Creating digital twins of parts of ecosystems is being explored on several occasions (BioDT, Edito, Local projects). These digital twins are created both to support operational practices, to inform policy and to strengthen scientific understanding. The practice of digital twinning often combines ecological modelling, machine learning or AI and direct linkage to data to create a near-real time or nowcasted representation of an ecological system. Creating a digital twin requires that (long term) data and models are available and FAIRly shared. In this session we will showcase several digital twins that are created for ecosystems and highlight how the practice of digital twinning has contributed to the purpose of the research. We especially would like to focus on the added value of long-term ecological research data, as are available at LTER-sites, in creating digital twins.