Quantitative Precipitation Forecasting Using Cloud-Based Techniques on AVHRR Data

Tri-yearly recurrence of flooding in Malaysia has made floods the most important significant natural disaster in the country in terms of cost and damage to property. Many hydrological and hydraulic flood models have been implemented but are yet to meet the requirement of a near real-time flood forecasting. This study envisage the “Nowcasting System ” through the operational coupling of quantitative precipitation forecasting (QPF) in remote sensing with hydrological and hydraulic oriented GIS to argument the problem. The paper presents the first phase of the study in the application of cloud model-based techniques in the extraction of precipitation from NOAA-AVHRR data for flood warning system. It discusses object oriented cloud classification techniques using Ecognition software to identify major cumulus cloud types. Cloud brightness temperature (TB) threshold levels are established and evaluated based on the digital number values in channels 1 & 2 of the AHVRR data. The intensity and dimensions of rain baring pixels are measured after a mask has been applied to extrude clouds and various enhancements and filtering processes. Estimates of expected precipitation are derived based on the assumption that every cloud pixel has a constant unit rain-rate of 3mm-h, which is appropriate for tropical precipitation over 2.5o x 2.5o areas around the equator