Search

La télédétection satellitaire est un outil employé couramment et avec succès en océanographie. Il n'en va pas de même en limnologie, où les applications sont encore rares.Par le moyen d'une revue bibliographique, nous tentons d'en analyser les raisons. Après une brève description de l'outil et des satellites en service, l'on met en évidence la spécificité des cibles aquatiques, caractérisées par une réflectance basse et une profondeur d'investigation variable. Ces particularités, jointes à la composition complexe des eaux intérieures, rendent impossible l'extension pure et simple, à la limnologie, des algorithmes développés en océanographie.Néanmoins, nous montrons que la télédétection a été utilisée dans l'étude du bassin versant des lacs, ainsi que pour la cartographie de leurs limites, de la végétation aquatique, des courants, de la thermique et de la couleur de l'eau. Des modèles empiriques, exprimant la matière en suspension ou les paramètres de qualité de l'eau, ont été calculés et appliqués avec succès dans certains lacs.On établit ensuite une typologie des difficultés rencontrées dans l'application de la télédétection à la limnologie : intrinsèques (complexité de la composition), technologiques (capteurs actuels non adaptés aux cibles aquatiques) et institutionnels (coûts élevés et manque de professionnels de ta télédétection dans les cercles limnologiques).Finalement, l'on présente quelques propositions pratiques dans la perspective des nouveaux véhicules spatiaux et capteurs des années 90, qui devraient permettre une exploitation de l'énorme potentiel de la télédétection en limnologie., Remote sensing has been used successfully in oceanography for many years, whereas applications in limnology have been comparatively modest. We attempt to discover why from a review of the literature.After a brief description of remote sensing and of satellites in operation, we stress the specificity of inland waters targets compared with "solid" targets : low reflectivity in the visible and near infrared, complex water composition (chlorophyll, mineral suspensoids and gelbstoff which are not covariant) and the variable depth of investigation. Hence, simple chlorophyll and suspended solid retrieval algorithms, developed for type-I waters, cannot be applied to inland waters.Nonetheless, remote sensing may be, and has been applied effectively to the study and management of drainage basins and to the mapping of lake limits, aquatic vegetation, water masses, currents, thermal structures and water colour. Site-specific, regression-type models have been computed to express suspended solid concentration and water quality parameters as functions of radiante. So far, most of the applications have been performed on large lakes (Laurentian Great Lakes, Tahoe, Chad, Biwa, Balaton, Léman), where not only LANDSAT but also CZCS and AVHRR data have been in operation.A typology of present difficulties in applying remote sensing to limnology is proposed : intrinsic (complexity of inland waters), technological (land resource satellite bands not designed for water sensing, revisit time too long) and institutional difficulties (cost of data and equipment, and scarcity of limnologists proficient in remote sensing techniques).

The Coastal Zone Color Scanner (CZCS)-retrieved chlorophyll-like pigments for the Mediterranean Sea show zones of high concentrations (10mgm-3) such as the shallow Kerkennah Shelf, off the southern Tunisian coast. This occurrence pointing to probable erroneous estimates, exaggerated by the effect of the bottom signal is discussed in the general framework of the applicability of water colour techniques to coastal areas, as a source of environmental data suitable for inclusion in a multi-purpose Geographic Information System (GIS). The state of knowledge on the Kerkennah Shelf test site, organized into a regional geo-referenced data base (including CZCS and TM imagery), is presented. A two-flow model has been applied to in situ reflectance measurements, to derive some information on the interplay between water colour and bottom typology. The results have been used to evaluate nature, intensity and extent of the sedimentary and vegetated sea bottom effect on the chlorophyll-like pigment concentrations derived from CZCS in the Kerkennah area. Finally, the applicability of satellite-derived ocean colour data to coastal zone mapping is discussed in some detail.

