Narendra N. Das

An Algorithm for Merging SMAP Radiometer and Radar Data for High-Resolution Soil-Moisture Retrieval

A robust and simple algorithm is developed to merge L-band radiometer retrievals and L-band radar observations to obtain high-resolution (9-km) soil-moisture estimates from data of the NASA Soil Moisture Active and Passive (SMAP) mission. The algorithm exploits the established accuracy of coarse-scale radiometer soil-moisture retrievals and blends this with the fine-scale spatial heterogeneity detectable by radar observations to produce a high-resolution optimal soil-moisture estimate at 9 km. The capability of the algorithm is demonstrated by implementing the approach using the airborne Passive and Active L-band System (PALS) instrument data set from Soil Moisture Experiments 2002 (SMEX02) and a four-month synthetic data set in an Observation System Simulation Experiment (OSSE) framework. The results indicate that the algorithm has the potential to obtain better soil-moisture accuracy at a high resolution and show an improvement in root-mean-square error of 0.015-0.02-cm3/cm3 volumetric soil moisture over the minimum performance taken to be retrievals based on radiometer measurements resampled to a finer scale. These results are based on PALS data from SMEX02 and a four-month OSSE data set and need to be further confirmed for different hydroclimatic regions using airborne data sets from prelaunch calibration/validation field campaigns of the SMAP mission.

Tests of the SMAP Combined Radar and Radiometer Algorithm Using Airborne Field Campaign Observations and Simulated Data

A soil moisture retrieval algorithm is proposed that takes advantage of the simultaneous radar and radiometer measurements by the forthcoming NASA Soil Moisture Active Passive (SMAP) mission. The algorithm is designed to downscale SMAP L-band brightness temperature measurements at low resolution ( ~ 40 km) to 9-km brightness temperature by using SMAPs L-band synthetic aperture radar (SAR) backscatter measurements at high resolution (1-3 km) in order to estimate soil moisture at 9-km resolution. The SMAP L-band SAR and radiometer instruments are designed to provide coincident observations at constant incidence angle, but at different spatial resolutions, across a wide swath. The algorithm described here takes advantage of the correlation between temporal fluctuations of brightness temperature and backscatter observed when viewing targets simultaneously at the same angle. Surface characteristics that affect the brightness temperature and backscatter measurements influence the signals at different time scales. This feature is applied in an approach in which fine-scale spatial heterogeneity detected by SAR observations is applied on coarser-scale radiometer measurements to produce an intermediate-resolution disaggregated brightness temperature field. These brightness temperatures are then used with established radiometer-based algorithms to retrieve soil moisture at the intermediate resolution. The capability of the overall algorithm is demonstrated using data acquired by the airborne passive and active L-band system from field campaigns and also by simulated global dataset. Results indicate that the algorithm has the potential to retrieve soil moisture at 9-km resolution, with the accuracy required for SMAP, over regions having vegetation up to 5- kg/m2 vegetation water content. The results show a reduction in root mean square error of volumetric soil moisture (40% improvement in the statistics) from the minimum performance defined as the soil moisture retrieved using radiometer measurements re-sampled to the intermediate scale.

Satellite-Based Precipitation Estimation and Its Application for Streamflow Prediction over Mountainous Western U.S. Basins

AbstractRecognizing the importance and challenges inherent to the remote sensing of precipitation in mountainous areas, this study investigates the performance of the commonly used satellite-based high-resolution precipitation products (HRPPs) over several basins in the mountainous western United States. Five HRPPs [Tropical Rainfall Measuring Mission 3B42 and 3B42-RT algorithms, the Climate Prediction Center morphing technique (CMORPH), Precipitation Estimation from Remotely Sensed Imagery Using Artificial Neural Networks (PERSIANN), and the PERSIANN Cloud Classification System (PERSIANN-CCS)] are analyzed in the present work using ground gauge, gauge-adjusted radar, and CloudSat precipitation products. Using ground observation of precipitation and streamflow, the skill of HRPPs and the resulting streamflow simulations from the Variable Infiltration Capacity hydrological model are cross-compared. HRPPs often capture major precipitation events but seldom capture the observed magnitude of precipitation ove...

