Next Article in Journal
Deep Learning for Flash Drought Detection: A Case Study in Northeastern Brazil
Previous Article in Journal
Endotoxins in Household Dust in Northern China: Associations with Home Environment Factors and Childhood Asthma and Allergies
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Atmosphere
Volume 15
Issue 7
10.3390/atmos15070760
atmosphere-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Research on Validation Method on Retrieval of Atmospheric Temperature and Humidity Profile Using a Microwave Sounder

by
Qiurui He
1,2,
Jiaoyang Li
3,*,
Ruiling Zhang
1,
Junqi Jia
1 and
Xiao Guo
4
1
School of Information Technology, Luoyang Normal University, Luoyang 471934, China
2
State Key Laboratory of Geo-Information Engineering, Xi’an 710054, China
3
School of Mechatronic Engineering and Automation, Shanghai University, Shanghai 200444, China
4
Hebi Institute of Engineering and Technology, Henan Polytechnic University, Hebi 458030, China
*
Author to whom correspondence should be addressed.
Atmosphere 2024, 15(7), 760; https://doi.org/10.3390/atmos15070760
Submission received: 18 April 2024 / Revised: 16 June 2024 / Accepted: 21 June 2024 / Published: 26 June 2024
(This article belongs to the Section Atmospheric Techniques, Instruments, and Modeling)
Browse Figures
Versions Notes

Abstract

:
The commonly used reference atmospheric profiles for the validation of retrieved atmospheric profiles for microwave sounders have bias compared with real atmospheric profile values, which is detrimental to the validation of the retrieval. Microwave sounder observations are the direct measurements of microwave radiation in atmospheric conditions and are a true representation of the status of the atmosphere. This paper proposed a validation method for the retrieved atmospheric temperature and atmospheric humidity profiles of the satellite-based microwave sounder using its own in-orbit observations. The validation experiments are performed both for the retrievals of the microwave temperature sounder-II (Xi’an Branch, China Academy of Space Technology, Xi’an, China. MWTS-II) and the microwave humidity and temperature sounder (National Space Science Center, Chinese Academy of Sciences, Beijing, China. MWHTS). The validation results show that the retrieved temperature profiles of MWTS-II have higher accuracy compared to the temperature profiles of ERA5 in the atmospheric pressure range of 3–30 hPa, and the accuracy of the rest of the pressure range is comparable between the profiles of ERA5 and the retrieved profiles. And the retrieved temperature profiles of MWHTS have higher accuracy compared to the temperature profiles of ERA5 in the atmospheric pressure level around 50 hPa and lower accuracy in the rest of the pressure levels. In addition, the retrieved humidity profiles of MWHTS have higher accuracy compared to the humidity profiles of ERA5 in the atmospheric pressure range of 350–925 hPa. The proposed validation method for the retrieved atmospheric temperature and atmospheric humidity profiles of MWHTS using its own observations is promising for improving the feasibility and reliability of the validation, and can be a good reference for the application of the satellite in-orbit observations and the optimization of the microwave sounders.

1. Introduction

Satellite-based microwave sounder is a key payload that can obtain the global distribution of atmospheric humidity and temperature profiles by using an inversion algorithm to convert the satellite-based observations to the temperature and humidity profiles [1]. The atmospheric humidity and temperature profiles retrieved from the microwave sounders are essential atmospheric parameters for the research of climate change, numerical weather prediction, meteorological disaster monitoring and forecasting, and other atmospheric applications [2,3,4]. The retrieval of high-precision temperature and humidity profiles based on passive microwave observations has always been a research hotspot in many meteorological and climate research groups [5,6,7], and the accuracy validation of the retrieved temperature and humidity profiles has attracted considerable attention from researchers and users in the field of atmospheric science.
The retrieval accuracy of the atmospheric profiles using the data of a satellite-based microwave sounder can be affected by many types of factors, such as the data quality of the microwave sounder, the retrieval algorithm, and the seasonal factors related to solar radiation [8,9]. However, the validation method is the key point to validate the accuracy and reliability of the retrieved profile. The atmospheric humidity and temperature profiles retrieved by microwave radiometers need to be verified with the value of real profiles. However, the true value of the global atmospheric humidity and temperature profiles is insufficient. In general, the validation of the retrieved profiles is performed by the comparison between the retrieved profiles and other reference profile data such as the reanalysis dataset produced by the assimilation system (e.g., the Reanalysis dataset from the European Center for Medium-Range Weather Forecasts (ECMWF) and the Reanalysis dataset from the National Center for Environmental Prediction (NCEP)), the direct measurement of the atmospheric humidity and temperature profiles detected by the radiosonde, and the retrieved profiles of other instruments that are operated at other frequencies [10,11,12]. Typically, the bias between the retrieved profiles and the reference profile data as well as the root mean square error (RMSE) are key indices to evaluate the retrieval results of the atmospheric humidity and temperature profiles. However, the reference data are also biased against the truth value of the real atmosphere, which may cause an unwanted bias for the validation of the retrieved profiles using reference data.
The reanalysis dataset, which is generated by the assimilation system with input data from multi-observation frequency and methods, is difficult to represent the real atmospheric parameters due to the unremovable errors of the assimilation system. The radiosonde is a balloon-borne instrument working in the atmosphere that sounds the atmospheric parameters including pressure, temperature, moisture, wind, etc. However, the radiosonde observations (RAOB) can have a large bias compared to real atmospheric data or are even invalid due to the systematic errors of the instruments, the drift of the balloon with the wind and other adverse weather conditions. In addition, there is no available RAOB in desert and marine areas that are inaccessible to humans. As for the drawback of setting the retrieved atmospheric temperature and humidity profiles using other instruments such as microwave and infrared sounders as the reference data for the validation, the error of the instruments and the retrieval algorithm used for obtaining these profiles can lead to the bias of these profiles compared to the real profile values. Therefore, the reliability of the validation results can be doubtful when the retrieved atmospheric profiles are verified using the above-mentioned reference data that deviate from the real atmospheric state. For instance, when using the reanalysis data to verify the retrieved atmospheric temperature and humidity profile of a specific pressure layer, the level of the bias between these two types of temperature and humidity profiles can be calculated, while it is hard to determine whether the reanalysis data or the inversion data are closer to the real temperature and humidity profiles.
To obtain more reliable validation results for the retrieved atmospheric temperature and humidity profiles observed by the microwave sounders, the selection of the reference data is of great importance. The microwave sounders sound the atmospheric parameters by measuring the radiation of atmospheric parameters during the emission, absorption, and scattering processes in the atmosphere. Therefore, the observations of microwave sounders reveal the real conditions of the Earth’s atmosphere, although it includes uncertainty caused by the instrument’s systematic, imperfect calibration, unfavorable atmospheric conditions, etc. Based on the literature review, the validation method of the retrieved atmospheric profile typically uses the reanalysis data, RAOB, or the retrieval results of other sounders working at different frequencies, or the combination of the above-mentioned multiple types of data, to improve the reliability of the validation results [7,13,14]. However, there are no reports on using microwave sounder observations as reference data to verify the retrieved results.
This study proposes to use the observation of microwave sounders as the reference data to perform the validation of the retrieved atmospheric profiles of microwave sounders for the purpose of providing more verification results for the application of microwave sounder retrieval of temperature and humidity profiles. First, the microwave observations and the atmospheric temperature and humidity profiles of the reanalysis dataset are fed into the statistical retrieval algorithm to retrieve the atmospheric temperature and humidity profiles. Then, the retrieved atmospheric profiles and the collocated profiles of the reanalysis dataset are fed into the radiative transfer model to calculate the simulated brightness temperatures, respectively, and these two types of simulated brightness temperatures are compared with the observed brightness temperatures of microwave sounder. Finally, the accuracy of these two types of simulated brightness temperatures and the weight function distribution characteristics of the microwave sounder channels are studied to verify whether the atmospheric humidity and temperature profiles retrieved by the microwave sounder or obtained from the reanalysis dataset are closer to the real data in a specific atmospheric pressure range.
The Fengyun-3D satellite is one of China’s second generation of polar-orbiting operational meteorological series satellites, which can achieve 3D observation of the atmosphere and obtain global and quantitative atmospheric and surface parameters under all-weather conditions [13]. The Fengyun-3D satellite has played an important role in global environment monitoring, disaster monitoring, and climate assessment. microwave temperature sounder-II (MWTS-II), which is used for the detection of atmospheric temperature profiles [14], and microwave humidity and temperature sounder (MWHTS), which is used for the simultaneous detection of atmospheric humidity and temperature profiles [15], are two primary passive microwave radiometers onboard Fengyun-3D satellite. In this paper, the validation tests are performed for the atmospheric profiles retrieved from the MWTS-II and MWHTS observations, respectively. The rest of this paper is organized as follows. Section 2 introduces the data and the preparation method used for this study. Section 3 presents the retrieval algorithm of the atmospheric temperature and humidity profiles, which is followed by the proposed validation method of the retrieved atmospheric profiles in Section 4. Section 5 provides the test results, and the conclusion is summarized in Section 6.

