ISCO Auto-samplers were stationed at the mouths of Arcadia Creek and Portage Creek. They recorded creek level and rainfall on 10-minute intervals. They were also programmed to collect water from the creeks according to a predetermined schedule. Ten wet weather events were monitored at each creek mouth, and one hundred dry weather baseline samples were collected. These samples were anaylzed for Total Phosphorus (TP), Total Suspended Solids (TSS) and Soluble Reactive Phosphorus (SRP) (25% of the samples). These data represent pollutant loading from each watershed into the Kalamazoo River. The ISCO at the mouth of Portage Creek represents loading from the Portage Creek, Axtell Creek and West Fork of Portage Creek watersheds.

To compute an annual loading estimate, a dry weather baseline load and a wet weather annual load were summed for each pollutant for each creek. The dry weather baseline load was calculated by determining the average dry weather TP and TSS concentration (ug/L and mg/L, respectively) measured at each ISCO. Those average concentrations were multiplied by an average flow at each station (cubic feet per second, cfs), a time duration (300 days) and a conversion factor. It is assummed that Kalamazoo County experiences an average of 300 dry days per year.

Annual wet weather loading was calculated by determining a TP and TSS pollutant load for each measured event. Those loads were graphed against the rainfall depths for each event. Regression lines were fit to these relationships. A rainfall frequency distribution, based on a 100-year dataset measured at the Kalamazoo Psychiatric Hospital, and the calculated load vs. rainfall relationships were utilized to determine an annual load of each pollutant at each station.

Eight wet weather events measured at the Portage ISCO were used. Of the ten measured events, two were grouped as one event because they occurred within a close time span of one another; and one event was based on snow melt runoff and, therefore, had no correpsonding rainfall depth. Graphs of these events illustrate the responses of TP and TSS concentrations to creek flow. Five wet weather events measured at the Arcadia ISCO were used. Of the ten measured events, two were groups into one event due to hydrograph behavior and timing; one event was based on snow melt; one event was based on upstream pond dredging activities and, therefore, had no corresponding rainfall depth; and one event was incompletely monitored. The latter event was incompletely monitored because the weir constructed at the Arcadia Creek outfall broke during the event. Very high creek flows (over 200 cfs) were measured before the weir broke, illustrating the hydraulic behavior of an urbanized watershed. The graph of this event illustrates a sharp rise in flow measured before the weir broke. The estimated annual TP and TSS loads for each watershed are shown below.