Estimation of Soil Moisture in Vegetation-Covered Floodplains with Sentinel-1 SAR Data Using Support Vector Regression

Soil moisture (SM) is a significant parameter influencing various environmental processes in hydrology, ecology, and climatology. SAR-derived remote sensing products can be valuable input features for estimating SM. In the past, the results often lacked a sufficient spatial resolution for a local application. With the new Sentinel-1 sensor it seems possible to derive more detailed SM-maps. Therefore, we utilized this sensor to test the applicability of a support vector regression (SVR) based method for SM retrieval for two different grassland-covered floodplains in north-east Germany. The model was then tested for its transferability. Moreover, it was operating exclusively with free and publicly available data. In situ data of volumetric SM were collected in 2015 for both study areas at the Elbe and Peene rivers. Remote sensing input data were VV and VH backscatter, and local incidence angle derived from Sentinel-1 images, as well as height, slope, and aspect derived from SRTM images. Additionally, the Landsat 8 NDVI product was included to compensate vegetation influences on SAR backscatter. Overall, the SVR is capable of estimating SM reasonably accurate (RMSE 9.7 and 13.8 Vol% for the individual study sites). Nevertheless, the performance highly depends on the in situ data—particularly, on the amount of samples (here 98 Peene and 71 Elbe) and the value range (10–100 Vol% Peene and 12–87 Vol% Elbe). Furthermore, the optimal input feature combination and the best performance differed between study sites. In summary, backscatter, elevation, and NDVI were the most important features for SM prediction. Models using only Radar-derived features where only 1.05 or 2.01 Vol% worse than models including optical data and are, therefore, able to estimate SM. Combining samples from both study sites slightly impaired the model. Due to varying site-conditions in terms of humidity and vegetation cover, the transferability of the SVR models is not possible for the studied sites.ZusammenfassungBodenfeuchtebestimmung in vegetationsbedeckten Flussauen mit Sentinel-1 SAR-Daten mittels Support-Vektor-Regression. Bodenfeuchte (BF) ist ein wichtiger Parameter, der verschiedene Umweltprozesse in der Hydrologie, Ökologie und Klimatologie beeinflusst. SAR-basierte Fernerkundungsprodukte können wertvolle Daten für die Abschätzung von BF sein. In der Vergangenheit fehlte oft eine ausreichende räumliche Auflösung für eine lokale Anwendung. Mit dem neuen Sentinel-1 Sensor ist es möglich, detailliertere BF-Karten abzuleiten. Daher haben wir Sentinel-1 verwendet, um die Anwendbarkeit der Support Vector Regression (SVR) Methode für die BF Abschätzung für zwei verschiedene Grasland-Überschwemmungsgebiete in Nordostdeutschland zu testen. Anschließend wurde das Modell auf seine Übertragbarkeit getestet. Darüber hinaus wurde ausschließlich mit freien und öffentlich zugänglichen Daten gearbeitet. In-situ-Daten des volumetrischen BF wurden im Jahr 2015 für die Untersuchungsgebiete an den Flüssen Elbe und Peene erhoben. Die Fernerkundungs-Eingabedaten waren VV- und VH-Rückstreuung und der lokale Einfallswinkel aus Sentinel-1-Bildern sowie Höhe, Neigung und Exposition ausSRTM-Bildern. Zusätzlich wurden Landsat 8 NDVI Produkte verwendet, um Vegetationseinflüsse auf die SAR-Rückstreuungen zu kompensieren. Insgesamt ist die SVR in der Lage, BF relativ genau zu schätzen (RMSE 9.7 und 13.8 Vol% für die einzelnen Untersuchungsgebiete). Dennoch hängt die Performance stark von den in-situ Daten ab – insbesondere von der Anzahl der Proben (hier 98 Peene bzw. 71 Elbe) und dem Wertebereich (10 - 100 Vol% Peene bzw. 12 - 87 Vol% Elbe). Darüber hinaus unterschieden sich die optimale Kombination von Eingabemerkmalen und die beste Leistung zwischen den einzelnen Untersuchungsgebieten. Zusammengefasst waren Rückstreuung, Höhe und NDVI die wichtigsten Merkmale für die BF-Vorhersage. Modelle, die ausschließlich von Radar abgeleitete Merkmale verwenden, sind nur um 1,05 bzw. 2,01 Vol% schlechter als Modelle mit optischen Daten und können daher BF bestimmen. Die Kombination von Daten aus beiden Untersuchungsgebieten führte zu einer leichten Beeinträchtigung des Modells. Die Übertragbarkeit der SVR-Modelle ist für die untersuchten Standorte aufgrund unterschiedlicher Standortbedingungen in Bezug auf Bodenfeuchtigkeit und Vegetationsdeckung nicht möglich.