2. Data and Model

2.1. Microwave Sounder

MWTS-II and MWHTS are both total power microwave radiometers operated under the cross-track scanning mode to receive microwave radiation of the atmosphere. MWTS-II has a total of 13 detection channels operated in the 50–60 GHz oxygen absorption band for the detection of atmospheric temperature profiles. MWTS-II has a scan angle of ±49.5° from the nadir and a total of 90 field-of-views in each scan line with a nadir resolution of 33 KM [16,17]. MWHTS sets up eight detection channels near the oxygen absorption line at 118.75 GHz, five detection channels near the water vapor absorption line at 183.35 GHz, and two detection channels in the atmospheric transparent windows at 89 GHz and 150 GHz, which allows MWHTS to achieve the simultaneous observation of atmospheric temperature and humidity profile and the parameters of cloud, rain, and surface. MWTS-II has a scan angle of ±53.35° from the nadir and a total of 98 fields of view in each scan line with a nadir resolution of 16 KM [18,19,20].
The characteristics of each detection channel of MWTS-II and MWHTS can be described by the weighting function of each channel. The weighting function is the theoretical basis for the design and monitoring of the characteristics of the detection channel, and the indicators of the sensitivity of each channel of the microwave sounder to different atmospheric layers. The atmospheric layer in which the peak of the channel weighting function is distributed indicates that the channel is most sensitive to that layer of the atmosphere; in other words, the majority of the atmospheric radiation measured by the channel comes from that atmospheric layer [21,22,23]. The channel weighting function of MWTS-II and MWHTS at the nadir view is calculated by inputting the U.S. standard atmosphere [24] to the radiative transfer for television and infrared observation satellite operational vertical sounder (RTTOV), in which the surface emissivity is set to 0.6, as shown in Figure 1.
As shown in Figure 1, the distribution of the peak of the weighting function of each channel indicates the range of atmospheric pressure mainly detected by each channel of MWTS-II and MWHTS. All 13 channels of MWTS-II and channels 2–9 of MWHTS have a non-uniform distribution of peak weight functions throughout the atmosphere, which enables the detection of atmospheric temperature between the surface and the top of the atmosphere. Channels 11–15 of MWHTS are developed to detect the water vapor in the range of 800–300 hPa, and window channels 1 and 10 are developed to detect the surface information according to their peak weighting function distribution. It is noted that the distribution characteristics of the channel weighting function are closely related to the atmospheric condition, e.g., the existence of clouds and rain could lead to the change in the shape of the distribution of the weighting function and the peak value of the weight function.

2.2. Atmospheric Data

The atmospheric data used in this study include temperature profile, humidity profile, cloud liquid water content profile, cloud ice water content profile, rainwater content profile, snow water content profile, surface temperature, surface humidity, surface pressure, and 10 m wind component. The above-mentioned atmospheric and surface data are named as the reference atmospheric (RA) data in this paper. The RA data can be used to calculate the weighting function and the simulated brightness temperature of the microwave sounder by using the radiative transfer model. And the atmospheric humidity and temperature profiles in the RA data, which can be used to build the statical retrieval model for humidity and temperature profiles, can also be used to assist in the validation of the retrieval accuracy of the retrieved profiles. The ERA5 reanalysis dataset of ECMWF [25,26,27], which is generated by assimilating various data sources, such as ground-based observations, RAOB, and satellite-based observations, is selected for this study as the source of RA data. ERA5 provides hourly estimated surface parameters with a resolution of 0.25° and the hourly estimated atmospheric parameters with 37 levels in the range of 1000–1 hPa. It is noted that the observations of MWTS-II and MWHTS have not been assimilated in the assimilation system of ERA5, so the observations of MWTS-II and MWHTS and ERA5 reanalysis data are independent of each other.

2.3. Data Preparation

This study uses the observed brightness temperature of MWTS-II and MWTHS onboard the FY-3D satellite over the ocean region (25° N–45° N and 160° E–200° E) from January 2020 to February 2021 to perform the retrieval study of the atmospheric temperature and humidity profiles. The development and validation of the retrieval algorithm and the simulation of the microwave sounder all need an effective collocation between the RA data and the observations of MWTS-II and MWHTS. The criteria of the collocation is that the time difference and absolute distance between the microwave sounder observations and the RA data should be less than 5 min and 10 KM. In addition, the RA data are input to the RTTOV to calculate MWTS-II and MWHTS’ simulated brightness temperatures which is named as the reference simulated brightness temperature (RSBT). Currently, RTTOV has been widely applied in diverse retrieval and assimilation systems. The observations of the passive visible, infrared and microwave-based downward-looking satellite-based sounders can be simulated using RTTOV by using known atmospheric status and observation geometry [28,29,30]. The data collocation is performed among the observations of satellite-based microwave sounders, the RA data and the RSBT to establish the collocation dataset. The collocated data samples in the collocation dataset from January 2020 to December 2020 are selected as the analysis dataset for the development of the retrieval algorithm, while the remaining data samples are selected as the validation dataset for the retrieval and the assisted validation of atmospheric temperature and humidity profiles.
According to the aforementioned data preparation procedure, an analysis dataset for MWTS-II with 681,844 samples and a validation dataset for MWTS-II with 113,833 samples are obtained by the collocation of the observations of RA data and MWTS-II data. Similarly, an analysis dataset for MWHTS with 1,101,240 samples and a validation dataset for MWHTS with 177,199 samples are obtained by the collocation of the observations of RA data and MWHTS data.

