Brightness Temperature Observations Research Articles

A Machine Learning Approach for Deriving Atmospheric Temperatures and Typhoon Warm Cores From FY‐3E MWTS‐3 Observations

AbstractMachine learning has gained an increasing popularity in the fields of satellite retrieval and numerical weather modeling. In this study, machine‐learning (ML) neural‐network (NN) models are utilized to retrieve atmospheric temperatures from observations of brightness temperature from the Microwave Temperature Sounder‐3 (MWTS‐3) onboard the China's first dawn‐dusk polar‐orbiting satellite Fengyun (FY)‐3E, with the ERA5 reanalysis serving as training data sets. The root mean square errors of the ML‐retrieved temperatures at all pressure levels are smaller than those obtained by a previously used traditional linear regression method compared to the ERA5 reanalysis over global oceans as well as radiosonde observations over land. Less than a 1‐week period of training data is usually sufficient for an ML NN model to converge in less than 50–100 iterations. The shortest time period of training data is 3‐days right before the testing data period. While the horizontal patterns and temporal evolutions of the ML‐retrieved warm cores of Typhoon Malakas (2022) and Typhoon Haikui (2023) in the upper troposphere compared favorably with those obtained by traditional regression methods as well as the ERA5 reanalysis in the testing periods. The vertical structures of ML‐retrieved warm cores extend further down to the middle and lower troposphere while those from the traditional regression method are confined in the upper troposphere. A comparison among ML results with additional training data sets across different seasons confirms the above conclusion.

Machine learning-based retrieval of total column water vapor over land using GMI-sensed passive microwave measurements

ABSTRACT The Global Precipitation Measurement (GPM) Microwave Imager (GMI) is a microwave (MW) radiometer that has near-global coverage and frequent revisit time. To date, operational total column water vapor (TCWV) data records from the GPM GMI sensor have been exclusively offered over oceanic regions. It is challenging to retrieve TCWV over land from satellite MW measurements because of varying land surface characteristics. In this paper, a novel Light Gradient Boosting Machine-based retrieval algorithm is proposed to derive TCWV over land from GMI-sensed MW brightness temperature (BT) observations. The GMI-observed MW BT at 18.7 GHz and 23.8 GHz, differential BT between 18.7 GHz and 23.8 GHz, latitude, longitude, and month are selected and utilized as the input variables of the retrieval approach, because of their strong correlation with satellite-sensed MW TCWV retrievals. Instead of surface emissivity data or radiative transfer model, we take into account the spatial and temporal elements, namely latitude, longitude, and month. The training of the retrieval method is performed based on ground-based TCWV estimates from worldwide 4,471 Global Navigation Satellite System (GNSS) stations in 2017. The performance of the newly proposed retrieval algorithm is independently validated in a worldwide coverage using reference TCWV from additional 4,341 GNSS stations in 2018–2020 and 605 radiosonde stations in 2017–2020. The newly retrieved TCWV estimates over land have a correlation coefficient of 0.76 and 0.83, a root-mean-square error (RMSE) of 5.82 mm and 6.02 mm, a relative RMSE of 34.91% and 34.36%, and a mean bias of 0.02 mm and −0.42 mm compared to reference TCWV from GNSS and radiosonde, respectively. The performance of the retrieval algorithm is satisfactory when compared to that of land-purpose TCWV of other satellite missions, though we have not used either surface emissivity data or radiative transfer model. This result increases confidence in retrieving TCWV over land from satellite-sensed MW BT measurements based on machine learning using ground-based TCWV observations. The newly developed retrieval algorithm has the potential for integration into operational satellite missions or meteorological services, thereby enhancing weather forecasting, climate modeling, and other relevant applications.

