1. Introduction
As the third pole of Earth, the Tibetan Plateau has a complex subsurface with extensive glaciers, permafrost, grass, and lakes. The complex surface and atmospheric processes over the region make the Tibetan Plateau a sensitive and vulnerable area for global climate and environmental change. The atmospheric temperature is an important variable in the climate change, and understanding its change and climate trends over the Tibetan Plateau are important for international research on global change (Wu et al. 2013).
Meteorological data are the basis for the study of atmospheric temperature climate trends over the Tibetan Plateau. China Meteorological Administration (CMA) and the Chinese Academy of Sciences jointly conducted two atmospheric observation experiments over the Tibetan Plateau in 1979 and 1998, respectively (Wu et al. 2004). More than 200 ground stations, more than 20 radiosonde stations, and other stations for observations of radiation and heat source were built inside and outside the Tibetan Plateau. From 2006 to 2009, Chinese and Japanese scientists jointly implemented the “New Generation Integrated Meteorological Observation Program on the Tibetan Plateau and its Surroundings” (Xu et al. 2019), whose main observation targets were to investigate the hydrological cycle and its influence over the Tibetan Plateau and its surrounding areas. In 2015, the 3rd CMA Tibetan Plateau Atmospheric Science Experiment was conducted from June to August in both 2015 and 2016 (Zhao et al. 2019). Nine and six new radiosonde stations were added in the central and western parts of the Tibetan Plateau, respectively.
Global reanalysis datasets are another type of data source that provide spatiotemporal continuous data for studying the characteristics of weather and climate change (Kalnay and Cai 2003; K. Yang et al. 2011; Zhou et al. 2004). A global reanalysis dataset was generated with a stable data assimilation system and weather forecast model setting to largely eliminate any unnatural effects on climate studies, such as replacements of data assimilation and/or forecast model from time to time, and satellite instrument dependent data biases. With rapid advances in satellite observing techniques and data assimilation, difference between reanalysis data and observations has been significantly reduced (Bao and Zhang 2019). However, although global reanalysis, such as the ERA5 reanalysis (Hersbach et al. 2020), assimilated AMSU-A TB observations and many other observations, the quality of the reanalysis of atmospheric temperature profile are affected on how well cloud detection, bias correction and quality control are implemented in the reanalysis data assimilation system. In fact, all-sky simulations could differ greatly from AMSU-A TB observations, especially on structural patterns and locations of cloud and precipitation (Niu and Zou 2022). This study is a first step to directly examining TB trends in the Tibetan Plateau using AMSU-A observations. A follow-on study is to examining TB trends over the Tibetan Plateau using AMSU-A all-sky simulations from several reanalysis datasets and from different climate model forecasts. A comparison of TB trends over the Tibetan Plateau among AMSU-A observations and all-sky simulations from reanalysis and climate model forecasts may offer new insights on the quality of global reanalysis and climate model forecasts, which were popularly used in weather and climate studies over the Tibetan Plateau.
Using reanalysis datasets, station observations, and climate simulations, many studies have demonstrated a rapid warming phase over the Tibetan Plateau due to an increase of greenhouse gases (Duan et al. 2016). A rapid reduction of ozone lead to a greater tropospheric warming and stratospheric cooling over the Tibetan Plateau than its surrounding areas (Zhang and Zhou 2008). The cloud radiation feedback process could be another reason for an accelerated warming after 1998 (Duan et al. 2016). In addition to the cloud radiation feedback, snow albedo feedback, which increases with terrain height, also contributed to the warming (Liu et al. 2009; X. Yang et al. 2011).
Conventional observation sites are very limited over the Tibetan Plateau and are mostly located on the eastern part of the Tibetan Plateau (Lu et al. 2015; Duan et al. 2016). The insufficiency of conventional observations also affects the accuracy of reanalysis data, leading to a larger uncertainty over the Tibetan Plateau than elsewhere (Bao and Zhang 2019). Polar-orbiting environmental satellite (POES) remote sensing instruments provide global temperature sounding observations of Earth and its atmosphere from space. The Microwave Sounding Unit (MSU) and the Advanced Microwave Sounding Unit-A (AMSU-A) observe thermal radiation at multiple channels located in the oxygen absorption band of 50–60 GHz to provide mainly information of the atmospheric temperature in the troposphere and low stratosphere. Other satellite temperature sounding instruments include the Stratospheric Sounding Unit (SSU) instrument that provides atmospheric temperature information from 20- to 55-km altitude (Wang et al. 2012; Nash and Saunders 2015; Zou et al. 2014), the Atmospheric Chemistry Experiment-Fourier Transform Spectrometer (ACE-FTS) (Bernath 2017) and the Global Ozone Monitoring by Occultation of Stars (GOMOS) (Sofieva et al. 2019; Hauchecorne et al. 2019). The latter two instruments provide information on atmospheric temperatures in the stratosphere and mesosphere. In this study, we employ AMSU-A observations of brightness temperature (TB) from NOAA-15, NOAA-18, NOAA-19, MetOp-A, and MetOp-B to investigate observed temperature changes in the troposphere and low stratosphere from 1998 to 2020, with emphasis on the Tibetan Plateau.
The earliest satellite-based global temperature trend studies made use of MSU and AMSU-A TB observations (Christy et al. 1998; Christy and Norris 2004; Mears et al. 2003; Vinnikov and Grody 2003; Ohring et al. 2005; Fu et al. 2004; Mears and Wentz 2005). They used MSU channel 2 and AMSU-A channel 5 to represent midtropospheric atmospheric temperatures around 5 km, MSU channel 3 and AMSU-A channel 7 to represent upper-tropospheric atmospheric temperatures around 10 km, and MSU channel 7 and AMSU-A channel 9 to represent atmospheric temperatures in the lower stratosphere around 17 km. Using ground- and satellite-based temperature observations of the upper-air atmosphere that were available for more than 40 years from 1979 to 2018, Steiner et al. (2020) showed a robust cooling of the stratosphere of about 1–3 K, and a robust warming of the troposphere of about 0.6–0.8 K over the last four decades (1979–2018). The satellite-based layer-average temperature trends were found to be consistent with those from vertically resolved radiosonde data records.
