Soil is a hidden ecosystem which harbours plant roots and countless microorganisms, vital for sustaining life aboveground. These belowground communities provide essential ecosystem services like soil stabilisation and organic matter decomposition. Soil is also one of the largest terrestrial carbon repositories, and land management strategies aimed at increasing organic matter inputs from plants, such as compost additions, can promote further soil carbon accumulation. Because organic carbon is important for soil water retention, this management may also help to increase resilience against more frequent and intense droughts. Although roots and microbial communities are largely acknowledged to play a key role in regulating the carbon cycle, there are still many open questions regarding the link between above- and belowground processes and ecosystem functions. Observing climate- and management-driven changes in the soil habitat is fundamental for understanding how ecosystems respond to environmental change.

The aim of this thesis is to explore the relationship between soil properties, plant communities, and soil microbial communities in response to environmental changes. The research builds on a meta-analysis of drought effects on grasslands, and a multifactorial field experiment which combined three years of precipitation reduction and a compost treatment in two Swedish grasslands. We analysed the response of roots and soil microbial communities to drought and compost amendments, and identified environmental factors behind their large spatial variability. Finally, we tested the effects of compost additions on soil carbon storage and its interactions with drought.

The results of the meta-analysis indicate that, on a global scale, grassland roots and shoots have diverging responses to drought duration and intensity, with long-term climate mediating that difference. At the local scale assessed in the field experiment, we observed that the spatial patterns of soil microbial communities were driven by soil properties and vegetation. Growing season drought affected roots only at trait level, but did not significantly affect microbial communities. Positive effects of compost on aboveground plant productivity and fungal growth were detectable after three years. Compost amendments also increased the percentage of total soil carbon, but no net increase in soil carbon stocks was detected. Spatial variability in roots and microbial communities was larger than the treatment effects, and was important in shaping microbial community composition and determining grassland responses to drought.

Taken together, these findings suggest that roots and microbial communities are likely to be tolerant to drought a within the timescale of this experiment, but we did not observe an increase soil carbon sequestration or drought resilience when adding compost. This thesis highlights the importance of considering soil processes as complementary to aboveground observations when studying carbon dynamics, predicting ecosystem responses to environmental change, and developing sustainable land management practices.