Event Aligned MODIS Dataset for Mangrove Canopy Response to Tropical Cyclones in the Bay of Bengal
Description
This dataset provides an event-aligned Earth observation record for analysing mangrove canopy responses to tropical cyclones in the Bay of Bengal. It integrates IBTrACS v04r01 cyclone tracks and quadrant wind radii, Global Mangrove Watch v3 mangrove masks, and MODIS Terra and Aqua NDVI and EVI time series with decoded MODIS quality-assurance flags. The dataset includes exposed and matched-control mangrove cells, wind-threshold exposure classes, storm-overlap indicators, contamination flags, and recovery or censoring metrics. Version 2 provides the corrected and updated dataset. The strict-QA seasonal baseline criterion was consistently applied using a minimum of 24 valid pre-landfall observations under the corresponding QA definition. This correction prevents unstable harmonic baseline estimation when observations pass the relaxed QA criterion but do not have sufficient strict-QA observations. Associated derived tables and outputs were updated accordingly. The dataset supports reproducible analysis of cyclone-related mangrove greenness responses, exposure-dependent effects, post-event recovery, robustness to quality filtering, and event-level heterogeneity. Analysis code is available at: https://github.com/RADiptho/mangrove-cyclone-greenness
Files
Steps to reproduce
1. Download and extract the Version 2 dataset. 2. Clone the analysis repository: https://github.com/RADiptho/mangrove-cyclone-greenness 3. Install the Python dependencies listed in `requirements.txt` or create the environment using `environment.yml`. 4. Place the downloaded dataset in the local `data/` directory as described in `data/README.md`. 5. Run the numbered scripts in the `scripts/` directory in sequence, beginning with data preparation and quality control and continuing through baseline estimation, matching, event-study analysis, inference, diagnostics, and figure generation. 6. For strict-QA analyses, the seasonal baseline requires at least 24 valid pre-landfall observations under the strict-QA definition. This criterion is implemented in the released code. The workflow uses IBTrACS v04r01 cyclone information, Global Mangrove Watch v3 mangrove extent, and MODIS Collection 6.1 MOD13Q1/MYD13Q1 observations. Large upstream source datasets are not duplicated in the repository and remain subject to their original providers' terms and licences.
Institutions
- Jahangirnagar UniversityDhaka Division, Dhaka