Soil Hydraulic Property Estimation Using Remote Sensing: A ReviewA review of recent developments related to soil hydraulic property estimation using remote sensing is presented. Several soil hydraulic parameter estimation techniques using proxi-mal-, air-, or satellite-based remotely sensed soil moisture, land surface temperature, and/or evapotranspiration time series have evolved over the past decades. In particular, micro-wave remote sensing of near-surface soil moisture has played a key role in this respect. Inverse modeling, regression techniques, data assimilation methods by utilizing soil–vegeta-tion–atmosphere transfer models, genetic algorithm based optimization, and uncertainty quantification using ensemble based approaches under synthetic and field conditions have been developed and adapted over these years. In this context multiple approaches for esti-mating effective soil hydraulic parameters at different spatial scales (e.g., point to remote sensing footprints) by various scaling methods are also summarized. Methods reviewed include traditional bottom-up approaches, such as homogenization of local scale properties up to footprint scale, as well as newer top-down approaches, such as use of inverse model-ing of remotely sensed soil moisture or temperature using genetic algorithms, Markov Chain Monte Carlo simulations with data assimilation, and multi-scale Bayesian neural networks. While new remote sensing platforms provide powerful tools and techniques for noninvasive soil hydraulic property estimation of the vadose zone at multiple scales, current limitations including vision for future developments are also summarized.
Abbreviations: AMSR-E, Advanced Microwave Scanning Radiometer; ANN, artificial neural network; BNN, Bayesian framework expanded multi-scale ANN technique; ESA, European Space Agency; ESTAR, Electroni-cally Scanned Thinned Array Radiometer; ET, evapotranspiration; GPR, ground penetrating radar; MCMC, Markov Chain Monte Carlo; NMCGA, Noisy Monte Carlo Genetic Algorithm; PALS, Passive and Active L-band Sensor; PSR, Polarimetric Scanning Radiometer; PTF, pedotransfer function; RS, remote sensing; SMAP, Soil Moisture Active and Passive; SMOS, Soil Moisture and Ocean Salinity; SVAT, soil–vegetation–atmosphere transfer.
Soil hydraulic properties (water retention, hydraulic conductivity) are by far the most important land surface parameters to govern the partitioning of soil moisture between infiltration and evaporation fluxes at a range of spatial scales. However, an obstacle to their practical application at the field, catchment, watershed, or regional scale is the difficulty of quantifying the “effective” soil hydraulic functions q(h) and K(h), where q is the soil water or moisture content, h (or y) is the pressure head, and K is unsaturated hydraulic conductivity. Proper evaluation of the water balance near the land–atmosphere boundary depends strongly on appropriate characterization of soil hydraulic parameters under field conditions and at the appropriate process scale. In recent years a number of approaches have been adopted to tackle this problem including: (i) bottom-up approaches, where larger-scale effective parameters are calculated by aggregating point-scale insitu hydraulic property measurements; (ii) top-down approaches, where effective soil hydraulic parameters are estimated by inverse modeling using remotely sensed soil moisture or soil temperature measurements; and (iii) artificial neural network approaches, where effective soil hydraulic parameters are estimated by exploiting the correlations with soil texture, topographic attributes, and vegetation characteristics at multiple spatial resolutions. In this review we summarize these recent developments.
6Soil Hydraulic FunctionsFor meso- and regional-scale soil–vegetation–atmosphere transfer (SVAT) schemes in hydroclimatic models, pixel dimensions may range from several hundred square meters to several square kilometers. Pixel-scale soil hydraulic parameters and their accuracy are criti-cal for the success of hydroclimatic and soil hydrologic models. Unsaturated flow typically uses closed-form functions to represent water retention characteristics and unsaturated hydraulic conductivities. Leij et al. (1997) provided a review of popular closed-form
A review of recent developments related to soil hydraulic property estimation using remote sensing is presented.
B.P. Mohanty, Dep. of Biological and Agricul-tural Engineering, Texas A&M Univ., College Station, TX 77843. *Corresponding author ([email protected]).
Vadose Zone J. doi:10.2136/vzj2013.06.0100Received 12 June 2013.
expressions for soil water retention and hydraulic conductivity. Gardner’s exponential model of hydraulic conductivity, Brooks and Corey, and van Genuchten soil water retention functions rep-resent some of the most widely used models.
