WindSat physically based forward model: atmospheric component

This paper established a geophysical retrieval algorithm for sea surface wind vector, sea surface temperature, columnar atmospheric water vapor, and columnar cloud liquid water from WindSat, using the measured brightness temperatures and a matchup database. To retrieve the wind vector, a chaotic particle swarm approach was used to determine a set of possible wind vector solutions which minimize the difference between the forward model and the WindSat observations. An adjusted circular median filtering function was adopted to remove wind direction ambiguity. The validation of the wind speed, wind direction, sea surface temperature, columnar atmospheric water vapor, and columnar liquid cloud water indicates that this algorithm is feasible and reasonable and can be used to retrieve these atmospheric and oceanic parameters. Compared with moored buoy data, the RMS errors for wind speed and sea surface temperature were 0.92 m s−1 and 0.88°C, respectively. The RMS errors for columnar atmospheric water vapor and columnar liquid cloud water were 0.62 mm and 0.01 mm, respectively, compared with F17 SSMIS results. In addition, monthly average results indicated that these parameters are in good agreement with AMSR-E results. Wind direction retrieval was studied under various wind speed conditions and validated by comparing to the QuikSCAT measurements, and the RMS error was 13.3°. This paper offers a new approach to the study of ocean wind vector retrieval using a polarimetric microwave radiometer.

A new multiparameter retrieval algorithm based on a backpropagation neural network (BPNN) has been developed for deriving integrated water vapor (WV) and cloud liquid water (CLW) contents over oceans from brightness temperatures (BTs) measured by the Multi-frequency Scanning Microwave Radiometer (MSMR) launched onboard Indian Remote Sensing satellite IRS-P4. The MSMR measures brightness temperatures in vertical and horizontal polarizations at 6.0-, 10.65-, 18.0-, and 21.0-GHz frequencies. The data are available at three spatial grid resolutions of 150, 75, and 50 km. In this paper, a BPNN has been trained using brightness temperatures simulated through radiative transfer model and simulated surface and atmospheric parameters. The present algorithm has been compared with the operational MSMR retrieval algorithm based on statistical regression using the same dataset. The validation of WV with in situ data (Vaisala radiosonde) is presented. Moreover, comparison of WV and CLW derived from MSMR using BPNN with the finished products from the Special Sensor Microwave/Imager and the Tropical Rainfall Measuring Mission Microwave Imager has also been carried out. The complexity of the BPNN in retrieval of geophysical products, individually and simultaneously, has also been discussed. Simultaneous retrieval of WV and CLW improves the results.

In 2000 Frank Wentz and Thomas Meissner described a model function for the microwave radiometer's brightness temperature of the ocean and intervening atmosphere. The oceanic and atmospheric parameters are the sea surface temperature the near-surface wind speed the columnar water vapor and the columnar cloud liquid water. We analyzed the sensitivity between the geophysical parameters and the microwave radiometer's brightness temperature with the model function. First we find that the sensitive bands of the sea surface temperature are 6.93GHz v-pol and 10.65GHz v-pol; the sensitive bands ofthe near-surface wind speed are 6.93GHz h-pol, 10.65GHz h-pol, 18.7GHz h-pol, 23.8GHz h-pol, and 36.5GHz h-pol; the sensitive bands of the columnar water vapor are 18.7GHz v-pol, 18.7GHz h-pol, 23.8GHz v-pol, 23.8GHz h-pol, 36.5GHz v-pol and 36.5GHz h-pol; the sensitive bands of the columnar cloud liquid water are 36.5GHz v-pol and 36.5GHz h-pol. Second we find that the brightness temperature is sensitive to incidence angle variations. One-degree error in specifying incidence angle produces a 2.5K error in brightness temperature. The 2.5K errors in brightness temperature produce 5K errors in sea surface temperature and 4.5m/s error in near-surface wind speed. Third we find that brightness temperature is not sensitive to the sea surface salinity in the 6 to 37 GHz band.

