Basic build (missing precip)

[lindsayplatt]

Here's where I got tonight. Unfortunately, there are some issues with missing/discontinued data, so I've commented them out to get a build. Not totally dead ends (I've left notes about the leads I found, but frustrating to not get it totally working). Besides getting the data, the other piece I skipped was selecting certain gages to display so that there are fewer spark lines. A fair number pop over flood stage and many go offline.

Running scmake('6_visualize/out/animation.gif') should get you this file:

1. Hourly precip. We can't use the built in functionality because it doesn't extend to PR. Looking back at the Maria viz, we used a different source but I cannot find that same data. I've tried looking into some other NOAA sources but there are very few things available for hourly and very recent. Here is some code from some of the exploration I did.

# Trying to download precip - I couldn't figure out how to get data after reading these files in

# Downloaded from here: https://water.weather.gov/precip/data_availability.php
untar(tarfile = 'ncep_stage_iv_source_files_20220916.tar')
precip_grb <- sf::gdal_read('ncep_stage_iv_source_files_20220916/pr_20220916_24h.grb2')

# Downloaded from here https://water.weather.gov/precip/download.php#browsedata
untar(tarfile = 'nws_precip_last12hours.tar')
precip_nc <- sf::gdal_read('nws_precip_last12hours_pr.nc')

2. River network. Unfortunately, I can't seem to get the nhdplustools::get_nhdplus() function to return anything for river geometries or other waterbodies. So, I've commented these out. This function documentation makes it seem like we can get the NHDPlus data, but by getting the whole set and then subsetting locally? I stopped after getting this download to function:

# Tried downloading PR rivers:
x <- download_nhdplusv2(outdir = "./tmp", url = paste0(
  "https://s3.amazonaws.com/edap-nhdplus/NHDPlusV21/", 
  "Data/NationalData/NHDPlusV21_NationalData_Seamless", 
  "_Geodatabase_HI_PR_VI_PI_03.7z"))

3. RDGs (Rapid Deployment Gages). I commented these out for now from the "outro" because Fiona is not yet a listed storm event in the STN viewer.

[jordansread]

Note that PR had a large percentage of flood forecasting gages go out during Hurricane Maria. It may be misleading to obscure gages that failed because the failures during Maria were a key part of the viz story and an important thing to bring attention to (see text on opportunities for gage hardening in that viz; those hardening investments may not be distributed equitably).

[cnell-usgs]

Following up on this, it doesn't seem like hourly precip data are retained anywhere past the most recent 3 hours. Then it switches to 6, 12, and daily.