Joint Centre for Mesoscale Meteorology (JCMM)
Short period forecasting of catchment-scale precipitation. Part I: The role of Numerical Weather Prediction
by M A Pedder, M Haile and A J Thorpe
A deterministic forecast of surface precipitation involves solving a time-dependent moisture balance equation satisfying conservation of total water substance. A realistic solution needs to take into account feedback between atmospheric dynamics and the diabatic sources of heat energy associated with phase changes, as well as complex microphysical processes controlling the conversion between cloud water (or ice) and precipitation. Such processes are taken into account either explicitly or via physical parameterisation schemes in many operational numerical weather prediction models, which can therefore generate precipitation forecasts which are fully consistent with the predicted evolution of the atmospheric state as measured by observations of temperature, wind, pressure and humidity.
This paper reviews briefly the atmospheric
moisture balance equation and how it may be solved in practice. We
present solutions obtained using the Meteorological Office Mesoscale version
of its operational unified numerical weather prediction (NWP) model, and
verify predicted precipitation rates against catchment-scale values based
on observations collected during an Intensive Observation Period (IOP)
of HYREX. Results highlight some limitations of an operational NWP
forecast in providing adequate time and space resolution, and its sensitivity
to initial conditions. It is argued that the large-scale model forecast
can nevertheless provide important information about the moist dynamical
environment which could be usefully incorporated into a higher resolution,
'storm resolving' prediction scheme.