[1] In this study we prove the feasibility of the advanced very high resolution radiometer (AVHRR) and Moderate Resolution Imaging Spectroradiometer sea surface temperature algorithms to derive operational lake surface water temperature (LSWT). A validation study covering 2 years was done using data from the AVHRR on NOAA 12, 15, 16, and 17, the MODIS on TERRA, and AQUA and with different method-ingested in situ data from different sized lakes. Best results were found for NOAA 16 nighttime data at Lake Geneva (bias of 0.18 K and standard deviation of 0.73 K) and TERRA nighttime data at Lake Constance (satellite-buoy bias of −0.08 K and standard deviation of 0.92 K). For all sensor families an overall scatter ranging from 0.9 to 1.6 K was found. Bias of MODIS is larger, −1.73 to 1.9 K, than the one of the AVHRR (−0.28 to 1.5 K). The current orbital configuration of the platforms used revealed the diurnal evolution of the lake surface temperature amplitude from space. The damped mixing found for a typical calm and clear-sky regime is different from open ocean conditions. As the main error source, we found undetected cloudy pixel. Furthermore, the physical difference between skin and bulk temperature, especially its relation to the diurnal thermocline, solar insolation, and wind stress contributes to the bias and scatter within the match-up data set. The data sets have been validated to allow further application for LSWT climatology and assimilation into numerical weather prediction models.

Bio-optical measurements were carried out on eutrophic Lake Zug and oligotrophic Lake Lucerne, to provide data and models for the in situ calibration of multispectral imagery collected in 1991, during an AVIRIS flight over Central Switzerland. The results indicate that Secchi depth, chlorophyll and total suspended solid concentrations, vertical extinction coefficient, absorption coefficient and irradiance reflectance in the PAR region can be used to discriminate between the two lakes. Dissolved organics concentration, scattering and total attenuation coefficients estimates, as well as backscattering probability, are less acceptable, because of instrumental limitations. Relationships between optical and water quality parameters were investigated and found to behave according to accepted theoretical frameworks. Both lakes exhibited vertical and horizontal heterogeneities in chlorophyll and turbidity, and had contrasting mean bio-optical characteristics: although suspended solid concentrations were similar, transparency was lower in Lake Zug, but reflectance was substantially higher in Lake Lucerne. Water colour determined by chromaticity analysis was blue-green in Lake Lucerne and green in Lake Zug. Reflectance spectra simulated through a three-component optochemical model did not completely match those derived from in situ measurements, because of lack of site-specific optical cross-sections for suspended minerals and dissolved organics. The monitoring of water quality in subalpine lakes by remote sensing, both with present and future technology, is discussed and considered as possible, provided that their optical behaviour is known, and preferably expressed by a valid optochemical model.

The current level of knowledge concerning Lake Erie is inadequate for the development of a predictive model of the system. Nevertheless, an outline of the submodels required for the development of a predictive model is presented. Further, as a result of many recent studies, the level of understanding of the Lake Erie system has increased immensely and major gaps in the present knowledge of the lake have been identified; useful areas for future research have been indicated. Many of the recent studies on Lake Erie are included in this special issue. The results of these investigations have been summarized and linked together where possible, to describe the lake and give an overview of the more important processes in Lake Erie.

Sediments were collected on a grid from Lake St. Clair in 1970 and 1974 and from Lake Erie in 1971 and analysed for organochlorine insecticides and PCBs. Suspended solids were centrifuged from pumped water in the Detroit River in 1974. Residues of DDE, TDE and DDT were highest in sediment from the Western Basin (70.3 ppb) Lake Erie and lowest in sediment from Lake St. Clair (6.6 ppb in 1970 and 2.6 ppb in 1974). The Central and Eastern Basins of Lake Erie contained residues that fell about mid-way between these means. TDE was the predominant component of ΣDDT and was followed by DDE at approximately half this residue. While DDT represented the lowest component in lake sediments, it was the highest component in suspended solids from the Detroit River. HEOD was present in the entire lake system but at only a fraction of the level of ΣDDT. The mean residue in Lake St. Clair was only 0.1 ppb while Lake Erie sediments contained 1–2 ppb. The highest mean residue (3.6 ppb) was in suspended solids in the Detroit River. Other organochlorine insecticides were only occasionally identified. PCB residues in sediment were approximately 3 times higher than the total organochlorine insecticide residue, being highest in the Western Basin (252 ppb) and lowest in Lake St. Clair (19 ppb in 1970 and 10 ppb in 1974). The Detroit River and Central and Eastern Basins of Lake Erie had intermediate residues. Residues in Lake St. Clair indicated a decline in ΣDDT, HEOD, and PCB residues between 1970 and 1974, and this coincided with the restrictions on the materials. A core sample from the Western Basin of Lake Erie indicated that ΣDDT, HEOD and PCB began to accumulate in sediments in 1958, 1953, and 1956, respectively. This appeared to indicate a lag period of 2–5 years after wide-spread use occurred.

