BY: Dr. John Kiefer, PE, PWS
Stream assessment and restoration approaches have advanced substantially during the last 20 years (NRCS 2007, Knighton 1998). While this has been especially true for streams with high sediment loads and perennial flow found in temperate climates, until recently, similar information was lacking for seasonally-wet warmlatitude streams including those of peninsular Florida. A few months ago, our research team from AMEC-BCI and the University of Florida (professors Joann Mossa and Bill Wise) completed an original three-year study of the natural streams of peninsular Florida aimed at providing much-needed tools necessary to facilitate better practical understanding and restoration of Florida’s streams (Kiefer 2010). AMEC-BCI’s scientists measured or derived more than 120 variables from 56 sites located between the Suwannee River and Lake Okeechobee. These streams drained watersheds ranging from less than 0.2 to greater than 300 square miles. This study was important because Florida’s combination of climate and physiography (geology, soils, vegetation, and topography) are unique to North America and stream corridors are formed and maintained largely by these factors.
First, we were curious to what extent existing approaches to classifying streams, if any, would apply best to Florida and quickly found substantial limitations to applying existing schemes. For example, classification systems based on channel shape, like the Rosgen classification system (Rosgen 1996), certainly have some merit in promoting understanding of streams in Florida. However, understanding channel shape alone provides an incomplete picture because fundamentally different erosion processes, some dominated by groundwater and some by surface scour, often lead to similarly shaped streams in Florida. This coincidental convergence of form is therefore not diagnostic of process. It is fundamentally important to understand processes and their geomorphic thresholds to make wise decisions concerning stream corridor conservation, management, and restoration.Processes are diffi cult to measure and the state has 1000’s of ungaged and unstudied streams. Therefore, we set out to derive a practical classification system targeted to Florida conditions that would be simple to apply based on factors that could be measured in a couple of days or less and that would have strong delineative associations with key processes and process thresholds.
We identified 15 different natural kinds of streams on the peninsula based on the distinguishing factors of their 120 variables related to their watershed size and soils, channel size and shape, floodplain size and shape, types of aquatic and riparian habitats, and wet season and dry season flow patterns. Peninsular Florida streams form in response to seasonal wet-dry cycles that interact with three categories of landscapes with very different runoff and infiltration capacities. Each landscape setting provides different combinations of dry season baseflow and wet-season flood pulses (Figure 1). The result of this is that the relative amount of flow reaching the stream via groundwater versus surface water pathways plays a fundamental role in the size and shape of the riparian corridor, its habitat structure and complexity, and the size and shape of the main stream channel in the corridor. In other words, streams belong to their watersheds.
BLACKWATER STREAMS
Systems with the least groundwater input and most surface water runoff occur in flatwoods landscapes. Flatwoods watersheds are dominated by NRCS Hydrologic Soil Group D soils that are poorly drained and limit groundwater infiltration while promoting substantial wetseason overland runoff. These kinds of watersheds produce dramatic fl uctuations in seasonal discharge and interannual flow patterns sensitive to rainfall. The resulting fluvial forms have high relative abundances of alluvial features within the channel, and if the watershed is at least several square miles in size, in their floodplains as well (Figure 2). Alluvial features are formed by sediment transport and deposition and they require periodic powerful flow spates to occur in floodplains. Such features include linear backswamps with silt and muck sediments, sandy or silty channel anabranches, natural sandy channel levees, and mucky oxbow lakes. Some examples of in-stream alluvial features include pools, sandy riffles, point bars at bends, mid-channel bars, and scour holes induced by large woody debris. Alluvial features create different habitat patches within the stream corridor playing a large role in the biodiversity of riparian systems.
Despite the fact that our classification was derived from numerous variables, we found that the stream types can be determined just using a handful of simple predictors that can be measured from publicly available information. For example, using our system, six types of flatwoods streams can be delineated by knowing just three variables; drainage area, longitudinal valley slope, and position in the drainage network (Figure 3). This means that even if a stream has been obliterated, and cannot be directly observed, its historic condition can be reasonably hindcast from a handful of remaining landscape variables. Based on a series of regressions developed from our data, once the stream type is known then its shape, channel dimensions, floodscape dimensions, types of alluvial features and associated distribution of in-stream and terrestrial habitats can be reasonably predicted. This is clearly a powerful and elegant set of tools to facilitate stream restoration designs in rural watersheds.