Different from past studies, we are interested in geographical differences in climate trends, even in different parts of the Tibetan Plateau in this research. A double difference method is first used to remove impacts of inter-satellite biases on climate trends (Xia and Zou 2021). A limb correction is then applied to allow use of AMSU-A TB observations at all fields-of-views (FOVs) (Goldberg et al. 2001; Tian et al. 2018). A global AMSU-A TB dataset 0.5° × 0.5° resolution is finally established by averaging daily observations at descending and ascending nodes. Characteristic features of climate trends at different time scales over the globe and the Tibetan Plateau are examined using the empirical ensemble mode decomposition (EEMD) method (Wu and Huang 2009; Qin et al. 2012).
The rest of paper is organized as follows: Section 2 provides a brief description of AMSU-A TB observations. Section 3 describes an inter-satellite bias calibration, a limb correction, a global gridded AMSU-A TB dataset for the 22 years from 1998 to 2020, as well as a brief description of the EEMD method. Results of limb correction effect on global TB observations and climate trends are discussed in section 4. In section 5, a scale decomposition of the time series of AMSU-A TB at several sounding channels is carried out at several arbitrarily selected grids over the Tibetan Plateau, pointing out clearly a modulation effect of the climate change of annual cycle on the warming trends over the Tibetan Plateau. Conclusions are given in section 6.
2. A brief description of AMSU-A observations
The detection frequency of AMSU-A is in the oxygen absorption band between 23.8 and 89.0 GHz. Based on the uniformity of oxygen, the AMSU-A 15 channels with different peak weighting function heights can detect the atmospheric temperature information from Earth’s surface to the stratosphere. AMSU-A adopts a cross-scanning mode, which can observe within the range of 48.33° scan angles from the satellite nadir position. Each scanning line has 30 FOVs and each FOV has a beamwidth of 3.3°. An FOV at nadir is circular with a diameter of 45 km. The height of a POES is generally about 833 km above Earth, and the width of the swath is about 2343 km.
As the first POES equipped with AMSU-A instrument, NOAA-15 was launched on 13 May 1998. Subsequent NOAA POES series from NOAA-16 to NOAA-19 had also equipped with AMSU-A instrument. Carried on the Suomi National Polar-Orbiting Partnership (SNPP) and NOAA-20 are a new microwave sounder named the Advanced Technology Microwave Sounder (ATMS). The European Organization for the exploitation of meteorological satellites (EUMESAT) launched three POESs in the MetOp series of polar-orbiting satellite program since 19 October 2006. They are named MetOp-A, MetOp-B, and MetOp-C. All three MetOp satellites were equipped with AMSU-A instruments. The TB observations at AMSU-A sounding channels 6–13 from NOAA-15, NOAA-18, and NOAA-19, MetOp-A, and MetOp-B are used for this study. These temperature-sounding channels are located in the upper troposphere and low stratosphere. The noise equivalent differential temperature (NEDT) is 0.25 K for channels 6–9 and 0.4 K for channel 10.
Figure 1 shows the time periods of available AMSU-A TB observations for each of channels 6–13 from the five POESs from 25 October 1998 to 30 November 2020. We plotted the data period for each channel at the pressure level of its peak weighting function (WF). When two POESs overlap, we choose the newly launched one to reduce impacts of instrument aging on data quality. NOAA-16 had a significant orbital drift and its AMSU-A data are not included in this study (Wang and Zou 2014). The operating time of NOAA-17 was too short to be used (Zou and Wang 2011). According to the noise information provided by NOAA integrated calibration/validation system (ICVS) website, the following AMSU-A data are not used when the noise is much greater than the observation signal: NOAA-15 channel 11 since April 2002, NOAA-15 channel 6 since 2006, NOAA-18 data since August 2005, NOAA-19 channel 8 since 2010, MetOp-A channels 7 and 8 after December 2009 and April 2016, respectively. Except for channels 10–13, which have a data gap from April 2002 to March 2005, the above said elimination of low-quality TB observations does not cause any data vacancy of channels 6–9 due to an overlapping time of NOAA-18 with NOAA-15.
Time period of available AMSU-A data at channels 6–13 from NOAA-15 (blue), NOAA-18 (black), NOAA-19 (cyan), MetOp-A (green), and MetOp-B (orange) from 25 Oct 1998 to 30 Nov 2020, and terrain height from 60° to 120°E at 35°N. Each channel is plotted at the pressure level of its peak WF.
Citation: Journal of Climate 36, 9; 10.1175/JCLI-D-22-0517.1
The terrain height over and around the Tibetan Plateau is shown in Fig. 2. It varies from 3 to 5 km over most areas of the Tibetan Plateau. The terrain height at 35°N along the zonally dotted line in Fig. 2 is shown in Fig. 1. It can be seen that the weighting function peaks of AMSU-A channel 6–13 are well above the Tibetan Plateau, so that the observed TB values are less influenced by the Tibetan Plateau’s surface inhomogeneity.
Terrain height over and around Tibetan Plateau. Also indicated are a vertical line at 85°E (for Figs. 12a,b) and a horizontal line at 35°N (for Figs. 12c,d), as well as three data points located at 35°N, 87.5°E (for Fig. 8) and 32°N, 85°E and 48°N, 85°E (for Figs. 9 and 10).
Citation: Journal of Climate 36, 9; 10.1175/JCLI-D-22-0517.1
3. Establishing a long-term AMSU-A gridded dataset
a. Limb correction
AMSU-A instruments have a cross-track scanning mode. AMSU-A received atmospheric emissions from larger scan angles go over a longer optical paths than those near nadir. Therefore, observations along a scanline reflect the atmospheric temperature information at different heights (Mears and Wentz 2009). This is illustrated in Fig. 3, which shows the vertical profiles of channel-6 weighting functions at FOVs 1, 4, 7, 10, 13 and 15. Weighting functions at FOVs 16–30 are similar to those of FOVs 15–1. FOVs 15 and 16 are nearest to the nadir and have the smallest scan angle. The weighting function at FOV 15 has its maximum value near 400 hPa. As scan angle increases from FOVs 15 to 1, the weighting functions shift upward. The weighting function at FOV 1 has its maximum value near 220 hPa. The resulting impact on TB observations is called the limb effect. As a result, weather features are concealed in TB observations by the prevailing scan pattern. An example is provided in Fig. 4a. Global TB observations of AMSU-A channel 8 at the MetOp-A ascending nodes on 1 January 2008 are dominated by a limb effect, especially in the tropics. The TB observations at larger scan angles are more than 8 K lower than those near the nadir. This is a combined result of the atmospheric temperature decreasing with altitude in the troposphere and weighting function peak increasing with altitude.
