search

lizzieinvancouver / temporalvar

0 stars 0 forks source link

add Bfin slopes and rerun #38

Closed donahuem closed 5 years ago

donahuem commented 5 years ago

Need to add a slope over the last 100 yrs and maybe a growth rate measure based on multiplicative rate of increase for the last 100 yrs to the summary output

Send new runs with these conditions

donahuem commented 5 years ago

Added SummaryOut:
-slope of linear regression for the period (i.e., stationary or nonstationary), excluding the first 10 years, if there are 10 or more years in the period -average(ln(N(t)/N(t-1)) for the period, excluding the first 10 years to avoid effect of initial conditions; does not calculate if there are less than 10 years in the period.

donahuem commented 5 years ago

@lizzieinvancouver First set of runs with these additional outputs.
Submitted batch job 62025212 Submitted batch job 62025233 Submitted batch job 62025263 Submitted batch job 62025274

lizzieinvancouver commented 5 years ago

@donahuem Nice! I pulled these and can work on plots with a little help. The files no longer have header rows that say what each column is so can you let me know what the last 12 columns names are? (I assume the first columns through tauIPmean are unchanged, but let me know if that assumption is incorrect). Thanks!

lizzieinvancouver commented 5 years ago

@donahuem Wait! One more issue (sorry); the files also have: 

#NAME?
--

In some of the cells (line 7 of SummaryOut_62025212-1.txt for example) and that is tricky to read in. Can we clean these up somehow in the creation script? I will see if I can figure a workaround for now....

lizzieinvancouver commented 5 years ago

@donahuem Me again! So I can read the files with the #NAME in them but the first one I looked at is odd: it says no two-species coexistence in period 1, but there is maybe 20 rows with two-species coexistence in period 2 ... which is odd. Maybe I have the column headers wrong ... will await further instructions.

lizzieinvancouver commented 5 years ago

Update: the #NAME? is an issue Lizzie added ... in opening and saving the files in Excel (just this file: SummaryOut_62025212-1.txt) she probably made all the -Inf values into #NAME?.

@lizzieinvancouver So ignore the #NAME? but Lizzie should re download: SummaryOut_62025212-1.txt

lizzieinvancouver commented 5 years ago

colnames are: jobID taskID runID period nperiods yout itertime ncoexist coexist1 coexist2 alpha1 alpha2 c1 c2 Rstar1 Rstar2 tauI1 tauI2 g1mean g2mean tauIP1_mean tauIP2_mean wetID dryID row R0_mean R0_media R0_autocorr Bfin1 Bfin2 slope_Bfin1 slope_Bfin2 avgPCgrowth1 avgPCgrowth1

donahuem commented 5 years ago

OK, I think i fixed output so that each period should print out and header is only at the top. Note that I added output columns for s_1, s_2, phi_1, phi_2

donahuem commented 5 years ago

Next batch of runs that I am trying for regular runs:
855518 855569 855591 855642

donahuem commented 5 years ago

When you have a chance to look at the output, can you check how frequently the "slopeBfin" and "avgPCgrowth" (last two columns) are NA or Inf.

lizzieinvancouver commented 5 years ago

SlopeBfin is never NA or -Inf, but avgPCgrowth is 60% of the time.