Born in Washington DC and raised in Central Illinois, Dr. Patrick McGuire has been at the University of Reading since 2017, first as a Land Surface Processes Computational Scientist and recently as a Land Surface Processes Research Scientist. Thanks to a background in physics and mathematics, further enriched with experience in geoinformatics, computational science, machine learning and climate modelling, Patrick McGuire here provides explanation for two key tools in his research: the data assimilation technique, which is employed to assimilate soil moisture observed by satellites into climate models in the European region; and the Joint UK Land Environment Simulator (JULES), used to model the soil and vegetation and to improve the estimates of soil physical property parameters.
Why did you get involved in EXPECT?
I was asked to work on this project by Brian Lawrence and Tristan Quaife and they know me for my expertise in land surface modelling. I’ve been working in this field since 2017 on various different projects that include JULES.
What is JULES?
It’s a land surface model. For example, with standalone JULES, we can take the weather data and use that as input to JULES. The weather data has the rainfall and the snowfall and the air temperature, and it has the cloud cover, and a few other variables. Then, JULES calculates how well the soil absorbs the water and how well the plants are growing. It also calculates the water budget, the carbon budget, and the energy balance.
What do you use data assimilation for?
To better estimate the physical parameters that characterise the soil. Once you have better, more accurate value for how much soil moisture you have, then you can estimate hydrological events like droughts.
What are the results so far?
The results indicate that our technique works, and it reduces the error on the estimation of soil moisture. It allows us to generalise that to different years that were not used in the training of the data assimilation technique for the actual soil moisture. Results seem rather robust. That would allow us to better estimate the impact of the weather on drought, among other things: if we can better estimate drought, then we can better estimate the drought feeds back to the weather. However, to do that we would need to our data assimilation capable technique into a coupled model that simulates the atmosphere and not only the land. To use a coupled model instead of just a land model is another step a little further in the future.