Search

This study shows how modern diatom assemblages, obtained from the surface sediments of 55 deep and shallow lakes in southern Quebec and representing a gradient in trophic state and specific conductivity, can be used to reconstruct the historical water column concentrations of total phosphorus (TP) over time. Inference models were developed with the aid of regression analyses based on weighted means. The TP models demonstrate statistical performances equal to or superior to TP models developed in the past for other regions in Canada and elsewhere in the world. A reconstruction of TP values, based on fossil diatom assemblages of Lac Nairne (Charlevoix, QC), was conducted. Comparing the TP inferred values with the measured values and those estimated using export coefficients, the reliability of the model was proven. In addition, the comparison of the results obtained from reconstructions performed with two inference models (Quebec and Ontario lakes) also revealed that models predicted TP and not a collinear variable. Resume : Cet article vise amontrer que les assemblages de diatomees modernes repertoriees dans les sediments de surface de 55 lacs du Quebec meridional, representatifs d'un gradient d'etats trophiques, permettent de reconstituer les teneurs histo- riques en phosphore total (PT) qui prevalaient dans la colonne d'eau des lacs du Quebec. Pour ce faire, un modele d'inference a ete developpe al'aide d'analyses de regressions basees sur les moyennes ponderees. Les performances statistiques du modele de PT sont equivalentes ou superieures aux modeles developpes pour des regions differentes au Canada ou ailleurs dans le monde. Une reconstitution des valeurs de PT, basee sur les assemblages de diatomees fossiles du lac Nairne (Charlevoix, QC) a ete menee. La comparaison des valeurs de PT inferees avec les valeurs mesurees ainsi que celles estimees al'aide des coefficients d'exportation a montre la justesse du modele. De plus, la comparaison des resultats provenant de reconstitutions menees avec deux modeles d'inference (lacs du Quebec et de l'Ontario) a egalement montre que les modeles predisent le PT et non une variable colineaire.

The effect of 33 environmental variables on the distribution of chironomids was studied in 60 lakes of northwestern Quebec (Canada). A detrended canonical correspondence analysis detected linearity in the chironomid assemblages, thus redundancy analysis was used to identify the variables affecting the chironomid communities. Lake depth, dissolved organic carbon (DOC), mean August air temperature, and surface water temperature were the four variables that best explained the distribution of chironomids. Partial least squares analysis was used to develop new inference models. Among models, the one for mean August air temperature had the highest coefficient of determination (r[Formula: see text] = 0.67) and the lowest root mean square error of prediction (= 1.17 °C). The results indicated that for downcore temperature reconstructions, it might be hard to dissociate the combined effects of temperature, DOC, and depth. Changes in taxa such as Heterotrissocladius brundini-type, Heterotrissocladius subpilosus-type, Heterotrissocladius grimshawi-type, Micropsectra radialis-type, Tanytarsus lugens-type, and Microtendipes can be attributed to changes in lake depth and (or) temperature. Changes in Heterotanytarsus, Dicrotendipes, Cryptotendipes, and Cryptochironomus might be attributed to shifts in temperature and DOC. Relationships among temperature, DOC, and lake depth should be studied in a "neo-ecology" design to better understand their impact on chironomid assemblage composition.

This paper presents a method designed to build species-tailored diatom–environment models. Using a pruning algorithm of artificial neural networks, powerful species-tailored models constrained to water temperature, water depth, and dissolved organic carbon were developed from a 109-lake training set from northwestern Canada and Alaska. The reasoning behind the approach is that the implementation of a single, gradient-based, organism–environment relationship should only use species that are comprehensively influenced by the variable of interest. By pruning species according to their relevance to each of the three studied variables, the cross-validated performances of all three models were significantly increased, suggesting that nonrelevant species have corrupting influences and need to be removed. The removal of corrupting species also suggests that palaeolimnological transfer functions based on an appropriate subset of useful species are more independent.

With depleted ozone levels and the possibility that climate change might lower dissolved organic matter (DOM) concentrations, the risk of exposure of fresh waters in Canada to ultraviolet B (UV-B) was evaluated. First, the distribution of a UV-B-sensitive zooplankton genus, Daphnia, was examined as a function of DOM concentration and maximum depth (Zmax) in 258 systems. Distribution was not restricted by UV-B, although very clear, shallow systems are underrepresented. Secondly, the depth at which 1% of surface radiance at 320 nm occurs (Z320,1%) was compared with Zmax in over 1000 aquatic systems in 15 ecozones to determine the proportions of optically clear systems (Zmax ≤ Z320,1%) and systems that may become clear (i.e., are "at risk") should DOM decrease by 50%. South of the treeline

