Detailed continuous collection system modeling today is commonly done using sophisticated models such as USEPA’s SWMM, DHI’s MOUSE and Wallingford Software’s InfoWorks. These models can be cumbersome for planning-level exercises however. Over the years, some modelers have continued to use STORM (Storage, Treatment, Overflow, and Runoff Model), a model developed by the US Army Corps of Engineers’ Hydrologic Engineering Center (HEC) in the 1970s. While SWMM has become a widely accepted standard and has been frequently updated, development of STORM ended after HEC ceased support in the early 1980s. STORM still maintains advantages over the more advanced models in certain situations, however. It was designed for long-term simulation while the other models were originally intended for single event analysis and are still more commonly used for that purpose. Long-term continuous simulation is particularly important in CSO studies, where key results of the analysis are overflow frequencies, typically of several times per year.

 

STORM's persistence testifies to the continuing utility of its approach: whereas an evaluation of five potential treatment plant pump rates in SWMM might require five individual simulations along with five post-processing passes through the output time series, HEC-STORM can evaluate the same scenarios with a single command line in its input file. STORM has many limitations relative to SWMM however: it cannot route flows, it does not compute overland or channel travel times, it cannot compute detailed hydraulics, and its runoff calculations are limited to Soil Conservation Service (now Natural Resources Conservation Service) curve numbers or the rational method.

 

NetSTORM adapts the central concepts of the original STORM program into a modern graphical user interface. It computes rainfall-runoff by applying the rational method at an hourly timestep, and computes overflow from a control structure based on a fixed treatment rate (T) and storage capacity (S):

 

Vt         = C i A Dt + St-1                (1)

Vtreat         = min (T Dt, Vt)                

St         = min (S, Vt - Vtreat)                

Vcso         = max (0, Vt – T Dt - St)        

 

where:

Vt          = inflow to structure during timestep t

C         = runoff coefficient 

i         = rainfall intensity (adjusted for initial abstraction)

A         = drainage area

Dt         = timestep (fixed at one hour)

St         = storage capacity used at time t

Vtreat          = “treated” outflow volume

Vcso          = untreated overflow volume

 

Using this methodology, the model can compute long-term planning level estimates of CSO volume or of capture efficiency for stormwater storage and pumping facilities. While the model’s accuracy on an event basis is less than that of a detailed hydrologic – hydraulic model, STORM can produce long-term statistics with great speed. This allows it to be very useful for purposes such as examining inter-annual trends in CSO, or identifying representative periods for long-term simulation using a more detailed model.

 

To compensate for the limitations of the rational method with regard to the non-linearity of runoff with rainfall intensity, NetSTORM incorporates a simple pervious area runoff algorithm. The pervious area fraction is computed as 1 – C. Runoff from pervious area (Q) is computed as:

 

Q = (1 – C) max (0, i – fc) A                

 

This approach yields results comparable with the Horton infiltration equation for cases where the Horton decay coefficient yields infiltration approaching the final infiltration rate (fc) within a moderate duration (e.g. k-1 < 1 hour). This reduces the number of model parameters from three (initial infiltration rate, decay coefficient, final infiltration rate) to one (final infiltration rate) without significantly reducing the resulting runoff for a long-term simulation. When pervious area runoff is included, equation (1) is re-written:

 

Vt = (C i + (1 – C) max (0, i – fc ) A Dt + St-1                        (6)