A new approach is developed to estimate atmospheric microwave radiation parameters (atmospheric radiation$T_{a}$and its optical thickness$\tau $) over the Arctic sea ice (SI) at the frequencies of the Advanced Microwave Scanning Radiometer 2 (AMSR2). The approach is based on the results of numerical modeling of the brightness temperature (BT) of vertically and horizontally polarized microwave radiation of the Arctic SI–atmosphere system under nonscattering conditions. The main point of the approach is the retrieval of$\tau $at 89 GHz ($\tau _{89}$) from BT polarization difference (PD) at 89 GHz under the assumption of constant SI emission PD at 89 GHz.$T_{a}$and$\tau $at the other AMSR2 frequencies and$T_{a}$at 89 GHz are estimated using$\tau _{89}$and atmospheric water vapor column (WVC), based on the results of numerical modeling and neural networks (NNs)-based inversion algorithms. Since the retrievals of WVC can be done over SI from the AMSR2 data with the previously developed algorithm, the suggested approach allows$T_{a}$and$\tau $retrieval at the AMSR2 channel frequencies using only AMSR2 measurements without any additional data. The method is restricted by the use of the “absorption only” form of the radiative transfer equation (RTE) and is not applicable in the months from April to October when cloud and precipitation scattering start to play an important role in microwave radiation transfer. The method allows estimating the SI effective emissivities from BT measurements by means of the atmospheric correction without any reanalysis or operational analysis data. It may be used both in SI surface property studies and in the determination of the tie points in the SI concentration (SIC) retrieval algorithms. The uncertainties of$T_{a}$and$\tau $estimation greatly depend on the error of the assumption of constant SI emission PD at 89 GHz.

In this study we have verified the algorithms for the Arctic sea ice concentration (SIC) retrieval from the data of the Russian satellite microwave imager/sounder MTVZA-GYa. The algorithms are based on the polarization differences (PD), measured by the MTVZA-GYa at the frequencies of 10.6 and 36.7 GHz. Verification is done with the SIC satellite product based on the data of the Advanced Microwave Scanning Radiometer 2 (AMSR2) produced with an advanced algorithm from PD measurements at 89 GHz. Verification is done using collocated in time and space AMSR2 and MTVZA-GYa measurements taken in 2020. Analysis of additional high resolution Sentinel-1 Synthetic Aperture Radar (SAR) data have shown that the usage of low frequency MTVZA-GYa PD measurements allow avoiding underestimation of SIC for some areas featured to high frequency AMSR2 PD measurements.

In this paper we present the algorithms for the atmospheric water vapor column (WVC) retrieval from the data of the Russian satellite microwave imager/sounder MTVZA-GYa. The algorithms are based on the numerical modeling of the brightness temperatures (BT) of the microwave radiation of the atmosphere-ocean system measured by MTVZA-GYa at the frequencies from 10.6 to 36.7 GHz. Recently developed models of microwave radiation interaction with environment are used for BT calculations for a wide range of environmental conditions. Two Neural Networks (NNs) are trained to solve an inverse problem of WVC retrieval from BT values - global NN and tropical NN. The global NN explores all the scanner frequency channels from 10.6 to 36.7 GHz. The tropical NN is based on the measurements at 10.6 and 18.7 GHz. This NN is applicable under conditions of high atmospheric water and specifically in Tropical Cyclones.

In this study we use numerical modeling of the atmosphere-ocean system microwave radiation to estimate the atmospheric upwelling and downwelling radiation along with its optical thickness at the frequencies of the Advanced Scanning Microwave Radiometer 2 (AMSR2) using the values of the total atmospheric water vapor content (WVC) and total cloud liquid water content (CLW). Numerical modeling is based on the solution of the radiation transfer equation under approximations of microwave range. Calculations are done for the database of the atmospheric parameter profiles complemented with the model profiles of cloud liquid water content. The results of model calculations allow parameterizing the atmospheric radiation parameters as functions of WVC and CLW, which can be retrieved from the AMSR2 measurement data under various conditions including Tropical Cyclones (TCs). These atmospheric parameters can be further used to derive the oceanic response to high winds, associated with TCs, from the AMSR2 data.

