Jobs:Job-00191

From CSDMS
Post-doc position in machine learning - watershed modeling
University of Kansas, Kansas, United States
Start reviewing process: 12 August 2020


Position overview:

The University of Kanas is excited to announce a postdoctoral research opportunity to advance large-scale watershed modeling by upscaling from detailed models using machine learning techniques and coupling multiple models to improve representation of near-channel biophysical processes. The postdoctoral researcher will use land unit characteristics and water quality outputs from existing Soil Water and Assessment Tool (SWAT) watershed models to train and validate machine learning models. This project is part of a large, multi-institutional effort that is focused on improving aquatic biological community health in the Upper Mississippi River Basin, which is intensively managed for row-crop agriculture. This postdoctoral position is located at the University of Kansas in Lawrence, Kansas, a mid-size college town located on the banks of the Kansas River. The successful candidate must have appropriate work authorization and be physically present at the defined work location with the United States by the start date of employment The position is funded for up to 2.5 years pending annual review of performance. To apply, submit requested application materials at https://employment.ku.edu/staff/17500BR.


Watershed models are common tools to characterize water quality responses to changes in land use and land management yet are often so heavily parameterized that their application is not extendible to nearby regions with similar yet not identical characteristics. Numerous watershed models of subbasins of the Mississippi River exist yet remain of limited value towards a larger regional analysis due to the lack of consistent parameterization or architecture. Although there are larger regional models they often too large to be computationally feasible to resolve the scales at which land management decisions are made for cost-benefit analysis where extensive model executions must occur. The focus of the research for this postdoctoral project will be to develop the intellectual and model base needed to create parsimonious models of the Mississippi River basin that can inform regional response plans to predicted patterns of climate and land use change.


Job Duties:
60% Advance watershed modeling. Homogenize high-resolution high-certainty SWAT models, create training data response library, apply machine learning models to predict unmodeled landscape unit response, build river network model (e.g. LISFLOOD), couple landscape output to river network model, model validation and uncertainty, execute coupled upscale model under scenarios of climate and land use change.

20% Planning and collaboration. Participate in planning, designing and conducting watershed modeling research projects under the direction of a faculty supervisor. Analyze and evaluate cross-model and individual model results from scenarios and provide interpretations. Collaborate with a diverse project team from multiple institutions across United States to define overall project goals, share data and communicate results.

20% Communicate science. Complete writing tasks, including literature review, required to publish research in peer-reviewed journals, contribute toward the preparation of technical reports, papers and/or records. Present results at scientific conferences and to stakeholders. Contribute to the development, preparation and submission of externally funded proposals related to the research program.


Application review begins Aug 12, 2020 and will continue until a pool of qualified applicants is received.


For more information and to apply: https://employment.ku.edu/staff/17000BR.

KU is an EO/AAE, full policy http://policy.ku.edu/IOA/nondiscrimination

Of interest for:
  • Hydrology Focus Research Group
  • Artificial Intelligence & Machine Learning Initiative