[1]  Fernando J. Aguilar,et al.  Accuracy assessment of lidar‐derived digital elevation models , 2008 .

[2]  Brian W. Barrett,et al.  Soil Moisture Retrieval from Active Spaceborne Microwave Observations: An Evaluation of Current Techniques , 2009, Remote. Sens..

[3]  Yunjin Kim,et al.  The Soil Moisture Active/Passive (SMAP) radar , 2008, 2008 IEEE Radar Conference.

[4]  Yuei-An Liou,et al.  Retrieving soil moisture from simulated brightness temperatures by a neural network , 2001, IEEE Trans. Geosci. Remote. Sens..

[5]  Yisok Oh,et al.  Quantitative retrieval of soil moisture content and surface roughness from multipolarized radar observations of bare soil surfaces , 2004, IEEE Transactions on Geoscience and Remote Sensing.

[6]  K. Tockner,et al.  Riverine flood plains: present state and future trends , 2002, Environmental Conservation.

[7]  J. A. Jones Global Hydrology: Processes, Resources and Environmental Management , 1997 .

[8]  Katarzyna Dabrowska-Zielinska,et al.  Assessment of Carbon Flux and Soil Moisture in Wetlands Applying Sentinel-1 Data , 2016, Remote. Sens..

[9]  Lorenzo Bruzzone,et al.  Soil moisture monitoring in mountain areas by using high‐resolution SAR images: results from a feasibility study , 2014 .

[10]  Martha C. Anderson,et al.  Landsat-8: Science and Product Vision for Terrestrial Global Change Research , 2014 .

[11]  Lorenzo Bruzzone,et al.  Estimating Soil Moisture With the Support Vector Regression Technique , 2011, IEEE Geoscience and Remote Sensing Letters.

[12]  Marc Voltz,et al.  Spatio-temporal distribution of soil surface moisture in a heterogeneously farmed Mediterranean catchment , 2006 .

[13]  D. Lobell,et al.  Moisture effects on soil reflectance , 2002 .

[14]  Andrew W. Western,et al.  Terrain and the distribution of soil moisture , 2001 .

[15]  Lorenzo Bruzzone,et al.  Estimation of Soil Moisture in Mountain Areas Using SVR Technique Applied to Multiscale Active Radar Images at C-Band , 2015, IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing.

[16]  Vladimir Vapnik,et al.  The Nature of Statistical Learning , 1995 .

[17]  Kun Shan Chen,et al.  A simple model for retrieving bare soil moisture from radar-scattering coefficients , 1995 .

[18]  Luca Brocca,et al.  Soil moisture spatial variability in experimental areas of central Italy , 2007 .

[19]  George P. Petropoulos,et al.  Satellite remote sensing of surface soil moisture , 2013 .

[20]  Christian Heipke,et al.  Quality assessment of digital surface models derived from the Shuttle Radar Topography Mission (SRTM) , 2001, IGARSS 2001. Scanning the Present and Resolving the Future. Proceedings. IEEE 2001 International Geoscience and Remote Sensing Symposium (Cat. No.01CH37217).

[21]  M. S. Moran,et al.  Appropriate scale of soil moisture retrieval from high resolution radar imagery for bare and minimally vegetated soils , 2008 .

[22]  M. Schmitz Measuring soil moisture. , 2000 .

[23]  Matthias Drusch,et al.  Sentinel-2: ESA's Optical High-Resolution Mission for GMES Operational Services , 2012 .

[24]  Marco Borga,et al.  The influence of soil moisture on threshold runoff generation processes in an alpine headwater catchment , 2010 .

[25]  Claudia Notarnicola,et al.  Estimation of surface soil moisture in alpine areas based on medium spatial resolution SAR time-series and upscaled in-situ measurements , 2014, Remote Sensing.

[26]  N. Baghdadi,et al.  Retrieving surface roughness and soil moisture from synthetic aperture radar (SAR) data using neural networks , 2002 .

[27]  E. Rodríguez,et al.  A Global Assessment of the SRTM Performance , 2006 .

[28]  Pascale C. Dubois,et al.  Measuring soil moisture with imaging radars , 1995, IEEE Trans. Geosci. Remote. Sens..

[29]  G. Schumann,et al.  Comparison of remotely sensed water stages from LiDAR, topographic contours and SRTM , 2008 .

[30]  J.,et al.  Accuracy assessment of lidar‐derived digital elevation models , 2008 .

[31]  Claudia Notarnicola,et al.  Soil moisture retrieval from remotely sensed data: Neural network approach versus Bayesian method , 2008, IEEE Transactions on Geoscience and Remote Sensing.

