SiB2 and river routing module

The SiB2 model is a grid-based model that is designed for use in atmospheric general circulation models. It calculates the water and energy balance equations at the land surface. In SiB2 model, precipitation consists of large-scale spatial uniform precipitation and convective spatial nonuniform precipitation. In most GCMs, a single (area averaged) figure for convective precipitation is produced for each grid area for each time step. The SiB2 model can use the output large-scale and convective precipitation from GCMs. In this study, we present an evaluation of the relative importance of subgrid variations in precipitation and in human activities. The emphasis is on the use of observations to drive a physically based model of the land surface water and energy balance. The large-scale and convective precipitations cannot be specified from the observed precipitation. We assume that the observed rainfall is spatially distributed according to a simple exponential function I(x) :

I(x) = ae^-bx + c (1)

where, I(x) is relative amount of rainfall as a function of fractional area of grid area x, (0 < x < 1). a, b, c are constants (Figure 1). The constants a, b, c are normalized so that the integration of I(x) over the whole grid is 1. If set a = b, there is c = e^a.

Figure 1: Precipiation area-amount relationships

2#2

Surface runoff and subsurface runoff are routed to the basin outlet through a channel network as described by Tang et al. (2006). The river basin and river network are abstracted from 10 km Digital Elevation Model (DEM). The surface runoff flows to the river channel following a hillslope, governed by one dimensional kinematics wave model. The subsurface runoff will connect to river channel with a groundwater reservoir. The river flow is governed by a kinematics wave model, taking into account the fraction of river channel.

The input data to the SiB2 model are hourly precipitation, temperature, vapor pressure, wind speed, shortwave downward radiation, and incoming longwave radiation. When the hourly input data are not supplied to the model, we partition these variables in the model time step based on daily precipitation, maximum and minimum temperatures, using the standard algorithms or empirical relationships (Allen et al., 1998; Jiménez et al., 1987; Revfeim, 1997; Cesaraccio et al., 2001). The daily wind speed is directly used as hourly wind speed.