It is not possible to study many scientific and natural phenomena in laboratory conditions, so they are expressed in the form of mathematical models and simulated with complex computer codes. Running a computer model with different inputs is called a computer experiment. Statistical issues have a special place in computer experiments.
In this article, while explaining the structure of these models, we introduce and explain the importance of variance-based sensitivity analysis. Sensitivity analysis is a set of methods that determine the effectiveness of input parameters on the output of the model with sensitivity indices. These indices are expressed based on the concepts of conditional variances. Since the mathematical form of these models is not explicitly known, the issue of estimating these indices with Monte Carlo-based methods is raised. On the other hand, execution time is a serious challenge in computer models. A special test point design based on pseudo-random numbers is proposed to reduce the model execution time. In order to address the practical aspect, the INCA-N model, which simulates the amount of nitrogen pollution entering river water, has been used to identify the variables affecting this factor threatening human health and the environment with sensitivity indices.