EelgrassFLWphenology.Rmd

---
title: "Eelgrass Flowering Phenology"
author: "Forest Schenck"
date: "5/25/2023"
output: pdf_document
editor_options: 
  chunk_output_type: console
---

#Packages
```{r}
library(tidyr)
library(dplyr)
library(lubridate)
library(ggplot2)
library(glue) #for plot function code; plot title function
library(purrr) #for plot function code; map function
#library(readxl) #install excel files

#GAMs
library(mgcv)

#Maps
library(maps)
library(mapdata)
library(mapproj)
```

#Data upload EG 2023
```{r}
d.flw.stage.2023 <- read.csv("C:/Users/Forest.Schenck/Downloads/eelgrass.flw.phnlgy.database.csv")

```

#Data upload EG 2024
```{r}
d.flw.stage.WHOI.2024 <- read.csv("C:/Users/Forest.Schenck/Desktop/FS Local FLW Data/whoi.eelgrass.flw.stage.data.csv")

d.flw.den.WHOI.2024 <- read.csv("C:/Users/Forest.Schenck/Desktop/FS Local FLW Data/whoi.eelgrass.density.data.csv")

d.flw.stage.RI.SNEP.2024 <- read.csv('C:/Users/Forest.Schenck/Desktop/FS Local FLW Data/RI.snep.eelgrass.flw.stage.data.master.csv')

d.flw.den.RI.SNEP.2024 <- read.csv("C:/Users/Forest.Schenck/Desktop/FS Local FLW Data/RI.snep.eelgrass.density.data.master.csv")

d.flw.stage.MA.SNEP.2024 <- read.csv('C:/Users/Forest.Schenck/Desktop/FS Local FLW Data/snep.eelgrass.flw.stage.data.csv')

d.flw.den.MA.SNEP.2024 <- read.csv('C:/Users/Forest.Schenck/Desktop/FS Local FLW Data/snep.eelgrass.density.data.csv')

#d.harvest.2024 <- read.csv("C:/Users/Forest.Schenck/Downloads/whoi.eelgrass.flw.harvest.data.csv")
```

#Data upload Hobos (temp/light) 2024
```{r}
d.t.AQ <- read.csv("C:/Users/Forest.Schenck/Desktop/FS Local FLW Data/Hobo/21177827_AQ.csv")

d.t.CB <- read.csv("C:/Users/Forest.Schenck/Desktop/FS Local FLW Data/Hobo/21258498_CB.csv")

d.t.CI <- read.csv("C:/Users/Forest.Schenck/Desktop/FS Local FLW Data/Hobo/20536645_CI.csv")

d.t.PC <- read.csv("C:/Users/Forest.Schenck/Desktop/FS Local FLW Data/Hobo/21177829_PC.csv")

d.t.DC <- read.csv("C:/Users/Forest.Schenck/Desktop/FS Local FLW Data/Hobo/21258490_DC.csv")

d.t.LH <- read.csv("C:/Users/Forest.Schenck/Desktop/FS Local FLW Data/Hobo/10942817_LH.csv")

d.t.SH <- read.csv("C:/Users/Forest.Schenck/Desktop/FS Local FLW Data/Hobo/21177828_SH.csv")

d.t.CC <- read.csv("C:/Users/Forest.Schenck/Desktop/FS Local FLW Data/Hobo/21982329_CC.csv")

d.t.WI <- read.csv("C:/Users/Forest.Schenck/Desktop/FS Local FLW Data/Hobo/21982330_WI.csv")

d.t.OB <- read.csv("C:/Users/Forest.Schenck/Desktop/FS Local FLW Data/Hobo/21982331_OB.csv")

d.t.WC <- read.csv("C:/Users/Forest.Schenck/Desktop/FS Local FLW Data/Hobo/21982332_WC.csv")

d.t.GB <- read.csv("C:/Users/Forest.Schenck/Desktop/FS Local FLW Data/Hobo/21982334_GB.csv")

```

#Phenology 2023
##data carpentry
```{r}
str(d.flw.stage.2023)

d.flw.stage.2023.sum <- d.flw.stage.2023 %>%
  mutate(date.collected = mdy(date.collected)) %>%
  group_by(Town, Subsite, date.collected) %>%
  summarise(spathe.stage.PS = sum(spathe.stage.PS, na.rm = T), spathe.stage.0 = sum(spathe.stage.0, na.rm = T), spathe.stage.1 = sum(spathe.stage.1, na.rm = T), spathe.stage.2 = sum(spathe.stage.2, na.rm = T), spathe.stage.3 = sum(spathe.stage.3, na.rm = T), spathe.stage.4 = sum(spathe.stage.4, na.rm = T), spathe.stage.5 = sum(spathe.stage.5, na.rm = T), spathe.stage.6 = sum(spathe.stage.6, na.rm = T)) %>%
  pivot_longer(!Town:date.collected, names_to = "flw.stage", values_to = "count") %>%
  mutate(flw.stage = as.factor(flw.stage)) %>%
  mutate(flw.stage = factor(flw.stage, levels=c ('spathe.stage.PS', 'spathe.stage.0', 'spathe.stage.1', 'spathe.stage.2', 'spathe.stage.3', 'spathe.stage.4', 'spathe.stage.5', 'spathe.stage.6')))

d.flw.stage.2023.sum = as.data.frame(d.flw.stage.2023.sum)


str(d.flw.stage.2023.sum)
```

##plot Bassett Island
```{r}
p <- ggplot(filter(d.flw.stage.2023.sum, Town == "Bourne" & Subsite == "A"), aes(x = date.collected, y = count, fill = flw.stage)) + 
  geom_col(position = "fill") +
  xlab("Date") +
  ylab("Proportion") +
  scale_fill_brewer() +
  theme_classic()

p

