Diel patterns of soil CO ₂ efflux vary among and within forests

Uneven and sparse high‐resolution temporal coverage of soil CO 2 efflux ( R S ) data limit our understanding of its variability across ecosystems, leaving fundamental questions with applied consequences for collecting and scaling site‐level observations. Processes driving R S can operate on the scale of hours or days, but in forests, data are often limited to weekly morning observations based on long‐standing assumptions of diel patterns and what controls them. Here, we re‐examine these assumptions using a new database of hourly R S measurements from many forests, consider their implications for sampling, and then parse controls on diel R S at one experimental forest. Across 31 globally distributed forests, mid‐morning measurements could represent daily fluxes on average, but diel flux patterns at individual sites deviated substantially from this central tendency. For example, at our study site, monitoring only between 9:00 a.m. and 12:00 p.m. would underestimate annual R S ∼12% (± 1.4), but the error varied across ridge, midslope, and valley floor topographic positions. Co‐located biotic and abiotic data (root respiration, turbulence, soil moisture, soil temperature, water table depth, and soil pCO 2 ) suggest near‐surface soil temperature drove these diel patterns, but curiously, if temperature were a universal mechanism, we may have expected our global dataset to reveal more consistent results among sites. Exploring these controls furthers our fundamental understanding of what drives diel R S across forests and provides insights for improved collection and upscaling of local R S observations for global estimates.