The frequency and intensity of extreme weather events (e.g. flood, drought) are predicted to increase for the foreseeable future and it is expected that these will negatively impact upon agroecosystem functioning. Our understanding of how grassland ecosystems respond to extreme weather events occurring at different times of the year, however, is lacking. To better understand the seasonal response of grassland to flooding, we subjected an agricultural grassland to an 8-week extreme flood event at three different temperatures (5 °C-winter, 15 °C-spring/autumn and 25°C-summer) and then followed its subsequent recovery for 9 weeks after floodwater removal. We focused on key indicators of ecosystem functioning including primary production, nutrient cycling, greenhouse gas (GHG) emissions, ammonia (NH3) volatilization, and soil microbial communities. The experiment used intact soil mesocosms (1 kg) with indigenous vegetation collected from a grassland with no previous history of flooding. Flooding reduced biomass production by 18% at 5 °C, 50% at 15 °C and 95% at 25 °C. Flooding also significantly disrupted elemental cycling (nitrogen, phosphorus and carbon) as evidenced by an increased release of P, Fe and NH4+ into the soil and overlying floodwater and large amounts of CH4 and NH3 released to the atmosphere (mainly during the flooding). These effects were more pronounced at higher temperatures (e.g. 45–700 kg CH4C ha−1 and 1–5 kg NH3N ha−1 at 15 and 25 °C, respectively). In addition, after floodwater removal this NH4+ was rapidly nitrified leading to large losses of N2O (1.0–14.2 kg N2ON ha−1 at 5–25 °C, respectively). Especially at higher temperatures, flooding resulted in a reduction in soil microbial biomass (more than 58% of the equivalent unflooded treatment at 25 °C) and changes in microbial community structure (assessed by PLFAs). Further, some of these changes persisted after flood removal including a loss of actinomycetes, arbuscular mycorrhizal fungi and fungi. Overall, we conclude that ecosystem responses to extreme weather events are critically dependent on temperature with those occurring at higher temperatures having a greater negative impact than those at the lowest temperature (5 °C). The large potential release of CH4 and N2O also suggests that flood events should be considered as a potential source of GHGs when comparing top-down and bottom-up calculations of national inventories, and that further work is needed to better refine GHG emission estimates for these events.