```

##plot Channel Shoal
```{r}
p <- ggplot(filter(d.flw.stage.2023.sum, Town == "Bourne" & Subsite == "B"), aes(x = date.collected, y = count, fill = flw.stage)) + 
  geom_col(position = "fill") +
  xlab("Date") +
  ylab("Proportion") +
  scale_fill_brewer() +
  theme_classic()

p
```

##plot Clark Island South
```{r}
p <- ggplot(filter(d.flw.stage.2023.sum, Town == "Duxbury"), aes(x = date.collected, y = count, fill = flw.stage)) + 
  geom_col(position = "fill") +
  xlab("Date") +
  ylab("Proportion") +
  scale_fill_brewer() +
  theme_classic()

p
```

##plot Annisquam River Mouth N
```{r}
p <- ggplot(filter(d.flw.stage.2023.sum, Town == "Gloucester"), aes(x = date.collected, y = count, fill = flw.stage)) + 
  geom_col(position = "fill") +
  xlab("Date") +
  ylab("Proportion") +
  scale_fill_brewer() +
  theme_classic()

p
```

#Production 2023
##data carpentry
```{r}
str(d.flw.stage)

d.seed.production <- d.flw.stage %>%
  mutate(date.collected = mdy(date.collected)) %>%
  mutate(seeds.counted = ifelse(Town == "Bourne" & Subsite == "A" & date.collected == "2023-06-06" & shoot.no <=5, "N", "Y")) %>%
  filter(seeds.counted == "Y") %>%
  filter(shoot.no != "NA") %>%
  group_by(Town, Subsite, date.collected, shoot.no) %>%
  summarise(no.seeds.sum = sum(no.seeds, na.rm = T))
  

d.seed.production= as.data.frame(d.seed.production)


str(d.seed.production)
```

##plot Bassett Island
```{r}
p <- ggplot(filter(d.seed.production, Town == "Bourne" & Subsite == "A"), aes(x = as.factor(date.collected), y = no.seeds.sum)) + 
  geom_boxplot() +
  xlab("Date") +
  ylab("no.seeds.per.shoot") +
  theme_classic()

p
```

##plot Channel Shoal
```{r}
p <- ggplot(filter(d.seed.production, Town == "Bourne" & Subsite == "B"), aes(x = as.factor(date.collected), y = no.seeds.sum)) + 
  geom_boxplot() +
  xlab("Date") +
  ylab("no.seeds.per.shoot") +
  theme_classic()

p
```

##plot Clark Island South
```{r}
p <- ggplot(filter(d.seed.production, Town == "Duxbury"), aes(x = as.factor(date.collected), y = no.seeds.sum)) + 
  geom_boxplot() +
  xlab("Date") +
  ylab("no.seeds.per.shoot") +
  theme_classic()

p
```

##plot Annisquam River Mouth N
```{r}
p <- ggplot(filter(d.seed.production, Town == "Gloucester"), aes(x = as.factor(date.collected), y = no.seeds.sum)) + 
  geom_boxplot() +
  geom_point() +
  stat_summary(geom = "point", fun = "mean", col = "red", size = 3, shape = 24, fill = "red") +
  xlab("Date") +
  ylab("no.seeds.per.shoot") +
  theme_classic()

p
```

#Phenology 2024
##data carpentry (prop)
```{r}
str(d.flw.stage.WHOI.2024)

d.flw.stage.RI.SNEP.2024 <- as.data.frame(d.flw.stage.RI.SNEP.2024) %>%
  mutate(spathe.stage.0 = as.integer(spathe.stage.0),
         site.id = if_else(is.na(site.id) == T, "NR", site.id))
str(d.flw.stage.RI.SNEP.2024)

d.flw.stage.MA.SNEP.2024 <- as.data.frame(d.flw.stage.MA.SNEP.2024) %>%
  mutate(no.seeds = as.numeric(no.seeds))
str(d.flw.stage.MA.SNEP.2024)


d.flw.stage.2024 <- dplyr::bind_rows(d.flw.stage.WHOI.2024, d.flw.stage.RI.SNEP.2024, d.flw.stage.MA.SNEP.2024)


d.flw.stage.2024.sum <- d.flw.stage.2024 %>%
  mutate(date.collected = mdy(date.collected, tz = "America/New_York")) %>%
  group_by(town, site.id, date.collected, shoot.no) %>%
  summarise(spathe.stage.PS = sum(spathe.stage.PS, na.rm = T), spathe.stage.0 = sum(spathe.stage.0, na.rm = T), spathe.stage.1 = sum(spathe.stage.1, na.rm = T), spathe.stage.2 = sum(spathe.stage.2, na.rm = T), spathe.stage.3 = sum(spathe.stage.3, na.rm = T), spathe.stage.4 = sum(spathe.stage.4, na.rm = T), spathe.stage.5 = sum(spathe.stage.5, na.rm = T), spathe.stage.6 = sum(spathe.stage.6, na.rm = T), spathe.stage.7 = sum(spathe.stage.7, na.rm = T), spathe.stage.unk = sum(spathe.stage.unk, na.rm = T)) %>%
  mutate(spathe.stage.0 = spathe.stage.PS + spathe.stage.0,
         spathe.stage.6 = spathe.stage.6 + spathe.stage.7,
         no.spathe = spathe.stage.0 + spathe.stage.1 + spathe.stage.2 + spathe.stage.3 + spathe.stage.4 + spathe.stage.5 + spathe.stage.6 + spathe.stage.unk,
         #stage.PS = spathe.stage.PS/no.spathe, #combined PS and 0
         stage.0 = spathe.stage.0/no.spathe,
         stage.1 = spathe.stage.1/no.spathe,
         stage.2 = spathe.stage.2/no.spathe,
         stage.3 = spathe.stage.3/no.spathe,
         stage.4 = spathe.stage.4/no.spathe,
         stage.5 = spathe.stage.5/no.spathe,
         stage.6 = spathe.stage.6/no.spathe,
         #stage.7 = spathe.stage.7/no.spathe, #combined PS and 0
         stage.unk = spathe.stage.unk/no.spathe) %>%
  dplyr::select(town:shoot.no, stage.0:stage.6, stage.unk) %>%
  pivot_longer(!town:shoot.no, names_to = "flw.stage", values_to = "prop") %>%
  mutate(flw.stage = as.factor(flw.stage),
         shoot.no = as.factor(shoot.no)) %>%
  mutate(flw.stage = factor(flw.stage, levels=c ('stage.0', 'stage.1', 'stage.2', 'stage.3', 'stage.4', 'stage.5', 'stage.6', 'stage.unk'))) %>%
  mutate(j = as.POSIXlt(date.collected, tz = "America/New_York")$yday + 1) #add 1 for leap year

d.flw.stage.2024.sum = as.data.frame(d.flw.stage.2024.sum)


str(d.flw.stage.2024.sum)
```


