Amy C Green
Use of available observations for estimating the rainfall field.
- Email: a.c.brown@ncl.ac.uk
Project title
Improving radar rainfall estimation for urban flood risk using Monte Carlo ensemble simulation
Supervisors
- School of Engineering
- School of Mathematics, Statistics and Physics
Project description
In urban drainage design and flood risk management, there is a need for rainfall data with a high spatiotemporal resolution. These data identify the spatial extent and distribution of high intensity rainfall.
Weather radar provides area-average rainfall estimates. These give good information on the spatial variation and movement of the precipitation field, but have many error sources. This leads to large uncertainties in rainfall amounts. Current methods for obtaining radar rainfall estimates are mainly physics-based. Reflectivity observations are treated as a covariate for rainfall rate. Various calibration techniques and corrections have been suggested. These often use nearby ground observations with the assumption that the rain gauges are ‘ground truth’.
In this project, we will explore and exploit the statistical space-time properties of radar reflectivity data and point based measurements. We focus on high rainfall intensities.
Using conditional random simulations, we will start by considering the spatio-temporal sampling differences between the two types of observations. This will give us an improved understanding of their dependence.
We will use this information to determine the best use of spatio-temporal reflectivity fields and point-based observations. We will consider replacing standard deterministic radar-gauge merging techniques with a stochastic ensemble of simulations conditioned on both the reflectivity field and point-based observations. This will enable us to investigate how best to use available observations for estimating the true rainfall field. This will ensure that information on the spatio-temporal dependence of the rainfall field is not lost or wasted.
Publications
- Fawcett L, Green AC. Bayesian posterior predictive return levels for environmental extremes. Stochastic Environmental Research and Risk Assessment 2018, 32(8), 2233-2252.
- Matthews J, Newman K, Green A, Fawcett L, Thorpe N, Kremer K. A decision support toolkit to inform road safety investment decisions. Proceedings of the Institution of Civil Engineers–Municipal Engineer June 2018, 172(1) 53-67. Thomas Telford Ltd.