Passive Microwave Soil Moisture Retrieval Research Articles

Coupling Normalization with Moving Window in Backpropagation Neural Network (BNN) for Passive Microwave Soil Moisture Retrieval

A common practice to capture the non-stationary characteristics of the time series data in Artificial Neural Network (ANN) is by randomly dividing the whole set of available data into training, validation and testing, i.e. the data in validation and testing are represented in the training data. Consequently, the usability of the developed model on data not represented by the training data used during the network model development process is always doubtful. In this work, we present a back-propagation neural network (BNN) model trained using one-day history data to predict soil moisture data at 1 km resolution for two future dates. Specifically, high soil moisture values were observed in the training data while the testing data were characterized by drier conditions due to minimal or no rainfall. Our model uses separate mean and standard deviation statistics values from the training and testing data, respectively, to the z-normalized data. With data pre-processed using this method, the BNN model next uses a moving window of size 4 km × 4 km to capture the spatial variability of the soil moisture throughout the 40 km × 40 km study area. The coupling of the normalization and moving window method managed to achieve average soil moisture with Root Mean Square (RMSE) of 3.67% and correlation coefficient, R2 of 0.89. By only using the suggested normalization without the moving window method, the BNN model managed to achieve an average RMSE of barely 5.82% with R2 = 0.83. When comparing with the normal practice of using the same mean and standard deviation statistics of the training data in the testing data, the retrieval accuracy of the BNN model deteriorates to 8.86% with R2 = 0.32. The experiment results demonstrated that the proposed coupling method performed better in terms of both RMSE and R2 for soil moisture retrieval.

Evaluation of six satellite- and model-based surface soil temperature datasets using global ground-based observations

The comprehensive evaluation of model- and satellite-based surface soil temperature (ST) products is a prerequisite for applications of these datasets in hydrology, ecology, and climate change, as well as in passive microwave soil moisture retrieval algorithms. Distinguished from existing regional validations, this study used ground soil temperature observations of approximately 800 stations from 5 sparse and 15 dense networks worldwide to fully assess six model- and satellite-based surface ST products from April 2015 to December 2017. The products consist of five model-based ST from the Modern-Era Retrospective Analysis for Research and Applications, version 2 (MERRA-2), the Goddard Earth Observing System Model version 5 Forward Processing (GEOS-5 FP), the Global Land Data Assimilation System (GLDAS) Noah, the ERA-Interim and the newly released ERA5 produced by the European Centre for Medium Range Weather Forecasts (ECMWF), and one satellite-based ST retrieved from the Advanced Microwave Scanning Radiometer 2 (AMSR2) by using the Land Parameter Retrieval Model (LPRM). The accuracy of these products was comprehensively assessed per availability of ground networks, soil temperature interval, soil moisture interval, land cover, climate zone, and elevation. The results show that the GEOS-5 exhibits the smallest averaged unbiased root mean square difference (ubRMSD) of 1.84 K among all the surface ST products. All model-based products show a high skill in capturing the temporal trend of ground observations with an averaged correlation coefficient larger than 0.97. The ERA5 surface ST obtains visible improvements compared to its predecessor ERA-Interim by showing smaller ubRMSD and absolute bias values. All model-based surface ST products generally show lower values than ground ST, while their bias tends to be warmer as soil temperature and soil moisture increase except for the highest temperature and moisture conditions. Moreover, unstable performance of most model-based products in shrubland and grassland, tropical, arid and cold climate, and high elevation regions is also demonstrated by larger ubRMSD values. The averaged ubRMSD of the satellite-based LPRM ST is 3.04 K, and more attentions should be paid to the impacts of elevations and underlying surfaces on improving this product. These new findings will be valuable for future refinement of ST datasets, algorithms used to estimate soil moisture from satellite data, and applications in various disciplines.

Effect of Structural Uncertainty in Passive Microwave Soil Moisture Retrieval Algorithm.