3. Retrieval Algorithm

In general, the retrieval algorithms for microwave remote sensing consist the physics-based and statistics-based retrievals. Physics-based retrieval is the direct inversion study of the radiative transfer equation, which has the advantages of clear physical principle and high inversion accuracy and is the fundamental way to improve the accuracy of retrieved atmospheric parameters by satellite-based microwave sounders [31,32,33]. The statistics-based retrieval uses the principle of data statistics and does not involve any physical modeling, which makes the statics-based retrieval method have the advantages of low computational effort and high computational efficiency, and easy to build [34,35,36]. However, it should be noted that the statistical inversion method has a strong dependence on the historical data. The low representativeness of historical data or insufficient data volume will lead to poor retrieval accuracy of the statistical retrieval method or even failure in retrieval [1]. As statistical tools develop by leaps and bounds, the employment of the statistics-based retrieval algorithm in microwave remote sensing has shown great potential and has been widely used for the retrieval of atmospheric profiles using microwave sounders.
Considering that the physical retrieval method needs to use the simulated brightness temperature calculated by the radiation transfer model to fit the observed brightness temperature of the satellite-based microwave sounder, and the verification method we proposed here also needs to use the radiation transfer model to calculate the simulated brightness temperature; to avoid the error transmission caused by the radiation transfer model, the statistical retrieval method is selected. The neural network algorithm is representative of the statistical inversion method. He et al. conducted a comparative study on the performance of the shallow neural network and deep neural network (DNN) in the retrieval of atmospheric parameters by a satellite-borne microwave sounder and verified that the DNN can obtain higher retrieval accuracy and higher stability in retrieval using microwave sounder data [1]. In this study, the DNN, which has outstanding performance in nonlinear modeling, is studied and developed as a statistics-based retrieval model for the inversion of atmospheric profiles using microwave sounder observations.
Taking the retrieval study of atmospheric temperature and humidity profiles using MWHTS observations as an example, the observed brightness temperature in the analysis dataset for MWHTS built in Section 2 is set as the input for training the DNN-based retrieval model, while the outputs of the DNN model are the corresponding atmospheric temperature and humidity profiles. For testing the trained DNN-based retrieval model, the observed brightness temperature in the validation dataset for MWHTS is used, and retrieved atmospheric temperature and humidity profiles can be obtained. The RMSE between the temperature and humidity profiles in the analysis dataset for MWHTS and the DNN’s predictions of the temperature and humidity profiles during training is used as an evaluation criterion to optimize and configure the parameters of the DNN during the training of the DNN. The input layer of DNN used in the DNN-based retrieval model has 15 neurons, which receive the observations of 15 channels of MWHTS, and the output layer has 74 neurons, which output the temperature and humidity profiles corresponding to the observations in the input layer. There is one hidden layer in the DNN and 1024 neurons in the hidden layer of the DNN. This configuration is selected after extensive testing with smaller and larger numbers of neurons. In the extensive testing, the trained DNN produces the best result in terms of reproducing the temperature and humidity profiles, which is evaluated by the RMSE between the predictions of the temperature and humidity profiles and the temperature and humidity profiles in the testing dataset. The rectified linear unit (ReLU) is selected as the activation function. RMSprop is selected as the optimizer, which is an adaptive learning rate algorithm for gradient-based optimization, and the learning rate is set to be 0.001. A sufficiently large number of epochs and an early stop method are used during the training of the DNN model to avoid underfitting and overfitting, respectively, in which 20% of the testing samples are used for the monitoring of RMSE in the early stop method. The training will be stopped if the RMSE is stable within 150 adjacent epochs. The flow chart of the DNN-based retrieval algorithm for the retrieval of atmospheric temperature and humidity profiles using MWHTS observations is shown in Figure 2.
Similarly, the DNN-based retrieval model for MWTS-II can be developed using the observed brightness temperature and temperature profiles in the analysis dataset for MWTS-II built in Section 2. The settings of the DNN model for MWTS-II are nearly the same as those of the DNN model for MWHTS, except the number of neurons of the input layer, hidden layer, and output layer of the DNN model of MWTS-II, which is set to be 13, 512 and 37. The atmospheric temperature profiles can be retrieved when the observed brightness temperature in the validation dataset for MWTS-II is fed into the DNN retrieval model for MWTS-II.

4. Validation of the Retrieval

4.1. Feasibility Analysis

The commonly used reference data, such as reanalysis data, RAOB, and retrieved temperature and humidity profiles using other sounders’ observations, have bias from the real atmospheric data [37,38]. Therefore, the validation of the retrieved atmospheric profiles of microwave sounders using these reference data cannot be precise. However, it is hard to obtain the real data of historical global high-resolution atmospheric parameters which can meet the requirements for the validation of the retrieved atmospheric parameters using microwave remote sensing. Therefore, it is more reliable and reasonable to perform cross-validation using multi-types of reference data simultaneously or induce new types of reference data sources for the validation of the retrievals.
For the satellite-borne microwave sounder, the observations in the 60 GHz and 118 GHz bands are mainly generated by the microwave radiation of the actual atmospheric temperature profile, and the observations in the 183 GHz band are mainly generated by the microwave radiation of the actual water vapor parameter. The observations in these three bands can indicate the real distribution characteristics of the atmospheric temperature and humidity parameters. Since satellite-based passive microwave observations indicate the actual atmospheric status in the microwave sensing band, microwave sounder observations can be used to evaluate the retrieval results of atmospheric profiles of the microwave sounder. In this study, the observations of microwave sounders are proposed to be used for the validation of the retrieved profiles. First, the RA data are input to the radiative transfer model to calculate the RSBT of the microwave sounder. Then, the retrieval values of the atmospheric profiles substitute the profile value in the RA data and this new set of RA data is input to the radiative transfer model to calculate the new set of RSBT, which is named the retrieved RSBT. Finally, the RSBT and retrieved RSBT are compared with the observed brightness temperature of the microwave sounder to decide which types of the above-mentioned atmospheric profiles used to calculate the brightness temperature are closer to the real atmospheric profile.
It is noted that there are always biases between the simulated and observed brightness temperatures. These biases mainly include the error of the sounder, the bias caused by the severe atmospheric environment, the uncertainty induced by the imperfect calibration, the error caused by the imperfect simulation model, the bias of the atmospheric parameters used in the calculation of the simulation model, etc. [39,40,41,42]. And these biases will influence the accuracy of the comparison of the RSBT and retrieved RSBT, which in turn affects the judgment of whether the retrieved or the reference profiles are closer to the real atmospheric profiles. Thus, the influence of these biases is the key to the feasibility analysis of the proposed validation method for retrieved atmospheric profiles using microwave observations. Here, these biases are classified into three categories: errors related to the sounder, errors related to the radiative transfer model, and errors related to the atmospheric parameters that are used in the calculation of the radiative transfer model.
The RSBT and retrieved RSBT are calculated using the same radiative transfer model in this study, so the errors related to the radiative transfer model for them are the same. In addition, the RSBT and retrieved RSBT are obtained for the same microwave sounder and compared with the same observations of the same microwave sounder, so the errors related to the sounder are the same. It can be concluded that the only difference between RSBT and retrieved RSBT is caused by the difference in the atmospheric parameters that are used in the calculation of the radiative transfer model for obtaining the RSBT and retrieved RSBT. Therefore, the conclusion of whether the RSBT or the retrieved RSBT is closer to the observed brightness temperature can be drawn if the difference between the observed brightness temperature and the RSBT or retrieved RSBT is smaller.
Based on the distribution of the channels’ weighting functions shown in Figure 1, each channel can detect a specific range of atmospheric pressures. Therefore, the comparison results of the RSBT and retrieved RSBT can not only reveal whether the profiles in the RA data or the retrieved profiles are closer to the real atmospheric profiles but also achieve the verification of the retrieval results in the specific atmospheric pressure range. To be specific, if a channel’s retrieved RSBT has higher accuracy compared to the corresponding RSBT in its detection pressure range, the retrieved atmospheric profiles of this channel of the microwave sounder are closer to the real atmospheric profiles compared to the reference profiles in the RA data. Therefore, the validation method proposed in this study includes two parts. One part is the comparison of the retrieved RSBT and the RSBT, which use the retrieved profiles using microwave observations and the profiles in the reference data to calculate the simulated brightness temperature, respectively. And the other part is the validation of the retrieved results according to the pressure range mainly detected by each channel of the microwave detector.

