What we found¶
We built a Karst Intrinsic Vulnerability Index (KIVI) at Mole Creek (25 m), applied it across Tasmania (100 m), and combined it with LIST land use, slope and CFEV conservation value, using the same code and fixed classes at both scales.
A KIVI for Tasmania. Susceptibility (Atlas category × exposure) times connectivity (D8 flow routing) runs at 25 m and 100 m with the same five classes. Across the state, 47% of karst cells are High and 9% Very high. Exposure is part of the score because it lifts the springs check from about the 96th to the 99th percentile.
Scale. A 2 m elevation model gives the same map as 25 m (0.99), and the 100 m grid changes it a little (0.89). The 100 m run scores Mole Creek higher, partly because its routing includes land beyond the 25 m buffer.
Sensitivity. Method choices move the map most. The six-input weighted sum gives a very different Mole Creek map (0.55), and adding the four ingredients changes the state’s ordering (0.70). Tasmanian land use and national land cover give the same Mole Creek ranking (0.98) and differ more across the state (0.84), because the Tasmanian layer separates plantations and irrigated pasture.
KIVI with land use. Multiplying KIVI by LIST 2021 pressure and the slope multiplier gives the risk map. Risk is highest in the west of Mole Creek, and Flowery Gully, Doctors Rocks and Gunns Plains lead the statewide ranking of larger systems. Over the DEA series from 1988 to 2024 the level of risk follows land use and the pattern stays fixed.
KIVI with conservation value. Combining KIVI with CFEV Integrated Conservation Value gives a four-class priority map. The top class covers 13% of the karst, mostly in the south-west mountains and at Mole Creek (Results, section 7).
What we would do next¶
Add ground truth. Ground-truthing at Mole Creek, with water-quality and dye-tracing records and records of karst damage and known contamination, would let us test the map against measured outcomes.
Route pressure along the flow paths. Pressure is a mean over each named system. Routing land use along the flow paths into each karst polygon describes what reaches the karst more closely.
Run each LIST snapshot. LIST has six snapshots (2002, 2009–10, 2013, 2015, 2019 and 2021), and running the model on each shows how risk moves with land use.
Keep the sinks. Flow routing that retains sinks (Whitebox, RichDEM) may suit karst, where sinks can be real sinkholes.
Update the Karst Atlas. Updated karst boundaries would improve every ingredient built on the Atlas.
Improve the drainage data. More accurate DEMs and dye-tracing would show both surface and subsurface drainage networks.
Add logging coupe data. Historical and planned coupe data would show past impacts and identify future risks.
Fill the missing inputs. Soil cover, depth to groundwater and sinkhole density would allow a literature-style index for Tasmania.
Flag the small systems. A confidence mark based on each system’s cell count shows where the 100 m map is least reliable.
Are we there yet?¶
Feasibility. Our project focused on the feasibility of developing a scalable, Python-based workflow that could meet our objectives while allowing us to test the sensitivity of different inputs and modelling choices. We feel that we have developed a workflow that demonstrates this feasibility, although we would not claim that our choices of inputs, class boundaries and weightings are necessarily the most appropriate. The workflow was designed so that these can be adjusted relatively easily as our understanding of karst vulnerability improves and more accurate or up-to-date datasets become available.
What we underestimated. The complexity of calculating karst vulnerability and risk associated with different land uses was an aspect of the project that we underestimated but feel that this is an important area for further research, particularly given the hidden nature of karst systems and the potential for land-use activities to create damage and contamination not immediately apparent at the surface. This reinforces the importance of continuing to investigate the Hidden Landscapes and Unseen Risks associated with karst geology.