Reintroducing radiometric surface temperature into the Penman‐Monteith formulation

Here we demonstrate a novel method to physically integrate radiometric surface temperature (TR) into the Penman-Monteith (PM) formulation for estimating the terrestrial sensible and latent heat fluxes (H and λE) in the framework of a modified Surface Temperature Initiated Closure (STIC). It combines TR data with standard energy balance closure models for deriving a hybrid scheme that does not require parameterization of the surface (or stomatal) and aerodynamic conductances (gS and gB). STIC is formed by the simultaneous solution of four state equations and it uses TR as an additional data source for retrieving the “near surface” moisture availability (M) and the Priestley-Taylor coefficient (α). The performance of STIC is tested using high-temporal resolution TR observations collected from different international surface energy flux experiments in conjunction with corresponding net radiation (RN), ground heat flux (G), air temperature (TA), and relative humidity (RH) measurements. A comparison of the STIC outputs with the eddy covariance measurements of λE and H revealed RMSDs of 7–16% and 40–74% in half-hourly λE and H estimates. These statistics were 5–13% and 10–44% in daily λE and H. The errors and uncertainties in both surface fluxes are comparable to the models that typically use land surface parameterizations for determining the unobserved components (gS and gB) of the surface energy balance models. However, the scheme is simpler, has the capabilities for generating spatially explicit surface energy fluxes and independent of submodels for boundary layer developments.

Active–Passive Soil Moisture Retrievals During the SMAP Validation Experiment 2012

The goal of this study is to assess the performance of the active-passive algorithm for the NASA Soil Moisture Active Passive mission (SMAP) using airborne and ground observations from a field campaign. The SMAP active-passive algorithm disaggregates the coarse-resolution radiometer brightness temperature (TB) using high-resolution radar backscatter (σo) observations. The colocated TB and σo acquired by the aircraft-based Passive Active Land S-band sensor during the SMAP Validation Experiment 2012 (SMAPVEX12) are used to evaluate this algorithm. The estimation of its parameters is affected by changes in vegetation during the campaign. Key features of the campaign were the wide range of vegetation growth and soil moisture conditions during the experiment period. The algorithm performance is evaluated by comparing retrieved soil moisture from the disaggregated brightness temperatures to in situ soil moisture measurements. A minimum performance algorithm is also applied, where the radar data are withheld. The minimum performance algorithm serves as a benchmark to asses the value of the radar to the SMAP active-passive algorithm. The temporal correlation between ground samples and the SMAP active-passive algorithm is improved by 21% relative to minimum performance. The unbiased root-mean-square error is decreased by 15% overall.

Sensitivity of Aquarius active and passive measurements temporal covariability to land surface characteristics

Active and passive microwave observations over land are affected by surface characteristics in different ways. L-band radar backscatter and radiometer measurements each have distinct advantages and problematic issues when applied to surface soil moisture estimation. Spaceborne radiometry has the advantage of better sensitivity to the geophysical parameter but suffers from coarse spatial resolution given limitations on antenna dimensions. Active sensing has the advantage of higher spatial resolution, but the measurements are, relative to radiometry, more affected by the confounding influences of scattering by vegetation and rough surfaces. Active and passive measurements can potentially span different scales and allow the combining of the relative advantages of the two sensing approaches. This strategy is being implemented in the NASA Soil Moisture Active Passive (SMAP) mission, which relies on the relationship between active and passive measurements to provide 9-km surface soil moisture estimates. The aim of this paper is to study the sensitivity of spaceborne L-band active and passive temporal covariations to land surface characteristics, in preparation for SMAP. A significant linear relationship (with slope β) is obtained between NASAs Aquarius scatterometer and radiometer observations across major global biomes. The error in β estimation is found to increase with land cover heterogeneity and to be unaffected by vegetation density (up to moderate densities). Results show that β estimated with two to eight months of Aquarius measurements (depending on vegetation seasonality) reflect local vegetation cover conditions under surfaces with complex mixture of vegetation, surface roughness, and dielectric constant.

Joint Sentinel-1 and SMAP Data Assimilation to Improve Soil Moisture Estimates

SMAP (Soil Moisture Active and Passive) radiometer observations at ~40 km resolution are routinely assimilated into the NASA Catchment Land Surface Model to generate the 9-km SMAP Level-4 Soil Moisture product. This study demonstrates that adding high-resolution radar observations from Sentinel-1 to the SMAP assimilation can increase the spatio-temporal accuracy of soil moisture estimates. Radar observations were assimilated either separately from or simultaneously with radiometer observations. Assimilation impact was assessed by comparing 3-hourly, 9-km surface and root-zone soil moisture simulations with in situ measurements from 9-km SMAP core validation sites and sparse networks, from May 2015 to December 2016. The Sentinel-1 assimilation consistently improved surface soil moisture, whereas root-zone impacts were mostly neutral. Relatively larger improvements were obtained from SMAP assimilation. The joint assimilation of SMAP and Sentinel-1 observations performed best, demonstrating the complementary value of radar and radiometer observations.

