Investigating the causes of variations of temperature and precipitation extremes using observations and climate model data

Understanding how climate extremes have changed in the recent past and how they will change in the future is crucial for preparing for future climate change. Natural variability and anthropogenic climate change are both known to significantly influence the frequency and severity of temperature and precipitation extremes, so it is vital to better understand these effects. A case study of frost days in southern Australia first illustrates how short-term variations in climate can sometimes override any global warming signal for a period of time. Some unexpected frost day ‘increases’ are found over 1981-2011, despite a longer-term warming signal. A link to the concurrent precipitation decline and associated reduced cloudiness is identified for south-eastern Australia. The remainder of this thesis investigates variations and trends in the fraction of area experiencing daily temperature and precipitation extremes in four continental regions: Europe, North America, Asia and Australia, and the Northern Hemisphere. A new method based on existing extreme indices is used to investigate changes in five temperature and precipitation components, including maximum and minimum temperature components, and total precipitation, heavy daily precipitation and wet/dry day components. Near-global increases in area experiencing much-above-average warm days and nights and decreases in the area experiencing much-above-average cool days and nights are found. Over the Northern Hemisphere regions, increases in the area affected by wet extremes are also identified. For Australia, the spatial extent of precipitation extremes is primarily influenced by natural variability. Using simulations from eight global climate models under different radiative forcing scenarios, an anthropogenic contribution to the increases in area affected by warm temperature extremes is identified for all regions. Over Europe, an anthropogenic contribution to the increases in area affected by a much-above-average contribut