USGS Water Resources Applications Software: OTIS
This page contains a set of Frequently Asked Questions (FAQ) for the OTIS software package.
Questions related to OTIS are divided into the following categories:False Convergence(or
Singular Convergence) — what do I do?
False Convergence(or
Singular Convergence). I noticed that the estimated dispersion coefficient is close to 0.01 — what do I do?
False Convergence(or
Singular Convergence) and none of the other entries in this FAQ apply to my situation — what do I do?
Installation files for the OTIS software may be downloaded from the Download section of the OTIS Web site. After obtaining the installation files, OTIS may be installed by following the instructions in Section 5.3 of the OTIS documentation.
The OTIS software is available for use, free-of-charge. Software and related materials (data and documentation) are made available by the U.S. Geological Survey (USGS) to be used in the public interest and the advancement of science. You may, without any fee or cost, use, copy, modify, or distribute this software, and any derivative works thereof, and its supporting documentation, subject to the USGS software User's Rights Notice.
The OTIS and OTIS-P executables are essentially "DOS-based" programs — as such, the Windows/Intel installation file available in the Download section of this web site includes executables that work under all flavors of Microsoft Windows (98, XP, Vista, 7, etc.). OTIS users running the 64-bit version of Windows 7 should follow the installation instructions available here.
Yes. OTIS users running the 32-bit version of Windows 7 can use the installation files that are available in the Download section of this Web site.
OTIS users running the 64-bit version of Windows 7 should install the software by performing the following tasks:
Note that OTIS users running the 64-bit version of Windows 7 should obtain and install the software as described in the bullets above (the installation files available in the Download section of this Web site will not work for the 64-bit version of Windows 7).
An executable version for Macintosh computers is not part of the official OTIS release. Executable binaries may be developed for the Mac by compiling the software as described in Section 5.4 of the OTIS documentation.
OTIS users running Linux or Windows on personal computers or Solaris on Sun Workstations can use the pre-compiled executables that are available in the Download section of this Web site; compilation for these operating systems/hardware platforms is not required.
Compilation is required in the following situations:
Yes, OTIS users frequently have applications in which the maximum dimensions of the problem under study exceed the maximum dimensions of the pre-compiled executables (see Section 5.5 of the documentation). An alternate version of OTIS with expanded maximum dimensions is therefore provided for Windows users. The maximum dimensions of the alternate version are set as follows: MAXREACH=1000, MAXPRINT=1000, MAXBOUND=20,000, MAXSEG=50,000, and MAXOBS=20,000.
Windows users who wish to use this alternate version should install the software by performing the following tasks:
OTIS relies on flat ASCII (text) files for input and output. At present, a graphical user interface for comprehensive management of OTIS input and output is not available. OTIS input files have traditionally been developed on the user's local computer system using a text-based editor. As an alternative, users may fill out the web-based forms available via the Generate Input section of this website. After providing the required information, users can download the resultant input files and run OTIS in the usual manner. OTIS output files are usually read into a spreadsheet or graphics program for plotting and analysis.
The OTIS solute transport model consists of two
operational models. The first model, OTIS, solves the
governing equations described in Section 2.2 of the
documentation using the
user-specified model parameters, flow information, and
system configuration. A second model, OTIS-P, is a modified
version of OTIS that uses non-linear regression to
determine a set of parameter estimates that optimally
describe observed concentration data. OTIS performs
predictive, forward simulations based on the user-specified
model parameters; OTIS-P is used for inverse modeling
exercises wherein optimal parameter estimates are obtained
by fitting the governing equations to user-specified
concentration data. As described in the Key Concepts
section of this FAQ, model users should work with
OTIS before attempting to use OTIS-P.
A variety of publications describe the techniques used to estimate physical transport parameters from tracer data, including the publications listed in the Additional Information and Application sections of this homepage. Suggested reading materials include:
essential references for starting an application of the transient storage concept.