4.2. Comparison of the Retrieved RSBT and the RSBT

Taking MWHTS as an example, the comparison process of the retrieved RSBT and the RSBT is as follows. First, as described in Section 3, the observed brightness temperature in the validation dataset for MWHTS is fed into the retrieval model for MWHTS to obtain the retrieved atmospheric temperature and humidity profiles. Then, the retrieved atmospheric profiles substitute the profiles in the RA dataset and the retrieved atmospheric dataset is established and fed into the RTTOV to calculate the retrieved RSBT. Finally, the RMSE between the retrieved RSBT and observed brightness temperature of each channel of MWHTS (i.e., retrieval-based RMSE), as well as the RMSE between the RSBT and the observed brightness temperature of each channel (i.e., reference RMSE) are calculated and compared. As a result, whether the RSBT or the retrieved RSBT is closer to the observed brightness temperature can be concluded. The comparison process of the retrieved RSBT and the RSBT is shown in Figure 3.
The comparison process of the retrieved RSBT and the RSBT for MWTS-II is the same as that of the MWHTS, except that the retrieved temperature profile of MWTS-II substitutes the temperature profile in the reference dataset to calculate the retrieved RSBT of MWTS-II. Then, the comparison of the retrieved RSBT and the RSBT of MWTS-II can be obtained.

4.3. Validation of the Retrieval Based on the Main Detection Pressure Range of Each Channel of the Microwave Sounder

The key to validating the retrieval results based on the detection range of pressure by each channel of the microwave sounder is to determine which pressure level is mainly detected by each channel of the microwave sounder. And it can be achieved by statistically characterizing the distribution of the peak of the channel weighting function of the microwave sounder. Although the weighting functions of MWTS-II and MWHTS as shown in Figure 1 can be used to determine the main detection pressure range for each channel, the calculation of these weighting functions is obtained under the clear sky condition and without the consideration of the influence of the cloud and rain. To better characterize the distribution of the peak value of the weighting functions of channels of MWTS-II and MWHTS under all weather conditions, the reference atmospheric profiles in the analysis dataset are used to calculate the weighting functions of MWTS-II and MWHTS in this study.
Take the determination of the main detection atmospheric pressure range of each channel of MWHTS as an example, the detailed process is as follows. The reference atmospheric profiles in the analysis dataset for MWHTS are fed into the RTTOV to calculate each channel’s weighting function of MWHTS, and one set of atmospheric profiles can obtain one set of weighting functions for each channel. Then, the atmospheric pressure level of each channel’s peak of the distribution of the weighting function is studied and counted, which in turn determines the range of the main atmospheric pressure for each detection channel. Based on the determination of the main detection atmospheric pressure range of each channel of MWHTS and the comparison between the retrieved RSBT and the RSBT of each channel of MWHTS, the validation of the retrieval results can be achieved by comparing the retrieved atmospheric profiles with the reference profiles in the validation dataset for MWHTS. Similarly, the main detection atmospheric pressure range of each channel of MWTS-II can be determined and the retrieved temperature profiles of MWTS-II can be validated. The validation method of the retrieved profiles based on the detection pressure range of each channel of microwave sounders is summarized in Figure 4.
Based on the description in Section 4.1 and Section 4.2, the validation method of the retrieved atmospheric profiles of MWHTS using its own observations is designed, and in turn, the retrieved atmospheric profiles are validated with the aid of the reference atmospheric profiles in a specific pressure range.

5. Retrieval Results and Validation

In order to realize the validation of the atmospheric profile retrieved by the microwave sounder based on the microwave sounder observations, multiple experiments are carried out here including the MWTS-II retrieval experiment of temperature profiles, validation experiment of MWTS-II retrieval results, MWHTS retrieval experiment of temperature and humidity profiles and validation experiment of MWHTS retrieval results. This section presents the validation results of retrieved atmospheric profiles of microwave sounders by comparing the traditional validation methods using the reanalysis data as the reference data and the proposed validation method using the microwave sounder observations as the reference data. In particular, four types of validation results are included: (1) the validation results of the retrieved temperature profile of MWTH-II using the ERA5 temperature profiles as the reference data and using the MWTS-II observed brightness temperature as the reference data, respectively, and (2) the validation results of temperature and humidity profiles of MWHTS using ERA5 temperature and humidity profiles as the reference data and using MWHTS observed brightness temperature as the reference data, respectively.

