Deformations monitoring in complicated-surface areas by adaptive distributed Scatterer InSAR combined with land cover: Taking the Jiaju landslide in Danba, China as an example

Abstract

Multi-temporal InSAR (MT-InSAR) technique provides a powerful tool for measuring large-scale surface displacements at high precision. However, traditional MT-InSAR approaches always employ a single global threshold to select measurement points (MPs), making it difficult to identify enough MPs in complex land cover areas (e.g., mountainous areas). In this study, to increase the number and spatial distribution density of distributed scatterer (DS) detection while maintaining its quality, we proposed a new MT-InSAR method, named Adaptive Distributed Scatterer InSAR Combined with Land Cover (ADSI-CLC). The key ideas of ADSI-CLC include the introduction of land cover maps as the constraint condition for DS adaptive selection and the employment of coherence weighted spatial adaptive filtering method to estimate optimal phase. Of these, DS adaptive selection method as the most pivotal idea realizes the unification of mathematical, physical, and geographical mechanism in the process of statistically homogeneous pixels (SHP) identification and adaptively adjusts the DS selection thresholds in different land cover areas in the process of DS selection. ADSI-CLC significantly increases the number and spatial distribution density of DS by using the adaptive selection method, and purifies the SHPs populations, and improves the efficiency of SHPs identification through introducing land cover maps. To demonstrate the effectiveness of the ADSI-CLC method, a case study of the Jiaju landslide using 19 scenes ALOS-1 PALSAR-1 L-band data (2006–2011) and 31 scenes Sentinel-1 C-band data (2014–2016) was carried out. The comparison between deformation measurements retrieved by ADSI-CLC and classical StaMPS-SBAS method showed that their distribution pattern is similar, but the ADSI-CLC method can provide more deformation details due to much higher (nearly 10 times) DS density. Quantitative comparisons between the ADSI-CLC results from ALOS-1 data and GNSS measurements also showed that the accuracy of this method is about 11.7 mm/yr in terms of root mean square error (RMSE). These results proved the reliability and effectiveness of ADSI-CLC and indicated its great potential for displacements monitoring in complicated-surface areas with large topographic fluctuations (e.g., mountainous areas).