Model users who wish to use their data to estimate
physical transport parameters should start with
OTIS before attempting to use OTIS-P. There are
two reasons for this recommendation. First, use of OTIS
allows you to get familiar with the basic model structure
and develop/debug your input files. The parameter and flow
files you develop as you use OTIS can then be used when you
move on to OTIS-P. Second, OTIS-P develops optimal
parameter estimates by comparing simulation results with
observed data. If your initial parameter estimates produce
a simulation that does not resemble the observed data, the
software will terminate before obtaining the optimal
estimates. It is therefore best to do a series of model
runs with OTIS so that the simulation is in the
ballpark
with respect to the observed data. The main
channel cross-sectional area, for example, may be adjusted
so that the timing of the simulated tracer curve
corresponds to that of the observed data (the main channel
cross-sectional area determines the advective velocity and
thus the timing of the simulation). Once there is a
reasonable match between the simulation and the data, one
can move on to OTIS-P.
The cross-sectional area of most streams and rivers varies spatially due to channel heterogeneities (for example, see Figure 3 of Bencala and Walters, 1983). Measurements of cross-sectional area obtained from surveys or stream gaging activities are therefore unlikely to provide accurate estimates of the average cross-sectional area for an entire stream reach. Accurate estimates of cross-sectional area may be obtained thorough the parameter estimation process, in which the governing transport equation is fit to observed tracer data (i.e. main channel cross-sectional area is adjusted so that the simulated traveltime corresponds to that of the tracer data). Measured values obtained from surveys or stream gaging provide initial estimates of main channel cross-sectional area; these estimates are then revised during the parameter estimation process.
Data from instantaneous slug injections can be used with OTIS to estimate physical transport parameters, provided enough data exists to define the upstream boundary condition. There are two ways to define the upstream boundary condition:
Use of #2 above is generally the preferred approach. The slug addition technique described in #1 above should always be used, however, so that an upstream boundary condition can be constructed in the event of sampling or laboratory problems associated with #2.
Note that there are theoretical reasons for using a continuous, constant rate injection rather than a slug injection. See:
Wagner, B.J., and Harvey, J.W., 1997, Experimental design for estimating parameters of rate-limited mass transfer: Analysis of stream tracer studies: Water Resources Research, v. 33 no. 7, p. 1731-1741, doi:10.1029/97WR01067
Despite these theoretical considerations, there are times when the slug injection method provides the best opportunity to collect a high quality data set (e.g. streams with high flow rates may not be amenable to continuous injections).
OTIS input files have traditionally been developed on the user's local computer system using a text-based editor. As an alternative, users may fill out the web-based forms available via the Generate Input section of this website. Even if you use the web-based forms, you may need to use a text editor to make small changes and fix mistakes. The following tips should be considered when using a text editor:
Din Tables 4-22 of the OTIS documentation) need to included a decimal point. For example, the PRTLOC variable should be entered as
50., not
50if you are requesting a print location at 50 m.
A downstream boundary condition is needed to solve the
equation that governs transport within the main channel. As
shown in Section 2.6 of the OTIS documentation, the
downstream boundary condition is set by specifying a fixed
dispersive flux at the downstream boundary, using the
user-specified value of DSBOUND. DSBOUND is typically set
to zero, such that the spatial concentration gradient
(∂C/∂x) is also zero. Because
concentrations within the study reach are likely to vary
spatially, specification of DSBOUND equal to zero may
result in erroneous simulation results if the downstream
boundary is placed too close to the locations of interest.
Model users therefore need to extend the modeled system
beyond the most downstream location of interest (e.g. the
most downstream sampling location) so that the error from
the downstream boundary condition is insignificant. There
are two ways to extend the modeled system: (1) make the
last reach of the modeled system extend beyond the last
print location, as shown in Application 1 of the OTIS
documentation (in Application 1, the modeled system is 669
m long and the last print location is at 619 meters); or
(2) include an extra dummy
reach. Note that the
length of the extended system is application dependent
— applications with higher dispersion coefficients
will generally need longer extensions. A general rule of
thumb for applications to streams and small rivers is to
add 100 segments to the modeled system, where each segment
is 1 m long. The accuracy of the model with respect to the
downstream boundary condition may be determined by
conducting a series of simulations, as shown in Section 4.1
of the OTEQ documentation (Runkel, 2010).
