P-model simulation at NEECF site

Site selection and meta data

mysites <- "efg"

Required meta information is specified for each site (in rows) and a number of variables:

• lat for latitude (decimal degrees)
• lon for longitude (decimal degrees) - this is only used for data ingestion but not for the P-model simulation with rsofun.
• elv for elevation (m a.s.l.)
• year_start and year_end specifying years covered by the simulation
• whc for the soil water holding capacity
• koeppen_code to group sites for evaluation by Koeppen-Geiger climate zones.

Let’s create this info.

siteinfo <- tibble(
  sitename = "efg",
  lon = 117.369258403,
  lat = 30.026229678,
  elv = 330,
  year_start = 2015,
  year_end = 2015,
  whc = 250
  )

Simulation settings

Specify additional simulation parameters that are identical for all site-scale simulations.

params_siml <- list(
  spinup             = TRUE,
  spinupyears        = 10,
  recycle            = 1,
  soilmstress        = TRUE,
  tempstress         = TRUE,
  calc_aet_fapar_vpd = FALSE,
  in_ppfd            = TRUE,
  in_netrad          = FALSE,
  outdt              = 1,
  ltre               = FALSE,
  ltne               = FALSE,
  ltrd               = FALSE,
  ltnd               = FALSE,
  lgr3               = TRUE,
  lgn3               = FALSE,
  lgr4               = FALSE
    )

Run prepare_setup_sofun() to define the simulation settings that contain all the information specified by the two steps above (meta info, and simulation parameters), global simulation parameters are wrapped inside an additional column params_siml, added to the site meta info dataframe.

siteinfo <- prepare_setup_sofun(siteinfo = siteinfo, params_siml = params_siml)

Define model parameters

First, let’s do it by hand (calibration of parameters is shown later).

params_modl <- list(
    kphio           = 0.09423773,
    soilm_par_a     = 0.33349283,
    soilm_par_b     = 1.45602286,
    vpdstress_par_a = 999,
    vpdstress_par_b = 999,
    vpdstress_par_m = 999
    )

Define soil parameters

For now, this is implemented as an illustration. Should be made site-specific. Values entered here take no effect.

df_soiltexture <- bind_rows(
  top    = tibble(layer = "top",    fsand = 0.4, fclay = 0.3, forg = 0.1, fgravel = 0.1),
  bottom = tibble(layer = "bottom", fsand = 0.4, fclay = 0.3, forg = 0.1, fgravel = 0.1)
)

Get input

Forcing data was collected as described in ingestr/vignettes_add/get_sitedata_nonetwork.Rmd.

secs_per_tstep <- 60 * 60 * 24
forcing <- read_csv("~/data/neecf/climate_data/climate_data_collected.csv") %>% 
  mutate(prec = prec / secs_per_tstep, ppfd = ppfd / secs_per_tstep)

## Parsed with column specification:
## cols(
##   date = col_date(format = ""),
##   doy = col_double(),
##   temp = col_double(),
##   vpd = col_double(),
##   prec = col_double(),
##   ppfd = col_double(),
##   ccov = col_double(),
##   co2 = col_double(),
##   fapar = col_double(),
##   patm = col_double()
## )

CO2 is separate.

df_co2 <- ingest_bysite(
  sitename  = "efg",
  source  = "co2_mlo",
  year_start= 2015,
  year_end  = 2015,
  verbose = FALSE
  )

## [1] "ftp://aftp.cmdl.noaa.gov/products/trends/co2/co2_mm_mlo.txt"

## /var/folders/f_/vvrf3z593z7_hnkqmjt6rwtw0000gq/T//RtmpKuhenr/file142337fcac0cf

Run the model

Run the model for all the sites specified in the first step.

## run for a single site
mod <- run_pmodel_f_bysite( 
  "efg", 
  siteinfo$params_siml[[1]], 
  siteinfo, 
  forcing, 
  df_soiltexture, 
  params_modl = params_modl, 
  makecheck = TRUE 
  )

Check.

forcing %>% 
  ggplot(aes(date, fapar)) + 
  geom_line()

forcing %>% 
  ggplot(aes(date, ppfd)) + 
  geom_line()

mod %>% 
  ggplot(aes(date, gpp)) + 
  geom_line()

mod %>% 
  ggplot(aes(date, vcmax25)) + 
  geom_line()