Low-frequency variability and teleconnections during Northern Hemisphere winters captured by satellite microwave brightness temperature observations

Low-frequency variability and teleconnections during Northern Hemisphere winters captured by satellite microwave brightness temperature observations

Thermal infrared shadow-hiding in GOES-R ABI imagery: snow and forest temperature observations from the SnowEx 2020 Grand Mesa field campaign

Abstract. The high temporal resolution of thermal infrared imagery from the Geostationary Operational Environmental Satellites R-series (GOES-R) presents an opportunity to observe mountain snow and forest temperatures over the full diurnal cycle. However, the off-nadir views of these imagers may impact or bias temperature observations, especially when viewing a surface composed of both snow and forests. We used GOES-16 and -17 thermal infrared brightness temperature observations of a flat snow- and forest-covered study site at Grand Mesa, Colorado, USA, to characterize how forest coverage and view angle impact these observations. These two geostationary satellites provided views of the study area from the southeast (134.1° azimuth, 33.5° elevation) and southwest (221.2° azimuth, 35.9° elevation), respectively. As part of the NASA SnowEx field campaign in February 2020, coincident brightness temperature observations from ground-based and airborne IR sensors were collected to compare with those from the geostationary satellites. Observations over the course of 2 cloud-free days spanned the entire study site. The brightness temperature observations from each dataset were compared to find their relative differences and how those differences may have varied over time and/or as a function of varying forest cover across the study area. GOES-16 and -17 brightness temperatures were found to match the diurnal cycle and temperature range within ∼ 1 h and ± 3 K of ground-based observations. GOES-16 and -17 were both biased warmer than nadir-looking airborne IR and ASTER observations. The warm biases were higher at times when the sun–satellite phase angle was near its daily minimum. The phase angle, the angle between the direction of incoming solar illumination and the direction from which the satellite is viewing, reached daily minimums in the morning for GOES-16 and afternoon for GOES-17. In morning observations, warm biases in GOES-16 brightness temperature were greater for pixels that contained more forest coverage. The observations suggest that a “thermal infrared shadow-hiding” effect may be occurring, where the geostationary satellites are preferentially seeing the warmer sunlit sides of trees at different times of day. These biases are important to understand for applications using GOES-R brightness temperatures or derived land surface temperatures (LSTs) over areas with surface roughness features, such as forests, that could exhibit a thermal infrared shadow-hiding effect.

Dynamical study of three African Easterly Waves in September 2021

AbstractThree convectively active African easterly waves (AEWs) that propagated south of the African easterly jet were observed over the northeast Atlantic Ocean in September 2021. Their evolution is studied using a suite of theoretical frameworks, as well as the European Centre for Medium‐range Weather Forecast reanalyses and satellite‐derived brightness temperature observations. The environment of these AEWs was sampled during the Cloud–Atmospheric Dynamics–Dust Interactions in West Africa campaign near Cape Verde with the goal to assess their potential for developing into tropical cyclones. We highlight the processes that inhibited the development of the first AEW (which evolved into tropical disturbance Pierre‐Henri) and that played a role in the development of the later two into tropical storms Rose and Peter on September 19, 2021. The three AEWs developed a so‐called “marsupial protective” pouch. For Peter and Rose, the pouch was associated with a vertically aligned vortex at low levels and efficiently protected the convective systems inside from dry and dusty air intrusion. The development of this low‐level vortex is associated with an interaction with the monsoon trough for Rose and with a vorticity center associated with a wave propagating north of the African easterly jet (AEJ) in the case of Peter. The presence of a dust flux toward the convective core near the surface is highlighted for Rose and Peter in spite of the presence of the protective marsupial pouch. On the other hand, Pierre‐Henri interacted positively with both the monsoon trough and an AEW north of the AEJ but failed to develop into a tropical cyclone. The wave north of the AEJ brought Saharan air layer air masses inside the pouch that led to a drying of the circulation that may explain the decrease in convective activity.

Isolated deep convections over the Tibetan Plateau in the rainy season during 2001–2020

