SAR for Urban Management: Coherence Estimation and Disaster Detection

 SAR for Urban Management: Coherence Estimation and Disaster Detection

Urban planning and planners use maps or technology tools to study the impact of land use development on urban areas and make the right decisions for the future growth of the cities. The Urban Footprint technique allows an analytical and consistent overview and helps to visualize existing land use options or estimate future outcomes. It also provides a faster, more efficient way to identify development scenarios and their impacts.

The Urban Footprint method can be used for policy making, urban planning, urban development, and monitoring. Several researchers have applied the CCD method (Coherent Change Detection) to detect and monitor changes related to natural and anthropic events or modifications of urban landscapes using SAR images.

The Sentinel-1 mission provides continuous all-weather, day-and-night all-year-round imagery of relatively medium-resolution C-band data. Providing imaging products for all global landmasses allows for comprehensive urban target interpretation regardless of geography or location. In particular, the interferometric wide (IW) swath mode is the primary operational mode over land. The data are publicly accessible and provide sufficient resources for land cover applications.

Coherent Change Detection (CCD)

Considering its applicability, we applied coherent change detection (CCD) using the intensity and phase (basic measurements of the backscattered signal) between two SAR images SENTINEL 1 SLC product (Single Look Complex) to identify target changes based on the coherence estimation rate of radar signals, using pre- and post-event coherence differences.

Nowadays, the application of CCD has been extended beyond the study of two coherent SAR pairs for coherence, and it is applied using time series analysis involving the estimated phase correlation between several SAR images acquired along a specific time frame and location to identify the temporal behaviour of these targets based on the stability of their returned radar signals. The coherence between two SAR images expresses the similarity of the radar reflection between them. Any changes in the complex reflectivity function of the scene manifest as a decorrelation in the phase of the appropriate pixels between the two images. The distance between the satellite antenna and the ground targets determines the phase of a SAR image.

Even very subtle changes in the scene from one image to the next can be detected, in a given pixel (a small area of the Earth’s surface), if the reflection and/or dielectric properties have changed between the two acquisitions (return of energy from a surface and/or the physical property of a material that determines how reflective that material), the coherence

the value of that target is reduced.

The coherence estimation is not used only for SAR interferometry (InSAR) applications, it has been used for other types of image processing and interpretation, which involves a stack of coregistered complex images. Interferometric coherence analysis is used for various applications, including coherence changes, to identify tsunami damage and detect liquefied areas. Coherence differential values and coherence ratios between co-event and pre-event are often used in detecting earthquake building damage using coherence.

Surfside-Florida

Almost two years ago, on June 24, 2021, the Champlain Tower, part of a condominium complex in Surfside, Florida, suddenly collapsed. In about 11 seconds, 55 of the building’s 136 units crumbled into a pile of rubble.

Coherence estimation in Surfside-Florida overlapping in Google Earth (June-July 2021)

For this specific event, we applied the change detection method in the residential zone of Surfside- Florida, United States. The image shows low coherence values depicted in black, indicating zones with subtle changes in the scene from one image to the next, during the two satellite acquisitions between the dates June 21 and July 03 of 2021, spanning twelve days. In the same image, it can be seen also the Champlain Tower zone, which seems to be engulfed in such a black area of change. Also, in the area it can be seen several other zones with low values (black) of coherence, showing that the methodology is sensitive and robust enough to detect subtle changes, though with far more consequences. In a previous article, we discussed all these changes which occur for diverse elements on the surface. As such, changes in vegetation: foliage, and weather-related vegetation changes are important sources for detecting changes. Ground disturbances due to natural causes such as earthquakes, tectonic activity, landslides, and sinkholes cause detectable changes to roads and infrastructure. Likewise, new urban developments, such as removing structures or building new ones.

The town of Surfside sits on a barrier island in the Atlantic Ocean, separated from mainland Miami by Biscayne Bay. The island experienced subsidence at about 1-3 mm/year between 1993 and 1999, which may have been the cause of the collapsing building. However, that sinking was not evenly distributed across the entire landmass. In a study published in 2020 in the journal Ocean & Coastal Management, it is suggested that the land beneath Champlain Towers South has been gradually sinking since the 1990s due to subsidence.

Another explanation is that the building collapse was caused by a void or a sinkhole that had opened up under the reinforced concrete pilings that sit beneath the building, causing some pilings to shift downward while others remained in place. Although we cannot say for sure the cause or source of the building collapse, either the subsidence or structural damage, poor building design, or the sinking land beneath the condo, they all have been flagged as possible triggers for the disaster. Regardless of the cause, it is clear that it has been picked up by the coherence estimation method, and it could have been flagged as a warning. Hence, it is necessary that beyond the current regulations and methods which monitor the stability of the infrastructure and land subsidence, to permanently seek new promising ways to detect these subtle but disastrous changes. Moreover, it is recommended to integrate these diverse geoscience techniques, which may lead to better preparedness and resilience within our cities and avoid disasters and life loss.

Coherence estimation in Surfside-Florida during June and July 2021 (View from above overlapping in Google Earth)

The coherence estimation using satellite data provides encouraging (to say at least) results. Even though it does not give the exact answer about the source(s) of such an event, it can contribute to finding clues why and when these events occur. It remains to be seen if it can be used to raise or flag (or even predict though this remains to be further studied) in a timely manner disaster before they happen. There are a lot of efforts internationally to study subsidence, e.g., the UNESCO Land Subsidence International Initiative (LaSII), a more scientific group, is using the integration of geological or geophysical models with satellite data to study and predict land subsidence occurrence and effects. But maybe we can discuss this new development in a future article. 

This article was written in memory of the victims of the Champlain Tower collapse in 2021.

Read more

This article is part of our series on SAR Remote Sensing Applications.

Featured image: This aerial view of the partially collapsed Champlain Towers South, north of Miami Beach, shows search-and-rescue personnel at the site.(Image credit: Chandan Khanna/AFP via Getty Images, taken from https://www.livescience.com/miami-dade-condo-collapse-subsidence-explained.html)
Article edited by Jerry Yao and Andrei Bocin-Dimitriu

Gabriela Quintana Sánchez

Related post