##Stage Proportion Plot Function (prop)
```{r}

stage.plot <- function(site.code) {
  d.flw.stage.2024.sum %>%
    filter(.data$site.id == .env$site.code) %>%
    ggplot() +
    aes(x = j, y = prop, color = flw.stage) +
    geom_point() +
    geom_smooth(se = FALSE) +
    xlab("Julian date") +
    ylab("Proportion") +
    coord_cartesian(xlim = c(100,300))+
    geom_hline(yintercept = 0.5, linetype = "dashed") +
    scale_x_continuous(breaks = c(121, 152, 182, 213, 244, 276)) +
    #scale_x_continuous(breaks = seq(100, 300, by = 10), minor_breaks = seq(100,300, by = 1)) +
    scale_colour_brewer(palette = "PuOr", direction = -1) +
    ggtitle(glue("Site: {site.code}")) +
    #theme(plot.title.position = "plot") +
    #theme(axis.ticks.length = unit(0.2, "cm"), axis.ticks.minor = element_line(size = 0.5, color = "grey")) #+
    theme_classic()
}

site.code <- c("SH", "CI", "PC", "LH", "PV", "AQ", "WB", "NB", "CB", "DC", "CC", "GB", "LI", "OB", "WC", "WI", "CL", "FG", "GI", "JB", "NR", "NK", "PI", "QU")

site.plots <- purrr::map(site.code, stage.plot)

site.plots[1:24]

```

##data carpentry (mean stage)
```{r}
str(d.flw.stage.2024)

d.flw.stage.2024.mean <- d.flw.stage.2024 %>%
  mutate(date.collected = mdy(date.collected, tz = "America/New_York")) %>%
  group_by(town, site.id, date.collected, shoot.no) %>%
  summarise(spathe.stage.PS = sum(spathe.stage.PS, na.rm = T), spathe.stage.0 = sum(spathe.stage.0, na.rm = T), spathe.stage.1 = sum(spathe.stage.1, na.rm = T), spathe.stage.2 = sum(spathe.stage.2, na.rm = T), spathe.stage.3 = sum(spathe.stage.3, na.rm = T), spathe.stage.4 = sum(spathe.stage.4, na.rm = T), spathe.stage.5 = sum(spathe.stage.5, na.rm = T), spathe.stage.6 = sum(spathe.stage.6, na.rm = T), spathe.stage.7 = sum(spathe.stage.7, na.rm = T), spathe.stage.unk = sum(spathe.stage.unk, na.rm = T)) %>%
  mutate(spathe.stage.0 = spathe.stage.PS + spathe.stage.0,
         spathe.stage.6 = spathe.stage.6 + spathe.stage.7,
         no.spathe = spathe.stage.0 + spathe.stage.1 + spathe.stage.2 + spathe.stage.3 + spathe.stage.4 + spathe.stage.5 + spathe.stage.6,
         stage.0 = (spathe.stage.0/no.spathe)*0,
         stage.1 = (spathe.stage.1/no.spathe)*1,
         stage.2 = (spathe.stage.2/no.spathe)*2,
         stage.3 = (spathe.stage.3/no.spathe)*3,
         stage.4 = (spathe.stage.4/no.spathe)*4,
         stage.5 = (spathe.stage.5/no.spathe)*5,
         stage.6 = (spathe.stage.6/no.spathe)*6,
         stage.sum = stage.0 + stage.1 + stage.2 + stage.3 + stage.4 +stage.5 + stage.6) %>%
  dplyr::select(town:shoot.no, stage.sum) %>%
  mutate(j = as.POSIXlt(date.collected, tz = "America/New_York")$yday + 1) #add 1 for leap year

d.flw.stage.2024.mean = as.data.frame(d.flw.stage.2024.mean)


str(d.flw.stage.2024.mean)
```

##Stage Proportion Plot Function (mean stage)
```{r}
p <-ggplot(data = d.flw.stage.2024.mean, aes(x = j, y = stage.sum, color = site.id)) +
  geom_point() +
  geom_smooth(se = FALSE) +
  xlab("Julian Day") +
  ylab("Mean Stage Value") +
  #scale_colour_brewer(palette = "PuOr") +
  scale_x_continuous(breaks = c(121, 152, 182, 213, 244, 276)) +
  geom_hline(yintercept = 4) +
  theme_classic()
  
p

#zoom plot
p <-ggplot(data = d.flw.stage.2024.mean, aes(x =j, y = stage.sum.mean, color = site.id)) +
  geom_point() +
  geom_smooth(se = FALSE) +
  xlab("Julian Date") +
  ylab("Mean Stage Value") +
  scale_colour_brewer(palette = "PuOr") +
  geom_hline(yintercept = 4) +
  coord_cartesian(ylim = c(3.5,4.5), xlim = c(190,225)) +
  scale_x_continuous(breaks = seq(190,230, by = 2)) +
  theme_classic()
p
```