[32]  F. Ulaby,et al.  Microwave Backscatter Dependence on Surface Roughness, Soil Moisture, and Soil Texture: Part II-Vegetation-Covered Soil , 1979, IEEE Transactions on Geoscience Electronics.

[33]  George P. Petropoulos,et al.  Turbulent Fluxes of Heat and Moisture at the Earth’s Land Surface: Importance, Controlling Parameters, and Conventional Measurement Techniques , 2013 .

[34]  W. Wagner,et al.  Soil moisture estimation through ASCAT and AMSR-E sensors: An intercomparison and validation study across Europe , 2011 .

[35]  Malcolm Davidson,et al.  GMES Sentinel-1 mission , 2012 .

[36]  A. Huete,et al.  A comparison of vegetation indices over a global set of TM images for EOS-MODIS , 1997 .

[37]  Richard K. Moore,et al.  Microwave Remote Sensing , 1999 .

[38]  Lorenzo Bruzzone,et al.  Multi-Objective Parameter Optimization in Support Vector Regression: General Formulation and Application to the Retrieval of Soil Moisture From Remote Sensing Data , 2012, IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing.

[39]  D. Vidal-Madjar,et al.  Theoretical study of the sensitivity of the microwave backscattering coefficient to the soil surface parameters , 1989 .

[40]  R.A.M. de Jeu,et al.  Using satellite based soil moisture to quantify the water driven variability in NDVI: A case study over mainland Australia , 2014 .

[41]  Gwo-Fong Lin,et al.  Effective forecasting of hourly typhoon rainfall using support vector machines , 2009 .

[42]  Heye Bogena,et al.  TERENO: German network of terrestrial environmental observatories , 2016 .

[43]  Rajat Bindlish,et al.  Parameterization of vegetation backscatter in radar-based, soil moisture estimation , 2001 .

[44]  G. Foody,et al.  Predictive relations of tropical forest biomass from Landsat TM data and their transferability between regions , 2003 .

[45]  M. S. Moran,et al.  RADAR REMOTE SENSING FOR ESTIMATION OF SURFACE SOIL MOISTURE AT THE WATERSHED SCALE 2 RADARSAT ERS SAR ERS ENVISAT ASAR JERS ALOS PALSAR , 2006 .

[46]  N. Baghdadi,et al.  Estimation of soil parameters over bare agriculture areas from C-band polarimetric SAR data using neural networks , 2012 .

[47]  Cuizhen Wang,et al.  Soil moisture extraction in sparse vegetated area using SAR and TM data , 2000, IGARSS 2000. IEEE 2000 International Geoscience and Remote Sensing Symposium. Taking the Pulse of the Planet: The Role of Remote Sensing in Managing the Environment. Proceedings (Cat. No.00CH37120).

[48]  G. Foody Geographical weighting as a further refinement to regression modelling: An example focused on the NDVI–rainfall relationship , 2003 .

[49]  N. Baghdadi,et al.  Soil parameters estimation over bare agriculture areas from C-band polarimetric SAR data using neural networks , 2012 .

[50]  I. Hajnsek,et al.  A tutorial on synthetic aperture radar , 2013, IEEE Geoscience and Remote Sensing Magazine.

[51]  Greg Hancock,et al.  An assessment of digital elevation models and their ability to capture geomorphic and hydrologic properties at the catchment scale* , 2010 .

[52]  Nazzareno Pierdicca,et al.  Soil moisture estimation over vegetated terrains using multitemporal remote sensing data. , 2010 .

[53]  Jiancheng Shi,et al.  The Soil Moisture Active Passive (SMAP) Mission , 2010, Proceedings of the IEEE.

[54]  W. Dorigo,et al.  Surface Soil Moisture Estimation : Significance, Controls and Conventional Measurement Techniques , 2013 .

[55]  Thomas J. Jackson,et al.  WindSat Global Soil Moisture Retrieval and Validation , 2010, IEEE Transactions on Geoscience and Remote Sensing.

[56]  Hugh G. Gauch,et al.  Winning the Accuracy Game , 2006 .

[57]  Paulin Coulibaly,et al.  Advances in soil moisture retrieval from synthetic aperture radar and hydrological applications , 2013 .

[58]  Peter Willett,et al.  Dissimilarity-Based Algorithms for Selecting Structurally Diverse Sets of Compounds , 1999, J. Comput. Biol..

[59]  Vladimir N. Vapnik,et al.  The Nature of Statistical Learning Theory , 2000, Statistics for Engineering and Information Science.

[60]  James E. Bartlett,et al.  Organizational research: Determining appropriate sample size in survey research , 2001 .

[61]  Mehrez Zribi,et al.  Operational performance of current synthetic aperture radar sensors in mapping soil surface characteristics in agricultural environments: application to hydrological and erosion modelling , 2008 .