A new simple two-scale model on the polarimetric microwave emission of ocean surface is derived at first, which can be expressed as an integral of weighting functions (M 0 and M 2) and ocean surface curvature spectrum coefficients (C 0 and C 2). This provides a simple way to investigate the effect of curvature spectrum on ocean emission. It is found that ocean waves with wavelengths both comparable to and much greater than the electromagnetic wavelength can contribute to the harmonics of ocean surface microwave emission, depending on the magnitude of the ocean surface spectrum in these length scales. Brightness temperature predictions differ significantly due to present diverse spectrum models, and thus a study on wave spectrum obtained inversely from brightness temperature measurements is necessary. From the ocean surface radiation data measured by polarimetric microwave radiometer, we derived an ocean wave spectrum with a wider wave number range, using the proposed two-scale model and constrained linear least-squares method. The derived ocean wave spectrum is useful for comparing with present diverse models.

The presence of snow cover affects the regional energy and water balance, thus having a significant impact on the global climate system. Temporal knowledge of the onset of snow melt and snow water equivalent (SWE) values are important variables in the prediction of flooding, as well as water resource applications such as reservoir management and agricultural activities. Microwave remote sensing techniques have been effective for monitoring snow pack parameters (snow extent, depth, water equivalent, wet/dry state). Coincident ground data, airborne polarimetric C-band (5.3 GHz) Synthetic Aperture Radar (SAR) and passive microwave radiometer data (19, 37 and 85 GHz) were collected on four dates (1 December 1997, 6 March 1998, 12 March 1998 and 9 March 1999) over two flight lines in Eastern Ontario, Canada. The multitemporal, multi-sensor data were analysed for changes in SAR polarimetric signatures and microwave brightness temperatures as a function of changing snow pack parameters. Results indicate that certain parameters such as linear polarizations and pedestal height are sensitive to changes in snow pack parameters, and respond differently to various snow conditions. SWE values derived from the passive microwave brightness temperatures compare well with ground measurements, with the exception of low snow volume and in the presence of significant ice layers.

The microwave signal at L-band is very sensitive to the soil moisture due to its penetrability. To analyze an algorithm to retrieve soil moisture at L-band, a simulation of microwave brightness temperature is conducted by using the τ-ω model in this study, and two methods of resampling are compared. One of the brightness temperature simulation is based on the point observation on ground, and the other is from the ground observation data of 1km resolution resampled from point. It turns out the latter method has a smaller error. An airborne L-band data from a Polarimetric L-band Microwave Radiometer (PLMR) acquired during the Hiwater experiment held in the Heihe River Basin in 2012 are used to validate the brightness temperature simulation. And the root-mean-square errors between L-MEB simulated and PLMR are 9K to 12K for V-polarization, and 6K to 8K at H-polarization respectively at different angles.

Radiative transfer simulations are performed to determine how water vapor and nonprecipitating cloud liquid water and ice particles within typical midlatitude atmospheres affect brightness temperatures TB's of moisture sounding channels used in the Advanced Microwave Sounding Unit (AMSU) and AMSU-like instruments. The purpose is to promote a general understanding of passive top-of-atmosphere TB's for window frequencies at 23.8, 89.0, and 157.0 GHz, and water vapor frequencies at 176.31, 180.3 1, and 182.31 GHz by documenting specific examples. This is accomplished through detailed analyses of TB's for idealized atmospheres, mostly representing temperate conditions over land. Cloud effects are considered in terms of five basic properties: droplet size distribution, phase, liquid or ice water content, altitude, and thickness. Effects on TB of changing surface emissivity also are addressed. The brightness temperature contribution functions are presented as an aid to physically interpreting AMSU TB's. Both liquid and ice clouds impact the TB's in a variety of ways. The TB's at 23.8 and 89 GHZ are more strongly affected by altostratus liquid clouds than by cirrus clouds for equivalent water paths. In contrast, channels near 157 and 183 GHz are more strongly affected by ice clouds. Higher clouds have a water impact on 157- and 183-GHz TB's than do lower clouds. Clouds depress TB's of the higher-frequency channels by suppressing, but not necessarily obscuring, radiance contributions from below. Thus, TB's are less closely associated with cloud-top temperatures than are IR radiometric temperatures. Water vapor alone accounts for up to 89% of the total attenuation by a midtropospheric liquid cloud for channels near 183 GHz. The Rayleigh approximation is found to be adequate for typical droplet size distributions; however, Mie scattering effects from liquid droplets become important for droplet size distribution functions with modal radii greater than 20 µm near 157 and 183 GHz, and greater than 30–40 µm at 89 GHz. This is due mainly to the relatively small concentrations of droplets much larger than the mode radius. Orographic clouds and tropical cumuli have been observed to contain droplet size distributions with mode radii in the 30–40-µm range. Thus, as new instruments bridge the gap between microwave and infrared to frequencies even higher than 183 GHz, radiative transfer modelers are cautioned to explicitly address scattering characteristics of such clouds.