Gardner–Russo Model:
1/(1 )
e1 1
( ) 1 exp2 2
S h h h+bé ùæ ö æ ö÷ ÷ç çê ú= + a - a÷ ÷ç ç÷ ÷ç çê úè ø è øë û
[1a]
shK K e-a= [1b]
Brooks–Corey Model:
Se(y) = (ay)−l when ay > 1 [2a]
Se(y) = 1 when ay £ 1 [2b]
K(y) = Ks(ay)−b when ay > 1 [3a]
K(y) = Ks when ay £ 1 [3b]
where ( 2) 2b=l + +
van Genuchten–Mualem Model:
rese
sat res
( ) 11 | |
m
nh
Sé ùq -q ê ú= = ê úq -q + ayê úë û
[4a]
( ) ( ) ( ){ } ( )2
s 1 1 1m mlmn n nK K
-é ù é ùy = - ay + ay + ayê ú ê úë û ë û [4b]
11m
n= -
where Se is degree of saturation; qs is saturated water content; qres is residual water content, Ks is the saturated hydraulic conductiv-ity; a , b, and l are related to pore-size distributions (note that the significance of these parameters may vary among different models); m and n are empirical parameters; and l is a parameter that accounts for the dependence of the tortuosity and the correla-tion factors on the water content. When these models are used in large heterogeneous scale processes, major questions remain about how to represent hydraulic properties over a heterogeneous soil volume and what averages of hydraulic property shape parameters to use for these models. Studies addressing the impact of averaging methods of shape parameters, parameter correlation, and correla-tion length on ensemble flow behavior in heterogeneous soils have been undertaken by various researchers in the past (e.g., Yeh et al., 1985; Mantoglou and Gelhar, 1987). Other studies develop-ing effective parameters that will predict ensemble behavior of the heterogeneous soils and investigating the effectiveness of the
“effective parameters” in terms of the degree of correlation between parameters for the steady-state and transient evaporation and
infiltration in unsaturated soil have been conducted (Mohanty and Zhu, 2007; Zhu and Mohanty 2002a,b, 2003a,b, 2004, 2006; Zhu et al., 2004, 2006). They also investigated how effective parameters under various flow scenarios and landscape heterogeneity (example scenarios of landscape heterogeneity are shown in Fig. 1a), such as the dryness of the fields, mosaics of soil textures across the land-scape, organization of subsurface soil layers, presence of plants or roots, water management practices as irrigation and drainage, and landscape configuration with various slope, aspect, and elevation may or may not differ.
6 Inverse Modeling and Effective Soil Hydraulic Parameterization
To address the land surface heterogeneity issue in the context of developing effective soil properties at a remote sensing footprint scale, Camillo et al. (1986) was among some of the early ones who envisioned that soil hydraulic properties may be estimated using microwave based soil moisture and soil temperature that drive water and energy fluxes near the land surface (Fig. 1b). Feddes et al. (1993) further asserted that the inverse modeling on the basis of soil moisture status at the top and bottom of the soil profile, cumulative areal evapotranspiration, and cumulative percolation to groundwater for estimating effective soil hydraulic parameters (based on small scale soil physics) at large scale is feasible under different land cover conditions. In addition, local scale studies by Ahuja et al. (1993) and Chen et al. (1993) showed successes in describing average soil hydraulic conductivity of the soil profile by using the changes in surface soil moisture. Using different soil textures, they observed that the strongest relationship between surface soil moisture and profile soil hydraulic conductivity exists for sandy soils and the relationship weakens with higher clay con-tent. In addition, they discovered hydroclimatic conditions with enhanced evaporation may also affect the parameters estimation. Burke et al. (1997, 1998) provided a calibration technique for soil water and energy balance model for estimating soil water reten-tion and hydraulic conductivity parameters. They investigated saturated hydraulic conductivity, bulk density, air entry pressure, and other soil hydraulic parameters. Hollenbeck et al. (1996) and Mattikalli et al. (1998) developed regression based tools for relat-ing airborne-based multi-temporal passive microwave (L band) brightness temperature values (which relates to soil moisture status) with saturated hydraulic conductivity and soil texture for remote sensing footprints across large regions. Chang and Islam (2000) developed an artificial neural network scheme for estimat-ing soil textural class from remotely sensed brightness temperature data during Washita 92 field campaign in Oklahoma. Uri et al. (1991) utilized soil moisture based on passive microwave radiom-etry (1.46 GHz) in the evaluation of drainage patterns at different locations of a large watershed leading to shallow subsurface soil hydraulic property estimation. With the availability of near sur-face soil moisture using passive and active microwave sensing from air-based sensors (e.g., Passive and Active L-band Sensor [PALS],
www.VadoseZoneJournal.org p. 3 of 9
Electronically Scanned Th inned Array Radiometer [ESTAR], and Polarimetric Scanning Radiometer [PSR]) and space-based sensors (e.g., Advanced Microwave Scanning Radiometer [AMSR-E], Soil Moisture and Ocean Salinity [SMOS], AQUARIUS) in the past decade, estimating soil hydraulic parameters under dif-ferent hydrologic and climatic conditions has been studied by several research groups. Among others, Santanello et al. (2007) parameterized soil hydraulic parameters and soil texture using a combination of passive microwave remote sensing soil moisture of NASA’s push-broom microwave radiometer (PBMR, L-band, 21 cm) and active microwave RADARSAT-1 (C-band; 5.6 cm) imagery, a parameter estimation algorithm (PEST; Doherty, 2004) and Noah Land Surface Model in a semiarid watershed and com-pared the estimation to pedotransfer function (PTF) based soil parameters. Th ey suggested that satellite based remotely sensed soil moisture data becoming widely available on a regular basis in the recent years at the global scale, the prospect of quantify-ing such eff ective parameters would be more a reality at multiple scales. Other studies (e.g., Hogue et al., 2005; Peters-Lidard et al.,
2008) have shown success in calibrating soil hydraulic properties at large scale using surface soil moisture from remote sensing. It is worth mentioning that there have been mixed results for esti-mating surface soil moisture using active microwave (radar-based) remote sensing, due to sensitivity of high frequency backscatter to the nature and degree of surface interaction. In contrast, pas-sive microwave sensing has been found to be most successful to measure land surface soil moisture and holds the most promise to estimate profi le soil hydraulic properties.
