Change point detection on time series satellite data for assessment of cyclonic impact on coastal mangroves

Understanding the cyclonic uncertainty impacting the mangrove community depends on detecting change points in time series. These transitions are classified as sudden changes or pattern shifts within the ecosystem. The date of sudden changes is indicated using Hurst t-statistics, whereas pattern shifts are identified through chaos equations and visualized with Burger plots. We use the Bayesian change point detection method to identify nonlinear responses and abrupt changes in mangrove health, indicating that small environmental stresses lead to large ecosystem changes over time. Statistical tests, including t-tests, p-values, and standard error computations, are employed to confirm the reliability of these change point evaluations. We assess the impact of cyclones Fani (May 3, 2019) and Amphan (May 20, 2020) on Henry Island Sundarban, West Bengal, India by analyzing four regions-of-interest (ROIs) using the Mangrove Vegetation Index, mangrove forest index (MFI), and leaf area index (LAI) time series Sentinel 2A/1C data; pre-cyclone (2016 to 2018), during-cyclone (2019 to 2020), and post-cyclone (2021 to 2022) periods. Post-Cyclone, ROI 1 exhibits partial recovery with a slight increase in the Hurst Exponent (HE) and stabilization in mangrove health by June 2019; ROI 2 shows a decrease in HE and fluctuating stability with change points on June 4, 2019, and July 13, 2020; ROI 3 experiences persistent stress with reduced long-term persistence, particularly on June 29, 2019, and May 24, 2020; and ROI 4 demonstrates substantial recovery with stabilized MFI but ongoing instability in LAI, with major change points on May 10, 2019, and October 7, 2019. Ground survey on ROI 4 reveals dominant species such as Ceriops benefit from the clustered structural support provided by dominating mangrove species such as Avicennia and Excoecaria. We confirm that the Hurst t-statistics method identifies the same change points as the Bayesian approach. Burger's chaotic maps visually support these findings by illustrating significant cyclone-induced disturbance patterns across ROIs.