Weighting functions of channel 6 at different scan angles of FOVs 1 (blue), 4 (cyan), 7 (green), 10 (red), 13 (magenta), and 15 (black) calculated by CRTM using the U.S. Standard Atmosphere, 1976 (COESA 1976) profile. The weighting function peak at FOV 15 (near nadir) is indicated by gray dashed line.
Citation: Journal of Climate 36, 9; 10.1175/JCLI-D-22-0517.1
Global distributions of channel-8 TB observations (a) before and (b) after the limb correction at the MetOp-A ascending nodes on 1 Jan 2008.
Citation: Journal of Climate 36, 9; 10.1175/JCLI-D-22-0517.1
The limb effect appears in all channels. Goldberg et al. (2001) proposed a limb correction algorithm to remove the scan-dependent patterns in AMSU-A TB observations. The limb effect in a targeting channel is removed by a linear regression of a few neighboring channels whose weight functions are right above and below the weighting function of the targeting channel. A detailed mathematical description is provided in Tian et al. (2018).
The global distribution of TB observations is similar to Fig. 4a, but the after limb correction is shown in Fig. 4b. The limb effect on TB observations at AMSU-A channel 8 seen in Fig. 4a is successfully removed. The weighting function has its maximum near 175 hPa. As expected, in January, TB observations in the Southern Hemisphere (SH) are significantly higher than those in the Northern Hemisphere (NH). High TB observations of channel 8 also appear along storm tracks in the NH, especially in the Northern Pacific. Due to the influence of storm axis, frontal surface often appears in this area (Chang and Fu 2002). The warmer stratospheric atmosphere invades into the troposphere along the frontal surface, resulting higher TB observations in storm track areas than their surrounding areas.
b. Inter-satellite bias calibration
A single satellite has limited lifetime. For climate research, we need a long time record of AMSU-A observations. Fortunately, the international community has continuously launched a number of polar-orbiting satellites carrying AMSU-A and AMSU-A like temperature sounders, which avoid any data gaps. However, the AMSU-A observations from different satellites cannot be simply connected. TB observations were obtained by a two-point calibration equation based on measured counts and underwent a careful calibration (Mo 2009), but are not absolutely calibrated (Zou et al. 2014). An inter-sensor bias correction is thus required when connecting AMSU-A observations from multiple satellites (Mears et al. 2002; Mo 2007).
One way to eliminate systematic data biases among different satellites is to select a reference satellite and adjust TB observations from other satellites to those of the reference satellite (Santer et al. 2017). In this study, NOAA-18 satellite is selected as the reference satellite. Observations from the NOAA-18 satellite (ONOAA-18) and those from another satellite (Osat, where the subscript represents any of NOAA-15, NOAA-19, MetOp-A, or MetOp-B) that are collocated with ONOAA-18 in space and time are first found through the simultaneous nadir overpass method. A double difference method is used to estimate the difference between NOAA-18 bias (μNOAA-18) and “sat” bias (μsat):
c. Gridded satellite AMSU-A dataset
A POES provides global remote sensing observations twice daily at two fixed local equatorial crossing times (LECTs) at its ascending and descending nodes that are about 12 h apart. Two POESs with different LECTs observe the same region at different local times. To reduce the impact of diurnal variation of temperature and satellite orbital drift to a certain extent (Po-Chedley et al. 2015; Wang and Zou 2014), AMSU-A observations at ascending and descending nodes of the same satellite are added together to produce daily observations. The impact of diurnal variation of temperature and satellite orbital drift can be reduced to a certain extent (Po-Chedley et al. 2015). Specifically, we convert the limb-corrected TB observations onto a gridded dataset at 0.5° × 0.5° latitudinal and longitudinal resolutions using a method of nested interpolation. First, the globe is divided into three horizontal resolutions of 0.5° × 0.5°, 4° × 4°, and 9° × 9°. The AMSU-A TB observations at both ascending and descending nodes within each grid box are averaged to form three gridded datasets. There is only a small amount of grid points near the equator that may not contain any observations at the 0.5° × 0.5° resolution. Using a polynomial interpolation method, the missing grid points at the 4° × 4° resolution will be obtained from the 9° × 9° resolution gridded dataset, and the missing grid points at the 0.5° × 0.5° resolution will be obtained from the 4° × 4° resolution gridded dataset. This daily global dataset of AMSU-A TB observations at the 0.5° × 0.5° resolution from December 1998 to November 2020 is used for the following study.
d. Subtracting annual cycle from satellite AMSU-A dataset
The ensemble empirical mode decomposition (EEMD) method is used to extract the TB data series from the highest to lowest frequencies (Huang and Wu 2008). It allows a decomposition of any data series into various components of decreasing frequencies without requiring a priori any fixed frequency basis functions which may introduce some artificial errors associated with the basis function specification. It uses extrema information of the riding waves in nonstationary time series and extracts successively the riding amplitude-frequency modulated oscillatory components from the highest to lowest frequencies, which incorporates naturally the adaptiveness and temporal locality into a data decomposition process. The EEMD method has already been widely applied in climate science, e.g., for studying climate trends in Ruzmaikin and Feynman (2009), Delgado et al. (2010), Breaker and Ruzmaikin (2011), Wu et al. (2011), and Qin et al. (2012), or for studying oscillations of certain periods in meteorological data in Ruzmaikin et al. (2007), Liang et al. (2008), Franzke (2009, 2010), Qian et al. (2011), Parey and Pachori (2012), and Lee and Ouarda (2010, 2011). A brief description of this method is provided below.
4. Global distributions and yearly variations of TB observations
a. Monthly mean TB distributions in 2008
Spatial distributions of the January climatology of TB observations at channels 6–10 in the Northern Hemisphere are provided in Fig. 5. The AMSU-A temperature-sounding channels 6–10 are located in the atmosphere from 400 to 50 hPa. The TB climatology is the average of TB observations in January months of the 22 years from 1998 to 2020. TB observations of channels 9 and 10 in the stratosphere are characterized by a circularly shaped cold TB center located around 80°N, 30°E, and a latitudinally distributed warm TB belt around 45°N with the its maximum located around 45°N, 165°E. Such a feature of the January TB climatology in the Northern Hemisphere is similar to that of the ERA5 air temperature at the peak weighting function level of the corresponding channel. The TB climatology (the ERA5 air temperature) is colder than 200 K (187 K) in high latitudes and warmer than 222 K (212 K) in middle latitudes. With a weighting function located between stratosphere and troposphere, channel-8 TB climatology displays a spatial distribution that resembles that of channel 9, but has a larger cold TB center in high latitudes (figure omitted). A dominating feature for the two AMSU-A channels 6 and 7 in the upper and middle troposphere is the two troughs of colder TBs observations located near the east coasts of Asia and the United States. The spatial distribution of the ERA5 air temperatures at 400 hPa matched quite well with that of channel-6 TB observations. The TB climatology decreased from more than 242 K (258 K) near the equator to about 220 K (224 K) in high latitudes and warmer than 222 K (212 K) in middle latitudes.