#Seed Production Per Shoot 2024

##data carpentry
```{r}
str(d.flw.stage.2024)

d.seed.production.2024.sum <- d.flw.stage.2024 %>%
  mutate(date.collected = mdy(date.collected)) %>%
  rowwise() %>%
  mutate(no.seeds.total = sum(no.seeds, no.ovules.w.stigma, na.rm = T)) %>%
  group_by(town, site.id, date.collected, shoot.no) %>%
  summarise(no.seeds.total = sum(no.seeds.total, na.rm = T)) %>%
  mutate(j = as.POSIXlt(date.collected)$yday + 1) %>% #add 1 for leap year
  filter(is.na(shoot.no) == FALSE)

d.seed.production.2024.sum = as.data.frame(d.seed.production.2024.sum)


str(d.seed.production.2024.sum)
```

##Seed production per Shoot Plot Function
```{r}

seeds.per.shoot.plot <- function(site.code) {
  d.seed.production.2024.sum %>%
    filter(.data$site.id == .env$site.code) %>%
    ggplot() +
    aes(x = j, y = no.seeds.total) +
    geom_point() +
    geom_smooth(se = TRUE) +
    coord_cartesian(xlim = c(100,300)) +
    scale_x_continuous(breaks = c(121, 152, 182, 213, 244, 276)) +
    xlab("Date") +
    ylab("Seeds per shoot") +
    ggtitle(glue("Site: {site.code}")) +
    theme(plot.title.position = "plot") +
    theme_classic()
}

site.code <- c("SH", "CI", "PC", "LH", "PV", "AQ", "WB", "NB", "CB", "DC", "CC", "GB", "LI", "OB", "WC", "WI", "CL", "FG", "GI", "JB", "NR", "NK", "PI", "QU")

site.plots <- purrr::map(site.code, seeds.per.shoot.plot)

site.plots[1:24]



p <-ggplot(data = d.seed.production.2024.sum, aes(x = date.collected, y = no.seeds.total, color = site.id)) +
  geom_point() +
  geom_smooth(se = FALSE) +
  xlab("Date") +
  ylab("Seeds per shoot") +
  scale_colour_brewer(palette = "PuOr") +
  theme_classic()
  
p

```


#Flw Density 2024
##data carpentry
```{r}
str(d.flw.den.WHOI.2024)

d.flw.den.RI.SNEP.2024 <- as.data.frame(d.flw.den.RI.SNEP.2024) %>%
  mutate(site.id = if_else(is.na(site.id) == T, "NR", site.id))
str(d.flw.den.RI.SNEP.2024)

d.flw.den.MA.SNEP.2024 <- as.data.frame(d.flw.den.MA.SNEP.2024)
str(d.flw.den.MA.SNEP.2024)

d.flw.den.2024 <- dplyr::bind_rows(d.flw.den.WHOI.2024, d.flw.den.RI.SNEP.2024, d.flw.den.MA.SNEP.2024)

d.flw.den.2024.clean <- d.flw.den.2024 %>%
  filter(is.na(shoot.type) == FALSE) %>%
  mutate(date.collected = mdy(date.collected)) %>%
  select(date.collected, site.id, quadrat.no, shoot.type, shoot.no) %>%
  spread(key = shoot.type, value = shoot.no) %>%
  filter(reproductive > 0 | vegetative > 0) %>% #only plots where eelgrass is present
  mutate(flw.m.sq = reproductive * 4)

```

##Flw Density Plot Function
```{r}
p <-ggplot(data = d.flw.den.2024.clean, aes(x = date.collected, y = flw.m.sq, color = site.id)) +
  geom_point() +
  geom_smooth(se = FALSE) +
  xlab("Date") +
  ylab("Repros per m2") +
  coord_cartesian(xlim = c(100,300)) +
  scale_x_continuous(breaks = c(121, 152, 182, 213, 244, 276)) +
  #scale_colour_brewer(palette = "PuOr") +
  theme_classic()
  
p
```


#Seed collection estimates 2024
## data carpentry
```{r}

str(d.harvest.2024)

d.harvest.2024.totals <- d.harvest.2024 %>%
  mutate(date.collected = mdy(date.collected)) %>%
  group_by(date.collected, site.id) %>%
  summarise(no.repro.shoots = sum(no.repro.shoots, na.rm = T))

d.seed.production.2024.mean <- d.seed.production.2024.sum %>%
  group_by(date.collected, site.id) %>%
  summarise(no.seeds.mean = mean(no.seeds.total, na.rm = T))

d.seed.harvest.2024 <- merge(d.seed.production.2024.mean, d.harvest.2024.totals, by = c("date.collected", "site.id")) %>%
  mutate(estimated.seeds.collected = no.seeds.mean * no.repro.shoots)



```

#Hobo Temp and Light 2024
##Data carpentry
```{r}
d.t.AQ <- d.t.AQ %>%
  mutate(site.id = "AQ")

d.t.CB <- d.t.CB %>%
  mutate(site.id = "CB")

d.t.CI <- d.t.CI %>%
  mutate(site.id = "CI")

d.t.DC <- d.t.DC %>%
  mutate(site.id = "DC")

d.t.LH <- d.t.LH %>%
  mutate(site.id = "LH")

d.t.PC <- d.t.PC %>%
  mutate(site.id = "PC")

d.t.SH <- d.t.SH %>%
  mutate(site.id = "SH")

d.t.CC <- d.t.CC %>%
  mutate(site.id = "CC")

d.t.GB <- d.t.GB %>%
  mutate(site.id = "GB")

d.t.OB <- d.t.OB %>%
  mutate(site.id = "OB")

d.t.WC <- d.t.WC %>%
  mutate(site.id = "WC")

d.t.WI <- d.t.WI %>%
  mutate(site.id = "WI")

d.t.2024 <- bind_rows(d.t.AQ, d.t.CB, d.t.CI, d.t.DC, d.t.LH, d.t.PC, d.t.SH, d.t.CC, d.t.GB, d.t.OB, d.t.WC, d.t.WI) %>%
  mutate(date.time = mdy_hm(date.time)) %>%
  filter(date.time > "2024-05-15 19:45") %>%
  filter(date.time < "2024-09-29 20:00") %>%
  mutate(j = as.POSIXlt(date.time)$yday + 1) %>% #calculating julian day, add 1 for leap year
  rename(light = Intensity.lum.per.sqft)

str(d.t.2024)

GDD.2024 <- d.t.2024 %>%
  group_by(site.id, j) %>%
  summarise(max.t = max(temp.C), min.t = min(temp.C), sum.light = sum(light)) %>%
  mutate(GDD = (max.t + min.t)/2 - 8) %>%
  group_by(site.id) %>% 
  mutate(GDD.sum = cumsum(GDD),
         light.sum = cumsum(sum.light))
```

##Plots
```{r}
  
p <-ggplot(data = GDD.2024, aes(x = j, y = GDD.sum, color = site.id)) +
  geom_point() +
  geom_smooth(se = FALSE) +
  xlab("Julian Day") +
  ylab("GDD") +
  coord_cartesian(xlim = c(100,300)) +
  scale_x_continuous(breaks = c(121, 152, 182, 213, 244, 276)) +
  #scale_colour_brewer(palette = "PuOr") +
  theme_classic()
p

p <-ggplot(data = GDD.2024, aes(x = j, y = light.sum, color = site.id)) +
  geom_point() +
  geom_smooth(se = FALSE) +
  xlab("Julian Day") +
  ylab("Light") +
  coord_cartesian(xlim = c(100,300)) +
  scale_x_continuous(breaks = c(121, 152, 182, 213, 244, 276)) +
  #scale_colour_brewer(palette = "PuOr") +
  theme_classic()
  
p

```


#Seed Density 2024
##Data Carpentry
```{r}

#Summarise to mean reproductive shoot density in plots where eelgrass present for each monitoring date
d.flw.den.2024.sd <- d.flw.den.2024 %>%
  mutate(date.collected = mdy(date.collected)) %>%
  select(date.collected, site.id, latitude, longitude, quadrat.no, shoot.type, shoot.no) %>%
  spread(key = shoot.type, value = shoot.no) %>%
  filter(reproductive > 0 | vegetative > 0) %>% #only plots where eelgrass is present
  mutate(flw.m.sq = reproductive * 4) %>%
  group_by(date.collected, site.id, latitude, longitude) %>%
  summarise(mean.flw.m.sq = mean(flw.m.sq))

#Summarise to mean seeds per flw for each monitoring date
d.seed.production.2024.sd <- d.seed.production.2024.sum %>%
  filter(!is.na(shoot.no)) %>%
  group_by(site.id, date.collected) %>%
  summarise(mean.seeds = mean(no.seeds.total))

#merge flw density and seed data sets, include dates where density measured but no seeds
d.seed.den.2024 <- merge(d.flw.den.2024.sd, d.seed.production.2024.sd, by = c("site.id","date.collected"), all.x = T)

#Cleaning up merged seed production data set
d.seed.den.2024.f <- d.seed.den.2024 %>%
  mutate(mean.seeds = replace(mean.seeds, is.na(mean.seeds), 0)) %>% #replacing dates with no seeds to 0
  mutate(seeds.m.sq = mean.flw.m.sq*mean.seeds, #calculating seeds per m2
         j = as.POSIXlt(date.collected)$yday + 1) #calculating julian day, add 1 for leap year

d.seed.den.2024.GDD <- left_join(d.seed.den.2024.f, GDD.2024, by = c("site.id", "j")) #Adding GDD to seed production data set

d.seed.den.2024.GDD <- d.seed.den.2024.GDD %>%
  filter(!is.na(GDD.sum))
  

```

##Plot by Date
```{r}
p <-ggplot(data = d.seed.den.2024.f, aes(x = j, y = seeds.m.sq, color = site.id)) +
  geom_point() +
  geom_smooth(se = FALSE) +
  xlab("Date") +
  ylab("Seeds per m2") +
  coord_cartesian(xlim = c(100,300)) +
  scale_x_continuous(breaks = c(121, 152, 182, 213, 244, 276)) +
  #scale_colour_brewer(palette = "PuOr") +
  #scale_y_continuous(trans = 'log10') +
  theme_classic()
  
p

seeds.per.m2.plot <- function(site.code) {
  d.seed.den.2024.f %>%
    filter(.data$site.id == .env$site.code) %>%
    ggplot() +
    aes(x = j, y = seeds.m.sq) +
    geom_point() +
    geom_smooth(se = TRUE) +
    coord_cartesian(xlim = c(100,300)) +
    scale_x_continuous(breaks = c(121, 152, 182, 213, 244, 276)) +
    xlab("Date") +
    ylab("Seeds per m2") +
    ggtitle(glue("Site: {site.code}")) +
    theme(plot.title.position = "plot") +
    theme_classic()
}

site.code <- c("SH", "CI", "PC", "LH", "PV", "AQ", "WB", "NB", "CB", "DC", "CC", "GB", "LI", "OB", "WC", "WI", "CL", "FG", "GI", "JB", "NR", "NK", "PI", "QU")

site.plots <- purrr::map(site.code, seeds.per.m2.plot)

site.plots[1:24]




p <-ggplot(data = d.seed.den.2024.f, aes(x =j, y = seeds.m.sq+1, color = site.id)) +
  geom_point() +
  geom_smooth(se = FALSE) +
  xlab("Julian Date") +
  ylab("Mean Stage Value") +
  scale_colour_brewer(palette = "PuOr") +
  geom_hline(yintercept = 4) +
  coord_cartesian(ylim = c(90,7000), xlim = c(170,240)) +
  scale_y_continuous(trans = 'log10') +
  scale_x_continuous(breaks = seq(170,240, by = 2)) +
  theme_classic()
p