Compared with the traditional microwave radiometer, one dimensional synthetic aperture microwave radiometer can effectively improve the spatial resolution of sea surface temperature retrieval. However, the spaceborne one dimensional synthetic aperture microwave radiometer is multiple incidence angle observation for the sea target and the incidence angle changes from 0° to 55°. In order to develop sea surface temperature inversion algorithms suited to one dimensional synthetic aperture microwave radiometer, it is necessary to evaluate the sensitivities of brightness temperature to oceanic and atmospheric environmental elements. Using the sea surface emissivity model and the atmospheric radiative transfer model, we construct a oceanic and atmospheric radiative transfer mode suited to the one dimensional synthetic aperture microwave radiometer. In this paper, the sensitivity changes of C-band vertical and horizontal polarization brightness temperature to oceanic and atmospheric environmental elements at different incidence angles are studied, and the corresponding sensitivity coefficients are calculated. The results indicate that the sensitivities of vertical and horizontal polarization brightness temperature to oceanic and atmospheric environmental elements show different characteristics. With the increase of the incidence angle, the sensitivity of the vertical polarization brightness temperature to the sea surface temperature is enhanced, and the sensitivity to the sea surface wind field is relatively weakened; but the horizontal polarization brightness temperature is opposite. The vertical and horizontal polarization brightness temperature errors caused by the errors of atmospheric water vapor content and cloud liquid water content increase with the increasing of the incidence angle, however, even at a large incidence angle of 55°, the errors of vertical and horizontal polarization brightness temperature are still less than 0.12 K. If the accuracy of sea surface temperature inversion is higher than 1 K, the calibration precision of one dimensional synthetic aperture microwave radiometer should be better than 0.6 K. Overall, the results of this paper are of great significance to the research of sea surface temperature retrieval with multiple incidence angle for one dimensional synthetic aperture microwave radiometer.

A ground-based combination microwave radiometer-satellite receiver operating at 28.5 GHz (wavelength = 1.05 cm) was employed to measure liquid water in clouds during the 1979-80 field season of the Sierra Cooperative Pilot Project (SCPP) in northern California. We report upon the ability of the instrument to detect small amounts (∼0.1 g m−3) of liquid water in non-precipitating clouds which may not be observed by other remote sensing systems. A successful cloud seeding experiment, performed after cloud liquid was detected by the instrument, is described.

Abstract. This empirical study demonstrates the feasibility of using 89-GHz Advanced Microwave Scanning Radiometer–Earth Observing System (AMSR-E) passive microwave brightness temperature data to detect heavily drizzling cells within subtropical marine stratocumulus. For the purpose of this paper, we define heavily drizzling cells as areas ≥ 6 km × 4 km with C-band Z > 0 dBZ; equivalent to > 0.084 mm h−1. A binary heavy drizzle product is described that can be used to determine areal and feature statistics of drizzle cells within the major marine stratocumulus regions. Current satellite liquid water path (LWP) and cloud radar products capable of detecting drizzle are either lacking in resolution (AMSR-E LWP), diurnal coverage (MODIS LWP), or spatial coverage (CloudSat). The AMSR-E 89-GHz data set at 6 km × 4 km spatial resolution is sufficient for resolving individual heavily drizzling cells. Radiant emission at 89 GHz by liquid-water cloud and precipitation particles from drizzling cells in marine stratocumulus regions yields local maxima in brightness temperature against an otherwise cloud-free background brightness temperature. The background brightness temperature is primarily constrained by column-integrated water vapor for moderate sea surface temperatures. Clouds containing ice are screened out. Once heavily drizzling pixels are identified, connected pixels are grouped into discrete drizzle cell features. The identified drizzle cells are used in turn to determine several spatial statistics for each satellite scene, including drizzle cell number and size distribution. The identification of heavily drizzling cells within marine stratocumulus regions with satellite data facilitates analysis of seasonal and regional drizzle cell occurrence and the interrelation between drizzle and changes in cloud fraction.

