fnc_get_soil: Soil-list creation

In rhabel/modLWFB90: R-package for applying LWF-Brook90 directly from FVA-BW servers

Soil-list creation

Description

This function is a wrapper of several functions and chunks of code and the main point of access for the final user. It takes a dataframe of coordinates of the points to be modeled and returns a list of soil data frames as required by LWFBrook90R. Adjustment options exist for the origin of soil data, the PTF to be used, whether MvG-parameters should be limited to a certain range, as well as all options for roots included in make_rootden and fnc_roots that can be passed down from here.

Usage

fnc_get_soil(
  df.ids,
  soil_option,
  PTF_to_use,
  BW_or_D = "BW",
  df.soils = NULL,
  limit_MvG = T,
  add_BodenInfo = T,
  incl_GEOLA = T,
  limit_humus = T,
  parallel_processing = F,
  bze_buffer = 12,
  meta.out = NA,
  limit_bodtief = NA,
  force_reg_rootsdepth = F,
  maxcores = NA,
  path_df.LEIT =
    "J:/FVA-Projekte/P01540_WHHKW/Programme/Eigenentwicklung/modLWFB90_data/LEIT_DB/LP_DB.rds",
  ...
)

Arguments

df.ids	either a data frame containing the following columns: ID_custom - a unique ID-column for assignment that all intermediate products as well as the output will be assigned to. easting and northing - coordinates in UTM EPSG:32632 OR: A simple feature with at the ID_custom-column. If df.ids is a simple feature, the polygon's centroids will be intersected with the Standortskartierung. A merging of information of partial coverage fractions of different STOK-polygons is not supported. So it is recommended to only use STOKA-Data as polygon-input layers, or to only use polygons that are small enough for the intersection process to not be an issue.
soil_option	whether regionalised BZE, STOK Leitprofil data (BZE), or own data should be used for modelling. While option BZE with a buffer of 50 shouldn't create many NAs. With option OWN, inusers can enter their own soil data (i.e. from lab or field experiments). If the option OWN is selected, the dataframes must be passed at df.soils.
PTF_to_use	the PTF to be used in the modeling process. Options are HYPRES, PTFPUH2, or WESSOLEK. Alternatively, if MvG parameters have been retrieved elsewhere (i.e. by lab analyses), OWN_PARMS can be selected to skip this.
df.soils	if OWN is selected at soil_option, a data frame must be provided here with data for each point or polygon in df.ids. The following columns are necessary: ID_custom - a unique ID matching the IDs of df.ids mat - number of soil layer starting with 1 counting upwards upper and lower - upper and lower boundaries of soil layers in cm humus - thickness of the humuslayer in m gravel - gravel content in volumetric percent bd - bulk density of fine earth in g/cm³ (not needed if PTF = "WESSOLEK") oc.pct - organic carbon content in mass percent sand - sand content in volumetric percent silt - silt content in volumetric percent clay - clay content in volumetric percent texture - optional - either texture or sand/silt/clay-percentages are needed Caution: If PTFs are to be applied, the columns required in fnc_PTF must be provided. Else, if PTF_to_use is set to OWN_PARMS, the following columns must be provided NA-free: ths, thr, alpha, npar, mpar, ksat, and tort. If roots are to be calculated, the columns required in fnc_roots must be provided. Otherwise they need to be stored in a column called rootden. If the nFK shall be calculated at some point, this will be done for the first 1m depth, so in this case one of the layers should end at 100cm depth.
limit_MvG	should the hydraulic parameters limited to "reasonable" ranges as described in fnc_limit. Default is FALSE.
add_BodenInfo	shall further soil info (nFK, PWP, FK, texture ...) be added to the soil-df, default is TRUE
incl_GEOLA	information from the Geowissenschaftliche Landesaufnahme will be used to get additional data on soil depth and max root depth. On top of that, Standortskartierung and GEOLA will be used for identifying soil types that will be modelled differently to include the effect of groundwater (Gleye / Auenboeden) or alternating Saturation (Stauwasserboeden). Default is TRUE
limit_humus	will be passed to fnc_PTF. See documentation for further information.
parallel_processing	the lists of dataframes are processed several times (adding roots, adding nFK information etc.). Default is F and runs with normal lapply statements. If many points are modelled, it is advised to set this to T, to activate parallel processing on several cores (as many as available). A BZE-based testrun with 32 GB RAM and 8 cores revealed a higher performance of parallel processing starting at the threshold of 250 points.
bze_buffer	whether buffer should be used in extracting points from BZE raster files if NAs occur in m, default is 12, because that way only the closest of the 25m raster cells gets found and we don't get multiple points from the same cell
limit_bodtief	max soil depth, default is NA and uses max soil depth as defined in df.LEIT, BZE or the GEOLA-dataset. If not NA, soil-dfs are created down to the depth specified here as depth in m, negative
force_reg_rootsdepth	there is a regionalised max-rootdepth, that takes into account certain soil conditions. However, it reduces the soil depth with roots significantly. Hence, by default, it is deactivated. If set to T, it will limit the lower roots depth to the regionalised values. #' @param meta.out if soil_option = "BZE" and bze_buffer != NA, this is a way to save the meta-information of the buffering process. meta.out must be a path to a .csv-file, where for each point the coordinates and distance of the buffered cell will be stored.
maxcores	when using a computer with multiple cores and parallel_processing = T, too many busy cores will kill the process because it creates too much data for the RAM. For 32GB RAM a maximum of 30 cores are recommended (when no other applications are open).
...	further function arguments to be passed down to fnc_roots. Includes all adjustment options to be found in make_rootden. Only exception is the roots functions' parameter maxrootdepth, which, if desired, has to be specified here as roots_max, because maximal root depth setting according to vegetation parameters will be complemented by root limitations from soil conditions. Settings can be either single values, applied to all soil data frames equally, or vector with the same length as df.ids specifying the roots setting for each modelling point. see example. If roots are counted and provided in df.soils as column rootden, set to table.

