In kylemonper/CamelsQuery: extract data from NCAR CAMELS model output

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  eval = F
)

library(CamelsQuery)

This guide walks through the how to:

1. download CAMELS data remotely
2. run and use the extract_huc_data function to query and visualize data from the CAMELS dataset
3. use the get_sample_data function to get usgs streamgauge data

download data.

Data can be download manually from https://ral.ucar.edu/solutions/products/camels. This package requires that specifically the following be downloaded from there:
- 1.2 CAMELS time series meteorology, observed flow, meta data (.zip)
- 2.0 CAMELS Attributes (.zip)

alternatively the download_camels() function can be used to automatically download and unzip the data into a folder of the user's choice

### specify new folder to be created
#~ note: this must be a nonexistant new folder within an already existing folder:
#~ here this is simply creating a camels_data folder within the users home directory

data_dir <- "~/Data/CAMELS"

# download_camels(data_dir)

The function requires three inputs

• basin directory (basin_dir)
• This is the location of the basin_data_public_v1p2 folder. From this directory you should be able to further navigate to desired daymet mean forcing data folders (labeled 01, 02, 03, etc) via : "~/home/basin_dataset_public_v1p2/basin_mean_forcing/daymet" , and the streamflow folders should be in: "~/home/basin_dataset_public_v1p2/usgs_streamflow" . This exact folder structure is required for the function to work properly
• attribute directory (attr_dir)
• location of .txt files for data attributes (camels_clim.txt, camels_geol.txt, etc)
• huc ids (huc8_names)
• a vector of 8 digit huc 8 ids to be queried

Running function

##~ directories
my_basin_dir <- "~/Data/CAMELS/basin_timeseries_v1p2_metForcing_obsFlow/basin_dataset_public_v1p2"
my_attr_dir <- "~/Data/CAMELS/camels_attributes_v2.0"
##~ list of hucs to query (provided as a vector)
my_huc8_names <- c("01144000", "01170100", "01169000")
# stream_gauges <- read_csv("../inst/extdata/USGS_trial_sites.csv")
# my_huc8_names <- stream_gauges$SiteNumber

### run function
##~ this returns a named list object with 9 items
camels_data <- extract_huc_data(basin_dir = my_basin_dir, 
                         attr_dir = my_attr_dir, 
                         huc8_names = my_huc8_names)

\newpage

Access output

view names of each list item

names(camels_data)

access each item

mean_forcing <- camels_data$mean_forcing_daymet
### an alternative using [[]] syntax: 
##~  mean_forcing <- camels_data[["mean_forcing_daymet"]]
## OR, because this is the first item in the list:
##~  mean_forcing <- camels_data[[1]]
##this returns a tibble/data frame containing the mean forcing data
glimpse(mean_forcing) 

## furthermore, we can see that each of the hucs we entered are present
unique(mean_forcing$ID)

\newpage

Visualize

library(ggplot2)
library(lubridate) # for dates
library(janitor) # clean column names

## first, rename huc ID's with location from camels_name, this isn't necessary, but makes for more informative labels
locs <- camels_data[["camels_name"]]
names(locs)

# clean column names
cleaned_forcing <- clean_names(mean_forcing)

## join camels_data (by ID) to bring in gauge names
cleaned_names <- left_join(cleaned_forcing, locs, by = c("id" = "gauge_id"))

### turn year, month, day columns into single "date" column
mean_forcing_date <- cleaned_names %>%
  ## join columns, forcing into year, month, day format
  mutate(date = ymd(paste(year, mnth, day, sep = "-")))
ggplot(mean_forcing_date, aes(date, prcp_mm_day)) +
  geom_line() +
  facet_wrap(~gauge_name)

Get USGS water quality data

## read in gauges of interest
# gauges <- readr::read_csv("../inst/extdata/USGS_trial_sites.csv")


### dataRetrieval functions require sites to be named using an "Agency-Site#" format, this code reformats the trial sites csv into this format
#~ eg: "USGS-01073319""
# gauges_new <- gauges %>% 
  # mutate(renamed_site = paste(SiteAgency, "-", SiteNumber, sep = ""))

## get first 20 site names
# site_names <- gauges_new$renamed_site
site_names <- paste0("USGS-", my_huc8_names)

sample_data <- get_sample_data(site_names)

## look at nitrogen species for all sites over time
N_spp <- sample_data %>% 
  filter(CharacteristicName == "Nitrogen, mixed forms (NH3), (NH4), organic, (NO2) and (NO3)")

ggplot(N_spp, aes(x = ActivityStartDate, y = ResultMeasureValue, color = MonitoringLocationIdentifier)) +
  geom_point() + labs(title = "Nitrogen, mixed forms (NH3), (NH4), organic, (NO2) and (NO3)")

kylemonper/CamelsQuery documentation built on Aug. 30, 2020, 12:34 p.m.