Besides volumetric soil moisture content (related to land surface emissivity), land surface temperature measurements have also been investigated to produce surface soil hydraulic properties. Using latent heat fl ux (evapotranspiration) as the matching variable, Gutmann and Small (2007) concluded that inverse modeling with soil skin temperature measured by remotely sensed infrared measurements provide more accurate soil hydraulic parameters for land surface models than soil texture based PTF schemes. In yet another follow-up study, Gutmann and Small (2010) evaluated the
Fig. 1. (a) Heterogeneous landscapes with diff erent vegetation, soils, and topography and (b) conceptual framework for inverse estimation of soil hydraulic properties using near-surface soil moisture time series data from remote sensing platform.
www.VadoseZoneJournal.org p. 4 of 9
use of MODIS based surface temperature for estimating landscape hydraulic properties, based on indirect relationship among soil temperature, soil moisture, and evapotranspiration. Th ey argued that under complex soil, topographic and vegetation heterogene-ity in large landscapes, dominant water fl ow processes including infi ltration, evaporation, transpiration, and surface runoff may vary across space and time scales and thus can be eff ectively repre-sented by landscape hydraulic properties instead of small-scale soil hydraulic properties. Th ey also showed that using the inversely esti-mated soil hydraulic properties decreases error in modeled latent and sensible heat fl uxes by more than 60% compared to standard texture based classifi cation. Using infrared-measured soil surface temperature and time domain refl ectometry–based soil water con-tents in a framework coupling HYDRUS_1D model with global optimizer (DREAM), Steenpass et al. (2010) inversely estimated van Genuchten–Mualem parameters for synthetic and real fi eld conditions. Th ey also compared the parameters obtained by using single objective (soil surface temperature) and multi-objective (soil surface temperature and water content) in their study.
Currently, soil moisture data from remote sensing (RS) are lim-ited to the near-surface soil layers. Under minimal vegetation cover, the maximum penetration depth of a microwave L-band sensor is about 5 cm. Many studies have been done to retrieve soil moisture profi les using these passive and active microwave brightness temperature data. Currently deployed SMOS mission by European Space Agency (ESA), AQUARIUS a NASA mis-sion, and the upcoming Soil Moisture Active and Passive (SMAP) mission of NASA in +2014 (Fig. 2a) are designed to populate soil moisture time series data across the globe for this decade (2010–2020). Soil moisture retrieved from these satellite platforms are further calibrated and validated at multiple scales using airborne microwave sensors and handheld or ground-based sensors (Fig. 2b) and pilot studies in various watersheds and hydroclimatic regions across the globe.
Researchers have taken this unique opportunity to develop and evaluate advanced inverse estimation methods for soil hydraulic properties using remote sensing data and their scaling behavior recently. Using time series of passive microwave (L-band) based soil moisture for airborne (e.g., ESTAR, PSR) and satellite (e.g., AMSR-E) footprints, Ines and Mohanty (2008a, b, 2009) developed Genetic Algorithm based inverse modeling and data assimilation methods for quantifying effective soil hydraulic parameters at the remote sensing footprint scale. Th ey conducted several numerical and fi eld studies for various possible landscape scenarios including diff erent soil textures (coarse to fi ne), ground-water table depths (shallow to deep), land covers (bare to vegetated), and climatic (arid to humid) conditions. Th e basic premises of their work include if surface soil moisture temporal dynamics hold the memory of hydrologic fl uxes (infi ltration, redistribution, evapo-transpiration) occurring across the entire vadose zone and thus can provide a representation of soil hydraulic properties of the entire
domain in an eff ective manner. Scope of their study also evaluated if the method is robust to handle various nonlinearities and uncer-tainties under diff erent hydrologic and climatic conditions. While theoretically the method can be adapted for any soil hydraulic conductivity function and land surface model, van Genuchten–Mualem model was used within SWAP model framework (van Dam et al., 1997) in their studies. Near-surface soil moisture data was used to derive eff ective q(h) and K(h) for the vadose zone for each remote sensing pixel assuming the constitutive functions of Mualem–van Genuchten and vertical soil water movement in the unsaturated zone defi ned by the Richards equation. A fl ow chart representation of the assimilation process is shown in Fig. 3. Sample results from this approach are shown in Fig. 4. Extending their original work, Ines and Mohanty (2008c) developed the concept of Noisy Monte Carlo Genetic Algorithm (NMCGA) to formulate an ensemble-based general method for estimating the eff ective soil hydraulic parameters and their uncertainties at the remote sensing footprint scale. Th e main assumption of the domain-dependent parameter estimation concept is based on the idea that the eff ective forms of the soil hydraulic functions (at the RS footprint) can be inferred by the derived eff ective soil hydraulic parameters from large-scale RS soil moisture data inversion with no lateral exchange of fl ow across adjacent RS footprints. In the synthetic case studies under pure (one soil texture) and mixed-pixel (multiple soil textures) conditions, NMCGA performed well in estimating the eff ective soil hydraulic parameters even with pixel complexities contributed by various soil types and land manage-ment practices (rainfed/irrigated). With the airborne and satellite remote sensing cases, NMCGA also performed well for estimating eff ective soil hydraulic properties so that when applied in forward stochastic simulation modeling it can mimic large-scale soil mois-ture dynamics. Th eir results also suggest a possible scaling down eff ect of the eff ective soil water retention curve q(h) at the larger satellite remote sensing pixel compared to airborne remote sensing pixel. Expanding on the inverse modeling and data assimilation idea, Shin et al. (2012) developed a layer-specifi c soil moisture assimilation scheme by interjecting subsurface soil moisture
Fig. 2. (a) NASA’s upcoming Soil Moisture Active Passive (SMAP) satellite mission to be launched in 2014. (b) NASA’s Twin Otter aircraft and ground sensor deployment during multi-scale cal/val fi eld campaigns of satellite-based microwave sensors.
www.VadoseZoneJournal.org p. 5 of 9
information. Th ey demonstrated that soil layers and vertical het-erogeneity (layering sequences) could impact the uncertainty of quantifying eff ective soil hydraulic parameters. Using layer-specifi c assimilation, they found further improvement of soil hydraulic parameter estimation and their accuracy.
