-
Notifications
You must be signed in to change notification settings - Fork 0
Home
MOSART was developed to be applicable across local, regional and global scales (Li et al., 2013). One of the major objectives of MOSART is to provide freshwater input for the ocean model to simulate convection and circulation. MOSART also provides an effective way of diagnosing the performance of the land model since the simulated river flow can be directly compared to the observations at the gauging stations under natural conditions (Li et al., 2015a). Moreover, MOSART serves as a cornerstone for further extensions to incorporate riverine transport and transformation of energy and biogeochemical fluxes under both natural and human-influenced conditions (Li et al., 2015b).
MOSART assumes that each spatial unit (e.g., lat/lon grids or watersheds) can be divided into three categories of geographic units: hillslopes that contribute both surface and subsurface runoff into tributaries, tributaries discharge into a single main channel, and the main channel provides links between the local spatial unit and upstream/downstream units. It further assumes that all the tributaries within a spatial unit can be treated a single hypothetical sub-network channel with a transport capacity equivalent to all the tributaries combined. Correspondingly, three routing processes are represented in MOSART: 1) hillslope routing: in each spatial unit, surface runoff is routed as overland flow into the sub-network channel, while subsurface runoff after generation directly enters the sub-network channel; 2) sub-network channel routing: the sub-network channel receives water from the hillslopes, routes water through itself and discharges into the main channel; 3) main channel routing: the main channel receives water from the sub-network channel and/or inflow from the upstream spatial units (if any), and discharges to its downstream spatial unit or the ocean. MOSART only takes positive runoff as inputs, although sometimes a land model may produce negative runoff items, for example, qgwl. As such, MOSART directly sends the negative runoff (either qsur, qsub or qgwl) from any grid at any time step to the corresponding basin outlet. In MOSART, the travel velocities of water across hillslopes, sub-network and main channel are all estimated using the Manning’s equation with different levels of simplifications. Generally the Manning’s equation is in the form as
where is the hydraulic radius [m]. If the water surface is sufficiently large or the water depth is sufficiently shallow, one can assume that . This is the case for both hillslope and sub-network channel routing. For the main channel, is given by , where is the wetted area [m2] defined as the part of the channel cross-section area below the water surface, is the wetted perimeter [m], the perimeter confines in the wetted area. is the friction slope accounting for the effects of gravity, friction, inertia and other forces on the water. If the channel slope is steep enough, the gravity force dominates over the others so one can approximate by the channel bed slope , which is the key assumption underpinning the kinematic wave method [Chow et al., 1988]. In Eq. (1), is the Manning’s roughness coefficient, which is mainly controlled by surface roughness and sinuosity of the flow path. For hillslopes, sub-network and main channels, a common continuity equation can be written as
where [m3/s] is the main channel flow from the upstream grid(s) into the main channel of current grid, so is zero for hillslope and sub-network routing. [m3/s] is the outflow rate from hillslope into the sub-network, or from the sub-network into the main channel, or from the current main channel to the main channel of its downstream grid (if not the outlet grid) or ocean (if current grid is already the basin outlet). [m3/s] is a source term, which is surface runoff generation rate for hillslopes, lateral inflow (from hillslopes) into sub-network channel and zero for main channel. It is assumed that surface runoff is generated uniformly across all the hillslopes. [m3/s] is a sink term accounting for the loss due to evaporation or infiltration during routing. Currently is set to zero for simplicity.
MOSART is supported by a comprehensive, global hydrography dataset at the 0.5 degree resolution. As such, the fundamental spatial unit of MOSART is a 0.5-degree lat/lon grid. The topographic parameters (such as flow direction, channel length, topographic and channel slopes etc.) were derived using the Dominant River Tracing (DRT) algorithm [Wu et al., 2011, 2012]. The DRT algorithm produces the topographic parameters in a scale-consistent way to preserve/upscale the key features of a baseline high-resolution hydrography dataset at multiple coarser spatial resolutions. Here the baseline high-resolution hydrography dataset is the 1km resolution Hydrological data and maps based on SHuttle Elevation Derivatives at multiple Scales (HydroSHEDS) [Lehner and Döll, 2004; Lehner et al., 2008]. The channel geometry parameters, e.g., bankfull width and depth, were estimated from empirical hydraulic geometry relationships as functions of the mean annual discharge. The Manning roughness coefficients for overland and channel flow were calculated as functions of landcover and water depth. For more details on the methodology to derive channel geometry and Manning’s roughness coefficients, please refer to Getirana et al. [2012]. The full list of parameters included in this global hydrography dataset is provided in the Table below.
Table 1 List of parameters in the global hydrography dataset
In CLM4.5 the routing model is River Transport Model (RTM) which is essentially based on the linear reservoir method, whilst MOSART in CLM5.0 is based on more physically-based kinematic wave method. MOSART and RTM differ in several major aspects: 1. Runoff treatment: RTM in CLM4.5 takes negative runoff hence producing negative streamflow sometimes, whilst MOSART in CLM5.0 takes nonnegative runoff only and always produces positive streamflow which is important for future extension of riverine heat and biogeochemical fluxes. 2. Input parameters: RTM in CLM4.5 only requires one layer of spatial variable channel velocity, whilst MOSART in CLM5.0 requires 13 parameters which are all available. 3. Outputs: RTM only produces streamflow simulation, whilst MOSART additionally simulates time-varying channel velocities and channel water depth and channel surface water variation.