Variability of the upper troposphere and lower stratosphere observed with GPS radio occultation bending angles and temperatures

Abstract

Recently, Lewis (2009) introduced a new method for the identification of tropopause heights (TPHs) from GPS radio occultation (RO) bending angles $(\alpha )$. The method uses a covariance transform to identify transitions in a $\mathit{ln}(\alpha )$ profile. Lewis validates the results with lapse rate tropopause (LRT) heights from one year of FORMOSAT-3/COSMIC data and radiosondes. In this study we apply the new method to the RO data sets from CHAMP/GRACE (2001–2009) and FORMOSAT-3/COSMIC (2006–2009). These results are the basis for TPH trend estimations for the time period between May 2001 and August 2009 (100 months) based on zonal monthly mean GPS RO data from CHAMP (2001–2008), GRACE (since 2006) and FORMOSAT-3/COSMIC (since 2006). Further, we compare the $\alpha$ based TPH trends with LRT height trends and discuss the differences, which are largest in the subtropical regions (20°–40°) on both the northern and southern hemisphere. A global increase of the TPH between 5 and 9 m/yr is found for both methods and different data sets (CHAMP/GRACE alone and CHAMP/GRACE plus FORMOSAT-3/COSMIC). The results for the TPH trends are linked with bending angle and temperature trends in the upper troposphere and lower stratosphere region. Generally, an upper tropospheric warming (bending angle decrease) and a lower stratospheric cooling (bending angle increase) is noted.

Introduction

The upper troposphere and lower stratosphere (UTLS) region is one of the key regions of the atmosphere with important links to the stratosphere-troposphere exchange as well as climate research. The determination of UTLS temperature and tropopause height trends are crucial for the monitoring of climate change processes. The global mean tropopause height, e.g., shows an increase in reanalyses and radiosonde observations during the last decades (Seidel and Randel, 2006), whereas an upper tropospheric warming and a lower stratospheric cooling is evident (Randel et al., 2009). The changes in the tropopause height are caused by different forcing mechanisms (Shepherd, 2002, Santer et al., 2004). One mechanism leading to an increase of the tropopause height is a warming of the troposphere (due to more CO₂) and a cooling of the lower stratosphere (due to less stratospheric ozone), both observed during the last decades. This might be an evidence for anthropogenic warming rather than natural warming from the sun (Shindell, 2008). The most important data source for the determination of tropopause parameters are temperature data from radiosondes whereas model analyses suffer from coarser vertical resolution. Despite of good vertical resolution of radiosonde measurements a global coverage is impossible.

Global Positioning System (GPS) radio occultation (RO) enables atmospheric information with high vertical resolution ($<1\text{km}$ in the tropopause region). The GPS RO technique requires no active calibration, is weather independent, and the occultations are almost uniformly distributed over the globe (Melbourne et al., 1994, Kursinski et al., 1997). The RO method exploits GPS signals received onboard a Low Earth Orbiting (LEO) satellite for atmospheric limb sounding. The GPS signals are influenced by the atmospheric refractivity field resulting in a time delay and bending of the signal. The atmospheric excess phase is the basic observable that is measured with millimetric accuracy (Wickert et al., 2001a) and is considered as a climate benchmark observable (Leroy et al., 2006, Ho et al., 2009).

A detailed general description to derive vertical atmospheric profiles from RO measurements is presented by Kursinski et al. (1997). Here, we give a brief summary: the GPS receiver onboard the LEO records phase and amplitude variations with high temporal resolution (e.g., 50 Hz) during an occultation event. By using high precision orbit information from the LEO and the occulting GPS satellite the atmospheric excess phase can be extracted which is related to a bending angle profile $(\alpha )$ as a function of the impact parameter a.

$$a=\mathit{nr}=n(z+R_c)$$

(n: refractive index, r: tangent radius, z: geometric height, and $R_c$: local radius of curvature).

Assuming a local spherically symmetric atmosphere the bending angles can be related to the refractive index n using an Abel transform. Finally, the atmospheric refractivity N is given by (Smith and Weintraub, 1953):

$$N=(n-1)·{10}^6=77.6\frac{p}{T}+3.73·{10}^5\frac{e_w}{T^2}$$

(p: total air pressure, T: air temperature, and $e_w$: water vapor pressure).

Due to the ambiguity between the dry and wet part in the refractivity (Eq. (2)), it is impossible to derive the temperature or water vapor independently. However, above that parts of the troposphere where water vapor can be neglected $(T<250\text{K})$, a dry atmosphere is a good assumption (Kursinski et al., 1997). Combined with the hydrostatic equation pressure and temperature profiles can be calculated from RO refractivity profiles.

