Local Channel trib network (nhd+ or custom project specific, small pump-store, etc)

[rburghol]

Overview

• Scripts:
  • Nest NHD+ Model setup: https://github.com/HARPgroup/vahydro/blob/master/R/modeling/json/nhdplus_nested.R
  • Use: Rscript.exe /c/usr/local/home/git/vahydro/R/modeling/json/nhdplus_nested.R
    • Outlet COMID (press ENTER to query by point):
• hydr variables to pass
  • Include and use code from passing OUTDGT (used for withdrawals), prototyped (and tested here: #31
• timeseries variable inclusion and linkage
  • initially, since the hsp2 hydr routine already has timeseries variables imported, we can map variables already gathered to full path names, and then to IX values in the ts Dict for use in specl()
• Workflow in place:
  • hydr(): calls load_nhd_simple() after hydr_get_ix() sets up variables/Dicts
  • load_nhd_simple() loads an NHD+ file (currently hard-wired to Beaver Creek nhd_simple_8566737.json )
  • Verify: Is flow hooked back in?

Concepts

• Small tribs, especially those with timestep challenges can be simulated effectively with a networked approximation
  • Assumption 1: water surface is continuous amongst all members of a contiguous (non-impounded) local stream network,
  • water surface is related to mean surface in network, (or center of water mass surface?)
  • water surface at any point is proportional to flows through network
  • during a timestep, water surface at any sub-reach is related to the head/flow rate at inlet of that reach, and outlet at that reach
  • water surface slope in any reach is related to land slope and relative inflow within that reach
  • mean water surface slope in the reach network is related to land slope and head?
  • if the local inflow per unit area is uniform across all reaches in a contiguous network, water surface slope should be equal to land surface slope, and head should vary according to local drainage area and mean flow through network?
• Free flowing streams
  • a virtual FTABLE can determine the approximate amount of water in the network at any point in time, and therefore estimate the head at any free flowing point in the network, given that head at any point in a network will be dependent on other points
  • the actual rate through a given sub-stream can be estimated from that substreams FTABLE
  • in this way, the sub-stream is almost like a monitoring point, rather than a node.
  • 2 iterations might be needed to first estimate mean surface water elevation given last timesteps demand, then a final estimate of demand at each point, which then subtracts all demands in the network
• Impoundments
  • when impoundments are over-flowing, they behave like free flowing streams (though their stage-storage amounts are considered inside the virtual FTABLE
  • when less than over-flowing, riser structure or other mandates that they be solved individually
  • Solving imp approaches:
  • impoundment nodes are the places that designate nodes, therefore, they designate the outlet of a virtual FTABLE group
  • impoundment nodes can flow into the next virtual FTABLE group
  • Other approach?

CBP Architecture Support

• Add a new DSN to hold runoff only (instead of runoff + upstream, DSN 11) - see https://github.com/HARPgroup/meta_model/issues/9

Data Acquisition

• nhdplusTools
  • Reference:
    • https://doi-usgs.github.io/nhdplusTools/reference/index.html
    • https://doi-usgs.github.io/nhdplusTools/articles/advanced_network.html
    • https://github.com/DOI-USGS/nhdplusTools/issues/125
    • https://github.com/HARPgroup/HSPsquared/issues/16
  • See comment below for working through this https://github.com/HARPgroup/HSPsquared/issues/16#issuecomment-1405312264
  • Functions:
    • plot_nhdplus()
    • Returns object:
      • catchment: the actual catchments
        • areaSqKM: local drainage area
      • flowline: the lines connecting catchments and waterbodies
        • comid: relates it to a catchment feature
        • fromNode: this flowlines upstream most point (node) number
        • toNode: this flowlines downstream most point, shared as the fromNode of the flowline that this flows into
      • network_wtb: waterbodies on the connected flow line network
      • off_network_wtb: waterbodies not on the network

[rburghol]

# get outlet, known comid (point above beaver creek in Rivanna/Mechums)

 bct <- nhdplusTools::get_nhdplus(comid='8567219')

# this gets tonode/fromnode, but I am unfamiliar with the number scheme, what is it referring to? It;s not a comid...
bct$fromnode
# [1] 200061236
bct$tonode
# [1] 200061237
# How to get next down comid?
# get next up?

# this does not work
nhdplusTools::get_tocomid(bct)
# Error in `rename()`:
# ! Names must be unique.
# ✖ These names are duplicated:
#   * "id" at locations 1 and 2.
# Run `rlang::last_error()` to see where the error occurred.

# this also fails, though it avoids an error, since hte comid is the same as requesting feature, and tocomid is 0
nhdplusTools::get_tocomid(as.data.frame(list(tonode = bct$tonode, comid=bct$comid, fromnode = bct$fromnode)), remove_coastal = FALSE, return_dendritic = FALSE )
#     comid tocomid    tonode  fromnode
# 1 8567219       0 200061237 200061236
# 
bct$comid
#  8567219

# notr this
nhdplusTools::navigate_nldi(bct$geometry)
# or this:
nhdplusTools::navigate_nldi(bct)
# both yield:
# Error in check_nldi_feature(nldi_feature) : 
#   Missing some required input for NLDI. Expected length 2 character vector or list with optional names: featureSource, featureID

[rburghol]

Got this partially working due to help at nhdplusTools issue queue:

library(nhdplusTools)
#> USGS Support Package: https://owi.usgs.gov/R/packages.html#support
library(sf)
#> Linking to GEOS 3.9.0, GDAL 3.2.1, PROJ 7.2.1

start_point <- st_sf(
  id = 1, 
  geom = st_sfc(
    st_point(
      c(-78.652281, 38.071266)
    ), 
    crs = 4269
  )
)

sf_use_s2(FALSE)
#> Spherical geometry (s2) switched off

domain <- st_buffer(st_as_sfc(st_bbox(start_point)), .1)

# grab some sample data and plot the domain
nhd <- plot_nhdplus(bbox = st_bbox(domain))

# Also plot the start point.
plot(st_geometry(st_transform(start_point, 3857)), 
     add = TRUE, col = "red", cex = 2)

start_point <- st_join(start_point, nhd$network_wtbd)

wb_out <- get_wb_outlet(lake_id = start_point$COMID, 
                        network = nhd$flowline)

plot_nhdplus(outlets = list(wb_out$comid), streamorder = 2)
# fails with error:
# Error occured while plotting:  Error in UseMethod("st_transform"): no applicable method for 'st_transform' applied to an object # of class "NULL"

# Warning messages:
# 1: Unknown or uninitialised column: `origin`. 
# 2: In get_plot_data(outlets, bbox, streamord

[rburghol]

# get an nhd feature with known comid
mc = nhdplusTools::get_nhdplus(comid='8566905')
# Get centroid of NHD catchment
mc_cent = st_centroid(mc$geometry)
# get lat and lon
mc_lon = mc_cent[[1]][[1]]
# [1] -78.66296
mc_lat = mc_cent[[1]][[2]]