A one-layer, two-dimensional computer model for the time-dependent, lakewide advective and diffusive transports as well as physical regeneration of total phosphorus in Lake Erie has been developed. It has 60 × 16 square grid cells of 6.67-km mesh size and uses a time step of 12 h. The model has been verified and found to be in reasonably good agreement with the Canada Centre for Inland Waters (CCIW) monitor cruise data for October, November, and December during which period there were high winds with high phosphorus concentrations being observed. Computed, daily averaged currents from a hydrodynamic model developed at CCIW are used for the transport terms and actual data for the lake boundary conditions. The physical regeneration is attributed to wave motions induced by wind. A formula is proposed which expresses the regenerated amount of total phosphorus in terms of wave orbital velocity and sediment mean grain size. A discussion is presented on the choice of the settling rate and the regeneration coefficient which produce satisfactory model results.

Geochemical samples from part of Lake Geneva were analyzed for 29oxides and trace elements. The variables and samples were subjected to R- and Q-mode analyses. The following techniques were applied in sequence: data transformation (normalization and standardization), data reduction (principal component and factor analysis), and automatic classification (dendrograph). The data were treated using various combinations of these techniques, and the resulting classifications evaluated by means of several criteria. The best classification of the samples is given by a cluster analysis performed on four principal components computed from standardized variables. The discriminatory power of the variables also was measured and determined to depend on their degree of intercorrelation. As a final result, the 29original variables were reduced to four components and the sediment samples classified into four facies, leading to easily interpretable geochemical maps.

In 259 samples of the topmost 3 cm of sediment taken from Lake Erie in 1971, total mercury values ranged from 8 to 2929 ppb [Formula: see text]. The highest levels were in the Western basin adjacent to the Detroit River, the rest of the lake showing increasing concentration from the coarser sediments of the Inshore Zone and cross-lake moraine sills out into the basin muds. A quartz correction was applied to compensate for texture and dilution by an inert constituent. The distribution of quartz-corrected mercury confirmed that the Detroit River is the predominant source of industrial mercury to the Western basin with subsequent transport across the Central basin to final sink in the Eastern basin. Further possible inputs were indicated from Erie, Pennsylvania and Buffalo, New York. Analysis of suspended sediment in the Detroit River showed no change in concentration between 1970 and 1974 with a mean 1974 level of 1060 ppb. Mercury loading was estimated at 7.2 m.t./yr to the sediments of the Western basin. A relationship was observed between total mercury and organic carbon, showing two distinct regressions: ’A’ with lower organic-bound mercury, is taken to represent the levels related to industrial and natural loadings existing prior to the introduction of mercury cell chlor-alkali plants; ’B’ applicable to highly contaminated sediments, is believed to be indicative of loadings derived from the chlor-alkali industry.

The phosphorus in 48 surficial Lake Erie sediment samples was present in three major forms: phosphorus associated with apatite, nonapatite inorganic phosphorus (NAIP), and organic phosphorus. The apatite was of natural, detrital origin. It existed as particles ranging from fine sand to clay in size but mostly as silt-sized particles and was concentrated in nearshore sediments. Both NAIP and organic phosphorus was concentrated in fine-grained sediments accumulating in offshore depositional areas. NAIP was associated with amorphous hydrated ferric oxide in the oxidized microzone but was present as vivianite (Fe3(PO4)2∙8H2O) and possibly other forms also in the reduced zone. The organic phosphorus content of the sediment was closely related to organic carbon content. The phyllosilicate, organic matter, and reactive iron and manganese components of the sediments existed in intimate association.