Northern Hemisphere distributions of January climatology of AMSU-A TB observations (color shading) at channels 6–7 and 9–10 and the ERA air temperatures (black curve) at 50, 100, 250, and 400 hPa.
Citation: Journal of Climate 36, 9; 10.1175/JCLI-D-22-0517.1
The climatology of TB observations in July shows a nearly zonally symmetric wavenumber-1 distributions of channels 9 and 10 in the stratosphere in the Northern Hemisphere (Fig. 6). The spatial distribution of TB observations at channels 6 and 7 are very different in high latitudes. The highest (lowest) TB observations appeared near the polar region for channel 7 (channel 6). We notice a zonal belt of high TBs around 35°N in the eastern half of the hemisphere. The ERA5 air temperature distributions at the peak weighting function levels of channels 9 and 10 match better with those of TB observations than those of channels 6 and 7. This is expected since TB observations in the troposphere are not only affected by atmospheric temperature, but also cloud and precipitation.
b. Yearly variations of zonal and monthly mean TB from 1999 to 2020
We show a yearly variation of the zonal and monthly mean TB observations at channel 6 in Fig. 7. The zonal and monthly mean TB profile in 1999, which is represented by a black curve, shows highest values of TB in the low latitudes, decreases slightly toward the North Pole, and decreases rapidly toward the South Pole. For clarity, the zonal and monthly mean TB profile in 1999 is subtracted from the zonal and monthly mean TBs from 1999 to 2020 in Fig. 7. In July (Fig. 7a), the observed TBs at channel 6 had experienced a general increasing in the tropics and middle latitudes in the SH since 1999. Specifically, TB observations near 58°S experienced a rapid increase of more than 0.8 K during from 1999 to 2003 and remained at similar magnitudes in the following years. TB observations in low latitudes between 30°S and 30°N had increased at a uniform speed for more than 0.6 K from 1999 to 2020. Compared with the zonal and monthly mean TBs in July 1999, the TBs in the NH middle latitudes were lower and higher before and after 2011, respectively. This explains why we see an overall negative global distribution of TB anomaly in 2008 (Fig. 6).
Temporal evolution of the monthly and latitudinal averaged TB observations at channel 6 in (a) July and (b) January from 1999 to 2020, in which the TB latitudinal profile in January 1999 (black solid curve) is subtracted. Also shown is the climatology profile (black dashed curve).
Citation: Journal of Climate 36, 9; 10.1175/JCLI-D-22-0517.1
The yearly variations of the zonal and monthly mean TB observations at channel 6 in January (Fig. 7b) are quite different from those in July (Fig. 7a). First, TB observations in the NH middle latitudes winter had increased since 1999 (Fig. 7b). Second, TB observations in high latitudes in the SH winter also experienced a rapid increase from 1999 to 2003 as in July, but with a larger magnitude of more than 1.4 K centered near 75°S. Third, the subtropical experienced a cooling before 2003. Fourth, the tropics had experienced a gradual warming since 2003 and merged with the warming zones in middle and high latitudes in both hemispheres by 2014. The zonal and monthly TB observations in January 2020 were nearly 1 K higher than those in January 1999 in the NH middle latitudes, the tropics, and the SH high latitudes.
5. Modulation effect of annual cycle in the Tibetan Plateau
Having examined the general features of TB variations, we focus on TB observations over the Tibetan Plateau. Figure 8 gives an example showing the temporal variation of TB observations for channel 6 at an arbitrarily selected data point (35°N, 87.5°E) over the Tibetan Plateau, the EEMD-extracted IMFs 1–7 (Fig. 8a) as well as the spectra for IMFs 1–7 (Fig. 8b). The peak spectral density values of IMFs 1–7 of TB observations all exceed the 95% significance level. A dominant feature in the TB data series is an annual cycle. The IMFs 1–4 basically represent the variations with periods less than 1.5 months. It is worth pointing out that an outlier that occurred on 27 December 2007 are extracted by using the first four IMFs. The fifth IMF describes mainly the seasonal variations with their periods of 2–3 months. The IMFs 1–5 represent 2–10-day synoptic scales, 25–35-day monthly scales, and 30–80-day intraseasonal scales. The sixth IMF captures mainly the half-year oscillations, and the seventh IMF represents mainly the annual cycle. We will take the sum of IMFs 6 and 7 as the annual cycle in the following discussions. The annual time scales are dominant frequencies and have similar magnitude as TB observations. This is a unique feature over the Tibetan Plateau (figures on other grid points are omitted). The potential causes could be associated with seasonal variations of latent heat release and precipitation over the Tibetan Plateau (Wu et al. 2007; Duan et al. 2013).
(a) TB observations (black) and the IMFs 1–7 (colored) as well as (b) their spectra for channel 6 at data point 35°N, 87.5°E for the time period from 25 Oct 1998 to 30 Nov 2020. Open circles are connected by solid line for IMFs 6 and 7. The same color convention is used for both (a) and (b). Dashed curves in (b) represent the 95% red noise significance level.
Citation: Journal of Climate 36, 9; 10.1175/JCLI-D-22-0517.1
As stated in section 2d, the EEMD method allows us to examine impact of the annual cycle on the climate trends. Figure 9 shows the spatial distributions of climate trends for channels 6–7 over and around the Tibetan Plateau with and without including the annual cycle. The left panels are the linear trends of original TB observations, and the right panels are the linear trends calculated from TB observations after removing the annual cycle. By comparing the trends with and without removing the annual cycle, we find that the annual cycle contributed greatly to the warming trends, especially over the Tibetan Plateau. Impacts of annual cycle on the climate trends over the Tibetan Plateau are much smaller in upper-level channels 8–13 (figures omitted) than low-level channels 6–7. The t-test results for detected linear trends suggests that the trends of channel-6 TB observations in the whole domain (Fig. 10a) and channel-7 TB observations only within a limited region in the middle of the Tibetan Plateau (Fig. 10c) pass the significant level. Once the annual cycle is removed, the detected trends of TB observations at both channels 6 and 7 pass the significant level (Figs. 10b,d).