The Tibetan Plateau (TP) is a prevalent region for convection systems due to its unique thermodynamic forcing. This study investigated isolated deep convections (IDCs), which have a smaller spatial and temporal size than mesoscale convective systems (MCSs), over the TP in the rainy season (June–September) during 2001–2020. The authors used satellite precipitation and brightness temperature observations from the Global Precipitation Measurement mission. Results show that IDCs mainly concentrate over the southern TP. The IDC number per rainy season decreases from around 140 over the southern TP to around 10 over the northern TP, with an average 54.2. The initiation time of IDCs exhibits an obvious diurnal cycle, with the peak at 1400–1500 LST and the valley at 0900–1000 LST. Most IDCs last less than five hours and more than half appear for only one hour. IDCs generally have a cold cloud area of 7422.9 km2, containing a precipitation area of approximately 65%. The larger the IDC, the larger the fraction of intense precipitation it contains. IDCs contribute approximately 20%–30% to total precipitation and approximately 30%–40% to extreme precipitation over the TP, with a larger percentage in July and August than in June and September. In terms of spatial distribution, IDCs contribute more to both total precipitation and extreme precipitation over the TP compared to the surrounding plain regions. IDCs over the TP account for a larger fraction than MCSs, indicating the important role of IDCs over the region.摘要本文利用卫星观测资料, 研究了2001–2020年雨季 (6–9月) 青藏高原上孤立深对流 (Isolated deep convections, IDCs) 的气候特征. IDCs定义为比中尺度对流系统 (Mesoscale convective systems, MCSs) 时空尺度小的对流. 结果显示, 每年雨季青藏高原上平均的IDC数量为54.2个, 主要分布在高原的南部. IDCs的初始时刻呈现明显的日循环, 在下午14–15时为峰值, 在上午9–10时为谷值. 大部分IDCs持续时间在5小时以内, 超过一半的IDCs仅持续1小时. IDCs的冷云平均面积约为7422.9km2, 其中包含65%的降水面积. IDC面积越大, 包含的强降水范围也越大. IDCs对青藏高原总降水的贡献约为20%–30%, 对极端降水贡献约为30%–40%, 在7月和8月的占比大于6月和9月. 在空间分布方面, 青藏高原上IDCs对总降水和极端降水的贡献大于周围平原地区. 青藏高原上IDCs对降水的贡献大于MCSs, 表明IDCs在该地区起着重要作用.

Joint Retrieval of Multiple Species of Ice Hydrometeor Parameters from Millimeter and Submillimeter Wave Brightness Temperature Based on Convolutional Neural Networks

Submillimeter wave radiometers are promising remote sensing tools for sounding ice cloud parameters. The Ice Cloud Imager (ICI) aboard the second generation of the EUMETSAT Polar System (EPS−SG) is the first operational submillimeter wave radiometer used for ice cloud remote sensing. Ice clouds simultaneously contain three species of ice hydrometeors—ice, snow, and graupel—the physical distributions and submillimeter wave radiation characteristics of which differ. Therefore, jointly retrieving the mass parameters of the three ice hydrometeors from submillimeter brightness temperatures is very challenging. In this paper, we propose a multiple species of ice hydrometeor parameters retrieval algorithm based on convolutional neural networks (CNNs) that can jointly retrieve the total content and vertical profiles of ice, snow, and graupel particles from submillimeter brightness temperatures. The training dataset is generated by a numerical weather prediction (NWP) model and a submillimeter wave radiative transfer (RT) model. In this study, an end to end ICI simulation experiment involving forward modeling of the brightness temperature and retrieval of ice cloud parameters was conducted to verify the effectiveness of the proposed CNN retrieval algorithm. Compared with the classical Unet, the average relative errors of the improved RCNN–ResUnet are reduced by 11%, 25%, and 18% in GWP, IWP, and SWP retrieval, respectively. Compared with Bayesian Monte Carlo integration algorithm, the average relative error of the total content retrieved by RCNN–ResUnet is reduced by 71%. Compared with BP neural network algorithm, the average relative error of the vertical profiles retrieved by RCNN–ResUnet is reduced by 69%. In addition, this algorithm was applied to actual Advanced Technology Microwave Sounder (ATMS) 183 GHz observed brightness temperatures to retrieve graupel particle parameters with a relative error in the total content of less than 25% and a relative error in the profile of less than 35%. The results show that the proposed CNN algorithm can be applied to future space borne submillimeter wave radiometers to jointly retrieve mass parameters of ice, snow, and graupel.

