Learning from past Land-Use Dynamics for Risk-Sensitive Development in the Future: A Remote Sensing based Analysis spanning the last Five Decades in the Ayeyarwady Delta of Myanmar
Anissa VOGEL, Institute of Geography, University of Cologne, Germany
Katharina SEEGER, Institute of Geography, University of Cologne, Germany
Helmut BRÜCKNER, Institute of Geography, University of Cologne, Germany
Dominik BRILL, Institute of Geography, University of Cologne, Germany
Frauke KRAAS, Institute of Geography, University of Cologne, Germany
Land-use and land-cover (LULC) change has affected about one third of the global land area in just six decades since 1960, with drastic consequences accumulating particularly in developing countries (Winkler et al. 2021). River deltas are hotspots of this Anthropocene transition (Nicholls et al. 2020). While they are among the most economically and ecologically valuable environments worldwide (Nienhuis et al. 2020), deltas are exposed to significant risk emerging from long-term processes of maldevelopment (Santos & Dekker 2021).
However, while international experts and decision makers are calling for “freely available, reliable, quantitative scientific information to further improve knowledge and understanding of river delta environments” (Kuenzer et al. 2019), the Ayeyarwady Delta in Myanmar remains one of the least studied megadeltas worldwide.
Against this background, this study applies a random forest machine learning approach using only freely available remote sensing data and open source software for the assessment of LULC dynamics in the Ayeyarwady Delta. A detailed, area-wide and consistent data set is provided, incorporating in total 50 optical and radar Landsat and Sentinel-1 satellite images, spanning nearly five decades (1973-2021) in 5-year intervals. The findings of this study revealed rapid transformation processes rooted in the colonial past of the delta and identified four major risk-related drivers of the LULC dynamics: the state-facilitated expansion of agricultural areas and irrigation facilities; large-scale aquaculture; urban expansion towards hazard-prone areas; the loss of mangroves as natural bioshields against high energy events.
The reliable long-term LULC data provided for the delta is a valuable data basis that contributes to a better understanding of disaster risk creating patterns and demonstrates that rapid action towards risk-sensitive development is required.
Mots clés : Ayeyarwady Delta|Remote Sensing|Random Forest|Land-Use and Land-Cover Change|Disaster Risk Drivers
A104930AV