The demand for comprehensive environmental assessment of river ecosystem has increased for engineers and scientists. Accurate and versatile water temperature models are required to meet this demand. A number of hydrological models take vegetation and soil characteristics into account, but very few temperature models do. The objective of this paper is to incorporate soil temperature and vegetation as input variables in a deterministic heat budget model. The CEQUEAU hydrological and water temperature model was used to simulate water temperature in Catamaran Brook, a small catchment located in central New Brunswick. The model was modified by incorporating soil temperature as a parameter influencing the temperature of interflow, using the so-called force-restore method. Crown closure was also incorporated in the model as a factor influencing locally advected water using a negative exponential function. The modified model simulated daily water temperatures better than the original model. Root-mean-square error for a period of 5 years decreased from 2.10°C with the original model to 1.77°C with the modified model. Nash coefficient increased from 0.78 with the original model to 0.82 with the modified model. An analysis of residuals showed that the modified model is sensitive to additional parameters such as crown closure, especially for short time scales during periods of higher discharge and during extreme meteorological and hydrological events such as tropical storms.Key words: stream temperature, hydrology, deterministic model, CEQUEAU, forestry.

The Eaton River precipitation network was set up as one of the Quebec International Hydrologic Decade projects in 1965. At the end of the decade, in 1975, 30 stations had been in operation for most of the period in this 248‐mi2 (643 km2) basin. After a principal components analysis of the data from 14 of the 30 stations on 10‐day precipitation totals, it was found that the stations could be divided into three groups, the composition and geographic distribution of which change from season to season. It was also possible to identify which stations were redundant and could be closed should it become necessary. Application of optimal interpolation, season by season, to a 30‐ and a 5‐station network showed that the precision of point interpolation varies more from season to season for the same network than from the 30‐ to the 5‐station network for the same season. In fact, errors of interpolation caused by microclimatic irregularities and by observational errors in the initial data are greater than those resulting from the reduction of the network from 30 to 5 stations.

No abstract

A stochastic model of water temperature has been established and then applied to data taken from measurements on the Sainte-Anne River, Québec; its development uses the Box–Jenkins modelling approach. As well, an existing deterministic water temperature model had been applied to the same data and a comparative study between the two models is made. This comparison contains a graphical daily analysis, a statistical analysis including the calculation of root-mean-square error, and a comparative performance study. With these criteria, no overall domination of one model over the other has been noticed. Nevertheless, an analysis based on the purpose of the model and on the data available allows choice of the appropriate model. Key words: model, temperature, river, stochastic, deterministic, performance.

Hydro-Québec is projecting to increase the hydroelectric production capacity of the St. Marguerite River by diversion of the tributaries Pékans and Carheil rivers of the Moisie River, the most productive salmon river of the whole Quebec. Along with substantial changes in hydrological regimes, this hydroelectric development is most likely to affect some physical environment factors such as the water temperature, which is of prime importance for the biotope and, in particular, for the salmon productivity. The objective of the present study is to simulate, over a long period of time, the river water temperatures under natural conditions as compare to those after the impoundment, to assess the consequences of the tributary diversion. We used the hydrological CEQUEAU model coupled with a temperature model.The temperature model developed is applicable to the ice-free period and calculates daily water temperatures in rivers by computing an energy budget to each element of the watershed. The energy budget considers the short-wave solar radiation, long-wave radiation, evaporation, and convection in the air as well as the advective heat of various inflows from surface runoff, interflow, and groundwaters. The estimation of the atmospheric thermal exchanges is based on the equations usually found in literature. The volumes of the various inflows are given by the hydrological model. The temperature model uses daily data for air temperature and monthly data for solar radiation, cloudiness, wind speed, and vapour pressure.The model has been applied to the Moisie River (Québec), using the measured values for the calibration. Both observed and calculated values show good agreement. The model was also used to simulate, over the whole watershed, the water temperatures for the 1961–1989 period and after the diversion. The results show that the tributary diversion contributed to increase the water temperature of the Moisie River and that this increase is gradually attenuated as we progress downstream. Key words: temperature, impacts, model, Moisie, Québec, diversion, hydrology.