Derived variables

[bpbond]

Implement derived variables; add L2 plots; debug code after running all 2024 data through

[Fausto2504]

@bpbond,

This is a how COMPASS calculates groundwater level below the soil surface.

For this, you need Aquatroll instalation info:
https://docs.google.com/spreadsheets/d/1O4sHvj2FO7EcWEm3WpKEZhFubGn8HCcUsz9EFXhQTXM/edit?gid=0#gid=0

Well measurements (chart explaining): https://docs.google.com/document/d/1N5mJfluWiq6zGNgjLacdm11zS27FqC3tRdE7gdGW9BE/edit?tab=t.0

Methods description:
"Groundwater level above and below the soil surface (GWL, m) was calculated using the relationship between fluid potential and hydraulic head referred to the soil surface. (35) Then, barometric pressure (P, kg m–1 s–1), groundwater density (ρ, kg m–3), gravity (g, m s–2), and depths of pressure sensors were measured during the sondes deployment (DPS) as follows:
GWL = (P / (p*g)) - Dps"
Reference: Machado-Silva et al., 2024 https://pubs.acs.org/doi/10.1021/acs.est.4c01115

R script:

Reading well dimensions to calculate wl_below the surface

well_dimensions <- read_sheet("https://docs.google.com/spreadsheets/d/1O4sHvj2FO7EcWEm3WpKEZhFubGn8HCcUsz9EFXhQTXM/edit#gid=0") %>%
mutate(location = case_when(transect_location == "Upland" ~ "UP",
transect_location == "Transition" ~ "TR",
transect_location == "Wetland" ~ "W"),
ground_to_sensor_cm = ring_to_pressure_sensor_cm - (well_top_to_ground_cm - bolt_to_cap_cm)) %>%
dplyr::select(site, location, ground_to_sensor_cm)

df_raw_depths <- inner_join(df_raw, well_dimensions, by = c("site", "location")) %>%
filter(!is.na(pressure_mbar))

df <- df_raw_depths %>%
mutate(density_gcm3_cor = ifelse(water_density > 0.95, water_density, 1),
pressurehead_m = (pressure_mbar * 100) / (density_gcm3_cor * 1000 * 9.80665),
wl_below_surface_m = pressurehead_m - (ground_to_sensor_cm / 100))

In the R script:
GWL = wl_below_surface_m
P = pressurehead_m
p = density_gcm3_cor
g = 9.80665
Dps = ground_to_sensor_cm
Note: Multiplications 100 and 1000 are unit conversions

[bpbond]

Thank you @Fausto2504 !

I will look at this and a script that @peterregier sent by email and let you guys know if I have questions.