The measurement of water quality parameters, such as chlorine and trihalomethane (THMs) concentrations, is usually performed in only some points of water distribution networks. Therefore, models are created to obtain an approximation to the real behavior of the system. However, in the trunk network of Bogota's water distribution system, the measurements are reported with a frequency greater than the retention time of the network, and they are only performed in a few points throughout the system, which limits model creation. Consequently, this document proposes a non-formal calibration using Monte Carlo simulations for a multispecies model of this network built in EPANET-MSX. The first-order models were adopted for the bulk and wall reactions of chlorine decay, and the linear model was used for TTHMs formation. The methodology consisted in the calibration of the yield of TTHMs coefficient, which was calculated for each pipe depending on a tracer analysis performed for the three water treatment plants (WTPs) of the network. The objective function was built in terms of descriptive statistics given the scarcity of measured data. In this manner, even though the calibration performance was satisfactory, it was found that the parameters for two of the WTPs are not identifiable nor sensitive. Moreover, given the limitations of the method and the scarcity of information, the estimated parameters are considered to have a high uncertainty level. Hence, a higher frequency of measurements is required for a more thorough and less uncertain calibration.