The distribution of modern diatom assemblages in surficial sediments of 64 Labrador (Canada) lakes across broad vegetational biomes was studied in order to explore species–environment relationships and to develop transfer functions for paleoenvironmental reconstruction. The study sites were situated along a latitudinal gradient (51°27' to 57°37' N) and classified according to six catchment vegetation types: wetland (peatland) forest, spruce/fir forest, lichen woodland, forest–tundra, coastal tundra, and tundra. Canonical correspondence analysis revealed that among 28 environmental variables determined for each site, water colour and alkalinity accounted for most of the variance in the diatom data. Using weighted-averaging partial least squares techniques, we developed transfer functions for inference of water colour (CLR) (r2jack= 0.85, root mean square error of prediction (RMSEP) = 0.18log(CLR + 1) or 1.51 Pt units) and alkalinity (ALK) (r2jack= 0.63, RMSEP = 0.25log(ALK + 1) or 1.78 µeq·L–1) from the percent abundance of the 132 most abundant diatom taxa. By determining diatom distribution in relation to more detailed vegetation types within the boreal forest zone (wetland forest, spruce/fir forest, and lichen woodland), this calibration data set demonstrated the potential of these assemblages for revealing more subtle changes in lake catchment vegetation over time.

Des donnees sur les caracteristiques limnologiques et chimiques obtenues pour 59 lacs dans le Yukon et la partie adjacente des Territoires du Nord-Ouest (Canada) ont ete interpretees a l'aide de l'analyse de regression lineaire et de l'analyse des composantes principales. Les sites d'etude sont des lacs qui representent une grande variete de regions ecoclimatiques, echelonnees sur de longs gradients de latitude (60°37'-69°35'N) et d'altitude (15 a 1387 m au-dessus du niveau de la mer). Les echantillons d'eau preleves dans chacun des lacs ont ete analyses pour determiner la concentration des ions traces, des metaux traces, des elements nutritifs et de la chlorophylle a. La plupart des lacs etaient pauvres en solutes (conductivite moyenne de 160 μS.cm -1 ) et leur pH variait de legerement acide a alcalin (etendue de pH = 5,9 a 9,3). La composition ionique a varie de Ca-Cl-Na pres de l'ocean Arctique a Ca-HCO 3 vers l'interieur des terres, ce qui reflete des differences dans les bassins hydrographiques locaux et la proximite de la mer. De maniere generale, les sites arctiques et alpins ont presente de nombreuses similitudes, mais des differences considerables en ce qui a trait a la chimie de l'eau ont ete observees entre les sites situes a l'interieur du Territoire du Yukon. Ces differences peuvent etre liees aux differences touchant les caracteristiques du substrat rocheux et de la vegetation du bassin versant.

Data on the water chemistry and limnology of 24 lakes located between Yellowknife (62 8279N, 1148219W) and Contwoyto Lake (658309N, 1108009W) in the central Northwest Territories were examined using principal components analysis and other statistical techniques. The study sites were mostly shallow ( Zmax = 2.5-25 m; mean = 8.2 m), nutrient-poor (3.4- 12.7 mg total phosphorus/L; mean = 6.6 mg · L -1 ), low-alkalinity lakes. Water was typically low in solutes (specific conductance near 0-100 mS · cm -1 ), with slightly acidic to alkaline pH (6.2-8.9). Levels of all nutrients and major ions showed identical trends of decreasing concentrations with increasing latitude, with the highest concentrations generally in lakes with conifer-forested catchments in the south. Resume : Les donnees sur les caracteristiques chimiques et limnologiques de 24 lacs situes entre Yellowknife (62 8279N, 1148219W) et le lac Contwoyto (658309N, 1108009W) dans le centre des Territoires du Nord-Ouest (T.N.-O.) ont ete examinees a l'aide de l'analyse des composantes principales. Les sites d'etude etaient pour la plupart des lacs peu profonds ( Zmax = 2,5 a 25 m; moyenne = 8,2 m), pauvres en elements nutritifs (3,4 a 12,7 mg · L -1 en phosphore totale, moyenne = 6,6 mg · L -1 ) et de faible alcalinite. De maniere typique, l'eau etait pauvre en solutes (conductivite electrique pres de 0 a 100 mS · cm -1 ), et son pH variait de legerement acide a alcalin (6,2 a 8,9). Les concentrations de tous les elements nutritifs et des ions majeurs ont montre des tendances identiques, c.-a-d. qu'ils diminuaient avec l'augmentation de la latitude, les concentrations les plus elevees se trouvant generalement dans les lacs meridionaux avec forets de coniferes dans le bassin versant. (Traduit par la Redaction)