Year and plot resampling did not resolve a change in rodent energy compensation in nine Portal plots from 2010–2020 to 2021–2025

We tested whether small granivores regained the high share of kangaroo-rat energy use that they replaced before 2010 at Portal, Arizona. We used four control plots and five kangaroo-rat exclosures from PortalData 6.117.0. We kept 391 matched, quality-controlled censuses from 1988 to 2025. Of these, 42 were from June 2021 to September 2025. Energy was estimated as 5.69 times capture mass to the power 0.75. Compensation was the pooled excess small-granivore energy on exclosures divided by control kangaroo-rat energy. Recent compensation was 29.0%, compared with 21.3% in January 2010 to January 2020 and 49.4% in July 1997 to December 2009. With 20,000 year and plot bootstrap draws, the recent-to-late change was +7.8 percentage points (95% interval −13.0 to +27.9). The recent-to-high change was −20.4 points (−48.3 to +8.2). Fixing the observed plots gave a recent-to-high interval below zero (−35.7 to −5.8 points). Bailey's pocket mouse supplied an estimated 23.4 points of recent compensation (6.7 to 42.8). Recent-to-late change remained unresolved in all planned sensitivity checks. The recent total energy ratio was 59.2%, but it included the baseline small-granivore share and kangaroo rats caught inside exclosures. In the eight retained 2025 censuses, those kangaroo-rat captures supplied 30.6% of the exclosure energy estimate. The earlier mass-fill rule reproduced all 3,078 archived plot totals. Raw inputs, code, hashes, results and figures are included. Limits are one site, five recent sampled years, the survey gap, plot differences, and a capture-based energy proxy. The data do not establish recovery or equal compensation across periods.

Portal Project · ecological compensation · species removal · desert rodents · community energy · temporal extension · bootstrap · Chaetodipus baileyi · sampling effort · open data