The podcast for the presntation by Dr Geert Jan van Oldenborgh is now available for download.
The Environment Institute’s Water Research Centre presented Dr Geert Jan van Oldenborgh from the Global Climate Research division of KNMI (Royal Netherlands Meteorological Institute) on Tuesday 22nd January 2013.
The talk was titled ‘Some examples how and why precipitation means and extremes are changing’ and covered mean precipitation trends, mean precipitation extremes and precipitation events (such as the Thailand floods, hourly precipitation extremes in the Netherlands & Hong Kong and more recent extremes in Manila and Queensalnd).
Dr. Geert Jan van Oldenborgh is senior researcher in the Global Climate Research division of KNMI. He is a climate analyst with a background in statistical analysis of observations, seasonal and decadal forecast verification, and climate event attribution. He combines these by applying seasonal forecast verification techniques to climate model output, verifying whether they are able to reproduce the observed changes. Recently, he used observations and models to consider the role of climate change on the Thailand and Manilla floods. As the author of the KNMI Climate Explorer he makes large amounts of climate observations, analyses and model output available for analysis to the wider climate-interested community. He is Lead Author for the IPCC WG1 AR5 (Chapter 11, Near-term projections and predictability, and Annex I Atlas).
Seasonal and annual mean precipitation trends are fairly well-observed and well-studied. This enables us to compare the observed trends to the trends simulated by the CMIP5 multi-model GCM ensemble used for the IPCC Fifth Assessment Report, using a verification statistic from weather and seasonal forecasting: reliability. It turns out that the mismatch between observed and modelled trends is larger than expected on the basis of natural variability and the spread between the models. The reasons for this are as yet unknown.
In summer 2011 severe flooding occurred in Thailand, a region where modelled and observed trends agree well. An investigation into the causes of these floods showed that no anthropogenic factor could be found in the meteorological aspects: the precipitation was high but not far out of the historical record, and no significant trend to more precipitation could be found in the observations. Climate model simulations covering the same period also showed no trend in mean or variability. However, there were clear anthropogenic factors on the ground that increased the risk of flooding.
Trends in extreme precipitation are harder to study, as good observations are harder to get and models have more problems reproducing extremes. We present one example of hourly precipitation extremes in the Netherlands and Hong Kong. In spite of the differences in climate, both scale exactly the same as a function of dew-point temperature. The rise in extremes in the Netherlands can be attributed to the rise in temperature, and hence global warming. In Hong Kong the increase in extreme hourly precipitation in the rainy season is not related to the mean temperature increase.
Extreme precipitation events on the daily scale also have very different characteristics from seasonal mean extremes. We discuss preliminary results on the recent flooding in Manilla in 2012 and apply the same method to the Queensland floods of 2011 to compare with results of the experts in the audience.