Leads are linear fractures formed by the deformation of the ice cover and are crucial areas for Arctic heat exchange between the ocean and the atmosphere. Accurately identifying leads is essential for studying climate change, polar ecosystems, and shipping navigation. Recent research on leads detection methods developed using Synthetic Aperture Radar (SAR) tend to ignore the developmental stage of the leads and the distribution context of leads (first-year ice or multiyear ice), which hinders the generalizability of these methods. In this paper, we propose a Stacking Approach for Lead Classification (SALC) integrating five specialized learners for five typical lead conditions. Ablation experiments reveal the effectiveness of each component of SALC: 1) the optimal preprocessing method is to input the incidence angle as a feature and apply the advanced thermal noise removal. In addition, the effect of inputting the incidence angle is significantly stronger than incidence angle correction for original images; 2) the SALC performs best when the data features include polarisation, texture, and upscaling features; 3) XGBoost is the most suitable meta classifier as compared to SVM and RF. The results show that SALC outperforms both single learners and traditional classifiers across various conditions and quality criteria. In addition, SALC can tolerate at least 15% of erroneous samples.

Integrated retrieval using the optimal estimation (OE) method iteratively finds a set of geographical parameters that best match the observations. However, this method becomes more challenging over the ice surface due to the highly sensitive parameters such as sea ice concentration (SIC) and multiyear ice concentration (MYIC). In this study, a new time constraint that captures the distinct temporal characteristics of SIC and MYIC is incorporated into the OE method. The integrated retrievals, using both the original and time-constraint OE method (referred to as OE and OE-Z, respectively), were conducted based on FengYun-3D (FY-3D) microwave radiation imager (MWRI) data. Compared to other radiometer-based SIC and MYIC products, OE-Z outperforms OE, with the correlations increasing from 0.91 to 0.96 for SIC and from 0.41 to 0.49 for MYIC. The time constraint in OE-Z effectively mitigates the anomalous retrievals in SIC and MYIC, resulting in smoother and more reasonable time series than OE. Improvements in SIC and MYIC lead to enhanced simulation of surface microwave emission, thus improving the retrieval of atmospheric parameters. In comparison with the MOSAiC total water vapor (TWV) measurements, the RMSE in OE-Z reduces from 1.72 to 1.66 kg/m2, and the correlation increases from 0.46 to 0.50. The simulated brightness temperature (TB) biases in OE-Z reduce from 0.71 to 0.31 K at 36 GHz and from −8.95 to −7.72 K at 89 GHz. This emphasizes the importance of imposing suitable constraints on highly sensitive parameters in integrated retrieval.

Antarctic sea ice is generally considered to be free from surface melt, as opposed to the Arctic sea ice where surface melt is prevalent during summer. In this work, we present recent high-resolution Sentinel-2 images which confirm the presence of melt ponds also on the Antarctic sea ice. We employ an optical melt pond fraction retrieval applied to Sentinel-3 data to investigate occurrences of surface melt in the vicinity of Lützow-Holm Bay where sea ice ponding has been confirmed in situ. Sea ice surface melt has the potential to disrupt passive microwave sea ice concentration product. We focus on an area of land fast ice with confirmed 100% ice concentration to highlight this effect on real data.

Understanding the Antarctic sea ice optical properties is the key to the accurate assessment of the radiative balance of the region. Although high resolution optical satellite data have recently become more available also over sea ice, most Antarctic remote sensing research is dedicated to land ice and glaciers, with sea ice classification studies still lacking. In the context of recent change with the two last years 2022 and 2023 showing record low Antarctic sea ice extent, high resolution sea ice optical properties are key to understanding the Antarctic sea ice evolution. In this work, we present high resolution surface classification based on the remote sensing Sentinel-2 MSI data. We retrieve sea ice albedo, fraction of blue ice and impurities. A case study scene features variable sea ice albedo, where pure sea ice alternates with darker sea ice featuring yellow substance contamination and potential surface melt.

Contrary to the sea ice surface melt during Arctic summer, melt ponds on top of the Antarctic sea ice during austral summer are thought to be rare. Their observations are limited to few occasional sightings and no regular dataset is published till now. In this work, we employ an existing remote sensing method which was already applied in the Arctic and for the first time detect potential surface melt areas on top of the Antarctic sea ice at a regional to global scale using Sentinel-3 data. We confirm the occurrence of melt ponds in Lützow-Holm Bay in the vicinity of the Shōwa Station observed in situ in January 2017 and detect surface melt in the following years 2018–2020. We also perform global processing of the Antarctic sea ice to detect further regions of potential surface melt.

The lack of in situ data has always posed challenges to remote-sensing-data product validation. Herein, the products of sea-ice concentration (SIC) data derived using the arctic radiation and turbulence interaction study (ARTIST) sea ice (ASI) algorithm were evaluated by comparing them with SICs from a high-resolution sea-ice aerial image obtained during the 27th China Antarctic expedition in January 2011. Results suggest that data obtained from the advanced microwave scanning radiometer for the earth-observing system (AMSR-E) underestimate SICs by 19%. We performed step-by-step comparisons among the aerial image, moderate-resolution-imaging spectroradiometer (MODIS), and AMSR-E SIC. These types of comparisons have not been made in previous validation studies. First, SICs acquired from MODIS-Terra imagery were acquired using a tie-point method and corrected by SICs derived from the aerial photography. Second, SICs of MODIS-Aqua images were trained based on the consistency of SIC results between MODIS-Terra and MODIS-Aqua over the selected region on the same day. Finally, the MODIS-Aqua SICs were employed to validate synchronous AMSR-E swath SIC products. The results show that the AMSR-E products underestimate SICs by 8.5% in the marginal ice zone in comparison with MODIS SICs. According to our further analysis between sea-ice types and AMSR-E biases, the higher the proportion of first-year ice, the smaller the AMSR-E SIC bias. In other words, results suggest that the higher the thin ice proportion, the more the AMSR-E underestimates the SIC.

