Spatiotemporal water dynamic modelling of Ramsar-listed lakes on the Victorian Volcanic Plains using Landsat, ICESat-2 and airborne LiDAR data

Abstract

Spatiotemporal dynamic information on surface water area and level is a prerequisite for effective wetland conservation and management. However, such information is either unavailable or difficult to obtain. In this study, for the first time, we leverage Landsat imagery, ICESat-2 and airborne LiDAR data to develop time series of water body dynamics over the last 35 years (1987–2021) using machine learning method on a cloud computing platform for lakes identified as international importance in the Western District Lakes Ramsar site in Victoria, Australia. Our results reveal distinct seasonal (dry and wet) variation patterns and long-term changes in trends of lake water areas and levels in response to seasonal rainfall variations and regional climate changes for the periods of before, during and after the Millennium Drought when southeast Australia experienced unprecedented dry conditions. Lake water bodies have not recovered to the status of pre-Millennium Drought, and many permanent Ramsar-listed lakes in the region have become to ephemeral lakes due to climate change. The outcome of this study provides a baseline to help understand the historical and ongoing status of the Ramsar-listed lakes in a warming and drying climate in support of the development of strategic plan to implement international obligations for wetlands protection under the Ramsar Convention.