I created a new anaconda environment for python 3.6 today. After installing the unofficial binary pyproj‑2.2.2‑cp36‑cp36m‑win_amd64.whl, the kernel crashed whenever invoking pyproj.Proj(), giving returning error -1073741571.

On my build, the issue was resolved by reverting to the previous version .

The following warning is produced during the first convolution step in EventsTable.ipynb:

This warning seems to only occur for the 6 hour duration events and only for some project areas. It does not appear to affect the results but should be investigated to identify the cause in case there is some underlying stability issue.

• Problem: In PrecipTable.ipynb, when data cannot be downloaded for a specific duration and recurrence interval, the final precipitation frequency table is incomplete. However, a complete table is needed to run EventsTable.ipynb.
• Solution: Any file that cannot be retrieved should be added to a list and the code should attempt to redownload that file at the end of the for-loop.

Currently, EventsTable_Statified.ipynb does not save any outputs. Add a method to calculate hydrographs and save to JSON files. @Mark-S-Bartlett recommends adding the nested hyetograph to describe the rainfall distribution and then running it through the SCS-CN equation.

The option to calculate the excess runoff for areas with stormwater infrastructure will be added in the branch: feature/reduced_pluvial.

Add legends to the plots in EventsTable_Statifed.ipynb. Currently, it is hard for the user to QC results without digging into the underlying code.

All local variables and functions should be lower case and for readability, all functions should have a description even if they seem self-explanatory.

Add a scope/summary of the hydromet_stratified.py below the libraries and above the functions. This way the user can get an overview of the contents and understand the workflow before diving into the individual functions.

It is good practice to add asserts/warningsto the code to make it more robust. For example, a warning could be added if the users enters a PMP that is unreasonably high or enters a return interval that is outside the bounds of the NOAA Atlas 14 precipitation frequency curve. @Mark-S-Bartlett recommends adding asserts/warnings for CN, PMP, and the mu parameter.

For method-specific variables like S_Limit, include a short description and an example value or range of acceptable values in the docstring. Similarly, when passing an array of parameters to a function, include an example:
. These types of descriptions make the code more readable and easier to identify issues.

When running reEventsTable.ipynb, plot_reduced_excess fails if it is passed a single duration.

In plot_runoff_distributions_final, the discharge range is hardcoded, ideally, you want this to be flexible for plotting in different areas, same in plot_max_potential_retentin_cond_runoff. Try to avoid hard coding when possible, can always pass the variable back to the user.

USACE recommends using confidence limits between 0.01 and 0.99 for calculating the mean curve. Consider reducing the range of confidence limits.

The polygons associated with the NOAA Atlas 14 volume and region do not perfectly overlap with the volume codes and their associated precipitation frequency rasters. For example, volume 8, region 4 and volume 9, region 1 have overlapping polygons but all of the se volume code raster is contained within volume 9, region 1 but does not cover the entire region. Part of the mw volume code extends into volume 9, region 1 but not across the entire portion associated with volume 8, region 4. For now, an error will be issued if the area of interest does not intersect the precipitation frequency raster which will instruct the user to manually identify the volume.

Read the docs needs to be updated to include documentation from PR #37 on the fluvial analysis.

This issue will be addressed in the branch: feature/DSS

Testing notebooks on binder, urllib & ftplib methods timing out when attempting to access Atlas 14 datasets on file server at ftp://hdsc.nws.noaa.gov/pub/hdsc/data/.

@Tmodurr -Have you had similar issues with NOAA ftps running on remote servers?

Some users have found that the PM-EventsTable.ipynb hangs or appears to not complete while running the EventsTable.ipynb for the 6 hour duration in some areas. This appears to only be a problem in areas that receive little rainfall and therefore have a small range in the precipitation frequency curve. This small range of precipitation will lead to a small range in the runoff amounts, which could cause a large number of events to be grouped into the same convolution bin, reducing the efficiency of the convolution procedure.

NOAA Atlas 14 volume numbers and their corresponding codes are stored within a cell of the PrecipTable.ipynb and should be saved as a json. This will be addressed in the NOAA_Atlas_Keys branch.

For PrecipTable.ipynb, intersect_temporal_areas function, line: vol_gdf = vol_gdf.to_crs(geo_df.crs):

The conversion to a new crs (via to_crs) doesn't work for the first polygon in the dataframe being converted.

If the original database was a shapefile, the problem is fixed by setting os.environ["SHAPE_RESTORE_SHX"]= 'YES'. Not sure how to fix this issue.

Thanks!

The reEventsTable.ipynb fails when there are no lateral inflow boundary conditions specified. As a solution, test to see if there are lateral inflow boundaries before trying to calculate the lateral inflow hydrographs.

When fitting the GEV to the NOAA Atlas 14 precipitation data and to the flow frequency curve in hydromet_stratified.py, for the bounds, could the loc/mu be greater than the average precipitation? It seems like this optimization may be over-constrained. @Mark-S-Bartlett recommends also revisiting.

The JSON forcing files created by the distaleventsTable.ipynb are missing the pluvial_BC_units key. This issue can be fixed by updating the combine_distal_results function within hydromet. py.

Creating this issue to incorporate improvements on sampling from CN values suggested by @Mark-S-Bartlett.

Add a plot that shows the overlap between the NOAA Atlas region and the input polygon. Print/display the results of the intersect_temporal_areas function as a part of this plot. Be sure the plot shows the intersection gdf not the original polygon gdf.