Accurate water vapour measurements are crucial for a better understanding of Arctic climate change and the global hydrological cycle. In the sparsely observed polar regions, passive microwave satellite-borne sensors allow obtaining the vertically integrated water vapour content of the atmosphere (Total Water Vapour, TWV) with whole Arctic or Antarctic coverage, respectively. However, the available retrievals work only over certain surface types. Over open ocean, microwave imagers like SSM/I, AMSR-2 or AMSR-E provide daily precipitable water. Over ice and land surfaces, another method based on data of the microwave humidity sounders AMSU-B and MHS allows TWV retrieval. Here, we aim to merge these complementary TWV retrievals, creating an Arctic-wide daily dataset of 50 km resolution with seamless coverage from the high Arctic to mid latitudes from 2002 to date.

Sea ice is an important quantity for the radiation budget of the Earth's climate system. In this paper we present a comparison of SMAP and SMOS measured brightness temperatures and translate a validated retrieval algorithm for the thickness of thin sea ice from SMOS to SMAP radiometer observations. For October to December all observations of SMAP in forward and backward direction are compared as daily averages and found with low RMSD of 2.25K and 2.42K in TB,vand TB,h, respectively. The mean of forward and backward observations from SMAP are mapped to SMOS equivalent brightness temperatures using a linear regression. SMAP was found to yield lower brightness temperatures compared to SMOS by about 5K in both polarizations while difference decreases with increasing brightness temperatures. An existing ice thickness retrieval for SMOS using averaged brightness temperatures in the incidence angles range of 40° to 50° is transferred to SMAP using another linear regression. Ice thicknesses retrieved from both sensors show good agreement with a correlation of r = 0.969 and RMSD of 3.31 cm for ice thicknesses from 1 cm to 50 cm.

Currently available sea ice thickness retrieval algorithms are compromised in summer in the presence of melt. This study presents a new approach to estimate sea ice thickness in summer in the presence of melt ponds. Analysis of field data obtained during RV “Polarstern” cruise ARK27/3 (August - October 2012) has shown a clear connection of ice thickness under melt ponds to their measured spectral albedo and to saturation value in the HSL colour space from field photographs. An empirical function has been derived from the data, which gives a possibility to access sea ice thickness information stored in the historic dataset of in situ imagery. A similar function has been applied to aerial imagery. The presented approach can be used (1) as fast, non-intrusive method to estimate sea ice thickness in situ; (2) for general circulation model input, validation, ice mass balance estimate; (3) for assessing summer sea ice thickness dynamics from historic data to be used in the context of Arctic change.

Melt ponds cover up to 50% of Arctic summer sea ice and strongly affect the local radiative balance. However, they are currently poorly represented in circulation models. One of the reasons is the lack of reliable large-scale observations. A challenge encountered during melt pond retrieval from optical satellite observations is the filtering of cloud which in the visible range of spectrum look similarly bright as the sea ice surface. Here we demonstrate the influence of undetected cloud on the melt pond retrieval and suggest a masking procedure based on a Bayesian classifier trained using coincident infrared observations of AATSR. The AATSR swath only covers the central half of the MERIS swath but the masking procedure extends the filtering scheme to the outer parts of the MERIS swath so that the full swath can be used for melt pond retrieval.

Multiyear ice (MYI) characteristics can be retrieved from passive or active microwave remote sensing observations. One of the algorithms that combine both observations to identify partial concentrations of ice types (including MYI) is the Environment Canada Ice Concentration Extractor (ECICE). However, cycles of warm-cold air temperature trigger wet-dry cycles of the snow cover on MYI surface. Under wet snow conditions, anomalous brightness temperature and backscatter, similar to those of first-year ice (FYI), are observed. This leads to misidentification of MYI as being FYI, hence decreasing the estimated MYI concentration suddenly. The purpose of this paper is to introduce a correction scheme to restore the MYI concentration under this condition. The correction is based on air temperature records. It utilizes the fact that the warm spell in autumn lasts for a short period of time (a few days). The correction is applied to MYI concentration retrievals from ECICE using an input of combined QuikSCAT and AMSR-E data, acquired over the Arctic region in a series of autumn seasons from 2003 to 2008. The correction works well by replacing anomalous MYI concentrations with interpolated ones. For September of the six years, it introduces over 0.1×106km2MYI area, except for 2005. Due to the regional effect of warm air spells, the correction could be important in the operational applications where ice concentrations are crucial on small scale and mesoscale.

