The class of problems at the lowest order of complexity, is the group of event, water-quantity applications. Of these problems, the more common are the sizing of conveyances, culverts, and floodways according to established design storm procedures, such as major-minor systems and zero-increase in runoff. Such first-order problems have in the past been based on approaches that model processes based only on the simple physics of wet weather flow. Earlier approaches were simple empirical curve fitting, such as the so-called rational formula and unit hydrograph methods (like SCS hydrology). Within these first-order problems there have arisen several common design problems of urban municipal water engineering that are not well suited to first-order approaches, such as sizing best management practices, e.g. storages or permeable pavement.   Underlying concepts and routines of first-order models did not include the recovery of storages and loss rates. Routines for water quality, erosion, scour and deposition are dependent in an important way on dry weather conditions,  routines that are an essential part of continuous modelling. Thus for the purposes of our discussion we could say that continuous water quality problems belong to a second level of complexity. (Similarly eco-systems models like WASP can be said to belong to a third order of complexity.)

Continuous modelling involves management of copious data. SWMM support modules (RAIN, TEMP, STATISTICS and COMBINE) relate to continuous models and to large and complex applications. Finally snowmelt is also an important part of continuous modelling in cold climates, although flooding due to snowmelt is usually thought to be a special application, often a complication additional to other problems.

Many textbooks start with a diagram such as Figure 1, schematizing dominant processes involved in real and modelled urban water systems. The underlying assumption is that there are statistically averaged fluxes and changes of state that can be estimated by the conservation equations of physics. Urbanization typically increases the amount and flux of stormwater runoff, pollutants and stressors entering urban receiving waters, and changes the energy fluxes. Major stressors found in receiving waters include: turbidity, elevated temperatures, increased stormwater flows, and its complement, reduced dry weather (or base) flows. Contaminants include oils and greases, heavy metals, nutrients, bacteria, suspended materials and organics. Impacts include loss of habitat and reduced bio-diversity.

In the past, with fewer folks and plenty of vibrant ecosystems, life was simpler. Urban stormwater drainage design provided rapid removal of stormwater from a development site to the nearest drain or storm sewer and watercourse. Impervious pavement was ideal. Consequences were entirely predictable: downstream flooding and pollution, and dry creeks. It gradually became clear that this situation could not be allowed to continue, because of detrimental environmental impacts, as well as the high economic and social costs of upgrading existing sewers to accept the increased flows (Beale, 1992).

Subsequent approaches to stormwater management therefore focused on attenuating surface runoff hydrographs, generally to provide detention and retention of excess stormwater runoff. However, this in turn brought water quality and environmental issues to the foreground, as shallow stormwater basins and receiving waters became eutrophic and excessively heated by direct sun. This brings us to the present situation: desirable aquatic ecosystems have been, and continue to be, seriously impacted, even destroyed, by traditional and currently popular practices of urban development.

The general design problem may be formulated as: find the optimum cost-effective array of BMPs to solve a list of external design problems. The problems may at the outset be formulated as a list of questions (James and Robinson, 1981 a, b), each with a question mark, termed design objectives herein, e.g.:

2.1 History of the world according to SWMM.

This necessarily brief review was written to help new users understand the peculiarities of SWMM code and manuals, and to plot a path through part of the literature. Development of the EPA Stormwater Management Model SWMM is described in "Introduction to the SWMM Environment", Chap. 1 in New Techniques for Modeling the Management of Stormwater Quality Impacts, 1993, pp. 1-28, published by Lewis Publishers. Parts of that paper are incorporated in the present discussion.

Most urban stormwater code has been written and distributed for commercial reasons. Development of SWMM, its ancestry and its continuing support, on the other hand, is probably unique. Apart from intermittent support of the USEPA, research groups at several different universities, including those of the author, the University of Florida, and Oregon State University, and engineers at agencies and in consulting offices, have spasmodically contributed ideas or more materially to the evolution of SWMM. In many ways, defining SWMM has become like nailing the proverbial jelly to a tree.

SWMM was unusual at the time - a major EPA funding effort devoted solely to hydrologic software development (another example was the conversion of HSP to HSPF, funded by EPA CEAM). From 1974 there was close collaboration with Canadians - notably the Ontario Ministry of the Environment, Environment Canada and several consultants (also an unusual if not unique effort). The following encapsulates some relevant dates:

1977 EXTRAN added by CDM. Documentation elaborated to include continuous simulation.
1981 Version 3 published by UoF. STORAG made generic, line IDs, metric capability, RAIN, TEMP and STATS blocks added, RECEIV block deleted, the first version of the present users manual and a separate EXTRAN manual issued.
1984 PCSWMM - first user-friendly personal computer version, distributed commercially with improved documentation by CHI.
1988 Version 4 (current major version) - USEPA public domain personal computer version. Free-format data entry, natural cross-sections added. Current (1998) users manual published. End of EPA support.
1993 PCSWMM ported to Windows environment
1993 SWMM News & Notes started by CHI.
1994 SWMM-USERS listserver started in January by W. James at U of Guelph (UoG).
1994 version 4.3 by EPA CEAM [Lahey Fortran with extended memory replaced 640kB RM Fortran as the standard.]
1997 version 4.4 by OSU and others.