Intercomparisons of Brightness Temperature Observations Over Land From AMSR-E and WindSat

The Advanced Microwave Scanning Radiometer-EOS (AMSR-E) on Aqua and WindSat on Coriolis instruments have collected multichannel passive microwave data over the global land and oceans since 2002 and 2003, respectively. AMSR-E on Aqua ceased operation in October 2011 due to a malfunction in the antenna scanning mechanism. AMSR-E and WindSat have similar frequencies, bandwidths, polarizations, incidence angles and instantaneous fields of view (IFOVs), but there are some differences in their configurations. The altitudes and local overpass times also differ between the AMSR-E and WindSat sensors. The time series of data from the two instruments have a long period of overlap, which can be used to intercompare and cross-calibrate the instrument data sets taking into account the instrument differences. This would allow retrieval of geophysical parameters using common algorithms that could take advantage of the increased time duration and sampling coverage afforded by combining data from the two sensors. In this paper, we focus on land applications and compare the multichannel data from these two sensors over land. Channels useful primarily for soil moisture and vegetation water content studies (i.e., ~ 6, ~ 10, ~ 18, and ~ 37 GHz at H- and V-pol) are used in the comparisons. To minimize differences caused by surface temperature effects related to local overpass times, only descending passes (with Equator crossing times for AMSR-E of 1:30 a.m. and WindSat 6:00 a.m.) are considered. Homogeneous and temporally stable sites such as Dome-C, Antarctica and the Amazon forest, and a flat and bare region in the Sahara desert are chosen to evaluate similarities and differences among comparable channel observations. Taking into consideration the sensor configurations and geophysical conditions during the descending overpasses, reasonably good agreement is observed between AMSR-E and WindSat measurements over the globe.

Uncertainty Estimates in the SMAP Combined Active–Passive Downscaled Brightness Temperature

NASAs Soil Moisture Active Passive (SMAP) mission objective is global mapping of surface volumetric soil moisture at 10-km resolution every two to three days and with accuracy of 0.04 cm3 cm-3 (one sigma). In order to achieve this resolution and accuracy, the SMAP utilizes L-band radar and L-band radiometer measurements. The instruments share a rotating 6-m mesh reflector antenna that scans across a 1000-km swath in order to meet the required data refresh rate. The Level-2 Active-Passive soil moisture product (L2_SM_AP) at 9 km is retrieved from the disaggregated/downscaled brightness temperature obtained by merging of active and passive L-band observations. The baseline L2_SM_AP algorithm disaggregates the coarse-resolution (~36 km) brightness temperatures of the SMAP L-band radiometer using the high-resolution (~3 km) backscatter data from the SMAP L-band radar with unfocused synthetic aperture processing. The inversion of brightness temperature to estimate surface soil moisture is more mature when compared with inversions of radar backscatter. This is the primary driver of the brightness temperature disaggregation approach to the combined active-passive surface soil moisture product. Furthermore, this approach allows some consistency with the coarse-resolution radiometer-only surface soil moisture product since the disaggregated brightness temperatures sums to the radiometer measurement. The disaggregated brightness temperature contains instrument errors (~0.7 dB for co-pol backscatter and ~1.0 dB for cross-pol backscatter, and ~1.3 K in brightness temperature) inherent in the radar and radiometer. Furthermore, the algorithm has two critical parameters that add uncertainty. Finally, correction of the land brightness temperature (used in the inversion) for water body contributions is a source of uncertainty. In this paper, we introduce analytical expressions for the SMAP downscaled brightness temperature due to all these sources of uncertainty. The expressions allow estimation of uncertainty (in kelvin) for each data granule of the SMAP L2_SM_AP product. Since the uncertainties depend on the given ground conditions, e.g., existing water body fraction and local algorithm parameters that depend on vegetation cover and landscape heterogeneity, it is necessary to evaluate the uncertainty for each data granule. In this paper, we show that the uncertainty expressions closely match Monte Carlo simulations with an overall difference of only ~0.1 K. Whereas Monte Carlo estimates of uncertainty can only be afforded for a nominal case (such as those typically reported in Algorithm Theoretical Basis Documents as uncertainty tables), the analytical expressions allow uncertainty estimates for every data granule. The expressions are now used to provide uncertainty standard deviation of downscaled brightness temperature at 9 km in the SMAP L2_SM_AP product. These standard deviations are useful for the following: 1) guidance on the expected level of error in the estimate brightness temperature due to the downscaling process and 2) observation error in direct radiance data assimilation.

Combining SMAP and Sentinel data for high-resolution Soil Moisture product

This presentation illustrates and discusses the possibility of SMAP-Sentinel combined product for the recovery phase of the SMAP mission post radar failure. Initial assessment and results are preliminary and show great promise.