12th National Monitoring Conference
Stream Temperature and Conductivity Predictions in the Delaware River Basin Using Machine Learning Methods
Speaker: Charuleka Varadharajan, Research Scientist, Lawrence Berkeley National Laboratory
Abstract: Floods and droughts are projected to increase over the next few decades due to climate change. These streamflow disturbances can worsen water quality by increasing salt, nutrient, and contaminant concentrations, which will have direct consequences for human and ecosystem health. Hence, it is important to understand and predict how water quality in streams and rivers will respond to new flow-disturbance regimes. Here, we describe results from a US Department of Energy (DOE) project that utilizes data-driven approaches to understand how changes in streamflow-disturbance events, ranging from floods to low-flow conditions, impact water quality over time. Our initial study is focused on water temperature and conductivity predictions in the Delaware River Basin, using a combination of simple machine learning and more complex deep learning models. Diverse data for these modeling efforts were synthesized using our custom data integration tool, BASIN-3D. Our preliminary results using individual monitoring station data indicate that typically these models perform well for predictions of average stream temperature values but are poor at predicting extreme values. New approaches to predicting the impacts of extreme events on water quality are needed to help water managers make optimal decisions in an uncertain future.
Speaker(s)
Charuleka Varadharajan
Charuleka Varadharajan is a scientist in the Energy Geosciences Division of the Earth and Environmental Sciences Area at Berkeley Lab. As a biogeochemist, she is interested in studying the nexus of carbon, water, and energy with a focus on understanding and limiting the impacts of human activities on water quality and climate. Her research involves the monitoring and mitigation of contaminants in water resources; the measurement and prediction of carbon fluxes in terrestrial and subsurface environments; and the management, synthesis, and analysis of diverse multi-scale environmental datasets. Her expertise spans various techniques for data collection and analysis, including laboratory experiments; x-ray synchrotron spectroscopy; sensor-based field data collection; web-based tools to integrate distributed datasets in real-time; and the use of geoinformatics, statistical, and wavelet-based data processing to analyze high spatial and temporal resolution data. She is currently interested in applying a combination of statistical, data mining, and machine learning approaches to groundwater and related datasets in California to gain insights that can help the state manage its groundwater sustainably. She had previously participated in data-driven scientific assessments of well stimulation (hydraulic fracturing) in California performed for federal and state agencies and was part of an expert committee advising the state of California on criteria for monitoring groundwater impacted by well stimulation. Varadharajan earned her PhD from the Massachusetts Institute of Technology and conducted her postdoctoral research at Berkeley Lab.