R/spline_uni.R

In ldbk/Rplanktonanalytic: What the Package Does (One Line, Title Case)

#' spline_uni
#'
#'Identification of annual bloom descriptors dates based on smoothing splines. 
#'The function automatically exludes those years containing less than  75% of
#'the possible observations for the temporal scale considered.
#' 
#' @param x a dataframe having the first 7 columns named as "location", "station", "date", "year", "month", "week", "day",
#' other columns corresponds to the species columns
#' @param ab_treshold quantile of the positive abundance distribution (default 0.75)
#' @param obs_year minimum number of samples for each year in which the species shows an abundance>0 (default 2)
#' @param s_param the smoothing parameter corresponding to the spar parameters
#' @param t_scale temporal scale resolution of the data, 52 for weekly data, 12 for monthly data (default 52)
#' of the smooth.spline function (default 0.35).
#' @param S.NAME name of the species column
#' 
#' @return a dataframe containing the annual time vector, the value vector and info
#' a character vector with 'Start', 'Max' and 'End' for the seasonal peak and NA
#' for the other date
#' @examples
#' \dontrun{
#'	  data("phytopknar")
#'    phytopknar_ret <- ret_time(phytopknar)
#'    phytopknar_ret_ord <- phytopknar_ret %>% dplyr::select(location,station,date,year,month,week,day,everything())
#'    spline_uni(phytopknar_ret_ord,ab_treshold=0.75,obs_year=2,s_param=0.35,t_scale=52,S.NAME="Cylindrotheca closterium")
#' }
#' @export
#' 
spline_uni<- function(x,ab_treshold=0.75,obs_year=2,s_param=0.35,t_scale=52,S.NAME=""){
  ## 1: remove NAs keep robust years
  x <- x %>% select(1:7,S.NAME)
  colnames(x)[ncol(x)]="value"
  df1 <- na.omit(x)
  ## 2: keep years containing N>=75% of possible observations
  ok_years <- 
  df1  %>% group_by(year) %>% tally() %>% filter(n>(t_scale*0.75)) %>% pull(year) 
  df2 <- df1 %>% filter(year %in% ok_years)
  ## 3: consider only years in which species has/have at least X observations > 0 
  year_ob <- 
  df2  %>% filter(value>0) %>% group_by(year) %>%
  tally() %>% filter(n>obs_year) %>% pull(year) 
  df3 <- df2 %>% filter(year %in% year_ob) %>% droplevels()
  y_levels <- unique(df3$year) 
  ## 4: calculate bloom abundance treshold 
  tre <- df3 %>% filter(value>0) %>% select(value) 
  tre <- quantile(tre$value,ab_treshold)
  
  #### loop spline_point for each year
  df_emp <- data.frame()
  for ( i in y_levels ) {
    
    dft <- df3 %>% filter(year==i)
    dfn <- spline_points(dft$value,dft$day,s_param=s_param)
    
    dfn$year=rep(i,nrow(dfn))
    dfn$species=rep(S.NAME,nrow(dfn))
    dfn$treshold=rep(tre,nrow(dfn))
    
    dfn1 <- dfn %>% select(species,year,time,value,treshold,everything())
    df_emp <- rbind(df_emp,dfn1) %>% as.data.frame()
  }
  return(df_emp)
}