Accurate temperature trend estimates are crucial to monitor decadal climate variability and to understand climate change forcing mechanisms. To construct consistent climate temperature records, long-term well-calibrated satellite temperature measurements with good temporal and spatial coverage from various missions are needed.

The global all-weather temperature distribution can be obtained from Global Positioning System (GPS) radio occultation (RO) data (Hocke 1997; Feng and

National Center for Atmospheric Research (NCAR) and University Corporation for Atmospheric Research (UCAR), P.O. Box 3000, Boulder, CO, USA e-mail: [email protected]

Herman 1999). GPS RO can provide high vertical resolution refractivity, which is a function of temperature, water vapor, and pressure. Because the fundamental observable of GPS RO sounding technique, the phase delay, is accomplished via precise measurement of time that is ensured by ultra-stable atomic clocks on the ground, GPS RO data are ideally suited for climate trend detection. The six-satellite FORMOSAT-3/Constellation Observing System for Meteorology, Ionosphere, and Climate (COSMIC) mission (denoted as COSMIC hereafter) was successfully launched in April 2006. COSMIC will provide approximately 2500 GPS RO soundings per day after it is fully deployed, offering more uniform temporal and spatial coverage than seen in previous GPS RO missions (e.g., CHAllenging Minisatellite Payload, CHAMP (Wickert et al. 2004) and Satelite de Aplicaciones Científicas-C, SAC-C (Hajj et al. 2004), both launched in 2000). Even though we have had GPS RO data from CHAMP and SAC-C since 2001 and from COSMIC since 2006, it is still unrealistic to use only seven years of GPS RO data to detect climate trend.

On board the National Oceanic and Atmospheric Administration (NOAA) series of polar-orbiting satellites, the Microwave Sounding Unit (MSU) and the Advanced Microwave Sounding Unit (AMSU) have also provided near all-weather temperature measurements at different atmospheric vertical layers since 1979 and 1998, respectively. Over the past decade, the roughly 30 years of MSU/AMSU measurements have been extensively used for climate temperature trend detection (Christy et al. 2000, 2003; Mears et al. 2003; Vinnikov and Grody 2003; Grody et al. 2004; Vinnikov et al. 2006; Zou et al. 2006). Even though these satellite missions use similar instruments (from NOAA6 to NOAA14 for MSU and from NOAA15 to NOAA18 for AMSU), the equatorial crossing times of the NOAA satellite orbits drift in local time after launch (Fig. 1), which leads to different temporal sampling of the MSU/AMSU measurements for each NOAA satellite. Because the

1979 1983 1987 1991 1995 1999 2003 2007

Fig. 1 NOAA satellite orbit drifts with local times after launch from 1978 to 2007. Numbers in the figure represent satellite series from NOAA6 to NOAA14

1979 1983 1987 1991 1995 1999 2003 2007


Fig. 1 NOAA satellite orbit drifts with local times after launch from 1978 to 2007. Numbers in the figure represent satellite series from NOAA6 to NOAA14

MSU/AMSU operational calibration coefficients were obtained from pre-launch datasets (Mo et al. 2001), the orbital changes on MSU/AMSU measurements after launch may not be completely accommodated by these calibration coefficients. Different MSU/AMSU missions may contain different measurement biases, which actually vary with times and locations due to on-orbit heating or cooling of the satellite component. This causes great difficulties for climate trend detection. Although different correction methods were proposed by different groups (e.g., Christy et al. 2000; Mears et al. 2003; Grody et al. 2004; Zou et al. 2006), due to lack of an absolute reference, only relative biases are corrected. The temperature trends can still vary as large as 0.1 K/decade when different satellite measurements are used as references (Christy et al. 2003). This leads to extra difficulties for the usage of MSU/AMSU temperature trends for the climate analysis (Christy et al. 2000, 2003; Mears et al. 2003; Grody et al. 2004; Karl et al. 2006; Zou et al. 2006).

The objective of this study is to show the usefulness of GPS RO data to inter-calibrate AMSU brightness temperatures (Tbs) by identifying the orbit-dependent biases for AMSU measurements from different satellite missions and to show the usefulness of the calibrated AMSU Tbs for calibrating other overlapping AMSU Tbs from different platforms. This is to demonstrate that both GPS RO and AMSU/MSU data can be used together to construct consistent climate temperature records. In this study, we use an AMSU fast radiative transfer model to convert COSMIC dry temperature (derived from GPS RO refractivity using the hydrostatic equation and assuming water vapor effect on refractivity is negligible) to synthetic AMSU Tbs for NOAA15, 16, and 18 (N15, N16, and N18), respectively. We focus on the comparison of AMSU temperature in the lower stratosphere (e.g., Tb for AMSU Ch9), where the moisture effect on GPS RO refractivity is negligible. The calibration coefficients (slopes and offsets) among COSMIC synthetic AMSU Tbs and N15, N16, and N18 AMSU Tbs are then constructed. We also construct the calibration coefficients among collocated NOAA AMSU Tb pairs (e.g., N15-N16, N16-N18, and N15-N18). The consistency of the calibration coefficients among COSMIC-NOAA pairs and NOAA-NOAA pairs is examined. We illustrate COSMIC-NOAA and NOAA-NOAA calibration methods in Sect. 2. The comparison results are presented in Sect. 3. We conclude this study in Sect. 4.

2 COSMIC-AMSU and AMSU-AMSU Calibration Methods

In this study, COSMIC RO dry temperature profiles from July 2007 are used to compute the synthetic AMSU Ch9 Tbs. All COSMIC RO dry temperature profiles were downloaded from the UCAR COSMIC Data Analysis and Archive Center (CDAAC)1. An AMSU fast forward model from the Cooperative Institute for Meteorological Satellite Studies, CIMSS, MWFcimss (Hal Woolf, CIMSS, personal communication, 2005) is used to project each COSMIC dry temperature profile into

1 From

synthetic microwave Tbs. The validation of microwave transmittance of this model is described in Woolf et al. (1999). The atmospheric contribution (weighting function) for AMSU Ch9 ranges mainly from 300-10 hPa and peaks near 110 hPa (not shown). The satellite viewing angle is set to nadir for our calculations. The means of AMSU Tbs for N15, N16, and N18 that are collocated with each COSMIC profile within 15 minutes and 50 km are computed. The COSMIC synthetic AMSU Tbs are used as calibration references to inter-calibrate AMSU Tbs from N15 (TbAMSU^i5), N16 (TbAMSU-N16), and N18 (TbAMSU-N18). Only AMSU pixels with satellite viewing angles ranging from -15° to 15° are included in our collocation procedure. The calibration coefficients (slopes and offsets) for each COSMIC-NOAA AMSU pair are then determined from the best fit for each collocated COSMIC and NOAA AMSU ensemble (see Sect. 3.1). This calibration approach is similar to that of Ho et al. (2007, 2009).

Raw counts and ancillary data for each AMSU pixel including viewing angle, location, and time for July 2007 are downloaded from National Environmental Satellite, Data, and Information Service (NESDIS). The operational calibration coefficients for each NOAA AMSU instrument are also downloaded and are used to convert AMSU raw counts into Tbs. The AMSU Tbs from N15, N16, and N18 within 15 minutes and 50 km are collected (e.g., N15-N16, N16-N18, and N15-N18, respectively). Only AMSU pixels with satellite viewing angles ranging from -15° to 15° are included. The calibration coefficients for each N15-N16, N16-N18, and N15-N18 pair are determined by their best linear fits (see Sect. 3.2).

Was this article helpful?

0 0

Post a comment