03-Data.Rmd

```{r setupinds, include=FALSE}
rm(list = ls()) ; invisible(gc()) ; set.seed(42)
library(knitr)
library(tidyverse)
library(V.PhyloMaker)
library(ggfortify)
library(ggtree)
library(sf)
library(leaflet)
library(readxl)
theme_set(bayesplot::theme_default())
opts_chunk$set(
  echo = F, message = F, warning = F, fig.height = 6, fig.width = 8,
  cache = T, cache.lazy = F)
crs <- '+proj=longlat +datum=WGS84' # leaflet CRS
```

# Field data

This chapter define sampled species and individuals. 
Specifically, this chapter details:

* Collected species
* Individuals from the METRADICA project
* Candidates with varying DBH and TWI
* Field maps

## Species

We selected 10 species (Tab. \@ref(tab:speciesTab)) common with the METRADICA project representative of the tree phylogeny in Paracou (Fig. \@ref(fig:speciesPhylo)).

```{r speciesTab}
species <- data.frame(taxon = c(
  "Casearia_javitensis", 
  "Chrysophyllum_prieurii", 
  "Conceveiba_guianensis",
  "Gustavia_hexapetala",
  "Jacaranda_copaia subsp. copaia",
  "Laetia_procera",
  "Protium_stevensonii",
  "Tachigali_melinonii",
  "Virola_michelii",
  "Virola_surinamensis"
)) %>% 
  separate(taxon, c("Genus", "Species"), sep = "_", remove = F)
kable(select(species, Genus, Species), caption = "Studied species.")
```

```{r speciesPhylo, fig.cap="Selected species phylogeny."}
paracou <- read_csv2("data/Paracou_P16_2020.csv") %>% 
  dplyr::select(Family, Genus, Species) %>% 
  unique() %>% 
  full_join(species) %>% 
  mutate(species = paste(Genus, Species), genus = Genus, family = Family) %>% 
  dplyr::select(species, genus, family, taxon) %>% 
  mutate(taxon = as.character(taxon))
# tree <- phylo.maker(sp.list = paracou, tree = GBOTB.extended, nodes = nodes.info.1, scenarios = "S3")
# save(tree, file = "save/phylogeny.Rdata")
load("save/phylogeny.Rdata")
fortify(tree$scenario.3) %>% 
  mutate(species = gsub("_", " ", label)) %>% 
  left_join(paracou) %>% 
  mutate(species = ifelse(is.na(taxon), NA, species)) %>% 
  separate(species, c("genus", "species")) %>% 
  mutate(species = ifelse(is.na(taxon), NA, paste0(substr(genus, 1, 1), ". ", species))) %>% 
  ggtree(aes(col = species), layout="circular") + 
  geom_tiplab2(aes(label = species), size = 3) +
  theme_tree(legend.position='right', legend.text = element_text(face = "italic")) +
  scale_color_discrete(guide = "none") +
  xlim(0, 175)
```