Evaporation and transpiration plays a complementary role to root zone soil moisture for soil water balance. Depending on root water uptake and soil evaporation, soil water may redistribute across soil profi le maintaining certain functional thresholds. Adding evapo-transpiration (ET) (estimated by S-SEBI and SEBAL algorithms
with MODIS data) to the objective function in the soil hydraulic property estimation scheme shown in Fig. 3, Shin et al. (2013) presented the improvement of eff ective soil hydraulic parameter estimation in comparison to values based on single objective func-tion (soil moisture content only) in diff erent hydroclimates. Note, however, that technical issues relating to ET measurements using remote sensing tools persist to date. Further development on ET estimation using remote sensing techniques in the future will refi ne the eff ective soil hydraulic property estimation. Uncertainties of water fl uxes in SVAT model by inverting soil moisture and evapo-transpiration in 22 hydroclimates were analyzed by Pollacco and
Fig. 3. Implementation of near-surface soil moisture assimilation scheme (aft er Ines and Mohanty, 2008a).
Fig. 4. Eff ects of (a, b, e, and f ) initial and bottom boundary conditions, (b, c, g, and f ) rooting depth, and (b, d, f, and h) root density on eff ective water retention q(h) and hydraulic conductivity K(h) estimation for the Oklahoma Southern Great Plains 1997 (SGP97) Site ARS-135 for a loam soil (BBC is bottom boundary condition, RZmax is maximum rooting depth, RZdensity is root density, and Dassim denotes predicted by the inverse modeling based near-surface soil moisture assimilation scheme) (aft er Ines and Mohanty, 2008a).
www.VadoseZoneJournal.org p. 6 of 9
Mohanty (2012). The primary conclusions from their study include uncertainties in simulated water fluxes increases as (i) the climate gets drier, (ii) the texture gets coarser, or (iii) the roots grow deeper. They attributed the uncertainty in recharge to soil moisture and transpiration decoupling. Soil moisture decoupling occurs when the information provided by surface flux is no longer representative of root zone flux. Transpiration decoupling occurs when there is substantially more water storage at depth. These new root zone ecohydrologic process understanding will enhance accuracy of effective soil hydraulic property estimation at the landscape scale in the future.
Besides microwave remote sensing, Vereecken et al. (2007) also suggested that combination of field-scale proximal-sensing techniques and hydrologic models provides the platform for accounting the upscaled behavior of the soil hydraulic prop-erties and flow processes in heterogeneous soils. In that spirit, Lambot et al. (2009) developed a technique using full-waveform integrated hydro-geophysical inversion of time-lapse off-ground ground penetrating radar (GPR). One of the advantages of this technique over the microwave techniques described earlier is deeper penetration depth.
6 Scaling Up or Down Soil Hydraulic Parameters
An alternative to the inverse estimation technique described above, Das et al. (2008) developed a Markov Chain Monte Carlo (MCMC) based data assimilation algorithm to derive upscaled land surface hydraulic parameters (from apriori local-scale param-eter distributions) for a SVAT model using time series data of satellite-measured atmospheric forcings (e.g., precipitation), and land surface states (e.g., soil moisture and vegetation). Their study focused especially on the evaluation of soil moisture measurements of the Aqua satellite based AMSR-E instrument using the MCMC-based scaling algorithm. Soil moisture evolution was modeled at a spatial scale comparable to the AMSR-E soil moisture product, with the hypothesis that the characterization of soil microwave emissions and their variations with space and time on soil surface within the AMSR-E footprint can be represented by an ensemble of upscaled soil hydraulic parameters. They demonstrated the fea-tures of the MCMC-based parameter upscaling algorithm (from field to satellite footprint scale) within a SVAT model framework to evaluate the satellite-based brightness temperature/soil moisture measurements for different hydro-climatic regions, and identified the temporal effects of vegetation (leaf area index) and other envi-ronmental factors on AMSR-E based remotely sensed soil moisture data. The SVAT modeling applied for different hydroclimatic regions revealed the limitation of AMSR-E measurements in high-vegetation regions. The study also suggested that inclusion of soil moisture evolution from the proposed upscaled SVAT model with AMSR-E measurements in data assimilation routine will improve the quality of soil moisture assessment in a footprint scale. The
technique also has the potential to derive upscaled parameters of other geophysical properties used in remote sensing of land surface states. The developed MCMC algorithm with SVAT model can be very useful for land–atmosphere interaction studies and further understanding of the physical controls responsible for soil mois-ture dynamics at different scales. Subsequently, Das et al. (2010) proposed an algorithm that uses the Karhunen–Loève expansion in conjunction with the MCMC technique, which employs mea-sured soil moisture values to characterize the saturated hydraulic conductivity distributions of an agricultural field at a 30-m resolu-tion. In a similar line, Montzka et al. (2011) developed a particle filter data assimilation scheme using ALOS and SMOS satellite based soil moisture data and HYDRUS-1D model for deriving van Genuchten soil hydraulic parameters.
