Multi‐Proxy, Multi‐Season Streamflow Reconstruction with Mass Balance Adjustment

Long history of river discharge, or streamflow, can be reconstructed from tree rings. These reconstructions help us understand the water cycle in the past, but they have not been widely used in water resources operations. This is because reconstructions are often annual (having only one data point per year). By combining different tree species and different features of tree rings (for example, ring width and stable isotope ratio), it is possible to reconstruct seasonal streamflow in addition to the annual one, and that is our first goal. But a major challenge arises: how do we ensure that the total flow volume of the seasonal reconstructions closely matches the annual one? This mass balance criterion is important to avoid misguiding water management decisions such as allocating water to different sectors. We develop a novel method to reconstruct seasonal and annual streamflow while accounting for mass balance at the same time. Our work is thus a step closer toward operational usability of streamflow reconstructions in water resources management.