Linking spatial inundation indicators and hydrological modelling to improve assessment of inundation extent

MODSIM2013, 20th International Congress on Modelling and SimulationEnvironmental water requirements (EWRs), including past inundation regimes, provide useful information for the conservation and management of wetlands systems. However, determining past inundation regimes is difficult as very few wetlands have historical records of water level or flow to determine EWRs. Generally, two approaches have been used to reconstruct wetland inundation regimes; remote sensing or hydrological modelling. Remote sensing methods have applied spectral indices of inundation to multi-temporal archival satellite imagery and developed simple empirical models to link inundation extent to rainfall or stream flow records. Hydrological models use a water balance approach and simple relationships between volume and area based on the wetland bathymetry to derive an inundation regime. While remote sensing approaches can provide a high spatial resolution of information, the temporal resolution is limited to an image every few weeks, and the usability of each image is subject to cloud cover. Hydrological models require some data in the region of interest for calibration, and often the model is not sensitive to the inundation variables of interest when calibrating to the data available. A study is presented where this is the case, where good results are achieved by a hydrological model when compared to downstream flow data, but very different wetland inundation regimes can be simulated. Due to this limitation, the remotely-sensed inundation regime was required as another source of information to calibrate the hydrological model, constraining the parameters involved in the modelled wetland storages. The remote-sensing constrained hydrological model provided greater confidence in the representation of processes occurring in the simulated wetland storages, and hence could be considered appropriate to be applied to a validation time period and for scenario testing. The findings of this research support the combination of remotely sensed indicators of inundation and hydrological modelling to increase confidence in the spatial representation of the dynamics of inundation. Understanding wetland inundation dynamics is critical to the understanding and management of threatened wetland ecosystems. Hence, the methods outlined in paper will ultimately assist in informing better management of wetlands.Matthew Scott Gibbs, Kenneth Clarke, Ben Taylo