methods_results.Rmd

---
title: "Assessing generality in trait-based ecology and theory by using experimental warming and transplant communities spanning a broad temperature gradient"

author: |
  | Brian J. Enquist^[Corresponding author: brian.enquist@uoa.com] $^1$, Brian S. Maitner$^1$, Aud H. Halbritter$^{4,5}$,
  | Richard J.Telford$^{4,5}$, Lorah Seltzer$^1$, Vanessa Buzzard$^1$,
  | Jonathan Henn, Sean Michaletz, Inge H.J. Althuizen$^4$,
  | Kine Blom$^4$, Alex Brummer, Julia Chacon$^1$,
  | Javiera Seaman Espinosa, Bárbara William Garcia$^4$, Ragnhild Gya$^{4,5}$,
  | Elisabeth Nesheim Hauge$^4$, Hai He, Francesca Jaroszynska$^4$,
  | Blake L. Joyce7, Kari Klanderud$^3$, Frida Knoop$^4$,
  | Hanna Lee, Rebecca Lehman, Miguel Muñoz Mazon,
  | Michelangelo Sergio Moerland, Linda Hovde Nordås$^4$,
  | Ahui Peng$^6$, Christine Pötsch$^4$, Claire Ponsac, Fei Ran$^6$,
  | Ruben Roos$^3$, Christien Steyn, Megan Kathleen
  | Sullivan, Xiangyang Sun$^6$, Jesslyn Tjendra$^4$, Yao Xiao$^6$,
  | Li Zhang$^6$, Xiaoxiang Zhao$^6$, Yan Yang$^6$ & Vigdis Vandvik$^{4,5}$

address: |
 | $^1$ Department of Ecology and Evolutionary Biology, University of Arizona, AZ, USA
 | $^2$ The Santa Fe Institute, Santa Fe, NM, USA
 | $^3$ Faculty of Environmental Sciences and Natural Resource Management, Norwegian University of Life Sciences, Norway,
 | $^4$ Department of Biological Sciences, University of Bergen, Norway,
 | $^5$ Bjerknes Centre for Climate Research, University of Bergen, Norway
 | $^6$ Institute of Mountain Hazards and Environment (CAS), China.
 | $^7$ University Information Technology Services, University of Arizona, AZ, USA


date: "`r format(Sys.Date(), '%d %B %Y')`"

output: 
  #pdf_document
  bookdown::pdf_book:
    dev: cairo_pdf
    toc: no
    number_sections: false
    latex_engine: lualatex
  fig_caption: yes
  extra_dependencies: ["flafter"]
  header-includes:
    - \usepackage{caption}
    - \usepackage{booktabs}
    - \usepackage{longtable}
classoption: a4paper
bibliography: Rmd/TDT2.bib
csl: Rmd/elsevier-harvard_rjt.csl


---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
library("drake")
library("cairoDevice")

```

\newpage

## Methods

This section contains the experimental design and all the references used for the data analysis of this manuscript.


```{r mountain, eval = TRUE, echo = FALSE, out.width = '90%', fig.cap="Experimental set up of the Mt. Gongga elevational experiment. There are four experimental study sites across the 1,130m elevational gradient. The arrows reflect the different tuff transplant experiments between these sites. The rectangles depict one of seven experimental blocks at each altitude, where squares reflect transplants and the hexagon an OTC). The colours reflect temperature treatments (orange = ca 1.75°C warmer, red = 5.3°C warmer, light blue =   ca 1.75°C colder, red = 5.3°C colder, green = control).  Within each of the seven blocks at each site, plots were randomly designated to six different experimental treatments: (i) passive warming with an open top chamber (OTC; all sites); (ii) transplanting to a site one step warmer along the gradient (warming; from sites High alpine, Alpine); (iii) Middle alpine transplanting to a site one step colder along the gradient (cooling; from sites Alpine, Middle alpine, and Lowland); (iv) transplanting down the entire gradient (extreme warming; from site High alpine); (v) transplanting up the entire gradient (extreme cooling, from site Lowland); (vi) transplanting within blocks within all sites (to control for the transplanting itself and a local transplant); (vii) an untouched control plot (Control; all sites). Thus, in every site, each OTC has a local unmanipulated control, and each transplanted turf has an “origin” site and a “destination” site, with two types of controls, local transplant and untouched plots."}

knitr::include_graphics("./images/ChinaMountain.pdf")

```


\newpage

Here is another conceptual figure.