The influence of sea ice thickness brightness temperatures and ice concentrations retrieved from passive microwave observations is quantified, using horizontally homogeneous sea ice thickness retrievals from ESA's SMOS sensor observations at high incidence angles. Brightness temperatures are influenced by thickness below 18 cm (89GHz) and 50 cm (1.4 GHz). Ice concentration retrievals reduced by ice thickness below 0.17 m and 0.33 m, with higher frequency algorithms being less influenced.

The Automatic oil Seep Location Estimator (ASLE) described in this paper is a system that can automatically estimate the locations of potential oil seeps using SAR images. The ASLE segments dark areas in SAR images, calculates direct features related to geometry and backscatter as well as contextual features like wind speed and direction for each dark object and uses them to classify the object as either a natural oil slick or a look-alike. The classification scheme has been implemented using a rule-based approach. The slick origins are detected and clustered together spatially to detect the feeding seep locations. A preliminary dataset of 25 images from ENVISAT's Advanced Synthetic Aperture Radar (ASAR) was used to test the algorithm. The results show the addition of contextual information helps reduce false positives in automatic slick detection.

The topographic differences in German Wadden Sea between 1999 and 2006 are compared and analyzed in this paper. The waterline method is used to generate topographic maps of this area with SAR images. 13 SAR images (from ERS-2) of the year 1999 and 14 SAR images (12 from ERS-2, 2 from ENVISAT_ASAR) of the year 2006 are analyzed. The SAR images captured at different dates are able to cover different tidal phases to represent the tidal flat. The waterlines are detected by a wavelet based edge detection algorithm, assigned with water level information from the operational water level model of the German hydrographic agency, corrected with water gauge data. The waterlines with water level information are interpolated into a regular grid to represent the intertidal topography of the Wadden Sea. The topographic maps, in large scale, are stable and correct. There are some significant changes at Gelbsand, Medemsand and Tertiussand. Sediment transport can be intended from these topographic maps.

A framework for the automatic detection of natural oil seeps using Synthetic Aperture Radar (SAR) images, implemented in Python, is presented. Dark objects are detected using morphological thresholding. For each object, features are computed, which are used to classify the object as either a natural oil slick or a look-alike. The classification scheme has been implemented using a rule-based approach. The slick origins are detected and clustered together spatially, in order to detect the seep origin. A dataset of 122 images from ENVISAT's Advanced Synthetic Aperture Radar (ASAR) were used to test the algorithm. In this paper, only preliminary results are reported.

The effect of Arctic sea ice on the radiative budget is of disproportional importance relative to its areal coverage. High albedo of sea ice is the driving force for a number of climatic feedbacks in the Arctic. In order to retrieve sea ice albedo from satellite and assess the state of the Arctic climate, important preparation steps are needed. They are the surface type classification and cloud clearing of the satellite scene. The MODIS sensor features spectral channels which can be used to distinguish all kinds of surfaces from snow-like surfaces, and further - separate snow, melting snow and ice, melt ponds and leads. In situ data is used to derive threshold values characterizing each family of surface types for MODIS bands 4 and 5. Separation of all the above mentioned snow-like surfaces from clouds is not a trivial task; MODIS bands 4, 5 and 6 are used to enhance the atmospheric reflectance in the NIR and screen out clouds from snow or ice-covered scene.

The waterline method is used to derive the topography of the tidal flats along the German coast by evaluation of SAR (Synthetic Aperture Radar) images. A series of SAR images taken at different water levels from 1996 to 1999 and 2009 is analyzed with a wavelet based edge detection algorithm to identify the border between the tidal flat and adjacent water areas. The waterlines detected are combined with corresponding water-levels to represent the topography on a irregularly spaced grid. The water levels are provided from a numerical tidal model and corrected by gauge data. Topographic maps of the inter tidal zone are produced after interpolating these data into a regular grid. We reproduced the maps from 1996 to 1999 and newly processed the data from 2009.

We developed a new method to monitor the ice sheet melting periods with passive microwave measurements. Our method combined the original brightness temperature time series with those simulated by time series simulation software package TIMESAT in order to obtain a time series with similar behaviour, but less noise. A generalized Gaussian model was used to classify the pixels to wet and dry snow. Based on the steep rise and drop of the new time series, we detected the onset and end of the Antarctic ice sheet melting from 1988 to 2008. The results indicated that the whole Antarctic experienced a deceasing melting during the 20 years. The Antarctic Peninsula region exhibited a long and stable melt occurrence, in comparison with which the Ross Ice Shelf has the nearly shortest melt duration and the largest variability in melt extent.

The Soil Moisture and Ocean Salinity (SMOS) Ice project explored the potential of retrieving sea-ice information from the SMOS satellite, a polar-orbiting L-band radiometer successfully launched in November 2009. Toward this end, radiance measurements were collected over the Northern Baltic during the Pol-Ice campaign. We test a simple ice-concentration retrieval algorithm on these data and compare the results with ARTIST Sea Ice (ASI) maps derived from the Advanced Microwave Scanning Radiometer on the Earth Observing System. All operational ice-concentration algorithms are based on the same principle which, for the campaign data, reduces to a linear scaling of the radiances because, effectively, only one channel was available. Because of biases introduced by the different footprint sizes of the two radiometers (airborne and satellite), the linear flight path, and pilot selection of preferred surface type, Pol-Ice and ASI concentrations were compared using three different levels of averaging. In the first case, the individual measurements from the airborne radiometer were compared with interpolated ASI values; in the second, they were averaged over the pixels in the ASI maps; and in the third, they were averaged by binning the ASI values in 1% intervals. The correlations were 0.59, 0.67, and 0.76, respectively. Because of the unique operating principle of SMOS, each ground point will be viewed at multiple effective angles within a short time span. It is proposed to exploit this extra information by interpolating to a single effective viewing angle.

