Climate change in the 21st century is predicted to push ecosystems across ecological thresholds, potentially resulting in abrupt ecosystem change into new and irreversible states. Ecological theory proposes that non-linear biotic responses are the result of a complex interplay of feedbacks, thresholds and interactions that operate over decades to thousands of years. As a result, standard ecological research methods are generally unable to quantify key ecological dynamics that are relevant for forecasting abrupt ecological change and there is a critical need to integrate long-term ecological data with process-based models. This will result in improved forecasting of climate-change impacts on ecosystems at both local- to global-scales. Such studies will play a critical role in understanding the ‘intrinsic’ factors (e.g. climate-vegetation feedbacks) that can result in non-linear biotic responses to climate change.Sediments are natural data-loggers that preserve the remains of plants and animals over thousands of years. They provide a unique resource for answering current high priority questions related to predicting future ecosystem change because they are the only way to obtain empirical information relevant for understanding long-term ecological dynamics and functioning. In this project I will develop an interdisciplinary framework that integrates state-of-the-art process-based modelling with new high-quality palaeoecological information to quantify the factors that result in non-linear responses to climate change. I will apply the framework to a major vegetation transition in the past: the sclerophyll-rainforest transition in north-east Australia that occurred between 10 and 7 thousand years ago. I will develop this case study for proof-of-concept of a new interdisciplinary framework. This will result in a greater understanding of non-linear biotic responses to climate change.