arctic_code.Rmd

---
title: "Untitled"
output: html_document
---

```{r}
library(tidyverse)
library(lmerTest)
library(lme4)
```


```{r}
#read in the unaltered phenophase data set
pheno_data<-read.csv("CCIN13215_20220120_tundra_phenology_database.csv")

#filter the dataset by flowering phenophase, control treatment, forb functional group, and species of interest (species that occur over the same time span with a lot of observations)
pheno_data_filt<-pheno_data %>% 
  filter(phenophase == "flower") %>% 
  filter(treatment == "CTL") %>% 
  filter(functional_group == "forb") %>% 
  filter(spp == "CARBEL" | spp == "DRALAC" | spp == "PAPRAD" | spp == "POLVIV" | spp == "PEDHIR" | spp == "SAXOPP" | spp == "SILACA" | spp == "STECRA")

#seeing how many species occur in both wet and dry
numbers<-pheno_data_filt %>% 
  group_by(spp,soil_moisture) %>% 
  summarise(n=n()) %>% 
  ungroup() %>% 
  filter(soil_moisture == "wet" | soil_moisture =="dry" | soil_moisture == "moist")

#only including wet and dry 
spp_filt[spp_filt == "moist"] <- "wet"
  

#filtering by three species of interest - occur in both wet and dry, have a similar time range of observations, and have a lot of observations
spp_filt<-pheno_data_filt %>% 
  filter(spp == "DRALAC" | spp == "PAPRAD" | spp == "POLVIV" | spp == "PEDHIR") %>% 
  filter(soil_moisture =="wet" | soil_moisture =="moist" | soil_moisture =="dry")

spp_filt[spp_filt == "moist"] <- "wet"

```


```{r}
#creating a dataset per species for t tests
polviv_ttest<-spp_filt %>% 
  filter(spp == "POLVIV")
dralac_ttest<-spp_filt %>% 
  filter(spp == "DRALAC")
paprad_ttest<-spp_filt %>% 
  filter(spp == "PAPRAD")
pedhir_ttest<-spp_filt %>% 
  filter(spp == "PEDHIR")

polviv_ttest[polviv_ttest == "moist"] <- "wet"
dralac_ttest[dralac_ttest == "moist"] <- "wet"
paprad_ttest[paprad_ttest == "moist"] <- "wet"
pedhir_ttest[pedhir_ttest == "moist"] <- "wet"


#removing blank values from the POLVIV dataset
polviv_ttest<-polviv_ttest %>% 
  filter(soil_moisture == "dry" | soil_moisture == "wet")
```


###check assumptions for t tests!!!!!!


```{r}
#performing t test for paprad
t.test(DOY~soil_moisture, data = paprad_ttest)
```

```{r}
#t test for polviv
t.test(DOY~soil_moisture, data = polviv_ttest)
```

```{r}
# t test for dralac
t.test(DOY~soil_moisture, data = dralac_ttest)
```


```{r}
#t test for pedhir
t.test(DOY~soil_moisture, data = pedhir_ttest)
```

```{r}
#attempting a t test with one site per species (site with the most observations)

#polviv shows a significant difference at both sites which contain both dry and wet soils, however the results are opposite
polviv_ttest2<-polviv_ttest %>% 
  filter(study_area == "Endalen")

t.test(DOY~soil_moisture, data = polviv_ttest2)

polviv_ttest3<-polviv_ttest %>% 
  filter(study_area == "Adventdalen")

t.test(DOY~soil_moisture, data = polviv_ttest3)
```

###check assumptions for linear model!!!!!


```{r}
#creating a linear model to test the effects of soil moisture on flowering date
#random intercept, fixed slope

mix_model_int<-lmer(DOY~soil_moisture * spp +
                      (1|study_area/subsite) + (1|year),
                    data = spp_filt)
summary(mix_model_int)
```

```{r}
#looking at the effects without species
mix_model_int2<-lmer(DOY~soil_moisture +
                      (1|study_area/subsite) + (1|year),
                    data = spp_filt)
summary(mix_model_int2)

```

```{r}
#model selection
AICc(mix_model_int, mix_model_int2)
```