On the use of dual-polarized multi-angular observations of P-band brightness temperature for soil moisture profile retrieval in thawed mineral soil

ABSTRACT This article investigated the possibility of remotely sensing the soil moisture profile in thawed soil from multi-angular dual-polarized brightness temperature (TB) observations at P-band frequencies of 750 MHz and 409 MHz using a modified Burke model. Moreover, it was found that an excellent agreement (coefficient of determination R2 = 0.999 and root-mean-square error (RMSE) no more than RMSE = 0.6 K) could be achieved between the Njoku coherent brightness temperature model and the modified Burke model by introducing a reflectivity from the air-soil interface that takes into account the phases of the multiple re-reflected waves in the underlying layers. Based on the modified Burke model, the depths from which apparent moisture and temperature could be retrieved in a dielectrically-inhomogeneous, non-isothermal soil were investigated, being approximately ten times less than the depth for which apparent soil temperature could be retrieved. In general, the thickness of the emitting layer depends on the TB look angle and polarization, along with the moisture and temperature profiles of the soil. It was also shown that due to the effect of the Brewster angle, the H-polarization of TB was twice as sensitive (4 K/1%) to changes in volumetric soil moisture than V-polarization (1.9 K/1%). Based on multi-angular (10°-50°) observations of TB at H- and V-polarizations, a method of moisture profile retrieval in the top 5–15 cm soil (depending on surface moisture) was proposed using an exponential fitting function, the parameters of which are found in the course of solving the inverse problem. A decrease in the sensing frequency from 750 MHz to 409 MHz makes it possible to increase the accuracy of soil moisture profiles retrieval by a factor of two, being from RMSE = 1.6% (R2 = 0.946) to RMSE = 0.85% (R2 = 0.982) in the top 15 cm layer of soil. The conducted investigation shows the promise of using P-band observations of TB for soil moisture profile retrieval.

Passive microwave remote-sensing-based high-resolution snow depth mapping for Western Himalayan zones using multifactor modeling approach

Abstract. Spatiotemporal snow depth (SD) mapping in the Indian Western Himalayan (WH) region is essential in many applications pertaining to hydrology, natural disaster management, climate, etc. In situ techniques for SD measurement are not sufficient to represent the high spatiotemporal variability in SD in the WH region. Currently, low-frequency passive microwave (PMW) remote-sensing-based algorithms are extensively used to monitor SD at regional and global scales. However, fewer PMW SD estimation studies have been carried out for the WH region to date, which are mainly confined to small subregions of the WH region. In addition, the majority of the available PMW SD models for WH locations are developed using limited data and fewer parameters and therefore cannot be implemented for the entire region. Further, these models have not taken the auxiliary parameters such as location, topography, and snow cover duration (SCD) into consideration and have poor accuracy (particularly in deep snow) and coarse spatial resolution. Considering the high spatiotemporal variability in snow depth characteristics across the WH region, region-wise multifactor models are developed for the first time to estimate SD at a high spatial resolution of 500 m × 500 m for three different WH zones, i.e., Lower Himalayan Zone (LHZ), Middle Himalayan Zone (MHZ), and Upper Himalayan Zone (UHZ). Multifrequency brightness temperature (TB) observations from Advanced Microwave Scanning Radiometer 2 (AMSR2), SCD data, terrain parameters (i.e., elevation, slope, and ruggedness), and geolocation for the winter period (October to March) during 2012–2013 to 2016–2017 are used for developing the SD models for dry snow conditions. Different regression approaches (i.e., linear, logarithmic, reciprocal, and power) are used to develop snow depth models, which are evaluated further to find if any of these models can address the heterogeneous association between SD observations and PMW TB. From the results, it is observed from the analysis that the power regression SD model has improved accuracy in all WH zones with the low root mean square error (RMSE) in the MHZ (i.e., 27.21 cm) compared to the LHZ (32.87 cm) and the UHZ (42.81 cm). The spatial distribution of model-derived SD is highly affected by SCD, terrain parameters, and geolocation parameters and has better SD estimates compared to regional and global products in all zones. Overall results indicate that the proposed multifactor SD models have achieved higher accuracy in deep snowpack (i.e., SD >25 cm) of the WH region compared to previously developed SD models.