As described in Section 3.6 of the OTIS documentation,
there are two ways to execute OTIS and OTIS-P: (1) by
entering otis
or otis-p
in a DOS window
(known as the Command Prompt
in Windows 7); this
method requires knowledge of DOS commands such as cd
or (2) by double clicking on the OTIS and OTIS-P
executables within the graphical Windows environment;
this method is not recommended, however,
as run-time error messages displayed on the screen will not
be visible following program execution.
Always check the file echo.out
to make
sure the software is reading the input files
correctly. A common mistake is to omit the decimal
point when specifying print locations, resulting in
incorrect values being written to the solute output file
(e.g. a print location of 200
, might be read in as
2.0
). Also make sure your lateral inflow and outflow
variables (QLATIN, QLATOUT) are specified using the correct
units; these flow variables are in terms of flow
per distance (within reach change in flow,
divided by reach length).
There are a variety of methods for interpreting the results of transport-based tracer analyses. One useful metric is Fmed, the median traveltime attributable to transient storage. Fmed may be calculated using Equation 14 in Runkel, 2002.
Perhaps. See the OTIS documentation (Section 3.6.1) and the FAQ entry immediately below this one.
This error occurs when OTIS can not find the control, parameter, and/or flow input file that it is looking for. All of these files must reside in the directory (folder) from which you are executing the OTIS command. The control file must be named 'control.inp'; names for the remaining input files must correspond to the filenames specified in the control file.
Perhaps. See the OTIS documentation (Section 3.6.2) and the FAQ entry immediately below this one.
This error occurs when OTIS-P can not find the control, parameter, flow, data, and/or STARPAC input file that it is looking for. All of these files must reside in the directory (folder) from which you are executing the OTIS-P command. The control file must be named 'control.inp'; names for the remaining input files must correspond to the filenames specified in the control file.
False Convergence(or
Singular Convergence) — what do I do?
False and singular convergence may often be eliminated by doing additional OTIS-P simulations. These additional simulations use the final parameter estimates from the previous OTIS-P run as the initial estimates for the current run (parameter estimates are taken from the STARPAC output file and used as initial estimates in revised parameter and/or flow files). Conducting multiple OTIS-P runs in this manner is also advisable in the case of parameter or residual sum of squares convergence; OTIS-P runs are repeated until the final parameter values and residual sum of squares are invariant from one run to the next. An example of this procedure is presented in Application 5 of the OTIS documentation.
False Convergence(or
Singular Convergence). I noticed that the estimated dispersion coefficient is close to 0.01 — what do I do?
The version of OTIS-P distributed here will not allow for estimates of the dispersion coefficient that are less than 0.01 (most OTIS users employ length units of meters, so this minimum value corresponds to 0.01 m2/sec). This minimum value is generally not problematic, but smaller values of the dispersion coefficient may be needed to fit tracer data from systems with low advective velocities. If estimates of the dispersion coefficient are close to 0.01, an alternate version of OTIS-P may be needed for successful parameter estimation. This alternate version is available upon request.
False Convergence(or
Singular Convergence) and none of the other entries in this FAQ apply to my situation — what do I do?
High quality data sets allow for the simultaneous estimation of up to four physical transport parameters (main channel cross-sectional area, dispersion coefficient, storage zone cross-sectional area, storage zone exchange coefficient). Unfortunately many data sets are not of sufficient quality to estimate all four parameters. One solution to this problem is to fix one or more of the parameters based on previous experience and/or expert knowledge, and estimate the remaining parameters using OTIS-P.
U.S. Department of the Interior |
U.S. Geological Survey
URL: https://water.usgs.gov/software/OTIS/faq/index.html
Page Contact Information: Rob Runkel
Page Last Modified: Monday, 12-Dec-2016 16:33:06 EST
