2_process_sw_gw_site_data.R

source('2_process/src/process_nwis_data.R')

p2_sw_gw_site_targets_list <- list(
  
  ## simplified vrn of all site in watersheds(regardless of whether there is relevant data at that site) + all site types
  tar_target(p2_site_in_watersheds_sf,
             p1_site_in_watersheds_sf %>% 
               select(site_tp_cd, site_no, Name, geometry)
             ),
  
  ## get just gw sites with outputed data for 2000-2020 (timeframe to change)
  tar_target(p2_nwis_dv_gw_data, 
             p1_nwis_dv_gw_data %>%
               left_join(p2_site_in_watersheds_sf, by = 'site_no') %>%
               st_as_sf() %>% 
               mutate(lon = st_coordinates(.)[,1], lat = st_coordinates(.)[,2]) %>% 
               filter(grepl('GW',site_tp_cd)) %>% 
               st_drop_geometry() %>% 
               ## quickly re-organizing cols
               select(!starts_with('X_'),starts_with('X_'))
             ),
  
  ## getting all sites along lake 
  tar_target(p2_sw_streamorder3_sites,
             sites_along_waterbody(p2_site_in_watersheds_sf,
                                   p2_lake_tributaries,
                                   lake_waterbody = FALSE)
  ),
  
  ## this takes a 5+ minutes due to time for buffer of tributaries to generate
  tar_target(p2_sw_in_lake_sites,
             sites_along_waterbody(p2_site_in_watersheds_sf,
                                   p2_saline_lakes_sf,
                                   lake_waterbody = TRUE)
             
  ),
  
  ## get just cont dv sw sites with outputed data for 2000-2022 with stream order category column
  tar_target(
    p2_nwis_dv_sw_data, 
    join_site_spatial_info(nwis_data = p1_nwis_dv_sw_data,
                           sites_sf = p2_site_in_watersheds_sf,
                           join_site_col = 'site_no') %>% 
      add_stream_order(nwis_sw_data = ., 
                       sites_along_streamorder3 = p2_sw_streamorder3_sites,
                       sites_along_lake = p2_sw_in_lake_sites) %>% 
      ## re-organizing cols so that measurements cols come after non-measurement cols
      select(!starts_with('X_'),
             starts_with('X_'))
  ),
  
  ## get just discrete sw sites with outputed data for 2000-2022 with stream order category column
  tar_target(
    p2_nwis_meas_sw_data, 
    join_site_spatial_info(nwis_data = p1_nwis_meas_sw_data,
                           sites_sf = p2_site_in_watersheds_sf,
                           join_site_col = 'site_no') %>% 
      add_stream_order(nwis_sw_data = ., 
                       sites_along_streamorder3 = p2_sw_streamorder3_sites,
                       sites_along_lake = p2_sw_in_lake_sites) %>% 
    ##  re-organizing cols so that measurements cols come after non-measurement cols
    select(!c('lat','lon'), c('lat','lon'))
  ),
  
  ## get just discrete gw sites with outputed data for 2000-2022 (no stream order category column)
  ## Note there are several sites for gw that we are keeping. GW, GW-HZ (Hyporheic-zone well), GW-MW (mult. wells), GW-CR (collector/ranney well), GW-TH (Test hole not completed as a well)
  tar_target(
    p2_nwis_meas_gw_data,
    join_site_spatial_info(nwis_data = p1_nwis_meas_gw_data,
                           sites_sf = p2_site_in_watersheds_sf,
                           join_site_col = 'site_no') %>% 
      ## both dfs have a site_tp_cd col so when joining, two versions are created. Resetti
      mutate(site_tp_cd = site_tp_cd.y) %>% 
      select(!contains(c('.x','.y'))) %>% 
      select(!c('lat','lon'), c('lat','lon'))
  )
)

  # # SW data -----------------------------------------------------------------
  # 
  # p1_nwis_dv_sw_data_sf <- p1_nwis_dv_sw_data %>% left_join(sites_simplified, by = 'site_no')
  # 
  # p1_nwis_dv_sw_data_sf$site_tp_cd %>% unique()
  # # "ST"     "ST-DCH" "LK"     "FA-DV"  "ST-CA"  NA       "SP" 
  # 
  # p1_nwis_dv_sw_data_sf %>% filter(site_tp_cd == 'FA-DV') %>% pull(site_no) %>% unique()
  # ## 1 site has SW data for site_tp_cd FA-DV
  # 
  # p1_nwis_dv_sw_data_sf %>% filter(site_tp_cd == 'SP') %>% pull(site_no) %>% unique()
  # ## 6 sites has SW data for site_tp_cd SP
  # 
  # p1_nwis_dv_sw_data_sf %>% filter(site_tp_cd == 'LK') %>% pull(site_no) %>% unique()
  # ## 10 sites has SW data for site_tp_cd SP
  # 
  # 
  # sw_sites <- p1_nwis_dv_sw_data_sf %>% 
  #   filter(site_tp_cd %in% c('ST','LK')) %>% 
  #   group_by(site_no,site_tp_cd, geometry) %>%
  #   summarize(n()) %>%
  #   st_as_sf() 
  # 
  # ## buffer tributaries
  # tributaries_buffered <- p2_lake_tributaries %>%
  #   st_buffer(1000)
  # 
  # ## buffer lakes
  # lakes_buffered <- p2_saline_lakes_sf %>% st_buffer(1000) 
  # 
  # ## join lake and tribs 
  # ## / UPDATE: This takes too long, going to skip
  # # lake_tributaries <- st_union(tributaries_buffered, lakes_buffered)
  # 
  # # filtered_sites_along_tribs<- sf::st_filter(x = sw_sites,y = tributaries_buffered,
  # #               .predicate = sf::st_is_within_distance,
  # #               dist = units::set_units(0, m))
  # ## /
  # 
  # filtered_sites_along_lk <- st_join(sw_sites, lakes_buffered, left = FALSE)
  # filtered_sites_along_tribs <- st_join(sw_sites, tributaries_buffered, left = FALSE)
  # 
  # ## st_join create tables from the lakes_buffered and tributaries buffered table - so jusing just the filter original table
  # filtered_sw_sites_sf <- sw_sites %>% filter(site_no %in% c(filtered_sites_along_tribs$site_no,filtered_sites_along_lk$site_no))
  # 
  # 
  # 
  # )