El Niño signals revealed by AMSU-A brightness temperature observations

El Niño signals revealed by AMSU-A brightness temperature observations

Inter-Calibration of Passive Microwave Satellite Brightness Temperature Observations between FY-3D/MWRI and GCOM-W1/AMSR2

Microwave sensors possess the capacity to effectively penetrate through clouds and fog and are widely used in obtaining soil moisture, atmospheric water vapor, and surface temperature measurements. Long time-series datasets play a pivotal role in climate change studies. Unfortunately, the lifespan of operational satellites often falls short of the needs of these extensive datasets. Hence, comparing and cross-calibrating sensors with similar configurations is paramount. The Microwave Radiation Imager (MWRI) onboard Fengyun-3D (FY-3D) is the latest generation of satellite-based microwave remote sensing instruments in China, and its data quality and application prospects have attracted widespread attention. To comprehensively assess the data quality of MWRI, a comparison of the orbital brightness temperature (TB) data between FY-3D/MWRI and Global Change Observation Mission 1st-Water (GCOM-W1)/Advanced Microwave Scanning Radiometer 2 (AMSR2) is conducted, and then a calibration model is established. The results indicate a strong correlation between the two sensors, with a correlation coefficient exceeding 0.9 across all channels. The mean bias ranges from −1.5 K to 0.15 K. Notably, the bias of vertical polarization is more pronounced than that of horizontal polarization. The TB distribution patterns and temporal evolutions are highly consistent for both sensors, particularly under snow and ice. The small intercepts and close-to-1 slopes obtained during calibration further demonstrate the minor data differences between the two sensors. However, the calibration process effectively reduces the existing errors, and the calibrated FY-3D/MWRI TB data are closer to GCOM-W1/AMSR2, with a mean bias approximately equal to 0 K and a correlation coefficient exceeding 0.99. The excellent consistency of the TB data between the two sensors provides a vital data basis for retrieving surface parameters and establishing long time-series datasets.

Identifying Seismic Anomalies via Wavelet Maxima Analysis of Satellite Microwave Brightness Temperature Observations

This study develops a wavelet maxima-based methodology to extract anomalous signals from microwave brightness temperature (MBT) observations for seismogenic activity. MBT, acquired via satellite microwave radiometry, enables subsurface characterization penetrating clouds. Five surface categories of the epicenter area were defined contingent on position (oceanic/terrestrial) and ambient traits (soil hydration, vegetal covering). Continuous wavelet transform was applied to preprocess annualized MBT readings preceding and succeeding prototypical events of each grouping, utilizing optimized wavelet functions and orders tailored to individualized contexts. Wavelet maxima graphs visually portraying signal intensity variations facilitated the identification of aberrant phenomena, including pre-seismic accrual, co-seismic perturbation, and postseismic remission signatures. The casework found 10 GHz horizontal-polarized MBT optimally detected signals for aquatic and predominantly humid/vegetative settings, whereas 36 GHz horizontal-polarized performed best for arid, vegetated landmasses. Quantitative machine learning methods are warranted to statistically define selection standards and augment empirical forecasting leveraging lithospheric stress state inferences from sensitive MBT parametrization.

Improving the representation of the atmospheric boundary layer by direct assimilation of ground‐based microwave radiometer observations

AbstractIn a joint effort, MeteoSwiss and Deutscher Wetterdienst (DWD) address the need for improving the initial state of the atmospheric boundary layer (ABL) by exploiting ground‐based profiling observations that aim to fill the existing observational gap in the ABL. We implemented brightness‐temperature observations from ground‐based microwave radiometers (MWRs) in our data assimilation systems using a local ensemble transform Kalman filter (LETKF) with Radiative Transfer for TIROS Operational Vertical Sounder, ground‐based (RTTOV‐gb) as a forward operator. We were able to obtain a positive impact on the brightness temperature first guess and analysis, as well as a slight impact on the ABL humidity, using two MWRs at MeteoSwiss. These results led to a subsequent operational implementation of the observing system at MeteoSwiss. Furthermore, we performed an extensive set of assimilation experiments at DWD to investigate further various aspects such as the vertical localisation of selected single channels. We obtained a positive impact on the 6‐hr forecast of ABL temperature and humidity by assimilating two channels employing a dynamical localisation based on the sensitivity functions of RTTOV‐gb but also with a static localisation in a single‐channel setup. Our experiments indicate the importance of vertical localisation when using more than one channel, although reliable improvements are challenging to obtain without a larger number of observations for both assimilation and verification.