Climate change is amplified in the Arctic region (north of 60 ° N) relative to elsewhere. By analyzing climate model simulations, it has been found that the largest contribution to Arctic amplification comes from temperature feedbacks, due to both the different warming profile in low and high latitudes and a larger temperature increase in longwave emission per unit of warming at colder background temperatures compared to tropical conditions. Satellite remote sensing offers the best way of deriving sea surface temperature (SST) in the Arctic, but given that the retrieval algorithms in the infrared (IR) are designed to compensate for the effects of the atmosphere, mainly water vapor, IR satellite-derived SSTs have larger uncertainties at high latitudes because the atmosphere is very dry and cold. So, the motivation of this study is to improve the algorithms to obtain more accurate SSTs which can be used to research the feedback mechanisms. To undertake the study, the matchup database (MUDB) for MODIS (Moderate resolution Imaging Spectroradiometer) on Aqua has been analyzed to characterize the differences between collocated and simultaneous satellite retrieved skin SSTs and in situ buoy temperatures, and to identify the main causes of the discrepancies. According to the radiative transfer simulations, the sea surface emissivity is proven to be significant in satellite SST retrievals at high latitudes due to the low atmospheric water vapor content, especially during winter. An Emissivity introduced Brightness Temperature Difference (EBΔT) is introduced to correct this emissivity effect in the algorithm. Furthermore, the reference temperature as a weight factor for brightness temperature (BT) difference between MODIS bands 31 and 32 has also been adjusted. The satellite SST biases and standard deviation are reduced by 0.41 K and 0.11 K after the corrections. The approach presented in this study is capable to make more appropriate atmospheric corrections for MODIS SST retrieval algorithm, leading to regional optimization of the SST retrievals. We report on the progress towards improving the satellite-derived SST with the expectation that the near two-decadal time series of MODIS SST fields will contribute to studying climate change in the Arctic.

Soil moisture retrieval from passive microwave remote sensing over densely vegetated areas, such as forests, has become a pertinent subject in the light of ESA's upcoming soil moisture and ocean salinity (SMOS) mission. Even though it is probably not possible to retrieve soil moisture over forested areas with the accuracy required by SMOS, understanding forest radiative transfer properties at L-band (1.1-1.7 GHz) is still important for soil moisture retrieval in partly forested pixels. The forward model to be used by the SMOS level 2 soil moisture algorithm is 'L-MEB', from 'L-band microwave emission of the biosphere'. This paper shows the results of L-MEB model calibration over two forest types in terms of various soil and vegetation parameters, together with an assessment of model performance in both forward and inverse mode. While, as expected, the retrieval of soil moisture over forest areas does not give results up to the standard required by SMOS, the forward model does result in an acceptable root mean square error in terms of 'reconstructed' brightness temperatures.

The responses of tropical clouds and water vapor to SST variations are investigated with simple numerical experiments. The fifth-generation Pennsylvania State University–National Center for Atmospheric Research (PSU–NCAR) Mesoscale Model is used with doubly periodic boundary conditions and a uniform constant sea surface temperature (SST). The SST is varied and the equilibrium statistics of cloud properties, water vapor, and circulation at different temperatures are compared. The top of the atmosphere (TOA) radiative fluxes have the same sensitivities to SST as in observations averaged from 20°N to 20°S over the Pacific, suggesting that the model sensitivities are realistic. As the SST increases, the temperature profile approximately follows a moist-adiabatic lapse rate. The rain rate and cloud ice amounts increase with SST. The average relative humidity profile stays approximately constant, but the upper-tropospheric relative humidity increases slightly with SST. The clear-sky mean temperature and water vapor feedbacks have similar magnitudes to each other and opposite signs. The net clear-sky feedback is thus about equal to the lapse rate feedback, which is about −2 W m−2 K−1. The clear-sky outgoing longwave radiation (OLR) thus increases with SST, but the high cloud-top temperature is almost constant with SST, and the high cloud amount increases with SST. The result of these three effects is an increase of cloud longwave forcing with SST and a mean OLR that is almost independent of SST. The high cloud albedo remains almost constant with increasing SST, but the increase in high cloud area causes a negative shortwave cloud radiative forcing feedback, which partly cancels the longwave cloud feedback. The net radiation decreases slightly with SST, giving a small net negative feedback, implying a stable, but very sensitive climate.