5.1. Validation Results of Temperature Profile Retrieved by MWTS-II

The observed brightness temperature data in the validation dataset for MWTS-II are fed into the retrieval model of MWTS-II to obtain 113,833 sets of retrieved temperature profiles. Then, the ERA5 temperature profiles in the validation dataset for MWTS-II are set as reference values to validate the retrieval values of the temperature profile. Figure 5 shows the comparison results between the retrieval values from three randomly selected cases of temperature profiles and the reference values of corresponding temperature profiles.
It can be observed from Figure 5 that the small variations in the structure of the reference temperature profiles can be captured by the three retrieved temperature profiles. For example, the retrieved temperature profiles at 100 hPa can follow the reference temperature profiles in terms of complex structural variations. This result indicates that the developed DNN-based retrieval algorithm is reliable to obtain the retrieved profiles which can be accord with practical physics. Meanwhile, it can be also observed that there are some sort of deviations between retrieved profiles and reference profiles in each case, for example, the difference between the retrieved profile 3 and the reference profile 3 is about 3.6 K at 550 hPa. To analyze the deviations between all the retrieved values of temperature profiles and the reference values, the mean bias and RMSE between the temperature profile reference values and the retrieved values are calculated and shown in Figure 6.
It can be seen from Figure 6 that the bias of MWTS-II between the retrieved temperature profiles and the reference temperature profiles is within 1 K, and the largest bias is about 0.9 K at around 5 hPa. As for the RMSE between the retrieved values and the reference values, except for the large value near the top of the atmosphere, the other pressures’ values remain around 2 K. Among them, the RMSE around 200 hPa has the largest value of 2.7 K. It is noted that the results of the retrieved temperature profiles of MWTS-II are validated by using ERA5 reanalysis data, rather than the real profiles, which means that the retrieved temperature profile may be more closed to the real profile compared with the EAR5 temperature profile. Thus, the observed brightness temperature of MWTS-II is further used as the reference data to validate the results of the temperature profile retrieved by MWTS-II.
Based on the validation method developed in Section 4, 113,833 samples are obtained from the retrieved RSBT calculated by MWTS-II retrieved temperature profile and the RSBT calculated by RA data, respectively. The RMSEs between these two types of simulated brightness temperature and observed brightness temperature are shown in Figure 7.
Figure 7 shows that the RMSE accuracy of the retrieved RSBT and the RSBT is close to each other for channels 1 to 10 of MWTS-II with a difference within 0.1 K. The accuracy of retrieved-based simulations is greater than that of reference simulations, and the differences are around 0.18 K, 0.26 K, and 1.2 K in channels 11 to 13, respectively. These results indicate that the retrieved temperature profiles of channels 1 to 10 of MWTS-II within the main detection pressure ranges are equivalent to the accuracy of ERA5 temperature profiles. Meanwhile, within the main detection pressure range of channels 11 to 13 of MWTS-II, the retrieved temperature profiles are closer to the real temperature profiles compared to the ERA5 temperature profiles. The channel weighting function of MWTS-II is calculated by using the reference atmospheric data in the analysis dataset for MWTS-II, and a total of 681,844 cases of weighting function calculation results are obtained. Table 1 shows the main detection pressure range of each channel determined by the analysis of the peak distribution height of the weighting function in each channel.
In this study, the distribution range of weighting function peak values is set as the main detection pressure range of each channel. However, the distribution of the weighting functions of each channel overlaps as shown in Figure 1, and the shapes and distributed peak values of the channel weighting function could be varied due to the change in weather conditions. Thus, it is possible that the same pressure range can be detected by different channels. It can be observed from Table 1 that the MWTS-II channels 1–10 have overlapped in the main detection range in some ways, especially the low-altitude detection channels 1–4 of MWTS-II.
The overlapping phenomenon of the main detection range of microwave sounder channels is not conducive to determining whether the retrieved temperature profile or the reference temperature profile is closer to the real profile within the specific pressure range through the comparison results between the retrieved RSBT and the RSBT. However, based on the retrieved RSBT and the RSBT comparison results shown in Figure 7 and the main detection range of each channel, the validation results can be provided by the comparison between the MWTS-II retrieved temperature profile and the reference temperature profile. In specific, the retrieved accuracy of the temperature profile by MWTS-II is equivalent to the accuracy of the reference temperature profile within the 30–1000 hPa pressure range due to the similar accuracies between the retrieved RSBT of MWTS-II from channel 1 to channel 10 and the corresponding RSBT. Since the accuracy of the retrieved RSBT of MWTS-II from channel 11 to channel 13 is greater than that of the RSBT, the accuracy of the retrieved temperature profile of MWTS-II is also greater than that of the reference temperature profile.

5.2. Validation Results of Temperature and Humidity Profiles Retrieved by MWHTS

A total of 177,199 data cases of retrieved temperature profiles and humidity profiles can be obtained by the MWHTS retrieval model when the observed brightness temperature of the validation dataset for MWHTS is fed into the model, and the temperature profiles and humidity profiles of the reference atmospheric data in the MWHTS validation dataset is used to validate the retrieved values. Figure 8 shows the comparison results between the retrieved values and the reference values by the three sets of randomly selected temperature and humidity profiles.
Based on the three retrieval cases discussed before, retrieving the humidity profile is more difficult than retrieving the temperature profile in terms of capturing the small changes in the profile structure. The main reason is that the spatial–temporal change in temperature is generally static while that of humidity profile is drastic, and it is difficult to describe the drastic changes in spatial–temporal chrematistics. To analyze the deviation level between the retrieved temperature and humidity profiles and the reference temperature and humidity profiles, the bias and RMSE between the reference values and the retrieved values are calculated and shown in Figure 9.
As for the temperature profile, the bias between the MWHTS retrieved profiles and the corresponding reference profiles is typically less than 3 K, except for that of the top of the atmosphere. The RMSE between the MWHTS retrieved profiles and the corresponding reference profiles are typically less than 3 K, except for those of the top of the atmosphere and the pressure of 400 hPa with the RMSE of 6 and 4 K, respectively. As for the humidity profile, the bias between the MWHTS retrieved profiles and the corresponding reference profiles is typically less than 4%, except for those of the pressure level ranging from 150 to 225 hPa with a bias of about 6%. The RMSE between the MWHTS retrieved profiles and the corresponding reference profiles is typically less than 20%. It should be noted that the above validation of the MWHTS retrievals of the temperature and humidity profiles uses the ERA5 reanalysis data, rather than the real profiles.
Next, the MWHTS observed brightness temperature is used to further validate the temperature and humidity profiles retrieved by MWHTS. The 177,199 data cases including RSBT and retrieved-based RSBT are calculated and obtained by using RA data and retrieved-based RA data in the validation dataset for MWHTS. Figure 10 shows the RSME results between these two types of simulated brightness temperature and the observed brightness temperature.
It can be seen from Figure 10 that the accuracy of RSBT is greater than that of retrieved RSBT in channels 1 to 10. For the temperature detection in channels 2 to 9, except for channels 3 and 6, the accuracy of RSBT is higher than that of retrieval RSBT. Among them, channels 7 to 9 are higher than 0.5 K while channels 2 and 4 are higher than 1 K. In humidity detection channels 11 to 15, the accuracy of the retrieved RSBT is higher than that of the RSBT, where the accuracies in channels 11 and 12 are the most obvious values of 1.18 K and 1.35 K, respectively. According to the accuracy comparison discussed above, the retrieved temperature profiles within the main detection range of channel 3 and channel 6 are closer to the real temperature profiles compared to the reference temperature profiles. However, in the remaining pressure range, the retrieved temperature profiles are not as good as the reference temperature profiles close to the real temperature profiles.
The main detection pressure range of each MWHTS channel is determined by the statistics using the peak distribution of weight function calculated by the analysis dataset for MWHTS, which can further provide the validation results of retrieved temperature and humidity profiles from MWHTS from the perspective of pressure. Table 2 shows the main detection pressure range of each channel of MWHTS.
According to the comparison results between of the retrieved RSBT and the RSBT of MWHTS, the accuracy of the retrieved RSBT from the temperature detection channel 6 is higher, and the main detection range of channel 6 is from 250 hPa to 775 hPa. The main detection range overlaps with the main detection pressure range of channel 5 and channels 7 to 9, and the accuracy of retrieved RSBT in these channels is not as good as the RSBT. Thus, this comparison result of channel 6 to evaluate the retrieval accuracy of the temperature profile cannot be used. Based on the accuracy of the retrieved RSBT in channel 3 being higher, and the main detection range of this channel being around 50 hPa, it is indicated that the retrieved temperature profile near 50 hPa is closer to the real temperature profile than the reference temperature profile, while the accuracy of the retrieved temperature profile is worse than the reference temperature profile in the rest of pressure range. Due to the accuracy of the retrieved RSBT being greater than that of the RSBT in humidity channels 11 to 15 of MWHTS, the humidity profiles retrieved by MWHTS are closer to the real humidity profiles than the reference humidity profiles in channels 11 to 15 within 350 hPa to 925 hPa.
Based on the above analysis, when the observed brightness temperature of the microwave sounder is used as the reference data to validate the retrieved results of the microwave sounder, more references on the validation results of the reference data based on reanalysis data can be provided. Specifically, when the temperature profile retrieved by MWTS-II is validated through the observed brightness temperature of MWTS-II within the pressure range of 30 hPa to 1000 hPa, the accuracy of the temperature profile retrieved by MWTS-II is basically the same as the accuracy of the temperature profile of ERA5. Then, the accuracy of the temperature profile retrieved by MWTS-II is higher than that of the temperature profile of ERA5 within the pressure range from 3 hPa to 30 hPa. When the temperature and humidity profiles retrieved by MWHTS are validated through the observed brightness temperature of MWHTS within the pressure around 15 hPa, the accuracy of the temperature profile retrieved by MWHTS is higher than the accuracy of the temperature profile of ERA5, while the accuracy of the retrieved temperature profile is lower than that of the temperature profile of ERA5 under the rest of the pressure range. Meanwhile, the accuracy of the humidity profile retrieved by MWHTS is higher than that of the humidity profile of ERA5 under the pressure range of 350 hPa to 925 hPa.

