Beskrivelse
Study sites Our study sites were located in Kastbjerg Ådal (river valley) in Eastern Jutland, Denmark. It is within the Natura 2000 and habitat area no. 223 appointed because of the wide stretch of fens and mires among other qualities. The water course is in good ecological status according to the Water Framework Directive. Nitrogen deposition in this area is low to moderate, 12.5-14.5 kgN/ha/yr (Ellermann et al. 2021). Meadows and fens dominate the study area, known for ‘the longest stretch of rich fen’ in Denmark. Large parts of the river valley are heavily degraded by drainage, fertilization and scrub encroachment, but there have also been recent efforts to restore the watercourse and the valuable rich fens in the valley. Most fens and wet meadows have been abandoned and are now increasingly dominated by tall grasses, tall forbs and willow scrub, but summer grazing occurs in some areas and efforts are made to ensure grazing in the most valuable fens. The drier meadows are typically mown by heavy machinery. The sites were selected to represent gradients in soil moisture from moist to wet and gradients in nutrient status or productivity from poor to rich and included rich fens with characteristic species, fens dominated by Juncus subnodulosus and by Equisetum fluviatile, drained fens encroached by Phragmites australis and natural meadows with characteristic species and encroached by Epilobium hirsutum and meadows characterized by clovers and cultural grasses. The nine sites were of 10 m2, each with ten 1 m2 plots. The 10 plots within each site had treatments assigned randomly. Despite the location in the same river valley, the sites were considered independent because of their different management history and starting conditions and a typical inter-site distance of c. 225 meters. The experiment was established in June 2017 and treatments were repeated monthly during summer and bimonthly during winter, depending on treatment. Responses were recorded in July 2019. Experimental set-up and treatments Each of the 9 sites were divided into ten 1 m × 1 m plots each with a 0.5 m × 0.5 m inner square and a surrounding plot buffer zone with a control and the following treatments: burning, mowing, trampling, intensive summer grazing (SI), intensive summer grazing with trampling (SIT), extensive summer grazing (SE), extensive summer grazing with trampling (SET, year-round grazing (YR), and year-round grazing with trampling (YRT). Treatments were allocated randomly to each plot with the restriction that the control plot was always in one corner. The experiment was multifactorial with respect to grazing and trampling, whereas burning and mowing were stand-alone treatments. Initial biomass in each plot was estimated at the beginning of the experiment in June 2017 as follows: all standing biomass and litter was removed from the plots by manual cutting at the soil surface and following the micro-topography. Bryophytes were harvested by hand plucking. Biomass, litter and bryophytes from the plot buffer zone were cut separately from the inner square. To estimate the species abundances, a representative sample of the inner square was sorted into litter and live biomass (including bryophytes) by species as sorting the complete biomass was not feasible. All species, litter and biomass from the buffer zone were dried at 55° C and weighed. Using the relative abundance of species in the representative sample and with respect to the weight of the total biomass in the inner square, we estimated the abundance of the species in the inner square. Burning was simulated in March 2018 and 2019. We used wooden boards to shield and adjacent areas were watered before burning the focal plot with a gas weed burner. We burned on a calm day following a dry period with frost to ensure minimum risk of igniting underlying peat and fire spreading over ground, but ensuring that the standing biomass and litter would be dry enough to ignite. This is not a simulation of a naturally occurring wildfire, but corresponds to the conditions that managers would prefer for prescribed conservation burning at larger scales. We simulated mowing as a biomass removal in June 2018. Biomass was removed uniformly across the whole plot in a height of c. 5 cm depending on microtopography. This corresponds to conservation mowing in management but without the added disturbance and pressure from machines. Trampling disturbance was applied using short stilts that could be attached to the field biologist’s boot. The surface of the stilt was 49 cm2 which corresponds to a pressure of 1.3-1.5 kg/cm2 with the added weight of the field biologist. This again corresponds to the pressure of a hoof of cattle weighing c. 300-400 kg. Trampling was applied by stepping into the field randomly 60 times once every month from May to September and was the same treatment in combination with intensive, extensive and year round grazing. Grazing was simulated by cutting the above-ground biomass using a 1 m2 frame divided into a 10 cm coordinate system using the letters A-J on the x-axis and the numbers 1-10 on the y-axis. We cut tufts of biomass within the coordinate system using a list of random combinations of letters and numbers. This system enables “ungrazed” individuals to flower and set seeds. Based on our experience with grazing as an agri-environmental management practice in Denmark, we defined intensive summer grazing as taking place between May and September with the goal of removing all standing biomass by September. Extensive summer grazing also takes place May-September, but we carried this out at half the intensity as intensive summer grazing. Year-round grazing obviously