SensorSpecies

This is a repository for the paper "Structure-based identification of sensor species for anticipating critical transitions" by Andrea Aparicio, Jorge X. Velasco, Claude H. Moog, Yang-Yu Liu, and Marco Tulio Angulo.

USER INSTRUCTIONS

I. To calculate the sensor score, early-warning score and (q^*, r_1) predictions for a mutualistic system, and obtain the network diagram (Figure 2):

1. Save in “data/“ the system’s interaction matrix in csv format ("sys_name.csv”). The matrix must be rectangular and have plants as rows and animals as columns.

2. In “SensorScore_EWS_calc/Sensor_score_calc.nb” enter the filename (without the extension), in the field fn (for Genebra:).

3. Run the main function. The results will be shown inline (for Genebra:).

II. To simulate a critical transition in one or more realizations of a mutualistic system with the parameters in Table S3 (nominal parameters)

1. Save in “data/“ the system’s interaction matrix in csv format ("sys_name.csv”). The matrix must be rectangular and have plants as rows and animals as columns.

2. Define
   kStart = initial realization number (must be > 0)
   kEnd = final realization number (must be > kStart)

3. Run the script “SensorScore_EWS_calc/SS_solveGen.py” as:
   python SS_solveGen.py sys_name kStart kEnd

4. Find the results in csv format in
   

• "data/data_sims/“:
  ⁃ the variance of the abundance of every species at every \mu step ("sysname_var_reps_k.csv")
  ⁃ the autocorrelation of the abundance of every species at every \mu step ("sysname_ac_reps_k.csv")
  ⁃ the mean of the abundance of every species at every \mu step ("sysname_var_mean_k.csv")
  ⁃ a sample of the abundance at every \mu step ("sysname_samp_reps_k.csv")

• "data/data_EWS/"
  ⁃ early-warning scores of every species (first plants and then animals) ("sysnameEWSk.csv")
  

III. To analyze the time-series obtained after simulating a critical transition on a mutualistic system, and obtain the plots in Figures 2 and 3

1. Make sure that the interaction matrix is saved as in I.1, that the data is stored as indicated in II.4, and have (kStart kEnd) from II.2 handy
2. Define kT: one particular realization to display (must be \in (kStart,kEnd])
3. In “SensorScore_EWS_calc/EW_score_calc.nb” enter
   ⁃ the filename (without the extension), in the field fn
   ⁃ kStart, KEnd and kT in the fields of the same name
   
4. Run the main function. The results will be shown inline.

IV. To generate the plots in Figure 1

1. Run the script “All_Figs/Toy_Mutualistic.py”

V. To generate the plots in Figure 4 and Figures S6-S14, find the corresponding script using the list below, and follow the instructions on its README file:

⁃ Fig.4, All_Figs/General_figs.nb
⁃ Fig.S4, SensorScore_EWS_calc/Sensor_score_calc.nb, SS_solveGen.py, EW_score_calc.nb
⁃ Fig.S5, SensorScore_EWS_calc/EW_score_calc.nb
⁃ Fig.S6, All_Figs/General_figs.nb
⁃ Fig.S7, All_Figs/General_figs.nb
⁃ Fig.S8, All_Figs/ Fig_S8_S9.nb
⁃ Fig.S9, All_Figs/ Fig_S8_S9.nb
⁃ Fig.S10, SensorScore_EWS_calc/ SS_solveGen.py, All_Figs/ Fig_S10-11-12.nb
⁃ Fig.S11, SensorScore_EWS_calc/ SS_solveGen.py, All_Figs/ Fig_S10-11-12.nb
⁃ Fig.S12, SensorScore_EWS_calc/ SS_solveGen.py, All_Figs/ Fig_S10-11-12.nb
⁃ Fig.S13, All_Figs/FigS13.nb
⁃ Fig.S14, All_Figs/FigS14.nb