```

##Plot by GDD
```{r}
p <-ggplot(data = d.seed.den.2024.GDD, aes(x = GDD.sum, y = seeds.m.sq+1, color = site.id)) +
  geom_point() +
  geom_smooth(se = FALSE) +
  xlab("GDD") +
  ylab("Seeds per m2") +
  #scale_colour_brewer(palette = "PuOr") +
  #scale_y_continuous(trans = 'log10') +
  theme_classic()
  
p

#to compare same sites but plotting by date
p <-ggplot(data = d.seed.den.2024.GDD, aes(x = j, y = seeds.m.sq, color = site.id)) +
  geom_point() +
  geom_smooth(se = FALSE) +
  xlab("Date") +
  ylab("Seeds per m2") +
  coord_cartesian(xlim = c(100,300)) +
  scale_x_continuous(breaks = c(121, 152, 182, 213, 244, 276)) +
  #scale_colour_brewer(palette = "PuOr") +
  #scale_y_continuous(trans = 'log10') +
  theme_classic()
  
p

```

##Plot stage prop by GDD
```{r}
d.flw.stage.2024.mean.GDD <- left_join(d.flw.stage.2024.mean, GDD.2024, by = c("site.id", "j"))

d.flw.stage.2024.mean.GDD <- d.flw.stage.2024.mean.GDD %>%
  filter(!is.na(GDD.sum))

p <-ggplot(data = d.flw.stage.2024.mean.GDD, aes(x = GDD.sum, y = stage.sum, color = site.id)) +
  geom_point() +
  geom_smooth(se = FALSE) +
  xlab("GDD") +
  ylab("Mean Stage Value") +
  #scale_colour_brewer(palette = "PuOr") +
  geom_hline(yintercept = 4) +
  theme_classic()
  
p


p <-ggplot(data = d.flw.stage.2024.mean.GDD, aes(x = j, y = stage.sum, color = site.id)) +
  geom_point() +
  geom_smooth(se = FALSE) +
  xlab("Julian Day") +
  ylab("Mean Stage Value") +
  #scale_colour_brewer(palette = "PuOr") +
  coord_cartesian(xlim = c(100,300)) +
  scale_x_continuous(breaks = c(121, 152, 182, 213, 244, 276)) +
  geom_hline(yintercept = 4) +
  theme_classic()
  
p
```

#Relative Seed Abundance 2024
##Data Carpentry
```{r}
str(d.seed.den.2024.f)

d.max.seed.2024 <- d.seed.den.2024.f %>%
  group_by(site.id) %>%
  summarise(max.seeds.m.sq = max(seeds.m.sq, na.rm = T))

d.RSA.2024 <- left_join(d.seed.den.2024.f, d.max.seed.2024, by = c("site.id")) %>% #Adding GDD to seed production data set
  mutate(RSA = seeds.m.sq/max.seeds.m.sq)

d.RSA.2024.GDD <- left_join(d.RSA.2024, GDD.2024, by = c("site.id", "j")) #Adding GDD to seed production data set

d.RSA.2024.GDD <- d.RSA.2024.GDD %>%
  filter(!is.na(GDD.sum))

```
##Plot by Date
```{r}

p <-ggplot(data = d.RSA.2024, aes(x = j, y = RSA, color = site.id)) +
  geom_point() +
  geom_line() +
  #geom_smooth(se = FALSE) +
  xlab("Date") +
  ylab("Relative Seed Abundance") +
  coord_cartesian(xlim = c(100,300)) +
  scale_x_continuous(breaks = c(121, 152, 182, 213, 244, 276)) +
  #scale_colour_brewer(palette = "PuOr") +
  #scale_y_continuous(trans = 'log10') +
  theme_classic()
  
p

RSA.plot <- function(site.code) {
  d.RSA.2024 %>%
    filter(.data$site.id == .env$site.code) %>%
    ggplot() +
    aes(x = j, y = RSA) +
    geom_point() +
    geom_line() +
    #geom_smooth(se = FALSE) +
    coord_cartesian(xlim = c(100,300)) +
    scale_x_continuous(breaks = c(121, 152, 182, 213, 244, 276)) +
    xlab("Date") +
    ylab("Relative Seed Abundance") +
    ggtitle(glue("Site: {site.code}")) +
    theme(plot.title.position = "plot") +
    theme_classic()
}

site.code <- c("SH", "CI", "PC", "LH", "PV", "AQ", "WB", "NB", "CB", "DC", "CC", "GB", "LI", "OB", "WC", "WI", "CL", "FG", "GI", "JB", "NR", "NK", "PI", "QU")

site.plots <- purrr::map(site.code, RSA.plot)

site.plots[1:24]
```
##Plot by GDD
```{r}
p <-ggplot(data = d.RSA.2024.GDD, aes(x = GDD.sum, y = RSA, color = site.id)) +
  geom_point() +
  geom_line() +
  #geom_smooth(se = FALSE) +
  xlab("GDD") +
  ylab("Relative Seed Abundance") +
  #scale_colour_brewer(palette = "PuOr") +
  #scale_y_continuous(trans = 'log10') +
  theme_classic()
  
p
```


