This year European Geosciences Union (EGU) General Assembly returned to Vienna, and included a few participants from the Hydro-JULES program. The conference was by far the largest I have attended with over 16,000 participants in total: even deciding which of the multiple parallel sessions to attend was difficult. Anyway, below are a few of the talks/sessions that I found interesting.
Monday: Rewetting Peatland
Peatlands cover around 3% of the world’s land area but account for around 30% of soil carbon. From a greenhouse gas perspective, they are CO2 sink but CH4 sources. If they are drained they become CO2 sources. What happens if they are re-flooded (e.g. for wildlife habitat)? The studied site became a massive source of CH4 for the first year before settling down. Nevertheless, the authors calculated that the re-flooded area would need to absorb C02 for the next 500 years (!) to offset its CH4. Of course, the drained land was a CO2 source anyway...
Tuesday: Measuring snow using neutron sensors
Neutrons, generated by cosmic rays and detected by sensors in the COSMOS-UK network, can be used to estimate the water content held within a snow pack. Actually this wasn’t a talk: it was my poster, pictured below :-)
Wednesday: Greenhouse gasses from European grasslands
Greenhouse gas (GHG) fluxes have been measured at a variety of managed grassland site across Europe, where typically sheep or cows are grazing. The sites were found to be CO2 sinks, N20 sources (particularly after fertilization), and could be either CH4 sources or sinks. The net GHG balance was always found to be negative, with the exception being a single year in which ploughing and reseeding took place. Actually, another talk in the same session highlighted the important of using GPS to track the location of cows when measuring GHG fluxes. The animals had preferred daytime and night-time locations that where only sometimes within the footprint of the detector.
River flow persistence was compared to streamflow climatology for use in seasonal (e.g. monthly) hydrological forecasts. Persistence was typically found to be more skilful for either, (i) forecasts initialised in summer and winter (rather than spring and autumn), (ii) for catchments with high base flow index, or (ii), for catchments with a low annual rainfall.
Friday: the Kalman filter
The Kalman filter is a method of approximating the true time series of a physical quantity measured by a noisy sensor. At each timestep the filter uses a weighted average of the sensor measurement and a predicted estimate based on a model of the system and the filter’s estimate at the previous timestep. A greater weight is accorded to the measurement or model prediction based on their uncertainties. Of course this is just the simple version and there are “Extended Kalman Filters”, “Ensemble Kalman Filters”, and probably lots more.
By John Wallbank 17.04.2019
Photos: the conference venue (upper left), my poster (upper right), the New Danube (lower left), and Capistran Chancel, St. Stephen's Cathedral (lower right)