Tree death counts showed excess variation among 20 m cells after species, size and 100 m block control in seven BCI census intervals
We tested whether species, starting tree size and broad location explain local variation in tree deaths in the Barro Colorado Island (BCI) 50-ha forest plot, Panama. We used eight censuses labelled 1982 to 2015. The selected data have 1,578,430 tree intervals from 394,360 trees, with 174,849 deaths. For each of seven adjacent census intervals, we kept the death total fixed within each species, diameter class and 100 m block. We then assigned deaths at random to trees within each group. Observed variation among 20 m cells was 2.50 to 3.24 times the exact expected variation under this model. Each interval had a Monte Carlo p value of 0.001 from 999 draws, and a Holm adjusted value of 0.007. Excess variation remained with 10 m or 50 m cells, trees at least 10 cm in diameter, and removal of each 100 m block in turn. An added check for half-year follow-up duration bins gave ratios of 2.40 to 3.17. The package has selected tree outcomes, source hashes, code, figures and checks. This is a conditional test in one forest. Fine habitat, tree condition and shared damage can all produce this result. The test does not identify a cause or prove that deaths are dependent after all risk factors are known.
tree mortality · spatial variation · Barro Colorado Island · conditional randomization · forest dynamics · open data
References
6- usesdoi:10.15146/5xcp-0d46
Source tree census archive, 2019 version, CC0.
- usesIn the Barro Colorado Island 50-ha plot (1982–2015), the stem shift toward dry-site tree species stopped by 2000–2005, as a drought legacy predicts, and did not continue as drying predicts sci:sha256:3693047d517c5acd0d809909ac925a8e5fda1a6227686a3f65e2996f56fc16e1
Reuses the retained unchanged tree census archive and source provenance; examines a separate question.
- doi:10.2307/2963497
- doi:10.1186/s40663-017-0103-1
- doi:10.1046/j.1461-0248.2003.00492.x
- doi:10.5479/si.26882632
Cited by
3 publicationsPublic catalog entries on Sciscape.
- usesExcess tree deaths in 20 m cells at Barro Colorado Island were positively correlated across adjacent census intervals after species, size and 100 m block control, 1985–2015
We reuse selected tree outcomes, record selection rules and source preparation code. We recalculate all conditional cell statistics and test temporal association, which the earlier work did not establish.
- usesAdjusted tree death maps had positive correlations between adjacent census intervals in 20 m cells at Barro Colorado Island, 1982–2015
We reuse its selected tree-interval outcomes and source reconstruction code. We extend its within-interval spatial benchmark with adjacent-interval comparisons.
- The mean loss of occupied 20 m cells did not persist at 100 individuals for 130 Barro Colorado woody species, 1985–2015
Prior test of spatial variation in tree deaths in the same forest plot; distinct from the occupancy target.