Introduction

• Updating peak flow data for the year 2001 and comparing the quality of that data to that collected by the EPA sponsored USGS data collection. Comparing the flow hydrographs at the spring system with level data collected at the Cave Creek sinks and Fieldstone dam operation records.
• Comparing storm volumes discharged at the spring system. The purpose of this report is to provide a preliminary analysis of these data so that a determination can be made if a more detailed analysis is justified and to aid in the determination of what tracer testing might be performed.

Review of Peak Flow Attenuation

It has been previously reported (Long Term Groundwater Monitoring Plan for Neal's Landfill, QAPjP Volume XXXI, Bloomington Project, Viacom Inc., April 2002, Section 1.5.2.2.1) that peak flow reductions have occurred at the spring system that discharges at the head of Conard's Branch. Data collected by USGS in 1983-84 was compared to data collected by Viacom in 1993-94 and 2000. The relationship between peak flow and total rain, given similar pre-storm conditions, was similar between 1983-84 and 1993 storm sets. The 2000 storm set showed less than 50% of the peak flows reached for the same size rain events compared to the previous years (Figure 1).

The data has now been compiled for the 2001 set of storms. As shown on Figure 2, the relationship is very close to the 2000 storm set, and is still less than 50% of the peak flows of the previous years for the same size rain event. Reductions of this magnitude imply a major change in basin characteristics, especially since the overall trend in the area has been an increase in building and paving, which would tend to increase peak flows.

Although there is dye trace data to the contrary, it was speculated that the building of storm water retentions basins along Cave Creek, particularly the >70 ac-ft Fieldstone dam, may have attenuated peak flows in that watershed. The Cave Creek watershed was speculated to contribute to storm flows at Conard's Branch, at least during the rising limb of the Cave Creek hydrograph. Although the overflow sinks at Cave Creek were traced and the dye did not appear in Conard's Branch, the dye was injected at the end of the storm recession. The other major basin change was the expansion of Roger's Quarry south of Oard Road. If the area intercepted by the quarry had been contributing to Conard's Branch, then not only would peak flows be attenuated but the volume of discharge during storms would also be reduced.

Cave Creek Sink Water Levels

Beginning in late April 2002, Viacom began recording water levels at the Cave Creek sinks. The idea was to compare rising and falling water levels at the sinks with flows at Conard's Branch and the timing of the closing of the discharge gates at Fieldstone dam. However, according to the engineer that operates Fieldstone dam, the motors operating the discharge gates have burned out and one gate is now always left open a small amount to allow the pond to act as a detention basin.

Figures 3 and 4 show the Cave Creek sink water levels and the flows at the Conard's Branch weir for the exceptionally rainy period of May 6-13, 2002. The shaded areas on the figures are periods when flows at Conard's Branch are declining while water levels at Cave Creek sinks are still rising. Given that relationship, it is unlikely that the Cave Creek sinks are major contributors to storm flows at the Conard's Branch springs. This also makes it unlikely that the detention basins on Cave Creek have influenced peak flows at Conard's Branch.

Preliminary Storm Volume Comparisons

Table 1 shows the storm volume calculated for the storm sets in 1993-94, 2000, and 2001 years along with the total rain which produced that storm. The flows are summed from the beginning of the rising limb of the hydrograph to the point of intersection of linear and non-linear components on the receding limb of the hydrograph. This is the conventional point in hydrograph analysis which denotes the end of the storm (know as t j in the discussion in Karst Geomorpholonv and Hvdrolouv, Ford and Williams, 1989, pages 193-203).

Figures 5, 6, & 7 show the linear regression for flow volume vs. total rain for the three storm sets. It appears that the 2000 and 2001 storm sets have about 1/3 less volume per comparable rain than the 1993-94 storm set.

In order to check the validity of the flow data, at least for the recent storm sets, storms from the 2000 and 2001 years were selected where there was both USGS flow data from the combined weir set up at the overflow springs and well 5A peak elevation data available. Those storms are shown on Table 2.

The storms that had less than a 22% difference between the Viacom flow value and the USGS flow value were chosen as representative of the best peak flow measurements. The storm on October 4, 2000 was also included even though there was a large discrepancy with the USGS flow data. Personnel visited the site during this storm, confirming the accuracy of the Viacom measurements and noting significant bypass around the USGS weirs that account for the discrepancy. Those set of storms are shown in Table 3.

The value of peak flow was regressed against the peak elevation in well 5A, as shown on Figure 8. That regression showed a correlation of 94%. The relationship of peak flow in gpm = (655)(maximum 5A level) - 482302 may now be used as a check on the accuracy of peak flows recorded in Conard's Branch.

The regression equation was used to predict a peak flow based on the maximum well 5A elevation for the 2001 storm set that had a complete well 5A record. The percent difference between the measured peak flow and the peak flow predicted by maximum well 5A level is shown in the last column of Table 4. Those storms that had less than or equal to 20% difference were chosen as a revised 2001 storm set. Figure 9 shows that revised storm set flow volume regressed against total rain. The correlation was improved from 58% (Figure 7) to 62%.

Table 5 shows a comparison between the 1993-94 and revised 2001 storm set based on respective relationships between total rain and peak flow. Based on these data, actual flow volume reduction on the order of 1/3 has occurred in addition to peak flow attenuation.

The data sets from all three comparison years have a fair amount of scatter. Much of the data is clustered around storms that had a total rain of less than 1 inch during all three years. Each of the three years also had a few larger storms. When regressing data such as this, the larger storms will mostly determine the best fit line. With just a few large storms for each year of data, the conclusions drawn at this point are somewhat tentative.

The comparisons/conclusions may be improved by attempting to eliminate some of the data scatter. Utilizing a multiple regression approach with other parameters such as pre-storm flow may provide a firmer conclusion.

Conclusions and Recommendations

Based on what was seen on the relationship with Cave Creek sink levels, it seems unlikely that storm flow retention basins on Cave Creek are having much of an effect on spring flow at Conard's Branch. Factors that would remove water from the spring's sub-surface drainage basin must be considered. One such factor was southeast site drainage changes made during site remediation that re- routed water that was sinking in the former Cattail Pond and Southeast Pond areas to flow off as surface drainage into Southwest Seep Branch. Another factor could be the quarry expansion south of Oard Road.

Recommended actions include the following:

• The collection of available data from the Roger's Quarry should proceed as intended.
• A more detailed examination of storm discharge volume taking into account the amount of flow diverted into the treatment plant and antecedent moisture conditions should be performed.
• The logging of Cave Creek sink levels can be discontinued.
• A dye trace from the Cave Creek sinks is not indicated.
• Dye tracing to delineate the basin contributing to storm flows should be undertaken.