Environmental modeling plays an important role in the decision-making process of environmental issues in all developed countries around the world. In particular, large air pollution models are used to simulate the distribution, transport, and transformation of pollutants in the atmosphere. These models can help scientists and policymakers predict the levels of the most dangerous pollutants (such as particulate matter (PM), nitrogen oxides (NOx), sulfur dioxide (SO \(_2\) ), and volatile organic compounds (VOC)) in different locations at various times. By doing so, it provides key insights into medium-term and long-term pollution trends. It is particularly useful in locating precisely where and when the pollution levels exceed some critical values, as well as in developing a reliable control system to keep them below these limits (of high importance, especially in the densely populated areas). Sensitivity studies, carried on in air pollution modeling are essential for understanding the reliability and robustness of the models under consideration and the accuracy of their output results. The Danish Eulerian Model (DEM) is the particular air pollution model, discussed here. It calculates the concentrations of a large number of pollutants and other chemical species in the air and their variation along certain time period, taking into account the main physical, chemical and photochemical processes in the atmosphere. Scalability properties of the parallel computer implementation of its sensitivity analysis code (SA-DEM), run on the most powerful supercomputers in Spain and Bulgaria (namely, IBM MareNostrum III and Discoverer), are presented in this paper.