Green roofs are a way for cities to mitigate environmental stressors, such as heatwaves and droughts. However, these environmental stressors can adversely affect green roof vegetation, causing challenges for plant growth and survival and subsequently reducing the ability of green roof systems to deliver critical ecosystem services, such as heat mitigation and nutrient cycling. Plant-associated microbes may facilitate the resilience and tolerance of green roof vegetation to climate-associated stress. However, despite their crucial role in plant growth and survival in natural ecosystems, there has been little research on plant-associated microbes in green roof systems. Plant choice on green roofs may also determine which microbes established in green roof growing media, and particular plant-microbial combinations may be more resilient to environmental stress. This project sought to characterize soil microbial community composition on green roofs across New York City with different plant palettes and assess how different combinations of green roof plant species and root-associated microbial assemblages responded to isolated and simultaneous heat and drought treatments. We surveyed green roofs planted with either Sedum species or with a mixedvegetation palette (i.e., wildflowers, grasses, and succulents). We found that mixedvegetation and Sedum green roofs had distinct soil bacterial and fungal communities (p < 0.0001) with a higher relative abundance of mycorrhizal fungi on mixed-vegetation roofs, and higher pathogen loads on Sedum roofs. Concurrently, we conducted a greenhouse experiment in which plants were grown from seed with live inocula collected from the two different types of vegetation on the green roofs we surveyed. Hoch et al. Green Roof Soil Microbial Communities We observed that plant species, soil inoculum, and abiotic stress treatment was correlated with shifts in soil fungal communities. This study demonstrated that soil microbial assemblages on green roofs are linked to the roof vegetation, and that they may facilitate green roof plants' tolerance and resilience to environmental stressors.

The projected increase in warming and drought severity (hotter and drier summers) in the U.S. Pacific Northwest (PNW) may negatively impact grassland plant composition and ecosystem function, with further implications for sustainable land management in the region. To test the vulnerability of Mediterranean grassland function to climate change, we quantified the response of grassland communities to multiannual warming (+2.5C) and drought (-40% precipitation) by quantifying plant species diversity, legume cover, and biogeochemical controls on and patterns of soil asymbiotic nitrogen fixation (ANF). We hypothesized that the effects of warming on plant functional diversity would increase soil ANF inputs by decreasing legume cover and soil nitrogen availability. Given that asymbiotic N fixers can increase soil organic carbon (C) and nitrogen (N) availability under drought, we hypothesized that the effect of drought on grassland plant cover correlated with increased soil ANF. We surveyed the vegetation and collected composite soil samples from five co-located plots under control (ambient), drought and warming conditions during the fall and spring seasons. In control and drought plots, we quantified the moderator effect of plant composition by comparing low-diversity (unmanipulated plant composition) and high-diversity (manipulated composition) grassland plots. We used a point intercept technique to survey plot-level plant community composition and calculate Shannon diversity index and percent cover of legumes (members of Fabaceae according to the Integrated Taxonomic Information System). We measured ANF by incubating collected soils with N-labeled dinitrogen (15N2), and quantified total soil C, total and available N, available phosphorus (P) and iron (Fe) pools, pH, and soil water holding capacity. Plant species diversity decreased significantly with warming and along the drought severity gradient. ANF response to warming varied by season and site, with rates increasing along the drought severity gradient in the fall but decreasing in the spring. Total soil inorganic N was the strongest predictor of ANF response to warming in the spring but not in the fall. Soil ANF response to drought increased with drought intensity; while soil ANF increased nearly twofold in the southernmost (warm and dry) site, ANF decreased in the northernmost (cool and wet) site. ANF response to drought also varied depending on plant diversity, where low-diversity grasslands had more predictable response to drought than high-diversity grasslands. Soil P availability and pH were the most important variables explaining ANF variability across vegetation types and sites. Our study highlights the importance of using soil-plant-atmosphere interactions to assess grassland ecosystem resilience to drought in the PNW.