Indirect estimation techniques using PTFs (e.g., Sharma et al., 2006) provide an effective alternative to direct measurements and/or to inverse modeling and data assimilation tools for large scale soil hydraulic parameters. This approach examines the effect of including topographic and vegetation attributes, besides pedo-logic attributes, on the prediction of soil hydraulic properties using PTFs. With the increasing availability of remote sensing products from air- and spaceborne sensors at different spatial scales, topographic and vegetation attributes are easily available from digital elevation models and normalized difference vegeta-tion index. Traditionally, PTFs have been used to estimate the required soil hydraulic parameters from other available or easily measurable soil properties. While most previous studies derive and adopt these parameters at matching spatial scales (1:1) of input and output data, Jana et al. (2007) developed a methodology to derive soil water retention functions at the point or local scale using the PTFs trained with coarser scale input data. This study was a novel application of an artificial neural network (ANN)-based PTF scheme across two spatial support scales within the Rio Grande basin in New Mexico. The ANN was trained using soil texture and bulk density data from the SSURGO database (scale 1:24,000) and then used for predicting soil water contents at differ-ent pressure heads with point-scale data (1:1) inputs. The resulting outputs were corrected for bias before constructing the soil water characteristic curve using the van Genuchten equation. A hierar-chical approach with training data derived from multiple clustered subwatersheds (with varying spatial extent) was used to study the effect of the increase in spatial extent. The results show good agree-ment between the soil water retention curves constructed from the ANN-based PTFs and field observations at the local scale near Las Cruces, NM. The robustness of the multiscale PTF methodology was further tested with a separate data set from the Little Washita watershed region in Oklahoma. Overall, ANN coupled with bias correction was found to be a suitable approach for deriving soil hydraulic parameters at a finer scale from soil physical properties at coarser scales and across different spatial extents. Later, using nonlinear bias correction and cumulative distribution function matching, Jana et al. (2008) developed a Bayesian framework to
www.VadoseZoneJournal.org p. 7 of 9
expand their multi-scale ANN technique (also called BNN) for estimating soil hydraulic parameters at multiple scales including their uncertainties. Further, Jana and Mohanty (2011) and Jana et al. (2012) evaluated their BNN technique using several remotely sensed and in situ land surface data in different hydroclimates. Their findings suggested that the BNN-based approach could be used for up or downscaling soil hydraulic parameters effectively.
More recently, using remote sensing data, Jana and Mohanty (2012) evaluated various soil hydraulic properties estimation and scaling scheme in the context of watershed hydrology. The equivalence of the upscaled parameters was tested by simulating water flow for the watershed pixels in HYDRUS-3D, and comparing the resultant soil moisture states with data from the ESTAR airborne sensor during the SGP97 hydrology experiment.
6 Current Challenges and Future Venues for Development or Improvement
Using inverse estimation and/or spatial scaling based on micro-wave remote sensing of soil moisture proved to be one of the most potent venues for soil hydraulic property estimation from local to regional scales. While these methods provide the initial break-through for estimating some of the most challenging properties of soil (effective hydraulic properties with large support volume) needed for many hydrologic, land–atmosphere interaction, and environmental applications, it is far from optimal. Among others, primary limitations of these approaches and possible ways forward to develop and improve them further include:
1. Penetration depth of microwave remote sensing is limited to approximately top few centimeters (?0–5 cm for L-Band radi-ometers) of soil depth, where many land surface disturbances and soil structures play a significant role resulting in errors in the dynamics of soil hydrologic processes. In the future, microwave sensors with different wavelengths (e.g., p-band) penetrating to deeper depths may able to minimize such errors. Deployment of such a sensor antenna aboard a satellite plat-form, however, is an engineering challenge in the near future. Alternatively, soil hydrologic models with data assimilation approach (e.g., ensemble Kalman Filter) by interjecting various local, proximal, or airborne soil water measurements at deeper depths and other surrogate variables could extend the accuracy of effective soil hydraulic parameterization schemes.
2. Accuracy of soil moisture retrieval using brightness tempera-ture (based on soil emissivity) is currently being affected by a number of factors such as type of sensor (i.e., bandwidth/frequency, polarization, scanning pattern, and footprint size), within-footprint heterogeneity of soil texture, vegetation, and landscape (topographic) features, as well as limited observation of vegetation water content and temperature (affected by both vegetation and soil temperature). Furthermore, for maintaining
a certain level of accuracy, microwave sensing for soil moisture is only limited to the regions with vegetation biomass less than a prescribed threshold value (e.g., L-band radiometer retrieval up to 4 Kg m−2) and does not work for densely vegetated areas. Future efforts should be geared at developing improved soil moisture retrieval algorithms categorized by remote sensing platforms by explicitly accounting for type of sensor, footprint size, and within-footprint heterogeneity, and by hydroclimatic regions. In this context, using other remote sensing spectrum (e.g., optical, thermal, hyper-spectral, full range of spectra), and proximal sensing (e.g., in situ sampling, spectroscopy, electro-magnetic induction, electrical resistivity tomography, GPR) tools for different soil attributes (moisture, texture, organic carbon, mineralogy, iron content, salinity, and other proxies), as described by Ben-Dor et al. (2009), Mulder et al. (2011), and Casa et al. (2013), can possibly be fused and translated to more complex soil hydraulic properties by developing PTFs. These values can subsequently be upscaled or downscaled matching to the application.