## METRADICA

This subparagraph details collected or wrong individuals in METRADICA (Tab. \@ref(tab:indMD)) and Marion's candidates (Tab. \@ref(tab:indMD), & Fig. \@ref(fig:indMDP16)),
but candidates of the project will be focused on P16 independently from her candidates and adjusted in the field. 

```{r indMD}
collected <- readxl::read_excel("data/202012_SelectionIndividus_V5.4.xlsx", sheet = "IndAMesurer") %>% 
  filter(TaxonActu %in% species$taxon) %>% 
  mutate(Statut = recode(Statut,
                         "AMesurerMD" = "Candidate",
                         "AMesurerMDPrioDK" = "Candidate",
                         "MesuréDK" = "Measured",
                         "MesuréDrought" = "Measured",
                         "MesuréMD" = "Measured",
                         "MesuréMD (MauvaiseSp)" = "Wrong",
                         "MesuréMD+Drought" = "Measured",
                         "MortMD" = "Wrong",
                         "MortParacou20" = "Wrong",
                         "MortParacou20PrioDK" = "Wrong",
                         "PasFaisable" = "Wrong",
                         ))
collected %>% 
  group_by(TaxonActu, Statut) %>% 
  summarise(N = n()) %>% 
  reshape2::dcast(TaxonActu ~ Statut, value.var = "N") %>% 
  mutate(TaxonActu = gsub("_", " ", TaxonActu)) %>% 
  kable(caption = "Sumary of individuals (from 202012_SelectionIndividus_V5.4.xlsx)")
```

```{r indMB}
tocollect <- readxl::read_excel("data/FTH2021_SelectionIndividus_V5.5.xlsx", sheet = "IndAMesurer") %>% 
  filter(TaxonActu %in% species$taxon) %>% 
  st_as_sf(coords = c("Xutm", "Yutm"),
           crs = '+proj=utm +zone=22 +datum=WGS84 +units=m +no_defs +ellps=WGS84 +towgs84=0,0,0') %>% 
  mutate(DBH = Circ/pi)
tocollect %>% 
  group_by(TaxonActu, Plot) %>% 
  summarise(N = n()) %>% 
  reshape2::dcast(TaxonActu ~ Plot, value.var = "N") %>% 
  mutate_all(funs(ifelse(is.na(.), " ", .))) %>% 
  kable(caption = "Sumary of candidates (from FTH2021_SelectionIndividus_V5.5.xlsx)")
```


```{r indMDP16, fig.cap="Marion's candidates in P16."}
palSp <- colorFactor("viridis", tocollect$TaxonActu)
limits <- st_read("data/OverallPlots/OverallPlots.shp", quiet = T)
leaflet(data = st_transform(tocollect, crs = crs)) %>%
  addTiles() %>% 
  addPolylines(data = st_transform(limits, crs = crs), col = "grey") %>% 
  addCircles(color = ~palSp(TaxonActu), opacity = 1,
             label = ~ TaxonActu, radius = ~ DBH/5) %>% 
  addLegend(pal = palSp, values = ~ TaxonActu)
```

## Candidates

We selected 15 individuals per species in P16 (10 + 5 extras) minimizing DBH variation while maximizing TWI variation (Fig. \@ref(fig:candFig), Tab. \@ref(tab:candTab), Fig. \@ref(fig:candMap)). We are lacking individuals for *Casearia javitensis*, *Protium stevensonii*, and *Virola surinamensis*.

```{r candidates}
candidates <- read_csv2("data/Paracou_P16_2020.csv") %>%
  left_join(species) %>%
  filter(!is.na(taxon)) %>%
  mutate(DBH = CircCorr/pi) %>% 
  st_as_sf(coords = c("Xutm", "Yutm"),
           crs = '+proj=utm +zone=22 +datum=WGS84 +units=m +no_defs +ellps=WGS84 +towgs84=0,0,0')
candidates$TWI <- raster::extract(raster::raster("data/TWI_1m.tif"), candidates)
candidates <- candidates %>% 
  group_by(taxon) %>% 
  mutate(N = n()) %>% 
  mutate(cand0 = ifelse(N <= 17, 1, 0)) %>% 
  mutate(m = median(DBH), l = quantile(DBH, 0.3), u = quantile(DBH, 0.6)) %>% 
  mutate(cand0 = ifelse(cand0 == 0 & DBH >= l & DBH <= u, 1, cand0))
cand <- as.data.frame(candidates) %>% 
  group_by(taxon) %>% 
  filter(cand0 == 1, N > 15) %>% 
  sample_n(15) %>% 
  unique()
candidates <- candidates %>% 
  mutate(candidate =  ifelse(idTree %in% cand$idTree, 1, 0)) %>% 
  mutate(candidate =  ifelse(N <= 15, 1, candidate)) %>% 
  mutate(candidate =  as.factor(candidate)) %>% 
  select(-cand0)
```

```{r candFig, fig.cap="DBH and TWI of candidates in P16."}
ggplot(candidates, aes(DBH, TWI, col = candidate, alpha = candidate)) +
  geom_point() +
  facet_wrap(~ taxon, scales = "free_x") +
  scale_color_manual("Candidate", values = c("red", "grey")) +
  scale_alpha_manual("Candidate", values = c(1, 0.5)) +
  theme(legend.position = c(0.8, 0.1)) +
  xlab("Diameter at breast height (DBH, cm)") +
  ylab("Topographic wetness index (TWI)")
```

```{r candTab}
as.data.frame(candidates) %>% 
  filter(candidate == 1) %>% 
  group_by(taxon, N) %>% 
  summarise(Nsel = n()) %>% 
  select(taxon, Nsel, N) %>% 
  kable(caption = "Candidates in P16.")
```

```{r candMap, fig.cap="Candidates in P16."}
palSp <- colorFactor("viridis", candidates$taxon)
leaflet(data = st_transform(filter(candidates, candidate == 1), 
                            crs = crs)) %>%
  addTiles() %>% 
  addPolylines(data = st_transform(filter(limits, Plot == 16), 
                                   crs = crs), col = "grey") %>% 
  addCircles(color = ~palSp(taxon), opacity = 1,
             label = ~ taxon, radius = ~ DBH/5) %>% 
  addLegend(pal = palSp, values = ~ taxon)
```

