,
Kompetenzzentrum Wasser Berlin gGmbH (KWB) [cph]| Title: | Fertilizer chemical risk assessment |
|---|---|
| Description: | This risk assessment is based on the TGD on risk assessment by the European Comission. Possible Endpoints are soil organisms, groundwater organisms and human health via food consumption.Every variable can be entered as probability distribution to include uncertainties or site specific variations. |
| Authors: | Malte Zamzow [aut, cre] ,
Kompetenzzentrum Wasser Berlin gGmbH (KWB) [cph] |
| Maintainer: | Malte Zamzow <[email protected]> |
| License: | MIT + file LICENSE |
| Version: | 0.0.0.9000 |
| Built: | 2025-04-01 03:16:21 UTC |
| Source: | https://github.com/KWB-R/kwb.fcr |
Estimation is based on the FCR prepared Monte Carlo table
add_AirWater(p)add_AirWater(p)
p |
Parameter table created with |
The Parameter table extended by a column for the partition coefficient. Unit is m³/m³.
Estimation is based on the FCR prepared Monte Carlo table
add_bcf(p, sub_info)add_bcf(p, sub_info)
p |
Parameter table created with |
sub_info |
The table containing additional substance information loaded
with |
The Parameter table extended by a column for the bio concentration factor.
Calculation is based on the FCR prepared Monte Carlo table
add_deposition(p)add_deposition(p)
p |
Parameter table created with |
The Parameter table extended by a column for the atmospheric Deposition. Unit is mg/(kg * d).
Estimation is based on the FCR prepared Monte Carlo table
add_DT50(p, sub_info)add_DT50(p, sub_info)
p |
Parameter table created with |
sub_info |
The table containing additional substance information loaded
with |
The estimation is based on the K_d value of a substance according to the technical guidance document on risk assessment Part II (table 8). If there is a distribution of K_d values, the median K_d is used for the estimation. In this function there is no distinction between biodegradbility classes. All substances are assumend to be inherently biodegradable to encourage the user of further half-life research.
The Parameter table extended by a column for the bio concentration factor.
Estimation is based on the FCR prepared Monte Carlo table
add_Henry(p)add_Henry(p)
p |
Parameter table created with |
The Parameter table extended by a column for Henry coefficient. Unit is (Pa*m3)/mol
Calculation is based on the FCR prepared Monte Carlo table
add_kbio(p)add_kbio(p)
p |
Parameter table created with |
The Parameter table extended by a column for the biodegredation rate. Unit is 1/d.
Estimation is based on the FCR prepared Monte Carlo table
add_Kd(p, sub_info)add_Kd(p, sub_info)
p |
Parameter table created with |
sub_info |
The table containing additional substance information loaded
with |
Theoretically the initial concentration in soil can become negative if the pollutant concentration in fertilizes is negative. A negative concentration makes obviously no sense in a real situation. However, long-term impact is assessed, a broad concentration range of the pollutant also leads to more uncertain results. Negative concentrations in one year are averaged out. For the regression of the sorption coefficient, the concentration must be positive. Thus, negative values are increased to the concentration in top soil that would appear after one day of deposition.
The Parameter table extended by a column for the sorption coefficient
Calculation is based on the FCR prepared Monte Carlo table
add_kleach(p)add_kleach(p)
p |
Parameter table created with |
The Parameter table extended by a column for the infiltration rate. Unit is 1/d.
Calculation is based on the FCR prepared Monte Carlo table
add_kplant(p)add_kplant(p)
p |
Parameter table created with |
The Parameter table extended by a column for the plant uptake rate. Unit is 1/d.
Calculation is based on the FCR prepared Monte Carlo table
add_kvolat(p)add_kvolat(p)
p |
Parameter table created with |
The Parameter table extended by a column for the volatilization rate. Unit is 1/d.
Calculation is based on the FCR prepared Monte Carlo table
add_PNEC_soil(p)add_PNEC_soil(p)
p |
Parameter table created with |
The Parameter table extended by a column for the Predicted no-effect concentration for soil organisms. Unit is mg/(kg Dry Matter).
Estimation is based on the FCR prepared Monte Carlo table
add_SoilWater(p)add_SoilWater(p)
p |
Parameter table created with |
The Parameter table extended by a column for the partition coefficient. Unit is m³/m³.