#GAM Trials Seed Density
##GAM J-day
```{r}
# SELECT AN ERROR DISTRIBUTION
  plot(density(d.seed.den.2024.f$seeds.m.sq))

  m1<-gam(seeds.m.sq~1, data=d.seed.den.2024.f)   
  m2<-gam(seeds.m.sq~1, data=d.seed.den.2024.f, family='nb') #NEGATIVE BINOMIAL
  m3<-gam(seeds.m.sq~1, data=d.seed.den.2024.f, family='tw') #TWEEDIE
  AIC(m1,m2,m3)
  fam<-'nb'

  
#GAM seeds by date
mod.j <- gam(seeds.m.sq~s(j, bs = 'cc'), data=d.seed.den.2024.f, family=fam, select=TRUE, method="REML") 
summary(mod.j)
plot(mod.j,
     xlim = c(127,295))

#GAM seeds by latitude
mod.lat <- gam(seeds.m.sq ~ s(latitude), data = d.seed.den.2024.f, family = fam, select = TRUE, method = "REML")
summary(mod.lat)
plot(mod.lat)

#GAM seeds by latitude and date
mod.j.lat <- gam(seeds.m.sq~te(j, latitude, bs = c('cc', 'ts'))
                 , data = d.seed.den.2024.f, family = fam, select = TRUE, method = "REML")

summary(mod.j.lat)
plot(mod.j.lat)

```

##GAM with GDD and total light
```{r}
# SELECT AN ERROR DISTRIBUTION
  plot(density(d.seed.den.2024.GDD$seeds.m.sq))

  m1<-gam(seeds.m.sq~1, data=d.seed.den.2024.GDD)   
  m2<-gam(seeds.m.sq~1, data=d.seed.den.2024.GDD, family='nb') #NEGATIVE BINOMIAL
  m3<-gam(seeds.m.sq~1, data=d.seed.den.2024.GDD, family='tw') #TWEEDIE
  AIC(m1,m2,m3)
  fam<-'nb'

  
#GAM seeds by date
mod.j <- gam(seeds.m.sq~s(j, bs = 'cc'), data=d.seed.den.2024.GDD, family=fam, select=TRUE, method="REML") 
summary(mod.j)
plot(mod.j,
     xlim = c(140,260))

#GAM seeds by latitude
mod.lat <- gam(seeds.m.sq ~ s(latitude, k = 5), data = d.seed.den.2024.GDD, family = fam, select = TRUE, method = "REML")
summary(mod.lat)
plot(mod.lat)

#GAM seeds by GDD
mod.GDD <- gam(seeds.m.sq ~ s(GDD.sum), data = d.seed.den.2024.GDD, family = fam, select = TRUE, method = "REML")
summary(mod.GDD)
plot(mod.GDD)

#GAM seeds by light
mod.light <- gam(seeds.m.sq ~ s(light.sum), data = d.seed.den.2024.GDD, family = fam, select = TRUE, method = "REML")
summary(mod.light)
plot(mod.light)

#GAM seeds by GDD, light independent
mod.GDD.light.ind<-gam(seeds.m.sq~s(GDD.sum) #GDD
               +s(light.sum) #Light
               ,data=d.seed.den.2024.GDD,family=fam,select=TRUE,method="REML")
summary(mod.GDD.light.ind)
plot(mod.GDD.light.ind)

#GAM seeds by GDD x light interaction
mod.GDD.light <- gam(seeds.m.sq ~ te(GDD.sum, light.sum), data = d.seed.den.2024.GDD, family = fam, select = TRUE, method = "REML")
summary(mod.GDD.light)
plot(mod.GDD.light)


#GAM Full GDD, light model
mod.full<-gam(seeds.m.sq~s(GDD.sum) #GDD
              +s(light.sum) #Light
              +s(j, bs = 'cc') #julian day
              +s(latitude, k = 5)
              +te(GDD.sum, light.sum) #light X GDD interaction
               ,data=d.seed.den.2024.GDD,family=fam,select=TRUE,method="REML")
summary(mod.full)
plot(mod.full)
qq.gam(mod.full) #IS THE ERROR DISTRIBUTION APPROPRIATE?
gam.check(mod.full) #CHECK TO SEE IF SMOOTHERS HAVE BEEN GIVEN SUFFICIENT EDF


