Evaluating flood inundation impact on wetland vegetation FPAR of the Macquarie Marshes, Australia

Abstract

Vegetation is a key measure of changes in wetland hydrology and ecology induced by flood inundation. The fraction of photosynthetically active radiation absorbed by the canopy (FPAR) is an indicator of presence and state of vegetation, which can be estimated from remote sensing data. This study investigates the spatio-temporal variations of FPAR under inundation by integrating remotely sensed images and hydrological data in a GIS environment. Taking the Macquarie Marshes as a study area, it aims to model the behavior of vegetation after an ecologically significant flood using Moderate Resolution Imaging Spectroradiometer (MODIS) FPAR products, so as to reveal the impacts of flood on wetland vegetation. Flow data from four gauging stations at upstream, mid-stream and downstream were analysed. The modified Normalized Difference Water Index (mNDWI) was used to extract inundation extent from MODIS images. Image ratio and lagged cross-correlation methods were employed to disclose the vegetation response to flooding. Vegetation in the Marshes went through considerable structural changes during the identified flood event. FPAR was relatively more stable in the inundated areas than the non-inundated areas. FPAR values in both inundated and non-inundated areas exhibited a similar trend in the “non-flood period”, but a different trend in the “flood period”. Lagged cross-correlation analysis shows diverse vegetation responses to flood inundation between the inundated and non-inundated areas. It proves the water retention of wetland inundation and confirms vegetation in the non-inundated areas more sensitive to soil moisture conditions. This study proposes a feasible method for efficiently studying vegetation dynamics of flood inundation at both spatial and temporal scales. These methodology and results can provide baseline information for eco-hydrological studies of floodplain-wetland ecosystems.