Fig 1.
Study region and monitoring design.
(a) Locations of the three study sites in Yunnan Province, southwestern China, representing temperate coniferous forest (CF), subtropical evergreen broadleaved forest (SF), and tropical forest (TF) along a climatic gradient. Provincial boundary data were obtained from Natural Earth (public domain; http://www.naturalearthdata.com/). (b) Schematic of the edge-to-interior transect design showing the open-area reference plot (plot 0) and forest plots at 1.5, 4.5, 12.5, 36.5, and 99.5 m from the edge. Each plot was a 3 m × 3 m quadrat. (c) Installation of the HOBO MX2301A logger used to measure air temperature at 1.5 m height. Photographs in panel (c) were taken by the authors.
Table 1.
Fixed effect estimates from log distance mixed effects models describing variation in maximum cooling intensity across edge distance, forest site, and their interaction at annual, seasonal, and monthly aggregations.
Fig 2.
Predictions of maximum cooling intensity as a function of edge distance at monthly and seasonal scales.
MCI is defined as the 5th percentile of daytime hourly temperature offsets derived from 10 min records, and more negative values indicate stronger cooling. Lines show model predictions for each forest site and shaded bands show 95% confidence intervals. Points show daily values for each distance along the transect.
Table 2.
Spatiotemporal variation in DEI across forest sites at monthly, seasonal, and annual scales.
Fig 3.
Predicted buffering slopes along the edge to interior gradient at monthly and seasonal scales.
Slopes were derived from regressions of lower tail daytime air temperatures in forest plots against the adjacent open reference. Values of the slope below 1 indicate buffering, and smaller values indicate stronger decoupling. Lines show fitted relationships with log transformed edge distance and shaded bands show 95% confidence intervals. Black markers indicate DEI, defined as the smallest distance at which the fitted slope enters the 95% confidence interval of the interior reference at 99.5 m.
Fig 4.
Maximum cooling intensity in coupled edge and decoupled interior zones across the climatic gradient.
Zones were defined relative to DEI for each forest and time window, and comparisons were limited to windows with detectable DEI. MCI represents the lower tail extreme of daytime temperature offsets, and more negative values indicate stronger cooling relative to the open reference. Boxplots show the median and interquartile range, and whiskers show the full range. Lowercase letters indicate significant differences between coupled and decoupled zones within a forest site, and uppercase letters indicate significant differences among forest sites within a zone.