Just Accepted
This article has been peer reviewed and accepted for publication. It is in production and has not been edited, so may differ from the final published form.
Habitat suitability modelling of the North Flinders Ranges thick-billed grasswren Amytornis modestus raglessi reveals dynamic shifts at the landscape scale
Abstract
Context The threatened subspecies of thick-billed grasswren Amytornis modestus raglessi, occupies the lower slopes and peripheral drainages of the North Flinders Ranges, South Australia, where it is dependent on dense chenopod shrublands as a preferred habitat. A decline in grasswren numbers was observed around 2012, after two preceding years of exceptionally high rainfall, and a profound reduction in observed numbers was evident in 2019, the second successive year of exceptionally hot and dry conditions, in areas where they had previously been numerous. Aims To use habitat suitability modelling to identify environmental factors influencing A. m. raglessi habitat suitability, and to use those findings to better understand possible drivers of grasswren decline and changes in distribution between the pre- and post-2012 periods. Methods Random forest modelling was used to predict grasswren habitat suitability in response to mapped environmental variables including remotely sensed vegetation, soil and landscape properties. Key environmental variables were identified from the modelling. Habitat suitability maps were produced for two separate periods, 1994-2011 and 2012-2023 and compared. An ornithological field survey was undertaken to validate the modelling, and vegetation time-series were examined to compare areas showing contrasting differences in habitat suitability. Key results The most important model predictors for grasswren habitat suitability were mapped soil properties and the minimum green vegetation cover value. The overall predicted area of habitat (suitability > 50%) declined 25% between the 1994-2011 and 2012-2023 periods. However, the changes included an expansion of high-suitability habitat in areas in the west, and a contraction in the south-eastern part of the distribution. Time-series vegetation data showed that lower bare ground cover, and higher non-green vegetation cover occurred in an area which showed a marked reduction in predicted habitat suitability. Implications The methods used in this study provide spatially explicit guidance for prioritizing conservation efforts for this subspecies and thick-billed grasswrens broadly. The demonstrated habitat improvement following reduced grazing at Witchelina illustrates the practical value of this modeling approach. Such methods are increasingly essential for land managers to understand biodiversity responses and species distributions under climate change.
WR24152 Accepted 02 September 2025
© CSIRO 2025