GEOPHYSICAL RESEARCH LETTERS, VOL. 29, NO. 2, 10.1029/2001GL013827, 2002
[6] The atmospheric-hydrological modeling system consists of a high-resolution regional atmospheric model coupled interactively to a hydrological model, and a routing module which is run off-line. The atmospheric model used is MC2 (Mesoscale Compressible Community Model), a 3-dimensional compressible regional model developed by Benoit et al. [1997]. MC2 uses the force-restore method (Deardorff, 1978) as the default treatment of land surface processes. The hydrological model is a modified version of the 1-dimensional column land surface model CLASS (Canadian Land Surface Scheme) with three soil layers and a “mosaic” approach for vegetation treatment, first developed by Verseghy [1991] and Verseghy et al. [1993] for atmospheric general circulation models. A higher effective resolution of land surface characteristics is obtained through the mosaic approach. The advantages of this approach are further discussed by Avissar and Pielke [1989], Koster and Suarez [1992], Wood et al. [1992], and Ducoudré et al. [1993]. In an earlier study (Wen et al., 2000a), we have coupled the original unmodified version of CLASS interactively to MC2 to study the effects of CLASS and force-restore method on the simulation of the intense precipitation of the Saguenay storm, at a spatial resolution of 5 and 10 km. Our study showed that the higher effective resolution of land surface characteristics achieved by CLASS means that MC2-CLASS at 10 km resolution can give better results than MC2-force restore at 5 km. The impact of land surface treatment on short-range precipitation simulation can thus be significant, especially in regions with complex vegetation characteristics.
[7] For this study, we have modified CLASS to allow for interflow and baseflow (Wen et al., 2000b). We will refer to the modified version as CLASS*. The original version of CLASS has no interflow, and a baseflow which is not appropriate for hydrological purpose. Our modification consists of the use of a field capacity threshold to allow for interflow generation, and the introduction of a reservoir at the bottom of the third CLASS soil layer. The latter is similar to the parameterization of drainage introduced by Dümenil and Todini [1992].
[8] The details of the initialization procedure of the coupled atmospheric-hydrological model (MC2-CLASS) is described in Wen et al. [2000a]. Initialization of MC2-CLASS* follows the same procedure. Soil moisture content is initialized using CMC (Canadian Meteorological Centre) analyses. With no modification to CLASS, the coupled model MC2-CLASS gives an effective precipitation that can be several times larger than the precipitation for certain grid squares. Effective precipitation is the component of rainfall that contributes to storm runoff, and is calculated by the coupled model. The large effective precipitation of MC2-CLASS is due to incorrect initialization of soil moisture based on CMC analyses. The initial soil moisture exceeds the field capacity, thus resulting in large amounts of free water in the soil column. This is especially true for sandy soil, leading to large drainage at the bottom of the column. With our modification to CLASS, the effective precipitation of MC2-CLASS* of these grid squares become more realistic, as the newly introduced reservoir at the bottom of the third CLASS soil layer moderates the baseflow, thus circumventing the initialization problem. Soil moisture initialization remains an important unsolved problem in weather prediction models. At this stage, we have no objective method for obtaining alternative values of initial soil moisture other than using CMC analyses, which may not be adequate for hydrological applications.
[9] The catchment scale routing module is based on the GUH (geomorphological unit hydrograph). The GUH was first introduced by Rodriguez-Iturbe and Valdés [1979] to link the catchment hydrologic response, given by the instantaneous unit hydrograph (IUH), with geomorphologic parameters of the catchment. The latter is assumed to be drained by a perfect Horton network. We have extended the GUH to a Horton network of order 4 and higher (Wen et al., 2001). All but one parameter of the GUH can be obtained from the basin geomorphology using GIS (Geographic Information System) methods. These parameters are Horton's bifurcation ratio, stream length ratio, stream area ratio, and the length of the highest order channel. The remaining parameter is a flow velocity, which can be obtained from calibration with historical data or methods appropriate for ungauged catchments. Wen et al. [2001] describe the GUH extension and discuss six methods to obtain the flow velocity. They have also tested the GUH using 252 flood events over 25 catchments in the Zhejian province of China with encouraging results. In the present study, the geomorphorlogic parameters of the Ha! Ha! River basin were obtained using standard GIS methods, and the flow velocity obtained using the method reviewed by Wen et al. [2001] for ungauged catchments.
Citation: A coupled atmospheric-hydrological modeling study of the 1996 Ha! Ha! River basin flash flood in Québec, Canada, Geophys. Res. Lett., 29(2), 10.1029/2001GL013827, 2002.