Plant composition and weather data during tallgrass prairie restoration

Tallgrass prairie communities were restored using the same live seed mix every other year from 2010-2020 to test the hypothesis that dissimilarity in planting-year climate statistically predicts dissimilarity in developing plant community composition, and planting-year precipitation and temperature variables explain cover of species and major compositional underlying community dissimilarity. This dataset contains percent cover of plant species establishing over time in six restored prairies established in six different years at the Konza Prairie Biological Station in Kansas (USA) and corresponding precipitation and temperature data for each year a prairie was restored. The experiment is referred to as the “Sequential Restoration Plots” and contains seven prairie communities (sequences) restored in an agricultural field over time. Plant species, the absolute cover of each species, and cover of compositional groups (all sown species, all volunteer species, all forb species, all grass species, sown C4 grasses, sown C3 species, and sown forb species) over the first three years of community establishment in each plot and subplot of the six sequences corresponding to planting years 2010, 2012, 2014, 2016, 2018, and 2020 (sequences I-VI) are provided in this dataset. Weather data contained in this dataset correspond the years each prairie sequence was restored. Annual precipitation and temperature summary variables were calculated to relate cover of species and compositional groups to planting year climate. Climate vectors (capturing intra-annual variability and extremes) were also created for each planting year to perform multiple regression on distance matrices to examine the relationship between dissimilarity in planting year precipitation and temperature conditions and dissimilarity in communities that developed over the first three years of restoration.

Methods
Plant communities were restored every other year in a former agricultural field from 2010 to 2020 at the Konza Prairie Biological Station (KPBS) in Riley County, Kansas USA (39° 06’ 09.69” N, 96° 36’ 07.21” W). The site had been cultivated for >70 years and contains soil classified as a silt loam formed by colluvial and alluvial deposits. Crop production was ceased in a 30-m x 100-m area (referred to as a sequence) of the agricultural field in the fall (October) prior to sowing prairie communities in the spring (first week of June) of a planting year (January-December). In each sequence, seeds of native plant species were sown into four 20-m x 20-m plots containing four 10-m x 10-m subplots. Plots were separated by a distance of 5 m. The logistical constraints of requiring continuous cultivation to promote the same starting soil conditions and the same prescribed fire regime for each plot restricted randomization plots assigned to different planting years throughout the study site. Thus, each sequence was installed adjacent to the sequence established the previous planting year. To foster statistical independence, we combined and mixed seeds separately for each plot and separated plots with 5-m buffer strips that were sown with a different composition of species.

Plant species composition surveys were conducted in permanent circular sampling areas in each subplot. Species composition was recorded in late spring and late summer each year after planting to capture maximum cover of early and late season species. Only late-summer sampling was carried out in the planting year. To be consistent with previous studies from this experiment (Manning and Baer 2018, Eckhoff et al. 2023), the cover of each species was visually estimated in 10 m2 sampling areas (1.78 m radius). Cover estimates were then assigned to a modified Daubenmire cover class (1 = 0-1%, 2 = 1-5%, 3 = 5-25%, 4 = 25-50%, 5 = 50-75%, 6 = 75-95%, 7 = 95-100%). The dataset contains the maximum cover class value of the two sampling times within a year, converted to the midpoint of each Daubenmire cover class range.

Climate variables were calculated from weather station data collected at KPBS. Data are accessible through the Konza Prairie LTER Data Repository (AWE01 dataset). From this dataset, planting-year climate summary variables were calculated from daily precipitation (mm) and temperature (°C) records. The selection of climate summary variables was based on potential to explain establishing plant species and communities. We chose precipitation and temperature in June and July (JJ) over growing season metrics because first-year plantings were not sown until the last week of May, and weather occurring in the first few months following sowing is critical to seed germination and establishment. Total annual precipitation, JJ precipitation, average yearly temperature, and JJ average temperature were calculated for each planting year.

Climate vectors (capturing intra-annual variability and extremes) were created for each planting year to perform multivariate analyses. For precipitation, we created vectors of total weekly and monthly precipitation throughout the entire year for each planting year, as well as vectors of total daily and total weekly precipitation during JJ of each planting year. For temperature, we created vectors of average weekly and monthly temperature throughout the entire year for each planting year, as well as vectors of average daily and weekly JJ temperature during each planting year.