Filter sites streamorder meas

[msleckman]

This is a small PR addressing issue #51

Note - I defined this as filtering in the title and issue, but in reality this is just creation of the stream order col for the nwis disc records dataset. No filtering is done until we visualize this data on maps.

Previously, the Stream Order 3 categorization was only done on the dv sw target p1_nwis_dv_sw_data.
This PR repeats this process for the field measurements (meas) sw dataset. - p1_nwis_meas_sw_data.
I therefore created a function - add_stream_order() which takes the dplyr code chunk that was previously directly in the target chunk for p2_nwis_dv_sw_data, and placed it in a function in sites_along_waterbody.R

In sum:

functions added: add_stream_order()
targets added: p2_nwis_dv_sw_data

Findings:

Number of unique field measurement sites per lake, split by stream order:

p2_nwis_meas_sw_data %>% select(lake_w_state,site_no,stream_order_category) %>% distinct() %>%
group_by(lake_w_state,stream_order_category) %>% summarize(count = n()) %>% 
ggplot(aes(lake_w_state, count))+
geom_bar(aes(fill = stream_order_category), position ='dodge', stat = 'identity')

Summary table of sites:

summarize_tbl <-p2_nwis_meas_sw_data %>% count(lake_w_state,stream_order_category, site_no) %>% group_by(lake_w_state, stream_order_category) %>% summarize(site_count = n(), total_meas_across_sites = sum(n), avg_meas_per_site = (total_meas_across_sites/site_count) %>% round(2)) 

summarize_tbl %>% arrange(desc(avg_meas_per_site)) %>% print(n = 27)

# A tibble: 27 × 5
# Groups:   lake_w_state [13]
   lake_w_state       stream_order_category site_count total_meas_across_sites avg_meas_per_site
   <chr>              <chr>                      <int>                   <int>             <dbl>
 1 Walker Lake,NV     along SO 3+                   15                    3449            230.  
 2 Carson Lake,NV     along SO 3+                   42                    6098            145.  
 3 Sevier Lake,UT     along SO 3+                   17                    2125            125   
 4 Pyramid Lake,NV    along SO 3+                   68                    8476            125.  
 5 Great Salt Lake,UT along SO 3+                  102                    9459             92.7 
 6 Carson Lake,NV     not along SO 3+               35                    3169             90.5 
 7 Walker Lake,NV     along lake                     1                      65             65   
 8 Great Salt Lake,UT along lake                    12                     752             62.7 
 9 Pyramid Lake,NV    not along SO 3+               90                    5541             61.6 
10 Walker Lake,NV     not along SO 3+               76                    4543             59.8 
11 Great Salt Lake,UT not along SO 3+               58                    3344             57.7 
12 Pyramid Lake,NV    along lake                     2                     114             57   
13 Carson Lake,NV     along lake                     2                      83             41.5 
14 Sevier Lake,UT     not along SO 3+                7                     253             36.1 
15 Owens Lake,CA      along SO 3+                   20                     542             27.1 
16 Mono Lake,CA       not along SO 3+                4                      83             20.8 
17 Mono Lake,CA       along SO 3+                    7                      85             12.1 
18 Malheur Lake,OR    along SO 3+                   38                     276              7.26
19 Owens Lake,CA      not along SO 3+               12                      76              6.33
20 Franklin Lake,NV   not along SO 3+                1                       6              6   
21 Malheur Lake,OR    not along SO 3+                8                      37              4.62
22 Honey Lake,CA      along SO 3+                    5                       8              1.6 
23 Lake Abert,OR      along SO 3+                    7                      11              1.57
24 Goose Lake,CA      along SO 3+                    1                       1              1   
25 Honey Lake,CA      not along SO 3+                3                       3              1   
26 Lake Abert,OR      not along SO 3+                1                       1              1   
27 Summer Lake,OR     not along SO 3+                1                       1              1   

Discrete sw measurements coverage over time (ignoring SO category):

• streamflow (cf/s)
  

• gage height (ft) - note gaps
  

code for scatterplots

library(ggplot2)
library(tidyverse)

targets::tar_load(p2_nwis_meas_sw_data)

top_5_lakes_meas_coverage<- c('Great Salt Lake,UT','Sevier Lake,UT','Walker Lake,NV','Pyramid Lake,NV','Carson Lake,NV')

plot_fun <- function(data, col){
  data %>%
    select(lake_w_state, stream_order_category, measurement_dt, discharge_va, gage_height_va) %>%
    mutate(measurement_dt = as.Date(measurement_dt)) %>% 
    ggplot(aes(y = .data[[col]], x = measurement_dt))+
    geom_point(aes(color = lake_w_state), size = 0.7)+ theme_bw()
}

sw_meas_time_plot_fun <- function(meas_data, measurement_col, split_vec){
  plot_top <- plot_fun(data = meas_data %>%
                         filter(lake_w_state %in% split_vec),
                       col = measurement_col)
  plot_rest <- plot_fun(data = meas_data %>%
                          filter(!lake_w_state %in% split_vec),
                        col = measurement_col)
  cowplot::plot_grid(plot_top, plot_rest, nrow = 2)
  
}

# streamflow
sw_meas_time_plot_fun(meas_data = p2_nwis_meas_sw_data, measurement_col = 'discharge_va', split_vec = top_5_lakes_meas_coverage)
# gage_height
sw_meas_time_plot_fun(meas_data = p2_nwis_meas_sw_data, measurement_col = 'gage_height_va', split_vec = top_5_lakes_meas_coverage)

[msleckman]

SW sites with stream order breakdown:

Note: Most sites have less than 100 records on the sw side.

[msleckman]

Discovered that there is a merge conflict because main no longer has 4_export.R, mapping/saline_lakes_mapping_script.R and saline_lakes_mapping_script.R. These files were removed in order to merge #58. However it causes conflict with #56 now.

Resolved by merging #60 first to main (commit msleckman@9b78a4b) The conflict is no longer present in this PR, nor in #58 since the commit histories are now in sync.

[cnell-usgs]

Is there meaning in the split in lakes in this plot?

[cnell-usgs]

This ran for me but I'd like you to address the comments below before merging. The most important is to have tidy, well-defined, targets. It would be better to create a site-level target that includes the streamorder variable and other site-level info, rather than adding directly to the long form data. This will make the pipeline more modular, and the site-level target will be useful in subsequent targets.

The comments about the buffer accuracy are important, but can be ignored for now.

[cnell-usgs]

Ah, ok I see my suggestion of creating a site-level target is partially addressed in #58

Address minor comments and go ahead and merge, those comments can be addressed on next PR

[msleckman]

Is there meaning in the split in lakes in this plot?

The split was to help vis the data. The top plot has the lakes with the most complete data, the bottom plot has the vis with the less complete data. See code dropdown at bottom of PR description.

[msleckman]

pipeline runs with latest changes following review. Merging.