Hydrologic Simulation Program - Fortran

Evaluation of Drainage Management Software

The following sections outline the input requirements, computational methods, and output options available with this software. The theoretical basis of this software was compared with the MTO Drainage Management Manual (1997). A summary of model capabilities, and the requirements for using this software for MTO design, analysis, or approvals is provided

Model Capabilities:

Functionality:

Use in MTO Drainage Management Practices:



Model Capabilities

What does it do?

The Hydrologic Simulation Program - FORTRAN (HSPF) is a continuous hydrologic model, which can be used to simulate a wide range of hydrologic and water quality processes. This evaluation focuses on the general capabilities of the software and hydrologic applications of the model. The water quality and erosion modules are not discussed here. Most of the descriptions and definitions are a synopsis from the HSPF Manual, which contains a more detailed discussion of the model. The model components of relevance are specifically: Hydrology: simulating continuous series of catchment runoff from pervious and impervious areas based upon input time series of meteorological data Channel Routing: modelling the routing of catchment outflows through connecting channels and adding flows at points of confluence Reservoir Routing: modelling the attenuation of inflow hydrographs (in time series form) through storage reservoirs for various outlet configurations.


How does it do it?

In general, HSPF uses the following methodologies for each of the main components:

Hydrology

The subcatchment runoff is simulated using the PERLND module for the pervious areas, and the IMPLND module for the impervious areas. HSPF divides the moisture supply available to a land segment into five main components: the surface runoff, interflow, groundwater flow, and evapotranspiration and percolation to deep groundwater, which are both lost from the system. The model to compute the various aspects of each of these components uses a number of hydrologic parameters. Brief descriptions of some of the relevant parameters for these modules have been provided in Appendix E. The default, maximum and minimum values of these parameters are specified in the User's Manual. Some of these parameters can be obtained directly from available watershed data (eg: area, latitude, mean elevation, etc). However, for most of these are calibration parameters, i.e., appropriate values can only be obtained by calibrating the model to streamflow measurements, since values may vary widely depending on the watershed characteristics. Based on the model calibration undertaken within the scope of the present study, and previous studies on the subject, some initial estimates for the parameters have been suggested in Appendix E, wherever possible.

Channel Routing

The hydraulic routing of the runoff produced by the PERLND and IMPLND modules is performed by the module HYDR, which simulates all the hydraulic processes occurring in open or closed channels (or reservoirs) in the watershed. The channel (or reservoir itself) is referred to as a RCHRES in HSPF. The physical characteristics of the RCHRES are specified in the UCI file, the values for surface area, stage-storage data, etc., being provided in the FtableS block. The routing of the flows is performed in the EXT SOURCES or NETWORK blocks, where the user specifies the runoff from the land segments entering a particular RCHRES. It should be noted that the hydraulic processes simulated in HSPF do not require separate calibration, if the physical data provided for the RCHRES are reasonable. Hence, the physical characteristics of a RCHRES should be defined with particular attention, so that it accurately represents the actual channel or reservoir.

Reservoir Routing

As indicated above, reservoir routing is completed using the HYDR block by defining the reservoir as a RCHRES.


File Management

Both the input and output data from HSPF are in the form of time series. Consequently, the HSPF software is planned around a time series management system operating on direct access principles. The simulation modules, which are accessed with the help of a User's Control Input (UCI) file, draw input from time series storage files and are capable of writing output to them. The "Run Interpreter" module (invoked by the UCI file) is used to read and interpret the UCI file. It sets up internal information instructing the system regarding the sequence of operations to be performed. It stores the initial conditions and the parameters for each operation in the appropriate file on disk and creates an instruction file which will ensure that time series are correctly passed between operations, where necessary. The Run Interpreter documents this information in an "ECHO" file, which also summarizes all errors that the program has encountered while processing the input data for a particular simulation. More detailed descriptions of the various components of the UCI files, and the input time series are provided in the following sections. The time series data are stored in a Watershed Data Management (WDM) file. Pre- and post-processing of the time series is performed with the help of a WDM file management software called ANNIE. ANNIE contains a set of procedures to organize, manipulate, and analyze the time series data in the WDM files.