3. Decoupling of surface and subsurface moisture in various hydroclimates, such as done by Pollacco and Mohanty (2012), Pollacco et al. (2013), and Shin et al. (2012), under various soil textures, plant rooting depths, and precipitation condi-tions, reduces accuracy of estimated soil hydraulic parameters using only surface soil moisture content observed by remote sensing in the objective function for parameter optimization. By adding weighted ET as a second criterion in the optimi-zation formulation improves the efficacy of the soil hydraulic parameter estimation method. However, measurement/esti-mation of ET at the corresponding (matching to soil moisture measurement) spatiotemporal scale using remote sensing is far from being perfect. Issues such as boundary layer turbulence effect, energy balance closure, and others remain to be further developed. Concerted efforts of measuring soil water, vegeta-tion water, and ET using co-located remote sensing platforms with matching resolution will provide maximum impetus for the progress soil hydrologic science and its applications at dif-ferent scales.
4. Spatial and temporal resolution of currently available remote sensing data may not be optimum for capturing all the hydro-logic processes (e.g., overland f low across remote sensing footprints) including the extreme scenarios (e.g., complete saturation and preferential flow during precipitation) within a remote sensing footprint resulting in (nonoptimal) soil hydrau-lic parameters, which may fail to describe the entire spectrum of hydrologic events. Upcoming satellite sensors as SMAP will generate multiresolution (3, 9, and 36 km) soil moisture products, providing the opportunity for generating multiscale effective soil hydraulic properties. In this context, however, it should be pointed out that computational cost for using fine spatial and temporal resolution data in inverse modeling and data assimilation schemes may possibly be far greater with only
www.VadoseZoneJournal.org p. 8 of 9
a small gain in accuracy of effective soil hydraulic parameters. Users are advised to make informed decisions on the accuracy required, and thus grid resolution (for downscaling or upscal-ing) based on specific application, extent of hydrologic domain, landscape complexity, and computational burden.
5. Discrepancy in soil hydrologic process scale based on established continuum-scale capillary bundle theory leading to Darcy and Richards type model formulations as opposed to the underlying dominant hydrologic processes at the support scale of remote sensing footprint. Although the formalized capillary-based governing flow equation may be best suited for local f low processes and provide the basis of currently available parameter estimation schemes, the physical basis of their adoption at the scale of the remote sensing footprint is debated. Alternative formulations based on thermodynamic principles, hydrologic functionality rules, or scale-appropriate water balance models across different irregular hydrologic units (e.g., catchments, watersheds, river basins) using existing or novel remote sensing attributes or indices may prove to be more effective for describing large scale soil hydrology.
Although limitations for remote sensing technology persist in terms of penetration depth, spatial and temporal resolutions, accuracy, and associated soil hydrologic process description at the footprint scale, the above studies show its great promise for estimating effective soil hydraulic properties and its application beyond continuum scale. The outlook of earth observing systems of various space agencies (e.g., NASA, ESA, JAXA) using satel-lite platforms is promising in the coming decade, and hydrologists and soil scientists should take the opportunity to use them more regularly for mapping soil hydrologic and other properties at the watershed, regional, continental, and global scales, supporting vari-ous Earth systems science applications.
ReferencesAhuja, L.R., O. Wendroth, and D. Nielsen. 1993. Relationship between initial
drainage of surface soil and average profile saturated conductivity. Soil Sci. Soc. Am. J. 57:19–25. doi:10.2136/sssaj1993.03615995005700010005x
Ben-Dor, E., S. Chabrillat, J.A.M. Dematte, G.R. Taylor, J. Hill, M.L. Whiting, and S. Sommer. 2009. Using imaging spectroscopy to study soil properties. Re-mote Sens. Environ. 113:S38–S55. doi:10.1016/j.rse.2008.09.019
Burke, E.J., R.J. Gruney, L.P. Simmonds, and P.E. O’Neill. 1998. Using a mod-eling approach to predict soil hydraulic properties from passive micro-wave measurements. IEEE Trans. Geosci. Rem. Sens. 36(2):454–460. doi:10.1109/36.662729
Burke, E.J., R.J. Gruney, L.P. Simmonds, and T.J. Jackson. 1997. Calibrating a soil water and energy budget model with remotely sensed data to obtain quantitative information about the soil. Water Resour. Res. 33:1689–1697. doi:10.1029/97WR01000
Casa, R.F. Castaldi, S. Pascucci, B. Basso, and S. Pignatti. 2013. Geophysical and hyperspectral data fusion techniques for in-field estimation of soil proper-ties. Vadose Zone J. 12. doi:10.2136/zj2012.0201.
Chang, D.-H., and S. Islam. 2000. Estimation of soil physical properties using re-mote sensing and artificial neural network. Remote Sens. Environ. 74:534–544. doi:10.1016/S0034-4257(00)00144-9
Chen, C., D.M. Thomas, R.E. Green, and R.J. Wagenet. 1993. Two-domain es-timation of hydraulic properties in macropore soils. Soil Sci. Soc. Am. J. 57:680–686. doi:10.2136/sssaj1993.03615995005700030008x
Das, N.N., B.P. Mohanty, and Y. Efendiev. 2010. Characterization of effective sat-urated hydraulic conductivity in an agricultural field using Karhunen‐Loève expansion with the Markov chain Monte Carlo technique. Water Resour. Res. 46:W06521. doi:10.1029/2008WR007100
Das, N.N., B.P. Mohanty, and E.G. Njoku. 2008. An Markov chain Monte Carlo algorithm for upscaled soil-vegetation-atmosphere-transfer modeling to evaluate satellite-based soil moisture measurements. Water Resour. Res. 44:W05416. doi:10.1029/2007WR006472
Doherty, J. 2004. PEST: Model independent parameter estimation. User manu-al. 5th ed. Watermark Numerical Computing, Brisbane, Australia.
