Le Rio Desaguadero relie le lac Titicaca (alt. 3 810 m) au lac Poopo (alt. 3 670 m) dans le bassin endoréique de l'Altiplano, en Bolivie, (fig. 1). Le bassin versant est constitué de terrains sédimentaires et volcaniques, avec une végétation steppique et aucune source de pollution organique. Le climat très variable présente d'une alternance de saisons humide et sèche. Le Rio Mauri fournit 90 % du flux de MES estimé à 6,6.106 t an-1 (d'après GUYOT et al., 1990).Des analyses de COT et d'autres paramètres chimiques ont été réalisées sur 15 stations en 3 campagnes (fig. 2 et tableau 1). Une ACP (fig. 3) révèle une corrélation entre COT et MES, correspondant à deux épisodes de crues (déc. 87 et fév. 88) (tableau 2). Des hypothèses concernant l'origine du carbone organique et les flux sont avancées (tableau 3). Un flux de base, principalement du COD provenant du lac Titicaca, représenterait 18 000 t an-1. Mais le COP, quoique fortement corrélé aux MES, n'a pas la même origine que celles-ci (fig. 4); il proviendrait principalement des zones connexes au Rio Desaguadero riches en végétation aquatique.La contribution de ces flux aux bilans du carbone organique paraît négligeable pour le lac Titicaca et peu importante pour le lac Poopo. Mais le maintien des interconnexions entre le Rio Desaguadero et ses zones humides doit être intégré dans les projets d'aménagement.
The Rio Desaguadero is the outlet of Lake Titicaca and flows into Lake Poopo in the endorheic watershed of the Altiplano in Bolivia (alt. 3 650 m); its main tributary is the Rio Mauri (fig. 1). The transport of dissolved and suspended solids has been recently evaluated by GUYOT et al. (1990). The data on Total Organic Carbon obtained for the first time in this hydrosystem allow to complement this investigation with an estimation of the organic carbon transport.The Altiplano is a basin filed with tertiary and quaternary sediments. The Eastern cordillera is constituted of primary rocks; the Rio Mauri flow from a tertiary volcanic watershed. The climate is rather cold, semi-arid (700 to 400 mm yr-1) and the vegetation is steppe-like with extensive pastures (sheep, lamas). The rainy season lasts from December to March. No direct pollution source can affect the TOC. The hydrology is very variable (see GUYOT et al., 1990). The base flow issues from Lake Titicaca and depends on the water level in the lake. The Total Suspended Solids (TSS) content is very high during floods. The annual flux of TSS is 6.6 106 t yr-1 at Ulloma, 90 % of transported sediment coming from the Rio Mauri.TOC was measured during 3 sampling series at 15 different stations (table 1). The 25 ml samples were acidified (pH 1) with H3 PO4 and sent to Lyon (France) to be analysed with a Dohrmann DC 80. In February 1988, Dissolved Organic Carbon was measured on filtered samples (Whatman GF/C decarbonated at 550°C). The major elements were analysed in La Paz (Bolivia).Raw TOC data are presented (fig. 2), with a Principal Components Analysis (PCA) for 9 parameters on all the samples (fig. 3). The first axis (52 % of total variance) is a scale of mineralization; it reveals a progressive concentration of dissolved elements in the Lake Poopo outlet (fig. 3B first cluster). The mineralization of these samples seems to be inversely correlated to the discharge flowing from Lake Titicaca for the corresponding series. The second axis (23 % of total variance) shows a correlation between the TOC and the TSS (fig. 3B, 4th cluster). This correlation is obvious for the samples of Dec. 87 and Feb. 88 (table 2), when TSS concentrations are high due to flash floods of some effluents. Equation (1) gives the hast regression of TOC vs TSS.The organic Carbon transported is constituted of dissolved (DOC) or Particulate Organic Carbon (POC). Estimation presented for DOC and POC transport at the three gauging stations (table 3) are based on the following hypotheses :In the Rio Desaguadero, DOC concentrations seam relatively stable (fig. 2D). Upstream of the Rio Mauri, the TOC transported at an average concentration of 11 mg. l-1 is assumed to be mainly DOC flowing out from Lake Titicaca. DOC transport may constitute a relatively constant base flow of organic carbon.Estimation of POC transport is much more difficult. During floods from tributaries, the rise in TOC concentrations correlated with an increase of TSS seems to be essentially due to the POC. Following a classical method (MEYBECK, 1982), a non-linear regression of percentages of POC In TSS vs TSS was calculed (equation 2), using the samples with high TSS concentrations (fig. 4A). Then, the POC fluxes were estimated using daily data of TSS for the gauging stations CA and UL. As most of the TSS come from the Rio Mauri results from this estimation that the flow rate of TOC from this stream would reach 7 t km-2yr-1. But this value according to MEYBECK (1982) is close to the export of a tropical rainforest. Thus, this hypothesis of a POC transport directly linked to the TSS flux must be rejected.POC and TSS, although correlated, have distinct origins. This is shown by the tact that the percentages of POC in TSS in the samples used to calculate equation (2) are much higher than those given by MEYBECK (1982) for world rivers (fig. 4B).A possible explanation for this phenomenon is that the POC could come from Rio Desaguadero Borders and wetlands, often rich in aquatic vegetation, and very extensive in some places (especially upstream of Nazacara and downstream of Eucaliptus). This particulate carbon might be flushed into the mainstream during rainy periods, while at the same time TSS flow down from tributaries. Hypotheses on POC transport (table 3) in the Rio Desaguadero are thus based on a POC/TSS correlation (equation (2)) valid only for the mainstream, the contribution of Rio Mauri to the total flux being estimated on the basis of e TOC export rate of 0.5 t km-2 yr-1.Contribution of these fluxes to organic carbon budgets seems to be negligible for Lake Titicaca, whose phytoplanctonic primary production is estimated at 4 106 t C yr-1 (RICHERSON et al., 1977). When considering only the labile fraction of TOC (ITTEKOT, 1988), the readily matabolizable organic carbon input in Lake Poopo hydrosystem could be evaluated to 36 000 t yr-1, corresponding to about 13 g C m-2yr-1; that might be unimportant for this presumably highly productive lake system.In conclusion, these preliminary data lead to hypotheses which consider the Rio Desaguadero as a very particular hydrosystem in organic carbon transport. If most of the particulate carbon comes from margins, the necessity for maintaining interconnexions between the main Carmel and the wetlands fn future regulation projects must be emphasized.