Estimation is based on the FCR prepared Monte Carlo table
add_SoilWater_reverse(p)add_SoilWater_reverse(p)
p |
Parameter table created with |
The Parameter table extended by a column for the partition coefficient. Unit is m³/m³.
Query and complementation of input variables
add_variables(p, info)add_variables(p, info)
p |
The input data table created with |
info |
Additional input data information, created with
|
The table p exetended by new variable columns needed for the risk assessment
Estimation via direct or indirect logarithmic (log 10?) linear regression including pH, organic carbon and soil concentration
BCF_regression( constant, beta_ph, beta_org, beta_conc, regType = "direct", pH, org_c, conc )BCF_regression( constant, beta_ph, beta_org, beta_conc, regType = "direct", pH, org_c, conc )
constant |
Log constant in linear Kd regression |
beta_ph, beta_org, beta_conc
|
Log coefficient for pH, organic carbon and soil concentration in log-linear Kd regression |
regType |
Chacter specifying the way the K_d value is calculated. Either "direct" which means that the regression directly leads to the sorption coefficient or"indirect" which means the regression estimates a concentration in pore water and the sorption coefficient is subsequently be calculated by the quotient of soil and pore water concentration. |
pH |
pH value |
org_c |
Organic carbon content in percent. |
conc |
Soil concentration in mg/kg |
The constant value, the regression parameters (beta) and the soil characteristics can be single numeric values or vectors of the same length.
A vector of bio concentration factors without unit. It is defined as concentration in plant dry matter divided by concentration in soil.
List of Variables into Monte Carlo simulation input table
create_mcs_input(data_list, nFields)create_mcs_input(data_list, nFields)
data_list |
A list of input variable tables. Each table must have columns "value_1", "value_2", "distribution", "site_specific", "pID" and "shift" |
nFields |
Number of Monte Carlo Cycles |
A matrix with all input variables from the list
Plot cumulative sum of RQ after application and compare it with initial state
CumSumSoil(v0, v, year_x, xmin = 0.001, xmax = 10)CumSumSoil(v0, v, year_x, xmin = 0.001, xmax = 10)
v0 |
Vector of initial risk quotient distribution |
v |
Vector of final risk quotient distribution of x years |
year_x |
Numarical value of the last year of the risk assessment |
xmin |
Minimum value of x-Axis |
xmax |
Maximumg value of x-Axis |
Cumulative sums are plotted and the underlying table of quantiles of the initial and final risk quotients
The initial soil concentration is calculated so that the average concetrantion of the first 30 days leads to a PEC_water that equals the PNEC water
get_c_i_from_PNEC_porewater(dat, info, nFields)get_c_i_from_PNEC_porewater(dat, info, nFields)
dat |
List with all the input variables. This list is produced by
function |
info |
he table containing additional substance information loaded
with |
nFields |
Number of Monte Carlo cycles |
The equation is a combination of the definition for PEC soil and PEC porewater the from TGD, back-calculating to c0.
Numeric vector of lenght nFields, containig the initial concentrations in µg/L
Predicted human consumption via crop consumption. The calculation is based on the average predicted soil concentration within the growing period.