In this study we demonstrate high potential of synergistic satellite data usage to investigate the Arctic saline sea ice. A simple geometrical optics model is explored to show the possibility to derive the First Year sea ice parameters best fitting the experimental data. Satellite Sentinel-1 Synthetic Aperture Radar (SAR) data are used to analyze the mesoscale sea ice structure and detect the age of the sea ice samples. Sentinel-3 altimeter data along with the Metop Advanced scatterometer (ASCAT) measurements of the normalized radar cross section (NRCS) are jointly used to build the NRCS dependencies on the incidence angle for the selected sea ice samples. Sea ice drift fields are used to investigate only those regions where the drift is negligible on a single day scale. It is shown that the suggested approach presents a powerful tool to study the Arctic sea ice properties.

An advanced method for sea ice concentration (SIC) retrieval, based on measurements of the Advanced Microwave Scanning Radiometer 2 (AMSR2), is validated. An advancement relates to tie point choice for the polarization difference in satellite measurements over open water. The approach is based on the results of physical modeling of the sea ice - ocean - atmosphere system and analysis of the AMSR2 measurements in the Arctic region. The validation is done for summer months using MODIS Sea Ice Extent data as a source of ground-truth information. Areas of 100% open water and 100% sea ice are compared for the MODIS Sea Ice Extent and the AMSR2 retrieved SIC. It is found that the AMSR2 algorithm systematically underestimates 100% SIC as compared with the MODIS 100% SIC in summer months.

An advanced algorithm for atmospheric water vapor column (WVC) retrieval from the Advanced Microwave Scanning Radiometer (AMSR) measurements over the Arctic sea ice (SI) and open ocean waters is presented. The algorithm is built on the physical modeling of the brightness temperature (BT) of the microwave radiation of the SI-open ocean-atmosphere system at the AMSR frequencies and polarizations. The BTs are calculated using a data set of the SI, atmospheric, and oceanic parameters changing in the range of their natural variability in the Arctic, and using the SI microwave emission coefficients varied according to the published experimental data. The inverse operator explores neural networks (NNs), trained on an ensemble of modeled BTs. The algorithm is applied both to the AMSR-E and to the AMSR2 measurement data. Validation of the algorithm is performed with radiosonde (r/s) WVC measurements from the four Arctic coastal stations at different SI conditions during 2014-2017. The results of the application of the new algorithm to satellite radiometer measurements are also compared with the Era-Interim reanalysis WVC, as well as with other satellite WVC products, based on the data of the Moderate Resolution Imaging Spectrometer (MODIS) and on the data of the Advanced Microwave Sounding Unit-B (AMSU-B) for 2008 and 2015. To justify the usage of the Era-Interim WVC as a reference data set for the algorithm accuracy estimation in the Arctic area, Era-Interim WVC is also compared with the r/s WVC measurements.

This work is aimed at the investigation of the application of general machine learning (ML) algorithms, used typically for image recognition, to ocean surface type categorizing, using Synthetic Aperture Radar (SAR) images as input data. The study is focused on the Convolutional Neural Network (CNN) based image segmentation. Its main goal is building the fully automatic pipeline, calculating (predicting) ocean surface (particularly ice) type from a SAR image.It is shown that the developed system has classification accuracy of up to 90% without any preliminary feature selection or region-based limitations. The advantages and drawbacks of the suggested approach are discussed.Sea ice - sea water discrimination and sea ice concentration estimation from Sentinel-1 SAR images are considered in more details due to their general availability and extensive coverage of the Arctic region.