Currently, many microwave radiometers have been successfully used in earth-observation and ocean-monitoring systems. In China, the scanning microwave radiometers (SMR) onboard the Haiyang-2A/2B (HY-2A/2B) satellites are multichannel radiometers capable of obtaining oceanic and atmospheric parameters, such as sea surface temperature (SST), sea surface wind speed (WS), water vapor (WV), and cloud liquid water (LW). As a successor to HY-2A, the HY-2B satellite was launched successfully on October 25, 2018. Compared with the HY-2A’s SMR, the HY-2B’s SMR is an improved microwave radiometer. This article is concerned with the instrument design of the HY-2B’s SMR and its improvements. The important improvements of the HY-2B’s SMR in regard to the feedhorns and cold-sky reflector (CSR) are verified, and the early in-orbit performance of the HY-2B’s SMR is evaluated. The results demonstrate that the HY-2B’s SMR exhibits an important improvement on the HY-2A’s SMR in terms of the quality of the brightness temperature (TB) products and calibration accuracy. The results also show that the HY-2B’s SMR has a good performance in terms of the global TB images and the retrieved products with a good agreement with the expectations.

Summary form only given. The WindSat payload on the Coriolis satellite produces approximately 200 MBytes of data per orbit, or 3.0 GB per day. The WindSat team has developed a Ground Data Processing Software (GDPS) to process 100% of available downlink data, providing accurate and precise data products as dictated by the WindSat mission requirements. The GDPS consists of four main processors: the Raw Data Record Processor (RDRP), the Temperature Data Record Processor (TDRP), the Sensor Data Record (SDRP), and the Environmental Data Record Processor (RDRP). The RDRP receives the WindSat data stream, including receiver, calibration, housekeeping and spacecraft attitude data. The RDRP decodes, unpacks, reformats and merges these data in preparation for further the remaining processing. The TDRP receives the formatted data and provides the first four critical functions. First the TDRP calibrates the raw counts into radiances using the calibration target measurements and the sensor calibration constants file. Then, each channel is geolocated using the scan angle, spacecraft attitude and ephemeris, and the antenna pointing offsets. Using the spacecraft ephemeris and an ionosphere model, the Faraday rotation is calculated. The Faraday rotation angle and the spacecraft attitude and ephemeris are then used to calculate various antenna parameters, such as the earth incidence angle, the polarization rotation angle and the surface compass azimuth angle, which describes the sensor look direction on the surface. These data are forwarded on to the SDRP. The next level of processing occurs in the SDRP, which forms the third and fourth elements Stokes parameters by taking the difference of the appropriate polarizations. The antenna pattern correction (APC) follows, using stored cross-coupling correction matrices and feed spillover efficiencies. Because each frequency set has a different beamwidth and relative pointing, the final stage of the SDRP resamples and averages the beams to yield collocated pixels at a common resolution. This is the highest level of phase 1 (brightness temperature) processing before the EDRP. The EDRP provides the WindSat Level 2 processing. The inputs to the EDRP are the calibrated, geolocated, co-registered WindSat Stokes parameters at 10.7, 18.7, and 37 GHz, and the vertical and horizontally polarized brightness temperatures at 6.8 and 23.8 GHz. These sets of 16 quantities form the basic measurement vector for the retrieval. The primary retrieval algorithm is an optimal estimation technique that simultaneously solves for ocean surface wind speed and direction, as well as sea surface temperature (SST) and atmospheric parameters such a precipitable water vapor, rain rate and cloud liquid water.