Post-graduate student talks and AGM 2017
LOCATION
School of GeoSciences
The University of Edinburgh, Grant Institute
The Kings Buildings, West Mains Road
Edinburgh
EH9 3JW
UK
SPEAKER | Katie Chowienczyk, Met Office.
TITLE | Weather Sensitivity Analysis and its Application in the Water Industry.
ABSTRACT | Work with Thames Water has produced a 'weather intelligence model' to predict water demand as a function of weather (Demand-WIM). This was used to provide operational demand forecasts to several water companies during the summer of 2015. Historic weather data was used to produce probability distribution functions of annual water demand. The model has recently been improved to make predictions during heatwaves and thunderstorms. An overview will be given of Demand-WIM models, their applications and future developments to improve their accuracy and capabilities.
------------------
SPEAKER | Jack McDonnell, Maynooth University.
TITLE | Case Study on Accuracy and Bias Correction of ECMWF Forecasts in Ireland.
ABSTRACT | Day-1 to Day-10 ECMWF deterministic forecasts between 2007 and 2013 were verified using observations from each of Met Eireann's 25 synoptic weather stations. The aim of the project is to include weather forecasts in a grass growth prediction model. The weather variables examined were rainfall, air and ground temperatures. The forecasts of temperature were good at all stations for up to 10 days. For rainfall, climatology and persistence outperformed the deterministic forecasts after 4 or 5 days. A number of weather variables show systematic biases at different locations. For temperature, a regression model using the ECMWF forecast and month gave the best results. For rainfall, where the forecast skill deteriorates, correction for bias gave the best performance.
SPEAKER | Katie Chowienczyk, Met Office.
TITLE | Weather Sensitivity Analysis and its Application in the Water Industry.
ABSTRACT | Work with Thames Water has produced a 'weather intelligence model' to predict water demand as a function of weather (Demand-WIM). This was used to provide operational demand forecasts to several water companies during the summer of 2015. Historic weather data was used to produce probability distribution functions of annual water demand. The model has recently been improved to make predictions during heatwaves and thunderstorms. An overview will be given of Demand-WIM models, their applications and future developments to improve their accuracy and capabilities.
------------------
SPEAKER | Jack McDonnell, Maynooth University.
TITLE | Case Study on Accuracy and Bias Correction of ECMWF Forecasts in Ireland.
ABSTRACT | Day-1 to Day-10 ECMWF deterministic forecasts between 2007 and 2013 were verified using observations from each of Met Eireann's 25 synoptic weather stations. The aim of the project is to include weather forecasts in a grass growth prediction model. The weather variables examined were rainfall, air and ground temperatures. The forecasts of temperature were good at all stations for up to 10 days. For rainfall, climatology and persistence outperformed the deterministic forecasts after 4 or 5 days. A number of weather variables show systematic biases at different locations. For temperature, a regression model using the ECMWF forecast and month gave the best results. For rainfall, where the forecast skill deteriorates, correction for bias gave the best performance.