Low precipitation due to climate change consistently reduces multifunctionality of urban grasslands in mesocosms

Abstract

Urban grasslands are crucial for biodiversity and ecosystem services in cities, while little is known about their multifunctionality under climate change. Thus, we investigated the effects of simulated climate change, i.e., increased [CO2] and temperature, and reduced precipitation, on individual functions and overall multifunctionality in mesocosm grasslands sown with forbs and grasses in four different proportions aiming at mimicking road verge grassland patches. Climate change scenarios RCP2.6 (control) and RCP8.5 (worst-case) were simulated in walk-in climate chambers of an ecotron facility, and watering was manipulated for normal vs. reduced precipitation. We measured eight indicator variables of ecosystem functions based on below- and aboveground characteristics. The young grassland communities responded to higher [CO2] and warmer conditions with increased vegetation cover, height, flower production, and soil respiration. Lower precipitation affected carbon cycling in the ecosystem by reducing biomass production and soil respiration. In turn, the water regulation capacity of the grasslands depended on precipitation interacting with climate change scenario, given the enhanced water efficiency resulting from increased [CO2] under RCP8.5. Multifunctionality was negatively affected by reduced precipitation, especially under RCP2.6. Trade-offs arose among single functions that performed best in either grass- or forb-dominated grasslands. Grasslands with an even ratio of plant functional types coped better with climate change and thus are good options for increasing the benefits of urban green infrastructure. Overall, the study provides experimental evidence of the effects of climate change on the functionality of urban ecosystems. Designing the composition of urban grasslands based on ecological theory may increase their resilience to global change.

Fig 1. Urban road verge grasslands mimicked in mesocosms in climate chamber experiment.

(a) Road verge grasslands often occur on heavily sealed ground with shallow soil and using few grass species. (b) Restored grasslands in road verges are species-rich and adapted to urban conditions to improve ecosystem services delivered. (c) In the mesocosm experiment, forb species used for restoring urban road verges were sown, and the communities developed under similar conditions (e.g., substrate characteristics, limited soil depth, and surrounding sealed space). (d) Mesocosm grasslands of varying functional compositions were submitted to two climate change scenarios in the climate chambers of an ecotron facility. Pictures of road grasslands (a) and (b) were taken in Munich, Germany (courtesy of Simon Dietzel); pictures (c) and (d) were taken at weeks 9 and 10 of mesocosm development in chambers submitted to RCP8.5 and RCP2.6, respectively. For a better overview of the communities’ development under the two tested scenarios, see S2 Fig.

Fig 2. Effects of forb proportion, precipitation, and climate change scenarios on indicators of grassland functioning.

(a) Aboveground biomass, (b) belowground biomass, (c) floral abundance, (d) vegetation height, (e) cover, and (f) soil respiration rate (means ± 95% confidence intervals) of mesocosm communities in the climate chambers of an ecotron facility. Data points are light-colored circles (n = 4).

Fig 3. Effects of forb proportion, precipitation and climate change scenarios on water regulation in grasslands.

(a) Fraction of retained, and (b) lost water by evapotranspiration [ET] (means ± 95% confidence intervals) of mesocosm communities in the climate chambers of the ecotron facility. A heavy-rain event was simulated to derive information on water regulation function. Data points are light-colored circles (n = 4).

Fig 4. Effects of forb proportion, precipitation and climate change scenario on grassland multifunctionality.

Two approaches to multifunctionality (MF) were calculated using eight indicator variables, including either aboveground biomass (a, b) or floral density (c, d) of mesocosm communities in the climate chambers of an ecotron facility. Averaged standardized values of the individual indicators contributed to averaged multifunctionality, whereas for the selected threshold, each function indicator exceeding 70% of the standardized maximum contributed to multifunctionality. Shown are means ± 95% confidence intervals. Data points are light-colored circles (n = 4).