get_PEC_human(p, d, food_only)get_PEC_human(p, d, food_only)
p |
Parameter table created with |
d |
The number of days of the growing period. |
food_only |
If TRUE, the predicted human consumption via food is multiplied by 2 to compensate the disregard of water consumption |
Numeric vector of human consumption via crops in µg/d
Average predicted environmental concentration (PEC) over d days after fertilizer application. The Calculation is based on the FCR prepared Monte Carlo table
get_PEC_porewater(p, d)get_PEC_porewater(p, d)
p |
Parameter table created with |
d |
The number of days after fertilizer application. Somewhere between 1 and the number of days of the growing period, since plant uptake is considered |
Numeric vector of concentrations per field in µg/L
Average predicted environmental concentration (PEC) over d days after fertilizer application. The Calculation is based on the FCR prepared Monte Carlo table
get_PEC_soil(p, d)get_PEC_soil(p, d)
p |
Parameter table created with |
d |
The number of days after fertilizer application. Somewhere between 1 and the number of days of the growing period, since plant uptake is considered |
Numeric vector of concentrations per field in mg/kg soil
Risk quotients and high risk scenarios with and without fertilization
get_risk(fertPEC, noFertPEC, PNEC)get_risk(fertPEC, noFertPEC, PNEC)
fertPEC |
A vector of predicted environmental concentrations at the end of simulation with fertilization |
noFertPEC |
A vector of predicted environmental concentrations at the end of simulation without fertilization. Both vectors must be based on the same environmental conditions during simulation. |
PNEC |
A numeric value, defining the predicted no-effect concentration |
Data frame with PECs, Risk Quotient and identification of high-risk scenarios for both simulations, with and without fertilization, as well as the risk quotient difference due to fertilization
Estimation via direct or indirect logarithmic linear regression including pH, organic carbon and soil concentration. (log10 is used for all variables)
Kd_regression( constant, beta_ph, beta_org, beta_conc, regType = "direct", pH, org_c, conc )Kd_regression( constant, beta_ph, beta_org, beta_conc, regType = "direct", pH, org_c, conc )
constant |
Log constant in linear Kd regression |
beta_ph, beta_org, beta_conc
|
Log coefficient for pH, organic carbon and soil concentration in log-linear Kd regression |
regType |
Chacter specifying the way the K_d value is calculated. Either "direct" which means that the regression directly leads to the sorption coefficient or"indirect" which means the regression estimates a concentration in pore water and the sorption coefficient is subsequently be calculated by the quotient of soil and pore water concentration. |
pH |
pH value |
org_c |
Organic carbon content in percent. |
conc |
Soil concentration in mg/kg |
The constant value, the regression parameter (beta) and the soil characteristics can be single numeric values or vectors of the same length.
A vector of sorption coefficients in L/kg
Based on the input data provided in Excel sheets, 3 different PECs are calculated. The number of years and number of agricultural fields with different properties (Number of Monte Carlo cycles) can be chosen.
longterm_PEC( dat, info, years, nFields, use_mixing_factor = FALSE, PNECwater_c_i = FALSE, food_only = TRUE, growing_period = 180, t_res = 1, traceBackVariables = FALSE, keep_c_course = TRUE )longterm_PEC( dat, info, years, nFields, use_mixing_factor = FALSE, PNECwater_c_i = FALSE, food_only = TRUE, growing_period = 180, t_res = 1, traceBackVariables = FALSE, keep_c_course = TRUE )
dat |
A List with all the input variables. This list is produced by
function |
info |
A table containing additional substance information loaded
with |
years |
Years of fertilizer application |
nFields |
Number of Monte Carlo cycles |
use_mixing_factor |
Not working yet! If TRUE, porewater is diluted by pollutant free groundwater for a more realistic estimation of risks in groundwater. However the TGD approach assumes porewater = groundwater for the assessment. |
PNECwater_c_i |
If TRUE, the initial concentration in soil is derived by a RQ = 1 for groundwater risk assessment (PEC_porewater = PNEC_porewater). Thus, the inital concentration is defined by the applied PNEC for groundwater. |
food_only |
If TRUE, the predicted human consumption via food is multiplied by 2 to compensate the disregard of water consumption |
growing_period |
Numeric value specifiyng the number of days with plant growth |
t_res |
Temporal resolution in days to be returned in the output matrix (no effect on the calculation). The default is 1 (every day) which can lead to a very large output matrix. The maximal reslution is 365 days, since every year is calculated seperately. |
traceBackVariables |
If TRUE, the variables of every simulation year are returned. Note: this can lead to very large lists. |
keep_c_course |
If FALSE, the course of concentration is not part of the output. In that case variabel t_res is meaningless. |
List with
table for all PEC types per year,
table for concentration course in top soil in the predefined temporal resolution,
optional: If traceBackVariables TRUE: list of distributed input variables for each year, if traceBackVariables FALSE: Input variables of the first year including initial situations for human consumption and inital concentration in porewater
Estimation is based on the FCR prepared Monte Carlo table. The output rate is split into one rate with plant uptake during plant growth and one rate without plant uptake during winter time.
one_year(p, growing_period, t_res)one_year(p, growing_period, t_res)
p |
Parameter table created with |
growing_period |
Numeric value specifiyng the number of days with plant growth |
t_res |
Temporal resolution to be returned in the output matrix (no effect on the calculation) |
A time period can be considered as growing period if 1) plants grow and 2) plants will be harvested after growing. Otherwise the uptaken pollutants will return to the soil.
