During the last decade, environmental flows have been measured with a growing set of instruments, which can now offer relatively high resolution in space and time. This growing amount of data can analysed following a "statistical approach to turbulence", rather well established in terms of theory and laboratory results. I will discuss three studies, two on the near bottom boundarly layer turbulence of the deep ocean, and one on the internal wave variability in Lake Geneva. I will try to show that this approach is particularly useful in the field, where the parameters of the flow cannot be controlled. A detailed description of the statistics of a natural flow can enable precise connections with established theories, and can even suggest new theoretical developments.