In software engineering of self-adaptive systems, uncertainties arise from the inherent complexity of these systems (i.e., the limitation to foresee all aspects related to their development and operation) and the unpredictability of the execution environments (i.e., the limitation to predict all the events), among other sources. Several taxonomies and specification languages have been proposed for uncertainties. However, they remain at a theoretical level and are hardly applied in practice. We believe the underlying problem is that researchers and practitioners lack best practices and concrete examples of how to apply the proposed taxonomies and languages in real systems. To address this gap, we contribute a catalog of uncertainties. All the uncertainties in the catalog have been extracted from concrete artifacts. Each uncertainty is described using the same template and associated with requirements specified in the RELAX language. Our objective is to provide an initial version of a catalog of uncertainties and requirements that can be (re)used as prototypical examples in different architectural solutions in various application domains. This will illustrate the strength and versatility of the chosen taxonomy and language, helping to facilitate their adoption. We believe this will also consolidate the terminology and provide a solid basis for tracing uncertainties from design to implementation and eventually to runtime. We envision that this catalog can be extended with more examples and enhanced by uncertainty patterns and corresponding adaptation strategies, providing much-needed guidance to practitioners.