Investigating the impact of the Millennium Drought on catchment water balance : A study of four catchments in Victoria, Australia

Abstract: Southeast Australia have between 1997-2009 experienced a severe drought, referred to as the Millennium Drought. During these years the region experienced a 11.4% decline in mean annual rainfall, an unprecedented decrease in runoff and a decline in soil moisture and groundwater storage. The drought officially ended in 2010 when one of the strongest La Nina-events on record occurred. However, it is still unknown how the behaviour of the catchments changed during the drought and if this change persists in the years following the drought. Changes in catchment behaviour and fluxes are commonly determined using a catchment water balance, where the change in groundwater storage is assumed to be neglectable when studying longer periods of time. However, studies have showed that this assumption might be inaccurate for catchments that experience a climatic disturbance such as a severe drought. This study investigates if including the change in groundwater storage by using spatial groundwater head data can improve the catchment water balance. This was done by assuming that specific yields are unknown and to be determined in a calibration. An unknown scalar applied to the evapotranspiration was used to try to account for the uncertainties in the known fluxes and was also to be determined in the calibration. Two different calibration schemes were considered: one assuming no delay in groundwater head response to climate and one accounting for the delay. The fluxes were determined for the period before, during and after the drought. The results were analysed to determine if the catchments showed a change in behaviour during and after the drought. The results showed that when not accounting for the delayed response of the groundwater head, at least one of the specific yields in the catchments became infinitely small. Including the delayed groundwater head response did improve one of the catchments significantly by producing plausible specific yields for all geological units. A conclusion of this is that including the change in groundwater storage could improve the water balance. However, for it to do so a thorough analysis of the groundwater and subsurface needs to be conducted. Further, the water balance error was the third biggest flux after rainfall and actual evapotranspiration suggesting that the evapotranspiration scalar reduced the actual evapotranspiration too much. All fluxes did decrease during the drought, by how much differed between the catchments and the water balance components. Two of the catchments showed a change in behaviour during the drought that persisted in the years following the drought. The most likely fluxes to have caused this were the change in runoff and groundwater storage. The other two catchments showed a smaller change in behaviour during the drought and an indication that it was on its way back to the same state as before the drought. The likely fluxes to have caused the small change in behaviour was runoff and actual evapotranspiration.