Regional high resolution version of the HD Model

by Sven Kotlarski and Stefan Hagemann

The Hydrological Discharge (HD) model globally simulates the lateral freshwater fluxes at the land surface. Fig. 1 shows a schematic overview of the HD model. The standard model version uses a daily time step and requires daily time series of surface runoff and drainage from the soil as input fields. As a general strategy the standard HD model computes the discharge at 0.5° resolution which corresponds to a typical average gridbox area of about 2500 km2. The model input fields of runoff and drainage resulting from the various GCM resolutions (such as T106 in this study) are therefore interpolated to the same 0.5° grid. In the HD model, the lateral waterflow is separated into the three flow processes of overland flow, baseflow and riverflow. Overland flow uses surface runoff as input, baseflow is fed by drainage from the soil and the inflow from other gridboxes contributes to riverflow. The sum of the three flow processes equals the total outflow from a gridbox. The model parameters are functions of the topography gradient between gridboxes, the slope within a gridbox, the gridbox length, the lake area and the wetland fraction of a particular gridbox.


The HD model is available to the scientific community --> Model distribution


In addition to the model version that is running off-line, the HD model is also part of the coupled atmosphere-ocean GCM ECHAM5-MPI-OM (Latif et al., 2003) that is operational at MPI-M. For very high resolution applications with the regional climate model REMO (Jacob, 2001), S. Kotlarski transfered the HD model from the global 0.5 grid to a regional 1/6 degree grid over Europe (see below).


For high resolution applications with the regional climate model REMO (Jacob, 2001), the HD model was transfered from the global 0.5° grid to a regional 1/6° grid over Europe (approx. 18 x 18 km horizontal resolution). The increase in resolution means a decrease in minimum travel time of water through a single grid box. Therefore, also the HD model time step had to be adjusted from 1 day in the original version to 1 hour in the 1/6° version. High resolution flow directions were derived for major European catchments and a set of model runs has been carried out. In all these runs, runoff and drainage input fields were produced by the regional climate model REMO. For validation purposes, the simulated discharge was compared to observations both on a long term basis (Fig.3, Fig.4) and for single flooding events (Fig. 5). The analysis reveals strong skills of the coupled system REMO-HD in reproducing observed features like the annual flow regime and timing and magnitude of single discharge peaks.


Back to Stefan`s Science Page

Back to Hydrological cycle

Back to Regional Climate Modelling

For some large rivers the simulated discharge of the previously used model of Sausen et al. (1991) and the HD-Model are compared with observed discharge which is provided by the GRDC (Dümenil et al., 1993). As input five years of daily values of runoff and drainage are taken from an ECHAM4-T42 (Roeckner et al., 1996) simulation using climatological SST.


Fig. 2 shows observed and simulated discharges for several rivers. The graphs show a considerable improvement of the simulated discharge comparing the HD- Model to the model of Sausen et al. For the Amazonas and the Mississippi this improvement is primarily based on the separation of flow processes. For the Donau this is based on the improved model topography. For the Jenissei the model of Sausen et al. looks apparently better than the HD- Model, but this is only a deception. We know that in ECHAM the snowmelt is computed about one month too late. Therefore a correct discharge simulation must simulate a snowmelt induced discharge also one month too late which is the case for the HD-Model at the Jenissei catchment.


Significant references for the HD model:


Hagemann S, Dümenil L (1998) A parameterization of the lateral waterflow for the global scale. Clim Dyn 14: 17-31

Hagemann S, Dümenil Gates L (2001) Validation of the hydrological cycle of ECMWF and NCEP reanalyses using the MPI hydrological discharge model. J Geophys Res 106: 1503-1510



The HD Model: A Parameterization of Lateral Discharge for the Global Scale

by Stefan Hagemann and Lydia Dümenil