Executive Summary - Watershed Protection and Flood Mitigation Project Phase II- Technical Analysis

Prepared by: Center for Watershed Protection, Aquafor Beech LTD., Lori Barg, and Robert Kort

This report is the second in a multi-phase project being undertaken by the State of Vermont, Agency of Natural Resources (ANR). Phase I of the project involved conducting a literature search and providing discussion and assessment of the impacts of land use change on stream ecology and how levels of change to a stream's hydrology and morphology affect aquatic ecosystems.

To help quantify the relationships between stream geomorphology and land use activities for Vermont conditions and to provide a technical foundation for possible future guidance governing stormwater management runoff control for growing watersheds, ANR commissioned this study under Phase II, Technical Analysis of the project. It is anticipated that Phase III of the project will involve the development of a stormwater management guidance manual for the State of Vermont and Phase IV will involve training and education on the implementation of the guidance.

ANR's goal for this phase of the project was to determine, in Vermont, the type and size of watershed hydrologic and geomorphic impact that could result from various watershed land use activities including, watershed development in the natural floodplain, various levels of urbanization, and logging activities.

This report documents multiple lines of evidence used to assess the above goal. The study methodology incorporated several complimentary components to derive relationships between and among the watershed land use activities and stream system health. The study methodology incorporated the following analyses in descending order of significance:

• Validation of an empirical approach quantifying the relationship between total basin imperviousness and the enlargement of stream channel cross-sectional area.

• Computation of current stream channel stability using a rapid geomorphic assessment technique.

• Comparing previously collected stream channel biological monitoring results with total basin imperviousness and the results of the previous two assessments.

• Comparing stream channel riparian cover as a percent of total channel length.

A total of 8 subwatersheds were investigated as part of the study. Data were collected in the field at 24 separate stream sampling locations (approximately 3 sampling locations per subwatershed). Land use data were provided by the Vermont Center for Geographic Information (VCGI) at the University of Vermont in combination with aerial photography obtained from the Vermont Mapping Program. Biological monitoring data were provided to the project team by the Vermont Agency of Natural Resources, Biomonitoring and Aquatic Studies Section.

Background on the Scope of the Study:

The first component of the study was to validate the empirical relationship of channel enlargement (as measured by cross-sectional area) as a function of total watershed impervious cover. Past investigations have found that channel enlargement is a function of basin imperviousness as well as the corresponding age of that impervious cover. This relationship can be defined by the function:

(Re)_post=((A_BFL)_post/(A_BFL)_pre)

where, Re is defined as the channel enlargement ratio, 'A' represents the cross-sectional area of the stream channel and the subscripts BFL, POST, and PRE refer to the bankfull stage, the post-disturbance condition, and pre-disturbance condition, respectively.

The age of the development is also a critical variable in the amount of channel enlargement. In general, the longer a channel is exposed to the forces causing accelerated channel erosion, the larger the channel cross-sectional area. The effect of the age of development is represented by the concept of a "relaxation period." This is defined as the period of time required for a channel to reach an "equilibrium" state in concert with the level of watershed alteration, where the channel erosion processes are in a relative balance with the watershed forces causing erosion.

The results of past investigations for channel enlargement and channel relaxation show strong correlations with basin imperviousness. The equation derived from past investigations for alluvial type (AL-Type) streams for the ultimate channel enlargement ratio is defined as:

(Re)_ULT= 0.00135(TIMP)² + 0.0167(TIMP) + 1.0

R² = 0.78, (n=38)

where, (Re)_ULT is defined as the channel enlargement ratio once a stream is in equilibrium with its watershed hydrologic parameters, and TIMP is the total basin impervious cover, in percent. Note that the square of the correlation coefficient shows a very strong relationship between basin imperviousness and channel enlargement for the 38 sites investigated.

The hypothesis being tested in this part of the study was to evaluate the cross-sectional area to impervious cover relationship for eight Vermont watersheds and statistically compare the findings with those of previous investigations. If it could be shown that channel enlargement ratios for Vermont streams were drawn from the same population as channel enlargement ratios for non-Vermont streams then the existing relationships could be used to help predict and assess stream morphological impacts associated with different land use modifications.

The second component of the investigation utilized a rapid geomorphic assessment (RGA) technique to define the current stability of stream channels. The technique used a number of visually observed factors to provide a semi-quantitative assessment of a stream's current stability, referred to as the stability index (SI). The primary purpose of the RGA was to corroborate the findings of the more quantitative channel enlargement assessment and to help define past or current modes of channel adjustment (i.e., aggradation, degradation, widening and/or plan form adjustment). The RGA notes whether change in channel form has occurred or is still occurring, however, it does not provide a measure of the rate of change.

