| AFFILIATIONS | - 1) Observatori De L Ebre (uni. Ramon Llull - Csic)), Roquetes (Spain) ,2) Estación Experimental Aula Dei (eead-csic)), Zaragoza (Spain) ,3) Departamento De Geología, Facultad De Ciencia Y Tecnología, Upv/ehu, Leioa (Spain)
|
| ABSTRACT | The Pyrenees, located at the isthmus that connects the Iberian Peninsula with the European continent, presents large climatic diversity, ranging from Atlantic to Mediterranean climates, with high environmental value and being particularly sensitive to global climate change. In parallel, this mountain range is the water tower for more than 15 million inhabitants, also providing this resource for industry, agriculture and ecosystems. Therefore, studying and understanding its evolution is crucial for its management. It is especially important to assess probable changes in the water cycle, particularly in surface flows, in the context of climate change. However, in addition to climatic considerations, non-climatic factors such as land use can have a significant impact on river flows. Gauging station streamflow records are the primary method for defining the hydrological regime of watercourses in this regard. These data can also be used to analyze possible hydrological changes over time, as well as trends in water resources availability, when they cover a sufficient time period and especially when they are supported with complementary tools like hydrological modeling. Hydrological modeling helps to understand the underlying processes by simulating variables that are difficult or impossible to observe (e.g. soil moisture, snowpack, or land evaporation) and by performing experiments impossible to conduct in the real world (e.g. fixing land use to assess the impacts of climate change only). However, all those valuable contributions are subjected to model uncertainty, an issue that should not be neglected and should be carefully assessed. We studied the historical evolution (1980–2013) of the natural river streamflows of the Pyrenees using observation values from non-influenced gauging stations and two hydrological model results (the fully distributed model SASER and the semi-distributed model SWAT). The comparison of observational data with models allows us to detect, evaluate, and analyze changes in flow rates, their trends, and attribution, as well as the main sources of uncertainty. The performance of the simulated time series from SASER and SWAT is evaluated using the non parametric Kling-Gupta efficiency test. After that, we computed monthly and seasonal statistics for two time periods (1980–2013 and 1990–2013) and computed the trends of the different statistics using Sen s slope estimator. The significance of the trends was estimated with the Mann-Kendall test on the pre-whitened time series, with the statistical significance tested at the 95% level. Finally, these trends have been analyzed using binary classification techniques and a visualization tool based on the contingency table. In most cases, no statistically significant trends have been observed, but when they do exist, we can attribute the trend, depending on the results of the models and observations, to climatic changes or changes in land use. Another thing that the results show is the acceleration of the effect of climate change on streamflows, as this trend is more clear in the shorter period than in the longer. Finally, applying two different models, we obtained different results. Thus, we conclude that the uncertainty is large and confirm that it was a good choice to use different models. This work is a contribution to the EFA210/16 PIRAGUA project. |