Retrieval of Desert Microwave Land Surface Emissivity Based on Machine Learning Algorithms

Based on the community radiative transfer model, ensemble tree-based random forest algorithm, and extreme gradient boosting tree algorithm, this study established a random forest retrieval model (RF) and an extreme gradient boosting tree retrieval model (XGBoost) for the microwave land surface emissivity by using ERA5 reanalysis data and the observed brightness temperature of 10.65 GHz vertical polarization from FY3C Microwave Radiation Imager-I. In addition, an optimized Bayesian regularized neural network retrieval model (M2_30NN) has also been established on the basis of the original neural network land surface emissivity retrieval model (M1_20NN). The results show that compared with the simulated brightness temperature of the original land surface emissivity, the simulated brightness temperature of the land surface emissivity from each retrieval model is not only significantly improved in the correlation coefficient with the observed brightness temperature (5.92% (M1_20NN), 4.23% (M2_30NN), 14.21% (RF), 13.07% (XGBoost)), but also in the evaluation indexes of root mean square error, mean absolute error and explained variance regression score in the training datasets. Furthermore, in terms of testing datasets and spatiotemporal independence test datasets, the retrieval results of RF and XGBoost models can capture the spatial distribution patterns that are consistent with observations well, and also show great numerical improvement compared with the original model. In general, the XGBoost retrieval model is the best, followed by the RF retrieval model.

Estimation of Lightning Flash Rate in Precipitation Features by Applying Shallow AI Neural Network Models to the GMI Passive Microwave Brightness Temperatures

AbstractThe satellite‐constellation passive‐microwave Brightness Temperature (TB) observations, with global coverage, and more additions from upcoming CubeSats, have been mainly used in surface precipitation retrievals. However, these observations can also be used to indicate the intensity of convective systems. This study attempts to relate Global Precipitation Mission (GPM) Microwave Imager (GMI) TBs to Geostationary Lightning Mapper (GLM) lightning flashes by using 4 years (02/2018–04/2022) of GPM Precipitation Feature (PF) database. GMI TBs are collocated to the GLM lightning counts, and to the ERA5 reanalysis 2‐m air temperature in PFs. Three Artificial Intelligent Neural Network Models (AI‐NN) are trained to classify PFs producing lightning in a 20‐min window respectively for land, ocean, or coast. The flash rate of the determined Lightning producing PFs (LPFs) is then quantified by three other AI‐NNs, each trained for one of the three regions. Though the models clearly capture the global geographical distribution of LPFs with a Probability Of Detection over 90%, high False Alarm Rates are found, ranging from 49.9% over land to 91.5% over the ocean. The importance of TB at each passive microwave channel varies regionally, corresponding to the different microphysical properties in various types of precipitation systems. The global lightning distribution is derived by applying the AI models to global PFs and is well compared to the lightning climatology from Lightning Imaging Sensor and Optical Transient Detector. This suggests that the use of passive microwave TBs can help to fill the gaps in lightning monitoring thanks to their global coverage.

A Diagnostic Study on the Formation of Double Eyewalls of Typhoon Lekima (2019) Based on POES Microwave and GOES Infrared Observations