Sea ice emissivity models for the L-band frequency range are described and then tested on Pol-Ice campaign field measurements. Pol-Ice was conducted in March 2007 in the Northern Baltic in preparation for the launch of the Soil Moisture and Ocean Salinity (SMOS) instrument, a satellite microwave radiometer operating at 1.4 GHz. The campaign comprised airborne measurements from the EMIRAD L-band radiometer and the E-M Bird ice thickness meter. Because of the translucency of sea ice at L-band, it is hoped that SMOS will render information on ice thickness. Three variations on radiative transfer are used to model EMIRAD brightness temperatures from collocated E-M Bird measurements: a single plane-parallel model, an ensemble of such models, and a ridged Monte Carlo model based on geometric optics that includes both the top- and bottom-surface topography. All three models accurately account for the instrument antenna pattern, relevant to satellite-mounted radiometers that sample a large and heterogeneous area. Because of ice growth processes, salinity and, by extension, permittivity are partial functions of ice thickness; thus, the models are further refined so that permittivity varies with ice thickness, which was necessary to correctly model the polarization difference. Other issues related to effective permittivities, an intermediate quantity in the models, are discussed. Analysis of partial correlations shows that ice ridging makes a significant contribution to the measured signal, commending further study using scattering models that are more appropriate to the scale of the ridging relative to the wavelength.

WindSat has provided an opportunity to investigate the first spaceborne passive fully polarimetric observations of the Earth's surface. In this paper, Arctic sea ice was investigated. The passive polarimetric data are provided in the form of the modified Stokes vector consisting of four parameters. The first two components of the modified Stokes vector are the vertically and horizontally polarized brightness temperatures, which have been continuously measured by various radiometers over the last three decades. The third and fourth Stokes components provide in formation on the degree of polarization of the emission. In this paper, three types of analysis are carried out: spatial (maps considering different azimuth angle intervals), temporal (time series of daily averaged Stokes components over a small selected azimuth angle range), and azimuthal (variations w.r.t. the azimuth angle over selected study areas). Analysis has shown the highest brightness temperature variations for the 37-GHz third Stokes component (>; 2 K) during summer. The next highest signals were observed for the 10.7-GHz third and fourth Stokes components (>; 1 K) during summer as well. The 37-GHz fourth Stokes component exhibited the least variability (>; 1 K). Spikes of up to 2 K were identified in the time series of the 37-GHz third Stokes component during mid-January 2004 (winter) over first-year ice regions. The near-surface air temperature of the European Center for Medium-Range Weather Forecasts model data and the Special Sensor Microwave/Imager National Aeronautics and Space Administration Team ice concentrations revealed that, during these events, the surface temperatures reached near melting levels and the retrieved ice concentrations were reduced to about 80%. Moreover, these observations also showed clear evidence of first harmonic azimuthal dependence. Geophysical parameters, such as temperature and ice leads, are likely to be the causes. The larger signals which occurred during summer were identified as being related to the ice surface temperatures being near melting.

The waterline method is used to derive the topography of the tidal flats along the German coast by evaluation of synthetic aperture radar (SAR) images. A series of about 70 European Remote Sensing Satellites SAR images of the German Bight taken at different water levels within four years is analyzed to detect the borderline between tidal flats and adjacent water areas using a wavelet-based edge-detection algorithm. After geocoding, the waterlines are combined with the corresponding water levels to represent the topography on an irregularly spaced grid. The water levels are taken from a numerical tide model and corrected with the measured gauge data. Interpolation of these data into a regular grid yields a topographic map of the intertidal zone. While the general practicability of this method has been demonstrated in previous studies for smaller test areas, this paper is the first attempt to generate maps of a large area on a yearly basis.

WindSat fully polarimetric passive microwave observations, expressed in the form of the Stokes vector, were analyzed over the Antarctic ice sheet. The vertically and horizontally polarized brightness temperatures (first two Stokes components) from WindSat are shown to be consistent with previous studies. Azimuthal modulations in the third and fourth Stokes components were analyzed and related to surface topography, roughness, and snow morphology. A second harmonic sine function of the azimuth angle was used to estimate the orientation angle of snow features, such as topographic slope and sastrugi. The results show good agreement with the orientations derived in a previous study using scatterometer data at similar frequencies. Seasonal variability in the third and fourth Stokes components is discussed. A consistent pattern of response emerged for 10.7 GHz. Under winter conditions, the large contribution of multiple volume scattering causes a high and regionally varying 10.7-GHz fourth Stokes signal. Under summer conditions, surface scattering dominates and results in a high 10.7-GHz third Stokes signal. The third and fourth Stokes observations at 37 GHz were found to correspond to the smaller penetration depth at this higher frequency, resulting in a low difference between the amplitudes of summer and winter. The study demonstrates the potential of the spaceborne fully polarimetric passive microwave radiometers in monitoring the thermal and morphological properties of large ice sheets.

