Spatio-temporal effects of inundation and climate on vegetation greenness dynamics in dryland floodplains

Monitoring floodplain vegetation response to water availability is essential for sustainable water resource management. However, the interpretation of these responses through space and time has been challenging due to the limitation of satellite-derived inundation extent, which is restricted by the return period of imagery and prevalence of clouds affected images during high flow events. We address this issue by utilising a predictive inundation model capable of reporting inundation at daily temporal resolution. Spatio-temporal patterns in vegetation greenness of four floodplain vegetation communities (Eucalyptus camaldulensis, Eucalyptus largiflorens, Eucalyptus coolabah and Duma florulenta) were quantified using Landsat-derived vegetation index and a combination of time series of floodplain inundation pattern, rainfall and soil moisture for the years 1989-2016. The results show that vegetation dynamics were highly variable between years and are closely associated with irregular rainfall patterns and overbank flooding. Linear mixed-effect models explained between 38% and 67% of the variability in vegetation productivity, with all involved variables being significant predictors (p < 0.05). The effects of rainfall, soil moisture and inundation were diverse and depended on vegetation type. Higher inundation frequency and low interflood period often corresponded to lower vegetation responses, which may relate to a mixed surface water and vegetation signal, or a short-term inhibitive effect of surface water on growth. These results highlight the importance of groundwater and lagged effects of rainfall and flow in explaining patterns of vegetation condition in semi-arid floodplains and should be considered in the design of environmental water monitoring programmes.