University of Edinburgh logo British Antarctic Survey logo National Oceanographic Centre logo University of Leeds logo

Projects

Development of Novel Rainfall Nowcasting Techniques for Southeast Asia

Background and Rationale

The Maritime Continent region of Southeast Asia contains the countries of Indonesia, Malaysia, Philippines and Papua New Guinea. It is a region of complex topography and a global hotspot for intense rainfall. Storms such as the squall line featured in Figure 1 are common across the region and are regularly the cause of landslides and flooding.

The skill of numerical weather forecasts on the 1-5 day timescale remains low, often with little skill beyond that provided by a daily persistence forecast (i.e. the weather tomorrow will be the same as today). “Nowcasting” uses detailed descriptions of the current weather (from satellite observations for example) and extrapolates up to 6 hours into the future to produce a short range weather forecast. In the Maritime Continent region, tropical convective systems can have lifetimes of several hours up to more than a day, which means there is an opportunity to make better use of the nowcasting techniques to aid weather forecasting.

Figure 1 Typical landscape of Java, Indonesia (left) and radar imagery of a squall line over Singapore on 8 August 2006 (right).

Project details and objectives

The simplest way to nowcast with geostationary satellite observations is to use brightness temperature at the ~11 um infrared channel to identify cold cloud tops. More complex systems use information from multiple channels; e.g. estimating convective rain rate from a combination of brightness temperature at both the infrared and water vapour channels with visible reflectivity during the day.

A number of countries in Southeast Asia, such as Indonesia and Malaysia, operate a precipitation radar network and the Met Office has been routinely producing high resolution (convection-permitting) deterministic and ensemble forecasts over the Maritime Continent since 2017.

Optical flow and machine learning approaches will be utilised to forward propagate the observations in time and to integrate the satellite and radar observations with numerical model forecasts. The integration of the radar and model forecasts is expected to lead to significant improvements in nowcasting of convective initiation, as it is expected to enable the identification of convergence lines using the convection-permitting model and visible channel satellite observations during the day. The integrated nowcasting system would be designed to optionally include lightning observations, which have been shown to significantly improve satellite-based nowcasts. Lightning sensors are currently subject to feasibility studies for upcoming Himawari satellites, and global lightning datasets are available for selected periods through the World Wide Lightning Location Network dataset.

Objectives:

  • Develop a blended nowcasting system that combines high quality Himawari satellite data (Figure 2) with ground-based radar and high resolution model forecasts
  • Produce quantitative rainfall nowcasts and short-term forecasts over Southeast Asia for a test period
  • Compare the skill of the new forecasts with existing weather forecasts and persistence forecasts

Figure 2 Left: Artist rendition of the Himawari-8 Satellite. Image credit: MELCO. Right: Satellite images of clouds over Indonesia and Malaysia from the Himawari-8 Satellite. Top: full-colour image. Bottom: infrared image, where very cold cloud tops that suggest heavy rainfall are coloured pink.

Training

You will be based at the University of Leeds and be supervised by Meteorology and Nowcasting experts from University of Leeds, University of Edinburgh and the Met Office, Exeter. You will gain strong expertise in using a mix of techniques in remote sensing, dynamical meteorology, weather forecasting and data analysis to develop and test a new weather forecasting tool.

Project outputs

The project will develop a step-change in the quality of short-term precipitation forecasts (up to 6-12 hours lead time), which will improve the ability of Southeast Asian governmental agencies to prepare for high impact weather. Although this could be perceived as a relatively short lead time to be practically useful, given the low skill of numerical weather forecasts, any advanced warning of severe weather would be socially and economically extremely valuable.

The PhD project would get to the proof-of-concept stage of development and provide evidence that the system is a significant improvement on existing forecasting methods that are based solely on numerical weather prediction. Beyond the time scale of this PhD, the longer term vision is to fully develop an operational product. In addition, the results will provide a set of high resolution observations of convective systems, which will be useful to validate and develop model forecasts and aid further research in convective processes.

Candidate description

We seek an enthusiastic student with a suitable Undergraduate and/or Masters Degree qualification equipped with quantitative skills. Suitable backgrounds may be in Informatics, Maths, Physics, Earth Sciences or Physical Geography. Some prior experience of computer programming in a language such as Python is desirable.

Further reading

Birch, C. E., S. Webster, S. C. Peatman, D. J. Parker, A. J. Matthews, Y. Li, M. E. Hassim, 2016: Scale interactions between the MJO and the Maritime Continent, J. Climate. doi: 10.1175/JCLI-D-15-0557.1.

Browning, K.A. and Collier, C., 1989. Nowcasting of precipitation systems. Reviews of Geophysics, 27(3), pp.345-370.

Chun‐Fung Lo, J and T. Orton (2016) The general features of tropical Sumatra Squalls, Weather, https://doi.org/10.1002/wea.2748.

de Coning, E., Gijben, M., Maseko, B. and van Hemert, L., 2015. Using satellite data to identify and track intense thunderstorms in South and southern Africa. South African Journal of Science, 111(7-8), pp.1-5.

Hill, P.G., and Co-Authors, 2020. How skilful are NWCSAF satellite nowcasting products for tropical Africa? Meteorological Applications, accepted.

Prudden, R., Adams, S., Kangin, D., Robinson, N., Ravuri, S., Mohamed, S. and Arribas, A., 2020. A review of radar-based nowcasting of precipitation and applicable machine learning techniques. arXiv preprint arXiv:2005.04988.Qian, J., 2008: Why precipitation is mostly concentrated over islands in the Maritime Continent. J. Atmos. Sci., 65, 1428–1441, doi:10.1175/2007JAS2422.1.