Matrix with number of columns equal to the rows of p and number of rows depending on the defined timesteps with t_res
The information of the Excel input files are used to prepare a Monte Carlo based data input for the calculation of a yearly concentration dynamic.
oneYear_matrix(dat, c_i, nFields, use_mixing_factor = FALSE)oneYear_matrix(dat, c_i, nFields, use_mixing_factor = FALSE)
dat |
List with all the input variables. This list is produced by
function |
c_i |
The initial concentration before fertilizer application |
nFields |
The number Monte Carlo Simulations. This can be seen as different agriculatural fields with different charactersitics. |
use_mixing_factor |
If TRUE, a mixing factor for porewater dilution by unpolluted groundwater is used to get the Risk quotient for Groundwater. The defahutl is FALSE. In that case pore water concentration is identical with groundwater concentration as is the approach in the Technical Guidance Document |
The pollutant concentration at t = 0 (c_0) already includes the fertilizer application
A Matrix with all available input data for the Assessment. The columns represent the input variabls, the rows correspond to the definied number of fields aka. number of Monte Carlo cycles.
This function combines a uniform distribution from minimum to maximum with the upper and lower tail of a normal distribution.
rderived(n, min, max, a = -Inf)rderived(n, min, max, a = -Inf)
n |
Number of samples to be drawn |
min |
Minimum value of normal comparables |
max |
Maximum value of normal comparables |
a |
Minimum of distribution (Default is -Inf) |
95%, 2.5% and 2.5% of all drawn values originate from the uniform distribution, the lower tail, and the upper tail of a normal distribution, respectively. The normal distribution is characterized by the mean of minimum and maximum. The standard deviation is derived from the 2.5% and 97.5% quantile definied by the minimum and maximum values. (sd = (max - mean) / 1.959963)
The number of Fields is multiplied by 38 for the uniform distribution, while it is multiplied by 1 for the tails of the normal distribution (38/40 = 95%). Finally, the defined number of fields is randomly drawn from the vector.
numeric vector of length n with randomly drawn values according to the distribution
Draw n samples from uniform, normal, truncated normal (-> positive only), lognormal or gamma distributions based on two input values
rdist(dist_name, value_1, value_2, n, shift = 0, seed = NULL)rdist(dist_name, value_1, value_2, n, shift = 0, seed = NULL)
dist_name |
Character vector specifying the Name of the distribution. Either "none", "uniform", "normal" "tnormal" (for truncated), "lognormal", "gamma", "derived", "tderived" or "logderived" |
value_1, value_2
|
Distribution parameters based on dist_name: |
n |
Number of samples to be drawn |
shift |
An numeric value defining a subsequent shift of the distribution The default is 0 (-> no shift) |
seed |
A numeric value to set the seed for random selection. The default is NULL -> no seed |
The values specified in value_1 and value_2 represent min and max for "uniform", mean and standard deviation for "normal" and "tnormal", log mean and log standard deviation (both natural logarithm) for "lognormal", shape and rate for"gamma". The shift value is especially interesting for lognormal or gamma distributions.
Numeric vector of length n with randomly drawn samples of the specified distribution
Reads Excel files with environmnetal proporties and substance coefficents.
read_fcr_input(input_path, pollutantName, siteName, fertilizerName = "none")read_fcr_input(input_path, pollutantName, siteName, fertilizerName = "none")
input_path |
Path of the folder "pollutants" and "sites" that contain the input data excel sheets |
pollutantName |
Name of Pollutant as in filename of the pollutant (prefix before "_sheet.xslx) |
siteName |
Name of the site as in filename of the environment sheet (prefix before "_sheet.xslx) |
fertilizerName |
Name of the fertilizer as in filename of the fertilizer sheet (prefix before "_sheet.xslx). If fertilizerName = "none", fertilizer application and pollutant concentration will be set to 0. |
Both input files must have the same path. The pollutant excel files must be named "'pollutantName'_sheet.xlsx". Furthermore, both excel tables must have the same number of columns and column names.
List of two items names "dat" and "info" and containing all the input variable data (information about distribution type, distribution parameters, type of variability and a unique variable ID) and additional information of how to simulate substance behavior.
