Novel Physically-based Streamflow Monitoring Methodology

Monitoring streamflow is critical for short- and long-term decision-making of multiple aspects of water uses. The monitored data, mostly collected by specialized agencies, are also used to underpin scientific studies related to rivers and multi-disciplinary research in water science. Most of the streamflow gaging stations considers flows as quasi-stationary processes fluctuating within a variability range. This view along with the technical limitations of the past have given rise to semi-empirical relationships, conventionally termed ratings, that consist of one-to-one dependencies between flow variables. During gradually-varied flows, however, streams are both unsteady and non-uniform therefore the stationarity assumption is not valid. Depending on the stream bed slope and rates of the flow changes, the actual dependencies among the flow variables are distinct for the rising and falling hydrograph limbs and from the steady-flow relationships used by conventional ratings. These non-unique dependencies, labeled hysteresis, produce “loops” in the relationships between any two flow variables and separation of the flow hydrographs, as shown below.  Corrections to the steady conventional ratings or more complex ratings are seldom used to recover the hysteretic loops associated with the gradually-varied flows. Recognizing the impact of hysteresis on the conventional monitoring methods, the USGS has searched for alternative approaches (i.e., use of the index-velocity method and numerical models) for sites where complex ratings are necessary. Our proposed research is an effort along this focal area.

The project goals aim to advance the state of practice by translating outcomes of our prior research to operations. Specifically, this research aims at: 1) substantiating improvements brought by a novel monitoring method accounting for temporal flow changes and riparian vegetation growth; 2) assessing the method’s accuracy; 3) providing additional datasets for real-time assimilation in streamflow forecasting models; and 4) informing data producers and users on cost/benefit considerations for adopting “fit-for-purpose” monitoring alternatives. The proposed research innovatively integrates the instrumentation used in the conventional Slope-Area and Index-Velocity methods into a new hybrid method whereby directly measured variables in conjunction with physically based relationships enable direct streamflow measurements in steady and unsteady flows. The new method will be proof-tested at a monitoring site equipped with conventional stage-discharge rating.

The outcomes of this research will produce readily accessible protocols, algorithms, and cost/benefit considerations to: a) document in real time multi-variable relationships associated with runoff event propagation and seasonal riparian roughness changes; b) setting foundations for a new, data-science approach for real time streamflow estimation that enables hydrologists to collect streamflow data more efficiently (by removing the expensive process of developing rating curves); c) inform producers and users of the streamflow data about levels of complexity they have to adopt for the data-acquisition system to optimally fulfil their priorities for operational and scientific applications; and, d) providing essential streamflow information on current or near‐current conditions to be assimilated during forecasting for updating the internal states of streamflow modeling.