Hydrologic modeling

Overview

Figure 1. Conceptual flow diagram of the Sac-SMA model and a schematic representation of model output. (Source: adapted from NOAA NWS 2002); Figure 6.3, State of the Science Report (Lukas and Payton 2020)

Hydrologic models are computer-based simulators used to characterize the likely behavior of real watersheds under user-specified conditions and inputs (e.g., current snowpack and soil moisture, future weather and climate, vegetation change). These models generally include meteorological inputs (e.g., precipitation and temperature), governing equations and physical laws, and model structure, such as the connectivity of watershed components such as tree canopy, snowpack, and subsurface water storage and flow. While all of these components are generally present in hydrologic models, there are large differences among models in how these components are represented, the ways in which runoff is calculated, and the spatial extent and resolution of the catchment areas in the model. There is no one approach or level of complexity that is optimal for all applications of hydrologic models.

Hydrologic models are widely used in the Colorado River Basin to study various aspects of hydrological processes and response (e.g., how runoff responds to wildfire, or climate change); for operational streamflow forecasting; and, coupled with other models, to generate hydrologic scenarios for long-term planning.

Hydrologic models can be broadly categorized into conceptual models and physical (or dynamical) models, although in reality there is more of a continuum. Conceptual models tend to have simple representations of watershed attributes and processes. The linkages between components are typically controlled by adjustable parameters whose values may be derived from observations, or deduced through calibration of the model. Physical models tend to be more complex, and spatial and temporal variations in watershed characteristics are more robustly incorporated, leading to a model that more closely reflects the physical workings of the actual watershed. That said, a simpler model can be better suited for forecasting than a more complex model.

Table 1. Summary of characteristics of three general classes of hydrologic models. Modified from Table 6.1 in Colorado River Basin Climate and Hydrology: State of the Science (Lukas and Payton, 2020)

	Bucket-style models	Land-surface models (LSM)	Watershed process models
Examples of models used in CRB applications	Sac-SMA (NOAA), SNOW-17 (NOAA), Monthly Water Balance Model (USGS)	VIC, Community Land Model (CLM), Noah-MP, SUMMA	WRF-Hydro (terrain routine), National Water Model (NWM); DHSVM
Model structure	Conceptual	Mostly physical	Physical
Spatial framework	Lumped (i.e., by watershed) or semi-distributed	Distributed (gridded)	Distributed (gridded)
Typical resolution	3 km - 50 km	500 m - 25 km	10 m - 500 m
Primary applications in the CRB	Operational streamflow forecasting, sensitivity analyses, coarse-scale climate-change impact analysis	Climate sensitivity analyses, climate change and variability impacts, streamflow forecasting	Hydrologic process studies (e.g., surface energy balance, surface-groundwater interactions, snow hydrology), climate variability and change studies

Data and tools

Key hydrologic models used for research and/or operational applications in the Colorado River Basin:

NOAA Sacramento Soil Moisture Accounting (Sac-SMA) Model

Sac-SMA is the primary model used for operational streamflow forecasting by the NOAA Colorado Basin River Forecast Center (CBRFC) and the other NOAA RFCs. Sac-SMA represents soil moisture and storage characteristics to effectively simulate streamflow. The Sac-SMA model is available in a number of coding languages, including Fortran, MATLAB, and R. Sac-SMA is used in NOAA CBRFC forecasts.

SNOW-17

SNOW-17 is a temperature-index model that uses precipitation, temperature, and the freezing level to simulate snowpack accumulation and ablation. It is typically run as a module paired with Sac-SMA, to add snowpack processes to Sac-SMA, such as for NOAA CBRFC’s forecasts. The input parameters include precipitation and temperature; precipitation is characterized as rain or snow depending on the temperature. NOAA has current Snow-17 Fortran code; the R code for Sac-SMA linked above includes SNOW-17 as a module.

Variable Infiltration Capacity (VIC) model (University of Washington)

VIC is a grid-based land-surface model (LSM) that solves the energy balance and water balance at the surface and subsurface using physical equations. VIC is open-source and is currently on its fifth major version; development and maintenance of the ‘official’ version of VIC is led by the Computational Hydrology Group in Civil and Environmental Engineering at the University of Washington. The main page linked above provides access to code and documentation for running VIC on multiple platforms.

Structure for Unifying Multiple Modeling Alternatives (SUMMA) (NCAR)

SUMMA is a hydrologic modeling approach that is built on a common set of physical equations and a common numerical solver, which together constitute the structural core of the model. Different modeling options can then be implemented within the structural core. The main page linked above provides access to the Fortran code and documentation for running SUMMA.

Additional resources

Colorado River Basin Climate and Hydrology: State of the Science report

Chapter 6(Hydrologic Models) of the State of the Science report provides expanded descriptions of hydrologic hydrologic modeling relevant to the Basin, including key models, calibration techniques, and model applications.

Chapter 8(Streamflow Forecasting) and Chapter 11 (Climate change-informed hydrology) describe how hydrologic models are used in those respective applications.

Chapter 4 (Observations - Weather and Climate) and Chapter 5 (Observations - Hydrology) describe the key sources of data used as inputs for hydrologic models, and to calibrate and validate the models, in applications for the Basin.