## Maps

Maps were automatically built using `sf` and `ggplot2` and save in the folder `maps`:

```{r mapsFun, echo=T}
contour <- st_read("data/ContourLinesPlots/ContourLinePlots.shp", quiet = T)
crs_rot = "+proj=omerc +lat_0=36.934 +lonc=-90.849 +alpha=0 +k_0=.7 +datum=WGS84 +units=m +no_defs +gamma=20"
make_map <- function(file, title, subplots){
  sublimits <- filter(limits, Plot == 16, Subplot %in% subplots) %>% 
    st_transform(crs = crs_rot)
  subcontour <- st_transform(contour, crs = crs_rot) %>% st_crop(sublimits)
  subcandidates <- filter(candidates, candidate == 1, SubPlot %in% subplots) %>% 
    st_transform(crs = crs_rot) %>% 
    mutate(label = paste0("P", Plot, "-", SubPlot, "-", TreeFieldNum, "_", substr(Genus, 1, 3), substr(Species, 1, 3), "_", round(DBH)))
  
  g <- ggplot() +
    geom_sf(data = subcontour, fill = NA, col = "lightgrey") +
    geom_sf(data = sublimits, fill = NA, col = "darkgrey", aes(text = subplots)) +
    geom_sf_text(data = sublimits, aes(label = Subplot), colour = "darkgrey") +
    geom_sf(data = subcandidates, col = "black") +
    ggrepel::geom_text_repel(
      data = subcandidates,
      aes(label = label, geometry = geometry),
      stat = "sf_coordinates",
      min.segment.length = 0
    ) +
    theme(axis.title = element_blank(), axis.text = element_blank(), 
          axis.ticks = element_blank(), axis.line = element_blank()) +
    scale_color_discrete(guide = "none") +
    ggtitle(title)
  ggsave(g, file = file, path = 'maps', width = 297, height = 420, unit = 'mm', dpi = 300)
}
```

```{r maps, eval=F}
make_map("P16NW.png", "Plot 16 North-West", c(1:3, 6:8, 11:13))
make_map("P16NE.png", "Plot 16 North-East", c(3:5, 8:10, 13:15))
make_map("P16SW.png", "P16 South-West", c(11:13, 16:18, 21:23))
make_map("P16SE.png", "P16 South-East", c(13:15, 18:20, 23:25))
```