6. Conclusions

This study uses microwave sounder observations to validate the accuracy of the atmospheric temperature and humidity profiles retrieved by microwave sounders. The simulated brightness temperature using the retrieved profiles of the microwave sounder and the reference atmospheric profiles of the ERA5 reanalysis dataset are compared and the main detection range of the atmospheric pressure is determined to perform the comparison study of the retrieved profiles and the ERA5 profiles in specific pressure range, which provide a good reference for the application of the retrieved atmospheric profiles. The validation experimental results show that the retrieved temperature profiles of MWTS-II have higher accuracy compared to the temperature profiles of ERA5 in the atmospheric pressure range of 3–30 hPa, and the accuracy of the rest of the pressure range is comparable between the profiles of ERA5 and the retrieved profiles. And the retrieved temperature profiles of MWHTS have higher accuracy compared to the temperature profiles of ERA5 in the atmospheric pressure level around 50 hPa and lower accuracy in the rest of the pressure levels. In addition, the retrieved humidity profiles of MWHTS have higher accuracy compared to the humidity profiles of ERA5 in the atmospheric pressure range of 350–925 hPa. According to the research results of this study, it can be concluded that the method of using spaceborne microwave sounder observations as the reference data to verify the retrieved profiles of satellite-based microwave sounders is feasible, and it can provide valuable references for the application of satellite-based microwave sounders to invert atmospheric parameters.
It is noted that the validation method of the retrieved profiles using microwave observations is essentially a comparison of the retrieved atmospheric profiles with the corresponding profiles in ERA5 reanalysis data. If we have more reference data such as the RAOB or other retrieved data from satellite-based instruments, more valuable and promising validation results may be obtained. Meanwhile, the results are obtained from the validation study of retrieved results of MWTS-II and MWHTS within the selected geographical range. Carrying out the verification of satellite-borne microwave sounder inversion parameters on a global scale is a systematic project, and it is also of great significance for the application of satellite-borne microwave sounder observations. Therefore, more types of reference atmospheric profile data will be induced to perform the validation for the retrieved atmospheric profiles, and more verification will be performed over a wider geographical area, which is the focus of our next research topic. In addition, the validation study is performed for the retrieval results of microwave sounder under all weather conditions and the study under specific weather conditions will be performed in the future, which can be promising to obtain more interesting validation results.

Author Contributions

Q.H. and J.L. designed the study; Q.H. and R.Z. carried out the experiments; J.L. presented some conclusions; Q.H. wrote the manuscripts; and Q.H., J.J. and X.G. edited the article. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by the National Natural Science Foundation of China under Grant No. 41901297, the China Postdoctoral Science Foundation under Grant No. 2021M693201, the Science and Technology Key Project of Henan Province under Grant No. 24210230012, the Special project of key research and development Plan of Henan Province under Grant No. 221111111700, and the State Key Laboratory of Geo-Information Engineering under Grant NO. SKLGIE2021-Z-3-2.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

Data are contained within the article.

Acknowledgments

The authors would like to thank NSMC for providing the MWHTS and MWTS-II observations, as well as ECMWF for providing the ERA5 reanalysis data.

Conflicts of Interest

The authors declare no conflicts of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

Abbreviations

AcronymMeaning
MWTS-IIMicrowave Temperature Sounder-II
MWHTSMicrowave Humidity and Temperature Sounder
ECMWFEuropean Center for Medium-Range Weather Forecasts
ERA5The fifth-generation ECMWF reanalysis data
NCEPNational Center for Environmental Prediction
RAOBRadiosonde Observations
RA dataReference Atmospheric data
RSBTReference Simulated Brightness Temperature
RTTOVRadiative Transfer for Television and Infrared Observation Satellite Operational Vertical Sounder
DNNDeep Neural Network
RMSERoot Mean Square Error

