Where does all the water go? Partitioning water transmission losses in a data-sparse, multi-channel and low-gradient dryland river system using modelling and remote sensing

- A combination of RS and field data allows hydrodynamic modelling in data-sparse basins.
- Hydrodynamic model simulated complex flooding in anabranching dryland rivers.
- Transmission losses in “channel country” are up to 68% of inflow (mean = 46%).
- Losses are a combination of evaporation, infiltration and terminal water storage.
- Lateral inflow is the largest source of uncertainty in water balance estimation.

SummaryDrylands cover approximately one-third of the Earth's surface, are home to nearly 40% of the Earth's population and are characterised by limited water resources and ephemeral river systems with an extremely variable flow regime and high transmission losses. These losses include actual evaporation, infiltration to the soil and groundwater and residual (terminal) water remaining after flood events. These critical components of the water balance of dryland river systems remain largely unknown due to the scarcity of observational data and the difficulty in accurately accounting for the flow distribution in such large multi-channel floodplain systems. While hydrodynamic models can test hypotheses concerning the water balance of infrequent flood events, the scarcity of flow measurement data inhibits model calibration, constrains model accuracy and therefore utility. This paper provides a novel approach to this problem by combining modelling, remotely-sensed data, and limited field measurements, to investigate the partitioning of flood transmissions losses based on seven flood events between February 2006 and April 2012 along a 180 km reach of the Diamantina River in the Lake Eyre Basin, Australia. Transmission losses were found to be high, on average 46% of total inflow within 180 km reach segment or 7 GL/km (range: 4-10 GL/km). However, in 180 km reach, transmission losses vary non-linearly with flood discharge, with smaller flows resulting in higher losses (up to 68%), which diminish in higher flows (down to 24%) and in general there is a minor increase in losses with distance downstream. Partitioning these total losses into the major components shows that actual evaporation was the most significant component (21.6% of total inflow), followed by infiltration (13.2%) and terminal water storage (11.2%). Lateral inflow can be up to 200% of upstream inflow (mean = 86%) and is therefore a critical parameter in the water balance and transmission loss calculations. This study shows that it is possible to constrain the water balance using hydrodynamic models in dryland river systems using remote sensing and simple field measurements to address the otherwise scarce availability of data. The results of this study also enable a better understanding of the water resources available for ecosystems in these unique multi-channel and large floodplain rivers. The combined modelling/remote sensing approach of this study can be applied elsewhere in the world to better understand the water balances and water transmission losses, important drivers of ecohydrological processes in dryland environments.

184