Linear trends of TB observations at (a),(b) channel 6 and (c),(d) channel 7 (left) without and (right) with the annual cycle extracted. The 3-km terrain height is indicated by black dashed curve. The point 33.0°N 92.0°E is indicated by a black cross symbol in (c). Trends in areas marked with green dots in (c) are significant at the 95% confidence level, but areas with magenta dots in (b) and (d) are not significant. The trends in the entire domain in (a) are significant at the 95% confidence level.
Citation: Journal of Climate 36, 9; 10.1175/JCLI-D-22-0517.1
Sen’s trends [shading; K (10a)−1] of TB observations and the MK significant test score Z (contours) at (a),(b) channel 6 and (c),(d) channel 7 (left) before and (right) after removing the annual cycle. The trends within areas with the MK test score Z > 2 pass the 95% significant test. The 3-km terrain height is indicated by the black dashed curve. The point 33.0°N, 92.0°E is indicated by a black cross symbol in (a). Trends in areas marked with green dots in (c) are significant at the 95% confidence level, but areas with magenta dots in (b) and (d) are not significant. The trends in the entire domain in (a) are significant at the 95% confidence level.
Citation: Journal of Climate 36, 9; 10.1175/JCLI-D-22-0517.1
The spatial distributions of the Sen’s trends (Fig. 10). The patterns of the Sen’s trends are quite similar to those of the linear trends shown in Fig. 9, which suggest that the linear trends calculated from TB observations are not affected by gross errors. The MK test for trend analysis reveals that the trends of TB observations at channel 6 pass the significant level (Fig. 10a), but are not statistically significant for channel 7 except over an area in the middle of the Tibetan Plateau centered around 33.0°N, 92.0°E. After removing the annual cycle, the trends of TB observations at both channels 6 and 7 pass the significant level (Figs. 10b,d).
To see more clearly on how the annual cycle increased the warming trends, we may take the TB data series at the point 33.0°N, 92.0°E located on the Tibetan Plateau and examine the yearly variations of annual cycle (Fig. 11). By comparing with the climatology of the annual cycle, we see that the maximum and minimum values and thus the amplitudes of the annual cycle vary yearly. The annual cycle in TB observations of both channels 6 and 7 at this point experienced a warming trend from 1999 to 2020. For channel 6, decadal variations of annual cycle increased the warming trend by 0.38 K (10a)−1, which is 47% of the total warming trend of 0.81 K (10a)−1. Here, the value of 0.38 K (10a)−1 is the decadal trend calculated from the interannual change part (i.e., the sum of IMFs 6 and 7) extracted from TB observations. For channel 7 (Fig. 11b), the warming trend of annual cycle at the point 33.0°N, 92.0°E is 0.20 K (10a)−1, which is 0.05 K (10a)−1 higher than the warming trend of the TB observations at this point [i.e., 0.15 K (10a)−1, see Fig. 8c]. In other words, the warming trend of channel-7 TB observations over the Tibetan Plateau is mainly caused by the climate trend of annual cycle.
Yearly variation of annual cycle extracted from TB observations at the point 33.0°N, 92.0°E for (a) channel 6 and (b) channel 7, the linear trends of the annual cycle (black dashed line), and the annual cycle climatology (shaded in cyan) for the time period from 25 Oct 1998 to 30 Nov 2020.
Citation: Journal of Climate 36, 9; 10.1175/JCLI-D-22-0517.1
(left) Linear trends and (right) uncertainty of TB observations along the fixed longitude of (a),(b) 85°E and (c),(d) the fixed latitude of 35°N without (dash–dotted curve) and with (solid curve) removing high-frequency oscillations (i.e., IMFs 1–5). The terrain height at 5′ resolution is indicated in gray shading (the top x axis).
Citation: Journal of Climate 36, 9; 10.1175/JCLI-D-22-0517.1
Since the EEMD method also produce a trend that is defined as the TB data series subtracting all its IMFs (Qin et al. 2012). The warming and cooling trends in AMSU-A TB observations at channels 6–9 deduced by using the EEMD method are shown in Fig. 13. Instead of one value of the linear trend for each channel, the EEMD provides a time-varying trend for each channel. As expected, the EEMD trends in Fig. 13 are monotonic curves, with each curve having at most one extremum within the time span of these TB data series. We will call the EEMD deduced trends as nonlinear trends.
The nonlinear trends derived from daily TB observations of channels 6–9 at grid points (a) 33.0°N, 92.0°E and (b) 35.0°N, 87.5°E from 25 Oct 1998 to 30 Nov 2020.
Citation: Journal of Climate 36, 9; 10.1175/JCLI-D-22-0517.1
Because the annual cycle is eliminated in deducing the EEMD trends, the signs of nonlinear trends (Fig. 13) are consistent with the linear trends of TB observations with the annual cycle extracted (left panels in Fig. 9). The warming trends for the lower-level tropospheric channel 6 are quite nonlinear. They increased from 1999 to 2012, and decreased from 2012 to 2020. Further investigation is needed to explain the nonlinear evolution of channel-6 trends. Being less influenced by the annual cycle, the nonlinear trends of the tropospheric channel 8 at both grid points of 33.0°N, 92.0°E and 35.0°N, 87.5°E are positive and those of the stratospheric channel 9 are mostly negative, which are consistent with globally averaged trends in Qin et al. (2012).
6. Summary and conclusions
The Tibetan Plateau plays a key role in global climate change and is an amplification area of climate change. The climate change over the Tibetan Plateau has an important impact on those in its surrounding and even the global climate and ecological environment. However, due to largely insufficient conventional observations over the Tibetan Plateau, especially in its western region. This study demonstrates that satellite microwave temperature sounding observations are important data sources for studying decadal changes of the atmosphere over the Tibetan Plateau. Through a limb correction, TB observations at all fields of view from cross-track radiometers AMSU-A onboard multiple POESs can be used for examining geographic distributions of global warming/cooling trends. It is also found that the annual cycle over the Tibetan Plateau had a significant interdecadal variation over the past two decades, which contributed to an enhanced tropospheric warming trend over the Tibetan Plateau.
