Real-time Hydrometeorological Observation Networks - Development Possibilities for the Early Warning System of the Mekong River Basin

Manual surface measurements have traditionally formed the basis for hydrometeorological observation networks. Although automation and real-time data communication has gradually increased, sufficient area coverage for minimizing uncertainties associated with spatial averaging of hydrometeorological variables is still relatively rare. The limited availability of real-time precipitation data for the Mekong River left bank (eastern) sub-basins of Lao PDR has been found to reduce the accuracy of flood forecasting. Also, the usage of surface data from Upper Mekong River Basin could be developed for better five-day water-level forecasts. During disasters like flash floods the lack of real-time hydrometeorological information for forecasting and advisory service contributes to unnecessary economic losses and human casualties. Therefore, a need to further develop hydrometeorological observation networks exists. Longer lead times and improved input accuracy for flood forecasting can be reached by increasing observations of atmospheric water i.e. measuring water before it precipitates on a river basin. As tropical storms and typhoons arriving from the South China Sea cause the most severe flooding situations in the Mekong River Basin, enhanced observations, tracking and modelling of these storms is recommended for accurate medium-term flood forecasts. In addition to satellite and aircraft observations, a very low frequency (VLF) long-range lightning detection network could be used to monitor convective activity and estimate rainfall associated with tropical systems over data-sparse sea areas. If suitable aircraft are available, dropsondes could enable even more detailed observations of tropical storms and their surrounding environment. Assimilation of any of these datasets into numerical weather prediction (NWP) models tends to improve both the storm track and precipitation forecasts. Weather radars play an important role in storm tracking and nowcasting over coastal and inland areas. Compared to conventional single polarization radars modern dual polarization weather radars can improve quantitative precipitation estimates (QPE) through hydrometeor classification and better data quality. Improving short term forecasts by assimilation of wind observations into NWP models such as the Fifth Generation Mesoscale Model (MM5) from weather radars is becoming evermore widespread. Finally, NWP model forecasts are used as input for medium term hydrological forecast models, while weather radar derived QPEs are combined with surface observations for improved nowcast input. In our paper, we summarize recent findings related to operational applicability of these methods for improving hydrometeorological forecasts. Moreover, we discuss hydrometeorological observation system issues related to spatial configuration and overall performance. 7 Annual Mekong Flood Forum (AMFF-7)