The automation of application deployment is vital today as manually deploying applications is too slow and error prone. For this reason, various deployment automation technologies have been developed that process deployment models to automatically deploy applications. However, in many scenarios, these deployment models have to be created in collaborative processes involving multiple participants that belong to independent organizations. For example, in data analytics scenarios, often external data scientists develop algorithms to process business-critical data of a company, while IT experts specify the technical infrastructure to deploy algorithms and data. However, as these deployment modeling processes are typically highly iterative and as the participating organizations may have competing interests, the degree of trust they have in each other is limited. Thus, without a guarantee of accountability, iterative collaborative deployment modeling is not possible in business critical domains. In this paper, we propose a decentralized approach that aims at achieving accountability in collaborative deployment modeling processes by utilizing public blockchains to store intermediate states of the collaborative deployment model in a way that guarantees its integrity and allows obtaining the history of changes it went through. The approach utilizes the same blockchain to establish the identity and authenticity of all participants of the process. We validate our approach by providing an architecture and a prototypical implementation of a blockchain-based deployment modeling environment based on the TOSCA standard.