Surface emissivity is an essential quantity to retrieve surface and atmospheric parameters from satellite measurements. The surface emissivity of the Arctic sea ice is calculated using Advanced Microwave Scanning Radiometer-Earth Observing System (AMSR-E) radiance. The method accounts for the variation of the penetration depth in the snow-covered ice with frequency, air temperature, and sea-ice temperature. The variation of emissivity for different frequencies at different seasons is noticed, together with their correlations.

Continuous monitoring of sea ice and its changes is mainly done by passive microwave sensors on satellites. One frequently used technique of retrieving sea-ice concentrations is the Arctic Radiation and Turbulence Interaction STudy Sea Ice (ASI) algorithm, which uses the near-90-GHz channels, here those of the Advanced Microwave Scanning Radiometer-Earth Observing System to calculate sea-ice concentrations. The ASI ice concentrations are compared with ice concentrations derived from the following: 1) the multispectral imager Enhanced Thematic Mapper Plus operating on Landsat and 2) from Envisat and Radarsat SAR images. In this paper, we focus on marginal ice zones, as the ice concentrations in those regions are in general observed with higher errors. First-year ice (bias: -1%-0% and rms error: 1%-4%) and young ice (bias: -4%-0% and rms error: 3%-9%) are fairly well recognized with little underestimation of ASI ice concentrations with respect to Landsat ice concentrations. New ice is identified with less accuracy by the ASI algorithm (bias: -16%-9% and rms error: 18.3%-26.2%). Averaged over all ice types, the bias ranges between -8.4% and 4.5%, and the rms error ranges between 2.0% and 17.4%. Discrepancies mainly occur in polynya areas (underestimation by ASI) and along the ice edge (overestimation by ASI). The results of the ASI-SAR comparison yield contrasting results. ASI underestimates the ice concentrations near the ice edge but overestimates them in some interior areas (bias: -2.9%-2.5% and rms error: 16.9%-20.1%). The discrepancies between both comparisons may be due to the different interaction mechanisms of the different sensor types, particularly with the newly formed ice.

We present a method for the retrieval of atmospheric and surface parameters (namely, surface wind speed, total water vapor, cloud liquid water, surface temperature, ice concentration, multiyear ice fraction) over the Arctic Ocean from brightness temperature measurements by the spaceborne microwave radiometer AMSR-E (Advanced Microwave Scanning Radiometer for EOS) on the satellite Aqua. We use an inverse method, with a forward model based on a fast radiative transfer for AMSR-E over open ocean which we have extended by including the possibility of ice-covered or partly ice-covered sea, using new data on sea ice emissivity at AMSR-E frequencies. The method performs reasonably well and can even retrieve cloud liquid water over ice, and ice concentration in the marginal ice zone in cloudy and humid conditions.

The radiation of the 3rdand 4thStokes components emitted by Arctic sea ice and observed by the spaceborne fully polarimetric radiometer WindSat is investigated. Two types of analysis are carried out, spatial (maps of different quadrants of azimuth look angles) and temporal (time series of daily averages over small selected ranges of azimuth angles). The 3rdStokes component at 37 GHz has shown the highest signal during early summer (Gt2 K). The next highest signals were observed at the 10.7 GHz 3rdand 4thStokes components during the summer months (Gt1 K). The 37 GHz 4thStokes component has shown the least variability (Lt1 K). The 10.7 GHz 4thStokes component has higher signal at the ice edge, confirmed from the sea ice concentration maps derived from Advanced Microwave Scanning Radiometer (AMSR-E) data, and similar to signals observed over global coastlines. From the comparison with near surface air temperature of the European Center for Medium-Range Weather Forecasts (ECMWF) model data and Scanning Multichannel Microwave/Imager (SSM/I) NASA Team ice concentrations it was concluded that the microphysical processes during early melting in the topmost sea ice layers are responsible for the high 37 GHz 3rdStokes signal.

The 150-GHz microwave window channel and the three water vapor channels, 183.3 plusmn 7, 183.3 plusmn 3, and 183.3 plusmn 1 GHz of the Defense Meteorological Satellite Programpsilas (DMSP) Special Sensor Microwave Imager/Sounder (SSMIS) are used to detect intense thunderstorms over the tropics (30degS-30degN). This application utilizes the advantages of conical scanning and high surface horizontal resolution of the SSMIS compared to the same measurements made with cross-track scanning radiometers. The criteria to detect intense tropical thunderstorms are derived by the combination of the brightness temperature differences between the three water vapor channels and brightness temperatures at the four frequencies. One year of SSMIS data (December 2005-November 2006) is surveyed for intense tropical thunderstorms. The most intense tropical thunderstorms are predominant over land and are mainly concentrated over Central and South America, tropical west Africa, and northern Australia. Over most regions these very intense tropical thunderstorms have a strong regional dependence and a pronounced seasonal migration.

A method to retrieve the surface emissivity of sea ice at the window channels of the Advanced Microwave Sounding Unit (AMSU) radiometers in the polar region is presented. The instruments are on the new-generation satellites of the U.S. National Oceanic and Atmospheric Administration (NOAA-15, NOAA-16, and NOAA-17). The method assumes hypothetical surfaces with emissivities zero and one and simulates brightness temperatures at the top of the atmosphere using profiles of atmospheric parameters, e.g., from the European Centre for Medium-Range Weather Forecasts (ECMWF) model runs, as input for a radiative transfer model. The retrieval of surface emissivity is done by combining simulated brightness temperatures with the satellite-measured brightness temperature. The AMSU window channels differ in surface penetration depths and, thus, in the surface microphysical parameters that they depend on. Lowest layer air temperatures from ECMWF are used to infer temperatures of emitting layers at different frequencies of sea ice. A complete yearly cycle of monthly average emissivities in two selected regions (first- and multiyear ice) is giving insight into the variation of emissivities in various development stages of sea ice.