Passive microwave sensors use a radiative transfer model (RTM) to retrieve soil moisture (SM) using brightness temperatures (TB) at low microwave frequencies. Vegetation optical depth (VOD) is a key input to the RTM. Retrieval algorithms can analytically invert the RTM using dual-polarized TB measurements to retrieve the VOD and SM concurrently. Algorithms in this regard typically use the τ-ω types of models, which consist of two third-order polynomial equations and, thus, can have multiple solutions. Through this work, we find that uncertainty occurs due to the structural indeterminacy that is inherent in all τ-ω types of models in passive microwave SM retrieval algorithms. In the process, a new analytical solution for concurrent VOD and SM retrieval is presented, along with two widely used existing analytical solutions. All three solutions are applied to a fixed framework of RTM to retrieve VOD and SM on a global scale, using X-band Advanced Microwave Scanning Radiometer-Earth Observing System (AMSR-E) TB data. Results indicate that, with structural uncertainty, there ensues a noticeable impact on the VOD and SM retrievals. In an era where the sensitivity of retrieval algorithms is still being researched, we believe the structural indeterminacy of RTM identified here would contribute to uncertainty in the soil moisture retrievals.

Passive/active microwave soil moisture change disaggregation using SMAPVEX12 data

The SMAPVEX12 (Soil Moisture Active Passive (SMAP) Validation Experiment 2012) experiment was conducted during June-July 2012 in Manitoba, Canada with the goal of collecting remote sensing data and ground measurements for the development and testing of soil moisture retrieval algorithms under varying vegetation and soil conditions for the SMAP satellite. The aircraft based soil moisture data provided by the passive/active microwave sensor PALS (Passive and Active L-band System) has a nominal spatial resolution of 1600 m. However, this resolution is not compatible with agricultural, meteorological and hydrological studies that require high spatial resolutions and this issue can be solved by soil moisture disaggregation. The soil moisture disaggregation algorithm integrates radiometer soil moisture retrievals and high-resolution radar observations and it can provide soil moisture estimates at a finer scale than the radiometer data alone. In this study, a change detection algorithm was used for disaggregation of coarse resolution passive microwave soil moisture retrievals with radar backscatter coefficients obtained from the higher spatial resolution UAVSAR (Unmanned Air Vehicle Synthetic Aperture Radar) at crop field scale. The accuracy of the disaggregated change in soil moisture was evaluated using ground based soil moisture measurements collected during SMAPVEX12 campaign. The results showed that soil moisture spatial variabilities were better characterized by the disaggregated change in soil moisture estimates at 5 m/800 m resolution as well as a good agreement between disaggregated estimates and in situ measurements. It also showed that VWC (Vegetation Water Content) did not have a big impact on disaggregation algorithm performance, with R2 of the disaggregated results ranging 0.628–0.794. The 5 m and 800 m resolution disaggregated soil moisture did no show significant difference in statistical performance variables.

Assessment of the impact of spatial heterogeneity on microwave satellite soil moisture periodic error

An accurate temporal and spatial characterization of errors is required for the efficient processing, evaluation, and assimilation of remotely-sensed surface soil moisture retrievals. However, empirical evidence exists that passive microwave soil moisture retrievals are prone to periodic artifacts which may complicate their application in data assimilation systems (which commonly treat observational errors as being temporally white). In this paper, the link between such temporally-periodic errors and spatial land surface heterogeneity is examined. Both the synthetic experiment and site-specified cases reveal that, when combined with strong spatial heterogeneity, temporal periodicity in satellite sampling patterns (associated with exact repeat intervals of the polar-orbiting satellites) can lead to spurious high frequency spectral peaks in soil moisture retrievals. In addition, the global distribution of the most prominent and consistent 8-day spectral peak in the Advanced Microwave Scanning Radiometer – Earth Observing System soil moisture retrievals is revealed via a peak detection method. Three spatial heterogeneity indicators – based on microwave brightness temperature, land cover types, and long-term averaged vegetation index – are proposed to characterize the degree to which the variability of land surface is capable of inducing periodic error into satellite-based soil moisture retrievals. Regions demonstrating 8-day periodic errors are generally consistent with those exhibiting relatively higher heterogeneity indicators. This implies a causal relationship between spatial land surface heterogeneity and temporal periodic error in remotely-sensed surface soil moisture retrievals.