References

  1. He, Q.; Wang, Z.; Li, J. Application of the Deep Neural Network in Retrieving the Atmospheric Temperature and Humidity Profiles from the Microwave Humidity and Temperature Sounder Onboard the Feng-Yun-3 Satellite. Sensors 2021, 21, 4673. [Google Scholar] [CrossRef] [PubMed]
  2. Weng, F.; Zou, X.; Wang, X.; Yang, S.; Goldberg, M.D. Introduction to Suomi National Polar-Orbiting Partnership Advanced Technology Microwave Sounder for Numerical Weather Prediction and Tropical Cyclone Applications. J. Geophys. Res. Atmos. 2012, 117, D19112. [Google Scholar] [CrossRef]
  3. Newman, S.; Carminati, F.; Lawrence, H.; Bormann, N.; Salonen, K.; Bell, W. Assessment of New Satellite Missions within the Framework of Numerical Weather Prediction. Remote Sens. 2020, 12, 1580. [Google Scholar] [CrossRef]
  4. Yang, Z.; Lu, N.; Shi, J.; Zhang, P.; Dong, C.; Yang, J. Overview of FY-3 Payload and Ground Application System. IEEE Trans. Geosci. Remote Sens. 2012, 50, 4846–4853. [Google Scholar] [CrossRef]
  5. Yang, J.X.; Lee, Y.-K.; Grassotti, C.; Garrett, K.; Liu, Q.; Blackwell, W.; Leslie, R.V.; Greenwald, T.; Bennartz, R.; Braun, S. Atmospheric Humidity and Temperature Sounding from the CubeSat TROPICS Mission: Early Performance Evaluation with MiRS. Remote Sens. Environ. 2023, 287. [Google Scholar] [CrossRef]
  6. Sivira, R.G.; Brogniez, H.; Mallet, C.; Oussar, Y. A Layer-Averaged Relative Humidity Profile Retrieval for Microwave Observations: Design and Results for the Megha-Tropiques Payload. Atmos. Meas. Tech. 2015, 8, 1055–1071. [Google Scholar] [CrossRef]
  7. He, Q.; Wang, Z.; He, J. Bias Correction for Retrieval of Atmospheric Parameters from the Microwave Humidity and Temperature Sounder Onboard the Fengyun-3C Satellite. Atmosphere 2016, 7, 156. [Google Scholar] [CrossRef]
  8. Li, J.; Qian, X.; Qin, Z.; Liu, G. Direct Assimilation of Chinese FY-3E Microwave Temperature Sounder-3 Radiances in the CMA-GFS: An Initial Study. Remote Sens. 2022, 14, 5943. [Google Scholar] [CrossRef]
  9. Qian, X.; Qin, Z.; Li, J.; Han, Y.; Liu, G. Preliminary Evaluation of FY-3E Microwave Temperature Sounder Performance Based on Observation Minus Simulation. Remote Sens. 2022, 14, 2250. [Google Scholar] [CrossRef]
  10. Zhang, Q.; Ye, J.; Zhang, S.; Han, F. Precipitable Water Vapor Retrieval and Analysis by Multiple Data Sources: Ground-Based GNSS, Radio Occultation, Radiosonde, Microwave Satellite, and NWP Reanalysis Data. J. Sens. 2018, 2018, 3428303. [Google Scholar] [CrossRef]
  11. Dee, D.P.; Uppala, S.M.; Simmons, A.J.; Berrisford, P.; Poli, P.; Kobayashi, S.; Andrae, U.; Balmaseda, M.A.; Balsamo, G.; Bauer, P.; et al. The ERA-Interim Reanalysis: Configuration and Performance of the Data Assimilation System. Q. J. R. Meteorol. Soc. 2011, 137, 553–597. [Google Scholar] [CrossRef]
  12. Chen, B.; Liu, Z. Global Water Vapor Variability and Trend from the Latest 36 Year (1979 to 2014) Data of ECMWF and NCEP Reanalyses, Radiosonde, GPS, and Microwave Satellite. J. Geophys. Res. Atmos. 2016, 121, 11442–11462. [Google Scholar] [CrossRef]
  13. Kan, W.; Han, Y.; Weng, F.; Guan, L.; Gu, S. Multisource Assessments of the FengYun-3D Microwave Humidity Sounder (MWHS) On-Orbit Performance. IEEE Trans. Geosci. Remote Sens. 2020, 58, 7258–7268. [Google Scholar] [CrossRef]
  14. Niu, Z.; Zhang, L.; Dong, P.; Weng, F.; Huang, W. Impact of Assimilating FY-3D MWTS-2 Upper Air Sounding Data on Forecasting Typhoon Lekima (2019). Remote Sens. 2021, 13, 1841. [Google Scholar] [CrossRef]
  15. He, J.; Zhang, S. Introduction of MWHTS onboard FY-3C Satellite and Typhoon Detecting. Remote Sens. Sci. 2015, 3, 2. [Google Scholar]
  16. Niu, Z.; Zou, X.; Ray, P.S. Development and Testing of a Clear-Sky Data Selection Algorithm for FY-3C/D Microwave Temperature Sounder-2. Remote Sens. 2020, 12, 1478. [Google Scholar] [CrossRef]
  17. Xia, X.; Zou, X. Combining FY-3D MWTS-2 with AMSU-A Data for Inter-Decadal Diurnal Correction and Climate Trends of Atmospheric Temperature. Remote Sens. 2021, 13, 3148. [Google Scholar] [CrossRef]
  18. Liu, K.; He, J.; Chen, H. Precipitation Retrieval From Fengyun-3D MWHTS and MWRI Data Using Deep Learning. IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens. 2022, 15, 7619–7630. [Google Scholar] [CrossRef]
  19. Wang, Z.; Li, J.; He, J.; Zhang, S.; Gu, S.; Li, Y.; Guo, Y.; He, B. Performance Analysis of Microwave Humidity and Temperature Sounder Onboard the FY-3D Satellite From Prelaunch Multiangle Calibration Data in Thermal/Vacuum Test. IEEE Trans. Geosci. Remote Sens. 2019, 57, 1664–1683. [Google Scholar] [CrossRef]
  20. Guo, Y.; He, J.; Gu, S.; Lu, N. Calibration and Validation of Feng Yun-3-D Microwave Humidity Sounder II. IEEE Geosci. Remote Sens. Lett. 2020, 17, 1846–1850. [Google Scholar] [CrossRef]
  21. He, Q. Study on Retrieving the Atmospheric Temperature and Humidity Profiles from Measurements of Microwave Humidity and Temperature Sounder on FY-3C Satellite. Ph.D. Thesis, Chinese Academy of Science, Beijing, China, 2017. [Google Scholar]
  22. Marzano, F.; Guido, V. Remote Sensing of Atmosphere and Ocean from Space: Models, Instruments and Techniques; Springer Science & Business Media: Berlin/Heidelberg, Germany, 2006. [Google Scholar]
  23. Andronache, C. Remote Sensing of Clouds and Precipitation; Springer: Cham, Switzerland, 2018. [Google Scholar]
  24. Oliver, J.E. Encyclopedia of World Climatology; Springer: Dordrecht, The Netherland, 2005. [Google Scholar]
  25. Hersbach, H.; Bell, B.; Berrisford, P.; Hirahara, S.; Horányi, A.; Muñoz-Sabater, J.; Nicolas, J.; Peubey, C.; Radu, R.; Schepers, D.; et al. The ERA5 Global Reanalysis. Q. J. R. Meteorol. Soc. 2020, 146, 1999–2049. [Google Scholar] [CrossRef]
  26. Hoffmann, L.; Gunther, G.; Li, D.; Stein, O.; Wu, X.; Sabine, G.; Heng, Y.; Konopka, P.; Muller, R.; Vogel, B.; et al. From ERA-Interim to ERA5: The Considerable Impact of ECMWF’s Next-Generation Reanalysis on Lagrangian Transport Simulations. Atmos. Chem. Phys. 2019, 19, 3097–3124. [Google Scholar] [CrossRef]
  27. Bell, B.; Hersbach, H.; Simmons, A.; Berrisford, P.; Dahlgren, P.; Horányi, A.; Muñoz-Sabater, J.; Nicolas, J.; Radu, R.; Schepers, D.; et al. The ERA5 Global Reanalysis: Preliminary Extension to 1950. Q. J. R. Meteorol. Soc. 2021, 147, 4186–4227. [Google Scholar] [CrossRef]
  28. Matricardi, M. A Principal Component Based Version of the RTTOV Fast Radiative Transfer Model. Q. J. R. Meteorol. Soc. 2010, 136, 1823–1835. [Google Scholar] [CrossRef]
  29. Saunders, R. RTTOV-7: Science and Validation Report; Met Office: Exeter, UK, 2002. [Google Scholar]
  30. Weng, F. Preface for the Special Issue of Radiative Transfer Models for Satellite Data Assimilation. J. Quant. Spectrosc. Radiat. Transf. 2020, 244, 106826. [Google Scholar] [CrossRef]
  31. Hu, H.; Weng, F. Influences of 1DVAR Background Covariances and Observation Operators on Retrieving Tropical Cyclone Thermal Structures. Remote Sens. 2022, 14, 1078. [Google Scholar] [CrossRef]