```{r concept-fig, eval = TRUE, echo = FALSE,  out.width = '100%', fig.cap="Conceptual figure showing the experimetnal design with traits changing along a gradienta and transplanting traits along the gradient. Trait imputation, bootstrapping and higher moments. And divergence - convergence."}

knitr::include_graphics("./images/TDT.png")

```


### Software

All analyses were done in R v. `r getRversion()` [@R] and used
traitstrap v. `r packageVersion("traitstrap")` [@traitstrap] to impute traits,
e1071 v. `r packageVersion("e1071")` [@e1071] to calculate moments,
and tidyverse v. `r packageVersion("tidyverse")` [@tidyverse].
Packages drake v. `r packageVersion("drake")` [@drake],
rmarkdown `r packageVersion("rmarkdown")` [@rmarkdown1] and
renv v. `r packageVersion("renv")` [@renv] were used to ensure reproducibility.
Code to reproduce all the analyses is archived at [https://github.com/richardjtelford/trait-driver-analysis](https://github.com/richardjtelford/trait-driver-analysis).

\newpage

## Results

This section contains all figures and tables of the main manuscript. The figures show responses in plant community composition and plant functional trait composition to climate  and climate change treatments treatments.


### Trait-climate correlations

This figure shows the correlation of trait mean and temperature, representing the elevational gradient.


```{r trait-climate, eval = TRUE, echo = FALSE, fig.cap="Relationship between functional traits and summer air temperature (2m above ground). The trait values are bootstrapped community weighted traits from control plots in 2016. The regression line is shown (grey dashed) and significant relationships are indicated with a solid red line. The shaded area displays the confidence interval. The traits are sorted from positive slope, no significant to negative slope."}
readd(moments_by_climate_plot)
```

\newpage

Regression results of the climate-trait relationship.

```{r trait-climate-table, eval = TRUE, echo = FALSE}
loadd(moments_by_climate_table)

moments_by_climate_table %>%
  kableExtra::kable(caption = "Results from regression of traits and summer air temperature. Shown are estimates, standard error, statistics and P value. Significant P values are indicated with asterisks: ** P value < 0.01, * p < 0.05")

```

\newpage


This is a principle componant analysis showing plant community composition for the different treatments in all sites.

```{r trait_ordination_plot, eval = TRUE, echo = FALSE, fig.width = 8,fig.height = 7, fig.cap="Multivariate principal component analysis of community composition for each climate treatment along the elevational gradient sites."}
readd(ordination)
```


\newpage

### Colonization and extinction

Here we present the number of species that colonize and go extinct in the different climate treatments.

```{r colo_ext_plot, eval = TRUE, echo = FALSE, fig.cap="Realized (closed bar) and expected (open bar) colonization (up) or extinction (down) for each climate treatment. Colonization and extinction are expressed as mean number of species per plot and treatment."}
readd(colo_extinction_plot)
```

Here we present the abundance of species that colonize and go extinct in the different climate treatments.

```{r colo_ext_plot2, eval = TRUE, echo = FALSE, fig.cap="Realized (closed bar) and expected (open bar) colonization (up) or extinction (down) for each climate treatment. Colonization and extinction are expressed as mean abundance per plot and treatment."}
readd(colo_ext_abundance_plot)
```

\newpage

### Convergence and divergence


```{r temp_trait_hist, eval = TRUE, echo = FALSE, fig.cap="Histogram for mean trait value over time."}
readd(temporal_trait_histograms)
```

```{r conv-div_plot, eval = TRUE, echo = FALSE, fig.height=7, fig.cap="Testing for convergence and divergence in each pair-wise trait value as a function of year since transplant. Temporal variation in effect sizes for each measured trait. The first column shows traits divergence and the second column trait convergence. The grey panels labeled with Positive slope, No slope and Negative slopes show the predicted outcome according to TDT, depending on the traits relationship with summer air temperature (see New Fig. 2). The colour of the lines indicates the transplant treatment: transplant to 1.75°C warmer (red), transplant to 0.5°C warmer (pink), OTC (orange), transplant to 0.5°C colder (light blue), transplant to 1.75°C colder (dark blue). Solid lines indicate a significant change in effect size over time."}
readd(conv_div_plot)
```

\newpage

```{r rda_plot, eval = TRUE, echo = FALSE, fig.cap="Testing for convergence in multivariate community (first column) or trait-space (second and third column) as a function of year since transplant. Results from  Principal Response Curve multivariate RDA analysis. "}
readd(TraitRDA)
```



```{r rda_result, eval = FALSE, echo = FALSE}

loadd(result_rda)
result_rda %>%
  kableExtra::kable(caption = "Results from RDA, testing difference in species or fixed/plastic trait composition over time among climate treatments and the origin control turfs using a permutation test with 999 permutations.")

```

