AbstractWhen in search of a solution for a problem in ecology or in the field of water resources the physical system has to be quantitatively specified. Such a specification is most concisely given under the form of a mathematical model. Not only important parameters and physical constants are provided, but also relationships between variables are presented. System theory gives a general answer concerning the constituent parts of the mathematical model representation. The elements are : a frame a structure and a parameter set. The frame specifies the input, output and state sets, the structure includes the functional relationships between set elements and the parameter set groups the physical constants. Whenever important ecological parameters like coefficient, etc., have to be found, parameter estimation permits to extract those constants out of the measurements. For a given frame, the parameter estimates will depend on the choice of the structure. The precise form of the structure is often a matter of debate. A parameter estimation procedure based on normalized sensitivity functions is developed. The scheme has the advantage of limiting errors which result from similar but wrong model structures. Robustness towards structural ambiguity is obtained.