Given the frequent overlap between biological plant invasion and ecological restoration efforts it is important to investigate their interactions to sustain desirable plant communities and modify long-term legacies both above and belowground. To address this relationship, we used natural reference, invaded, and constructed vernal pools in the Central Valley of California to examine potential changes in direct and indirect plant effects on soils associated with biological invasion and active restoration ecosystem disturbances. Our results showed that through a shift in vegetation composition and changes in the plant community tissue chemistry, invasion by non-native plant species has the potential to transform plant inputs to soils in vernal pool systems. In particular, we found that while non-native litter decomposition was driven by seasonal and interannual variability, associated with changes in precipitation, the overall decomposition for non-native litter was drastically lower than native species. This shift has important implications for long-term alterations in plant-based inputs to soils in a negative feedback to nutrient cycling. Moreover, these results were independent of historic active restoration efforts. Despite the consistent shift in plant litter decomposition rates and community composition, we did not detect associated shifts in belowground function associated with invasion by non-native plants. Instead, soil C:N ratios and microbial biomass did not differ between invaded and reference naturally occurring pools but were reduced in the manipulated restored pools independent of invasion levels. Our results suggest that while there is an observed invasive positive feedback aboveground, this trajectory is not necessarily represented belowground and restoration legacies were still dominant ten years after practices were applied. Restoration practices that limit invasive plant feedbacks and account for soil legacy recovery, therefore offer the best solution for disturbed ephemeral ecosystems.

All data presented here were collected associated from a long-term restoration vernal pool field site located on Travis Air Force Base, in Solano Co. California. Field samples  including plant materials and soils were collected and processed as described in the associated paper. 

There are datasets associated with this paper, associated with different types of samples and sample processing. These include: leaf litter decomposition data (ChurchillFaist_Decomposition), soil properties (ChurchillFaist_SoilMetrics), plant community composition (ChurchillFaist_PlantMultiVariate), and plant tissue chemistry data (ChurchillFaist_PlantUniVariate). Column names and explanations are listed below for each discrete dataset.

ChurchillFaist_Decomposition: Dataframe containing weights associated with litter bag deployments, Year- refers to the year decomposition study was conducted, Pool- an ordinal assigned to identify the on-site location of the pool sampled, PoolType- a classification (factor) associated with the type of plant community, Location- a category within each pool associated with the position of litter bag relative to the microtopography of each pool (bottom, transition, edge), LitterPlacement- a category of either above or below the existing litter layer as a placement for each litter bag, Species- the species of litter used for decomposition; LOMU is invasive, PLCA is native, Season- refers to the time of year when litter bags were deployed (winter or summer), PropLost- the proportion of the litter mass lost relative to the original weight, PercentMassLoss- the percent of litter mass lost relative to the original weight

ChurchillFaist_SoilMetrics: Dataframe associated with soil samples collected during April 2011, Pool- an ordinal assigned to identify the on-site location of the pool sampled, PoolType- a classification (factor) associated with the type of plant community growing in the pool, Location- a category within each pool associated with the position of litter bag, SoilCN- ratio of soil carbon to nitrogen for soils collected in 2011, SM- soil moisture of subsample collected simultaneous to other metrics (%), pH- soil pH measured using 1:3 soil to water ratio, Sand- soil texture fraction component sand (%), Silt- soil texture fraction component silt (%), Clay- soil texture fraction component clay (%), MicrobialBiomass- microbial biomass (mg Carbon / g dry weight soil)

ChurchillFaist_PlantMultiVariate: Dataframe associated with plant count data measured in the growing season of 2011, Year- refers to the year species composition surveys were conducted, Pool- an ordinal assigned to identify the on-site location of the pool sampled, PoolType- a classification (factor) associated with the type of plant community growing in the pool, Location- a category within each pool associated with the position of litter bag relative to the microtopography of each pool (bottom, transition, edge), All remaining columns are counts associated with individual plant species codes. 

ChurchillFaist_PlantUniVariate: This dataframe contains plant chemistry trait information for species present at the field site; samples collected in 2011 (Note that not all species included here were present in 2011, and there are species also present in 2011 that were not sampled for chemistry due to limited sample availability), PlantType- category for either Native or Invasive type of plants, FxnlGroup- functional group for plant species, either grass or forb in this dataset, Species- scientific name of plant species, Code- Shorthand abbreviation for species used in PlantMultiVariate dataframe, Nitrogen- plant tissue percent nitrogen, Carbon- plant tissue percent carbon, Cellulose- plant tissue percent cellulose, Lignin- plant tissue percent lignin, CN- ratio of plant tissue carbon and nitrogen, LN- ratio of plant tissue lignin and nitrogen, Cell_N- ratio of plant tissue cellulose and nitrogen. Codes for species that were present in count surveys but not analyzed for tissue chemistry are also included here.