Development of an orographic rain enhancement parameterisation

Operational NWP and climate models are run at various horizontal resolutions from the order of a kilometre or so to several hundred kilometres. The underlying topography is therefore not always well represented, causing an under-prediction of the amount of orographic rain enhancement occurring during the passage of frontal systems as shown in Smith et al (2012). A parameterisation scheme to represent rain enhancement due to sub-grid orography via the seeder feeder mechanism is being developed with the aid of the Kinematic Driver (KiD) model developed at the Met Office. The KiD model uses a full microphysics scheme (here we use the Morrison scheme), while idealized wind fields from linear theory assuming constant N and U with height represent the flow over a small scale 2 dimensional hill. Advective transport and particle sedimentation are therefore allowed while avoiding the complexity caused by feed-backs between dynamics and microphysics.

First a set of simulations of neutral flow over a well represented 2D Agnesi hill with various widths (up to 20 km) and heights (less than 900 m) were performed. An orographic cloud forms within a 2 km deep moist layer adjacent to the surface, and the RH in this layer was given various different values. For now only warm microphysical processes are allowed, and autoconversion was switched off for this set of simulations so that none of the water was rained out. The amount of orographic water produced by each of these simulations was successfully estimated by integrating the displacement of the streamline above the lifting condensation level at each height, and multiplying this by the rate of production of adiabatic liquid water per metre of ascent.

Another set of KiD simulations were run with auto-conversion switched back on and a large scale cloud above 2 km in order to provide the large scale seeder rain present during the passage of a frontal system. In the absence of any large scale cloud the hill must be large enough to produce some seeder raindrops which then initiate the accretion process by which most of the rain mass is produced. When a large scale cloud is present aloft, this is no longer necessary and smaller scale hills are able to enhance the surface rain rates via accretion alone. It is demonstrated that accretion rates are generally a factor of 100 larger than the autoconversion rates for these simulations with a large scale cloud. This is also true for larger scale hills which are able to produce rain independently: once autoconversion in the orographic cloud has produced sufficient seeder raindrops, accretion takes over in the production of additional rain mass.

Simplifying assumptions about the nature of the variation of the sub-grid surface height about the grid-box mean make it possible to estimate the amount of orographic water missing from an NWP grid-box at each level. This sub-grid orographic water is added to the liquid water mixing ratio used to calculate the accretion rate, thereby enhancing the rain rate. So far the sub-grid rain enhancement scheme has been tested using the KiD model for a Gaussian hill of half width 5 km and height 400 m using various horizontal grid spacings, with simulations in which the hill is fairly well resolved, partially resolved and also fully sub-grid. As in a gravity wave drag scheme, the standard deviation of the surface heights about the grid-box mean is used to represent the amount of sub-grid orography. These initial tests using the KiD model have given promising results.