During the 1976 and 1977 growing seasons the growth of wild rice was quantified at four sampling stations on the Mississippi River near the Clay Boswell Steam Electric Station at Cohasset, Minnesota. Cluster analysis was used to illustrate the intercorrelations and seasonal trends which existed among the water and sediment chemical variables. All biological, chemical, and physical variables were corrected for any time dependency, and discriminant analysis isolated those factors which could separate the four sampling stations. The discriminant functions were composed of biological (wild rice density), chemical (alkalinity), and physical (water temperature, water depth) factors previously known to influence wild rice growth. A model of the growth of wild rice was predicted effectively for each of the four sampling stations by combining (i) a time-independent equation formed from a multiple regression analysis of the time-corrected biological, chemical, and physical variables isolated as being significant in the discriminant analysis, versus the time-corrected dry weights per wild rice plant, and (ii) a time-dependent equation formed by fitting a logistic equation to the overall mean weights per wild rice plant at the four sampling stations.
展开▼
