Sentinel-1 C-band Synthetic Aperture Radar Research Articles

Monitoring firn and wet snow on mountain glaciers: polarization and orbit effects

ABSTRACT Mountain glaciers are sensitive to climate variability and can be of great importance for downstream populations due to their hydrological significance. Synthetic Aperture Radar (SAR) images are often used to monitor the characteristics of glaciers based on the backscattering coefficient. However, the influence of satellite orbit and polarization when collecting images for wide regions has not been well considered. This study focuses on the extraction of wet snow in summer and firn in winter in West Kunlun Shan and the Tibet Interior Mountains by using Sentinel-1 C-band SAR data. The investigated regions have different climate patterns. We compare backscatter coefficient distributions for wet snow and firn, derived from maximum likelihood classification under various polarizations, alongside their respective ratios and show that polarization has a minor impact on the identification and monitoring of both wet snow and firn. However, a comparison of the wet snow ratios at different satellite orbits reveals notable differences between ascending and descending orbits in summer. We furthermore show, by analyzing weather stations on glaciers, that such effect can be related to the different acquisition time and different temperatures in the morning and afternoon and therefore to the orbit. In contrast, firn ratios across different orbits show less variation in winter, and the monitoring results consistently align with the patterns of ablation and accumulation typical under both climatic influences. These findings demonstrate glaciers’ sensitivity to temperature fluctuations and the radar wave’s responsiveness to surface characteristics. Consequently, when employing SAR for glacier monitoring, it is crucial to consider the influence of orbit and polarization, in combination with temperature variations, and whether the season is winter or summer.

Estimating vegetation water content from Sentinel-1 C-band SAR data over savanna and grassland ecosystems

Studying vegetation water content (VWC) dynamics is essential for understanding plant growth, water and carbon cycles, and ecosystem stability. However, acquiring field-based VWC estimates, consistently through space and time, is challenging due to time and resource constraints. This study investigates the potential of Sentinel-1 C-band Synthetic Aperture Radar (SAR) data for estimating VWC in natural ecosystems in central Brazil. We assessed (i) how well Sentinel-1 SAR data can capture variations in VWC over three different vegetation types (i.e. dry and waterlogged grasslands, and savannas) and (ii) how the studied vegetation types respond to seasonal dry periods in terms of water content. Field data from 82 plots, distributed across the three vegetation types and revisited in four different seasons, were used to calibrate and validate a model for VWC estimation. The calibrated model, with an R 2 of 0.52 and RMSE of 0.465 kg m−2, was then applied to Sentinel-1 SAR backscatter data to generate monthly VWC maps for grassland and savanna ecosystems at 30 m spatial resolution between April 2015 and September 2023. These maps, combined with rainfall and potential evapotranspiration data, provided insights into how the studied vegetation types respond to water shortage during the dry season at the community scale. More specifically, savannas showed to be better able to retain higher levels of water content during the dry season, probably due to a higher water holding capacity of the woody component together with its deep-root system ability to access deeper groundwater. This research demonstrates the potential of Sentinel-1 SAR data for monitoring VWC in natural ecosystems, allowing for future studies to assess ecosystems’ response to drought events and changes in their functioning, ultimately supporting land management decisions.

A Study on C-band Synthetic Aperture Radar Soil Moisture Estimation Based on Machine Learning Using Soil Physics, Topography, and Hydrological Information

In this study, we applied machine learning to estimate soil moisture levels in South Korea by harnessing data from the Sentinel-1 C-band synthetic aperture radar (SAR). Our approach incorporated not only the relationship between backscattering coefficients and soil moisture but also diverse physical characteristics. This encompassed topographic information, soil physics data, and antecedent precipitation which is a hydrological factor influencing the initial condition of soil moisture. We applied a variety of machine-learning techniques and conducted a comprehensive analysis to compare the performance of each model.