Output Options

As noted above, post-processing of the time series is performed with the help of a WDM file management software called ANNIE. ANNIE contains a set of procedures to organize, manipulate, and analyze the time series data in the WDM files. Hence outputs can be generated in both detail and in summary forms at the user's discretion for any time series generated by the model.

Top of page   Top of page


Functionality

Experience Required

The HSPF code is written almost entirely in ANSI standard Fortran, and consequently implementation on a wide variety of computers is possible. Since HSPF and ANNIE are both DOS-based programs, the user should possess basic familiarity with the interface. The programs are well documented, and have extensive online help and user's manuals. A complete understanding of the hydrologic processes occurring in a watershed, and experience with watershed modelling, are also required for appropriate selection of input parameters, and computational modules to be used for a particular watershed model. Moreover, before attempting to use the program for hydrologic modelling, the user should establish a thorough familiarity with the HSPF manual, and study existing literature on the subject of HSPF modelling. A formal training program for HSPF modelling, or a period of training with some experienced HSPF users are also recommended.

Using the program

Installing

HSPF is supplied as a .zip file, which is downloaded from the US EPA web site. It is installed by following detailed instructions provided in the.zip file. This takes about 30 minutes.

Learning

The program provides extensive Help files, which echo the hard copy manual and include test files, which can be used as tutorials. It also includes the application experience documentation, which guides the user in selecting appropriate procedures and parameters. However, as noted above, a period of formal training is normally required to become familiar with HSPF.

Using

The program is DOS based and all input is created through a text editor according to pre-defined formats. Time series files are created using procedures within HSPF or by using ANNIE. The simulations are executed using the DOS command line. Results are analysed with ANNIE. Input errors and computation errors are indicated by messages included in an echo file, which records the steps in the model execution. Execution errors are indicated as DOS messages referring to Fortran program code errors. These can be deciphered using the HSPF Programmer's Manual.


Comparison to Previous Version

The current active version is 10.0. A beta version of Release 11.0 is available. There have been numerous changes and developments of the program since its initial release in the mid-1970's. However, the basic hydrologic procedures remain the same.

General Considerations

The following notes should be considered when using this program on MTO projects or projects requiring approval by MTO.

The historic meteorological data for the area being modelled are typically organized into continuous time series and stored in the WDM file. The following meteorological data are relevant to an HSPF simulation:

  • Precipitation
  • Air Temperature
  • Wind Movement
  • Solar Radiation
  • Evaporation
  • Dew Point Temperature
  • Cloud Cover

It should be noted that HSPF simulations can be conducted with internal time-steps, which range from 1 minute to 24 hours, over a prolonged period of time. However, the most accurate meteorological data available in Canada for extended time periods is hourly data. This normally limits the catchment area that can be modelled accurately using HSPF. If hourly precipitation data is being used for a particular simulation, then catchments with time of concentration less than one hour would potentially not be modelled accurately using this program.

The UCI file is used to invoke the various simulation and utility modules, which simulate the hydrologic processes and routing occurring in a watershed. The subcatchment runoff is simulated using the PERLND module for the pervious areas, and the IMPLND module for the impervious areas. HSPF divides the moisture supply available to a land segment into five main components: the surface runoff, interflow, groundwater flow, and evapotranspiration and percolation to deep groundwater, which are both lost from the system. A number of hydrologic parameters are used by the model to compute the various aspects of each of these components. The default, maximum and minimum values of these parameters are specified in the User's Manual. Some of these parameters can be obtained directly from available watershed data (e.g. area, latitude, mean elevation, etc). However, most of these are calibration parameters, i.e., appropriate values can only be obtained by calibrating the model to stream flow measurements, since values may vary widely depending on the watershed characteristics.

The important calibration parameters in the PERLND and IMPLND subroutines used for producing desired shapes of hydrographs, peak flows, etc. are:

  • LZSN: lower zone nominal storage
  • UZSN: upper zone nominal storage
  • IRC: interflow recession parameters
  • INTFW: interflow inflow parameter

INTFW alters the peak and shape of the hydrograph, but does not alter runoff volumes, since this parameter is used to determine what fraction of the moisture infiltrates, and what goes into interflow.

IRC mainly affects the peak and the shape of the hydrograph, and is mainly used for finer adjustments of the recession limbs of hydrographs from large watersheds where storm events extend over a period of days.