AbstractMicrowave observations of brightness temperature (TB) from multiple polar‐orbiting operational environmental satellite imagers and infrared observations of geostationary operational environmental satellite Himawari‐8 imager are used to investigate characteristic features of secondary eyewall evolution of Typhoon Lekima (2019). It is found that microwave observations showed a longer duration time of concentric double eyewalls than Advanced Himawari Imager (AHI) infrared observations, due to a stronger cloud penetration capability of microwave than infrared observations. The AHI infrared observations with high temporal and horizontal resolutions captured several actively propagating inner spiral rainbands before the secondary eyewall formation (SEF) of Lekima (2019). They are characterized by a wavenumber‐1 asymmetry propagating radially outward and wavenumbers‐2 and ‐3 asymmetries propagating anticlockwise. The wavenumber‐2 asymmetry approximates an azimuthal phase speed of wavenumber‐2 vortex Rossby wave. Later, an outer spiral rainband experienced an obvious axisymmetrization, then merged with multiple inner spiral rainbands propagating radially outward from the primary eyewall. The stagnation radius of these inner spiral rainbands coincides with the location of the highly symmetric outer spiral rainband, which accelerated the SEF. It is suggested that the SEF of Typhoon Lekima (2019) results from a concerted activity of inner and outer spiral rainbands.

Deep learning estimation of northern hemisphere soil freeze-thaw dynamics using satellite multi-frequency microwave brightness temperature observations.

Satellite microwave sensors are well suited for monitoring landscape freeze-thaw (FT) transitions owing to the strong brightness temperature (TB) or backscatter response to changes in liquid water abundance between predominantly frozen and thawed conditions. The FT retrieval is also a sensitive climate indicator with strong biophysical importance. However, retrieval algorithms can have difficulty distinguishing the FT status of soils from that of overlying features such as snow and vegetation, while variable land conditions can also degrade performance. Here, we applied a deep learning model using a multilayer convolutional neural network driven by AMSR2 and SMAP TB records, and trained on surface (~0-5 cm depth) soil temperature FT observations. Soil FT states were classified for the local morning (6 a.m.) and evening (6 p.m.) conditions corresponding to SMAP descending and ascending orbital overpasses, mapped to a 9 km polar grid spanning a five-year (2016-2020) record and Northern Hemisphere domain. Continuous variable estimates of the probability of frozen or thawed conditions were derived using a model cost function optimized against FT observational training data. Model results derived using combined multi-frequency (1.4, 18.7, 36.5 GHz) TBs produced the highest soil FT accuracy over other models derived using only single sensor or single frequency TB inputs. Moreover, SMAP L-band (1.4 GHz) TBs provided enhanced soil FT information and performance gain over model results derived using only AMSR2 TB inputs. The resulting soil FT classification showed favorable and consistent performance against soil FT observations from ERA5 reanalysis (mean percent accuracy, MPA: 92.7%) and in situ weather stations (MPA: 91.0%). The soil FT accuracy was generally consistent between morning and afternoon predictions and across different land covers and seasons. The model also showed better FT accuracy than ERA5 against regional weather station measurements (91.0% vs. 86.1% MPA). However, model confidence was lower in complex terrain where FT spatial heterogeneity was likely beneath the effective model grain size. Our results provide a high level of precision in mapping soil FT dynamics to improve understanding of complex seasonal transitions and their influence on ecological processes and climate feedbacks, with the potential to inform Earth system model predictions.

Improving AHI Radiance Assimilation over Land with the Surface Skin Temperature Constrained by Station Observations and Its Impact for Quantitative Precipitation Forecasts

Abstract In this study, a new way to assimilate clear-sky Advanced Himawari Imager (AHI) surface-sensitive brightness temperature (TB) observations over land is investigated for improving quantitative precipitation forecasts (QPFs) in eastern China. To alleviate problems arising from inaccurate surface temperature in radiance simulations, surface station observations of land surface skin temperature (LSST) together with conventional and AMSU-A observations are assimilated to improve AHI surface-sensitive TB simulations of the Community Radiative Transfer Model (CRTM) before AHI data assimilation. First, the Gridpoint Statistical Interpolation (GSI) three-dimensional variational (3DVar) system is updated with the additional control variable of surface temperature and its background error covariances. Second, surface temperature and emissivity sensitivity checks are designed for the quality control of the surface-sensitive AHI channels. Finally, the impacts of a two-time data assimilation strategy are assessed for a local convection rainfall case and a synoptic-scale precipitation case. The experiment in which AHI data are assimilated after assimilating LSST data (ExpL2) outperforms the traditional experiment in which the LSST is not updated (ExpL) in terms of its 24-h QPF skill score. A better description of atmospheric instability and moisture convergence forcing is obtained in ExpL2 than in ExpL. Both experiments show additional low-level temperature and humidity adjustments compared to the experiment that does not assimilate AHI surface-sensitive channels (ExpNL). Lower AHI TB simulation biases are found in the ExpL2 experiment, which improve the analyzed field and subsequent QPFs. The results in this study suggest the importance of proper utilization of LSST observations for AHI surface-sensitive TB assimilations over land.

