Data and code from: Decades of historical outbreak cycles in a multivoltine insect reveal a plastic phenological response to climate change
Abstract
Many organisms overwinter in a specific life stage, which means their phenology must be well-timed with seasonal changes in the environment. As environments warm, we expect a delay in the onset of winter. For organisms where temperature is the primary driver of development rate, warming environments also mean faster development. If temperature dependence in the development rate of individuals does not also change, this will cause a mismatch between seasonal timing of the critical overwintering stage and the onset of winter. There are two biological mechanisms by which ectotherms can adapt their phenology in the face of climate change to maintain correct timing: i) organisms undergo evolution of the development thermal reaction norm, and ii) organisms have a plastic response in their development to multiple environmental cues. Here we use high resolution records of densities of the smaller tea tortrix (Adoxophyes honmai) over multiple decades across 9 locations in Japan to infer temperature-dependent changes in development rates over both time and space. The pest insect displays regular single-generation limit cycles, which provides a unique opportunity to infer changes in developmental rates directly from historical records of natural populations. The last half century has seen a temperature increase of about 1 ℃ across Japan, and our analyses show that populations slowed development on average by 16% to maintain the correct timing of the overwintering stage. Development rates measured from common garden experiments reveal that the change is not due to evolution. Our results build on recent laboratory studies to suggest that there is substantial plasticity in developmental thermal reaction norms that may explain how the phenology of ectotherms will respond to climate warming in natural systems.
https://doi.org/10.5061/dryad.573n5tbhg
Description of the data and file structure
The data and scripts archived here contain everything needed to reproduce the results of: Decades of historical outbreak cycles in a multivoltine insect reveal a plastic phenological response to climate change by R.A. Smith, W.A. Nelson, T. Yamanaka, Yasushi Sato, Takeshi Kamimuro, Ryosuke Omata, and Ottar Bjornstad. The scripts are written with relative file paths that assume: (1) The 'data' and 'scripts' directories are stored in the same directory and (2) each script is run from the directory in which it resides.
Files and variables
File: data.zip
Description:
development data.csv -- Laboratory measured development rates measured by Yasushi Sato (For three subpopulations of A. honmai, for A. dubia, and for A. orana). Columns in the dataset are: Species (A. honmai, A. dubia, or A. orana), Population (location), Sex (M or F), Stage (developmental stage), Temperature (degrees C), DevTime (days), PhotoPeriod (hours light and hours dark), Source (categorical variable indicating where the data came from).
MeanTemp9Stations.Yamanaka.xlsx -- Daily temperature data for the 9 sites in the study. This excel file has a different sheet for each location in the study. Each sheet has two columns: date (format: YYYY-MM-DD), MeanTemp (degrees C).
mean_temp_9stations.csv -- same as 'MeanTemp9Stations.Yamanaka.xlsx' but in csv format. This dataset has three columns: location (name of location), date (format:YYYY-MM-DD), MeanTemp (degrees C).
MeanTempLong.Yamanaka.xlsx -- Daily temperature data for nearby sites; used for imputing missing temperature data. This excel file has a different sheet for each location. Each sheet has two columns: date (format: YYYY-MM-DD), MeanTemp (degrees C).
mean_temp_9stations_long.csv -- same as 'MeanTempLong.Yamanaka.xlsx' but in csv format. This dataset has three columns: location (name of location), date (format:YYYY-MM-DD), MeanTemp (degrees C).
tea_plantation_approx_locations.csv -- Locations of sites used in the study. This file has three columns: location (name of location), lat (latitude), long (longitude).
TortrixRecords9Stations.Yamanaka.xlsx -- Counts of adults collected either in light or pheromone traps at each of the 9 sites. This excel file has a different sheet for each location in the study. All sheets have the same format. The first row has a single entry indicating whether the data are pheromone or light trap census data. The second row contains the variable names. The first two columns are: Month (full Month name), Date (day of the month data were collected). Remaining columns names are the year in which the census was conducted and therefore the total number of columns varies by location. Entries in these columns are the number of adult moths observed, with missing data coded as NA.
tortrix_records_9stations.csv -- Same as 'TortrixRecords9Stations.Yamanaka.xlsx' (with the exception of some data cleaning that is done in the script in 'data.R') but in csv format. There are four columns in this dataset: location (site that data were collected), method (type of trap used: either light or pheromone), date (format: YYYY-MM-DD), count (number of adult moths observed, NA for missing data).
Yacha.Pheromone.Dat_20240404.csv -- Additional counts of adults from pheromone traps in collected in Kanaya. This dataset has 35 columns. The first two columns are: month (format: full name of month), day (day of the year). The remaining columns are labeled by the year the data were collected and range from 1991 to 2023, inclusive.
Code/software
Data analysis for this paper was run in R. The scripts folder in 'Software' related work (Zenodo) contains the following:
- data.R -- A script that reads in the data, cleans it, and infers cycle length on the phi scale for each annual time series at each site.
- functions.R -- A collection of functions that are used by various scripts.
- last_cycle_analysis -- A folder that contains (1) a script that runs the analysis with a dataset in which each annual time series is truncated to remove the last population cycle, and (2) an Rmarkdown file that generates the plots from this analysis (Appendix S2 from paper).
- manuscript_figures -- A folder containing a script for each figure presented in the manuscript.
- Rdata -- A folder that holds output from the scripts as Rdata files. The original Rdata files are included here, but can be regenerated by running the scripts.
- stats.R -- Statistical analysis of data presented in the main paper.