Feddes, R.A., M. Menenti, P. Kabat, and W.G.M. Bastiaanssen. 1993. Is large-scale inverse modeling of unsaturated flow with areal average evaporation and surface soil moisture as estimated from remote sensing feasible? J. Hy-drol. 143:125–152. doi:10.1016/0022-1694(93)90092-N
Gutmann, E.D., and E.E. Small. 2010. A method for the determination of the hy-draulic properties of soil from MODIS surface temperature for use in land‐sur-face models. Water Resour. Res. 46:W06520. doi:10.1029/2009WR008203
Gutmann, E.D., and E.E. Small. 2007. A comparison of land surface model soil hydraulic properties estimated by inverse modeling and pedotransfer func-tions. Water Resour. Res. 43:W05418. doi:10.1029/2006WR005135
Hogue, T.S., L. Bastidas, H. Gupta, S. Sorooshian, K. Mitchell, and W. Emmerich. 2005. Evaluation and transferability of the Noah land surface model in semi-arid environments. J. Hydrometeorol. 6:68–84. doi:10.1175/JHM-402.1
Hollenbeck, K.J., T.J. Schmugge, G.M. Hornberger, and J.R. Wang. 1996. Identi-fying soil hydraulic heterogeneity by detection of relative change in passive microwave remote sensing observations. Water Resour. Res. 32:139–148. doi:10.1029/95WR02916
Ines, A.V.M., and B.P. Mohanty. 2008a. Near-surface soil moisture assimilation for quantifying effective soil hydraulic properties under different hydro-climatic conditions. Vadose Zone J. 7:39–52. doi:10.2136/vzj2007.0048
Ines, A.V.M., and B.P. Mohanty. 2008b. Near-surface soil moisture as-similation to quantify effective soil hydraulic properties using genetic algorithm. 1. Conceptual modeling. Water Resour. Res. 44:W06422. doi:10.1029/2007WR005990.
Ines, A.V.M., and B.P. Mohanty. 2008c. Parameter conditioning with a noisy Monte Carlo genetic algorithm to estimate effective soil hydraulic properties from space. Water Resour. Res. 44:W08441. doi:10.1029/2007WR006125.
Ines, A.V.M., and B.P. Mohanty. 2009. Near-surface soil moisture assimilation to quantify effective soil hydraulic properties using genetic algorithm. 2. With air-borne remote sensing during SGP97 and SMEX02. Water Resour. Res. 45. doi:10.1029/2008WR007022
Jana, R.B., B.P. Mohanty, and E.P. Springer. 2007. Multi-scale pedo-transfer functions for soil water retention. Vadose Zone J. 6:868–878. doi:10.2136/vzj2007.0055
Jana, R., B.P. Mohanty, and E. Springer. 2008. Bayesian implementation of multi-scale pedo-transfer functions with non-linear bias corrections. Water Resour. Res. doi:10.1029/2008WR006879.
Jana, R., and B.P. Mohanty. 2011. Enhancing PTFs with remotely sensed data for multi-scale soil water retention estimation. J. Hydrol. doi:10.1016/i.jhy-drol.2010.12.043.
Jana, R., and B.P. Mohanty. 2012. A comparative study of multiple approaches to soil hydraulic parameter scaling applied at the hillslope scale. Water Re-sour. Res. 48:W02520. doi:10.1029/2010WR010185.
Jana, R., B.P. Mohanty, and Z. Sheng. 2012. Upscaling soil hydraulic parameters in the Picacho Mountain region using Bayesian neural networks. Trans. AS-ABE 55:463–473.
Lambot, S., E. Slob, J. Rhebergen, O. Lopera, K.Z. Jadoon, and H. Vereecken. 2009. Remote estimation of the hydraulic properties of a sand using full-waveform integrated hydrogeophysical inversion of time-lapse, off-ground GPR data. Vadose Zone J. 8:743–754. doi:10.2136/vzj2008.0058
Leij, F.J., W.B. Russell, and S.M. Lesch. 1997. Closed-form expressions for water retention and conductivity data. Ground Water 35:848–858. doi:10.1111/j.1745-6584.1997.tb00153.x
Mantoglou, A., and L.W. Gelhar. 1987. Stochastic modeling of large-scale tran-sient unsaturated flow systems. Water Resour. Res. 23:37–46. doi:10.1029/WR023i001p00037
Mattikalli, N.M., E.T. Engman, T.J. Jackson, and L.R. Ahuja. 1998. Microwave remote sensing of temporal variations of brightness temperature and near-surface soil water content during a watershed-scale field experiment, and its application to the estimation of soil physical properties. Water Resour. Res. 34:2289–2299. doi:10.1029/98WR00553
Mohanty, B.P., and J. Zhu. 2007. Effective hydraulic parameters in horizontally and vertically heterogeneous soils for steady state land-atmosphere inter-action. J. Hydrometeorol. 8:715–729. doi:10.1175/JHM606.1
Montzka, C., M. Moradkhani, L. Weihermuller, H.-J.H. Franssen, M. Canty, and H. Vereecken. 2011. Hydraulic parameter estimation by remotely-sensed top soil moisture observations with the particle filter. J. Hydrol. 399:410–421. doi:10.1016/j.jhydrol.2011.01.020
www.VadoseZoneJournal.org p. 9 of 9
Mulder, V.L., S. de Bruin, M.E. Schaepman, and T.R. Mayr. 2011. The use of remote sensing in soil and terrain mapping- A review. Geoderma 162:1–19. doi:10.1016/j.geoderma.2010.12.018
Peters-Lidard, C.D., D.M. Mocko, M. Garcia, J.A. Santanello, M.A. Tischler, M.S. Moran, and Y.H. Wu. 2008. Role of precipitation uncertainty in the estimation of hydrologic soil properties using remotely sensed soil moisture in a semiarid environment. Water Resour. Res. 44:W05S18. doi:10.1029/2007WR005884
Pollacco, J.A., and B.P. Mohanty. 2012. Uncertainties of water fluxes in SVAT models: Inverting surface soil moisture and evapotranspiration retrieved from remote sensing. Vadose Zone J. 11. doi:10.2136/vzj2011.0167.
