flowerGAM.R

In ecodata-technology/ecodatamisc: EcoData Technologies

# An attempt at flowering over time plots using generalized additive models to fit data

# Packages
library("readr")
library("tidyverse")
library("mgcv")
library("lubridate")
library("tidymv")
library("ecodatamisc")

# import and manipulate
fc_dat <- read_csv("./Data/Output/Flower_Cherelle.csv") %>%
  mutate(julian_day = yday(Observation_Date), .after = "Observation_Date") %>%
  mutate(across(Year, ~as.factor(.x)),
         y1 = Total_Flowers_Above_2m + Total_Flowers_Under_2m)

# select 2018 only
fc_2018 <- fc_dat %>%
  filter(Year == "2018")

# fit gam
gam2018 <- gam(y1 ~ s(julian_day), data = fc_2018, method = "REML")
summary(gam2018)

# plot with plot_gam
gamplot(gam2018, ylab = "Total Flowers", xlab = "Julian Day (2018)", ylims = c(0, 400))


#### Pollination Timing ####

hp_dat <- read_csv("./Data/Output/Hand_Pollination.csv") %>%
  filter(Pollination_Timing %in% c("early", "middle", "late"))

hp_dat%>%
  group_by(Year, Pollination_Week) %>%
  summarize(date_min = yday(min(Pollination_Date)),
            date_max = yday(max(Pollination_Date))) %>%
  mutate(Pol_Timing = recode(Pollination_Week,
                             Week_1 = "early", Week_2 = "middle", Week_3 = "late"),
         .after = "Pollination_Week")