Downscaling of passive microwave soil moisture retrievals based on spectral analysis

ABSTRACTThe retrieval of soil moisture from passive microwave remote-sensing data is presently one of the most effective methods for monitoring soil moisture. However, the spatial resolution of passive microwave soil moisture products is generally low; thus, existing soil moisture products should be downscaled in order to obtain more accurate soil moisture data. In this study, we explore the theoretical feasibility of applying the spectral downscaling method to the soil moisture in order to generate high spatial resolution soil moisture based on both Moderate Resolution Imaging Spectroradiometer and Fengyun-3B (FY3B) data. We analyse the spectral characteristics of soil moisture images covering the east-central of the Tibetan Plateau which have different spatial resolutions. The spectral analysis reveals that the spectral downscaling method is reliable in theory for downscaling soil moisture. So, we developed one spectral downscaling method for deriving the high spatial resolution (1 km) soil moister data from the FY3B data (25 km). Our results were compared with the ground truth measurements from 15 selected experimental days in 16 different sites. The average coefficient of determination (R2) of the spectral downscaling increased nearly doubled than that of the original FY3B soil moisture product. The spectral downscaled soil moister data were successfully applied to examine the water exchange between the land and atmosphere in the study regions. The spectral downscaling approach could be an efficient and effective method to improve the spatial resolution of current microwave soil moisture images.

Four decades of microwave satellite soil moisture observations: Part 2. Product validation and inter-satellite comparisons

Soil moisture is widely recognized as an important land surface variable that provides a deeper knowledge of land-atmosphere interactions and climate change. Space-borne passive and active microwave sensors have become valuable and essential sources of soil moisture observations at global scales. Over the past four decades, several active and passive microwave sensors have been deployed, along with the recent launch of two fully dedicated missions (SMOS and SMAP). Signifying the four decades of microwave remote sensing of soil moisture, this Part 2 of the two-part review series aims to present an overview of how our knowledge in this field has improved in terms of the design of sensors and their accuracy for retrieving soil moisture. The first part discusses the developments made in active and passive microwave soil moisture retrieval algorithms. We assess the evolution of the products of various sensors over the last four decades, in terms of daily coverage, temporal performance, and spatial performance, by comparing the products of eight passive sensors (SMMR, SSM/I, TMI, AMSR-E, WindSAT, AMSR2, SMOS and SMAP), two active sensors (ERS-Scatterometer, MetOp-ASCAT), and one active/passive merged soil moisture product (ESA-CCI combined product) with the International Soil Moisture Network (ISMN) in-situ stations and the Variable Infiltration Capacity (VIC) land surface model simulations over the Contiguous United States (CONUS). In the process, the regional impacts of vegetation conditions on the spatial and temporal performance of soil moisture products are investigated. We also carried out inter-satellite comparisons to study the roles of sensor design and algorithms on the retrieval accuracy. We find that substantial improvements have been made over recent years in this field in terms of daily coverage, retrieval accuracy, and temporal dynamics. We conclude that the microwave soil moisture products have significantly evolved in the last four decades and will continue to make key contributions to the progress of hydro-meteorological and climate sciences. (C) 2017 Elsevier Ltd. All rights reserved.

