Spatiotemporal Visualisation of PS InSAR Generated Space–Time Series Describing Large Areal Land Deformations Using Diagram Map with Spiral Graph

Abstract

The space–time series carry information on temporal and spatial patterns in observed phenomena. The reported research integrates computational, visual and cartographic methods to support visual analysis of space–time series describing terrain surface movement. The proposed methodology for space–time series visualisation can support their analysts in investigating space–time patterns using transformation, clustering, filtration and visualisation. The presented approach involves spiral graphs for representation time dimension and cartographic visualisation through proportional point symbol map for representation of spatial dimension. The result is an intuitive visualisation of space–time series, conveying the sought-after spatio-temporal information. For practical tests, we used space–time series obtained by permanent scatterers interferometry (PS InSAR) to monitor the Earth’s surface movement above the underground gas storage (UGS) Tvrdonice, the Czech Republic. An UGS is characterised by periodic injection and withdrawal of natural gas, which induces periodic movement of the terrain above it. We have verified that our visualisation method provides the required pattern information and is easy to use.

Monitoring Non-Linear Ground Motion above Underground Gas Storage Using GNSS and PSInSAR Based on Sentinel-1 Data

“…For example, from 2011 to 2022, the minimum fullness of the virtual UGS ranged from 26 March to 1 May, and the maximum fullness ranged from 6 October to 9 December. Therefore, the working cycle period may not be exactly 365 days; in our case, it fluctuated from 342 to 388 days [26].…”

“…We compiled a theoretical filling curve for the UGS, verified by the course of fullness cycles of the so-called virtual UGS defined for the Czech Republic [26]. The whole cycle usually starts at the turn of April to May, initiating with gas injection into the UGS, and the injection is usually over at the turn of October to November, when withdrawal starts.…”

“…The irregular component (noise) can be removed, for example, by a moving average (provided that unwrapping errors [36] do not occur in the data or are removed). This step is especially very suitable for visualization needs [26]. The data relating to each PS can be viewed as a time series of measurements.…”

Research on land surface change and 3D visualization based on WebGis and computer graphics

“…Both methods mainly use static displays, such as line graphs, heat maps, and scatter plots, making it difficult to show continuous surface three-dimensional changes over time and lacking continuous three-dimensional animation evolution processes. Using temporal and spatial information to restore the monitoring process is more intuitive [10].…”

PSI Spatially Constrained Clustering: The Sibari and Metaponto Coastal Plains

“…A common choice is to use clustering algorithms [25][26][27][28], whose underlying assumption (widely accepted by the scientific community) is that the more PS show a coherent displacement, the more reliable the observed effect is. A popular choice for the remote sensing community is the DBSCAN algorithm (Density-Based Spatial Clustering of Applications with Noise) [29][30][31], especially for its efficiency in retrieving clusters with arbitrary shape and its computational efficiency. Nevertheless, as DBSCAN operates in the feature space, it can neglect important constraints provided by spatial proximity, which can, in principle, improve clustering results.…”