[62]  K. Vorontsov Splitting and similarity phenomena in the sets of classifiers and their effect on the probability of overfitting , 2009, Pattern Recognition and Image Analysis.

[63]  C. Ottlé,et al.  Surface soil moisture estimation from the synergistic use of the (multi-incidence and multi-resolution) active microwave ERS Wind Scatterometer and SAR data , 2003 .

[64]  N. Baghdadi,et al.  Soil Moisture Profile Effect on Radar Signal Measurement , 2008, Sensors.

[65]  Roger D. De Roo,et al.  A semi-empirical backscattering model at L-band and C-band for a soybean canopy with soil moisture inversion , 2001, IEEE Trans. Geosci. Remote. Sens..

[66]  Yann Kerr,et al.  Soil moisture retrieval from space: the Soil Moisture and Ocean Salinity (SMOS) mission , 2001, IEEE Trans. Geosci. Remote. Sens..

[67]  S. Islam,et al.  Estimation of Soil Physical Properties Using Remote Sensing and Artificial Neural Network , 2000 .

[68]  M. S. Moran,et al.  Estimating soil moisture at the watershed scale with satellite-based radar and land surface models , 2004 .

[69]  Emanuele Santi,et al.  Soil moisture mapping using Sentinel-1 images: Algorithm and preliminary validation , 2013 .

[70]  Fabio Castelli,et al.  Mutual interaction of soil moisture state and atmospheric processes , 1996 .

[71]  Mac McKee,et al.  Multi-time scale stream flow predictions: The support vector machines approach , 2006 .

[72]  Divya Tomar,et al.  Prediction of Profitability of Industries using Weighted SVR , 2011 .

[73]  Andrea Cavallaro,et al.  Sensor Capability and Atmospheric Correction in Ocean Colour Remote Sensing , 2015, Remote. Sens..

[74]  Yann Kerr,et al.  Soil Moisture , 1922, Botanical Gazette.

[75]  M. S. Moran,et al.  Soil moisture evaluation using multi-temporal synthetic aperture radar (SAR) in semiarid rangeland , 2000 .

[76]  Pariva Dobriyal,et al.  A review of the methods available for estimating soil moisture and its implications for water resource management , 2012 .

[77]  Irena Hajnsek,et al.  Soil Moisture Estimation Under Low Vegetation Cover Using a Multi-Angular Polarimetric Decomposition , 2013, IEEE Transactions on Geoscience and Remote Sensing.

[78]  Paolo Ferrazzoli,et al.  A parametric study on passive and active microwave observations over a soybean crop , 1999, IEEE Trans. Geosci. Remote. Sens..

[79]  Sylvie Le Hégarat-Mascle,et al.  Soil moisture estimation from ERS/SAR data: toward an operational methodology , 2002, IEEE Trans. Geosci. Remote. Sens..

[80]  Emanuele Santi,et al.  A Comparison of Algorithms for Retrieving Soil Moisture from ENVISAT/ASAR Images , 2008, IEEE Transactions on Geoscience and Remote Sensing.

[81]  Richard Stone,et al.  A Simple Model , 1951 .

[82]  Bernhard Schölkopf,et al.  A tutorial on support vector regression , 2004, Stat. Comput..

[83]  R. Knight,et al.  Soil Moisture Measurement for Ecological and Hydrological Watershed‐Scale Observatories: A Review , 2008 .

[84]  C. C. Schmullius,et al.  Tabular summary of SIR-C/X-SAR results: synthetic aperture radar frequency and polarization requirements for applications in ecology and hydrology , 1997, IGARSS'97. 1997 IEEE International Geoscience and Remote Sensing Symposium Proceedings. Remote Sensing - A Scientific Vision for Sustainable Development.

[85]  Claudia Notarnicola,et al.  Use of radar and optical remotely sensed data for soil moisture retrieval over vegetated areas , 2006, IEEE Transactions on Geoscience and Remote Sensing.

[86]  M. McKee,et al.  SOIL MOISTURE PREDICTION USING SUPPORT VECTOR MACHINES 1 , 2006 .

[87]  Lorenzo Bruzzone,et al.  Estimation of Soil Moisture in an Alpine Catchment with RADARSAT2 Images , 2011 .

[88]  N. Baghdadi,et al.  Retrieving surface roughness and soil moisture from SAR data using neural networks. , 2002 .

[89]  A. Kalra,et al.  Estimating soil moisture using remote sensing data: A machine learning approach , 2010 .

[90]  Sar Jiankang Estimating soil moisture using the Danish polarimetric , 2004 .

[91]  刘焱序 Liu Yanxu,et al.  Advances in soil moisture retrieval from remote sensing , 2019 .