This work describes the operational geographic information system - Arctic portal - developed at the Satellite Oceanography Laboratory of the Russian State Hydrometeorological University and a new, created with this instrument, database of the Polar lows (PL) and less intensive mesocyclones (MC) observed over the entire Arctic region for the period of 2017-2018. The database is created by the synergistic use of satellite data of spectral and microwave radiometers and scatterometers implemented in the Arctic portal allowing most reliable PL and MC identification. This database presents a complicated structure of PL and MC parameters for the whole period of their lifetime. The database is of particular interest as it covers not only the Northern seas, but also the Eastern sector of the Russian Arctic, where PL events are poorly investigated, and where the formation of the MCs has become possible due to recent sea ice retreat.

The work presents a verification of the Weather Research and Forecasting (WRF) model simulations of the surface wind speed (SWS) and track for the eight Polar lows (PL) observed in the Nordic seas in 2009-2010. As compared to the measurements of satellite Advanced Scatterometer (ASCAT) and Advanced Scanning Microwave Radiometer (AMSR-E) the model reproduces well the occurrence and general direction of the PL propagation, but systematically underestimates the mean SWS. Besides, the simulated PL tracks often deviate from those observed from the satellites by up to several hundreds of kilometers. A number of numerical experiments have been conducted to analyze the effect of using different resolutions, parameterization schemes, and dates of the simulation start.

New empirical geophysical model functions (GMFs) have been developed to interpret 89-GHz brightness temperature measurements over cold Arctic seas. Careful data screening is applied to the Advanced Microwave Scanning Radiometer 2 (AMSR2) multifrequency measurements to exclude atmospheric scattering and large absorption impacts, to estimate the 89-GHz sea surface emissivity and its relative changes with surface wind speed (SWS). Matched-up wind speeds are directly derived from the AMSR2 low-frequency measurements. The GMFs are obtained from the AMSR2 level 1R 89 GHz measurements corrected for the atmospheric impact with physical models and atmospheric parameters derived from the AMSR2 data. A carefully selected database encompasses mostly cold sea conditions (<;4 °C) and a large range of wind speed conditions, including extratropical cyclone and polar low high-wind events over the Nordic Seas. Resulting GMFs contrast with previously proposed ones manifesting larger SWS signal at horizontal and positive SWS signal at vertical polarization.

In present study Cold-Air Outbreaks (CAOs) associated with storm winds and extreme air-sea exchange conditions are diagnosed over the Chukchi Sea during autumn-winter periods (Sept-Dec) 2014-2016 using multisensor satellite measurements and high-resolution reanalysis dataset. Analysis of the fields of geophysical parameters from satellite and model data showed that in the autumn-winter period cold-air outbreaks are systematically (on average 2 cases/month) observed over the Chukchi Sea with the near sea surface wind speed W ≥ 20 m/s. At the end of November - early December 2014 and 2015 anomalously prolonged (7-13 days) CAOs were identified, accompanied by severe/storm winds and waves, rapid formation and drift of sea ice. In both cases, the cold advection from the polar ice cap was caused by a system of extratropical cyclones over the Bering Sea and the Gulf of Alaska, which strengthens the Aleutian Low, and the Arctic anticyclone. Unlike typical CAOs over the Northwestern Pacific Ocean, covering the lower troposphere layer up to 3-5 km, intense cold advection over the Chukchi Sea was noted mainly in the boundary layer of the atmosphere below 1 km.

In this study Era-Interim re-analysis data on the atmospheric total water vapor content (WVC), total cloud liquid water content (CWC), ice water content (IWC) and profiles of the atmospheric meteorological parameters for the whole 2015 year are used to analyze WVC, IWC and CWC seasonal variability over the sea ice and sea water surfaces and estimate the influence of the atmospheric integrated water parameters on the total microwave radiation of the sea ice and sea water - atmosphere system at the frequencies and polarizations of the Advanced Microwave Scanning Radiometer 2 (AMSR2). The brightness temperature (BT) of the sea ice - atmosphere system microwave radiation for the AMSR2 channel characteristics is modeled for the conditions of non-scattering atmosphere. The arrays of BTs are calculated using published data on the sea ice emissivities and Era-Interim re-analysis data on the sea ice concentration (SIC) and temperature.