The polar regions are among those where the least information is available about the current and predicted states of surface and atmosphere. We present advances in a method to retrieve the total water vapor (TWV) of the polar atmosphere from data from spaceborne microwave radiometers such as the Advanced Microwave Sounding Unit B (AMSU-B) on the polar-orbiting satellites of the National Oceanic and Atmospheric Administration (NOAA), NOAA-15, -16, and -17. The starting point of the retrieval is a recently proposed algorithm that uses the three AMSU-B channels centered around the 183-GHz water vapor line and the window channel at 150 GHz, and that can retrieve the TWV with little dependence on the surface emissivity. This works up to TWV values of about 7 kg/m2. We extend the retrievable range toward higher TWV values by including the window channel at 89 GHz. However, now, the algorithm needs information on the surface emissivity, which we have extracted from emissivity measurements over sea ice and open water during the Surface Emissivities in Polar Regions-Polar Experiment campaign. The resulting algorithm can retrieve TWV up to about 15 kg/m2, with reduced accuracy as compared to the original algorithm. It now allows the monitoring of the TWV over the central Arctic sea ice and over Antarctica, and the surrounding sea ice during most of the year with a spatial resolution of about 50 km. Such TWV fields can show details which might be missed out by standard weather model analysis data.

The process of determining the geographic latitude and longitude (lat/lon) of the center point of the footprint is called geolocation, which is currently suboptimal for the Advanced Microwave Scanning Radiometer for Earth Observing System (AMSR-E) Level 1 data provided by Japan Aerospace Exploration Agency (JAXA). Here, we present a study for improving the geolocation. The viewing angles (the nadir angle and the scan angle) that define the boresight direction of the instrument are optimized, and new lat/lon coordinates are calculated. The optimization method is based on minimizing the differences between ascending and descending swaths. The results of the calculated viewing angles here have an overall standard deviation of 0.005deg, which is 170 m on ground for the nadir angle and 70 m for the scan angle. The residual geolocation error ranges between 425 m (89 GHz) and 1425 m (6 GHz). The averaged repositioning between JAXA and our geolocation ranges between 3.5 and 7 km, i.e., in the order of one footprint size at 89 GHz. A comparison with a similar study, performed by Wentz et al., shows a good agreement in the viewing angles; the root mean square of the differences is 0.0083deg (ap 283 m) for the nadir angle and 0.0132deg (ap 191 m) for the scan angle.

We compare total water vapor (TWV) data retrieved from two different kinds of satellite data: From the microwave radiances of AMSU-B (Advanced Microwave Sounding Unit B) and from visible spectra of GOME (Global Ozone Monitoring Experiment) and SCIAMACHY (Scanning Imaging Absorption spectroMeter for Atmospheric CHartographY). The retrieval from AMSU-B data only works for dry polar atmosphere but independent of daylight, while retrieval from GOME/SCIAMACHY works globally but depends on daylight. The correlation between the two data sets is moderate to high (0.5 to 0.9) with the exception of late summer and autumn, when it is below 0.4. The most likely reason for the low correlation is that the microwave algorithm is least reliable in autumn because the frequency dependence of the sea ice emissivity (which enters into the algorithm) is modified by the summer surface melt and refreeze in a yet unknown way. The TWV data retrieved from GOME/SCIAMACHY have a slight dry bias with respect to TWV data from AMSU-B, as fully cloudy footprints are excluded from the retrieval with GOME/SCIAMACHY. As the two retrieval methods have different strengths and weaknesses but have only little mutual bias, TWV data from both methods can complement each other.

The effect of ice on polarimetric 4thStokes component observations is investigated using WindSat data over Antarctica. The difference in the magnitude of the signal observed during (July 2003) and summer (February 2004) months are investigated using a second harmonic sine function of the azimuth look angle. The seasonal variations are further investigated by a time series of the 4,th Stokes component for a location in Wilkes Land in east Antarctica and compared with a time series observed by the ERS scatterometer (ESCAT) from a previous work. The paper discusses the potential of a polarimetric radiometer in providing information about scattering and thermal properties of a snow ice pack.

The process of determining the geographic latitude and longitude (short:lat/lon) of the center point of the footprint is called geolocation, which is currently suboptimal for AMSR-E (Advanced Microwave Scanning Radiometer for Earth Observing System) Level 1 data provided by the Japan Aerospace Exploration Agency (JAXA). Here we present a study for improving the geolocation. The viewing angles (off-nadir angle and initial scan angle) that define the boresight direction of the instrument are optimized and new lat/lon coordinates are calculated. The optimization method is based on minimizing differences between ascending and descending swathes. The results of the here calculated viewing angles have an overall standard deviation of 0.005°, which is 170 m on ground for the off-nadir angle and 70 m for the initial scan angle. The averaged repositioning between JAXA and our geolocation is 6.3+1.5 km. A comparison with a similar study, performed by Frank Wentz from RSS (Remote Sensing Systems), shows good agreement in the viewing angles, the RMS of the differences is 0.0083° (≈ 283 m) for the off-nadir angle and 0.0132° (≈ 191 m) for the initial scan angle.