Total-variation regularized U-Net for wildfire burned area mapping based on Sentinel-1 C-Band SAR backscattering data

Previous studies have shown that Synthetic Aperture Radar (SAR) is able to detect burned areas, serving as a key data source for monitoring active wildfires in situations where optical sensors are hindered by dense smoke or cloud cover. Radar remote sensing provides rich and useful data on historical burned areas, which are critical for large-scale wildfire burned area mapping. This study aims to unveil the potential inherent in Sentinel-1 C-Band SAR data and to investigate the impact of reference masks on wildfire burned area mapping. SAR images frequently exhibit disruptive speckle noise, which manifests as isolated pixels or small regions, thereby hindering the seamless identification of burned areas through SAR data. To mitigate noise and enhance the connectivity of SAR-based burned area delineation, we propose a novel approach: a Total-Variation (TV) regularized U-Net model, tailored to learn from noisy pseudo masks derived from Sentinel-1 C-Band SAR backscattering data. To validate its efficacy, we assembled a dataset encompassing 16 geographically diverse wildfire events, incorporating SAR-based kMap, a SAR-based binary label (designated as SARREF), and an optical-based binary label (referred to as OptREF). The kMap, serving as a relative indicator of SAR backscattering change, is defined as the divergence between post-fire SAR backscattering and the temporal average, normalized by the temporal standard deviation of pre-fire SAR time series. Utilizing SAR-based kMap as input and SARREF or OptREF as labels, we systematically explored the impact of TV regularization on SAR-based burned area mapping. Our experimental findings highlight several key insights: 1) Vanilla U-Net considerably elevates the F1 score from 0.61 to 0.66 when compared to threshold-based approaches; 2) The TV-regularized U-Net exhibits a notable capacity for enhancing both accuracy and the connectivity of SAR-based burned areas. When trained with noisy SARREF, it achieves a substantial F1 score of 0.69 and an IoU score of 0.53. 3) In scenarios involving OptREF, the vanilla U-Net attains peak performance with the highest F1 (0.75) and IoU (0.61) scores, with the expected outcome of TV regularization yielding no additional enhancement.

Soil moisture retrieval from Sentinel-1 using a first-order radiative transfer model—A case-study over the Po-Valley

Soil moisture is an important variable controlling many land surface processes and is used to quantify precipitation, drought, flooding, irrigation and other factors that influence decision making and risk-assessment. This paper presents the retrieval of high resolution (∼1 km) soil moisture data from Sentinel-1 C-band Synthetic Aperture Radar (SAR) backscatter measurements using a new bistatic radiative transfer modeling framework (RT1) previously only tested for scatterometer data. The model is applied over a diverse set of landcover types across the entire Po-Valley in Italy over a 4-year time-period from 2016 to 2019. The performance of the soil moisture retrievals is analyzed with respect to the ERA5-Land reanalysis dataset. The model parameterisation and retrieval method are chosen such as to constitute a trade-off between a physically plausible and a computationally feasible modeling approach. The results demonstrate the potential of RT1 for the retrieval of high-resolution soil moisture data from SAR time series.

A Globally Applicable Method for NDVI Estimation from Sentinel-1 SAR Backscatter Using a Deep Neural Network and the SEN12TP Dataset

Vegetation monitoring is important for many applications, e.g., agriculture, food security, or forestry. Optical data from space-borne sensors and spectral indices derived from their data like the normalised difference vegetation index (NDVI) are frequently used in this context because of their simple derivation and interpretation. However, optical sensors have one major drawback: cloud coverage hinders data acquisition, which is especially troublesome for moderate and tropical regions. One solution to this problem is the use of cloud-penetrating synthetic aperture radar (SAR) sensors. Yet, with very different image characteristics of optical and SAR data, an optical sensor cannot be easily replaced by SAR sensors. This paper presents a globally applicable model for the estimation of NDVI values from Sentinel-1 C-band SAR backscatter data. First, the newly created dataset SEN12TP consisting of Sentinel-1 and -2 images is introduced. Its main features are the sophisticated global sampling strategy and that the images of the two sensors are time-paired. Using this dataset, a deep learning model is trained to regress SAR backscatter data to NDVI values. The benefit of auxiliary input information, e.g., digital elevation models, or land-cover maps is evaluated experimentally. After selection of the best model configuration, another experimental evaluation on a carefully selected hold-out test set confirms that high performance, low error, and good level of spatial detail are achieved. Finally, the potential of our approach to create dense NDVI time series of frequently clouded areas is shown. One limit of our approach is the neglect of the temporal characteristics of the SAR and NDVI data, since only data from a single date are used for prediction.

