hours along the transect at the mouth 

 of the Chesapeake Bay. This 36-hour 

 intensive survey was critical to es- 

 tablishing the previously undefined 

 boundary at the mouth of the bay both 

 for nutrients and hydrodynamics. 



The set of bay-wide nutrient 

 data collected during the July inten- 

 sive survey will be used to verify a 

 predictive water quality model of the 

 tidal Chesapeake Bay. In order to 

 make the model function properly, 

 model coefficients or rates must be 

 determined and then read into the 

 model's computer code. These model 

 values will be obtained from the 

 July data as well as from experiments 

 carried out this May and August by 

 bay research institutions. Examples 

 of these model rates are grazing 

 coefficient for zooplankton, benthic 

 dissolved oxygen demand, light ex- 

 tinction coefficient and ratios of N 

 and P to chlorophyll. The model 

 selected for application to the bay 

 is the model developed by Dr. H. S. 

 Chen of the Virginia Institute of 

 Marine Science. 



ANALYTICAL APPROACH 



WATER QUALITY MODELING 



The need for mathematical de- 

 scriptors of the processes which in- 

 teract to generate the trophic condi- 

 tion of the Chesapeake Bay system was 

 reflected in the CBP Eutrophication 

 Work Plan of late 1977 (Pheiffer et 

 al. 1977). The plan called for the 

 selection of water quality assessment 

 tools which would develop loads from 

 the tributary basin and "translate 

 material loads into eutrophication 

 levels." 



Many predictive models of storm 

 runoff pollution have been developed 

 and reported in the literature over 



the past decade. They range from 

 complex, computer-based models of 

 rainfall/washoff to a simple statis- 

 tical relationship between streamflow 

 (or runoff) and aerial pollutant 

 yield rates. Modelers generally 

 classify the former (complex) type as 

 deterministic models and the latter 

 (simple) as parametric models. The 

 trade-offs between these two general 

 classes of modeling approaches have 

 been described as an inverse rela- 

 tionship between the risk of not 

 representing the system versus the 

 difficulty in obtaining a solution 

 (Figure 4). In other words, the 

 level of effort involved with the 

 set-up, calibration, verification and 

 production utilization of a model 

 should be justified by the level of 

 significance required of the results, 

 the quality and extent of the cali- 

 bration/verification data base and, 

 above all, the availability of the 

 resources necessary to perform the 

 work (Smullen 1980). 



The level of modeling selected 

 for the non-tidal drainage basin of 

 Chesapeake Bay is referred to as HSPF 

 or Hydrological Simulation Program in 

 FORTRAN. This state-of-the-art mod- 

 eling package was developed by EPA 

 Environmental Research Laboratory at 

 Athens, Georgia. The HSPF model is a 

 continuous simulation model which 

 simulates the movement of water and 

 associated pollutants on land sur- 

 faces as well as the dispersionary 

 and flow characteristics of conserva- 

 tive and non-conservative constitu- 

 ents in branching stream systems and 

 rivers. Constituents modeled include 

 conservative minerals, temperature, 

 BOD, chlorophyll a, organic and or- 

 tho-phosphorus, ammonia, nitrate, ni- 

 trite, dissolved oxygen and coliform 

 bacteria. It also considers nutrient 

 cycles, zooplankton and algal growth. 



The work plan proposed to be 

 implemented under the EPA/Northern 



195 