Global Mapping of Microwave Emissivity and Emitting Depth in Arid Areas Using GMI Observations

AbstractOver arid areas, observations of brightness temperatures by passive microwave radiometers are affected by the variation of the emitting depth with wavelengths. When this variation is unaccounted for, it limits the assimilation of passive microwaves over deserts in Numerical Weather Prediction models and it causes large errors in passive microwave retrievals of land surface temperatures. The emitting depths, along with the corresponding emissivities, are estimated from 10 to 89 GHz, using the non‐Sun‐synchronous observations of the Global Precipitation Mission Microwave Imager to reconstruct the monthly diurnal cycle of brightness temperature. The soil temperature profile is modeled using a two‐term Fourier decomposition based on the ERA5 surface temperature. The combination of the observation and the modeled temperature allows for an estimation of the microwave effective emitting depth. The emitting depth is estimated to be up to 25 cm at 36 GHz, resulting in large differences between the surface temperature and the effective emitting temperature. The variation of emitting depth with frequency is parameterized, and a companion data set provides the necessary parameters to calculate the emitting depth for arid areas between 10 and 89 GHz, globally. The benefit of this parameterization is quantified, with an application to the modeling of observations from the Special Sensor Microwave Imager Sounder over arid areas.

IMERG Precipitation Improves the SMAP Level-4 Soil Moisture Product

Abstract The NASA Soil Moisture Active Passive (SMAP) mission Level-4 Soil Moisture (L4_SM) product provides global, 9-km resolution, 3-hourly surface and root-zone soil moisture from April 2015 to the present with a mean latency of 2.5 days from the time of observation. The L4_SM algorithm assimilates SMAP L-band (1.4 GHz) brightness temperature (Tb) observations into the NASA Catchment land surface model as the model is driven with observation-based precipitation. This paper describes and evaluates the use of satellite- and gauge-based precipitation from the NASA Integrated Multi-satellitE Retrievals for the Global Precipitation Measurement (IMERG) products in the L4_SM algorithm beginning with L4_SM Version 6. Specifically, IMERG is used in two ways: (i) The L4_SM precipitation reference climatology is primarily based on IMERG-Final (Version 06B) data, replacing the Global Precipitation Climatology Project Version 2.2 data used in previous L4_SM versions, and (ii) the precipitation forcing outside of North America and the high latitudes is corrected to match the daily totals from IMERG, replacing the gauge-only, daily product or uncorrected weather analysis precipitation used there in earlier L4_SM versions. The use of IMERG precipitation improves the anomaly time series correlation coefficient of L4_SM surface soil moisture (versus independent satellite estimates) by 0.03 in the global average and by up to ∼0.3 in parts of South America, Africa, Australia, and East Asia, where the quality of the gauge-only precipitation product used in earlier L4_SM versions is poor. The improvements also reduce the time series standard deviation of the Tb observation-minus-forecast residuals from 5.5 K in L4_SM Version 5 to 5.1 K in Version 6. Significance Statement Soil moisture links the land surface water, energy, and carbon cycles. NASA Soil Moisture Active Passive (SMAP) satellite observations and observation-based precipitation data are merged into a numerical model of land surface water and energy balance processes to generate the global, 9-km resolution, 3-hourly Level-4 Soil Moisture (L4_SM) data product. The product is available with ∼2.5-day latency to support Earth science research and applications, such as flood prediction and drought monitoring. We show that a recent L4_SM algorithm update using satellite- and gauge-based precipitation inputs from the NASA Integrated Multi-satellitE Retrievals for the Global Precipitation Measurement (IMERG) products improves the temporal variations in the estimated soil moisture, particularly in otherwise poorly instrumented regions in South America, Africa, Australia, and East Asia.