Most Florida streams occur in the flatwoods, but some of the most interesting streams drain landscapes that allow for substantially more groundwater interaction. These conditions tend to occur in catchments with at least 45% NRCS Hydrologic Soil Groups A and C. Such landscapes are strongly associated with xeric uplands on well-drained sandy “highlands.” These sandy uplands, formed from ancient dunescapes, allow for more rainfall infiltration into the surficial aquifer and the slower groundwater discharge to the channel reduces the overall stream flow variability versus that of the flashier flatwoods streams. The main effect is that the highlands streams have less well-developed and fewer alluvial features present in their floodplains than flatwoods streams draining similar size watersheds. In fact, it typically takes a 20 square mile watershed or larger in the highlands to generate suffi cient flood power necessary to develop floodplain alluvial features of the size and complexity that a flatwoods watershed draining only 5 square miles can produce. Alluvial floodplain features are formed in association with wet-season pulses that are more powerful and frequent in the flatwoods. These pulses depend heavily on the overall poor wet-season infiltration capacity of flatwoods watersheds. Even during the wet season, much more infiltration capacity exists in the highlands catchments which literally absorb much of the potential energy from wet season rains as they seep into the ground. This means less energy is routinely available to sculpt highlands floodplains.
Interestingly, flatwoods and highlands streams have very similar bankfull channel size and shape in association with drainage area. The similarities in bankfull channels probably occur because these channels are maintained by routine flows common in the flatwoods and highlands, typically occurring for about 25% of the year in perennial Florida streams. Evidently, these kinds of flow durations are typically sustained during the wet season irrespective of whether the pathway is via surface runoff, the surficial aquifer, or a combination thereof. Just like in the flatwoods, three kinds of highlands stream systems can be delineated based on a small number of variables; drainage basin area and valley slope. We have also developed regression-based design approaches to properly dimension and design highlands stream channels, their floodplains, and their key riparian habitat features.
In some respects highlands streams and flatwoods streams could be viewed as belonging to a super-class of blackwater streams that vary based largely on differences along a continuum in their wet-season hydrology, watershed size, valley slope and the location and types of their associated alluvial features. These streams are called blackwater systems because during the wet season their water is heavily and darkly stained by organic acids from mucky wetlands and sandy-organic upland soils. Simply classifying all 10 blackwater systems based on their water quality color would miss an opportunity to describe natural kinds of stream systems that differ signifi cantly in many important physical habitat features that depend on thresholds in fluvial forces.
SPRING RUNS
Florida’s crystal clear artesian spring runs are in a class by themselves (Figure 4). Spring runs fed by Florida’s immense karst aquifers are dominated by groundwater discharge and exhibit some of the most constant flow regimes found in nature (Figure 1). Runs with at least 70% of their flow coming from the aquifer not only lack the alluvial floodplain features found in the flatwoods streams, they also exhibit fundamentally different bankfull channel processes when compared to the blackwater channels. Florida spring runs tend to be straighter and wider than blackwater streams with similar bankfull discharge, but this pattern is not ubiquitous. The bed sediments of many of Florida’s spring runs differ substantially from those of flatwoods and highlands streams. Blackwater channel bed materials are typically dominated by fi ne to medium sand, much of which is carried to the stream from its watershed. While sand occurs on most spring run beds, the dominant material often consists of biologically derived sediments. This material consists of fi ne, fl uffy cohesive organics with bits of detritus and sometimes abundant snail shells (Figure 5). It is often layered over by a thin veneer of fi ne sand hiding the fact that most of the bed consists of soft organic material up to several feet thick in some cases. The organic flocc is sometimes limited to the channel margins and frequently supports dense meadows of submerged aquatic vegetation (SAV).
The flocc derives largely from biologically-mediated processes from within the stream itself including algal and plant detritus and calcareous snail shell debris. Essentially, this means that Florida spring runs are channels that course through bed materials of their own making. Compared to blackwater streams, they rarely produce alluviating flood flows and their discharge is typically well-contained within the bankfull channel. This qualified spring runs as perpetually-full gullies. The closer one looks at these spectacular groundwater dominated systems, the more unique their fluvial geomorphology becomes apparent. Five types of spring runs can be delineated simply by knowing their median flow. One critical threshold involves types we have referred to as High Magnitude and Medium Magnitude Spring Runs. High Magnitude Spring Runs generally occur at median flow greater than 15 cfs, which appears to be a threshold important to maintain channel widths that prevent complete tree canopy closure and allow enough light to reach the water surface to support at least patchy SAV meadows.
A NEW KIND OF STREAM
In addition to the spring runs, one of the other stream types we observed depends on a unique association between groundwater discharge and biological controls. This type of channel system has not been described elsewhere in scientifi c journals, which we have dubbed “Root- Step Channels.” This unusual fluvial form depends on subterranean erosion through thick sandy soils in the surficial aquifer, a process called groundwater sapping, which creates a narrow and steeply sloped valley with steady seepage flows. This combination of gentle groundwater hydrology and a gradual erosive process allows massive live roots to grow over the narrow stream bed forming living weirs anywhere from a few inches to 3 feet high and up to several feet wide (Figure 6). During the wet season, the enhanced stream flow cascades copiously over the rootweirs, forming a series of steps separating deeper pools in between. The banks of root-step channels are also biological in origin, often comprised of thick moss collars, dense fibrousroot mats, and peat accumulations that have grown and formed over the original sandy hillslope, considerably narrowing the stream channel.
