Including Erodium counts reduced the winter plant CV ratio for rodent removal versus access on eight Portal plots in 1989–2008

We tested whether complete rodent exclusion changed annual plant count variation differently in winter and summer at Portal, Arizona. We used eight plots with stable rodent and ant treatments in 1989–2008. All 5,120 fixed-quadrat survey records were complete. The sample had 320 plot-season counts and 343,559 annual plants. The coefficient of variation (CV) is sample SD divided by mean. The full-period winter-to-summer ratio of rodent treatment ratios was 1.001, with a nominal 95% plot-model interval of 0.780–1.283. A planned taxon check changed the winter removal-to-access CV ratio from 0.845 with Erodium cicutarium counts to 1.036 without them. The added paired test gave a ratio of these ratios of 0.815. Its interval was 0.658–0.950 when plots and two-year time blocks were resampled together. Including Erodium changed the mean treatment ratio more than the SD treatment ratio. The full winter CV reduction itself was uncertain under joint plot-and-time resampling. Log-count and half-period checks also changed the interpretation. The taxon result is count accounting, not the field effect of removing Erodium. Intervals are exploratory and depend on plot and time assumptions at one site. The package preserves source data, provenance, code, all results, figures, and independent checks.

annual plants · rodent exclusion · Erodium cicutarium · temporal variation · coefficient of variation · Portal Project · reproducible ecology