A linear model of the 1997–2021 Cedar Creek data did not show lower rain sensitivity at higher planted richness

We tested whether planted species richness reduced the proportional biomass response to early-season rain in the Cedar Creek Biodiversity II grassland experiment. We rebuilt live biomass for 154 fixed plots in 1997–2021 and quality-controlled daily weather from three nearby stations. We used April–June rain so that the exposure preceded the harvests, including the early July harvest in 2008. We fitted annual richness slopes, then tested their rain response with controls for experiment age and heat. The rain test has 25 years, not 3,850 independent plot-year observations. Across all plots, geometric mean biomass increased by 7.4% per 100 mm of rain (95% interval 2.4% to 12.7%). The fitted biomass ratio of 16-species to one-species plots changed by +3.6% per 100 mm (interval −1.5% to +8.9%; p = 0.158). A negative change would indicate a smaller proportional rain response at greater richness. A direct group contrast and all planned checks also had intervals that included zero. The data did not show a clear decrease in rain sensitivity at greater richness. They do not prove equal sensitivity or test all forms of drought resistance. Rain was observational, the response model was linear, and the data came from one managed site. The package includes unchanged source files, provenance, code, all results, figures, a local plan, and independent checks.

biodiversity · grassland · rainfall · biomass · Cedar Creek · temporal replication · climate sensitivity