Role of Surface Roughness Length on Simulation of Cyclone Aila - Monitoring and Prediction of Tropical Cyclones in the Indian Ocean and Climate Change - page 251

In the present numerical weather prediction models, the surface roughness
length is treated as a fixed parameter for each grid point (Wang, 2009). The
roughness length is determined for a specific environment where the numerical
model developed. A simplistic treatment (i.e., using the default value) may be
appropriate when modelling the slowly varying processes (such as land use/
land cover changes or seasonal variations in leaf volume or snow cover etc.).
The surface roughness length at a grid point highly depends on and is derived
from the underlying land-use/land cover maps. Wieringa (1986) provided the
roughness lengths (m) for various type of land cover maps. For example, the
water or ice bodies have the least roughness length of order of 10 -4 m followed
by mown grass (10 -2 m), long grass/rocky ground (0.05 m), pasture land (0.2
m), suburban housing (0.6 m) and forest/cities (1-5 m). Hence, the default
value may not be suitable for the experimentation domain/region, where the
underlying surface is different. If surface roughness length is determined
empirically at a grid point near the boundary between regions having different
land surface characteristics, such as along a coastline or at the edge of an urban
area, it is obvious to expect different values.
The surface roughness length is used in the atmosphere models to determine
exchange coefficients for momentum, sensible and latent heat fluxes. These
fluxes play a crucial role in development and maintenance of a tropical cyclone.
The momentum flux affects the secondary circulations. It determines the effect
of surface friction on angular momentum transport and has significant role in
improving the radial inflow and reducing the tangential winds. A change in
surface roughness lengths first affects surface wind speeds and then affects
structural change in a system. The central-pressure simulations depend on a
surface roughness length. Wada and Kohno (2010) showed that maximum
difference in the simulated central pressure is noticed when the simulated
cyclone undergoes intensification. Understanding the interaction between the
land surface parameters such as roughness length and the tropical cyclone is,
therefore, critically important. Therefore, in this study, the impact of surface
roughness lengths on simulation of tropical cyclones is studied for TC Aila
(23-26 May 2009) using a high resolution mesoscale Advanced Research
Weather Research and Forecasting (WRF-ARW) model.
2. Data Used and Methodology
An attempt is made to understand the interaction of surface roughness length
(Zo) on the simulation of TC Aila in terms of track, intensity, structure and
rainfall. A set of three experiments have been conducted with (i) default Zo
values known as CNTL (ii) Zo is decreased by 50% named Zo-50 and (iii) Zo
is increased by 50% called Zo+50. The initial and boundary conditions to the
model are taken from the analyses and forecast fields of NCEP-GFS (National
Centers for Environmental Predictions-Global forecasting system). The
experiments are conducted with two domains with a grid distance of 9 and 3