```

#GAM Trials Stage Proportion
##GAM with GDD
```{r}
d.flw.stage.2024.GAM <- d.flw.stage.2024.mean.GDD %>%
  group_by(site.id, j) %>%
  summarise(stage.sum = mean(stage.sum), GDD.sum = mean(GDD.sum), light.sum = mean(light.sum))

# SELECT AN ERROR DISTRIBUTION
  plot(density(d.flw.stage.2024.GAM$stage.sum))

  m1<-gam(stage.sum~1, data=d.flw.stage.2024.GAM)   
  m2<-gam(stage.sum~1, data=d.flw.stage.2024.GAM, family='nb') #NEGATIVE BINOMIAL
  m3<-gam(stage.sum~1, data=d.flw.stage.2024.GAM, family='tw') #TWEEDIE
  AIC(m1,m2,m3)

#GAM stage by GDD
mod.GDD <- gam(stage.sum ~ s(GDD.sum), data = d.flw.stage.2024.GAM, select = TRUE, method = "REML")
summary(mod.GDD)
plot(mod.GDD)

qq.gam(mod.GDD) #IS THE ERROR DISTRIBUTION APPROPRIATE?
gam.check(mod.GDD) #CHECK TO SEE IF SMOOTHERS HAVE BEEN GIVEN SUFFICIENT EDF

#GAM stage by j-day
mod.j <- gam(stage.sum ~ s(j, bs = 'cc'), data = d.flw.stage.2024.GAM, select = TRUE, method = "REML")
summary(mod.j)
plot(mod.j,
     xlim = c(148,256))

#GAM stage by light
mod.light <- gam(stage.sum ~ s(light.sum), data = d.flw.stage.2024.GAM, select = TRUE, method = "REML")
summary(mod.light)
plot(mod.light)


#GAM stage by GDD x light interaction
mod.GDD.light <- gam(stage.sum ~ te(GDD.sum, light.sum), data = d.flw.stage.2024.GAM, select = TRUE, method = "REML")
summary(mod.GDD.light)
plot(mod.GDD.light)

#GAM Full GDD, light model
mod.full<-gam(stage.sum~s(GDD.sum) #GDD
              +s(light.sum) #Light
              +s(j, bs = 'cc') #julian day
              +te(GDD.sum, light.sum) #light X GDD interaction
               ,data=d.flw.stage.2024.GAM,select=TRUE,method="REML")
summary(mod.full)
plot(mod.full)
qq.gam(mod.full) #IS THE ERROR DISTRIBUTION APPROPRIATE?
gam.check(mod.full) #CHECK TO SEE IF SMOOTHERS HAVE BEEN GIVEN SUFFICIENT EDF
```

#Site Map
```{r}
setwd("C:/Users/Forest.Schenck/Desktop")


tiff(filename = "site.map.tiff", height = 5600, width = 5200, units = "px", res = 800, compression = "lzw")
map("worldHires", "usa", xlim = c(-71.85, -69.95), ylim = c(41.25, 42.729), col = "grey", fill = T, lwd = 0.01)
points(mapproject(list(x = -70.85289, y = 42.53003)), pch = 19, cex = 1, col = "black") #AQ
points(mapproject(list(x = -70.80848, y = 42.55635)), pch = 19, cex = 1, col = "black") #WB
points(mapproject(list(x = -70.68275, y = 42.65720)), pch = 19, cex = 1, col = "black") #CB
points(mapproject(list(x = -70.67422, y = 42.66801)), pch = 19, cex = 1, col = "black") #DC
points(mapproject(list(x = -70.68233, y = 41.98567)), pch = 19, cex = 1, col = "black") #PC "lightblue"
points(mapproject(list(x = -70.61506, y = 42.00343)), pch = 19, cex = 1, col = "black") #SH "darkgreen"
points(mapproject(list(x = -70.64148, y = 42.00146)), pch = 19, cex = 1, col = "black") #CI
points(mapproject(list(x = -70.95276, y = 42.43701)), pch = 19, cex = 1, col = "black") #LH
points(mapproject(list(x = -70.65669, y = 42.59809)), pch = 19, cex = 1, col = "black") #NB
points(mapproject(list(x = -70.91592, y = 42.42002)), pch = 19, cex = 1, col = "black") #PI
points(mapproject(list(x = -71.72232, y = 41.34719)), pch = 19, cex = 1, col = "black") #Shady Harbor
points(mapproject(list(x = -71.63434, y = 41.36879)), pch = 19, cex = 1, col = "black") #The Arches
points(mapproject(list(x = -71.65471, y = 41.36441)), pch = 19, cex = 1, col = "black") #Ninigret Kayak
points(mapproject(list(x = -71.516065, y = 41.380007)), pch = 19, cex = 1, col = "black") #Coastal Field Lab
points(mapproject(list(x = -71.50799, y = 41.38976)), pch = 19, cex = 1, col = "black") #Great Island
points(mapproject(list(x = -71.36642, y = 41.49758)), pch = 19, cex = 1, col = "black") #Conanicus
points(mapproject(list(x = -71.39762, y = 41.49342)), pch = 19, cex = 1, col = "black") #Fort Getty
points(mapproject(list(x = -71.320328, y = 41.581383)), pch = 19, cex = 1, col = "black") #Prudence Island
points(mapproject(list(x = -70.657151, y = 41.741083)), pch = 19, cex = 1, col = "black") #Onset
points(mapproject(list(x = -70.673056, y = 41.524995)), pch = 19, cex = 1, col = "black") #Garbage Beach
points(mapproject(list(x = -70.647106, y = 41.690511)), pch = 19, cex = 1, col = "black") #Wings Cove
points(mapproject(list(x = -70.911218, y = 41.595157)), pch = 19, cex = 1, col = "black") #Clarks Cove
points(mapproject(list(x = -70.824288, y = 41.582794)), pch = 19, cex = 1, col = "black") # West Island
points(mapproject(list(x = -70.648689, y = 41.611945)), pch = 19, cex = 1, col = "black") #Little Island
points(mapproject(list(x = -70.55405, y = 41.44615)), pch = 19, cex = 1, col = "black") #Hart Haven
points(mapproject(list(x = -70.6254, y = 41.45903)), pch = 19, cex = 1, col = "black") #Lake Tashmoo
points(mapproject(list(x = -70.07658, y = 41.28507)), pch = 19, cex = 1, col = "black") #Monomoy
points(mapproject(list(x = -69.99902, y = 41.33083)), pch = 19, cex = 1, col = "black") #Wauwinet
points(mapproject(list(x = -69.9558412, y = 41.7516166)), pch = 19, cex = 1, col = "black") #Pleasant Bay
dev.off()
```