Anik Daigle

Multivariate analysis of the low-flow regimes in eastern Canadian rivers

Abstract A characterization of the low-flow regimes of 175 eastern Canadian rivers based on multivariate analysis of hydrological indices (HIs) is presented. Principal component analysis (PCA) was used to identify eight highly informative and low-correlated HIs amongst 67 low-flow HIs reported in the literature, and to test their ability to describe regional characteristics and differences among low-flow regimes at the 175 stations. It was found that eight HIs can provide a regional description of the main low-flow characteristics almost equivalent to using all HIs related to low flows. The PCA also identified regional similarities and differences among the geographical and hydrological regions within the Province of Quebec, as well as between the Atlantic provinces of Canada. Multivariate analysis proved to be an efficient tool that can complement expert knowledge in the selection of criteria for instream flow assessments in eastern Canada, by quantitatively indicating indices carrying high information, and by ensuring low redundancy in the selected subset of variables (HIs). Citation Daigle, A.,St-Hilaire, A., Beveridge, D., Caissie, D. & Benyahya, L. (2011) Multivariate analysis of the low-flow regimes in eastern Canadian rivers. Hydrol. Sci. J. 56(1), 51–67.

Climate change and resilience of tributary thermal refugia for salmonids in eastern Canadian rivers

Abstract River water temperature regimes are expected to change along with climate over the next decades. This work focuses on three important salmon rivers of eastern Canada, two of which warm up most summers to temperatures higher than the Atlantic salmon lethal limit (>28°C). Water temperature was monitored at 53 sites on the three basins during 2–18 summers, with about half of these sites either known or potential thermal refugia for salmon. Site-specific statistical models predicting water temperature, based on 10 different climate scenarios, were developed in order to assess how many of these sites will remain cool enough to serve as refugia in the future (2046–2065). The results indicate that, while 19 of the 23 identified refugia will persist, important increases in the occurrence and duration of temperature events in excess of 24°C and 28°C, respectively, in the mainstems of the rivers, will lead to higher demands for thermal refugia in the salmonid populations. Editor Z.W. Kundzewicz; Associate editor T. Okruszko

Multivariate Modelling of Water Temperature in the Okanagan Watershed

Initiatives for the protection of river ecosystems must include the monitoring of key flow and water quality variables, as well as clear and quantifiable management goals. One variable which strongly influences water quality is water temperature, and its modification arising from human activity should be incorporated into ecosystem protection guidelines. This work, conducted as part of the Canadian National Agri-Environmental Standards Initiative (NAESI) program, presents a preliminary study investigating statistical regression methods and a geostatistical approach to model key water temperature characteristics that could assist in the development of standards. Water temperature time series recorded at 16 sites in the Okanagan watershed were used to develop the models. Monthly maxima were modelled for the period of April-September 2007 using four predictors: the site longitude, the drainage basin maximum altitude, the local slope at the station, and the log of the mean substrate diameter. Four types of multivariate regressions of monthly maxima were produced, and a leave-one-out resampling approach was used to validate the models. Relative Bias, Root Mean Square Errors (RMSE) and a corrected Akaike Information Criterion (AICc) were calculated for each month. Models gave RMSE values between 0.9°C and 2.1°C for the monthly maxima. All models generally performed best between May and July. Geostatistical interpolation of maxima was also performed in a multivariate physiographic space reduced to two orthogonal dimensions using canonical correlation analysis (CCA). Examples of interpolated maps show that the approach can be used to discriminate between warm and cool streams.

Statistical downscaling of precipitation and temperature using sparse Bayesian learning, multiple linear regression and genetic programming frameworks

This study attempted to investigate two approaches to downscale temperature and four approaches to downscale precipitation. The first approach was an implementation of multiple linear regression (MLR) in the form of backward stepwise regression. The second approach applied canonical correlation analysis (CCA) with a sparse Bayesian learning (SBL) approach called relevance vector machine (RVM). For precipitation downscaling, two additional approaches which combined genetic programming (GP) as the predictor processing method with sparse Bayesian learning (SBLGP) and multiple linear regression (MLRGP) were also presented. The results showed that SBL outperformed MLR in downscaling temperature. For all stations, temperature was better downscaled than precipitation. For precipitation downscaling, the SBLGP approach outperformed all other approaches. MLRGP, on the other hand, did not bring about much improvement in the results and was in many cases outperformed by MLR.

Comparison of direct statistical and indirect statistical-deterministic frameworks in downscaling river low-flow indices

ABSTRACT This work explores the ability of two methodologies in downscaling hydrological indices characterizing the low flow regime of three salmon rivers in Eastern Canada: Moisie, Romaine and Ouelle. The selected indices describe four aspects of the low flow regime of these rivers: amplitude, frequency, variability and timing. The first methodology (direct downscaling) ascertains a direct link between large-scale atmospheric variables (the predictors) and low flow indices (the predictands). The second (indirect downscaling) involves downscaling precipitation and air temperature (local climate variables) that are introduced into a hydrological model to simulate flows. Synthetic flow time series are subsequently used to calculate the low flow indices. The statistical models used for downscaling low flow hydrological indices and local climate variables are: Sparse Bayesian Learning and Multiple Linear Regression. The results showed that direct downscaling using Sparse Bayesian Learning surpassed the other approaches with respect to goodness of fit and generalization ability. Editor D. Koutsoyiannis; Associate editor K. Hamed

Optimization methodology for a river temperature monitoring network for the characterization of fish thermal habitat

ABSTRACT A methodology for planning an optimized river water temperature monitoring network is presented. The methodology is based on sampling of the physio-climatic variability of the region to be monitored. Physio-climatic metrics are selected to describe the study region, based on principal component analysis. The sites to be monitored are then identified based on a k-means clustering in the multidimensional space defined by the selected metrics. The methodology is validated on an existing dense water temperature network in Haute-Savoie, France. Different configurations of more or less dense network scenarios are evaluated by assessing their ability to estimate water temperature indices at ungauged locations. An optimized network containing 83 sites is found to provide satisfactory estimations for seven ecologically and biologically meaningful thermal indices defined to characterize brown trout thermal habitat.