Mapping Surface Organic Soil Properties in Arctic Tundra Using C-Band SAR Data

Surface soil organic carbon (SOC) content is among the first-order controls on the rate and extent of Arctic permafrost thaw. There is a large discrepancy in current SOC estimates in Arctic tundra, where sparse measurements are unable to capture SOC complexity over the vast tundra region. Synthetic aperture radar (SAR) data are sensitive to surface vegetation, roughness, and moisture conditions, and may provide useful information on surface SOC properties. Here, we investigated the potential of multitemporal Sentinel-1 C-band SAR data for regional SOC mapping in the Arctic tundra through principal component analysis (PCA). Multiple <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">in situ</i> SOC datasets in the Alaska North Slope were assembled to generate a consistent surface (0–10 cm) SOC and bulk density dataset ( <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">n</i> = 97). The radar VV backscatter shows a strong correlation with surface SOC, but the correlation varies greatly with surface snow, moisture, and freeze/thaw conditions. However, the first principal component (PC1) of radar backscatter time series from different years shows spatial consistency representing dominant and persistent surface backscatter behavior. The PC1 also shows a strong linear correlation with surface SOC concentration ( <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">R</i> = 0.65, <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">p</i> <0.01), and an exponential relationship with bulk density ( <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">R</i> = −0.65, <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">p</i> <0.01). The resulting predicted SOC maps show much lower soil bulk density and higher SOC concentration in the southern shrub tundra area than in the northern coastal region, consistent with <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">in situ</i> data. Our analysis shows that it is possible to separate the effects of different factors on the radar backscatter response using PCA and multitemporal SAR data, which may lead to more effective satellite-based methods for Arctic SOC mapping.

Multi-Source Mapping of Forest Susceptibility to Spruce Budworm Defoliation Based on Stand Age and Composition across a Complex Landscape in Maine, USA

Spruce budworm (Choristoneura fumiferana; SBW) outbreaks in the northeastern USA and Canada are recurring phenomena leading to large-scale mortality of spruce (Picea sp.) and balsam fir (Abies balsamea (L.) Mill.) forests as susceptibility to SBW is primarily determined by the availability of host species and their maturity. Our study examined several satellite remote sensing (Sentinel-1 C-band synthetic aperture radar (SAR), PALSAR L-band SAR, and Sentinel-2 multispectral) and site variables over space and time to develop a method to produce large-scale SBW stand impact types and susceptibility maps in Maine, USA. We used two machine-learning algorithms (Random Forest, RF; Multi-Layer Perceptron, MLP) to map SBW host species where RF produced better results than MLP. Our best model with site (elevation and aspect) and Sentinel-2 data attained an overall accuracy (OA) of 83.4%. However, the addition of SAR variables did not improve the model further. Combining host species data with age data retrieved from Land Change Monitoring, Assessment, and Projection (LCMAP) products, we demonstrated that SBW susceptibility map (based on stand impact types) could be produced with an OA of 88.3%. The fine spatial resolution (20 m) maps derived from our study provide reliable products for landscape-level SBW interventions in the region.

Utilising Sentinel-1’s orbital stability for efficient pre-processing of sigma nought backscatter

For already more than seven years, the Sentinel-1 C-band Synthetic Aperture Radar (SAR) mission has been providing indispensable information for monitoring bio-geophysical parameters at fine temporal and spatial scales. As many applications require backscatter datacubes as input, enormous amounts of data have to be radiometrically and geometrically corrected to be in a common, Earth-fixed reference system. Pre-processing workflows accomplishing this task have already been established and are implemented in several software suites. However, typically, these workflows are computationally expensive which may lead to prohibitively large costs when generating multi-year Sentinel-1 datacubes for whole continents or the world. In this paper, we discuss existing approaches for generating sigma nought and projected local incidence angle (PLIA) data and present simplifications of the overall workflow relying on the unprecedented orbital stability of Sentinel-1. Propagating orbital deviations through the complete Sentinel-1 pre-processing pipeline helped us to simulate and identify PLIA as a static layer per relative orbit. The outcome of these simulations also provided the necessary information to replace iterative root-finding algorithms for determining the time of closest approach (TCA), i.o.w. the azimuth index, with a linear one — at no expense of radiometric accuracy. All experiments were performed using an in-house developed toolbox named wizsard, which made it possible to speed up Sentinel-1 pre-processing by approximately 4–5 times with respect to the Sentinel Application Platform (SNAP). This could pave the way for producing quality-curated, large-scale backscatter datacubes at continental and global scales in acceptable time frames.

