Picture: Dead flowers due to lack of water. Credit: Pawel Czerwiński (Unsplash).

It’s an odd thing to do but we open this report not with a hello but a farewell. Sadly, one of our Drought research program leaders, Mike Roderick, will be retiring after a decade of contribution to Centre of Excellence drought research through ARCCSS and CLEX. He has been a key mentor to many over this time and has made a career out of asking apparently simple questions that turn out to have profound implications. Not surprisingly, Mike doesn’t leave us with a whimper but a bang. Just prior to his retirement announcement, Mike presented a virtual seminar on drought, bushfires and climate. It attracted more than 350 people and now holds the record for the largest seminar attendance at the ANU’s Research School of Earth Sciences. It is a reflection of the esteem in which he is held.

In the meantime, over the past few challenging months the Drought program has maintained a flow of high quality research. Our work to improve the quality of land surface models has borne fruit in ways that will significantly improve our understanding of drought processes and how these may change in the future.

We recently implemented a new model of plant hydraulics into the Community Atmosphere Biosphere Land Exchange (CABLE) land surface model to help us  robustly project future drought impacts on Australian vegetation. The researchers constrained the vegetations’ sensitivity to drought using hydraulic and physiological traits measured in a manipulative drought experiment conducted on Australian tree species originating from across a wide rainfall gradient. The model results agreed well with real world drought impacts derived from satellite observations. This shows that it is now possible to predict the risk of tree death – at large scales and this could have important consequences for conservation and management of Australian forests and woodlands.

The program has also improved how photosynthesis is represented in land surface models. CLEX researchers and colleagues found that adding vegetation canopy architecture with zenith angle variations significantly improved photosynthesis prediction in light-limited ecosystems. The results closely matched real world observations and significantly improved photosynthesis prediction in light-limited ecosystem, finding enhanced photosynthesis in the bottom canopy layers.

Our modelling work has also extended to partnerships with the Heatwave and Cold Air Outbreak research program, where we specifically focussed on the new CMIP6 models. In the first instance, we successfully addressed the error compensation issue for temperature extremes by defining a novel performance metric that identifies those models that can simulate temperature extremes well and simulate them well for the right reasons. We found there was a noticeable improvement between the current CMIP6 compared to CMIP5 models.

In a second partnership with Heatwave and Cold Air Outbreak program, we assessed the ability of CMIP6 climate models to simulate the climate of Australia and the new scenarios for 21st Century climate change. They showed small improvements over the previous generation (CMIP5; developed in ~2012), including better reproduction of land and marine heatwaves and sea-level rise as well as improved relationships between Australia’s climate drivers and rainfall.

Using the latest CMIP6 models, the drought program looked at how well these models simulated observed drought, and how droughts may alter in the future with climate change. The models indicated southwestern Australia and parts of southern Australia will see longer and more intense droughts due to a lack of rainfall caused by climate change. But Australia is not alone. Across the globe several important agricultural and forested regions in the Amazon, Mediterranean and southern Africa can expect more frequent and intense rainfall droughts.  While some regions like central Europe and the boreal forest zone are projected to get wetter and suffer fewer droughts, those droughts they do get are projected to be more intense when they occur. The key to these robust results was factoring in rainfall variability and average rainfall.

None of the improvements and assessments of climate models can be verified without robust datasets, so improving these remains a cornerstone of our work. A major challenge in climate science is getting accurate rainfall estimates in areas where ground-based rain gauges are scarce. A CLEX researcher and international colleagues developed a hybrid approach to estimate recent rainfall that combines satellite-based rainfall estimates with satellite-based soil moisture estimates. When this approach was tested against independent rain gauge measurements it showed notable improvements, in a range of metrics when compared to other existing satellite-derived estimates.

Another challenge is understanding how energy and water are distributed at the Earth’s surface. This is a key part of predicting climate extremes such as drought, heatwaves, and flooding. Depending on the landscape and prevailing conditions, rainfall might join rivers, get stored in soils or evaporate. At the same time, absorbed energy from the sun heats the air, soil or powers evaporative processes. While a range of observational data sets of each of these key processes exists, each has its own limitations. Collectively they do not necessarily obey the laws of conservation of energy or mass. To overcome this restriction, we developed a method for combining multiple observational datasets of the energy and water budgets at the land surface from different sources into a single hybrid dataset that conserves energy and mass.

Direct observations with our colleagues at Western Sydney University have also revealed how trees may respond to future higher levels of atmospheric carbon dioxide. The results were unexpected and fascinating. In the first experiment of its kind applied to mature native forest, as well as the first in the Southern Hemisphere, the researchers exposed a 90-year old eucalypt woodland on Western Sydney’s Cumberland Plain to elevated carbon dioxide levels. As expected, the trees took in about 12% more carbon under the enriched CO2 conditions but they did not grow any faster. A carbon-tracking analysis showed that the extra carbon absorbed by the trees was quickly cycled through the soil and returned to the atmosphere, with around half the carbon being returned by the trees themselves, and half by fungi and bacteria in the soil. These findings have global implications: models used to project future climate change, and impacts of climate change on plants and ecosystems, currently assume mature forests will continue to absorb carbon over and above their current levels, acting as carbon sinks. This suggests those sinks may be weaker or absent for forests on low-nutrient soils.

On far longer timescales that the observations noted above, CLEX led an examination of how palaeoclimate evidence could provide rich insights into the Indian Ocean Dipole to discover that strong Indo-Pacific variability was important in breaking droughts over the last 1000 years. Palaeoclimate evidence for hydroclimate changes during the last millennium highlights the importance of interannual IOD and ENSO variability in providing the rainfall that breaks droughts in regions that are impacted by these modes of variability. We contributed to a major perspectives piece in Nature Climate Change on flash droughts – an area where a lot more research is needed.

Far from home, the Drought program took the opportunity to examine an annual drought in South America and Mexico, to see if it could give broader insight into predicting droughts more broadly. This region sees high precipitation rates from May to July and then again from August to October but between these two peaks, from July to August, rainfall is at its lowest for the entire year. The researchers found the reversal of onshore and offshore winds during the development of the drought, and how the wind was forced by the steep mountainous terrain were two key characteristics. However, this interaction was so complex that while the development of the drought was predictable to a point it was difficult to summarise this development in a single theory. As always, the complexity around what causes droughts continues no matter where you are.