Most of the root weirs are formed by tree species that thrive in groundwater seepage systems, such as sweet bay and dahoon. Root-step channels are normally found only in highlands watersheds, generally less than a few square miles, with valley slopes typically at least 1% (which is quite steep for Florida’s sandy soils). Without their intense biological controls, these otherwise sandy valleys would erode more deeply, but this process is arrested because the channel is enveloped and stabilized by living biological banks and root structures. Only minor sandy alluvial features exist between the root weirs and these systems have very high flow resistance factors (Manning’s n values of 0.25 for any engineers reading). The plants providing this stabilization benefi t from it because it preserves the shallow seepage conditions under which they thrive.
SUMMARY
Our classification system is hierarchical, starting with watershed characteristics that determine the total volumes of discharge and proportion of that discharge that reaches the stream as baseflow versus routine wet-season flood pulses. That hierarchy is critical because the differences in hydropattern and source of water to the stream greatly affect the channel and floodplain size and complexity. The physiographic setting can be readily assigned simply by knowing the NRCS hydrologic soil groups which are derived from widely available soils maps.
Once the physiographic region is assigned (flatwoods, highlands, karst), the next hierarchy of scale necessary to delineate the stream type is its watershed size and longitudinal valley slope. Stream power is a function of the change in elevation times the flow volume. Watersheds deliver the flow volume and the change in elevation is essentially the valley slope. Therefore, valley slope and drainage basin size provide easily measured surrogates for processes that are associated with fluvial forms which are highly dependent on stream power thresholds. Those forms include essentially all of the aforementioned alluvial habitats that provide each stream type with its own suite of biological components. A lack of high stream power levels, which occurs in the streams dominated by groundwater flow, allows biological processes to control key aspects of fluvial form in ways that cannot occur in higherenergy runoff-dominated flow regimes where physics exert more direct infl uence on channel shape and dimension. In simple terms, Florida’s warmth, wet climate, mild winters, and generally low-gradient landscapes provide a special setting for strong biological controls to exist and these controls can dominate aspects of fluvial form especially in portions of the landscape that allow for the hydropattern to be dominated by steady groundwater discharge lacking intense wet-season spates. Where the wet season pulses are strong, the physics of sediment transport takes on a much bigger role in the formation of riparian habitats.
Most Florida streams are not completely dominated by groundwater flow regimes, and under these circumstances, when valley slope and drainage area are looked at in concert the streams sort graphically along a classification gradient associated with stream power. Most stream types within a physiographic region can be delineated based on knowledge of just two or three easily measurable variables which can be accurately calculated from publically-available topographic and soils maps. However, for very small streams it is prudent to survey a small section of the valley to verify longitudinal slope.
In summary, our team has successfully achieved a good deal of practical solutions for defi ning Florida streams in a manner that is expected to increase understanding of their fluvial functions and the processes that maintain them. Our classification is not merely descriptive, but is sufficiently quantitative to properly identify key threshold associations among hydropatterns, basin scale, valley shape and scale, channel dimension, channel shape, and the abundance and types of in-stream and floodscape habitats. In other words, we have developed a set of empirical tools to greatly aid in the assessment and design of stream restoration projects and conservation management planning.
It has been a distinct pleasure and a humbling experience to have worked in many beautiful streams across the state the past few years. It is time now to roll up our sleeves and start restoring! This is important, because when we were trying to find relatively undisturbed streams to study, we ended up rejecting 75% of the sites that came up in our randomized selection process because they were clearly altered in some fashion or were obliterated by ditches.
The results of our study describing Florida’s natural streams and offering guidance concerning their restoration will be published by the FIPR Institute later this year. In the meantime, we hope this brief summary has raised some interest and you are encouraged to call me with any questions.
REFERENCES
Kiefer, J.H. 2010. Hydrobiogeomorphology of Fluvial Systems in Peninsular Florida: Implications to Classifi cation, Conservation, and Restoration. Doctoral Dissertation. University of Florida. 436 p.
Knighton, D. 1998. Fluvial Forms and Processes: A New Perspective. Arnold, London. 383 p.
NRCS. 2007. Stream Restoration Design. National Engineering Handbook, Part 654. Washington DC.
Rosgen,D. 1996. Applied River Morphology. Wildland Hydrology. Pagosa Springs, CO.
| < Prev | Next > |
|---|