Interpreting Sentinel-1 SAR Backscatter Signals of Snowpack Surface Melt/Freeze, Warming, and Ripening, through Field Measurements and Physically-Based SnowModel

The transition of a cold winter snowpack to one that is ripe and contributing to runoff is crucial to gauge for water resource management, but is highly variable in space and time. Snow surface melt/freeze cycles, associated with diurnal fluctuations in radiative inputs, are hallmarks of this transition. C-band synthetic aperture radar (SAR) reliably detects meltwater in the snowpack. Sentinel-1 (S1) C-band SAR offers consistent acquisition patterns that allow for diurnal investigations of melting snow. We used over 50 snow pit observations from 2020 in Grand Mesa, Colorado, USA, to track temperature and wetness in the snowpack as a function of depth and time during snowpack phases of warming, ripening, and runoff. We also ran the physically-based SnowModel, which provided a spatially and temporally continuous independent indication of snowpack conditions. Snowpack phases were identified and corroborated by comparing field measurements with SnowModel outputs. Knowledge of snowpack warming, ripening, and runoff phases was used to interpret diurnal changes in S1 backscatter values. Both field measurements and SnowModel simulations suggested that S1 SAR was not sensitive to the initial snowpack warming phase on Grand Mesa. In the ripening and runoff phases, the diurnal cycle in S1 SAR co-polarized backscatter was affected by both surface melt/freeze as well as the conditions of the snowpack underneath (ripening or ripe). The ripening phase was associated with significant increases in morning backscatter values, likely due to volume scattering from surface melt/freeze crusts, as well as significant decreases in evening backscatter values associated with snowmelt. During the runoff phase, both morning and evening backscatter decreased compared to reference values. These unique S1 diurnal signatures, and their interpretations using field measurements and SnowModel outputs, highlight the capacities and limitations of S1 SAR to understand snow surface states and bulk phases, which may offer runoff forecasting or energy balance model validation or parameterization, especially useful in remote or sparsely-gauged alpine basins.

Monitoring peatland water table depth with optical and radar satellite imagery

Peatland water table depth (WTD) and wetness have widely been monitored with optical and synthetic aperture radar (SAR) remote sensing but there is a lack of studies that have used multi-sensor data, i.e., combination of optical and SAR data. We assessed how well WTD can be monitored with remote sensing data, whether multi-sensor approach boosts explanatory capacity and whether there are differences in regression performance between data and peatland types. Our data consisted of continuous multiannual WTD data from altogether 50 restored and undrained Finnish peatlands, and optical (Landsat 5–8, Sentinel-2) and Sentinel-1 C-band SAR data processed in Google Earth Engine. We calculated random forest regressions with dependent variable being WTD and independent variables consisting of 21 optical and 10 SAR metrics. The average regression performance was moderate in multi-sensor models (R 2 43.1%, nRMSE 19.8%), almost as high in optical models (R 2 42.4%, nRMSE 19.9%) but considerably lower in C-band SAR models (R 2 21.8%, nRMSE 23.4%) trained separately for each site. When the models included data from several sites but were trained separately for six habitat type and management option combinations, the average R 2 was 40.6% for the multi-sensor models, 36.6% for optical models and 33.7% for C-band SAR models. There was considerable site-specific variation in the model performance (R 2 −3.3–88.8% in the multi-sensor models ran separately for each site) and whether multi-sensor, optical or C-band SAR model performed best. The average regression performance was higher for undrained than for restored peatlands, and higher for open and sparsely treed than for densely treed peatlands. The most important variables included SWIR-based optical metrics and VV SAR backscatter. Our results suggest that optical data works usually better than does C-band SAR data in peatland WTD monitoring and multi-sensor approach increases explanatory capacity moderately little.

