Job Announcement: Postdoctoral fellow

Position: Postdoctoral fellow (1-2 years)

Topic: Developing a process-guided deep learning model for streamflow in Wisconsin

A recent decision by the Wisconsin Attorney General requires natural resource managers to regulate high-capacity wells to protect streams and trout fisheries, but we lack a strong understanding of whether groundwater withdrawals interact with climate change to affect streamflow and temperature. The persistence of the state’s world-class trout fisheries depends on cold water sustained by groundwater input and a natural hydrologic regime. Air temperatures are increasing, but stream warming can be buffered by groundwater. In fact, many trout populations in Wisconsin are currently thriving despite warming conditions due to high groundwater levels associated with recent increases in precipitation. This increased precipitation, however, often occurs as extreme events that can lead to floods. Depending on their timing and intensity, floods may threaten successful emergence of trout fry or trout survival. Imposed on top of these climatic interactions is a second anthropogenic driver of hydrologic change: increasing high-capacity wells, which can decrease local groundwater levels and recharge to streams, leading to decreased flow and warmer water temperatures during critical summer low flow periods. Under current high-groundwater conditions, streams remain flowing and cold, but what will happen if precipitation patterns shift or groundwater use intensifies? For effective decision making into the future, managers need tools that can simulate these interactions between hydrologic regimes, temperature, and groundwater pumping for use in assessments of well permit applications and to manage sustainable wild trout fisheries.

We will use trout population time series to estimate indices of annual recruitment and adult abundances for each population. We will use three analytical approaches to test whether antecedent weather conditions influenced trout populations in the past and to predict the influence of future climate and groundwater use: (1) groundwater-surface water hydrologic modeling of past and future streamflow and temperature, using new field data to evaluate parameter uncertainty; (2) hierarchical population dynamics models with catchment-level covariates (climatic and geospatial attributes); and (3) adaptation of models developed under (2) to future weather and groundwater use scenarios developed in (1). 

We are seeking an interested candidate to build a process-guided deep learning model
for streamflow in Wisconsin. The candidate would assist in grant-writing in July 2021 and be available to begin a postdoctoral position spring or summer of 2022 .
If interested, contact Dana Lapides (danalapides@gmail.com).