2.1.1 Derivatives of SWMM

Thus the public domain program known as SWMM has been under more-or-less continuous development for about three decades, almost exclusively under Wayne Huber's leadership. The original manuals were also written by Huber, and produced occasionally in response to EPA contract agreements. But, under Reaganomics, USEPA support fell away over the years, and in the late 80s, it virtually dried up. In the early 1990s a number of derivative codes became available in the private domain, some of which favor a design-office environment (whereas PCSWMM is perhaps more aimed at a University environment, is web-oriented, and costs much less). PCSWMM uses standard SWMM code, whereas most others use highly modified code.

2.2 SWMM documentation

For each major issue of the documentation, the evolving code was bundled up and given a new version number. Software archaeologists should take care not to infer that the historic versions of the documentation reflect the then contemporary SWMM code; available documentation clearly lags code development by several years (twelve at the end of the millennium).

Starting in 1977, the author and his group ported the code to minicomputer and later to microcomputer systems. Free-format input, 80-column output, screen-oriented graphics, error-checking and transparent file handling was added. The documentation was rewritten in user-friendly style, the code stored on diskettes, and the package distributed by CHI as PCSWMM. Version 4 of SWMM appeared with published documentation in August 1988. It was an EPA response to PCSWMM which was a commercial product, and included most of the attributes of PCSWMM as well as a wide range of enhancements. The new EPA documentation reflected many of the contributions of PCSWMM, particularly the tables of input data requirements. It was configured for operation on hard disk, but retained much of the big-batch main frame architecture. It was now much better configured for continuous simulation and time-series management, incorporating a RAIN block and a TEMPERATURE block.

SWMM documentation is currently somewhat difficult to read, and has been repackaged by the present writer into two Guides:
1. Hydrology includes all service modules and RUNOFF;
2. Hydraulics comprises the other three modules (TRANS, EXTRAN and STORAGE).
PCSWMM provides a hyperlinked help system that acts as on-line documentation.

2.3 Serial Literature

The SWMM environment is a natural consequence of active participation in scientific, technical and engineering conferences, symposia, seminars, workshops and other meetings. Besides workshops and short courses given by (i) the University of Florida, (ii) the USEPA, and (iii) the writer, the more common meetings include irregular six-monthly user group conferences in the US and Canada. Recently, in the 1990s, meetings have been held annually in Toronto. A study of the lists of papers, will inevitably detect some repetition over the years. Fortunately, many consulting engineers, planners, geographers, aquatic biologists and related professionals concerned with urban development and its impact on aquatic environments have contributed to the rich literature. Graduate students have been an important feature.

Proceedings of these conferences are generally difficult to locate in libraries or elsewhere, and the author has made available (Biblio’98, also as part of PCSWMM98) abstracts of most of the papers. They cover many topics of interest to users of SWMM, and nicely encapsulate the changing emphasis over the past two decades. Topics range from concerns with water flows in a remote-batch-mainframe environment, to interdisciplinary ecosystems concerns in the evolving networked-workstation design environment.

2.4 SWMM capability

SWMM comprises four important service modules (none of which are needed for simple first-order applications) and four hydraulics/hydrology modules. Their relationship is shown in the next two Figures. We omit discussion of the EXECUTIVE module since knowledge of the EXEC module is unnecessary using PCSWMM.

SWMM may be run for an unlimited number of time steps, including very few. An assessment of urban runoff problems and estimates of the effectiveness of abatement procedures (such as different arrays of best management practices, BMPs) can be performed. Trade-offs among various control options, such as infiltration and storage may be evaluated. If desired, critical events from the long period of simulation may be selected for detailed analysis. There's no cost-benefit analysis.

A coarse schematization greatly reduces the complexity of the model (amount of entries required for subcatchments and channels). PCSWMM allows the use of virtually any local raingage data.

Transactions between the modules are handled by means of interface files, one of the four types of files used by SWMM as shown in the Figure. Output consists basically of computed hydrographs and pollutographs, presented over the whole event at a specified interval of time steps (e.g., every time step). This imposes a severe handicap on the modelling: no feedback is allowed between the modules.

Infiltration capacity is regenerated during dry periods by two methods, the better of which is based on the work of Green and Ampt. Monthly evaporation totals are used to regenerate depression storage on both pervious and impervious areas and are also considered an initial "loss" for each time step with rainfall. Computations are bypassed during dry periods if infiltration and depression storage regeneration is complete.

For all input datafiles, line identifiers are required, comment lines are allowed, lines may be 230 columns long and wrap around, and are written in free format. This allows the use of existing datafiles as templates (they are loaded up into an editor, numbers replaced, added or deleted as required to build your specific datafile). A generic example is given below:

It is sometimes difficult to tell in a session where PCSWMM ends and SWMM starts. For now, all that you need to know is that PCSWMM is a shell surrounding a SWMM engine which runs in a DOS window: