Applying Machine Learning and Time-Series Analysis on Sentinel-1A SAR/InSAR for Characterizing Arctic Tundra Hydro-Ecological Conditions

Abstract

Synthetic aperture radar (SAR) is a widely used tool for Earth observation activities. It is particularly effective during times of persistent cloud cover, low light conditions, or where in situ measurements are challenging. The intensity measured by a polarimetric SAR has proven effective for characterizing Arctic tundra landscapes due to the unique backscattering signatures associated with different cover types. However, recently, there has been increased interest in exploiting novel interferometric SAR (InSAR) techniques that rely on both the amplitude and absolute phase of a pair of acquisitions to produce coherence measurements, although the simultaneous use of both intensity and interferometric coherence in Arctic tundra image classification has not been widely tested. In this study, a time series of dual-polarimetric (VV, VH) Sentinel-1 SAR/InSAR data collected over one growing season, in addition to a digital elevation model (DEM), was used to characterize an Arctic tundra study site spanning a hydrologically dynamic coastal delta, open tundra, and high topographic relief from mountainous terrain. SAR intensity and coherence patterns based on repeat-pass interferometry were analyzed in terms of ecological structure (i.e., graminoid, or woody) and hydrology (i.e., wet, or dry) using machine learning methods. Six hydro-ecological cover types were delineated using time-series statistical descriptors (i.e., mean, standard deviation, etc.) as model inputs. Model evaluations indicated SAR intensity to have better predictive power than coherence, especially for wet landcover classes due to temporal decorrelation. However, accuracies improved when both intensity and coherence were used, highlighting the complementarity of these two measures. Combining time-series SAR/InSAR data with terrain derivatives resulted in the highest per-class F1 score values, ranging from 0.682 to 0.955. The developed methodology is independent of atmospheric conditions (i.e., cloud cover or sunlight) as it does not rely on optical information, and thus can be regularly updated over forthcoming seasons or annually to support ecosystem monitoring.

Machine Learning-Based Active Layer Thickness Estimation Over Permafrost Landscapes by Upscaling Airborne Remote Sensing Measurements with Cloud-Computing Geotechnologies

“…Thus, the use of GEE removed the need for powerful and expensive local computing power (i.e., high-performance computing; HCP), which is otherwise necessary for scaling high-resolution predictions over large areas with RS big data [35]. This study now joins a dense collection of literature demonstrating the large-scale modeling capabilities of GEE for environmental applications and information generation over permafrost rich landscapes [33,[68][69][70][71]. As such, the use of GEE, along with other cloud-based geospatial platforms (e.g., Microsoft Azure, Amazon Web Services, etc.…”

“…Interferometric SAR (InSAR) techniques in particular have proven effective for ALT estimation and monitoring [32]. The InSAR technique relies on both the phase and amplitude information contained in the SAR signal, acquired from a pair (or more) of multi-temporal SAR images [33]. By comparing the timing of the radar return between each image, the distance between the sensor and ground can be quantified.…”

“…EO image sources included multi-frequency SAR, from the C-band Sentinel-1 and L-band ALOS PALSAR satellites, and electro-optical imagery from Landsat-8. These EO datasets were selected since SAR can penetrate vegetation and soil surfaces and interact with sub-surface hydrological conditions [33], and Landsat-8 supplies both reflectance and thermal energy measurements which can help differentiate frozen from unfrozen ground [50]. EO image preprocessing followed that of Merchant et al [39], and included fundamental steps such as speckle filtering and orthorectification of SAR, and cloud-masking of Landsat-8 imagery.…”

Recognizing the Shape and Size of Tundra Lakes in Synthetic Aperture Radar (SAR) Images Using Deep Learning Segmentation

“…Although the manual analysis of SAR data is sensor-independent, and relatively robust against misclassification [35,40,41], it is time-consuming; therefore, automated methods are preferable. Some studies have proposed using machine learning for characterizing Arctic tundra landscapes, including lakes [30,42]. To the best of our knowledge, stateof-the-art deep learning methods have not been previously applied to the problems of automatic tundra lakes recognition from SAR, and, thus, one of the primary aims of this study was to investigate the potential of deep learning for accurate tundra lake shape and size recognition.…”

Quantitative estimation of sentinel-1A interferometric decorrelation using vegetation index

“…However, the short-wavelength characteristics bring a range of limitations to its application, especially in vegetated areas, because C-band signals normally do not penetrate the surface or top layer of the forest (i.e., leaves and twigs) (Anh and Hang, 2019). Phenological changes to vegetation-or even leaf and branch movement-can lead to decorrelation (Merchant et al, 2022), which is a primary error source that limits the capability of InSAR for deformation mapping in areas with low coherence (Liang et al, 2021b) and what's more, its data handing, processing, and interpretation are barriers preventing a rapid uptake of SAR data by application specialists and non-expert domain users in the field of agricultural monitoring (Kumar et al, 2022).…”