Reconstructing the dynamics of intermittent coastal inlets in Australia form publicly available satellite imagery: Early results from InletTracker, a new open-source Python toolkit

Heimhuber V1

1University Of New South Wales


Despite their high ecological and socio-economic significance, there is often a lack of knowledge on the long-term dynamics of intermittent coastal inlets such as those found at intermittently open and closed lakes and lagoons (ICOLLs) or barrier islands. Australia alone is home to 305 (21%) of the 1477 globally mapped ICOLLs and for many of them, there are no long-term records of their time-evolving entrance state, which represents a challenge for coastal managers and decision-makers. Here, we present early results from InletTracker, a new tool for reconstructing the dynamics of intermittent coastal inlets or similar landforms over the last 30+ years from publicly available Landsat 5, 7 and 8 and Sentinel-2 satellite imagery. InletTracker is a Google Earth Engine enabled python tool that uses a novel least cost path finding approach to trace inlet channels along and across the barrier/berm, and then analyses the resulting transects to infer the minimum channel width and whether an inlet is open or closed. The tool is easy to use and provides users with data on the location and shape of entrance channels, the width of entrance channels at the bottleneck (i.e., at the throat) and the open vs. closed state of an inlet for all sufficiently cloud-free satellite images. In a previous study, the skill and limitations of InletTracker were demonstrated based on a validation exercise that involved 12 ICOLLs in Southeastern and Southwestern Australia. Here, we applied the public version of the tool to 20 intermittent coastal inlets that are either of particularly high socio economic importance (i.e., Murray River mouth, SA) or have undergone significant changes between 1986 to present (e.g., Lake Illawarra, NSW). Our results revealed large differences in the frequency, duration, size and seasonality of inlet openings across the 20 demonstration sites and highlight the dramatic morphological changes that some of the inlets have undergone since 1986. The data provided by this experiment will enable coastal managers and researchers to gain new insights into dynamics and drivers of intermittent coastal inlets. Further, the improved understanding of the long-term entrance dynamics will enable a more evidence-based approach to entrance management, in particular for sites for which this type of information did not yet exist.


Valentin (Tino) Heimhuber is an environmental researcher & engineer working on the integrated management of river systems, wetlands, and estuaries under climate change and ongoing population growth. His research focuses on the use of large satellite datasets and machine learning to develop new tools for improved management of water resources and water-dependent ecosystems. Valentin is particularly interested in using data to find the right balance between human and environmental water requirements in catchments, rivers and coastlines.