takes place during the whole year (here administered May-September and November, January and March) with the goal of removing all standing biomass by the end of winter (March) before the beginning of a new growing season. We used the initial standing biomass (June 2017) as a measurement of plot productivity and estimated the amount of biomass to be removed during “grazing” as c. 20 % of the initial productivity each month May-September in intensive plots and with all standing biomass “grazed” in September. For extensive plots, we estimated removed biomass as c. 10 % of the initial productivity each month May-September leaving some standing biomass in September. Year-round grazing biomass removal was estimated as c. 10 % of yearly productivity removed every month May-September and November and 20 % removed in January and March resulting in no standing biomass at the end of the winter. As expected plot productivity changed as a result of the treatments, the amount of biomass removed had to be adjusted throughout the experiment. In practice, we aimed for removing twice the amount of biomass in intensive plots relative to extensive plots within the same site and always ensuring that no standing biomass was left in intensive plots in September, c. 50 % of the standing biomass was left in extensive plots in September and no standing biomass was left in year-round grazing plots in March (see actual removed biomass by treatment in Appendix A). All treatments were applied to the whole plot (1 m × 1 m), while the biomass response was only measured in the inner square (0.5 m × 0.5 m), leaving a buffer zone between plots with different treatments. Response variables A full plot (1 m × 1 m) species list was recorded in the field at the end of the experiment. From this total plot richness, vascular plant plot richness, bryophyte plot richness and number of indicator species per plot were calculated. Indicator species of conservation status are species considered moderately to very sensitive towards habitat degradation as defined by Fredshavn et al. (2010, see Appendix C). Indicator species are often adapted to relatively infertile habitats revealed by low Ellenberg N values and high Grime’s S values reflecting tolerance to nutrient shortage. Mean plot Grime’s C and S values (Grime et al. 1989) were calculated based on vascular plant species lists. We converted Grime’s life strategies to numerical values based on Ejrnæs and Bruun (2000). We performed a Nonmetric Multi-dimensional Scaling analysis (NMDS) on the presence-absence of vascular plant and bryophyte species at the end of the experiment using the function metaMDS in R-package ‘vegan’ (Oksanen et al. 2017) in R version 4.0.3 (R Core Team 2017), using Sørensen dissimilarity and a four-dimensional solution (k =4). The plot coordinates at the three first NMDS axes were extracted (NMS4 was discarded as noise) and these, along with the four richness variables as well as Grime’s C and S values, were used as response variables in Linear Mixed Models (LME) as described in ‘Statistical analyses’. Supplementary to regression models of single response variables we carried out a quadratic discriminant analysis (QDA) as described in ‘Statistical analyses’ using the change in six indicators during the course of the experiment. The difference between plot species richness at the beginning and end of the experiment was calculated based on the species lists from sorted initial biomass and end biomass (0.5 m × 0.5 m). Start-end differences were also calculated separately for vascular plant species richness, bryophyte species richness, richness of indicator species, the ratio between biomass of forbs and graminoids (grasses, sedges and rushes) and Grime’s C and S mean site values. Explanatory and co-variables Leaf nitrogen, carbon and phosphorous were determined from plot level sampling of leaf plates of grasses, i.e., the most abundant species group across sites. Fresh leaf plates were collected at the beginning and end of the project and then dried, ground and analyzed in the lab. Soil moisture (% volumetric water content) was measured as the mean of four measurements per plot at the beginning and end of the project using a FieldScout TDR 300 Soil Moisture Meter. The total number of species found in each site was used as a co-variable in species richness models reflecting the local species pool. Data processing All species names were checked for synonyms using the national database arter.dk. References: Ejrnæs, R. and H. H. Bruun (2000). "Gradient analysis of dry grassland vegetation in Denmark." Journal of Vegetation Science 11(4): 573-584. Ellermann, T., R. Bossi, J. Nygaard, J. H. Christensen, P. Løfstrøm, C. Monies, C. Geels, I. E. Nielsen and M. B. Poulsen (2021). Atmosfærisk deposition 2019. NOVANA. Aarhus, Aarhus Universitet, DCE - Nationalt Center for Miljø og Energi. Fredshavn, J., R. Ejrnæs and B. Nygaard (2010). "Teknisk anvisning for kortlægning af terrestriske naturtyper. TA-N3, Version 1.04. Fagdatacenter for Biodiversitet og Terrestriske Naturdata, Danmarks Miljøundersøgelser. 18 s. ." Grime, J. P., J. G. Hodgson and R. Hunt (1989). Comparative plant ecology: a functional approach to common British species. London, Unwin Hyman. Oksanen, J., F. G. Blanchet, R. Kindt, P. Legendre, R. B. O'Hara, G. L. Simpson, P. Solymos, M. H. H. Stevens and H. Wagner (2017). "Package 'vegan': Community Ecology Package. Version 2.4-3. http://cran.r-project.org/web/packages/vegan/vegan.pdf." R Core Team (2017). R: A language and environment for statistical computing. Vienna, Austria, R Foundation for Statistical Computing.
Dato for tilgængelighed | 2022 |
---|---|
Forlag | Zenodo |