Detecting floodplain inundation frequency using MODIS time-series imagery

Floodplains play an important role in riverine hydrological and ecological environments. Accurate estimate of the spatial and temporal inundation frequency patterns is an critical step to understand how the inundation conditions affect local hydrology and floodplain ecosystems. This study presents a methodology to detect spatial-temporal changes in inundation extent associated with flood inundation frequency based on MODIS 8-day composite time-series imagery (MOD09A1) from 2000 to 2010. It takes Chowilla floodplain located in the northwestern Murray-Darling Basin of Australia as a study area. The GIS-based framework includes extraction of flow peaks from time-series of observed flow data, inundation detection from MODIS using the Open Water Likelihood index, inundation validation using modified Normalized Difference Water Index on Landsat TM image, and flood extent and frequency mapping. The study shows that the maximum inundated areas that have been inundated at least once are 12.5% of the total study area. The areas with different inundated frequencies were calculated and linked with ecological significant flood return periods. The areas that were inundated twice in the 11 years, which is equivalent to a 1-in-5 Average Recurrence Interval, occupies 27.8% of total inundated area. The results of this study provide an important indicator for the health condition and spatial distribution of some dominant species in the floodplain, such as river red gum. This, in turn, improves knowledge of eco-hydrology characteristics of floodplain ecosystem.