PREEVENTS machine learning model optimization Project Log

2/25/2019

Jeff to Do by March 8

• Finalize preparing DHSVM config files.

1a. Transfer DHSVM files to AMS computer with WGet command

1b. Confirm parameters and parameter range to use in MCMC method

1c. Document justification for parameter range and parameter values

• Run model

2a. Run model for 3 year period with rain-to-snow temp as a parameter

2b. Run model for 3 month period that has no snow, turn off rain-to-snow parameter

Jeff To Do by March 4

• Get DHSVM model ready

1a. Make a Hydroshare resource to publish daily and hourly WRF-PNNL data for Sauk. - done

1b. Compare Sauk input parameters to Skagit report parameters. See this Skagit example

1c. Run model with full grid of WRF climate forcings.

• Get parameters ready

2a. Choose list of parameters that either Kelleher suggests or you have good reason

2b. Write down table and text describing parameters (1 paragraph). Update this file

2c. Make a version of FAST.py and DREAM.py for your parameters. See this example. Do pull request. Make new versions.

• Get time period ready

3a. Plot streamflow distribution. Pick 5 or less years that span to the tails of the full distribution of data.

3b. Make plots and write up (1 paragraph) for presenting.

Christina and Shiv To Do by March 4

Size machine for Model run time (from Jeff)

Start FAST

Run Decision Tree

Run DREAM

2/12/2019

List of resources:

• Paper Outline: Overleaf.com https://www.overleaf.com/project/5c608d8a3d92286beaaf796b
• Sauk Model instance: https://www.hydroshare.org/resource/2b69ca6ca3b54cdda0eea0892456d37e/
• Sauk Model instance/Climate forcings for Livneh WRF: https://www.hydroshare.org/resource/31bd8cc9eb1e4f14b1e760c08f596920/
• Preliminary project work: https://github.com/ChristinaB/Incubating-a-DREAM
• and preliminary paper outline: https://github.com/ChristinaB/Incubating-a-DREAM/wiki
• References list: https://github.com/ChristinaB/Incubating-a-DREAM/wiki/References
• Code for running methods: Combining artificial and human intelligence to improve mountain flood prediction: Code to be updated
• Code for running methods: Github repo

Christina To Do by Feb 28

1. Update paper tables in to Latex
2. Add References

Jeff To Do by Feb 28

1. Make a Hydroshare resource to publish daily and hourly WRF-PNNL data for Sauk.
2. Make a Hydroshare resource "Little 694" Clearwater watershed WRF-PNNL data and DHSVM setup.
3. Writing Task: Test out adding DHSVM and WRF-PNNL details to Overleaf.
4. Review Shiv's code and make a list of questions/requests.