## Extra FTH

Candidate individuals from the P16 were not enough for the FTH field campaign. 
We added individuals from the P6:

```{r, eval=F}
missings <- data.frame(
  spcode = c("Casjav", "Congui", "Jaccop", "Proste", "Tacmel", "Virmic", "Virsur"),
  inds = c(5, 1, 4, 4, 4, 1, 4),
  Genus = c("Casearia", "Conceveiba", "Jacaranda", "Tetragastris", "Tachigali", "Virola", "Virola"),
  Species = c("javitensis", "guianensis", "copaia", "panamensis", "melinonii", "michelii", "surinamensis")
)
p6 <- src_sqlite(file.path("/home/sylvain//Documents/BIOGECO/PhD/data/Paracou/",
                     "trees", "Paracou.sqlite")) %>% 
  tbl("Paracou") %>% 
  filter(CodeAlive == 1) %>% 
  filter(Plot == 6) %>%
  filter(CensusYear == 2017) %>% 
  collect() %>% 
  mutate(DBH = CircCorr/pi)
metradica <- read_excel("data/fth/MetradicaP6_dejafait.xlsx") %>% 
  rename(TreeFieldNum = FieldNr, SubPlot = Subplot) %>% 
  select(Plot, SubPlot, TreeFieldNum) %>% 
  mutate(metradica = 1)
candidates <- p6 %>% 
  left_join(metradica) %>% 
  filter(is.na(metradica)) %>% 
  mutate(spcode = paste0(substr(Genus, 1, 3), substr(Species, 1, 3))) %>% 
  mutate(spcode = recode(spcode, "Tetpan" = "Proste")) %>% 
  left_join(missings %>% 
  select(spcode) %>% 
  mutate(tocollect = 1)) %>% 
  filter(tocollect == 1) %>% 
  filter(DBH > 10, DBH < 60) %>% 
  st_as_sf(coords = c("Xutm", "Yutm"),
           crs = '+proj=utm +zone=22 +datum=WGS84 +units=m +no_defs +ellps=WGS84 +towgs84=0,0,0')
```

```{r}
contour <- st_read("data/ContourLinesPlots/ContourLinePlots.shp", quiet = T)
crs_rot = "+proj=omerc +lat_0=36.934 +lonc=-90.849 +alpha=0 +k_0=.7 +datum=WGS84 +units=m +no_defs +gamma=20"
limits <- st_read("data/OverallPlots/OverallPlots.shp", quiet = T)
sublimits <- filter(limits, Plot == 6, Subplot %in% 1:4) %>% 
    st_transform(crs = crs_rot)
subcontour <- st_transform(contour, crs = crs_rot) %>% 
  st_crop(sublimits)
subcandidates <- candidates %>% 
    st_transform(crs = crs_rot) %>% 
    mutate(label = paste0("P", Plot, "-", SubPlot, "-", TreeFieldNum, "_", substr(Genus, 1, 3), substr(Species, 1, 3), "_", round(DBH)))
g <- ggplot() +
  geom_sf(data = subcontour, fill = NA, col = "lightgrey") +
  geom_sf(data = sublimits, fill = NA, col = "darkgrey", aes(text = Subplot)) +
  geom_sf_text(data = sublimits, aes(label = Subplot), colour = "darkgrey") +
  geom_sf(data = subcandidates, col = "black") +
  ggrepel::geom_text_repel(
    data = subcandidates,
    aes(label = label, geometry = geometry),
    stat = "sf_coordinates",
    min.segment.length = 0
  ) +
  theme(axis.title = element_blank(), axis.text = element_blank(), 
        axis.ticks = element_blank(), axis.line = element_blank()) +
  scale_color_discrete(guide = "none")
ggsave(g, file = "P6.png", path = 'maps', width = 297, height = 420, unit = 'mm', dpi = 300, bg = "white")
```

```{r extrafthmap, fig.height=10, fig.width=10}
g
```