The proof-of-concept GPS RO experiment GPS/MET (GPS/METeorology) performed between 1995 and 1997 has demonstrated for the first time the potential of GPS based limb sounding from LEO satellites for deriving atmospheric temperature and tropospheric water vapor profiles (Kursinski et al., 1997, Rocken et al., 1997). Further missions with RO experiments followed (Ørsted, SAC-C), but the German CHAMP (CHAllenging Minisatellite Payload) mission launched in July 2000 was the first RO experiment providing RO data continuously in an operational manner between May 2001–September 2008 (Wickert et al., 2001b, Schmidt et al., 2005a). The U.S.-German GRACE (Gravity Recovery And Climate Experiment) mission continuous this first long-term RO data set since mid-2006 (Wickert et al., 2009). A new milestone for sounding the atmosphere with GPS RO was the launch of the six-satellite FORMOSAT-3/COSMIC (Constellation Observing System for Meteorology, Ionosphere, and Climate) mission in 2006 (Cheng et al., 2006, Anthes et al., 2008). Further RO missions whose data are not considered in this study are the European Metop (since 2006) and the German TerraSAR-X (since 2007) missions.

In the past, several comparison studies between GPS RO data and radiosondes as well as cross-validations with other satellite sensors have been performed (Hajj et al., 2004, Kuo et al., 2005, Steiner et al., 2007, Gobiet et al., 2007). Special comparisons of tropopause parameters between CHAMP and ECMWF (European Centre for Medium-Range Weather Forecasts) or NCEP (National Centers for Environmental Prediction) can be found in, e.g., Borsche et al. (2007), Schmidt et al., 2004, Schmidt et al., 2005b.

For a climatological validation of stratospheric temperatures in the altitude range from 10 to 30 km between ECMWF and CHAMP for 2001–2004, see Gobiet et al. (2005). An extended comparison study between CHAMP and ECMWF for the lower stratosphere temperatures and geopotential heights was performed by Schmidt et al. (2008b) and the effects of main changes in the ECMWF assimilation scheme in 2006 were also discussed.

The general potential of GPS RO data for climate monitoring has been shown with simulation studies (Leroy et al., 2006), as well as with the CHAMP data set from 2001 to 2007 (Schmidt et al., 2008a). More recently Ho et al. (2009) compared CHAMP refractivity trends (2002–2006) from different data centers. This study can be considered as a reference for the discussion of the different errors relevant for climatological (trend) investigations with GPS RO data. Steiner et al. (2009) discuss temperature trends based on GPS/MET (1995/97) and CHAMP (2002–2008) data.

As mentioned above the determination of temperatures as a function of geometrical heights inserts several assumptions (spherical symmetry, dry air) and a priori information for the integration of the hydrostatic equation. Therefore, to accomplish climate studies with RO data the direct use of bending angle information seems to be appropriate. The RO bending angle has been successful assimilated into operational weather forecast systems and demonstrates significant impacts (Healy and Thepaut, 2006), whereas Ringer and Healy (2008) exhibit the potential of using bending angles for climate monitoring.

Recently, Lewis (2009) introduced a new method for the identification of tropopause heights (TPHs) from GPS RO bending angle $(\alpha )$ profiles (Section 3) and applied the technique to FORMOSAT-3/COSMIC data between May and November 2008 and radiosondes. We apply this method for the complete CHAMP and GRACE data sets (until August 2009) as well as for the reprocessed FORMOSAT-3/COSMIC data from April 2006–March 2009 provided by UCAR (University Corporation for Atmospheric Research) and compare the $\alpha$ based TPHs (called TPH$(\alpha )$ in the following) with conventional lapse rate tropopause (LRT) heights defined by the WMO (1957).

The paper is structured as follows. Section 2 shows the data base used in this study. In Section 3 we give an overview over the method of Lewis (2009) and Section 4 exhibits the validation of the new TPH$(\alpha )$ method with LRT heights. Section 5 summarizes linking mechanisms between TPHs, tropospheric and stratospheric variations. Sections 6 Determination of zonal monthly means, 7 Statistical methods describe the methods for the determination of zonal means and the statistics applied to the different data sets. In Section 8 tropopause height trends based on the different TPH algorithms are discussed. Finally, Section 9 presents some results showing the linkage between the TPH trends and bending angles as well as temperature trends in the upper troposphere and lower stratosphere (UTLS) region from the GPS RO data sets.