Low fruit crops at Lopé overlapped more than a five-year block model predicted for 40 species in 1986–2018

We asked if low fruit crops remain aligned after a model keeps slow changes in each species' low-crop frequency. We screened open MASTREE+ V2 records for groups with at least five taxa and 20 complete years. The screen retained 23 groups from eight source studies. Low crop meant annual output below 10% of each taxon's record mean. Sixteen groups had defined low-crop variance ratios. Their study-weighted geometric mean ratio was 1.477 relative to independent timing (95% study bootstrap interval 1.088–2.298). The largest effect was at Lopé National Park, Gabon: 40 species in 1986–2018. The variance of the low-crop species count was 29.72. Full-record independence predicts 6.01. A model that keeps each species' low-crop frequency within five-year blocks predicts 17.91. The observed value was 1.66 times this block-model expectation. None of 4,999 block-model draws reached it (upper-tail probability 0.0002; Holm-adjusted value 0.0032 across 16 group tests). The other five local groups did not pass the adjusted 5% threshold. The Lopé result held for crop thresholds of 5–25% and block lengths of 3–10 years. These site checks were exploratory. The records overlap earlier evidence of a long decline in Lopé fruiting. They do not provide independent evidence of that decline. The block model retains coarse changes, but shared trends within blocks and common measurement changes can remain. Visual crop indices do not measure food energy or animal responses. The package contains exact source bytes, balanced panels, code, figures, exact model checks, and reproduction instructions.

crop failure · fruit production · species synchrony · MASTREE+ · Lopé · conditional permutation · ecosystem dynamics · reproducible ecology