 °C and 5.3 hPa and reanalysis by 1.0 °C and 17.5 hPa, respectively. Results of multiple comparisons demonstrate a significant difference for$p_t$between the NNR and COSMIC data. Systematic biases are more significant in low-pressure level than in high-pressure level for reanalysis and radiosonde in terms of seasonal average differences. As for annual mean difference pressure, the COSMIC is higher than ESRL and NNR data by 3.4 hPa–6.5 hPa and 10 hPa, respectively. Besides, the COSMIC and other two data sets are in the best agreement for$T_t$and$p_t$with maximum number of occultation events in 2008. Lastly, quasi-biennial period of correlation coefficients between the ESRL and NNR data sets from 2007 to 2011 requests further verification.

Introduction

The tropopause, lying between troposphere and stratosphere, attracted increasing consideration recently. The World Meteorological Organization (WMO) (1957) has defined the tropopause as “the lowest level at which the lapse rate decreases to 2 °C/km or less, provided that the average lapse rate between this level and all higher levels within 2 km does not exceed 2 °C/km”, which is also known as thermal tropopause. Its properties serve as a hint of climate change. Both observations and numerical models show that tropopause altitude has increased since 1979 (Santer et al., 2003a, Santer et al., 2003b). The tropopause also plays a key role in the upper troposphere and lower stratosphere (UTLS) by affecting chemical tracers exchange (Holton et al., 1995).

In subtropics, the tropopause is of high oscillation due to frequent occurrence of double tropopause (Randel et al., 2007a). It is also sensitive to the expansion of the Hadley Cell (HC) because the boundary of the HC locates there (Lu et al., 2007). Unlike in the tropical and high latitudinal region, $p_t$ is fluctuating rapidly in subtropics with maximum gradient at ~35° (Son et al., 2011), where the NNR data sets underestimate$p_t$.

To precisely identify tropopause properties in subtropics, we use the Global Positioning System (GPS) Radio occultation (RO) observations. This technique has been applied to study the atmosphere utilizing its high vertical resolution, high precision and global coverage (Kursinski et al., 1997, Nishida et al., 2000, Wickert et al., 2001, Hajj et al., 2002, Randel et al., 2003, Anthes, 2011, Guo et al., 2011a, Guo et al., 2011b). Kishore et al. (2009) validated the first year of COSMIC temperature observations by comparing them with three operational analyses. They found the largest differences in the polar region and the maximum differences (2–4 °C) in the tropics tropopause. He et al. (2009) demonstrated a very low mean difference between COSMIC and Vaisala-RS92 and Shanghai radiosonde systems in the UTLS. They also found diurnal radiative effects might cause large temperature biases. In addition, the synoptic scale variability increases biases between COSMIC and radiosonde data for collocation time and distance mismatch (Sun et al., 2010).

Both of $T_t$and$p_t$are also available from the NNR, which have been widely used (Hoerling et al., 1991, Hoinka, 1998, Hoinka, 1999, Highwood et al., 2000, Randel et al., 2007b). Randel et al. (2000) once examined the tropopause of radiosonde and the NNR data in the tropical. Their correlation coefficients of interannual anomalies were about 0.64–0.65 over the period 1957–1997. Birner (2010) argues that underlying reasons are poor centered differences and partial WMO definition without thickness criterion. Son et al. (2011) shows global distributions of tropopause differences between NNR and COSMIC data. They point that the largest bias exists in the subtropics.

The main objective of this paper is to extend the comparison between the COSMIC data, NNR reanalysis and ESRL observations from August 2006 to December 2011 on three timescales. Data sets and method are described in Section 2. Results are discussed in Section 3. In Section 4, we draw a conclusion.