## Extra Post - FTH

Individuals were still missing after the FTH field campaign. 
We added individuals from the P16:

```{r}
inds <- googlesheets4::read_sheet("https://docs.google.com/spreadsheets/d/1-INnZ563VkNBH8rnSwfWieuGwPo9x9GRSV6jXNzatmM/edit?usp=sharing",
                          "individuals") %>% 
  rename(Individual = FieldCode) %>%
  dplyr::select(Genus, Species, Plot, SubPlot, TreeFieldNum) %>% 
  unique()

missings_hydroitv <- data.frame(
  spcode = c("Casjav", "Proste", "Tacmel"),
  Nhydroitv = c(6, 6, 1),
  Genus = c("Casearia", "Protium", "Tachigali"),
  Species = c("javitensis", "stevensonii", "melinonii")
)

missings_metradica <- data.frame(
  spcode = c("Casjav", "Chrpri", "Proste", "Proste", "Virmic", "Virsur"),
  Nmetradica = c(1, 1, 7, 3, 4, 4),
  Genus = c("Casearia", "Chrysophyllum", "Protium", "Protium", "Virola", "Virola"),
  Species = c("javitensis", "prieurii", "stevensonii", "stevensonii", "michelii", "surinamensis"),
  habitat = c("TF", "BF", "BF", "TF", "BF", "BF")
)

missings <- full_join(missings_hydroitv, missings_metradica) %>% 
  mutate(Nhydroitv = ifelse(is.na(Nhydroitv), 0, Nhydroitv)) %>% 
  mutate(N = Nmetradica) %>% 
  rowwise() %>% 
  mutate(N = ifelse(spcode == "Casjav", 6, N))

candidates <- read_csv2("data/Paracou_P16_2020.csv") %>%
  left_join(species) %>%
  filter(!is.na(taxon)) %>%
  mutate(DBH = CircCorr/pi) %>% 
  left_join(inds %>% 
              dplyr::select(Plot, SubPlot, TreeFieldNum) %>% 
              mutate(done = 1)) %>% 
  mutate(done = ifelse(is.na(done), 0, done)) %>% 
  left_join(dplyr::select(missings, Genus, Species) %>% 
              mutate(missing = 1)) %>% 
  filter(missing == 1) %>% 
  st_as_sf(coords = c("Xutm", "Yutm"),
           crs = '+proj=utm +zone=22 +datum=WGS84 +units=m +no_defs +ellps=WGS84 +towgs84=0,0,0') %>% 
  st_intersection(st_read("data/TopographicLevels/TopographicLevels.shp", quiet = T)) %>% 
  mutate(topo = recode(TypeTopo, "BasFond" = "BF", "Plateau" = "TF", "Pente" = "P")) %>% 
  group_by(Genus, Species) %>% 
  mutate(m = median(DBH), l = quantile(DBH, 0.25), u = quantile(DBH, 0.75)) %>% 
  mutate(l = ifelse(Genus == "Casearia", 0, l)) %>% 
  filter(DBH >= l, DBH <= u)
# candidates %>% 
#   group_by(Species, Genus) %>% 
#   summarise(N = n()) %>% 
#   left_join(select(missings, Genus, Species, Nhydroitv, Nmetradica))
```

```{r}
contour <- st_read("data/ContourLinesPlots/ContourLinePlots.shp", quiet = T)
crs_rot = "+proj=omerc +lat_0=36.934 +lonc=-90.849 +alpha=0 +k_0=.7 +datum=WGS84 +units=m +no_defs +gamma=20"
limits <- st_read("data/OverallPlots/OverallPlots.shp", quiet = T)
sublimits <- filter(limits, Plot == 16, Subplot %in% 1:25) %>% 
    st_transform(crs = crs_rot)
subcontour <- st_transform(contour, crs = crs_rot) %>% 
  st_crop(sublimits)
subcandidates <- candidates %>% 
    st_transform(crs = crs_rot) %>% 
    mutate(label = paste0("P", Plot, "-", SubPlot, "-", TreeFieldNum, "_", substr(Genus, 1, 3), substr(Species, 1, 3), "_", round(DBH), "_", topo))
g <- ggplot() +
  geom_sf(data = subcontour, fill = NA, col = "lightgrey") +
  geom_sf(data = sublimits, fill = NA, col = "darkgrey", aes(text = Subplot)) +
  geom_sf_text(data = sublimits, aes(label = Subplot), colour = "darkgrey") +
  geom_sf(data = subcandidates, col = "black") +
  ggrepel::geom_text_repel(
    data = subcandidates,
    aes(label = label, geometry = geometry),
    stat = "sf_coordinates",
    min.segment.length = 0,
    size = 3
  ) +
  theme(axis.title = element_blank(), axis.text = element_blank(), 
        axis.ticks = element_blank(), axis.line = element_blank()) +
  scale_color_discrete(guide = "none")
ggsave(g, file = "P16extra.png", path = 'maps', width = 297, height = 420, unit = 'mm', dpi = 300, bg = "white")
```

```{r postfthmap, fig.height=10, fig.width=10}
g
```