  32. Xu, J.; Ma, Z.; Hu, H.; Weng, F. A Cloud-Dependent 1DVAR Precipitation Retrieval Algorithm for FengYun-3D Microwave Soundings: A Case Study in Tropical Cyclone Mekkhala. IEEE Geosci. Remote Sens. Lett. 2023, 20, 1–5. [Google Scholar] [CrossRef]
  33. Carminati, F.; Migliorini, S. All-Sky Data Assimilation of MWTS-2 and MWHS-2 in the Met Office Global NWP System. Adv. Atmos. Sci. 2021, 38, 1682–1694. [Google Scholar] [CrossRef]
  34. Han, Y.; Yang, J.; Hu, H.; Dong, P. Retrieval of Oceanic Total Precipitable Water Vapor and Cloud Liquid Water from Fengyun-3D Microwave Sounding Instruments. J. Meteorol. Res. 2021, 35, 371–383. [Google Scholar] [CrossRef]
  35. Hu, J.; Wu, J.; Petropoulos, G.P.; Bao, Y.; Liu, J.; Lu, Q.; Wang, F.; Zhang, H.; Liu, H. Temperature and Relative Humidity Profile Retrieval from Fengyun-3D/VASS in the Arctic Region Using Neural Networks. Remote Sens. 2023, 15, 1648. [Google Scholar] [CrossRef]
  36. Yuan, Q.; Shen, H.; Li, T.; Li, Z.; Li, S.; Jiang, Y.; Xu, H.; Tan, W.; Yang, Q.; Wang, J.; et al. Deep Learning in Environmental Remote Sensing: Achievements and Challenges. Remote Sens. Environ. 2020, 241, 111716. [Google Scholar] [CrossRef]
  37. Chen, X.; Valencia, R.; Soleymani, A.; Scott, K.A. Predicting Sea Ice Concentration With Uncertainty Quantification Using Passive Microwave and Reanalysis Data: A Case Study in Baffin Bay. IEEE Trans. Geosci. Remote Sens. 2023, 61, 1–13. [Google Scholar] [CrossRef]
  38. Papa, M.; Mattioli, V.; Montopoli, M.; Casella, D.; Rydberg, B.; Marzano, F.S. Investigating Spaceborne Millimeter-Wave Ice Cloud Imager Geolocation Using Landmark Targets and Frequency-Scaling Approach. IEEE Trans. Geosci. Remote Sens. 2022, 60, 1–17. [Google Scholar] [CrossRef]
  39. Dee, D.P. Bias and Data Assimilation. Q. J. R. Meteorol. Soc. 2005, 131, 3323–3343. [Google Scholar] [CrossRef]
  40. Auligné, T.; McNally, A.P.; Dee, D.P. Adaptive Bias Correction for Satellite Data in a Numerical Weather Prediction System. Q. J. R. Meteorol. Soc. 2007, 133, 631–642. [Google Scholar] [CrossRef]
  41. Dee, D.P. Variational Bias Correction of Radiance Data in the ECMWF System. In Proceedings of the ECMWF Workshop on Assimilation of High Spectral Resolution Sounders in NWP, Reading, UK, 28 June–1 July 2004. [Google Scholar]
  42. Dee, D.P.; Uppala, S. Variational Bias Correction of Satellite Radiance Data in the ERA-Interim Reanalysis: Variational bias correction in ERA-Interim. Q. J. R. Meteorol. Soc. 2009, 135, 1830–1841. [Google Scholar] [CrossRef]
Atmosphere 15 00760 g001
Figure 1. The weighting function of each channel of MWTS-II and MWHTS with channel frequency setting. (a) MWTS-II. (b) MWHTS.
Figure 1. The weighting function of each channel of MWTS-II and MWHTS with channel frequency setting. (a) MWTS-II. (b) MWHTS.
Atmosphere 15 00760 g001
Atmosphere 15 00760 g002
Figure 2. The structure of the developed DNN model.
Figure 2. The structure of the developed DNN model.
Atmosphere 15 00760 g002
Atmosphere 15 00760 g003
Figure 3. The comparison of the retrieved RSBT and RSBT of MWHTS.
Figure 3. The comparison of the retrieved RSBT and RSBT of MWHTS.
Atmosphere 15 00760 g003
Atmosphere 15 00760 g004
Figure 4. The determination of the main detection pressure range of the microwave sounder and validation of its inversion results.
Figure 4. The determination of the main detection pressure range of the microwave sounder and validation of its inversion results.
Atmosphere 15 00760 g004
Atmosphere 15 00760 g005
Figure 5. Comparison results of retrieved temperature profiles and reference temperature profiles.
Figure 5. Comparison results of retrieved temperature profiles and reference temperature profiles.
Atmosphere 15 00760 g005
Atmosphere 15 00760 g006
Figure 6. Bias and RMSE between the retrieved temperature profile and the reference profile.
Figure 6. Bias and RMSE between the retrieved temperature profile and the reference profile.
Atmosphere 15 00760 g006
Atmosphere 15 00760 g007
Figure 7. Brightness temperature RMSE comparison between the retrieved RSBT and the RSBT of MWTS-II.
Figure 7. Brightness temperature RMSE comparison between the retrieved RSBT and the RSBT of MWTS-II.
Atmosphere 15 00760 g007
Atmosphere 15 00760 g008
Figure 8. Comparison of the MWHTS retrieved profiles and the corresponding reference profiles. (a) Temperature profiles. (b) Humidity profiles.
Figure 8. Comparison of the MWHTS retrieved profiles and the corresponding reference profiles. (a) Temperature profiles. (b) Humidity profiles.
Atmosphere 15 00760 g008
Atmosphere 15 00760 g009
Figure 9. Bias and RMSE between the MWHTS retrieved profiles and the corresponding reference profiles. (a) Temperature. (b) Humidity.
Figure 9. Bias and RMSE between the MWHTS retrieved profiles and the corresponding reference profiles. (a) Temperature. (b) Humidity.
Atmosphere 15 00760 g009
Atmosphere 15 00760 g010
Figure 10. Brightness temperature RMSE comparison between the retrieved RSBT and the RSBT of MWHTS.
Figure 10. Brightness temperature RMSE comparison between the retrieved RSBT and the RSBT of MWHTS.
Atmosphere 15 00760 g010
Table 1. Detection pressure range of each channel of MWTS-II.
Table 1. Detection pressure range of each channel of MWTS-II.
ChannelMain Detection Pressure Range (hPa)ChannelMain Detection Pressure Range (hPa)
1775–1000870–125
2775–1000930–70
3775–9251030
4600–7751110
5250–450125–7
6175–250133
7125–175
Table 2. Detection pressure range of each channel of MWHTS.
Table 2. Detection pressure range of each channel of MWHTS.
ChannelMain Detection Pressure Range (hPa)ChannelMain Detection Pressure Range (hPa)
1850–10009775–1000
23010775–1000
35011350–750
410012350–850
525013450–850
6250–77514500–925
7775–100015550–925
8775–1000
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

He, Q.; Li, J.; Zhang, R.; Jia, J.; Guo, X. Research on Validation Method on Retrieval of Atmospheric Temperature and Humidity Profile Using a Microwave Sounder. Atmosphere 2024, 15, 760. https://doi.org/10.3390/atmos15070760

AMA Style

He Q, Li J, Zhang R, Jia J, Guo X. Research on Validation Method on Retrieval of Atmospheric Temperature and Humidity Profile Using a Microwave Sounder. Atmosphere. 2024; 15(7):760. https://doi.org/10.3390/atmos15070760

Chicago/Turabian Style

He, Qiurui, Jiaoyang Li, Ruiling Zhang, Junqi Jia, and Xiao Guo. 2024. "Research on Validation Method on Retrieval of Atmospheric Temperature and Humidity Profile Using a Microwave Sounder" Atmosphere 15, no. 7: 760. https://doi.org/10.3390/atmos15070760

APA Style

He, Q., Li, J., Zhang, R., Jia, J., & Guo, X. (2024). Research on Validation Method on Retrieval of Atmospheric Temperature and Humidity Profile Using a Microwave Sounder. Atmosphere, 15(7), 760. https://doi.org/10.3390/atmos15070760

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop