Study of Hydrological Processes for Five Catchments in Puna and Montane Forest Ecosystems in Bolivia Using Data Mining Techniques

Thursday, 9 June 2016
Mauricio Florencio Villazon1, Lina Gabriela Terrazas1,2, Ana Cecilia Escalera1, Pablo Pardo1 and Patrick Willems2, (1)San Simon University, Civil Engineering, Cochabamba, Bolivia, (2)Katholieke Universiteit Leuven, Leuven, Belgium
Abstract:
In this work the hydrological processes of five catchments in Bolivia, two in Puna ecosystem and three catchments in a montane forest area were simulated with model structures identified and calibrated with the VHM approach (Willems, 2014; Willems et al., 2014). The possibility to modify the prefixed structure allowed applying extensions and the use of multi criteria evaluation helped to understand the effect in model performance for different Andean ecosystems.

The two catchments in Puna ecosystem are located in a range of 4400 to 4050 m.a.s.l. and have two years of continuous hourly data of discharge and precipitation, the monitoring activity is done under the IMHEA (Initiative of Hydrological Monitoring of Andean Ecosystems); and the other tree catchments 2900 to 600 m.a.s.l. have 14 years of hourly data series and have been monitoring by the SEARPI. Due to problems of infiltration and siltation of the gauged structure (inline weir) identified during the calibration process, one catchment in Puna ecosystem was discarded.

The most parsimonious calibrated structures could represent cumulative flow and low flows in a good way during the calibration. The extension to overland flow routing included modification of this sub model parameters and application of the kinematic wave approximation, the results were improvement in NSE (Nash & Sutcliffe Efficiency) and peak flow representation. The optimum NSE was 0.64, 0.69 and 0.76 for the montane forest ecosystem and 0.78 for Puna ecosystem. The comparison in parameter estimation let us conclude about the hydrological characteristics of the different ecosystems.

Willems, P., 2014. Parsimonious rainfall–runoff model construction supported by time series processing and validation of hydrological extremes – Part 1: Step-wise model-structure identification and calibration approach. J. Hydrol. 510, 578–590. doi:10.1016/j.jhydrol.2014.01.017

Willems, P., Mora, D., Vansteenkiste, T., Taye, M.T., Van Steenbergen, N., 2014. Parsimonious rainfall-runoff model construction supported by time series processing and validation of hydrological extremes – Part 2: Intercomparison of models and calibration approaches. J. Hydrol. 510, 591–609. doi:10.1016/j.jhydrol.2014.01.028.