Abstract Specialized translation means that translators must quickly acquire a certain level of specialized knowledge. For this purpose, they must develop documentation skills and have a range of effective knowledge resources at their disposal, such as terminological knowledge bases (TKBs). From a cognitive perspective, Meyer et al. (1992) state that the conceptual categories in TKBs should be structured in a similar way to how they are related in the brain. Lexicographic and terminographic tools should account for the dynamic way in which we conceptualize the world around us, and that is the reason why TKBs are valuable tools in specialized translation (Tercedor et al. 2012). Furthermore, a conceptually-structured TKB, in which terms are linked to concepts based on non-language-specific criteria, is a valuable resource for specialized translators. In this way, not only is there coherent cross-referencing, but also linguistic data can be added and manipulated without altering the quality and consistency of the conceptual design (Giacomini 2014). EcoLexicon (http://ecolexicon.ugr.es) (Faber et al. 2014), a TKB on the environment science, is a good example of a modern terminology management tool. It specifically targets user knowledge acquisition through different types of multi-modal and contextualized information, in order to respond to both cognitive and communicative needs. However, EcoLexicon has certain issues still to be resolved, such as occasional noise and information overload. This paper analyzes the effectiveness of EcoLexicon for specialized translation and discusses certain problems derived from an overly simple definition of generic-specific relations. In this line, we explore and assess ‘umbrella concepts’ as a means of restricting the sense of hyponymy. Moreover, we describe the context and methodology for creating them and increasing their number. Our study resulted in the specification of a new and enhanced set of ‘umbrella concepts’ that have improved the conceptual structure of the EcoLexicon knowledge base. References Giacomini, L. (2014). Testing User Interaction with LSP e-Lexicographic Tools: A Case Study on Active Translation of Environmental Terms. In Proceedings of the 12th edition of the Konvens Conference, pp. 77–85. Hildesheim, Germany. Faber, P., León-Araúz, P., Reimerink, A. (2014). Representing Environmental Knowledge in EcoLexicon. In Bárcena, E., Read, T. & Arhus, J. (Eds.) Languages for Specific Purposes in the Digital Era, Educational Linguistics 19, pp. 267-301. Berlin, Heidelberg: Springer. Meyer, I., Bowker, L. & Eck, K. (1992). COGNITERM: An Experiment in Building a Knowledge-Based Term Bank. In Proceedings of the Fifth EURALEX International Congress (EURALEX ’92), pp. 159-172. Tampere, Finland: Tampereen Yliopisto. Tercedor, M., López Rodríguez, C. I. & Faber, P. (2012). Working with Words: Research Approaches in Translation-oriented Lexicographic Practice. TTR: Traduction, Terminologie, Rédaction, XXV (1), pp. 181–214.