The performance of weather filters in satellite passive microwave sea ice concentration retrieval algorithms is analyzed under conditions of high winds. The analysis is based both on numerical modeling results and on the measurements of the Advanced Microwave Scanning Radiometer 2 (AMSR2) over high wind areas. It is shown that when the threshold values for the gradient ratios are used to classify measurements as being taken over open water, sea surface wind speed geophysical model function (GMF) uncertainty plays an important role in such weather filtering. The GMF uncertainty at higher AMSR2 frequencies is mostly pronounced at high winds. The gradient ratios are calculated using the radiative transfer modeling for the conditions of non-precipitating atmosphere and an ensemble of atmospheric meteorological profiles combined with ocean parameter data. The threshold values for the gradient ratios are set up basing on numerical calculations.

Numerical simulation of brightness temperatures of the Advanced Microwave Scanning Radiometer 2 (AMSR2) over the sea ice-open ocean-atmosphere system is fulfilled for non-precipitating conditions using a database of atmospheric meteorological parameter profiles, model profiles of cloud liquid water and published experimental data for sea ice emissivities. The results of numerical experiment show that polarization difference (PD) at high frequency channels can help to discriminate between the sea ice and sea water. At that high wind speeds and optically thick atmospheres lead to misinterpretation of sea ice basing on the only PD values. This complicates sea ice concentration retrievals under bad weather conditions, associated with high winds and atmospheres with high values of total atmospheric absorption. In this study we use numerical simulation results to define the weather limits of applicability of the current sea ice concentration algorithms based on satellite passive microwave PD measurements.

Polar lows (PLs), emerging over the sea ice edge, are studied using multisensor data information, surface analysis maps and reanalysis data. PLs over the Western (the Greenland, the Norwegian and the Barents Seas), and the Eastern (the Chukchi, the East Siberian and the Laptev Seas) parts of the Arctic are considered. It is shown that currently operating satellite instruments, taken separately, cannot provide means for PL confident detection. Whereas multisensor satellite data, including passive and active microwave data products along with infrared and visible measurements, analyzed simultaneously, provide an opportunity to monitor PLs, developed under most environmental conditions, including the vicinity to the sea ice edge. Geophysical parameters are retrieved from satellite passive microwave data with advanced algorithms. PLs are detected using the complex geo-informational system, making advantage from simultaneous analysis of different satellite and model data.

With the sea-ice decline over the eastern part of the Eurasian Arctic (EEA), polar mesocyclones (MCs) and their most intensive representatives-polar lows (PLs)-can occur over more open-water areas. Visible and infrared MODIS images, active and passive microwave spaceborne instruments, and ERA Interim reanalysis data are combined and used to analyze the synoptic situations and to infer the factors influencing MC appearance and evolution over the Kara Sea, the Laptev Sea, the East Siberian Sea, and the Chukchi Sea. In recent years, the Arctic more often loses its summer sea-ice cover, and PLs may more commonly emerge within open-water Eastern Arctic regions during fall and summer months. This conclusion is derived basing on the analysis of more than 150 MCs for August, September, October, and November months over the EEA for the period of 2003-2015. PL and MC distribution and characteristics necessitate to be studied with a multisensor approach since neither reanalysis data nor data from a single instrument can provide sufficient means for their confident detection and analysis.

Measurements of the Advanced Microwave Scanning Radiometer 2 (AMSR2) onboard the Global Change Observation Mission-Water 1 (GCOM-W1) satellite at 6.925 and 7.3 GHz and both linear polarizations over tropical cyclones (TCs) during 2012-2014 are used to derive a new geophysical function relating the brightness temperature to the sea surface wind speed (SWS) in extreme conditions. Similar sensitivity to the SWS at close C-band frequencies allowed correcting for the atmospheric contributions to the microwave radiance and estimating the brightness temperature (TB) at the surface under TCs, combining theoretical modeling and measured TBanalyses. Estimated oceanic TB's were regressed against the wind speeds from the Best Track Archive to derive a new geophysical model function for the wind speed excess emissivity at AMSR2 C-band microwave frequencies.