Value

Returns a list of soil data frames completely processed to be further used by run_multisite_LWFB90 or run_LWFB90

Examples

# STOK based, filled with BZE if STOK unavaioable
fnc_get_soil(df.ids = test.ids.bds,
             soil_option = "STOK_BZE",
             PTF_to_use = "HYPRES",
             rootsmethod = "hartmann")

# same example, but with a uniform depth of 2m
fnc_get_soil(df.ids = test.ids.bds,
             soil_option = "STOK_BZE",
             PTF_to_use = "HYPRES",
             rootsmethod = "hartmann",
             limit_bodtief = -2)

# roots methods can be set individually
fnc_get_soil(df.ids = test.ids.bds,
             soil_option = "BZE",
             PTF_to_use = "HYPRES",
             rootsmethod = "betamodel",
             limit_bodtief = -2,

             maxrootdepth = c(-1, -2, -1, -2, -0.5),
             beta = 0.98)


# points in gaps of STOK-maps are not modelled
fnc_get_soil(df.ids = test.ids.bds,
                soil_option = "STOK",
                PTF_to_use = "HYPRES",
                rootsmethod = "betamodel",
                beta = 0.95)


# sample data frame with "own" data
df.own.test <- data.frame("ID_custom" = c(rep("E",4), rep("D",4), rep("C",4), rep("B",4), rep("A",4)),
                          "mat" = rep(1:4, 5),
                          "upper" = c(0,5,10,30, 0,10,50,60, 0,5,10,30, 0,10,50,60, 0,15,20,70),
                          "lower" = c(5,10,30,60, 10,50,60,100, 5,10,30,60, 10,50,60,100, 15,20,70,100),
                          "sand" = rep(30, 20),
                          "silt" = rep(40, 20),
                          "clay" = rep(30, 20),
                          "oc.pct" = c(5,2,1,0,4,1,0,0, 5,2,1,0,4,1,0,0,8,2,0,0),
                          "bd" = c(1.1,1.2,1.4,1.5,1,1.2, 1.2, 1.3, 1.1,1.2,1.4,1.5,1,1.2, 1.2, 1.3, 1.1,1.4, 1.6, 1.6 ),
                          "gravel" = c(1,5,5,20, 0,0,10,15, 5,8,8,10, 20,20,25,60,0,0,0,0) ,
                          "humus" = c(rep(0.04,4), rep(0.06, 4), rep(0.04,4), rep(0.1, 4), rep(0.05, 4)))

fnc_get_soil(df.ids = test.ids.bds,
                soil_option = "OWN",
                PTF_to_use = "HYPRES",
                df.soils = df.own.test,
                rootsmethod = "betamodel",
                beta = 0.95)


# perforremance test
df.ids <- readRDS("H:/FVA-Projekte/P01540_WHHKW/Daten/Ergebnisse/Modellierung_BW/input_data/ids_master_tr_1.rds") %>%
  mutate(id_standard = paste0(easting, northing),
         tranche = fnc_relateCoords(.)$tranche)
for(i in c(10,50,250, 1000, 5000, 10000)){
  tic(paste0(i, " STOK Punkte ohne Parallel:"))
  test <- fnc_get_soil(df.ids = df.ids[1:i,],
                       soil_option = "BZE",
                       PTF_to_use = "HYPRES",
                       parallel_processing = F)
  toc(log = T)
  log.txt <- tic.log(format = TRUE)
  tic.clearlog()

  write(log.txt[[1]],
        file = "H:/FVA-Projekte/P01540_WHHKW/Programme/Eigenentwicklung/paketanwendung/time.txt",
        append=TRUE)
  rm(log.txt)

  tic(paste0(i, " STOK Punkte mit Parallel:"))
  test2 <- fnc_get_soil(df.ids = df.ids[1:i,],
                       soil_option = "STOK",
                       PTF_to_use = "HYPRES",
                       parallel_processing = T)
  all.equal(test, test2)
  toc(log = T)

  log.txt <- tic.log(format = TRUE)
  tic.clearlog()

  write(log.txt[[1]],
        file = "H:/FVA-Projekte/P01540_WHHKW/Programme/Eigenentwicklung/paketanwendung/time.txt",
        append=TRUE)
  rm(log.txt)
}

rhabel/modLWFB90 documentation built on Nov. 21, 2024, 3:28 a.m.