Natural interannual variability in precipitation complicates revealing drought effects on restored communities

Background/Question/Methods

Restoring degraded lands aims to increase plant diversity and ecosystem functions, but variation in climate contributes to highly variable outcomes. Climate alters plant composition and ecosystem functioning in native prairie, but the effects of climate variability, particularly drought, on developing prairie are not well understood. We aimed to elucidate the effect of drought on plant community development and productivity in restored tallgrass prairie by installing rainfall reduction (66% interception) shelters paired with ambient rainfall control plots at the onset of a restoration. Rainout shelters were erected during the growing season for the first three years of restoration (2018-2020). An unexpected natural drought occurred in the first year of the experiment, so we used data from restorations established in 2010 (average precipitation) and 2012 (another natural drought) using identical restoration and data collection methods. Plant species composition and richness was measured in the first three years of all restorations by estimating the maximum percent cover of 1 m2 permanent quadrats. Aboveground net primary productivity (ANPP) was measured in 0.1 m2 quadrats. Composition was analyzed with a repeated measures permutational multivariate analysis of variance (PERMANOVA), and visualized using non-metric multidimensional scaling (NMDS). Species richness and ANPP were analyzed with mixed models.

Results/Conclusions

In the drought experiment, age significantly affected plant composition (P=0.001), richness (P< 0.001), and ANPP (P< 0.001), but there was no effect of drought treatment. We attributed no effect of drought treatment to the severe natural drought in the establishment year that caused a drought signature in the ambient precipitation treatment. Comparison of similar-aged communities restored in different years showed that community composition was affected by an interaction between restoration year and age (P=0.001). All communities changed with age and were separated by establishment year corresponding with variation in precipitation. Richness was also affected by an interaction between restoration year and age (P=0.001). Richness increased over time in all restorations, but richness in the prairie restored under average precipitation was higher than prairies restored in drought conditions. Total ANPP was affected by an interaction between restoration year and age (P=0.001), increasing with age of restoration and annual precipitation. Comparing prairies restored under contrasting climate revealed that drought in the establishment year can have large effects on short-term restoration outcomes, whereas, the effects of drought through direct manipulation of rainfall is complicated by climate variability creating more frequent drought years that can mask the effect of imposed drought treatments.