With a single instrument, AMSU-A, of satellite TB observations, it is difficult to know the causes of uneven warming over the Tibetan Plateau. The larger warming over the middle region of the Tibetan Plateau could be a combined effect of higher terrain and climate change of cloud and precipitation over this region. Further efforts in this regard will be made using the Special Sensor Microwave Imager/Sounder instrument (SSMIS) TB observations and cloud retrieval products (e.g., liquid water path) from the U.S. Air Force Defense Meteorological Satellite Program polar-orbiting satellites F16, F17, and F18. The SSMIS combines the Special Sensor Microwave Temperature sounder (SSMT), the Special Sensor Microwave/Water Vapor sounder (SSM/T2) and Special Sensor Microwave Imagers (SSM/I) into a single sensor so that thermally emitted radiation from Earth and the atmosphere can be simultaneously measured at all channels of SSMT, SSMT/2, and SSM/I. The SSMIS TB observations and retrieval products since 2005 to present are available to the general public. Unfortunately, SSMIS lower atmospheric sounding channels have data noise of complicated structures in F16 TB observations since 25 April 2013, and also contain significant across-track high-frequency striping noise spikes in F16 TB observations starting on 20 October 2017. Appropriate noise mitigation techniques were already developed and tested to work well for removing SSMIS LAS channels data noise (Dong and Zou 2022a,b). Given the convenience of SSMIS to include not only temperature sounding channels similar to AMSU-A, but also imager channels that are sensitive to cloud and precipitation, the potential causes of the annual time scales being the dominant frequencies may be further investigated with regards to cloud and precipitation over the Tibetan Plateau.
In future research, we will also explore the potential applications of this global satellite dataset, such as the impact of global warming on the structural and intensity changes of various weather systems, such as northeast China cold vortices and hurricanes, over the past two decades. Another interesting research topics is the mutual influences between low-frequency variability and climate change.
Acknowledgments.
This research was supported by the National Key R&D Program of China under the support of Grant 2018YFC1507302.
Data availability statement.
The data are available on request from the first author (Z. Qin; qzk_0@nuist.edu.cn).
REFERENCES
Bao, X., and F. Zhang, 2019: How accurate are modern atmospheric reanalysis for the data-sparse Tibetan Plateau region. J. Climate, 32, 7153–7172, https://doi.org/10.1175/JCLI-D-18-0705.1.
Bernath, P. F., 2017: The atmospheric chemistry experiment (ACE). J. Quant. Spectrosc. Radiat. Transfer, 186, 3–16, https://doi.org/10.1016/j.jqsrt.2016.04.006.
Breaker, L. C., and A. Ruzmaikin, 2011: The 154-year record of sea level at San Francisco: Extracting the long-term trend, recent changes, and other tidbits. Climate Dyn., 36, 545–559, https://doi.org/10.1007/s00382-010-0865-4.
Carminati, F., and S. Migliorini, 2021: All-sky data assimilation of MWTS-2 and MWHS-2 in the Met Office global NWP system. Adv. Atmos. Sci., 38, 1682–1694, https://doi.org/10.1007/s00376-021-1071-5.
Chang, E. K. M., and Y. Fu, 2002: Interdecadal variations in Northern Hemisphere winter storm track intensity. J. Climate, 15, 642–658, https://doi.org/10.1175/1520-0442(2002)015<0642:IVINHW>2.0.CO;2.
Christy, J. R., and W. B. Norris, 2004: What may we conclude about global tropospheric temperature trends? Geophys. Res. Lett., 31, L06211, https://doi.org/10.1029/2003GL019361.
Christy, J. R., R. W. Spencer, and E. S. Lobl, 1998: Analysis of the merging procedure for the MSU daily temperature time series. J. Climate, 11, 2016–2041, https://doi.org/10.1175/1520-0442-11.8.2016.
COESA, 1976: U.S. Standard Atmosphere, 1976. NOAA, 227 pp.
Delgado, J. M., H. Apel, and B. Merz, 2010: Flood trends and variability in Mekong River. Hydrol. Earth Syst. Sci., 14, 407–418, https://doi.org/10.5194/hess-14-407-2010.
Demaree, G. R., and C. Nicolis, 1990: Onset of Sahelian drought viewed as a fluctuation-induced transition. Quart. J. Roy. Meteor. Soc., 116, 221–238, https://doi.org/10.1002/qj.49711649110.
Dong, H., and X. Zou, 2022a: Mitigation of significant data noise in F17 SSMIS observations since October 2017. Remote Sens., 14, 1684, https://doi.org/10.3390/rs14071684.
Dong, H., and X. Zou, 2022b: Mitigation of systematic noise in F16 SSMIS LAS channels observations for tropical cyclone applications. Remote Sens., 14, 3171, https://doi.org/10.3390/rs14133171.
Duan, A., M. Wang, Y. Lei, and Y. Cui, 2013: Trends in summer rainfall over China associated with the Tibetan Plateau sensible heat source during 1980–2008. J. Climate, 26, 261–275, https://doi.org/10.1175/JCLI-D-11-00669.1.
Duan, A., Z. Xiao, and G. Wu, 2016: Characteristics of climate change over the Tibet Plateau under the global warming during 1979–2014. Climate Change Res., 12, 374–381, https://doi.org/10.12006/j.issn.1673-1719.2016.039.
Franzke, C., 2009: Multi-scale analysis of teleconnection indices: Climate noise and nonlinear trend analysis. Nonlinear Processes Geophys., 16, 65–76, https://doi.org/10.5194/npg-16-65-2009.
Franzke, C., 2010: Long-range dependence and climate noise characteristics of Antarctic temperature data. J. Climate, 23, 6074–6081, https://doi.org/10.1175/2010JCLI3654.1.
Fu, Q., C. M. Johanson, S. G. Warren, and D. J. Seidel, 2004: Contribution of stratospheric cooling to satellite-inferred tropospheric temperature trends. Nature, 429, 55–58, https://doi.org/10.1038/nature02524.
Goldberg, M. D., D. S. Crosby, and L. Zhou, 2001: The limb adjustment of AMSU-A observations: Methodology and validation. J. Appl. Meteor. Climatol., 40, 70–83, https://doi.org/10.1175/1520-0450(2001)040<0070:TLAOAA>2.0.CO;2.
Hauchecorne, A., and Coauthors, 2019: A new MesosphEO data set of temperature profiles from 35 to 85 km using Rayleigh scattering at limb from GOMOS/ENVISAT daytime observations. Atmos. Meas. Tech., 12, 749–761, https://doi.org/10.5194/amt-12-749-2019.
Hersbach, H., and Coauthors, 2020: The ERA5 global reanalysis. Quart. J. Roy. Meteor. Soc., 146, 1999–2049, https://doi.org/10.1002/qj.3803.