The third level of investigation involved the comparison of previously collected biological monitoring data with the corresponding level of impervious cover. The Vermont Agency of Natural Resources, Biomonitoring and Aquatic Studies Section and the Vermont Department of Fish and Wildlife provided the project team with macroinvertebrate and fish biological monitoring data covering a twelve year period (1986-1998). This analysis was intended to support the more quantitative geomorphological investigation of channel enlargement and channel stability and was not intended as a statistical evaluation of Vermont biological monitoring data.

The final element of the study involved comparing stream channel riparian cover length for each of the selected streams to assess whether or not riparian cover length was a factor in overall physical or biological condition. The methodology utilized aerial photography to estimate the extent of forest buffers in each subwatershed. The extent of the buffer was defined as the length of the forest buffer divided by the total stream length.

Methodology

The project team employed a ten step methodology to collect and analyze the data. As stated above, data were collected in eight Vermont subwatersheds. Table E.1 presents the basic project methodology.

Table E.1 Basic Project Methodology for the State of Vermont - Watershed Hydrology Protection and Flood Hazard Mitigation Project - Phase II, Technical Analysis

Step 1: Select a list of potential candidate subwatersheds representing a range of land use activities

Step 2: Compile historic data on candidate streams (cross-sectional data, biomonitoring, etc)

Step 3: Select a "short list" of streams with historical cross-sectional data, past biomonitoring data, and desired range of land use activities; conduct field screening of potential sites

Step 4: Select the final list of eight subwatersheds for field assessment

Step 5: Produce base mapping of selected stream reaches (land use/land cover mapping to compute total basin impervious cover (TIMP) and identification of stream location)

Step 6: Conduct field assessment of selected stream reaches (cross-sectional data and rapid geomorphic assessment at 24 cross-section locations -- 3 in each of the eight selected subwatersheds)

Step 7: Compile and analyze biomonitoring data for selected streams

Step 8: Conduct riparian buffer assessment of streams within the urbanized subwatersheds

Step 9: Conduct data analysis to define channel enlargement relationships, channel stability class, and stream bedload analysis

Step 10: Evaluate correlations between geomorphic parameters, biomonitoring and land use change as measured by TIMP

The first step was to select an initial candidate list of subwatersheds that met a range of land uses, had past biological monitoring data, and likely had historic stream cross-sectional surveys data (for estimating the pre-disturbance bankfull area, (A_BFl)_PRE). Next, a data collection effort was conducted to obtain past biomonitoring information, historic cross-sectional information, and current and past land use information. Candidate sites were then field reviewed to eliminate those where possible conflicts existed. The final selection of subwatersheds and streams involved input from the Project Steering Committee and included reference subwatersheds, subwatersheds with a range of urban/suburban development densities, a subwatershed where recent logging activity had occurred, and two subwatersheds where upland development was present. Table E.2 lists the final subwatersheds selected for data collection and assessment.

Table E.2 Final Subwatersheds Selected for Data Collection and Assessment
Stream Name	Town	Dominant land use	Impervious Cover1	Approx. Drainage Area (Sq. Mi.)
Cold River	Clarendon	Reference	<1%	20.7
Dowsville Brook	Duxbury	Logging	6%2	6.4
Moon/Tenney Brooks	Rutland	Urban	13 and 6% resp.	5.3 and 4.4
Potash Brook	S. Burlington	Urban	22%	7.4
Roaring Brook	Sherburne	Upland dev.	6%2	5.4
Smith	Goshen	Reference	<1%	3.2
Stevens Brook	St. Albans	Urban	13%	6.9
W. Branch Little River	Stowe	Upland dev.	2%2	24

¹ subwatershed impervious cover and drainage area at downstream most sampling location