Risk Aggregation Calculation of Delta Risk Quotient and Maximum Risk
risk_aggregation(df_risk)risk_aggregation(df_risk)
df_risk |
A dataframe created by |
The Delta Risk Quotient (RQ) is the average increase of high-risk scenarios caused by fertilization. A high-risk scenarios are defined as scenraios with a final RQ higher than 1. The increase refers to a comparison between two similar scenarios, the only difference is the application of fertilizer.
The maximum risk is the average RQ of the upper 5% of the RQ-Distribution. It is meant to describe the upper tail of the distribution.
Further information can be found in the report of the Eurpean Horizon 2020 project Nextgen "Assessment and risk analysis of NextGen demo case solutions" (p.151)
A list of 2, the delta risk quotient and the maximum risk
Risk Interpretation Assigns a risk to the aggregated risk parameters
risk_interpretation(delta_RQ, risk_max)risk_interpretation(delta_RQ, risk_max)
delta_RQ |
Numeric vector of length 1: Difference of of high-risk simulations between scenarios with and without fertilization |
risk_max |
Numeric vector of length 1: Mean value of the Upper five percent of RQ distribution |
Further information can be found in the report of the Eurpean Horizon 2020 project Nextgen "Assessment and risk analysis of NextGen demo case solutions" (p.151)
A list of 2, the delta risk quotient and the maximum risk
This function draws random values from a uniform distribution (minimum to maximum) combined with the upper and lower tail of a log normal distribution. Minimum and maximum values are transformed to a log10 scale first.
rlderived(n, min, max)rlderived(n, min, max)
n |
Number of samples to be drawn |
min |
Minimum value of normal comparables |
max |
Maximum value of normal comparables |
95%, 2.5% and 2.5% of all drawn values originate from the log-uniform distribution, the lower tail, and the upper tail of a log-normal distribution, respectively. The log-normal distribution is characterized by the mean of log-minimum and log-maximum. The standard deviation is derived from the 2.5% and 97.5% quantile defined by the minimum and maximum values. sd = (log10(max) - log10(mean)) / 1.959963
The number of Fields is multiplied by 38 for the uniform distribution, while it is multiplied by 1 for the tails of the log-normal distribution (38 / 40 = 95%) Finally, the defined number of fields is randomly drawn from the vector.
numeric vector of length n with randomly drawn values according to the distribution
Truncate normal distribution at a minimum or maximum value
rtnorm(n, mean, sd, a = -Inf, b = Inf)rtnorm(n, mean, sd, a = -Inf, b = Inf)
n |
Number of samples to be drawn |
mean |
Mean value of normal distribution |
sd |
Standard deviation of Normal distribution |
a |
Minimum of distribution (Default is -Inf) |
b |
Maximum of distribution (Default is Inf) |
numeric vector of length n with randomly drawn numbers of the truncated normal distribution
Plot cumulative sum of RQ after application and compare it with initial state
shadingPlot( mat_xRow, ymin = 0, ymax = max(mat_xRow), resolution = 0.01, xlab = "", ylab = "", main = "" )shadingPlot( mat_xRow, ymin = 0, ymax = max(mat_xRow), resolution = 0.01, xlab = "", ylab = "", main = "" )
mat_xRow |
A matrix where every row is a distribution for one x value |
ymin |
Minimum value of y-Axis |
ymax |
Maximumg value of y-Axis |
resolution |
The widths of quantiles used for shading (default is 1%) |
xlab, ylab, main
|
Óptional definitions of axis and title |
A plot with overlapping transperant quantiles
Calculation of the pollutant concentration in top soil
temp_c_profile(conti_input, output_rate, c_i, t_max, t_res, t_beg = 0)temp_c_profile(conti_input, output_rate, c_i, t_max, t_res, t_beg = 0)
conti_input |
Numeric vector of atmospheric depositions in mg/kg topsoil |
output_rate |
Numeric vector of absolute pollutant decay rates. |
c_i |
Numeric vector of initial concentrations in top soil |
t_max |
Numeric value with the last considered timestep (unit depends on the unit of the decay rate) |
t_res |
Temporal resolution (numeric value) of the timesteps returned (unit depends on the unit of the decay rate) |
t_beg |
Numeric value of the first considered timestep. This value does not influence the calculation itself, it is only used for row names. |
Matrix with columns corresponding to the length of the provided numeric vectors and rows depending on the defined timesteps with t_max and t_res