Quantification of hydrological alterations of the Parana River (The Plata basin) downriver of dams

ALBA PUIG; ALBA PUIG; HÉCTOR F. OLGUÍN SALINAS; HÉCTOR F. OLGUÍN SALINAS; JUAN BORÚS; JUAN BORÚS

Niteroi, Rio de Janeiro

Simposio; 7th HYBAM Scientific Meeting "Large tropical rivers under climate and land use changes. Evolution of hydrological, sedimentary and biogeochemical cycles in the critical zone"; 2017

Institut de Recherche pour le Développement (IRD, Francia), la Universidad Federal Fluminense (UFF) y Servicio Geológico de Brasil (CPRM)

Currently, the non-stationarity in hydrological series is recognized in many rivers and there is an increasing interest in differentiate effects of climate variability and anthropogenic effects, especially by dams. These analyses are difficult to perform in large river systems such as the Parana, which shows hydrological variability at multiple scales, anthropogenic effects (such as deforestation) and a large number of dams in its upper basin (being difficult to have little altered hydrologic series that may be comparable). This river originates in the tropical zone of Brazil and finally drains, after an extensive fluvio-deltaic complex, into a binational estuary, representing a relevant cross-border basin. This is the hydrological region in Brazil showing more dams (similar proportion of reservoirs and run-of-the-river dams), accumulating a considerable potential of downstream hydrological alteration. In this work we attempt to progress in distinguish between effects of all dams and those associated to climate variability or land-use changes, by analyzing daily flows of this large river in the zone of the last dam (Yacyreta, Argentina-Paraguay). We have obtained data of "naturalized" flows (supplied by ONS, Brazil; eliminating effects of dam regulation, reservoir evaporation and consumptive uses) of distal dams in Brazil: Itaipu (on the Parana R.) and Salto Caxias (on the Iguazu R.), as well as regulated flows at Yacyreta and the estimated contribution of the river reach from Itaipu until this dam (provided by INA, Argentina). Changes in the hydrological regime due mainly to dams were quantified by indicators (IHA) comparing (with non-parametric approach) "naturalized" values of incoming flows at Yacyreta with its regulated effluents. Changes due to climate variability or land-use changes were quantified by the IHA in the "naturalized" flow series. The presence of significant change-points was analyzed using a non-parametric multivariate method applied to basic IHA. Multivariate canonical analyses contributed to distinguish changes in the IHA by regulation and by climatic periods. In the "naturalized" series a first change-point was detected in the early 70´s, separating the historical period of the wet one, and a second point, around the beginning of the century, separating the wet from the dry periods. Then, these abrupt changes (coincident with those detected in the incoming flows to the Delta in a previous work) were independent of dam effects, being probably due to climate variability. In the "naturalized" series, the flow duration curves as well as the monthly medians along an average hydrological year showed rather parallelism between periods. In contrast, the regulation strongly decreased the slope of the flow duration curve and the monthly seasonality, thus modifying the hydrological functioning. The regulation also caused increase in minimum and base flows, number of reversals (rise/fall), rise rate, variability in date of minimum flow, as well as decrease in fall rate. These changes imply a "homogenization" pattern at monthly scale, mainly due to reservoirs, but an increase in daily variability, mainly due to run-of-the-river dams. The effect of the regulation in the last period (dry) explained a higher variation in the basic IHA that the difference between opposite climatic periods (wet and dry) of "naturalized" series. These results highlight the convenience of analyze the cumulative effect of all dams and the application of IHA indicators and different multivariate methods.