WindSat provides an opportunity to explore the passive microwave polarimetric signatures of land surfaces. In order to accommodate the large sensor footprint, large homogeneous regions with unique features were used. These included forest, rangeland, desert, and agricultural conditions. WindSat observations at horizontal and vertical polarizations over land surfaces were found to be well calibrated and consistent with other passive microwave sensors. Isotropic regions (e.g., Amazon rainforest) had no polarimetric response at all azimuth angles. Results showed that land surfaces with aligned features (topography or row structured vegetation) produced systematic variations in the third and fourth Stokes parameters. These responses were found to be in good agreement with previous sea surface studies. Analysis of the temporal trends of the variation in polarimetric measurements for a specific azimuth angle could be attributed to the crop growth cycle in the agricultural region. Further analyses will seek to isolate specific features that could be used in applications such as soil moisture retrieval.

Polar regions are composed of different surface types, e.g., open water, land ice and different types of sea ice. As a consequence the retrieval of near surface atmospheric parameters (temperature, humidity etc..) is restricted due to the high and highly variable surface emissivity except for open water. The surface emissivities are retrieved using the brightness temperatures at window channels of the cross-track scanning advanced microwave sounding unit (AMSU) satellite radiometer. The emissivities are practically constant up to local zenith angles of 45deg. Comparison of temperature profiles retrieved from AMSU is done with the radiosonde measurements. The root mean square deviations from the measurements is 2 K. Near the surface it increases up to 6 K.

Recent progress in spatial resolution enhancement of sea ice concentrations obtained by microwave remote sensing has been stimulated by two new developments: First, the new sensors AMSR (Advanced Microwave Scanning Radiometer) on MIDORI-II and AMSR-E on AQUA offer horizontal resolutions of 6x4 km at 89 GHz. This is nearly three times the resolution of the standard sensor SSM/I at 85 GHz (15x13 km). The sampling distance at the high frequencies is 12.5 km at SSM/I and 5 km at the AMSR-E instrument. Second, a new algorithm enables the estimation of sea ice concentrations from the channels near 90 GHz, despite the enhanced atmospheric influence in these channels. This allows to fully exploit their horizontal resolution which is two to three times finer than the one of the channels near 19 and 37 GHz. These frequencies are used by the most widespread algorithms for sea ice retrieval, the NASA Team and Bootstrap algorithms. These two developments are combined to determine operationally sea ice concentration maps. The used ASI (Artist Sea Ice) algorithm combines a model for retrieving the sea ice concentration from SSM/I 85 GHz data proposed by Svendsen et al. (1) with an ocean mask derived from the 18-, 23-, and 37-GHz AMSR-E data using two weather filters and the Bootstrap Algorithm. The AMSR-E sea ice concentration data are projected into grids of sampling sizes down to 3 km. Hemispherical and regional maps are provided daily at www. iup.physik.uni-bremen.de.

The common method to detect deep convective clouds is from satellite infrared (IR) measurements, which is based on thresholds of cloud-top temperatures. However, thick cirrus clouds with high cloud tops are difficult to screen out using IR methods, resulting in an overestimation of deep convective cloud fractions. Two aircraft cases with simultaneous Millimeter-wave Imaging Radiometer, Multispectral Atmospheric Mapping Sensor, and ER-2 Doppler radar measurements during the Convection and Moisture Experiment 3 in August 1998 are analyzed to investigate the influence of high thick cirrus clouds on two previously developed IR methods. In contrast, a microwave method based on the brightness temperature differences between the three water vapor channels around 183.3 GHz of the Advanced Microwave Sounding Unit-B (AMSU-B) (183.3/spl plusmn/1,183.3/spl plusmn/3, and 183.3/spl plusmn/7 GHz) can screen out high thick cirrus clouds efficiently. The tropical deep convective cloud fractions (30/spl deg/S-30/spl deg/N) estimated by the IR methods and the AMSU-B method are compared. Although their geographical distributions are in well agreement with each other, the total fractions detected by the IR methods are about 2-3.5 times greater than that detected by the AMSU-B method. Moreover, the overestimation of deep convective cloud fractions by the IR method (11-μm brightness temperature less than 215 K) can result in a displacement in the detected location of the deep convective clouds. The average thick cirrus clouds cover 2.5 times the area of the deep convective clouds that generates them.

The effects of cloud structures on microwave radiances at frequencies from 89-190 GHz are investigated by simulations using the Goddard cumulus ensemble model data as input for a radiative transfer model. It was found that the brightness temperatures at these frequencies have different sensitivities to clouds with a tilted structure. The different sensitivities to altitude and amount of hydrometeors allow the estimation of the canting angle and tilt direction of tilted clouds using brightness temperatures at the water vapor channels at 183.3 /spl plusmn/ 1 and 183.3 /spl plusmn/ 7 GHz. The estimated canting angle and tilt direction are in agreement with the model situation. This method provides a potential to estimate tilted convective structures from microwave radiometric observations at 183.3 /spl plusmn/ 1 and 183.3 /spl plusmn/ 7 GHz. It is applied to a tilted storm observed from the National Aeronautics and Space Administration's ER-2 aircraft flying at about 20 km on August 26, 1998 during the third Convection and Moisture Experiment using the observed downlooking brightness temperatures at the water vapor channels of a Millimeter-wave Imaging Radiometer. The estimated results are in good agreement with the realistic storm situation obtained from the simultaneous observations of the ER-2 Doppler radar. This method also provides information about the vertical displacement of cloud structure and thereby to estimate the accurate location of surface rainfall. This is important when validating precipitation retrieval based on observations of the ice scattering above surface rainfall against surface rain observations using the microwave frequencies sensitive to high altitudes.

