PubAg

Main content area

Systematic increase in model complexity helps to identify dominant streamflow mechanisms in two small forested basins

Author:
David, Paula C., Oliveira, Debora Y., Grison, Fernando, Kobiyama, Masato, Chaffe, Pedro L. B.
Source:
Hydrological sciences journal 2019 v.64 no.4 pp. 455-472
ISSN:
2150-3435
Subject:
basins, hydrologic models, rain, river flow, runoff, soil, stream flow, water storage, watersheds
Abstract:
This study shows how the use of increasing model complexity allows us to hypothesize about dominant streamflow mechanisms in two small Brazilian forested basins. Nine different structures from SUPERFLEX, an objective framework to systematically increase hydrological model complexity, were tested and we extended the flexible modelling methodology to error models as well. We show that applying a rigorous methodology in a model evaluation framework, with residual analysis and control of model complexity, is essential for testing a model as a hypothesis for dominant hydrological controls. Our results indicate that the model architecture was more important than the increase in the number of model parameters. Better performing models were those with a parallel structure, which confirms our a priori belief about the dominant runoff mechanisms of the studied catchments, characterized by a rapid response to rainfall, but also a constant river discharge fed by water storage on the thick soil layer.
Agid:
6369691