```{r rda_proportion, eval = FALSE, echo = FALSE}

loadd(rda_proportions)
rda_proportions %>%
  kableExtra::kable(caption = "Proportion convergence towards destination site after 4 years of climate warming or cooling.", digits = 2)

```



```{r skewness_plot, eval = TRUE, echo = FALSE, fig.cap="xxx"}
readd(skewness_plot)
```

```{r half_happymoment_plot, eval = TRUE, echo = FALSE, fig.cap="Change in mean and skewness for leaf area, thickness, phosphorus, and C and N isotopes in each treatment and over time. Shown are mean values for each time period and climate treatment with standard error."}
readd(half_happymoment_plot)
```


\newpage
## Supplementary

This section contains all plots and tables of the supplementary material of this manuscript.


### Trait correlation

```{r trait-cor, eval = TRUE, echo = FALSE, fig.height=7, fig.cap="Correlation coefficient between each trait combination. Correlation coefficients are shown for trait combinations with a significnat correlarion."}
readd(trait_corr)
```



### Trait change over time

```{r hist-all-1, eval = TRUE, echo = FALSE, fig.height=7, fig.cap="Histogram for mean trait values over time."}
readd(temporal_trait_histograms_all_1)
```

```{r hist-all-2, eval = TRUE, echo = FALSE, fig.height=7, fig.cap="Histogram for mean trait values over time."}
readd(temporal_trait_histograms_all_2)
```

```{r hist-all-3, eval = TRUE, echo = FALSE, fig.cap="Histogram for mean trait values over time."}
readd(temporal_trait_histograms_all_3)
```

<!-- ### Colonization and extinction over time -->

```{r col-ext-time-plot, eval = FALSE, echo = FALSE, fig.cap="Number of species colonizing and going extinct over time in each climate treatment."}
readd(col_ext_over_time_plot)
```


\newpage
### Convergence and divergence for traits with no relationship along climate gradient

```{r conv-div-no-slope-plot, eval = TRUE, echo = FALSE, fig.height=4, fig.width=6, fig.cap="Testing for convergence and divergence in each pair-wise trait value as a function of year since transplant. Temporal variation in effect sizes for each measured trait. The first column shows traits divergence and the second column trait convergence. The grey panels labeled with Positive slope, No slope and Negative slopes show the predicted outcome according to TDT, depending on the traits relationship with summer air temperature (see New Fig. 2). The colour of the lines indicates the transplant treatment: transplant to 1.75°C warmer (red), transplant to 0.5°C warmer (pink), OTC (orange), transplant to 0.5°C colder (light blue), transplant to 1.75°C colder (dark blue). Solid lines indicate a significant change in effect size over time."}
readd(conv_div_no_slope_plot)
```


### Stats for convergence and divergence analysis

```{r conv-div_table, eval = FALSE, echo = FALSE}
loadd(treatment_effect_table)
treatment_effect_table %>%
   kableExtra::kable("latex", longtable = TRUE, booktabs = TRUE, caption = "Results Testing for convergence and divergence in each pair-wise trait value as a function of year since transplant. Significant P values are indicated with asterisks: ** P value < 0.01, * p < 0.05") %>% 
  kableExtra::kable_styling(latex_options = c("hold_position", "repeat_header"))
```


\newpage


### Eucledian distance analysis

```{r eucledian_plot, eval = TRUE, echo = FALSE, fig.cap="Euclidean distance in functional trait composition since transplant (2012 - 2016) for control plots and treatments at each site. The colours indicate the different treatments control (grey), transplant to 1.75°C warmer (red), transplant to 0.5°C warmer (pink), OTC (orange), transplant to 0.5°C colder (light blue), transplant to 1.75°C colder (dark blue). The two columns show two scenarios of trait plasticity. Divergence reflects a conservative scenario, where treatments are compared to the origin control and do not allow for trait plasticity. Contrary, convergence reflects full plasticity where treatments are compared to the destination controls."}
readd(euclidean_distance_plot)
```


```{r eucledian_table, eval = FALSE, echo = FALSE}
loadd(euclidean_dist_table)

euclidean_dist_table %>%
  kableExtra::kable(caption = "Results testing difference between eucledian distance for control plots and each treatment at each site between 2012 and 2016.")

```



### All the happy moments

```{r full_happymoment_plot, eval = TRUE, echo = FALSE, fig.cap="Change in mean, variance, skewness, kurtosis and range for leaf area, thickness, phosphorus, and C and N isotopes in each treatment and over time. Shown are mean values for each time period and climate treatment with standard error."}
readd(full_happymoment_plot)
```


## References