A budget for phosphorus retention in Lake Erie during 1970 has been drawn up. Phosphorus associated with sand-sized particles (> 63 μm) was excluded. Retention of apatite phosphorus (AP) and nonapatite phosphorus (NAP) within each of the three basins of the lake (Western, Central, and Eastern) was estimated in each case by two independent methods. The two estimates of retention of total phosphorus within the entire lake agreed to within 9.5% of their mean. The estimates of AP retention within the Western basin also agreed well. However, the input–output balance estimates of AP retention in the Central basin and of NAP retention in the Western and Central basins greatly exceeded the corresponding sedimentation values, while in the Eastern basin this pattern was reversed. The AP discrepancy was attributed to unobserved transport of AP from the Central to the Eastern basin via nearshore currents past Long Point, while the NAP discrepancies were largely attributed to eastward movement of NAP through the lake within about 2 m of the lake bottom. The total amount of phosphorus retained within the lake in 1970 was 41,000 metric tons, or about 92% of input. AP of natural origin and inert nature accounted for approximately 44% of the total phosphorus retained. A reduction in anthropogenic phosphorus input of about 15,000 metric tons/yr, as called for in the Great Lakes Water Quality Agreement, will reduce present annual NAP loading to close to pre-1850 values.

On étudie l'influence des rejets de deux stations d'épuration sur la teneur des sédiments côtiers en phosphore, fer et carbone organique. Ces deux stations sont situées sur le littoral lémanique, l'une à Lausanne (Vidy), l'autre à Morges. La quantité de phosphore des sédiments situés dans la zone de rejet de Vidy excède dix fois la teneur naturelle. Les très fortes valeurs trouvées également pour le fer indiquent qu'une pattie du complexe ferro-phosphaté formé lors de l'épuration est relâchée avec les effluents. Les effets de la station de Morges sont moins marqués et ne sont pas visibles dans la distribution géographique du phosphore total dans les sédiments. En revanche, l'utilisation des rapports [fer extractible]/[phosphore non-apatite] et [carbone organique]/[phosphore organique] montre que les effluents de la station de Morges ont enrichi les sédiments côtiers en phosphore à presque deux fois la teneur naturelle. Ces rapports semblent être de meilleurs indicateurs de l'enrichissement des sédiments que les valeurs du P total.

On the basis of extensive echosounding and grab sampling, three major units have been recognized in Lake Erie: till and bedrock, glaciolacustrine clay, and postglacial muds. These units represent the late glacial and postglacial evolution of the basin and occur in an offshore younging sequence. The main basin of the lake is subdivided by residual glacial moraines into four depositional basins: Western, Sandusky, Central, and Eastern basins. The sediment texture has been defined by moment measures (mean, standard deviation, skewness, and kurtosis), the trends of which are related to the mixing of two primary grain-size populations in the sand- and clay-size ranges. A third grain-size mode in the silt size, composed of fine quartz with some carbonate, has been recognized. This mode has a modifying effect on the symmetry of the two primary populations and may, to some extent, be sufficiently abundant to behave as a discrete population. The trends in the textural characteristics, particularly skewness and kurtosis, have been utilized to define energy regimes at the sediment–water interface which indicates three distinct sedimentary or hydraulic regions: 1) Western basin region — Fine-grained sediment accretion in shallow water related to an imbalance in sediment budget, with high input loadings of fine-grained sediment, and deficit in coarse materials, with an excess of input over sediment export to the Central basin region. This results in net sediment accretion in shallow water with texture in disequilibrium with environmental energy, which produces mixing and suspension, followed by redeposition; 2) Central basin region — West to East coarsening of sediment in textural equilibrium with hydraulic energy, as it relates to increasing fetch under westerly and southwesterly prevailing winds; 3) Eastern basin region — Deepwater basin with sediments showing decreasing size offshore with increasing water depth. The deepwater sediment is modified by the influx of substantial quantities of the silt-size material derived from shoreline erosion in the north shore of the Central basin region.The interrelationships of parameters indicate textural dependence on mineralogic composition, particularly important being the relationship of clay concentration to mean grain size. This has particular value in modelling the physical behavior of clay-associated geochemical elements such as phosphorus.