UZSN and the LZSN will affect the affect the runoff volumes, since they determine what fraction of the moisture becomes available for direct runoff.

The baseflow calibration is achieved with the help of the following parameters:

  • DEEPFR: fraction of groundwater inflow, which enters deep groundwater and is lost from the system
  • BASETP: fraction of potential evapo-transpiration, which can be satisfied from baseflow if enough is available
  • AGWETP: fraction of potential evapo-transpiration, which can be satisfied from active groundwater storage if enough is available.

Increasing DEEPFR, BASETP, AGWETP decreases the baseflow, since these represent water that is lost from the groundwater storage.

Top of page   Top of page


Use in MTO Drainage Management Practice

Conditions for Use in MTO Projects or Projects Requiring MTO Approval

For the design of bridges/culverts or use in watershed/subwatershed studies for areas with Time of Concentration greater than 1 hour:

  1. Design flows must be derived from a frequency analysis of continuous simulated flows based on at least 25 years of hourly meteorological input data (rainfall, temperature, radiation, wind speed, etc.). This requirement is consistent with the length of flow record required to complete an accurate single station flood frequency analysis to estimate the 100 year event. For further details on the frequency analysis, reference should be made to "Hydrology of Floods in Canada: A Guide to Planning and Design," W.E. Watt, Editor-in-Chief, Associate Committee on Hydrology, NRC, 1989.
  2. The model must be calibrated with at least 5 years of data and verified with an additional 5 years of data from a stream flow gauge located on the same watershed or an adjacent watershed with similar characteristics. This is a generally accepted minimum requirement detailed in "Technical Guidelines: Flood Plain Management in Ontario," Section E2.4, MNR, 1985.

For the design of bridges/culverts or use in subwatershed/master drainage studies for drainage areas with Time of Concentration less than 1 hour:

  1. The design flows must be derived from a frequency analysis of flows simulated with HSPF based on at least 25 years of meteorological input data (rainfall, temperature, radiation, wind speed, etc.) with a time step equal to or less than the Time of Concentration of the catchment. This would typically be from 5 to 15 minutes.
  2. In the absence of suitable continuous meteorological input data with an the appropriate time step, it is possible to use HSPF in a single event simulation mode using "design events" provided that the soil moisture conditions would yield runoff volumes that are consistent with other accepted hydrologic design methods. This can be accomplished by applying the SCS storage equation to calculate soil-moisture conditions for the specific land use for a specific catchment area.
  3. If HSPF is used to calculate flows for a design storm, it must be demonstrated that the resulting peak flows are consistent with those derived from another method acceptable to MTO (e.g. OTTHYMO model, MTO Watershed Classification method, etc.).
  4. HSPF must be calibrated and validated with appropriate meteorological data from the same watershed or an adjacent watershed with similar characteristics. The data must be 5 years of continuous data for calibration and an additional 5 years of data for validation in both continuous simulation, and in single event design applications. These are generally accepted minimum requirements detailed in "Technical Guidelines: Flood Plain Management in Ontario," Section E2.4, MNR, 1985.

For the design of stormwater management facilities (generally for areas with Time of Concentration less than 1 hour):

  1. If the conditions indicated under Conditions 2 above have been satisfied, the HSPF model must be used to calculate storages and discharges for quantity control facilities designed to match post-development flows to pre-development flows derived using HSPF.
  2. If pre-development target flows calculated with HSP-F have been provided from a watershed plan, then HSP-F can be used to calculate storages and discharges for flood control facilities provided that a short time step data input is used in the post development HSP-F simulations.

Minimum Requirements for Inclusion in Report

This section provides some guidance to designers, analysts and reviewers on the type of information and minimum requirements for reporting the results of an HSPF analysis. This by no means should be considered a comprehensive list of the information included in a drainage report.

Summary of input parameters

A summary of input parameters needs to be provided for each subwatershed considered. A summary of which computational methods were used for simulation should be included.

Tabular and Graphical Output

The report should include both tabular and graphical output showing the calibration/verification results. Similarly, Tabular and graphical summaries of design flows derived from the models should be provided.

Top of page   Top of page

Back to Top