Pollacco, J.A.P., B.P. Mohanty, and A. Efstratiadis. 2013. The weighted objec-tive function selector algorithm and its application to invert surface soil moisture and evapotranspiration in SVAT models. Water Resour. Res. doi:10.1002/wrcr.20554.
Santanello, J.A., C.D. Peters-Lidard, M.E. Garcia, D.M. Mocko, M.A. Tischler, M.S. Moran, and D.P. Thoma. 2007. Using remotely-sensed estimates of soil moisture to infer soil texture and hydraulic properties across a semi-arid watershed. Remote Sens. Environ. 110:79–97. doi:10.1016/j.rse.2007.02.007
Sharma, S.K., B.P. Mohanty, and J. Zhu. 2006. Including topography and veg-etation attributes for developing pedo transfer functions in Southern Great Plains of USA. Soil Sci. Soc. Am. J. 70:1430–1440. doi:10.2136/ss-saj2005.0087
Shin, Y., B.P. Mohanty, and A.V.M. Ines. 2012. Soil hydraulic properties in one-dimensional layered soil profile using layer-specific soil moisture assimila-tion scheme. Water Resour. Res. 48:W06529. doi:10.1029/2010WR009581
Shin, Y., B.P. Mohanty, and A.V.M. Ines. 2013. Estimating effective soil hydraulic properties using spatially distributed soil moisture and evapotranspiration. Vadose Zone J. 12. doi:10.2136/vzj2012.0094.
Steenpass, C., J. Vanderborght, M. Herbst, J. Simunek, and H. Vereecken. 2010. Estimating soil hydraulic properties from infrared measurements of soil sur-face temperatures and TDR data. Vadose Zone J. 9:910–924. doi:10.2136/vzj2009.0176
Uri, U., P. Tambet, G.G. Shukin, V.A. Kashevar, Y.V. Kybakov, and Y.M. Shulgina. 1991. Use of SHF-radiometry during the evaluation of the land-reclamation state of drained earth (based on experiments in Estonia). Trudy Metody
Aktivony I Passivnony Radiolokatsii v Meterologii. Gidrometeoizdat, Lenin-grad. p. 46–50.
van Dam, J.C., J. Huygen, J.G. Wesseling, R.A. Feddes, P. Kabat, P.E.V. van Wal-sum, et al. 1997. Theory of SWAP version 2.0. Rep. 71. Dep. Water Resour., Wageningen Agric. Univ., Wageningen, the Netherlands.
Vereecken, H., R. Kasteel, J. Vanderborght, and T. Harter. 2007. Upscaling hy-draulic properties and soil water flow processes in heterogeneous soils: A review. Vadose Zone J. 6:1–28. doi:10.2136/vzj2006.0055
Yeh, T.-C.J., L.W. Gelhar, and A.L. Gutjahr. 1985. Stochastic analysis of unsatu-rated flow in heterogeneous soils: 1. Statistically isotropic media. Water Resour. Res. 21:447–456. doi:10.1029/WR021i004p00447
Zhu, J., and B.P. Mohanty. 2002a. Spatial averaging of van genuchten hydrau-lic parameters for steady state flow in heterogeneous soils. Vadose Zone J. 1:261–272.
Zhu, J., and B.P. Mohanty. 2002b. Upscaling of hydraulic properties for steady state evaporation and infiltration. Water Resour. Res. 38:1178. doi:10.1029/2001WR000704.
Zhu, J., and B.P. Mohanty. 2003a. Effective hydraulic parameters for steady state flows in heterogeneous soils. Water Resour. Res. 39(8):1227. doi:10.1029/2002WR001831
Zhu, J., and B.P. Mohanty. 2003b. Upscaling of hydraulic properties of hetero-geneous soils In: Methods of scaling in soil physics. CRC Press, Boca Raton, FL. p. 97–117.
Zhu, J., and B.P. Mohanty. 2004. Soil hydraulic parameter upscaling for steady-state flow with root water uptake. Vadose Zone J. 3:1464–1470.
Zhu, J., and B.P. Mohanty. 2006. Effective scaling factor for transient infiltra-tion in heterogeneous soils. J. Hydrol. 319:96–108. doi:10.1016/j.jhy-drol.2005.07.004
Zhu, J., B.P. Mohanty, A.W. Warrick, and M.Th. van Genuchten. 2004. Corre-spondence and upscaling of hydraulic functions for heterogeneous soils. Vadose Zone J. 3:527–533.
Zhu, J., B.P. Mohanty, and N.N. Das. 2006. On the effective averaging schemes of hydraulic properties at the landscape scale. Vadose Zone J. 5:308–316. doi:10.2136/vzj2005.0035
