The habitat quality index applied to New Mexico streams

The accuracy of two trout biomass (standing stock) prediction models, developed for Wyoming streams by Binns & Eiserman (1979), was evaluated for New Mexico streams inhabited by brown trout, Salmo trutta L. and rainbow trout, Oncorhynchus mykiss Walbaum. Thirty-two representative sites in 15 different streams were sampled under summer low-flow conditions in 1988 and 1989. The 11 phyiscal, chemical, and biological variables used in original models were used as independent variables for simple and multiple regression analysis designed to predict total trout biomass. Model I of Binns and Eiserman proved to be of limited utility; it explained 53% of the variation in total trout biomass at each of the New Mexico sites (kg ha(-1) = 8.916 + 0.830/Model I/). Only 9.5% of the biomass variations was explained by Model II. Statistical analysis showed that trout biomass was significantly correlated with nitrate-nitrogen concentration and macroinvertebrate diversity in Model I. Because both variates are time consuming to estimate, Model I may not be any more cost-effective than sampling trout directly. The low predictive power of Model II probably indicates that it is limited to the geographical area of field measurement origin.