November 30, 2018 | Published by | ,

The Drought program has been strongly focused on

  • evaluating and improving climate models, and
  • developing a drought database for documenting drought and for benchmarking model performance.

One of the key issues with climate models has been the need to evaluate the processes in the models to see how they compare with the current knowledge and assumptions about how the physical system works. The standard approach has been to compare the outputs of a model or multiple models against independent observational datasets.

When models diverge from each other or the observations, the difficulty has been isolating the precise cause and whether it is a single process that caused the divergence or interactions between two seemingly unrelated processes.

Working with international colleagues, CLEX scientists have put forward an open-source, model agnostic, modular framework called the multi-assumption architecture and testbed (MAAT). This has been proposed as a generic framework to systematically explore how different underlying model assumptions, hypotheses and parameters lead to predicted model behaviour. In the first applications, the MAAT has been successfully tested on a simple groundwater model and a widely used leaf scale photosynthesis model. We expect further applications in coming years.

Modelling extreme heat events and how they might change has also proven to be challenging. It turns out that drought and heat-waves share many processes in common. Hence it is no surprise that members of the drought group have been working with colleagues in Heatwaves and Cold Air Outbreaks Research Program, to test how well climate models simulated hot and dry conditions in wet regions.

In research led by CLEX research associate, Anna Ukkola, CMIP5 model simulations were compared with the observations during heat waves made at flux towers world-wide. The results confirmed that CMIP5 models do not replicate the surface-atmosphere coupling as recorded in the observations. One important consequence is that CMIP5 models tend to amplify the temperature during heat extremes and underlie the need to improve model simulations under extreme conditions.

In a second application of the flux tower database (known as FLUXNet), CLEX research led by PhD student, Ned Haughton, aimed to investigate how predictable the flow of heat, moisture and carbon is between the land surface and the atmosphere. The long-held assumption is that vegetation type (e.g. tropical forest, grassland, boreal forest, etc.)  was the best predictor of the land-atmosphere fluxes. The research did not find that. Instead they found that predictability was in fact unpredictable. We are still working through the scientific implications of this result.

The world-wide FLUXNet database has over the years proved invaluable for evaluating models. However, FLUXNet is essentially a point-based database and does not directly evaluate the lateral flows that are central to Hydrology (e.g. runoff).

Over the past six months, Dongqin Yin has been leading research to evaluate a new Hydrologic Reanalysis product called the CDR (Climate Data Record) made available by Hydrologists at Princeton University. This represents the most comprehensive Hydrologic reanalysis available to date. The data are available at around for the global land surface at 50 km resolution for monthly period from 1984 to 2010. Initial results have shown that the CDR fills an important gap in the available databases and publications on the use of this database are currently being prepared. We expect to make use of the CDR for the duration of CLEX.

As part of our global linkages, we also welcomed Lorenz Keysser an undergrad student from ETH Zurich who spent four months at the University of New South Wales hub working with Annette Hirsch, Sonia Seneviratne (ETH) and Benoit Guillod (ETH) to evaluate the land use scenarios derived by integrated assessment models.

Meanwhile, research associate, Martin De Kauwe co-led an international workshop held at Biosphere 2 in Arizona that aimed to integrate evidence streams to describe the CO2 Fertilisation effect on the global carbon sink.

The chief outcome of the workshop will be a review paper, alongside a number of spin-off manuscripts on topics including: global photosynthesis trends, scaling of leaf-to-canopy estimates of photosynthesis, reconciliation of water use efficiency estimates from FACE experiments and tree ring isotopes, evidence for satellite greening and how vegetation communities are changing in response to elevated CO2.