A new method for radio-frequency interference (RFI) contamination identification over open oceans for the two C-subbands and X-band of Advanced Microwave Scanning Radiometer 2 (AMSR2) channel measurements is suggested. The method is based both on the AMSR2 brightness temperature (TB) modeling and on the analysis of AMSR2 measurements over oceans. The joint analysis of TBspectral differences allowed to identify the relations between them and the limits of their variability, which are ensured by the changes in the environmental conditions. It was found that the constraints, based on the ratio of spectral differences, are more regionally and seasonally independent than the spectral differences themselves. Although not all possible RFI combinations are considered, the developed simple criteria can be used to detect most RFI-contaminated pixels over the World Ocean for AMSR2 measurements in two C-subbands and the X-band.

An approach for detecting and tracking polar lows (PLs) is developed based on satellite passive microwave data from two sensors: Special Sensor Microwave Imager (SSM/I) on board the Defense Meteorological Satellite Program satellite and Advanced Microwave Scanning Radiometer-Earth Observing System (AMSR-E) on board the Aqua satellite. This approach consists of two stages. During the first stage, the total atmospheric water vapor fields are retrieved from SSM/I and AMSR-E measurement data using precise Arctic polar algorithms, applicable over open water and having high retrieval accuracies under a wide range of environmental conditions previously developed. During the second stage, the vortex structures are detected by visual analysis in these fields, and PLs are identified and tracked. A few case studies are comprehensively conducted based on multisensor data usage. SSM/I and AMSR-E measurements and other satellite data, including visible, infrared, and synthetic aperture radar images, scatterometer wind fields, surface analysis maps, and reanalysis data, have been used for PL study. It has been shown that multisensor data provide the most complete information about these weather events. Through this, advantages of satellite passive microwave data are demonstrated.

New algorithms for total atmospheric water vapor content (Q) and total cloud liquid water content (W) retrieval from satellite microwave radiometer data, based on neural networks (NNs) and applicable for high-latitude open-water areas, were developed. For algorithm development, a radiative transfer equation numerical integration was carried out for Special Sensor Microwave/Imager (SSM/I) and Advanced Microwave Scanning Radiometer-Earth Observing System (AMSR-E) channel characteristics for nonprecipitating conditions over the open ocean. Sets of sea surface temperatures less than 15°C, surface winds, and radiosonde (r/s) reports collected by Russian research vessels served as input data for integration. It was shown that NNs perform better than the conventional regression techniques.Qretrieval algorithms were validated both for the SSM/I and AMSR-E instruments using satellite radiometric measurements collocated in space and time with polar station r/s data. The resulting SSM/I and AMSR-E retrieval errors proved to be 1.09 and 0.90 kg/m2correspondingly. For SSM/IQretrievals, the algorithms were compared with the Wentz global operational algorithm. This comparison demonstrated the advantages of NN-based polar regional algorithms in comparison with the Wentz global one. The retrieval errors proved to be 1.34 and 1.90 kg/m2( ~ 40% worse) for the NN and Wentz algorithms correspondingly.

Satellite and in situ data were examined for insights into the behavior of water vapor, cloud liquid water and wind speed during formation and evolution of synoptic-scale and mesoscale cyclones and cold air outbreaks - weather systems, which are usually accompanied by gale winds and intensive air-sea interaction. Satellite measurements carried out at visible, infrared and microwave ranges were collected over the Northern Pacific and Northern Atlantic Oceans in winter allowed investigating both the large-scale structural features (the main and the secondary fronts, etc.) and the small- and fine- scale details of the frontal boundaries, organized convection in the marine boundary layer of the atmosphere, etc. Multisatellite approach improved temporal resolution and the possibility to trace the location and characteristics of weather systems including fast moving and fast evolving systems.