Team Meeting 5/5/2025

[rburghol]

What Do Our Successes & Failures Look Like?

What are our expectations of our individual models in different months?

How does the Simple LM and Stormvol Compare monthly (bar charts)

[COBrogan]

Future to-do / Wishlist that may help breakdown models:

• Comparing WDM files (e.g. precipitation summaries over certain land segments) for the amalgamated scenarios vs the PRISM or NLDAS2 only WDM files
• Bar plots of the $R^2$ values of the ratings in comparison to obs. vs. predicted plots Methods Review Plots HARParchive#1404 (comment)
  • Some inital boxplots visualizing this are here
• Individual events - Using Workflow geo Storm Volume Plots HARParchive#1402 event plotting method, can we see an individual storm (flow + precip) and compare directly to the method that was selected? And how did the model perform?
  • We should be able to include one or more model scenarios
  • We should include points showing the start/end point of each storm
• Extended low and high flows - How does the model compare between 1-, 7-, 30-, 90-day low and high flows?
• Pulse Flows - Adding into the model summary RMD, can we include a new plot that shows the IHA high + low pulse counts (perhaps automatically found via IHA::group4() function. Monthly metrics will be found in IHA::group1() or IHA::group2())
• Can we look at flow and precip boxplots for each dataset side-by-side? Looking at Seasonal Precipitation Plots HARParchive#1407 boxplots, can we add the USGS gage flow?
  • Can we add model flow to this? Can we add model precip to this?
  • Can we have a function that plots a boxplot for a given series of months for each raw data source (prism, NLDAS2, daymet), USGS flow, and maybe model flow?
    • For now, model precip is only available for a landsegment. We'd only be able to compare precip between the PRISM only model, the NLDAS2 only model, and the amalgamated models until we run a daymet only model OR summarize the amalgamated models by USGS gages

[COBrogan]

@rburghol , I made some tweaks to @nathanielf22 's code to include model outputs and storm start/end dates. Now we should be able to quickly hop to a storm or a period of time and check model performance more specifically against precip and flow.

library("hydrotools")
basepath='/var/www/R'
source(paste(basepath,'config.R',sep='/'))
source("https://raw.githubusercontent.com/HARPgroup/HARParchive/refs/heads/master/HARP-2024-2025/stormEvent_plot_any.R")

storm_plot("usgs_ws_01646000", date = "2003-01-01", bufferDays = 30,
           modelScenarios = c("pubsheds","mubsheds","lubsheds","subsheds"),
           riverseg = "PM7_4581_4580")

[COBrogan]

Summer, 2007

storm_plot(hydrocode = "usgs_ws_01646000", date = "2007-06-01", bufferDays = 60,
            modelScenarios = c("pubsheds","mubsheds","lubsheds","subsheds"),
            riverseg = "PM7_4581_4580")