Huang, N. E., and Z. Wu, 2008: A review on Hilbert-Huang transform: Method and its applications to geophysical studies. Rev. Geophys., 46, RG2006, https://doi.org/10.1029/2007RG000228.
Huth, R., and M. Dubrovsky, 2021: Testing for trends on a regional scale: Beyond local significance. J. Climate, 34, 5349–5365, https://doi.org/10.1175/JCLI-D-19-0960.1.
Kalnay, E., and M. Cai, 2003: Impact of urbanization and land-use change on climate. Nature, 423, 528–531, https://doi.org/10.1038/nature01675.
Kendall, M. G., 1975: Rank Correlation Methods. 4th ed. Griffin, 202 pp.
Kroodsma, R. A., D. S. McKague, and C. S. Ruf, 2012: Inter-calibration of microwave radiometers using the vicarious cold calibration double difference method. IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens., 5, 1006–1013, https://doi.org/10.1109/JSTARS.2012.2195773.
Lee, T., and T. B. M. J. Ouarda, 2010: Long-term prediction of precipitation and hydrologic extremes with nonstationary oscillation processes. J. Geophys. Res., 115, D13107, https://doi.org/10.1029/2009JD012801.
Lee, T., and T. B. M. J. Ouarda, 2011: Prediction of climate nonstationary oscillation processes with empirical mode decomposition. J. Geophys. Res., 116, D06107, https://doi.org/10.1029/2010JD015142.
Lettenmaier, D. P., E. F. Wood, and J. R. Wallis, 1994: Hydro-climatological trends in the continental United States, 1948–88. J. Climate, 7, 586–607, https://doi.org/10.1175/1520-0442(1994)007<0586:HCTITC>2.0.CO;2.
Liang, M.-C., K.-F. Li, R.-L. Shia, and Y. L. Yung, 2008: Short-period solar cycle signals in the ionosphere observed by FORMOSAT-3/COSMIC. Geophys. Res. Lett., 35, L15818, https://doi.org/10.1029/2008GL034433.
Lins, H. F., and J. R. Slack, 1999: Streamflow trends in the United States. Geophys. Res. Lett., 26, 227–230, https://doi.org/10.1029/1998GL900291.
Liu, X., and Coauthors, 2009: Elevation dependency of recent and future minimum surface air temperature trends in the Tibet Plateau and its surroundings. Global Planet. Change, 68, 164–174, https://doi.org/10.1016/j.gloplacha.2009.03.017.
Lu, N., K. E. Trenberth, J. Qin, K. Yang, and L. Yao, 2015: Detecting long-term trends in precipitable water over the Tibet Plateau by synthesis of station and MODIS observations. J. Climate, 28, 1707–1722, https://doi.org/10.1175/JCLI-D-14-00303.1.
Mann, H. B., 1945: Nonparametric tests against trend. Econometrica, 13, 245–259, https://doi.org/10.2307/1907187.
Mears, C. A., and F. J. Wentz, 2005: The effect of diurnal correction on satellite-derived lower tropospheric temperature. Science, 309, 1548–1551, https://doi.org/10.1126/science.1114772.
Mears, C. A., and F. J. Wentz, 2009: Construction of the remote sensing systems V3.2 atmospheric temperature records from the MSU and AMSU microwave sounders. J. Atmos. Oceanic Technol., 26, 1040–1056, https://doi.org/10.1175/2008JTECHA1176.1.
Mears, C. A., M. C. Schabel, F. J. Wentz, B. D. Santer, and B. Govindasamy, 2002: Correcting the MSU middle tropospheric temperature for diurnal drifts. 2002 IEEE Int. Geoscience and Remote Sensing Symp., Toronto, Canada, Institute of Electrical and Electronics Engineers, 1839–1841, https://doi.org/10.1109/IGARSS.2002.1026272.
Mears, C. A., M. C. Schabel, and F. J. Wentz, 2003: A reanalysis of the MSU channel 2 tropospheric temperature record. J. Climate, 16, 3650–3664, https://doi.org/10.1175/1520-0442(2003)016<3650:AROTMC>2.0.CO;2.
Mo, T., 2007: Postlaunch calibration of the NOAA-18 advanced microwave sounding Unit-A. IEEE Trans. Geosci. Remote Sens., 45, 1928–1937, https://doi.org/10.1109/TGRS.2007.897451.
Mo, T., 2009: A study of the NOAA-15 AMSU-A brightness temperatures from 1998 through 2007. J. Geophys. Res., 114, D11110, https://doi.org/10.1029/2008JD011267.
Nash, J., and R. Saunders, 2015: A review of stratospheric sounding unit radiance observations for climate trends and reanalyses. Quart. J. Roy. Meteor. Soc., 141, 2103–2113, https://doi.org/10.1002/qj.2505.
Nguyen, P., and Coauthors, 2018: Global precipitation trends across spatial scales using satellite observations. Bull. Amer. Meteor. Soc., 99, 689–697, https://doi.org/10.1175/BAMS-D-17-0065.1.
Niu, Z., and X. Zou, 2022: Comparison among all-sky simulations, FY-3E MWTS-3 and FY-4A AGRI observations of the first Typhoon Malakas in 2022. Earth Space Sci., 9, e2022EA002498, https://doi.org/10.1029/2022EA002498.
Ohring, G., B. Wielicki, R. Spencer, B. Emery, and R. Datla, 2005: Satellite instrument calibration for measuring global climate change: Report of a workshop. Bull. Amer. Meteor. Soc., 86, 1303–1314, https://doi.org/10.1175/BAMS-86-9-1303.
Parey, A., and R. B. Pachori, 2012: Variable cosine windowing of intrinsic mode functions: Application to gear fault diagnosis. Measurement, 45, 415–426, https://doi.org/10.1016/j.measurement.2011.11.001.
Po-Chedley, S., T. J. Thorsen, and Q. Fu, 2015: Removing diurnal cycle contamination in satellite-derived tropospheric temperatures: Understanding tropical tropospheric trend discrepancies. J. Climate, 28, 2274–2290, https://doi.org/10.1175/JCLI-D-13-00767.1.
Qian, C., C. Fu, and Z. Wu, 2011: Changes in the amplitude of the temperature annual cycle in China and their implication for climate change research. J. Climate, 24, 5292–5302, https://doi.org/10.1175/JCLI-D-11-00006.1.
Qin, Z., X. Zou, and F. Weng, 2012: Comparison between linear and nonlinear trends in NOAA-15 AMSU-A brightness temperatures during 1998–2010. Climate Dyn., 39, 1763–1779, https://doi.org/10.1007/s00382-012-1296-1.