² impervious cover estimate includes an "equivalent" impervious value

Subwatershed impervious cover was computed at each of the 24 stream sampling points. Impervious cover was derived using the VCGI's geographic information system and review of aerial photography. An "equivalent" impervious cover value was estimated for those land uses where the hydrologic alteration was not attributed to impervious cover (e.g, logging activities). In these cases, a runoff coefficient approach, based on Natural Resource Conservation Service Methods (NRCS), was used to derive the equivalent impervious value.Stream geomorphic data were collected in the field at 24 cross-section locations. The types of data collected at each station included, longitudinal channel slope, cross-sectional area, various measurements for channel depth and width, semi-quantitative assessments of channel stability using the RGA approach, stream substrate pebble data, and stream bank soil data. Stream data were analyzed using a series of spreadsheet models to calculate bankfull flowrate (Q_BFL), current cross-sectional area at bankfull stage, and Manning's roughness coefficient. Next, historical information of channel geometry (from older bridge construction plans, for example) and historical impervious cover estimates (from past aerial photography) were used to estimate the bankfull cross-sectional area for the historic channel [(A_BFL)_PRE]. The resulting ratio of current cross-sectional area to historic cross-sectional area (Re_i) was used to calculate an ultimate channel enlargement ratio (Re_ULT). These data were then compared to channel enlargement data from non-Vermont streams using statistical tests. The RGA data were used to compute the stability index for each stream.Biological monitoring data for macroinvertebrate and fish were assembled and evaluated as a function of subwatershed imperviousness. Biomonitoring data were presented for each stream and each sampling period. Only the overall biological "Community Assessments" for macroinvertebrates and fish are presented.Summary of Results:Channel Enlargement AssessmentTable E.3 lists a summary of the resulting data from the channel enlargement assessment for nine Vermont streams (note, Moon Brook and Tenney Brook are within the same subwatershed). The "observed" values were compared to "predicted" values derived from the non-Vermont Enlargement Curve to determine if they were drawn from the same population. Statistical tests for variance and mean were performed for these data and found to be statistically significant at the 95% confidence level.

Table E.3 Summary of Channel Enlargement Assessment
Basin	Site	Historic Channel Survey Data	Current Channel Survey Data	[(Re)ULT]OBS	(ABFL)PRE

ABFL (ft2)

ti (yrs)

TIMP (%)

(Re)i

ABFL (ft2)

ti (yrs)

TIMP (%)

(Re)i

(ft2)

Cold	CLD4	Reference Stream	201.2	46.7	2.0	Reference Stream
	CLD5	Reference Stream	52.2	80.5	1.0	Reference Stream
Cold (Gould)	GLD6	Reference Stream	110.3	80.5	1.0	Reference Stream
Dowsville	DOW1	Reference Stream	13.5	46.7	5.8	Reference Stream

	DOW2	60.5	19.5	1.0	1.00	105.5	23.4	5.8	1.04	1.91	60.5
	DOW3	55.2	19.5	1.0	1.00	51.1	23.4	5.8	1.04	1.01	55.2

Moon	MOO1	33.8	19.1	9.3	1.07	41.3	53.7	13.0	1.35	1.39	31.7
	MOO2	51.3	19.8	7.7	1.05	37.4	49.7	13.0	1.32	0.84	48.7
Tenney	TEN1	39.9	4.3	1.0	1.00	57.7	49.6	6.0	1.11	1.50	39.9
Potash	POT1	47.1	14.1	14.4	1.08	75.6	41.5	22.0	1.61	2.18	43.5
	POT2	48.5	14.1	14.4	1.08	63.6	41.5	22.0	1.61	1.78	44.8
	POT3	40.2	13.1	10.6	1.10	59.9	42.7	20.0	1.81	1.76	36.4
Roaring	ROA1	106.9	25.0	1.5	1.01	124.2	30.6	6.0	1.17	1.29	105.9
	ROA2	103.4	25.0	1.5	1.01	165.2	28.0	7.0	1.07	1.78	102.4

	RBT1	Reference Stream	28.6	46.7	2.0	Reference Stream
Smith	SMI1	Reference Stream	53.6	80.5	1.0	Reference Stream
	SMI2	Reference Stream	53.6	80.5	1.0	Reference Stream
	SMI3	Reference Stream	51.9	80.5	1.0	Reference Stream

Stevens	STB7	26.8	41.7	8.8	1.15	35.6	48.9	11.0	1.24	1.65	23.3
	STB8	28.6	40.2	8.3	1.13	30.4	48.9	11.0	1.24	1.30	25.3
	STB9	72.7	33.1	12.0	1.18	60.3	52.8	13.0	1.34	1.05	61.5
West Branch	WBL1	303.8	32.0	2.0	1.02	379.0	55.0	2.0	1.03	1.28	299.2
	WBL2	336.5	32.0	2.0	1.02	433.0	55.0	2.0	1.03	1.32	331.4
	WBL3	227.3	43.3	3.0	1.00	216.4	55.0	3.0	1.02	0.99	226.9

A_BFL= Bankfull channel cross-sectional area; t_I = area weighted average age of disturbance;

TIMP = Total Basin Imperviousness; (Re)_i = Enlargement Ratio at time t_I(i.e., current cross-section);

[(Re)_ULT]_OBS = Ultimate channel Enlargement Ratio, based on observed survey data;

(A_BFL)_PRE = Pre-disturbance channel bankfull channel cross-sectional area

The original channel enlargement curve for alluvial type streams was revised by integrating the Vermont data into the original database and undertaking a curve fitting process. The following second order polynomial provided the best fit for the data:

Revised Equation for Channel Enlargement Incorporating Vermont Data

(Re)_ULT = 0.0013(TIMP)² + 0.0168(TIMP) +1.0

(R² =0.83, n= 52)

Channel Stability Assessment

Results of the channel stability assessment are presented in Table E.4. The RGA process was originally developed for application in older urban watersheds that had been under riparian vegetation management programs and, consequently, largely denuded of wooded species. As such, metrics indicative of early geomorphic alteration were not incorporated into the original RGA Protocol. In consideration of the above, a modified RGA protocol was developed for Vermont to include the additional parameters: the number of Large Organic Debris pieces (NLOD) observed within the channel and riparian zone, the number of debris jams (NJAMS) and the number of complete riffle lines (NRIFF). The results are contained within the modified RGA data presented in Table E.4.

Table E.4 Summary of Channel Stability Assessment Using the Modified Rapid Geomorphic Assessment Form
Basin	Site	RGA FACTOR	Stability Index(1)	Stability Class	Channel Type

		AI	DI	WI	PI	(SI)
Cold	CLD4	0.14	0.20	0.14	0.13	0.15	Stable	AL(Ar)
	CLD5	0.14	0.00	0.00	0.38	0.13	Stable	AL(Ar)
Cold (Gould)	GLD6	0.14	0.20	0.29	0.13	0.19	Stable	AL(Ar)
Dowsville	DOW1	0.67	0.00	0.43	0.13	0.31	Transitional	AL(Ar)
	DOW2	0.14	0.00	0.71	0.38	0.31	Transitional	AL(Ar)
	DOW3	n/a	n/a	n/a	n/a	n/a	n/a	AL(Ar)
Moon	MOO1	0.67	0.40	0.88	0.63	0.64	In Adjustment	AL
	MOO2	0.71	0.00	0.86	0.63	0.55	In Adjustment	AL
Tenney	TEN1	0.33	0.17	0.63	0.63	0.44	In Adjustment	AL
Potash	POT1	0.57	0.20	0.86	0.50	0.53	In Adjustment	AL(Ar)
	POT2	0.33	0.60	0.83	0.43	0.55	In Adjustment	AL(Ar)
	POT3	0.60	0.00	1.00	0.60	0.55	In Adjustment	RB
Roaring	ROA1	0.20	0.00	0.83	0.17	0.30	Transitional	RB(Ar)
	ROA2	0.33	0.17	0.57	0.20	0.31	Transitional	AL(Ar)
	RBT1	0.14	0.00	0.71	0.33	0.30	Transitional	AL(Ar)
Smith	SMI1	0.17	0.20	0.29	0.00	0.16	Stable	AL(Ar)
	SMI2	0.00	0.00	0.38	0.00	0.09	Stable	AL(Ar)
	SMI3	0.00	0.20	0.33	0.00	0.13	Stable	AL(Ar)
Stevens	STB7	0.57	0.90	0.70	0.43	0.65	In Adjustment	AL(Ar)
	STB8	0.57	0.17	0.25	0.29	0.32	Transitional	AL
	STB9	0.14	0.17	0.50	0.29	0.27	Transitional	AL(Ar)
West Branch	WBL1	0.71	0.80	0.56	0.75	0.70	In Adjustment	AL
	WBL2	0.43	0.88	0.56	0.75	0.65	In Adjustment	AL
	WBL3	0.43	0.80	0.83	0.88	0.53	In Adjustment	AL(Ar)

(1) SI = Modified Stability Index for Vermont Conditions

AI = Aggradation Factor; DI = Degradation Factor;

WI = Widening Factor; PI = Planimetric Adjustment Factor;

n/a = not available; AL = Alluvial; Ar = Armored; RB = Rock Bed with alluvial banks;

The RGA protocol was applied to 23 sites surveyed in this study, with the exception of Site DOW3, A simple linear correlation analysis was undertaken relating the Stability Index to Total Basin Imperviousness (TIMP) for 20 of the 23 sites (W. Branch of Little River was excluded from the analysis because of past gravel mining operations) as follows:

SI = 0.158(TIMP)^0.413, R= 0.75, n = 20

The above relation was found to be statistically significant at the 95% confidence level for variance and mean.

Biological Monitoring Analysis

Table E.5 lists a generalized assessment of the biological monitoring data for the nine Vermont streams evaluated in this. The results suggest that these Vermont streams can be related to their contributing impervious cover and fall into one of two categories. The generally "good" streams, from a biological community assessment perspective, fall into an impervious cover range of 6% and less. The "poor" streams have impervious cover of 12% or greater.