The Advanced Microwave Scanning Radiometer for EOS (AMSR-E) was launched on May 4, 2002 aboard the NASA's Earth Orbiting System (EOS) Aqua Satellite. This new instrument measures the Earth's radiation in two orthogonally polarized channels at 6 frequencies extending from 6.925 to 89.0 GHz. Here we study the influence of surface radiation on retrieval of cloud liquid water (CLW) and precipitable water vapor (PWV) using multifrequency polarized measurements of AMSR-E. Based on this preliminary investigation, a physical method to simultaneously retrieve CLW and PWV is proposed to build.

Microwave radiometers were operated on a research aircraft during low-level flights in the Arctic in March 2001. Data from these instruments are used to calculate surface emissivity at millimetre wavelengths. This calculation includes the retrieval of an effective temperature from multiple channels centred at 183 GHz. Emissivity spectra will be presented for a range of surface types, including open water, new ice (nilas), first year, multi-year, and glacial ice. These show that the emissivity of older ice types is lower at 24 GHz, decreases further at higher frequencies, but usually starts to increase by 157 or 183 GHz. These results will help exploit data from satellite instruments, such as AMSU, in the Arctic.

The recently suggested sub-millimeter sensors for the global observation of cirrus clouds all bear several channels widely separated from each other and spanning the range from 180 to 880 GHz. While the potential for retrieving the integrated ice water content and the mean effective particle size of ice clouds has been investigated, the quantitative retrieval of particle shape and orientation information is still an active research area. Here we investigate the influence of the parameters of a Gaussian distribution of the horizontal particle orientation on the single scattering cross section. It is only slightly affected, but the polarization signal is reduced to 90% of the completely horizontally oriented case if the standard deviation /spl sigma//sub /spl beta// of the horizontal distribution is 15/spl deg/ and reduced to 50% if /spl sigma//sub /spl beta// = 30/spl deg/.

A method to detect the cloud signature (mainly the cloud liquid water) over the sea ice-covered Weddell Sea in the Austral summer season is presented. By using the polarization differences at the two high frequency channels (i.e., 37 and 85 GHz) of the Special Sensor Microwave/Imager (SSM/I), a new quantity called R-factor is defined. Using the R-factor, the atmospheric signal can be easily separated from the surface signal and, more importantly, the surface signal and its variation can be strongly suppressed, especially in regions with low ice concentrations. In regions with high ice concentrations, other sea ice parameters like snow cover play an important role as indicated by simulations using in situ measured sea ice emissivities and observed by the SSM/I. Under the assumption that the sea ice parameters remain sufficiently stable within a short period (e.g.. ten days), a method is proposed to determine the background term from SSM/I measurements, allowing the detection of the cloud signature. A comparison with a known SSM/I cloud liquid water algorithm over the open ocean shows a high degree of correlation (0.958) among the cloud signatures detected by the two algorithms. On January 2 and 3, 1996, a low pressure system moved into the sea ice-covered Weddell Sea. Its cloud signature detected using the R-factor method compares well with coincident observations from both visible and infrared sensors.

A method to derive the sea ice edges from the 37- and 85-GHz channels of the spaceborne Special Sensor Microwave/Imager (SSM/I) with higher resolution than operational ice concentration analyses is developed. A combined ice-atmosphere model to evaluate atmospheric influences showed that under clear sky conditions the error of the ice edge position is 3.1 km on average. The method performs well up to total liquid water content (TWC) of 125 g/m/sup 2/ and overestimates the ice extent when TWC increases up to 250 g/m/sup 2/. The algorithm is suitable to classify open water, the interior ice pack, and thin ice (nilas) or low ice concentrations caused by a mixture of open water and pancake ice. Simulations showed that interpolation and resolution-matching techniques may improve the retrieval results with emphasis on interpolation. Validation of the algorithm with high-resolution infrared data shows errors of the ice edge on average less than 5 km. Comparison with the NASA Team ice concentration algorithm in the Weddell Sea, Antarctica, shows best agreement with the contour line of 33% ice concentration. The suitability of the method for mesoscale investigations is demonstrated with a time series of a meander structure in the marginal ice zone (MIZ).

A space variant image restoration algorithm has been developed with the aim of improving the spatial resolution of SSM/I (Special Sensor Microwave/Imager) passive microwave imagery. Due to the conical scanning of the instrument the relative geometry of the data samples changes over the scan. This change is accounted for by using a space variant point-spread-function in the restoration algorithm. Application of this algorithm to a scene from the Weddell Sea results in an image with enhanced ice edge and coast definition. As a result ice concentration estimates near the edge agree more closely with higher resolution (optical) data from AVHRR.>