Identification of the parameters describing primary production from continuous oxygen signals

Abstract

The parameters of a simple dissolved oxygen (DO) model have been estimated on a day to day basis from continuous oxygen measurements. Oxygen was measured in a shallow ditch (0.38 m) dominated by bentic algae. The main objective was to study the identifiability of the parameters P_max and I_opt in the photosynthesis vs. irradiance (P/I) curve, in order to determine the seasonal variations of the parameters and to study relationships with environmental conditions. For this purpose joint parameter confidence regions, parameter correlation, and the eigenvalues of the Fisher information matrix have been determined. The results show that P_max can be estimated fairly well, whereas for I_opt the confidence regions are much larger. In particular for days in winter and spring extremely long elongated confidence regions can be found. However, it can be concluded that for a sufficient number of days, both parameters in the P/I curve can be estimated sufficiently accurately to study seasonable variability and dependency with temperature and light conditions.