Table E.5 Comparison of Biological Monitoring to Subwatershed Imperviousness
Stream Name	Subwatershed Current Impervious Cover (%)	Macro-invertebrate Bio-monitoring - Overall Community Assessment*	Fish Bio-monitoring - Overall Community Assessment*
Roaring Brook	6	Fair	Excellent
Stevens Brook	13	Poor	Poor - Fair
Dowsville Brook	6	Good - Excellent	Good
Potash Brook	22	Poor - Fair	Fair - Good
Tenney Brook	6	Fair - Good	Good - Excellent
Moon Brook	13	Poor	Fair
Smith Brook	<1	Excellent	-
Cold River	<1	Good	-
West Branch Little River	2	Good - Fair	Good

* represents an average of all biomonitoring presented in Table 4.1

 

Riparian Cover Analysis

The results of the riparian cover analysis are presented in Table E.6. Forest buffers were identified based on aerial photography for each watershed A simple methodology was used to estimate the extent of forest buffers in each subwatershed. The extent of the buffer was defined as the length of the forest buffer divided by the total stream length. The criteria used to determine the length of stream and buffer were:

• The stream length represents the total length of perennial streams based on USGS quad sheets.
• A forest buffer is defined as at least a 50' width of forest cover along the stream, with at least 20' of forest cover on each side of the stream.

Based on methodology performed, the results presented in Table E.6 yield no conclusive results to suggest that the extent of riparian cover has an undue influence on biological or physical stream quality. It should be noted that the assessment was conducted for only those streams with measurable development within a fairly modest range of impervious cover (~6 to 22%).

Table E.6 Forest Buffer Length as a Fraction of Total Stream Length
Stream	Section	Buffer Fraction
Moon Brook	Lower	30%
	Upper	35%
Tenney Brook	--	55%
Stevens Brook	Lower	35%
	Upper	20%
Potash Brook	Lower	20%
	Middle	20%
	Upper	25%
A forest buffer is defined as at least a 50' width of forest cover along the stream, with at least 20' of forest cover on each side of the stream.

Conclusions:

The methodology and data analyses support a suite of conclusions on the findings of this study. The project team identified the following six major conclusions as a result of our work on the geomorphological, and biological assessments:

• The key hypothesis of this study was to test whether stream geomorphological assessment techniques, that had been developed and tested in regions outside of Vermont, were valid for Vermont conditions. Specifically, two assessment techniques were evaluated: the Rapid Geomorphic Assessment technique that defines stream stability via a stability index value (SI) and the relationship of channel enlargement ratio [(Re)_ULT] to total basin imperviousness. The study results confirmed that both of these techniques could be applied with statistical significance to Vermont conditions.

An Enlargement Ratio equation and curve developed using stream geomorphological data from outside of Vermont was tested for inclusion with data from the Vermont streams investigated in this study and found to be statistically valid for the total population of data-points. This conclusion supports that there is now a statistically valid tool for Vermont conditions to help predict channel enlargement as a function of watershed imperviousness.

• The channel enlargement ratio [(Re)_ULT] for the nine Vermont streams was found to be somewhat related to total basin imperviousness (R² = 0.34). The overall channel enlargement equation and curve present a strong correlation between enlargement ratio and total basin imperviousness (R²= 0.83).
• The channel stability index (SI) conducted using the Rapid Geomorphic Assessment technique for the nine Vermont streams was also found to be strongly related to total basin imperviousness (R² = 0.78). The slightly lower correlation coefficient is not surprising given the qualitative nature of the data collection protocol for SI versus the more quantitative nature for (Re)_ULT data collection and analysis.
• The concept of "equivalent impervious cover," where land uses that alter the hydrologic characteristics of watershed cover without creating impervious cover (e.g., logging and upland development land uses) are equated to an equivalent amount of imperviousness, was found to be a meaningful measure. The resulting channel enlargement and stability index in subwatersheds where this method was employed did not deviate significantly from those subwatersheds where conventional imperviousness was the indicator of hydrologic change.
• The assessment of biological community health, relying on Vermont biomonitoring data, showed a general relationship of decreasing biological community health with increasing watershed impervious cover. However, since no statistical tests were conducted, the strength of this conclusion should be weighed against the more rigorous statistical tests that were performed for channel enlargement and channel stability class.
• The methodology used to perform the analysis of the possible benefits of riparian cover on stream biological or physical quality yielded inconclusive results. The possible benefits associated with adjacent wetlands, the level of detail associated with this portion of the study, and/or the comparison between streams with only a modest difference in impervious could have impacted the study findings.