The chemical and physical characteristics of six cores from Lake Erie are described. The concentrations of Si, Al, K, Na, and Mg, which represent the major mineral species in the sediment, were uniform in each core. Surface enrichments of Hg, Pb, Zn, Cd, Cu, Organic-C, N, and P were observed at each location, due mainly to the increasing anthropogenic loading of these elements to the sediments since 1850. Concentration profiles for Mn, Fe, and S were related to sediment Eh and are due to mobilization of these elements in the pore waters.Natural and anthropogenic inputs of nutrients and heavy metals were calculated. About 60% of the total loading of these elements is being deposited in the Eastern basin. It is calculated that 30 million metric tons of fine-grained sediment accumulates in the offshore basins of the lake, each year. The major source of the sediment input is erosion of the shoreline bluffs, with the north shore between Erieau and Long Point contributing 21 million metric tons annually. A sediment mass balance suggests that a large part of this bluff material is transported up to 150 km and is deposited in the Eastern basin.

Un bilan d'entrees-sorties du phosphore total pour le Leman entre 1964 et 1983 est presente. Dans l'ensemble, le lac realise l'elimination nette de Ptot a un taux moyen de 0,2 an-1, mais avec de fortes fluctuations, les annees a sedimentation nette alternant avec celles a regeneration nette. Les conditions meteorologiques jouent un role capital dans cette alternance. L'application au Leman du modele simple de Vollenweider et des normes OCDE est discutee. L'approximation du coefficient d'elimination Φ a l'aide du coefficient de renouvellement de l'eau conduit a une forte surestimation de l'elimination. Pour tenir compte de la variabilite de Φ, l'on simule l'evolution du stock selon divers scenarios en suivant une approche stochastique. Il s'avere qu'une variabilite realiste de Φ (50 %) peut provoquer des fluctuations interannuelles du stock de plusieurs milliers de tonnes. Dans les conditions de charge actuelle (1200 t an- 1), ou meme apres une reduction de 400 t an- 1 suite a une interdiction des phosphates dans les detergents, le lac ne pourra pas atteindre l'objectif d'un stock a 2 000-3 000 t.

Moment and graphic size parameters were computed on 420 surfical sediment samples from Lake Geneva (Switzerland). The sediment ranges from sand to clay-size and comprises a significant amount of precipitated silt-size calcite. Comparison of parameters shows that mean, sorting, and skewness may be estimated either by moment measures or by Folk and Ward's descriptors, without essential differences in interpretation. As for kurtosis, moment and graphic parameters provide different information, and both should be used separately as interpretive tools.

The wind speed dependence of air/water gas exchange in natural systems is controversial, since field measurements failed to verify such a dependence, whereas wind/wave tunnel data show a strong increase of the gas exchange rate with wind speed. This turns out to be no real discrepancy, since the mechanisms controlling gas exchange are not well understood yet and a simultaneous view on recent wind tunnel results shows that other parameters beside the windspeed, dependent on the water surface waves, are influencing the exchange. In this paper we present gas exchange rates obtained with the Tritium/Helium-3 method from 4 lakes. The measured transfer velocities range from 0.4 to 11 m/d. Despite the large uncertaincies both of the exchange rates and the friction velocity, a significant windspeed dependence is found. Finally we describe how to improve measurements in order to get better parametrization and present our investigation started in autumn 1982 at Lake Sempach, Switzerland. In the middle of the lake a buoy is anchored measuring all important meteorological and limnological parameters. The study is closely linked with a project of the EAWAG, during which the internal mixing of the lake is studied with the same data sets.