Abstract:
This thesis investigates the redistribution of rainfall over low hills due to wind flow. This study is motivated by previous empirical studies which found that wind can cause the spatial distribution of rain to decrease on the windward side of the hill and increase on the leeward side. This spatial distribution differs from other rain processes, which typically give an increase of rain on the windward side as one moves up the hill. This "carryover" process conserves the total amount of rainfall, the spatial variation is due only to the kinetic effects of the wind.
The aim of this current work was to establish the sensitivity of cnrryover to different environmental conditions. An attempt was made to validate our numerical model by comparing output to field data obtained previously on Norfolk Island.
The rain model was used traces the trajectories of raindrops through a 3 dimensional time-independent wind flow. Wind flow used in the model ranged from simple potential flow to the more comprehensive wind flow obtained by the RAMS numerical model.
We found that the redistribution has a linear relationship to the far-field horizontal velocity. At low horizontal velocities the stability becomes important by inducing large vertical velocities. Redistribution for three-dimensional flow is smaller than that for two-dimensional flow with the same stability since air can also flow around the hill. For three-dimensional flow the redistribution is relatively constant and independent of FrH. Stable flow causes a weaker secondary depletion/enhancement region downwind of the first. The lee downdraft at cloud height results in an additional increase of rain to the lee of the hill. Thiseffect reduces as cloud height increases.
The enhancement due to carryover increases sharply for drops less than I mm in diameter. For drops larger than I mm the enhancement is relatively constant. Hence, as higher rain intensities have relatively fewer smaller drops, the redistribution decreases with rain rate. Changing from a Marshall-Palmer drop size distribution to a Gamma the drop size distribution (with fewer drops in the smaller range) did not affect carryover appreciably since the smaller drops do not contribute significantly to the rain intensity distribution. Microphysical processes such as evaporation and coalescence, which alter the distribution of liquid water between the drop sizes, had little effect on carryover.
Time dependent effects on carryover can be ignored since the wind flow over a hill causes no appreciable delay in the arrival of drops on the surface of the hill.
Comparison of model output to data from a raingauge network on Norfolk Island for specific rain events lasting of the order of a few hours did not give good agreement. This was probably due to the uncertainty with which the actual wind flow field was known as well as the rain gauge network having a spatial density lower than the spatial variability of the terrain.
Overall this present work allows for the importance of carryover to be evaluated for more general terrain and environmental conditions.