## Leaf variation

For the leaf variation within individual we sampled across the 27 *Virola michelii* from the P16 that followed our requirements (DBH between 30th and 60th quantiles):

```{r samplingleafvar, fig.cap="Candidates in P16."}
# data
candidates <- read_csv2("data/Paracou_P16_2020.csv") %>%
  filter(Genus == "Virola", Species == "michelii") %>% 
  mutate(DBH = CircCorr/pi) %>% 
  filter(DBH >= quantile(DBH, 0.3), DBH <= quantile(DBH, 0.6))  %>% 
  st_as_sf(coords = c("Xutm", "Yutm"),
           crs = '+proj=utm +zone=22 +datum=WGS84 +units=m +no_defs +ellps=WGS84 +towgs84=0,0,0') %>% 
  mutate(Ind = paste0("P", Plot, "_", SubPlot, "_", TreeFieldNum))

# html map
limits <- st_read("data/OverallPlots/OverallPlots.shp", quiet = T)
leaflet(data = st_transform(candidates, crs = crs)) %>%
  addTiles() %>% 
  addPolylines(data = st_transform(filter(limits, Plot == 16),
                                   crs = crs), col = "grey") %>%
  addCircles(opacity = 1, label = ~ Ind, radius = ~ DBH/2) 

# paper map 
contour <- st_read("data/ContourLinesPlots/ContourLinePlots.shp", quiet = T)
crs_rot = "+proj=omerc +lat_0=36.934 +lonc=-90.849 +alpha=0 +k_0=.7 +datum=WGS84 +units=m +no_defs +gamma=20"
sublimits <- filter(limits, Plot == 16) %>% 
    st_transform(crs = crs_rot)
subcontour <- st_transform(contour, crs = crs_rot) %>% 
  st_crop(sublimits)
subcandidates <- st_transform(candidates, crs = crs_rot) %>% 
    mutate(label = paste0("P", Plot, "-", SubPlot, "-", TreeFieldNum, "_", substr(Genus, 1, 3), substr(Species, 1, 3), "_", round(DBH)))
g <-  ggplot() +
  geom_sf(data = subcontour, fill = NA, col = "lightgrey") +
  geom_sf(data = sublimits, fill = NA, col = "darkgrey") +
  geom_sf_text(data = sublimits, aes(label = Subplot), colour = "darkgrey") +
  geom_sf(data = subcandidates, col = "black") +
  ggrepel::geom_text_repel(
    data = subcandidates,
    aes(label = label, geometry = geometry),
    stat = "sf_coordinates",
    min.segment.length = 0
  ) +
  theme(axis.title = element_blank(), axis.text = element_blank(), 
        axis.ticks = element_blank(), axis.line = element_blank()) +
  scale_color_discrete(guide = "none") +
  ggtitle("Plot 16 - Virola michelii")
ggsave(g, file = "P16_virmic.png", path = 'maps', width = 297, height = 420, unit = 'mm', dpi = 300, bg = "white")

# gps
# candidates %>%
#   mutate(name = paste0("P", Plot, "-", SubPlot, "-", TreeFieldNum, "_", substr(Genus, 1, 3), substr(Species, 1, 3), "_", round(DBH))) %>% 
#   dplyr::select(name, geometry) %>%
#   st_write(dsn = "maps/P16_virmic.gpx", delete_dsn = T, driver = "GPX", layer = "waypoints")
```