Mystery of Jordan Hydrology
The Jordan Sanstone Formation is Cambrian in age and ranges from 80-100 feet in thickness in Minnesota. Hydrostratigraphically, it is comprised of two units. The lower unit is a fine clastic component that is anywhere from 5-50 feet in thickness. Hydraulically, it is often considered to be a confining unit, as small and poorly connected pore spaced inhibit substantial flow through this unit. Hydraulic conductivities for the lower section range from less than 1 to 35 ft/d, with a mean of 17.4 ft/d. Significant cavities are rare in this unit, but they do occur locally where fossils and carbonate-rich intraclasts have been dissolved away. Mesoscopic fractures can be observed, but they are generally no larger than 1mm wide. The upper unit is a coarse clastic sediment occurring in thicknesses of 20-80 feet. It is associated with much higher hydraulic conductivities, ranging from less than 1 to 93 ft/d, with a mean of 43.2 ft/d. This increased average hydraulic conductivity is attributed to significant fracture flow through this upper portion. Vertical flat joints can be observed in the uppermost 20 feet of the Jordan, and these conduits extend tens of vertical feet through the section and have inch-scale opennings. The Jordan is often considered to be an intergranular aquifer, where flow through pore space explains groundwater movement through the rock. Runkel and others (2003) suggest that "flow along fractures may actually be volumetrically dominant in certain settings." This complicates our understanding of groundwater movement through the Jordan because fracture flow is, at best, poorly understood and no broadly applicable models currently exist.*
Not the end of the road
It would require dedicated and lengthy investigation into the specific hydrogeology of SMC Watershed to accurately understand local groundwater flow and the influence that landuse and land cover change has had on recharge to the spring discharging out of the Jordan. However, using what we do know about the general hydraulic characteristics of the Jordan as an aquifer, we can hypothesize a few reasonable scenarios. For the sake of the following models, it is assumed that both hydrostratigraphic units of the Jordan are present. This assumption is valid because the Jordan can be up to 100 feet thick, and we know that Seven Mile Creek cuts through 210 vertical feet from the uplands down to the Minnesota River. No lower contact is observed at the bluffs in the park, which is below the spring, and Kasota, a dolostone representing an upper contact, is mined directly across the river but not present in the park. The major hydrologic units considered in the hypotheticals below are the Jordan Sandstone and the overlying glacial till.
The units are directly connected as one hydrogeologic unit.
This assumption implies that there is no confining unit separating the upper unit of glacial till with the lower sandstone. Nothing is said here of fracture flow versus intergranular flow, rather it is simply assumed that the Jordan Formation is a hydrologically homogeneous unit, expressing the same hydraulic conductivity throughout. In this case, we could conclude that land use and land cover changes have significantly impacted the spring. Tile drainage has substantially reduced recharge to the spring, and what little flow is present year-round respresents groundwater that has slipped between the network of tiling. Based on the findings of Runkel and others (2003), treating the Jordan as a hydraulically homogeneous unit is ill-advised, and we would expect the rate of percolation through the upper section of the formation to far exceed that of the lower.
The units are separated by a completely impermeable confining layer.
This assumption treats the lower portion of the glacial till as a confining unit. In this case, we would conclude that historical land use change has had no direct impact on recharge to the spring. However, that is not to say that there has been no impact at all. The upland region serves as a recharge zone for some aquifer, and certainly that unit would have felt a decline in recharge following the installation of tiling. Due to the poorly sorted nature of glacial till and associated low hydraulic conductivities, this is a reasonable assumption. It may be less reasonable to consider the till to be completely impermeable, but if seepage into the Jordan is barely measurable our model would still be accurate.
The units are separated by a confining unit, but they are hydraulically connected by fracture flow through vertical joints and stress-relief fractures.
Fracture flow would allow the upland to serve as a significant source of recharge to the underlying sandstone aquifer. In this model, groundwater percolating down from the upland recharge zone would exploit conduits of infinitely high hydraulic conductivity, formed as stress-relief fractures that penetrated the confining layer. Thus, we would conclude that land use change could have had a great impact on the flow of the spring. The severity of which would be dependent on by the proximity of drainage tiling to major fractures versus the proximity of wetlands and lakes. This model makes two reasonable assumptions. One, it assumes the till to be a confining unit, which is consistent with known hydraulic properties of some glacial tills. Two, it assumes significant vertical jointing penetrating both hydrologic units. There is evidence to support this assumption on two levels. Firstly, large flat vertical joints were observed by Runkel and others (2003) in the upper section of the Jordan. Secondly, stress-relief fractures are commonly found in bluff settings, where rock has room to expand into a vacated ravine.
The units are separated by a leaky confining unit.
This model assumes that there is some measurable recharge to the spring from the upland region. While a known confining unit exists between the till and sandstone, either small fractures or an increased permeability exist locally in the confining layer. In this case, the degree to which that layer is leaking would determine the historical impact of land use change. There could be a substantial range of contributions depending upon specific conditions, but it is important to note the validity of this model. No confining unit is completely impermeable, and that is even more true for a unit of glacial origin.
Closing Thoughts
In evaluating the above hypotheses, it would be wise to consider two more processes potentially at work in the hydrology of SMC Watershed. Sandstone commonly exhibits horizontal bedding plane fractures, which can create lenses of high hydraulic conductivity. In the Jordan, Runkel and others (2003) observed such fractures in deep bedrock settings in the fine clastic section. These fractures can serve as conduits linking regional, rather than local, recharge to spring flow. Bedding plane fractures are essential to this model, as discharge on the order of 1-3 cfs would likely not be achieved by intergranular flow alone. This assumes that recharge is not originating in the upland region and travelling stratigraphically downward, but originating outside the watershed of Seven Mile Creek and flowing laterally through one portion of the Jordan alone.
The second process deals with pressure heads in the aquifers underlying SMC Watershed. Even if there is substantial vertical fracturing or significant porous connectivity between the two units, local or regional heads may inhibit groundwater flow from the upland region directly down to the spring's aquifer. The spring water may be originating in an aquifer beneath the elevation of the spring, either deeper in the Jordan or in stratigraphically lower units, and it is forced upward by pressure.
*Runkel, and others. 2003, Hydrogeology of the Paleozoic Bedrock in Southeastern Minnesota: Minnesota Geological Survey Report of Investigations 61, 105p., 2 pls.