Ruzmaikin, A., and J. Feynman, 2009: Search for climate trends in satellite data. Adv. Data Sci. Adapt. Anal., 1, 667–679, https://doi.org/10.1142/S1793536909000266.
Ruzmaikin, A., M. L. Santee, M. J. Schwartz, L. Froidevaux, and H. M. Pickett, 2007: The 27-day variations in stratospheric ozone and temperature: New MLS data. Geophys. Res. Lett., 34, L02819, https://doi.org/10.1029/2006GL028419.
Santer, B. D., and Coauthors, 2017: Causes of differences in model and satellite tropospheric warming rates. Nat. Geosci., 10, 478–485, https://doi.org/10.1038/ngeo2973.
Saunders, R., and Coauthors, 2018: An update on the RTTOV fast radiative transfer model (currently at version 12). Geosci. Model Dev., 11, 2717–2737, https://doi.org/10.5194/gmd-11-2717-2018.
Sen, P. K., 1968: Estimates of the regression coefficient based on Kendall’s Tau. J. Amer. Stat. Assoc., 63, 1379–1389, https://doi.org/10.1080/01621459.1968.10480934.
Sofieva, V. F., F. Dalaudier, A. Hauchecorne, and V. Kan, 2019: High-resolution temperature profiles retrieved from bichromatic stellar scintillation measurements by GOMOS/Envisat. Atmos. Meas. Tech., 12, 585–598, https://doi.org/10.5194/amt-12-585-2019.
Steiner, A. K., and Coauthors, 2020: Observed temperature changes in the troposphere and stratosphere from 1979 to 2018. J. Climate, 33, 8165–8194, https://doi.org/10.1175/JCLI-D-19-0998.1.
Tian, X., X. Zou, and S. Yang, 2018: A limb correction method for the microwave temperature sounder 2 and its applications. Adv. Atmos. Sci., 35, 1547–1552, https://doi.org/10.1007/s00376-018-8092-8.
Vinnikov, K. Y., and N. C. Grody, 2003: Global warming trend of meantropospheric temperature observed by satellites. Science, 302, 269–272, https://doi.org/10.1126/science.1087910.
Wang, L., C.-Z. Zou, and H. Qian, 2012: Construction of stratospheric temperature data records from stratospheric sounding units. J. Climate, 25, 2931–2946, https://doi.org/10.1175/JCLI-D-11-00350.1.
Wang, W., and C.-Z. Zou, 2014: AMSU-A-only atmospheric temperature data records from the lower troposphere to the top of the stratosphere. J. Atmos. Oceanic Technol., 31, 808–825, https://doi.org/10.1175/JTECH-D-13-00134.1.
Wu, G., J. Mao, and A. Duan, 2004: Recent progress in the study on the impacts of Tibetan Plateau on Asian summer climate. Acta Meteor. Sin., 62, 528–540.
Wu, G., and Coauthors, 2007: The influence of mechanical and thermal forcing by the Tibetan Plateau on Asia climate. J. Hydrometeor., 8, 770–789, https://doi.org/10.1175/JHM609.1.
Wu, G., A. Duan, X. Zhang, Y. Liu, Y. Ma, and K. Yang, 2013: Extreme weather and climate changes and its environmental effects over the Tibetan Plateau (in Chinese). Chin. J. Nature, 35, 167–171.
Wu, Z., and N. E. Huang, 2009: Ensemble empirical mode decomposition: A noise-assisted data analysis method. Adv. Adapt. Data Anal., 1, 1–41, https://doi.org/10.1142/S1793536909000047.
Wu, Z., N. E. Huang, J. M. Wallace, B. V. Smoliak, and X. Chen, 2011: On the time- varying trend in global-mean surface temperature. Climate Dyn., 37, 759, https://doi.org/10.1007/s00382-011-1128-8.
Xia, X., and X. Zou, 2021: Combining FY-3D MWTS-2 with AMSU-A data for inter-decadal diurnal correction and climate trends of atmospheric temperature. Remote Sens., 13, 3148, https://doi.org/10.3390/rs13163148.
Xu, X., L. Dong, Y. Zhao, and Y. Wang, 2019: Effect of the Asian water tower over the Qinghai-Tibet Plateau and the characteristics of atmospheric water circulation. Chin. Sci. Bull., 64, 2830–2841, https://doi.org/10.1360/TB-2019-0203.
Yang, K., X. Guo, J. He, J. Qin, and T. Koike, 2011: On the climatology and trend of the atmospheric heat source over the Tibetan Plateau: An experiment supported revisit. J. Climate, 24, 1525–1541, https://doi.org/10.1175/2010JCLI3848.1
Yang, X., Y. Hou, and B. Chen, 2011: Observed surface warming induced by urbanization in east China. J. Geophys. Res., 116, D14113, https://doi.org/10.1029/2010JD015452.
Yue, S., P. Pilon, and G. Cavadias, 2002: Power of the Mann–Kendall and Spearman’s rho tests for detecting monotonic trends in hydrological series. J. Hydrol., 259, 254–271, https://doi.org/10.1016/S0022-1694(01)00594-7.
Zhang, R., and S. Zhou, 2008: The air temperature change over the Tibetan Plateau during 1979–2002 and its possible linkage with ozone depletion. Acta Meteor. Sin., 6, 916–925.
Zhao, P., and Coauthors, 2019: The Tibetan Plateau surface–atmosphere coupling system and its weather and climate effects: The third Tibetan Plateau atmospheric science experiment. J. Meteor. Res., 33, 375–399, https://doi.org/10.1007/s13351-019-8602-3.
Zhou, L., and Coauthors, 2004: Evidence for a significant urbanization effect on climate in China. Proc. Natl. Acad. Sci. USA, 101, 9540–9544, https://doi.org/10.1073/pnas.0400357101.
Zou, C.-Z., and W. Wang, 2011: Intersatellite calibration of AMSU-A observations for weather and climate applications. J. Geophys. Res., 116, D23113, https://doi.org/10.1029/2011JD016205.
Zou, C.-Z., H. Qian, W. Wang, L. Wang, and C. Long, 2014: Recalibration and merging of SSU observations for stratospheric temperature trend studies. J. Geophys. Res. Atmos., 119, 13 180–13 205, https://doi.org/10.1002/2014JD021603.
Zou, X., 2012: Climate trend detection and its sensitivity to measurement precision. Adv. Meteor. Sci. Tech., 2, 41–43.