Livestock grazing is often used as a land management tool to maximize vegetation diversity in grassland ecosystems worldwide. Prior research has shown that cattle grazing benefits native annual forb species in California’s coastal prairies, but drought and increasing aridity may alter this relationship. In 2016 and 2017, we resurveyed the vegetation structure, native annual forb cover, and native annual forb richness in ten grazed and ungrazed prairies that were originally measured in 2000 and 2001 along a 200-km gradient from Monterey to Sonoma counties in California. We found that grazed prairies continued to have significantly lower vegetation height and thatch depth than ungrazed prairies, and that shrub encroachment over the 15-year period was significantly greater in ungrazed prairies. Furthermore, grazed prairies continued to have greater native annual forb richness (4.9 species per site) than ungrazed sites (3.0 species per site), but that native annual forb richness declined by 2.8 species per site in grazed prairies and 0.1 species per site in ungrazed prairies between survey periods. We suggest that severe drought and increasing aridity may be driving declines in native annual forb richness in grazed prairies. The species we recorded only in earlier surveys were disproportionately wetland-associated and had higher average specific leaf area than species that remained through the second survey period. Finally, cover of native annual species increased regardless of whether prairies were grazed, suggesting that the high precipitation in 2017 may have benefitted the native annual forb species that persisted at sites between surveys. Our study shows that weather conditions affect the outcomes of land management strategies.

In 2000–2001, we surveyed 26 paired grazed and ungrazed coastal prairie sites across 400 km of coastline between San Luis Obispo and Mendocino counties in California, USA. In 2016–2017, we resampled ten sites along 200 km in the central area of the prior survey extent (Monterey to Sonoma Counties) where we could obtain permission from landowners for access and grazing patterns had been maintained since 2000-2001.

At all sites, paired plots were less than 2 km apart, on similar slopes and aspects, and transects were placed at least 5 m from fence lines to avoid edge effects. Cattle were the only domestic grazing animals present at the sites, there was no evidence of prior mechanical soil disturbance, and soils were deeper than 40 cm and not ultramafic (serpentine). We sampled from April through May, traveling from southern sites to northern sites to follow the peak flowering phenology of most species. At each site, we relocated T1 50-m line transects using a Garmin eTrex 20 GPS. There were five transects each in the grazed and ungrazed portions at each site. 

We quantified community composition by recording the identity of each species that intersected a 1.8-mm-diameter pin at 1-m intervals along the transects. We recorded each species that intercepted the pin once. Our nomenclature and species origin information follow the Jepson Manual and online Jepson eFlora database. At 5-m intervals, we measured vegetation height using a piece of paper dropped onto foliage and recording the lowest point, and thatch depth (i.e., build-up of dead plant biomass) by pushing a 1.8-mm pin to the soil surface and measuring the tallest standing dead biomass. We calculated shrub cover as the number of woody shrubs intercepts along each transect as a percent of the 50 intercepts total along the transect. Finally, we carefully searched for and quantified the cover of all native annual forbs (to the nearest cm²) present within a 1-m belt transect centered over the 50-m transect.

We compiled specific leaf area (SLA, leaf area/leaf dry mass) for 40 of the 55 native annual forbs species from the TRY Plant Trait Database and data collected by others. We also collected data from species present in and near our sites during spring and summer 2018. Most of these data came from grazed plots, since this is where the species were more prevalent, though we selected individuals that showed no evidence of recent grazing. When data for a single species was available from multiple sources, we averaged the values, which are available here. We also gathered the US Army Corps of Engineer wetland indicator status of the native annual forbs.

All data are shared as .csv files.