Coastal land-use mapping along Johore Straits using Sentinel 1-SAR data with maximizing parameterization of machine learning classification

In tropical regions, cloud cover accounts for a major obstacle to detect the coastal land-use while adopting remote sensing technology. The advent of the latest Sentinel-1 C-band synthetic aperture radar (SAR) satellite provides advantages to collect data in all-weathered conditions. In this study, Sentinel-1 images are processed using Lee filter and GLCM-Mean texture analysis in order to enhance the classification results. Several sets of parameters have been tested and these resulted in the optimum overall accuracy by Neural Network with 79.00% in 2015 and 68.29% in 2019. In contrast, Support Vector Machine classifiers obtained overall accuracies of 77.44% and 71.26% in 2015 and 2019 respectively. The results were accessed and it is found that Support Vector Machine outperforms the Neural Network classifier in discriminating data with high heterogeneity properties. Besides, Support Vector Machine has more consistent results in parameter testing compared to Neural Network.

Estimation of Aboveground Carbon Density of Forests Using Deep Learning and Multisource Remote Sensing

Forests are crucial in carbon sequestration and oxygen release. An accurate assessment of forest carbon storage is meaningful for Chinese cities to achieve carbon peak and carbon neutrality. For an accurate estimation of regional-scale forest aboveground carbon density, this study applied a Sentinel-2 multispectral instrument (MSI), Advanced Land Observing Satellite 2 (ALOS-2) L-band, and Sentinel-1 C-band synthetic aperture radar (SAR) to estimate and map the forest carbon density. Considering the forest field-inventory data of eastern China from 2018 as an experimental sample, we explored the potential of the deep-learning algorithms convolutional neural network (CNN) and Keras. The results showed that vegetation indices from Sentinel-2, backscatter and texture characters from ALOS-2, and coherence from Sentinel-1 were principal contributors to the forest carbon-density estimation. Furthermore, the CNN model was found to perform better than traditional models. Results of forest carbon-density estimation validated the improvements effectively by combining the optical and radar data. Compared with traditional regression methods, deep learning has a higher potential for accurately estimating forest carbon density using multisource remote-sensing data.

ALOS-2 L-band SAR backscatter data improves the estimation and temporal transferability of wildfire effects on soil properties under different post-fire vegetation responses

Remote sensing techniques are of particular interest for monitoring wildfire effects on soil properties, which may be highly context-dependent in large and heterogeneous burned landscapes. Despite the physical sense of synthetic aperture radar (SAR) backscatter data for characterizing soil spatial variability in burned areas, this approach remains completely unexplored. This study aimed to evaluate the performance of SAR backscatter data in C-band (Sentinel-1) and L-band (ALOS-2) for monitoring fire effects on soil organic carbon and nutrients (total nitrogen and available phosphorous) at short term in a heterogeneous Mediterranean landscape mosaic made of shrublands and forests that was affected by a large wildfire. The ability of SAR backscatter coefficients and several band transformations of both sensors for retrieving soil properties measured in the field in immediate post-fire situation (one month after fire) was tested through a model averaging approach. The temporal transferability of SAR-based models from one month to one year after wildfire was also evaluated, which allowed to assess short-term changes in soil properties at large scale as a function of pre-fire plant community type. The retrieval of soil properties in immediate post-fire conditions featured a higher overall fit and predictive capacity from ALOS-2 L-band SAR backscatter data than from Sentinel-1 C-band SAR data, with the absence of noticeable under and overestimation effects. The transferability of the ALOS-2 based model to one year after wildfire exhibited similar performance to that of the model calibration scenario (immediate post-fire conditions). Soil organic carbon and available phosphorous content was significantly higher one year after wildfire than immediately after the fire disturbance. Conversely, the short-term change in soil total nitrogen was ecosystem-dependent. Our results support the applicability of L-band SAR backscatter data for monitoring short-term variability of fire effects on soil properties, reducing data gathering costs within large and heterogeneous burned landscapes.

Decadal changes of Campbell Glacier Tongue in East Antarctica from 2010 to 2020 and implications of ice pinning conditions analyzed by optical and SAR datasets

