IMPROVED MOSQITO APPLICATION PROCEDURES
Report No FR0443

MARCH 1994

SUMMARY

I BENEFITS

Improved default values and applications procedures will increase the reliability of MOSQITO predictions and give users greater confidence in the use of the model. A better understanding of which data are the most important for model calibration and verification will make data collection and, hence, model application more efficient and cost-effective.

II OBJECTIVES

The MOSQITO model contains global input default parameter values to facilitate initial model building. These values have been estimated from data collected during the development of the model. As the model begins to have more widespread use it is appropriate to:

Review the sewer flow quality database used to develop MOSQITO and identify any deficiencies;

Produce improved default values for MOSQITO input and calibration variables incorporating recently collected data; and

Review and improve MOSQITO application procedures, including data collection, calibration and verification.

III REASON

MOSQITO is now a commercially available software product. Due to the differing nature of sewer catchments, some of the data required for calibration and verification of a MOSQITO model must be collected locally. However, some types of data cannot be readily collected. Also, MOSQITO may sometimes only be used with default values. The reliability of the default values in MOSQITO is, therefore, important. New data have become available to update the original default values and application procedures, as recommended in previous studies.

IV CONCLUSIONS

An assessment of all of the available data has been made and updated default values for MOSQITO have been produced. Sensitivity testing of the MOSQITO parameters has been carried out and the results used to develop updated procedures for the use of MOSQITO. Model calibration should seek to establish a reliable representation of sediment deposits and sediment erosion, transport and deposition, particularly under storm flow conditions. MOSQITO results are highly sensitive to three key parameters used to describe the behaviour of sediments. These are the sediment fraction settling velocity, pipe sediment effective shear stress and the efficiency factor used to represent sediment partitioning at CSOs with storage tanks. Further research and data collection are required to reinforce the reliability of current default values and user guidance associated with these parameters, which are inherently impractical to measure in typical model applications studies.

V RECOMMENDATIONS

It is recommended that the new default values and the new procedures for using MOSQITO are employed in future application studies. It is also recommended that these default values and procedures are regularly updated as more data become available from model application studies. More research is required to improve parameter estimates and user guidance in the use of settling velocity, shear strength and tank efficiency factors in calibration. A model user group should be employed to co-ordinate the feedback of user experience and identified software and user support deficiencies to the model developers.

VI RESUME

The work carried out to review and improve MOSQITO application procedures is described, including a review of the current database, a table of improved MOSQITO defaults values and the sensitivity studies carried out. The implications of these studies are assessed, identifying updated procedures for the use of MOSQITO for data collection and for the calibration and verification of a model. Finally, conclusions and recommendations from this study for the future use and development of MOSQITO are presented.

an opportunity should be sought to compare the results obtained using these values against further site specific data;

the default values should be regularly updated as new applications data become available.

Four problem areas with the derivation and use of the default parameters have been identified.

  1. Settling velocity and shear stress are impractical to measure in the course of applications studies. Other sediment parameter estimates have different associated levels of measurement error. It is recommended that, in model calibration, the greatest emphasis is placed on the use of the most reliably measured parameters.
  2. The definition of sediment characteristic default values by users should ensure that the resulting sediment description is volumetrically correct.
  3. The addition of foul flow sediment attached pollutants to foul flow event mean concentrations represents a conceptual flaw in the model which will result in an overestimation of foul flow pollutant loads. This potential error has been reduced by the new, lower sediment potency factors. It is recommended that the model is amended to rectify this problem.
  4. The results for bacterial parameter event mean concentrations are highly variable between sites and between events at individual sites. Hence, it is recommended that the global default values should be used with caution and replaced with local estimates where possible.

Copies of the Report are available from FWR, price £25.00 less 20% to FWR Members