A novel approach to validate satellite soil moisture retrievals using precipitation data

AbstractA novel approach is proposed that attempts to validate passive microwave soil moisture retrievals using precipitation data (applied over India). It is based on the concept that the expectation of precipitation conditioned on soil moisture follows a sigmoidal convex‐concave‐shaped curve, the characteristic of which was recently shown to be represented by mutual information estimated between soil moisture and precipitation. On this basis, with an emphasis over distribution‐free nonparametric computations, a new measure called Copula‐Kernel Density Estimator based Mutual Information (CKDEMI) is introduced. The validation approach is generic in nature and utilizes CKDEMI in tandem with a couple of proposed bootstrap strategies, to check accuracy of any two soil moisture products (here Advanced Microwave Scanning Radiometer–EOS sensor's Vrije Universiteit Amsterdam‐NASA (VUAN) and University of Montana (MONT) products) using precipitation (India Meteorological Department) data. The proposed technique yields a “best choice soil moisture product” map which contains locations where any one of the two/none of the two/both the products have produced accurate retrievals. The results indicated that in general, VUA‐NASA product has performed well over University of Montana's product for India. The best choice soil moisture map is then integrated with land use land cover and elevation information using a novel probability density function‐based procedure to gain insight on conditions under which each of the products has performed well. Finally, the impact of using a different precipitation (Asian Precipitation‐Highly‐Resolved Observational Data Integration Towards Evaluation of Water Resources) data set over the best choice soil moisture product map is also analyzed. The proposed methodology assists researchers and practitioners in selecting the appropriate soil moisture product for various assimilation strategies at both basin and continental scales.

Parametric exponentially correlated surface emission model for L-band passive microwave soil moisture retrieval

Surface soil moisture is an important parameter in hydrology and climate investigations. Current and future satellite missions with L-band passive microwave radiometers can provide valuable information for monitoring the global soil moisture. A factor that can play a significant role in the modeling and inversion of microwave emission from land surfaces is the surface roughness. In this study, an L-band parametric emission model for exponentially correlated surfaces was developed and implemented in a soil moisture retrieval algorithm. The approach was based on the parameterization of an effective roughness parameter of Hp in relation with the geometric roughness variables (root mean square height s and correlation length l) and incidence angle. The parameterization was developed based on a large set of simulations using an analytical approach incorporated in the advanced integral equation model (AIEM) over a wide range of geophysical properties. It was found that the effective roughness parameter decreases as surface roughness increases, but increases as incidence angle increases. In contrast to previous research, Hp was found to be expressed as a function of a defined slope parameter m=s2/l, and coefficients of the function could be well described by a quadratic equation. The parametric model was then tested with L-band satellite data in soil moisture retrieval algorithm over the Little Washita watershed, which resulted in an unbiased root mean square error of about 0.03m3/m3 and 0.04m3/m3 for ascending and descending orbits, respectively.

Space-based passive microwave soil moisture retrievals and the correction for a dynamic open water fraction

Abstract. The large observation footprint of low-frequency satellite microwave emissions complicates the interpretation of near-surface soil moisture retrievals. While the effect of sub-footprint lateral heterogeneity is relatively limited under unsaturated conditions, open water bodies (if not accounted for) cause a strong positive bias in the satellite-derived soil moisture retrieval. This bias is generally assumed static and associated with large, continental lakes and coastal areas. Temporal changes in the extent of smaller water bodies as small as a few percent of the sensor footprint size, however, can cause significant and dynamic biases. We analysed the influence of such small open water bodies on near-surface soil moisture products derived from actual (non-synthetic) data from the Advanced Microwave Scanning Radiometer for the Earth Observing System (AMSR-E) for three areas in Oklahoma, USA. Differences between on-ground observations, model estimates and AMSR-E retrievals were related to dynamic estimates of open water fraction, one retrieved from a global daily record based on higher frequency AMSR-E data, a second derived from the Moderate Resolution Imaging Spectroradiometer (MODIS) and a third through inversion of the radiative transfer model, used to retrieve soil moisture. The comparison demonstrates the presence of relatively small areas (<0.05) of open water in or near the sensor footprint, possibly in combination with increased, below-critical vegetation density conditions (optical density <0.8), which contribute to seasonally varying biases in excess of 0.2 (m3 m−3) soil water content. These errors need to be addressed, either through elimination or accurate characterisation, if the soil moisture retrievals are to be used effectively in a data assimilation scheme.

Improved Understanding of Soil Surface Roughness Parameterization for L-Band Passive Microwave Soil Moisture Retrieval

Surface roughness parameterization plays an important role in soil moisture retrieval from passive microwave observations. This letter investigates the parameterization of surface roughness in the retrieval algorithm adopted by the Soil Moisture and Ocean Salinity mission, making use of experimental airborne and ground data from the National Airborne Field Experiment held in Australia in 2005. The surface roughness parameter is retrieved from high-resolution (60 m) airborne data in different soil moisture conditions, using the ground soil moisture as input of the model. The effect of surface roughness on the emitted signal is found to change with the soil moisture conditions with a law different from that proposed in previous studies. The magnitude of this change is found to be related to soil textural properties: in clay soils, the effect of surface roughness is higher in intermediate wetness conditions (0.2-0.3 v/v) and decreases on both the dry and wet ends. Consequently, this letter calls for a rethink of surface roughness parameterization in microwave emission modeling.

Evaluation of the SMOS L-MEB passive microwave soil moisture retrieval algorithm

Soil moisture will be mapped globally by the European Soil Moisture and Ocean Salinity (SMOS) mission to be launched in 2009. The expected soil moisture accuracy will be 4.0 %v/v. The core component of the SMOS soil moisture retrieval algorithm is the L-band Microwave Emission of the Biosphere (L-MEB) model which simulates the microwave emission at L-band from the soil–vegetation layer. The model parameters have been calibrated with data acquired by tower mounted radiometer studies in Europe and the United States, with a typical footprint size of approximately 10 m. In this study, aircraft L-band data acquired during the National Airborne Field Experiment (NAFE) intensive campaign held in South-eastern Australia in 2005 are used to perform the first evaluation of the L-MEB model and its proposed parameterization when applied to coarser footprints (62.5 m). The model could be evaluated across large areas including a wide range of land surface conditions, typical of the Australian environment. Soil moisture was retrieved from the aircraft brightness temperatures using L-MEB and ground measured ancillary data (soil temperature, soil texture, vegetation water content and surface roughness) and subsequently evaluated against ground measurements of soil moisture. The retrieval accuracy when using the L-MEB ‘default’ set of model parameters was found to be better than 4.0 %v/v only over grassland covered sites. Over crops the model was found to underestimate soil moisture by up to 32 %v/v. After site specific calibration of the vegetation and roughness parameters, the retrieval accuracy was found to be equal or better than 4.8 %v/v for crops and grasslands at 62.5-m resolution. It is suggested that the proposed value of roughness parameter H R for crops is too low, and that variability of H R with soil moisture must be taken into consideration to obtain accurate retrievals at these scales. The analysis presented here is a crucial step towards validating the application of L-MEB for soil moisture retrieval from satellite observations in an operational context.

The added value of spaceborne passive microwave soil moisture retrievals for forecasting rainfall‐runoff partitioning

Using existing data sets of spaceborne soil moisture retrievals, streamflow and precipitation for 26 basins in the United States Southern Great Plains, a 5‐year analysis is performed to quantify the value of soil moisture retrievals derived from the Tropical Rainfall Measuring Mission (TRMM) Microwave Imager (TMI) X‐band (10.7 GHz) radiometer for forecasting storm event‐scale runoff ratios. The predictive ability of spaceborne soil moisture retrievals is objectively compared to that obtainable using only available rainfall observations and the antecedent precipitation index (API). The assimilation of spaceborne observations into an API soil moisture proxy is demonstrated to add skill to the forecasting of land surface response to precipitation.