ABSTRACT Changes in ice shelf dynamics significantly impact the discharge rate of grounded ice, sea ice formation, and marine environments. In particular, changes of a sub-shelf pinning point induce complex dynamics of an ice shelf. In this study, we investigated decadal changes (2010–2020) in the area, ice velocity, and grounding line position of Campbell Glacier Tongue (CGT) in East Antarctica, which has an ice pinning point at the southwestern end, and analyzed the effect of the pinning conditions on the dynamics of CGT. Panchromatic band images of Landsat-7 Enhanced Thematic Mapper Plus and Landsat-8 Operational Land Imager, COSMO-SkyMed X-band synthetic aperture radar (SAR), and Sentinel-1 C-band SAR datasets were employed. The surface area of CGT digitized from the optical and SAR imagery decreased by 8% in the last decade, during which four ice calving events that caused icebergs with an area range of 3.8–5.4 km2. The largest calving at the ice pinning point occurred in March 2014, which created an iceberg with 4 km2. The ice velocity of CGT was measured via the normalized cross-correlation-based image matching of the Landsat panchromatic images. The ice velocity along the flow direction and its anomaly of CGT hardly changed in the hinge zone from 2011 to 2020. The ice velocity increased by ~7% and the flow direction steered abruptly ~7° to the east in the freely floating zone after the 2014 calving at the ice pinning point. Considering the retreat of the local grounding line at the pinning point analyzed by COSMO-SkyMed double-differential interferometric SAR, the advancing glacier tongue came in contact with the sloping eastern flank of the locally elevated seabed at the pinning point after the calving, and the ice-bed contact area decreased, causing a sudden acceleration and eastward rotation of ice flow. It is expected that continuous calving at the ice pinning point of CGT may lead to an abrupt changes in glacier tongue dynamics again, and continuous monitoring is necessary.

Global seasonal Sentinel-1 interferometric coherence and backscatter data set

This data set is the first-of-its-kind spatial representation of multi-seasonal, global C-band Synthetic Aperture Radar (SAR) interferometric repeat-pass coherence and backscatter signatures. Coverage comprises land masses and ice sheets from 82° Northern to 79° Southern latitudes. The data set is derived from multi-temporal repeat-pass interferometric processing of about 205,000 Sentinel-1 C-band SAR images acquired in Interferometric Wide-Swath Mode from 1-Dec-2019 to 30-Nov-2020. The data set encompasses three sets of seasonal (December-February, March-May, June-August, September-November) metrics produced with a pixel spacing of three arcseconds: 1) Median 6-, 12-, 18-, 24-, 36-, and 48-days repeat-pass coherence at VV or HH polarizations, 2) Mean radiometrically terrain corrected backscatter (γ0) at VV and VH, or HH and HV polarizations, and 3) Estimated parameters of an exponential coherence decay model. The data set has been produced to obtain global, spatially detailed information on how decorrelation affects interferometric measurements of surface displacement and is rich in spatial and temporal information for a variety of mapping applications.

Fine-resolution mapping of the circumpolar Arctic Man-made impervious areas (CAMI) using sentinels, OpenStreetMap and ArcticDEM

ABSTRACT Man-made impervious areas map is of great demand in environmental and urbanization studies since impervious areas are considered as a key indication of urbanization, especially for circumpolar Arctic. However, to date, finer resolution and spatially continuous impervious areas information remains scarce in the Arctic. In this study, we developed an accurate and complete circumpolar Arctic Man-made impervious areas (CAMI) map at a resolution of 10 m by combining Sentinel-1 C-band Synthetic Aperture Radar, Sentinel-2 multispectral images, OpenStreetMap, and ArcticDEM via Google Earth Engine platform. A random forest classifier model was trained and used to generate corresponding impervious areas map for the year 2020. The evaluation results suggested that CAMI was the most accurate with an overall accuracy of 86.36% and kappa coefficient of 70.73% as against the three existing impervious areas products. Based on the generated map and OpenStreetMap, we estimated that total impervious areas area in the Arctic has achieved 807.80 , of which roads, industrial and resident land were three major land use types, accounting for 54.08%, 17.85% and 10.34%, respectively. The CAMI map will support for new application and provide advanced insight into the infrastructure vulnerability evaluation and environmental sustainability in the Arctic.

Integrating SAR and Optical Remote Sensing for Conservation-Targeted Wetlands Mapping

