In nine Portal plots, the energy compensation score was 27% in 2021–2026 and 49% in 1997–2010

We tested if low compensation for kangaroo-rat exclusion persisted after the January 2020 cutoff of an earlier study. We used PortalData 6.117.0 and nine plots with fixed rodent treatments. We estimated relative energy use from body mass. The score is the extra small-granivore energy on exclosures divided by kangaroo-rat energy on controls, with sums over matched censuses. The later accepted data span June 2021 to January 2026. They contain 42 matched censuses in five annual blocks. The score was 26.7%, compared with 49.2% in July 1997 to January 2010 and 21.5% in February 2010 to January 2020. The later-minus-1997–2010 difference was -22.6 percentage points. A 95% interval from year blocks, with plots fixed, was -37.0 to -7.7 points. An added interval that also resampled plots was -48.3 to +5.8 points. Thus the lower estimate is supported for these fixed plots, but a decrease is not established under the broader plot-resampling model. Bailey's pocket mouse supplied most of the compensation in both periods. Biomass, fixed species masses, month balance, and source quality checks kept the same direction. We removed redundant source rows and preserved the audit. The package contains raw data, hashes, code, results, figures, and offline reproduction steps. Limits include one site, few plots, sampling gaps, imperfect exclusion, and a mass-based energy proxy. This is a data extension with a different estimator, not an exact replication or proof of a new community state.

Portal Project · desert rodents · species loss · functional compensation · long-term experiment · energy use · plot variation · data quality · reproducible ecology