Environmental managers are interested in the composition of species in waterways in northern Queensland, Australia. Traditionally, the sampling of the species composition has been recorded on a regular time frequency. However, the locations of the sampling sites are limited in access making sampling quite expensive. We propose a new and novel approach to selecting sampling times based on the dynamics associated with the populations of species residing in the site. This employs the use of the predator-prey models, such as the Lotka-Volterra differential equations, to simulate species composition. Results will be provided for each of the following optimality criteria, A optimal, E optimal, D optimal, and predictive performance.