The Prairie Pothole Region (PPR) contains numerous depressional wetlands known as potholes that provide habitats for waterfowl and other wetland-dependent species. Mapping these wetlands is essential for identifying viable waterfowl habitat and conservation planning scenarios, yet it is a challenging task due to the small size of the potholes, and the presence of emergent vegetation. This study develops an open-source process within the Google Earth Engine platform for mapping the spatial distribution of wetlands through the integration of Sentinel-1 C-band SAR (synthetic aperture radar) data with high-resolution (10-m) Sentinel-2 bands. We used two machine-learning algorithms (random forest (RF) and support vector machine (SVM)) to identify wetlands across the study area through supervised classification of the multisensor composite. We trained the algorithms with ground truth data provided through field studies and aerial photography. The accuracy was assessed by comparing the predicted and actual wetland and non-wetland classes using statistical coefficients (overall accuracy, Kappa, sensitivity, and specificity). For this purpose, we used four different out-of-sample test subsets, including the same year, next year, small vegetated, and small non-vegetated test sets to evaluate the methods on different spatial and temporal scales. The results were also compared to Landsat-derived JRC surface water products, and the Sentinel-2-derived normalized difference water index (NDWI). The wetlands derived from the RF model (overall accuracy 0.76 to 0.95) yielded favorable results, and outperformed the SVM, NDWI, and JRC products in all four testing subsets. To provide a further characterization of the potholes, the water bodies were stratified based on the presence of emergent vegetation using Sentinel-2-derived NDVI, and, after excluding permanent water bodies, using the JRC surface water product. The algorithm presented in the study is scalable and can be adopted for identifying wetlands in other regions of the world.

Semi-automated tracking of iceberg B43 using Sentinel-1 SAR images via Google Earth Engine

Abstract. Sentinel-1 C-band synthetic aperture radar (SAR) images can be used to observe the drift of icebergs over the Southern Ocean with around 1–3 d of temporal resolution and 10–40 m of spatial resolution. The Google Earth Engine (GEE) cloud-based platform allows processing of a large quantity of Sentinel-1 images, saving time and computational resources. In this study, we process Sentinel-1 data via GEE to detect and track the drift of iceberg B43 during its lifespan of 3 years (2017–2020) in the Southern Ocean. First, to detect all candidate icebergs in Sentinel-1 images, we employ an object-based image segmentation (simple non-iterative clustering – SNIC) and a traditional backscatter threshold method. Next, we automatically choose and trace the location of the target iceberg by comparing the centroid distance histograms (CDHs) of all detected icebergs in subsequent days with the CDH of the reference target iceberg. Using this approach, we successfully track iceberg B43 from the Amundsen Sea to the Ross Sea and examine its changes in area, speed, and direction. Three periods with sudden losses of area (i.e., split-offs) coincide with periods of low sea ice concentration, warm air temperature, and high waves. This implies that these variables may be related to mechanisms causing the split-off of the iceberg. Since the iceberg is generally surrounded by compacted sea ice, its drift correlates in part with sea ice motion and wind velocity. Given that the bulk of the iceberg is under water (∼30–60 m freeboard and ∼150–400 m thickness), its motion is predominantly driven by the westward-flowing Antarctic Coastal Current, which dominates the circulation of the region. Considering the complexity of modeling icebergs, there is a demand for a large iceberg database to better understand the behavior of icebergs and their interactions with surrounding environments. The semi-automated iceberg tracking based on the storage capacity and computing power of GEE can be used for this purpose.

A synergic method of Sentinel-1 and Sentinel-2 images for retrieving soil moisture content in agricultural regions

Remote sensing techniques, which provide timely and accurate soil moisture information to improve food production, are becoming increasingly popular among agricultural production in China and many other regions. It also shows its importance in monitoring agricultural basins where the meteorological and topographic conditions vary greatly. In this case, the reliability of soil moisture content (SMC) monitoring for theses basins could be promised. The main objective of this paper is to develop a synergic method of radar and optical data, as well as an on-site sampling scheme, to accurately estimate SMC in the basin. The reason is that in the winter wheat production area of Dawen River basin in China, there are few conventional SMC testing equipment. Considering the characteristics of different crops and the conditions of crops in different growth periods, the VV polarization mode of Sentinel-1 C-band Synthetic Aperture Radar (SAR) images was used to retrieve SMC. Combined with the Normalized difference vegetation index (NDVI) extracted from Sentinel-2 MSI optical images and the optimized Water Cloud Model (WCM), the surface backscattering coefficient was obtained by eliminating vegetation moisture content (VMC), and the SMC model was established. The results showed that from March to May in 2020, the coefficient of determination (R2) between the estimated VMC and the field observation was 0.655, and the Root Mean Square Error (RMSE) between the two was 0.334. In the inversion algorithm model, the R2 between the retrieved SMC and the field measurement was 0.506, and the RMSE was 3.725.